Add Vitastor support

Makefile

@@ -24,7 +24,6 @@ endif
 
 PC_BIOS_FW_PURGE_LIST_IN = \
 	hppa-firmware.img \
-	hppa-firmware64.img \
 	openbios-ppc \
 	openbios-sparc32 \
 	openbios-sparc64 \
@@ -32,8 +31,7 @@ PC_BIOS_FW_PURGE_LIST_IN = \
 	s390-ccw.img \
 	s390-netboot.img \
 	u-boot.e500 \
-	.*[a-zA-Z0-9]\.dtb \
-	.*[a-zA-Z0-9]\.dts \
+	.*\.dtb \
 	qemu_vga.ndrv \
 	slof.bin \
 	opensbi-riscv.*-generic-fw_dynamic.bin \
@@ -58,7 +56,7 @@ $(BUILDDIR): submodule
 deb kvm: $(DEBS)
 $(DEB_DBG): $(DEB)
 $(DEB): $(BUILDDIR)
-	cd $(BUILDDIR); dpkg-buildpackage -b -us -uc
+	cd $(BUILDDIR); dpkg-buildpackage -b -us -uc -j32
 	lintian $(DEBS)
 
 sbuild: $(DSC)

debian/changelog

@@ -1,239 +1,8 @@
-pve-qemu-kvm (9.2.0-5) bookworm; urgency=medium
+pve-qemu-kvm (8.1.2-4+vitastor1) bookworm; urgency=medium
 
-  * pve backup: backup-access api: simplify bitmap logic
+  * Add Vitastor support
 
- -- Proxmox Support Team <support@proxmox.com>  Fri, 04 Apr 2025 16:15:58 +0200
-
-pve-qemu-kvm (9.2.0-4) bookworm; urgency=medium
-
-  * various async snapshot improvements, inclduing using a dedicated IO thread
-    for the state file when doing a live snapshot. That should reduce load on
-    the main thread and for it to get stuck on IO, i.e. same benefits as using
-    a dedicated IO thread for regular drives. This is particularly interesting
-    when the VM state storage is a network storage like NFS. It should also
-    address #6262.
-
-  * pve backup: implement basic features and API in preperation for external
-    backup provider storage plugins.
-
- -- Proxmox Support Team <support@proxmox.com>  Thu, 03 Apr 2025 17:00:34 +0200
-
-pve-qemu-kvm (9.2.0-3) bookworm; urgency=medium
-
-  * revert changes to the High Precision Event Timer (HPET) to fix performance
-    regression
-
- -- Proxmox Support Team <support@proxmox.com>  Wed, 26 Mar 2025 09:56:01 +0100
-
-pve-qemu-kvm (9.2.0-2) bookworm; urgency=medium
-
-  * fix assertion failure when migrating a VM with multiple disks on a
-    replicated ZFS.
-
- -- Proxmox Support Team <support@proxmox.com>  Mon, 24 Feb 2025 17:33:34 +0100
-
-pve-qemu-kvm (9.2.0-1) bookworm; urgency=medium
-
-  * update submodule and patches to QEMU 9.2.0
-
- -- Proxmox Support Team <support@proxmox.com>  Tue, 04 Feb 2025 08:49:20 +0100
-
-pve-qemu-kvm (9.1.2-3) bookworm; urgency=medium
-
-  * async snapshot: explicitly specify raw format when loading the VM state
-    file
-
-  * vma create: rework CLI parameters for passing disk to a more structured
-    style and use that to allow explicitly specifying the format
-
- -- Proxmox Support Team <support@proxmox.com>  Fri, 24 Jan 2025 16:12:34 +0100
-
-pve-qemu-kvm (9.1.2-2) bookworm; urgency=medium
-
-  * adapt machine version deprecation for Proxmox VE release and support
-    cycle.
-
- -- Proxmox Support Team <support@proxmox.com>  Fri, 17 Jan 2025 16:34:06 +0100
-
-pve-qemu-kvm (9.1.2-1) bookworm; urgency=medium
-
-  * update submodule and patches to QEMU 9.1.2
-
-  * improve error handling and edge cases with fleecing backups.
-
- -- Proxmox Support Team <support@proxmox.com>  Wed, 11 Dec 2024 16:47:21 +0100
-
-pve-qemu-kvm (9.0.2-4) bookworm; urgency=medium
-
-  * async snapshot: ensure any dynamic vCPU-throttling applied for
-    auto-converge gets always disabled again after finishing the snapshot.
-
- -- Proxmox Support Team <support@proxmox.com>  Sun, 10 Nov 2024 11:23:09 +0100
-
-pve-qemu-kvm (9.0.2-3) bookworm; urgency=medium
-
-  * pick up fix for VirtIO PCI regressions
-
-  * pick up stable fixes for 9.0, including fixes for VirtIO-net, ARM and
-    x86(_64) emulation, CVEs to harden NBD server against malicious clients,
-    as well as a few others (VNC, physmem, Intel IOMMU, ...).
-
- -- Proxmox Support Team <support@proxmox.com>  Fri, 06 Sep 2024 16:21:42 +0200
-
-pve-qemu-kvm (9.0.2-2) bookworm; urgency=medium
-
-  * actually update submodule to QEMU 9.0.2. The previous release was still
-    based on 9.0.0 by mistake.
-
- -- Proxmox Support Team <support@proxmox.com>  Wed, 07 Aug 2024 10:16:01 +0200
-
-pve-qemu-kvm (9.0.2-1) bookworm; urgency=medium
-
-  * update submodule and patches to QEMU 9.0.2. While our version had most
-    stable fixes included already, there are new fixes for VirtIO and VGA
-    display screen blanking (#4786)
-
-  * backport fix for a regression with the LSI-53c895a controller and one for
-    the boot order getting ignored for USB storage
-
- -- Proxmox Support Team <support@proxmox.com>  Mon, 29 Jul 2024 18:59:40 +0200
-
-pve-qemu-kvm (9.0.0-6) bookworm; urgency=medium
-
-  * fix a regression in the zeroinit block driver that prevented importing and
-    cloning disks to RBD storages which are not using the krbd setting
-
- -- Proxmox Support Team <support@proxmox.com>  Mon, 08 Jul 2024 16:11:15 +0200
-
-pve-qemu-kvm (9.0.0-5) bookworm; urgency=medium
-
-  * backport fix for CVE-2024-4467 to prevent malicious qcow2 image files from
-    already causing bad effects if being queried via 'qemu-img info'. For
-    Proxmox VE, this is an additional safe guard, as currently it directly
-    creates and manages the qcow2 images used by VMs and does not allow
-    unprivileged users to import them
-
-  * fix #4726: code cleanup: avoid superfluous check in vma backup code
-
- -- Proxmox Support Team <support@proxmox.com>  Wed, 03 Jul 2024 13:13:35 +0200
-
-pve-qemu-kvm (9.0.0-4) bookworm; urgency=medium
-
-  * fix crash after saving a snapshot without including VM state when a VirtIO
-    block device with iothread is configured.
-
-  * fix edge case in error handling when opening a block device from PBS fails
-
-  * minor code cleanup in backup code
-
- -- Proxmox Support Team <support@proxmox.com>  Mon, 01 Jul 2024 11:26:11 +0200
-
-pve-qemu-kvm (9.0.0-3) bookworm; urgency=medium
-
-  * fix crash when doing resize after hotplugging a disk using io_uring
-
-  * fix some minor issues in software CPU emulation (i.e. non-KVM) for ARM and
-    x86(_64)
-
- -- Proxmox Support Team <support@proxmox.com>  Wed, 29 May 2024 15:55:44 +0200
-
-pve-qemu-kvm (9.0.0-2) bookworm; urgency=medium
-
-  * fix #5409: backup: fix copy-before-write timeout
-
-  * backup: improve error when copy-before-write fails for fleecing
-
-  * fix forwards and backwards migration with VirtIO-GPU display
-
-  * fix a regression in pflash device introduced in 8.2
-
-  * revert a commit for VirtIO PCI devices that turned out to cause more
-    potential security issues than what it fixed
-
-  * move compatibility flags for a new VirtIO-net feature to the correct
-    machine type. The feature was introduced in QEMU 8.2, but the
-    compatibility flags got added to machine version 8.0 instead of 8.1. This
-    breaks backwards migration with machine version 8.1 from a 8.2/9.0 binary
-    to an 8.1 binary, in cases where the guest kernel enables the feature
-    (e.g. Ubuntu 23.10).
-    While that breaks migration with machine version 8.1 from an unpatched to
-    a patched binary, Proxmox VE only ever had 8.2 on the test repository and
-    9.0 not yet in any public repository.
-
- -- Proxmox Support Team <support@proxmox.com>  Fri, 17 May 2024 17:04:52 +0200
-
-pve-qemu-kvm (9.0.0-1) bookworm; urgency=medium
-
-  * update submodule and patches to QEMU 9.0.0
-
- -- Proxmox Support Team <support@proxmox.com>  Mon, 29 Apr 2024 10:51:37 +0200
-
-pve-qemu-kvm (8.2.2-1) bookworm; urgency=medium
-
-  * update submodule and patches to QEMU 8.2.2
-
- -- Proxmox Support Team <support@proxmox.com>  Sat, 27 Apr 2024 12:44:30 +0200
-
-pve-qemu-kvm (8.1.5-5) bookworm; urgency=medium
-
-  * implement support for backup fleecing
-
- -- Proxmox Support Team <support@proxmox.com>  Thu, 11 Apr 2024 17:46:48 +0200
-
-pve-qemu-kvm (8.1.5-4) bookworm; urgency=medium
-
-  * fix live-import for certain kinds of VMDK images that rely on padding
-
-  * backup: avoid bubbling up first error if it's an ECANCELED one, as those
-    are often a result of cancling the job due to running into an actual
-    issue.
-
-  * backup: factor out & clean up gathering device info into helper
-
- -- Proxmox Support Team <support@proxmox.com>  Tue, 12 Mar 2024 14:08:40 +0100
-
-pve-qemu-kvm (8.1.5-3) bookworm; urgency=medium
-
-  * backport fix for potential deadlock during QMP stop command if the VM has
-    disks attached through VirtIO-Block and IO-Thread enabled
-
-  * fix #4507: add patch to automatically increase NOFILE soft limit
-
- -- Proxmox Support Team <support@proxmox.com>  Wed, 21 Feb 2024 20:11:23 +0100
-
-pve-qemu-kvm (8.1.5-2) bookworm; urgency=medium
-
-  * work around for a situation where guest IO might get stuck, if the VM is
-    configure  with iothread and VirtIO block/SCSI
-
- -- Proxmox Support Team <support@proxmox.com>  Fri, 02 Feb 2024 19:41:27 +0100
-
-pve-qemu-kvm (8.1.5-1) bookworm; urgency=medium
-
-  * update to 8.1.5 stable release, including more relevant fixes like:
-    - virtio-net: correctly copy vnet header when flushing TX
-    - hw/pflash: implement update buffer for block writes
-    - Fixes to i386 emulation and ARM emulation.
-
- -- Proxmox Support Team <support@proxmox.com>  Fri, 02 Feb 2024 19:08:13 +0100
-
-pve-qemu-kvm (8.1.2-6) bookworm; urgency=medium
-
-  * revert attempted fix to avoid rare issue with stuck guest IO when using
-    iothread, because it caused a much more common issue with iothreads
-    consuming too much CPU
-
- -- Proxmox Support Team <support@proxmox.com>  Fri, 15 Dec 2023 14:22:06 +0100
-
-pve-qemu-kvm (8.1.2-5) bookworm; urgency=medium
-
-  * backport workaround for stuck guest IO with iothread and VirtIO block/SCSI
-    in some rare edge cases
-
-  * backport fix for potential deadlock when issuing the "resize" QMP command
-    for a disk that is using iothread
-
- -- Proxmox Support Team <support@proxmox.com>  Mon, 11 Dec 2023 16:58:27 +0100
+ -- Vitaliy Filippov <vitalif@yourcmc.ru>  Mon, 04 Dec 2023 01:40:09 +0300
 
 pve-qemu-kvm (8.1.2-4) bookworm; urgency=medium
 

debian/control

@@ -59,6 +59,7 @@ Depends: ceph-common (>= 0.48),
          libspice-server1 (>= 0.14.0~),
          libusb-1.0-0 (>= 1.0.17-1),
          libusbredirparser1 (>= 0.6-2),
+         vitastor-client (>= 0.9.4),
          libuuid1,
          ${misc:Depends},
          ${shlibs:Depends},

debian/patches/bitmap-mirror/0001-drive-mirror-add-support-for-sync-bitmap-mode-never.patch

@@ -27,18 +27,18 @@ Signed-off-by: Ma Haocong <mahaocong@didichuxing.com>
 Signed-off-by: John Snow <jsnow@redhat.com>
 Signed-off-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
 Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
-[FE: rebased for 9.1.2]
+[FE: rebased for 8.1.1]
 Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
 ---
- block/mirror.c                         | 99 ++++++++++++++++++++------
+ block/mirror.c                         | 98 +++++++++++++++++++++-----
  blockdev.c                             | 38 +++++++++-
  include/block/block_int-global-state.h |  4 +-
  qapi/block-core.json                   | 25 ++++++-
  tests/unit/test-block-iothread.c       |  4 +-
- 5 files changed, 142 insertions(+), 28 deletions(-)
+ 5 files changed, 142 insertions(+), 27 deletions(-)
 
 diff --git a/block/mirror.c b/block/mirror.c
-index 2afe700b4d..c3d4be9b15 100644
+index d3cacd1708..1ff42c8af1 100644
 --- a/block/mirror.c
 +++ b/block/mirror.c
 @@ -51,7 +51,7 @@ typedef struct MirrorBlockJob {
@@ -50,7 +50,7 @@ index 2afe700b4d..c3d4be9b15 100644
      BlockMirrorBackingMode backing_mode;
      /* Whether the target image requires explicit zero-initialization */
      bool zero_target;
-@@ -73,6 +73,8 @@ typedef struct MirrorBlockJob {
+@@ -65,6 +65,8 @@ typedef struct MirrorBlockJob {
      size_t buf_size;
      int64_t bdev_length;
      unsigned long *cow_bitmap;
@@ -59,9 +59,9 @@ index 2afe700b4d..c3d4be9b15 100644
      BdrvDirtyBitmap *dirty_bitmap;
      BdrvDirtyBitmapIter *dbi;
      uint8_t *buf;
-@@ -723,7 +725,8 @@ static int mirror_exit_common(Job *job)
+@@ -705,7 +707,8 @@ static int mirror_exit_common(Job *job)
+     bdrv_child_refresh_perms(mirror_top_bs, mirror_top_bs->backing,
                               &error_abort);
- 
      if (!abort && s->backing_mode == MIRROR_SOURCE_BACKING_CHAIN) {
 -        BlockDriverState *backing = s->is_none_mode ? src : s->base;
 +        BlockDriverState *backing;
@@ -69,7 +69,7 @@ index 2afe700b4d..c3d4be9b15 100644
          BlockDriverState *unfiltered_target = bdrv_skip_filters(target_bs);
  
          if (bdrv_cow_bs(unfiltered_target) != backing) {
-@@ -824,6 +827,16 @@ static void mirror_abort(Job *job)
+@@ -809,6 +812,16 @@ static void mirror_abort(Job *job)
      assert(ret == 0);
  }
  
@@ -86,7 +86,7 @@ index 2afe700b4d..c3d4be9b15 100644
  static void coroutine_fn mirror_throttle(MirrorBlockJob *s)
  {
      int64_t now = qemu_clock_get_ns(QEMU_CLOCK_REALTIME);
-@@ -1020,7 +1033,8 @@ static int coroutine_fn mirror_run(Job *job, Error **errp)
+@@ -997,7 +1010,8 @@ static int coroutine_fn mirror_run(Job *job, Error **errp)
      mirror_free_init(s);
  
      s->last_pause_ns = qemu_clock_get_ns(QEMU_CLOCK_REALTIME);
@@ -96,7 +96,7 @@ index 2afe700b4d..c3d4be9b15 100644
          ret = mirror_dirty_init(s);
          if (ret < 0 || job_is_cancelled(&s->common.job)) {
              goto immediate_exit;
-@@ -1309,6 +1323,7 @@ static const BlockJobDriver mirror_job_driver = {
+@@ -1251,6 +1265,7 @@ static const BlockJobDriver mirror_job_driver = {
          .run                    = mirror_run,
          .prepare                = mirror_prepare,
          .abort                  = mirror_abort,
@@ -104,7 +104,7 @@ index 2afe700b4d..c3d4be9b15 100644
          .pause                  = mirror_pause,
          .complete               = mirror_complete,
          .cancel                 = mirror_cancel,
-@@ -1327,6 +1342,7 @@ static const BlockJobDriver commit_active_job_driver = {
+@@ -1267,6 +1282,7 @@ static const BlockJobDriver commit_active_job_driver = {
          .run                    = mirror_run,
          .prepare                = mirror_prepare,
          .abort                  = mirror_abort,
@@ -112,7 +112,7 @@ index 2afe700b4d..c3d4be9b15 100644
          .pause                  = mirror_pause,
          .complete               = mirror_complete,
          .cancel                 = commit_active_cancel,
-@@ -1719,7 +1735,10 @@ static BlockJob *mirror_start_job(
+@@ -1658,7 +1674,10 @@ static BlockJob *mirror_start_job(
                               BlockCompletionFunc *cb,
                               void *opaque,
                               const BlockJobDriver *driver,
@@ -123,10 +123,10 @@ index 2afe700b4d..c3d4be9b15 100644
 +                             BlockDriverState *base,
                               bool auto_complete, const char *filter_node_name,
                               bool is_mirror, MirrorCopyMode copy_mode,
-                              bool base_ro,
-@@ -1734,10 +1753,39 @@ static BlockJob *mirror_start_job(
- 
-     GLOBAL_STATE_CODE();
+                              Error **errp)
+@@ -1670,10 +1689,39 @@ static BlockJob *mirror_start_job(
+     uint64_t target_perms, target_shared_perms;
+     int ret;
  
 -    if (granularity == 0) {
 -        granularity = bdrv_get_default_bitmap_granularity(target);
@@ -166,7 +166,7 @@ index 2afe700b4d..c3d4be9b15 100644
      assert(is_power_of_2(granularity));
  
      if (buf_size < 0) {
-@@ -1878,7 +1926,9 @@ static BlockJob *mirror_start_job(
+@@ -1804,7 +1852,9 @@ static BlockJob *mirror_start_job(
      s->replaces = g_strdup(replaces);
      s->on_source_error = on_source_error;
      s->on_target_error = on_target_error;
@@ -176,10 +176,10 @@ index 2afe700b4d..c3d4be9b15 100644
 +    s->bitmap_mode = bitmap_mode;
      s->backing_mode = backing_mode;
      s->zero_target = zero_target;
-     qatomic_set(&s->copy_mode, copy_mode);
-@@ -1904,6 +1954,18 @@ static BlockJob *mirror_start_job(
-      */
-     bdrv_disable_dirty_bitmap(s->dirty_bitmap);
+     s->copy_mode = copy_mode;
+@@ -1825,6 +1875,18 @@ static BlockJob *mirror_start_job(
+         bdrv_disable_dirty_bitmap(s->dirty_bitmap);
+     }
  
 +    if (s->sync_bitmap) {
 +        bdrv_dirty_bitmap_set_busy(s->sync_bitmap, true);
@@ -193,10 +193,10 @@ index 2afe700b4d..c3d4be9b15 100644
 +        }
 +    }
 +
-     bdrv_graph_wrlock();
      ret = block_job_add_bdrv(&s->common, "source", bs, 0,
                               BLK_PERM_WRITE_UNCHANGED | BLK_PERM_WRITE |
-@@ -1986,6 +2048,9 @@ fail:
+                              BLK_PERM_CONSISTENT_READ,
+@@ -1902,6 +1964,9 @@ fail:
          if (s->dirty_bitmap) {
              bdrv_release_dirty_bitmap(s->dirty_bitmap);
          }
@@ -206,7 +206,7 @@ index 2afe700b4d..c3d4be9b15 100644
          job_early_fail(&s->common.job);
      }
  
-@@ -2008,35 +2073,28 @@ void mirror_start(const char *job_id, BlockDriverState *bs,
+@@ -1919,31 +1984,25 @@ void mirror_start(const char *job_id, BlockDriverState *bs,
                    BlockDriverState *target, const char *replaces,
                    int creation_flags, int64_t speed,
                    uint32_t granularity, int64_t buf_size,
@@ -231,23 +231,19 @@ index 2afe700b4d..c3d4be9b15 100644
 -                   MirrorSyncMode_str(mode));
 -        return;
 -    }
--
-     bdrv_graph_rdlock_main_loop();
 -    is_none_mode = mode == MIRROR_SYNC_MODE_NONE;
      base = mode == MIRROR_SYNC_MODE_TOP ? bdrv_backing_chain_next(bs) : NULL;
-     bdrv_graph_rdunlock_main_loop();
- 
      mirror_start_job(job_id, bs, creation_flags, target, replaces,
                       speed, granularity, buf_size, backing_mode, zero_target,
                       on_source_error, on_target_error, unmap, NULL, NULL,
 -                     &mirror_job_driver, is_none_mode, base, false,
--                     filter_node_name, true, copy_mode, false, errp);
+-                     filter_node_name, true, copy_mode, errp);
 +                     &mirror_job_driver, mode, bitmap, bitmap_mode, base,
-+                     false, filter_node_name, true, copy_mode, false, errp);
++                     false, filter_node_name, true, copy_mode, errp);
  }
  
  BlockJob *commit_active_start(const char *job_id, BlockDriverState *bs,
-@@ -2063,7 +2121,8 @@ BlockJob *commit_active_start(const char *job_id, BlockDriverState *bs,
+@@ -1970,7 +2029,8 @@ BlockJob *commit_active_start(const char *job_id, BlockDriverState *bs,
                       job_id, bs, creation_flags, base, NULL, speed, 0, 0,
                       MIRROR_LEAVE_BACKING_CHAIN, false,
                       on_error, on_error, true, cb, opaque,
@@ -255,13 +251,13 @@ index 2afe700b4d..c3d4be9b15 100644
 +                     &commit_active_job_driver, MIRROR_SYNC_MODE_FULL,
 +                     NULL, 0, base, auto_complete,
                       filter_node_name, false, MIRROR_COPY_MODE_BACKGROUND,
-                      base_read_only, errp);
+                      errp);
      if (!job) {
 diff --git a/blockdev.c b/blockdev.c
-index 6740663fda..38fa63155c 100644
+index e6eba61484..a8b1fd2a73 100644
 --- a/blockdev.c
 +++ b/blockdev.c
-@@ -2781,6 +2781,9 @@ static void blockdev_mirror_common(const char *job_id, BlockDriverState *bs,
+@@ -2848,6 +2848,9 @@ static void blockdev_mirror_common(const char *job_id, BlockDriverState *bs,
                                     BlockDriverState *target,
                                     const char *replaces,
                                     enum MirrorSyncMode sync,
@@ -271,7 +267,7 @@ index 6740663fda..38fa63155c 100644
                                     BlockMirrorBackingMode backing_mode,
                                     bool zero_target,
                                     bool has_speed, int64_t speed,
-@@ -2799,6 +2802,7 @@ static void blockdev_mirror_common(const char *job_id, BlockDriverState *bs,
+@@ -2866,6 +2869,7 @@ static void blockdev_mirror_common(const char *job_id, BlockDriverState *bs,
  {
      BlockDriverState *unfiltered_bs;
      int job_flags = JOB_DEFAULT;
@@ -279,7 +275,7 @@ index 6740663fda..38fa63155c 100644
  
      GLOBAL_STATE_CODE();
      GRAPH_RDLOCK_GUARD_MAINLOOP();
-@@ -2853,6 +2857,29 @@ static void blockdev_mirror_common(const char *job_id, BlockDriverState *bs,
+@@ -2920,6 +2924,29 @@ static void blockdev_mirror_common(const char *job_id, BlockDriverState *bs,
          sync = MIRROR_SYNC_MODE_FULL;
      }
  
@@ -309,7 +305,7 @@ index 6740663fda..38fa63155c 100644
      if (!replaces) {
          /* We want to mirror from @bs, but keep implicit filters on top */
          unfiltered_bs = bdrv_skip_implicit_filters(bs);
-@@ -2894,8 +2921,8 @@ static void blockdev_mirror_common(const char *job_id, BlockDriverState *bs,
+@@ -2965,8 +2992,8 @@ static void blockdev_mirror_common(const char *job_id, BlockDriverState *bs,
       * and will allow to check whether the node still exist at mirror completion
       */
      mirror_start(job_id, bs, target,
@@ -320,7 +316,7 @@ index 6740663fda..38fa63155c 100644
                   on_source_error, on_target_error, unmap, filter_node_name,
                   copy_mode, errp);
  }
-@@ -3039,6 +3066,8 @@ void qmp_drive_mirror(DriveMirror *arg, Error **errp)
+@@ -3114,6 +3141,8 @@ void qmp_drive_mirror(DriveMirror *arg, Error **errp)
  
      blockdev_mirror_common(arg->job_id, bs, target_bs,
                             arg->replaces, arg->sync,
@@ -329,7 +325,7 @@ index 6740663fda..38fa63155c 100644
                             backing_mode, zero_target,
                             arg->has_speed, arg->speed,
                             arg->has_granularity, arg->granularity,
-@@ -3058,6 +3087,8 @@ void qmp_blockdev_mirror(const char *job_id,
+@@ -3135,6 +3164,8 @@ void qmp_blockdev_mirror(const char *job_id,
                           const char *device, const char *target,
                           const char *replaces,
                           MirrorSyncMode sync,
@@ -338,7 +334,7 @@ index 6740663fda..38fa63155c 100644
                           bool has_speed, int64_t speed,
                           bool has_granularity, uint32_t granularity,
                           bool has_buf_size, int64_t buf_size,
-@@ -3098,7 +3129,8 @@ void qmp_blockdev_mirror(const char *job_id,
+@@ -3183,7 +3214,8 @@ void qmp_blockdev_mirror(const char *job_id,
      }
  
      blockdev_mirror_common(job_id, bs, target_bs,
@@ -349,10 +345,10 @@ index 6740663fda..38fa63155c 100644
                             has_granularity, granularity,
                             has_buf_size, buf_size,
 diff --git a/include/block/block_int-global-state.h b/include/block/block_int-global-state.h
-index eb2d92a226..f0c642b194 100644
+index da5fb31089..32f0f9858a 100644
 --- a/include/block/block_int-global-state.h
 +++ b/include/block/block_int-global-state.h
-@@ -158,7 +158,9 @@ void mirror_start(const char *job_id, BlockDriverState *bs,
+@@ -152,7 +152,9 @@ void mirror_start(const char *job_id, BlockDriverState *bs,
                    BlockDriverState *target, const char *replaces,
                    int creation_flags, int64_t speed,
                    uint32_t granularity, int64_t buf_size,
@@ -364,10 +360,10 @@ index eb2d92a226..f0c642b194 100644
                    BlockdevOnError on_source_error,
                    BlockdevOnError on_target_error,
 diff --git a/qapi/block-core.json b/qapi/block-core.json
-index fd3bcc1c17..48ba32049f 100644
+index 2b1d493d6e..903392cb8f 100644
 --- a/qapi/block-core.json
 +++ b/qapi/block-core.json
-@@ -2178,6 +2178,15 @@
+@@ -2145,6 +2145,15 @@
  #     destination (all the disk, only the sectors allocated in the
  #     topmost image, or only new I/O).
  #
@@ -383,7 +379,7 @@ index fd3bcc1c17..48ba32049f 100644
  # @granularity: granularity of the dirty bitmap, default is 64K if the
  #     image format doesn't have clusters, 4K if the clusters are
  #     smaller than that, else the cluster size.  Must be a power of 2
-@@ -2220,7 +2229,9 @@
+@@ -2187,7 +2196,9 @@
  { 'struct': 'DriveMirror',
    'data': { '*job-id': 'str', 'device': 'str', 'target': 'str',
              '*format': 'str', '*node-name': 'str', '*replaces': 'str',
@@ -394,7 +390,7 @@ index fd3bcc1c17..48ba32049f 100644
              '*speed': 'int', '*granularity': 'uint32',
              '*buf-size': 'int', '*on-source-error': 'BlockdevOnError',
              '*on-target-error': 'BlockdevOnError',
-@@ -2499,6 +2510,15 @@
+@@ -2471,6 +2482,15 @@
  #     destination (all the disk, only the sectors allocated in the
  #     topmost image, or only new I/O).
  #
@@ -410,7 +406,7 @@ index fd3bcc1c17..48ba32049f 100644
  # @granularity: granularity of the dirty bitmap, default is 64K if the
  #     image format doesn't have clusters, 4K if the clusters are
  #     smaller than that, else the cluster size.  Must be a power of 2
-@@ -2547,7 +2567,8 @@
+@@ -2521,7 +2541,8 @@
  { 'command': 'blockdev-mirror',
    'data': { '*job-id': 'str', 'device': 'str', 'target': 'str',
              '*replaces': 'str',
@@ -421,10 +417,10 @@ index fd3bcc1c17..48ba32049f 100644
              '*buf-size': 'int', '*on-source-error': 'BlockdevOnError',
              '*on-target-error': 'BlockdevOnError',
 diff --git a/tests/unit/test-block-iothread.c b/tests/unit/test-block-iothread.c
-index 20ed54f570..4f50a99334 100644
+index d727a5fee8..8a34aa2328 100644
 --- a/tests/unit/test-block-iothread.c
 +++ b/tests/unit/test-block-iothread.c
-@@ -755,8 +755,8 @@ static void test_propagate_mirror(void)
+@@ -757,8 +757,8 @@ static void test_propagate_mirror(void)
  
      /* Start a mirror job */
      mirror_start("job0", src, target, NULL, JOB_DEFAULT, 0, 0, 0,
@@ -434,4 +430,4 @@ index 20ed54f570..4f50a99334 100644
 +                 false, BLOCKDEV_ON_ERROR_REPORT, BLOCKDEV_ON_ERROR_REPORT,
                   false, "filter_node", MIRROR_COPY_MODE_BACKGROUND,
                   &error_abort);
- 
+     WITH_JOB_LOCK_GUARD() {

debian/patches/bitmap-mirror/0002-drive-mirror-add-support-for-conditional-and-always-.patch

@@ -24,10 +24,10 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  1 file changed, 18 insertions(+), 6 deletions(-)
 
 diff --git a/block/mirror.c b/block/mirror.c
-index c3d4be9b15..7b6f7c0068 100644
+index 1ff42c8af1..11b8a8e959 100644
 --- a/block/mirror.c
 +++ b/block/mirror.c
-@@ -694,8 +694,6 @@ static int mirror_exit_common(Job *job)
+@@ -682,8 +682,6 @@ static int mirror_exit_common(Job *job)
          bdrv_unfreeze_backing_chain(mirror_top_bs, target_bs);
      }
  
@@ -36,9 +36,9 @@ index c3d4be9b15..7b6f7c0068 100644
      /* Make sure that the source BDS doesn't go away during bdrv_replace_node,
       * before we can call bdrv_drained_end */
      bdrv_ref(src);
-@@ -805,6 +803,18 @@ static int mirror_exit_common(Job *job)
-     bdrv_drained_end(target_bs);
-     bdrv_unref(target_bs);
+@@ -788,6 +786,18 @@ static int mirror_exit_common(Job *job)
+     block_job_remove_all_bdrv(bjob);
+     bdrv_replace_node(mirror_top_bs, mirror_top_bs->backing->bs, &error_abort);
  
 +    if (s->sync_bitmap) {
 +        if (s->bitmap_mode == BITMAP_SYNC_MODE_ALWAYS ||
@@ -55,7 +55,7 @@ index c3d4be9b15..7b6f7c0068 100644
      bs_opaque->job = NULL;
  
      bdrv_drained_end(src);
-@@ -1763,10 +1773,6 @@ static BlockJob *mirror_start_job(
+@@ -1699,10 +1709,6 @@ static BlockJob *mirror_start_job(
                         " sync mode",
                         MirrorSyncMode_str(sync_mode));
              return NULL;
@@ -66,7 +66,7 @@ index c3d4be9b15..7b6f7c0068 100644
          }
      } else if (bitmap) {
          error_setg(errp,
-@@ -1783,6 +1789,12 @@ static BlockJob *mirror_start_job(
+@@ -1719,6 +1725,12 @@ static BlockJob *mirror_start_job(
              return NULL;
          }
          granularity = bdrv_dirty_bitmap_granularity(bitmap);

debian/patches/bitmap-mirror/0003-mirror-add-check-for-bitmap-mode-without-bitmap.patch

@@ -16,10 +16,10 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  1 file changed, 3 insertions(+)
 
 diff --git a/blockdev.c b/blockdev.c
-index 38fa63155c..204cf6fad1 100644
+index a8b1fd2a73..83d5cc1e49 100644
 --- a/blockdev.c
 +++ b/blockdev.c
-@@ -2878,6 +2878,9 @@ static void blockdev_mirror_common(const char *job_id, BlockDriverState *bs,
+@@ -2945,6 +2945,9 @@ static void blockdev_mirror_common(const char *job_id, BlockDriverState *bs,
          if (bdrv_dirty_bitmap_check(bitmap, BDRV_BITMAP_ALLOW_RO, errp)) {
              return;
          }

debian/patches/bitmap-mirror/0004-mirror-switch-to-bdrv_dirty_bitmap_merge_internal.patch

@@ -16,10 +16,10 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  1 file changed, 4 insertions(+), 7 deletions(-)
 
 diff --git a/block/mirror.c b/block/mirror.c
-index 7b6f7c0068..2b1c07095d 100644
+index 11b8a8e959..00f2665ca4 100644
 --- a/block/mirror.c
 +++ b/block/mirror.c
-@@ -809,8 +809,8 @@ static int mirror_exit_common(Job *job)
+@@ -792,8 +792,8 @@ static int mirror_exit_common(Job *job)
               job->ret == 0 && ret == 0)) {
              /* Success; synchronize copy back to sync. */
              bdrv_clear_dirty_bitmap(s->sync_bitmap, NULL);
@@ -30,7 +30,7 @@ index 7b6f7c0068..2b1c07095d 100644
          }
      }
      bdrv_release_dirty_bitmap(s->dirty_bitmap);
-@@ -1971,11 +1971,8 @@ static BlockJob *mirror_start_job(
+@@ -1892,11 +1892,8 @@ static BlockJob *mirror_start_job(
      }
  
      if (s->sync_mode == MIRROR_SYNC_MODE_BITMAP) {
@@ -43,4 +43,4 @@ index 7b6f7c0068..2b1c07095d 100644
 +                                         NULL, true);
      }
  
-     bdrv_graph_wrlock();
+     ret = block_job_add_bdrv(&s->common, "source", bs, 0,

debian/patches/bitmap-mirror/0006-mirror-move-some-checks-to-qmp.patch

@@ -12,7 +12,7 @@ uniform w.r.t. backup block jobs.
 
 Signed-off-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
 Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
-[FE: rebase for 8.2.2]
+[FE: rebase for 8.0]
 Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
 ---
  block/mirror.c             | 28 +++------------
@@ -21,12 +21,12 @@ Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
  3 files changed, 70 insertions(+), 59 deletions(-)
 
 diff --git a/block/mirror.c b/block/mirror.c
-index 2b1c07095d..f5787b380c 100644
+index 00f2665ca4..60cf574de5 100644
 --- a/block/mirror.c
 +++ b/block/mirror.c
-@@ -1763,31 +1763,13 @@ static BlockJob *mirror_start_job(
- 
-     GLOBAL_STATE_CODE();
+@@ -1699,31 +1699,13 @@ static BlockJob *mirror_start_job(
+     uint64_t target_perms, target_shared_perms;
+     int ret;
  
 -    if (sync_mode == MIRROR_SYNC_MODE_INCREMENTAL) {
 -        error_setg(errp, "Sync mode '%s' not supported",
@@ -62,10 +62,10 @@ index 2b1c07095d..f5787b380c 100644
  
          if (bitmap_mode != BITMAP_SYNC_MODE_NEVER) {
 diff --git a/blockdev.c b/blockdev.c
-index 204cf6fad1..79d47b1920 100644
+index 83d5cc1e49..060d86a65f 100644
 --- a/blockdev.c
 +++ b/blockdev.c
-@@ -2857,7 +2857,36 @@ static void blockdev_mirror_common(const char *job_id, BlockDriverState *bs,
+@@ -2924,7 +2924,36 @@ static void blockdev_mirror_common(const char *job_id, BlockDriverState *bs,
          sync = MIRROR_SYNC_MODE_FULL;
      }
  

debian/patches/extra/0001-monitor-qmp-fix-race-with-clients-disconnecting-earl.patch

@@ -48,7 +48,7 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  6 files changed, 59 insertions(+), 5 deletions(-)
 
 diff --git a/include/monitor/monitor.h b/include/monitor/monitor.h
-index c3740ec616..7f38ce6b8b 100644
+index 965f5d5450..e04bd059b6 100644
 --- a/include/monitor/monitor.h
 +++ b/include/monitor/monitor.h
 @@ -16,6 +16,7 @@ extern QemuOptsList qemu_mon_opts;
@@ -60,7 +60,7 @@ index c3740ec616..7f38ce6b8b 100644
  void monitor_init_globals(void);
  void monitor_init_globals_core(void);
 diff --git a/monitor/monitor-internal.h b/monitor/monitor-internal.h
-index cb628f681d..93dbd62fc2 100644
+index 252de85681..8db28f9272 100644
 --- a/monitor/monitor-internal.h
 +++ b/monitor/monitor-internal.h
 @@ -151,6 +151,13 @@ typedef struct {
@@ -78,10 +78,10 @@ index cb628f681d..93dbd62fc2 100644
  
  /**
 diff --git a/monitor/monitor.c b/monitor/monitor.c
-index 56786c0ccc..30071d0c8a 100644
+index dc352f9e9d..56e1307014 100644
 --- a/monitor/monitor.c
 +++ b/monitor/monitor.c
-@@ -116,6 +116,21 @@ bool monitor_cur_is_qmp(void)
+@@ -117,6 +117,21 @@ bool monitor_cur_is_qmp(void)
      return cur_mon && monitor_is_qmp(cur_mon);
  }
  
@@ -104,7 +104,7 @@ index 56786c0ccc..30071d0c8a 100644
   * Is @mon is using readline?
   * Note: not all HMP monitors use readline, e.g., gdbserver has a
 diff --git a/monitor/qmp.c b/monitor/qmp.c
-index 5e538f34c0..eb181d5979 100644
+index 6eee450fe4..c15bf1e1fc 100644
 --- a/monitor/qmp.c
 +++ b/monitor/qmp.c
 @@ -165,6 +165,8 @@ static void monitor_qmp_dispatch(MonitorQMP *mon, QObject *req)
@@ -135,7 +135,7 @@ index 5e538f34c0..eb181d5979 100644
      qobject_unref(rsp);
  }
  
-@@ -461,6 +473,7 @@ static void monitor_qmp_event(void *opaque, QEMUChrEvent event)
+@@ -478,6 +490,7 @@ static void monitor_qmp_event(void *opaque, QEMUChrEvent event)
  
      switch (event) {
      case CHR_EVENT_OPENED:
@@ -144,7 +144,7 @@ index 5e538f34c0..eb181d5979 100644
          monitor_qmp_caps_reset(mon);
          data = qmp_greeting(mon);
 diff --git a/qapi/qmp-dispatch.c b/qapi/qmp-dispatch.c
-index 176b549473..790bb7d1da 100644
+index 555528b6bb..3baa508b4b 100644
 --- a/qapi/qmp-dispatch.c
 +++ b/qapi/qmp-dispatch.c
 @@ -117,16 +117,28 @@ typedef struct QmpDispatchBH {
@@ -180,16 +180,16 @@ index 176b549473..790bb7d1da 100644
      aio_co_wake(data->co);
  }
  
-@@ -253,6 +265,7 @@ QDict *coroutine_mixed_fn qmp_dispatch(const QmpCommandList *cmds, QObject *requ
+@@ -231,6 +243,7 @@ QDict *coroutine_mixed_fn qmp_dispatch(const QmpCommandList *cmds, QObject *requ
              .ret        = &ret,
              .errp       = &err,
              .co         = qemu_coroutine_self(),
 +            .conn_nr    = monitor_get_connection_nr(cur_mon),
          };
-         aio_bh_schedule_oneshot(iohandler_get_aio_context(), do_qmp_dispatch_bh,
+         aio_bh_schedule_oneshot(qemu_get_aio_context(), do_qmp_dispatch_bh,
                                  &data);
 diff --git a/stubs/monitor-core.c b/stubs/monitor-core.c
-index 1894cdfe1f..d74d0459f0 100644
+index afa477aae6..d3ff124bf3 100644
 --- a/stubs/monitor-core.c
 +++ b/stubs/monitor-core.c
 @@ -12,6 +12,11 @@ Monitor *monitor_set_cur(Coroutine *co, Monitor *mon)
@@ -201,6 +201,6 @@ index 1894cdfe1f..d74d0459f0 100644
 +    return -1;
 +}
 +
- void qapi_event_emit(QAPIEvent event, QDict *qdict)
+ void monitor_init_qmp(Chardev *chr, bool pretty, Error **errp)
  {
  }

debian/patches/extra/0002-scsi-megasas-Internal-cdbs-have-16-byte-length.patch

@@ -0,0 +1,69 @@
+From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
+From: Guenter Roeck <linux@roeck-us.net>
+Date: Tue, 28 Feb 2023 09:11:29 -0800
+Subject: [PATCH] scsi: megasas: Internal cdbs have 16-byte length
+MIME-Version: 1.0
+Content-Type: text/plain; charset=UTF-8
+Content-Transfer-Encoding: 8bit
+
+Host drivers do not necessarily set cdb_len in megasas io commands.
+With commits 6d1511cea0 ("scsi: Reject commands if the CDB length
+exceeds buf_len") and fe9d8927e2 ("scsi: Add buf_len parameter to
+scsi_req_new()"), this results in failures to boot Linux from affected
+SCSI drives because cdb_len is set to 0 by the host driver.
+Set the cdb length to its actual size to solve the problem.
+
+Signed-off-by: Guenter Roeck <linux@roeck-us.net>
+Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
+(picked-up from https://lists.nongnu.org/archive/html/qemu-devel/2023-02/msg08653.html)
+Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
+---
+ hw/scsi/megasas.c | 14 ++------------
+ 1 file changed, 2 insertions(+), 12 deletions(-)
+
+diff --git a/hw/scsi/megasas.c b/hw/scsi/megasas.c
+index 32c70c9e99..984b6a3145 100644
+--- a/hw/scsi/megasas.c
++++ b/hw/scsi/megasas.c
+@@ -1781,7 +1781,7 @@ static int megasas_handle_io(MegasasState *s, MegasasCmd *cmd, int frame_cmd)
+     uint8_t cdb[16];
+     int len;
+     struct SCSIDevice *sdev = NULL;
+-    int target_id, lun_id, cdb_len;
++    int target_id, lun_id;
+ 
+     lba_count = le32_to_cpu(cmd->frame->io.header.data_len);
+     lba_start_lo = le32_to_cpu(cmd->frame->io.lba_lo);
+@@ -1790,7 +1790,6 @@ static int megasas_handle_io(MegasasState *s, MegasasCmd *cmd, int frame_cmd)
+ 
+     target_id = cmd->frame->header.target_id;
+     lun_id = cmd->frame->header.lun_id;
+-    cdb_len = cmd->frame->header.cdb_len;
+ 
+     if (target_id < MFI_MAX_LD && lun_id == 0) {
+         sdev = scsi_device_find(&s->bus, 0, target_id, lun_id);
+@@ -1805,15 +1804,6 @@ static int megasas_handle_io(MegasasState *s, MegasasCmd *cmd, int frame_cmd)
+         return MFI_STAT_DEVICE_NOT_FOUND;
+     }
+ 
+-    if (cdb_len > 16) {
+-        trace_megasas_scsi_invalid_cdb_len(
+-            mfi_frame_desc(frame_cmd), 1, target_id, lun_id, cdb_len);
+-        megasas_write_sense(cmd, SENSE_CODE(INVALID_OPCODE));
+-        cmd->frame->header.scsi_status = CHECK_CONDITION;
+-        s->event_count++;
+-        return MFI_STAT_SCSI_DONE_WITH_ERROR;
+-    }
+-
+     cmd->iov_size = lba_count * sdev->blocksize;
+     if (megasas_map_sgl(s, cmd, &cmd->frame->io.sgl)) {
+         megasas_write_sense(cmd, SENSE_CODE(TARGET_FAILURE));
+@@ -1824,7 +1814,7 @@ static int megasas_handle_io(MegasasState *s, MegasasCmd *cmd, int frame_cmd)
+ 
+     megasas_encode_lba(cdb, lba_start, lba_count, is_write);
+     cmd->req = scsi_req_new(sdev, cmd->index,
+-                            lun_id, cdb, cdb_len, cmd);
++                            lun_id, cdb, sizeof(cdb), cmd);
+     if (!cmd->req) {
+         trace_megasas_scsi_req_alloc_failed(
+             mfi_frame_desc(frame_cmd), target_id, lun_id);

debian/patches/extra/0003-ide-avoid-potential-deadlock-when-draining-during-tr.patch

@@ -55,10 +55,10 @@ Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
  1 file changed, 6 insertions(+), 6 deletions(-)
 
 diff --git a/hw/ide/core.c b/hw/ide/core.c
-index 08d9218455..20d8c0cf66 100644
+index 07971c0218..6a74afe564 100644
 --- a/hw/ide/core.c
 +++ b/hw/ide/core.c
-@@ -456,7 +456,7 @@ static void ide_trim_bh_cb(void *opaque)
+@@ -444,7 +444,7 @@ static void ide_trim_bh_cb(void *opaque)
      iocb->bh = NULL;
      qemu_aio_unref(iocb);
  
@@ -67,7 +67,7 @@ index 08d9218455..20d8c0cf66 100644
      blk_dec_in_flight(blk);
  }
  
-@@ -516,6 +516,8 @@ static void ide_issue_trim_cb(void *opaque, int ret)
+@@ -504,6 +504,8 @@ static void ide_issue_trim_cb(void *opaque, int ret)
  done:
      iocb->aiocb = NULL;
      if (iocb->bh) {
@@ -76,7 +76,7 @@ index 08d9218455..20d8c0cf66 100644
          replay_bh_schedule_event(iocb->bh);
      }
  }
-@@ -528,9 +530,6 @@ BlockAIOCB *ide_issue_trim(
+@@ -516,9 +518,6 @@ BlockAIOCB *ide_issue_trim(
      IDEDevice *dev = s->unit ? s->bus->slave : s->bus->master;
      TrimAIOCB *iocb;
  
@@ -86,7 +86,7 @@ index 08d9218455..20d8c0cf66 100644
      iocb = blk_aio_get(&trim_aiocb_info, s->blk, cb, cb_opaque);
      iocb->s = s;
      iocb->bh = qemu_bh_new_guarded(ide_trim_bh_cb, iocb,
-@@ -754,8 +753,9 @@ void ide_cancel_dma_sync(IDEState *s)
+@@ -742,8 +741,9 @@ void ide_cancel_dma_sync(IDEState *s)
       */
      if (s->bus->dma->aiocb) {
          trace_ide_cancel_dma_sync_remaining();

debian/patches/extra/0003-tcg-Reset-free_temps-before-tcg_optimize.patch

@@ -1,82 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Richard Henderson <richard.henderson@linaro.org>
-Date: Sat, 7 Dec 2024 18:14:45 +0000
-Subject: [PATCH] tcg: Reset free_temps before tcg_optimize
-MIME-Version: 1.0
-Content-Type: text/plain; charset=UTF-8
-Content-Transfer-Encoding: 8bit
-
-When allocating new temps during tcg_optmize, do not re-use
-any EBB temps that were used within the TB.  We do not have
-any idea what span of the TB in which the temp was live.
-
-Introduce tcg_temp_ebb_reset_freed and use before tcg_optimize,
-as well as replacing the equivalent in plugin_gen_inject and
-tcg_func_start.
-
-Cc: qemu-stable@nongnu.org
-Fixes: fb04ab7ddd8 ("tcg/optimize: Lower TCG_COND_TST{EQ,NE} if unsupported")
-Resolves: https://gitlab.com/qemu-project/qemu/-/issues/2711
-Reported-by: wannacu <wannacu2049@gmail.com>
-Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
-Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
-Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
-(cherry picked from commit 04e006ab36a8565b92d4e21dd346367fbade7d74)
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
----
- accel/tcg/plugin-gen.c          | 2 +-
- include/tcg/tcg-temp-internal.h | 6 ++++++
- tcg/tcg.c                       | 5 ++++-
- 3 files changed, 11 insertions(+), 2 deletions(-)
-
-diff --git a/accel/tcg/plugin-gen.c b/accel/tcg/plugin-gen.c
-index 0f47bfbb48..1ef075552c 100644
---- a/accel/tcg/plugin-gen.c
-+++ b/accel/tcg/plugin-gen.c
-@@ -275,7 +275,7 @@ static void plugin_gen_inject(struct qemu_plugin_tb *plugin_tb)
-      * that might be live within the existing opcode stream.
-      * The simplest solution is to release them all and create new.
-      */
--    memset(tcg_ctx->free_temps, 0, sizeof(tcg_ctx->free_temps));
-+    tcg_temp_ebb_reset_freed(tcg_ctx);
- 
-     QTAILQ_FOREACH_SAFE(op, &tcg_ctx->ops, link, next) {
-         switch (op->opc) {
-diff --git a/include/tcg/tcg-temp-internal.h b/include/tcg/tcg-temp-internal.h
-index 44192c55a9..98f91e68b7 100644
---- a/include/tcg/tcg-temp-internal.h
-+++ b/include/tcg/tcg-temp-internal.h
-@@ -42,4 +42,10 @@ TCGv_i64 tcg_temp_ebb_new_i64(void);
- TCGv_ptr tcg_temp_ebb_new_ptr(void);
- TCGv_i128 tcg_temp_ebb_new_i128(void);
- 
-+/* Forget all freed EBB temps, so that new allocations produce new temps. */
-+static inline void tcg_temp_ebb_reset_freed(TCGContext *s)
-+{
-+    memset(s->free_temps, 0, sizeof(s->free_temps));
-+}
-+
- #endif /* TCG_TEMP_FREE_H */
-diff --git a/tcg/tcg.c b/tcg/tcg.c
-index 0babae1b88..4578b185be 100644
---- a/tcg/tcg.c
-+++ b/tcg/tcg.c
-@@ -1489,7 +1489,7 @@ void tcg_func_start(TCGContext *s)
-     s->nb_temps = s->nb_globals;
- 
-     /* No temps have been previously allocated for size or locality.  */
--    memset(s->free_temps, 0, sizeof(s->free_temps));
-+    tcg_temp_ebb_reset_freed(s);
- 
-     /* No constant temps have been previously allocated. */
-     for (int i = 0; i < TCG_TYPE_COUNT; ++i) {
-@@ -6120,6 +6120,9 @@ int tcg_gen_code(TCGContext *s, TranslationBlock *tb, uint64_t pc_start)
-     }
- #endif
- 
-+    /* Do not reuse any EBB that may be allocated within the TB. */
-+    tcg_temp_ebb_reset_freed(s);
-+
-     tcg_optimize(s);
- 
-     reachable_code_pass(s);

debian/patches/extra/0004-migration-block-dirty-bitmap-fix-loading-bitmap-when.patch

@@ -0,0 +1,48 @@
+From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
+From: Fiona Ebner <f.ebner@proxmox.com>
+Date: Fri, 28 Jul 2023 10:47:48 +0200
+Subject: [PATCH] migration/block-dirty-bitmap: fix loading bitmap when there
+ is an iothread
+
+The bdrv_create_dirty_bitmap() function (which is also called by
+bdrv_dirty_bitmap_create_successor()) uses bdrv_getlength(bs). This is
+a wrapper around a coroutine, and thus uses bdrv_poll_co(). Polling
+tries to release the AioContext which will trigger an assert() if it
+hasn't been acquired before.
+
+The issue does not happen for migration, because there we are in a
+coroutine already, so the wrapper will just call bdrv_co_getlength()
+directly without polling.
+
+Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
+---
+ migration/block-dirty-bitmap.c | 6 ++++++
+ 1 file changed, 6 insertions(+)
+
+diff --git a/migration/block-dirty-bitmap.c b/migration/block-dirty-bitmap.c
+index 032fc5f405..e1ae3b7316 100644
+--- a/migration/block-dirty-bitmap.c
++++ b/migration/block-dirty-bitmap.c
+@@ -805,8 +805,11 @@ static int dirty_bitmap_load_start(QEMUFile *f, DBMLoadState *s)
+                      "destination", bdrv_dirty_bitmap_name(s->bitmap));
+         return -EINVAL;
+     } else {
++        AioContext *ctx = bdrv_get_aio_context(s->bs);
++        aio_context_acquire(ctx);
+         s->bitmap = bdrv_create_dirty_bitmap(s->bs, granularity,
+                                              s->bitmap_name, &local_err);
++        aio_context_release(ctx);
+         if (!s->bitmap) {
+             error_report_err(local_err);
+             return -EINVAL;
+@@ -833,7 +836,10 @@ static int dirty_bitmap_load_start(QEMUFile *f, DBMLoadState *s)
+ 
+     bdrv_disable_dirty_bitmap(s->bitmap);
+     if (flags & DIRTY_BITMAP_MIG_START_FLAG_ENABLED) {
++        AioContext *ctx = bdrv_get_aio_context(s->bs);
++        aio_context_acquire(ctx);
+         bdrv_dirty_bitmap_create_successor(s->bitmap, &local_err);
++        aio_context_release(ctx);
+         if (local_err) {
+             error_report_err(local_err);
+             return -EINVAL;

debian/patches/extra/0004-target-i386-Reset-TSCs-of-parked-vCPUs-too-on-VM-res.patch

@@ -1,149 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: "Maciej S. Szmigiero" <maciej.szmigiero@oracle.com>
-Date: Thu, 12 Dec 2024 15:51:15 +0100
-Subject: [PATCH] target/i386: Reset TSCs of parked vCPUs too on VM reset
-
-Since commit 5286c3662294 ("target/i386: properly reset TSC on reset")
-QEMU writes the special value of "1" to each online vCPU TSC on VM reset
-to reset it.
-
-However parked vCPUs don't get that handling and due to that their TSCs
-get desynchronized when the VM gets reset.
-This in turn causes KVM to turn off PVCLOCK_TSC_STABLE_BIT in its exported
-PV clock.
-Note that KVM has no understanding of vCPU being currently parked.
-
-Without PVCLOCK_TSC_STABLE_BIT the sched clock is marked unstable in
-the guest's kvm_sched_clock_init().
-This causes a performance regressions to show in some tests.
-
-Fix this issue by writing the special value of "1" also to TSCs of parked
-vCPUs on VM reset.
-
-Reproducing the issue:
-1) Boot a VM with "-smp 2,maxcpus=3" or similar
-
-2) device_add host-x86_64-cpu,id=vcpu,node-id=0,socket-id=0,core-id=2,thread-id=0
-
-3) Wait a few seconds
-
-4) device_del vcpu
-
-5) Inside the VM run:
-# echo "t" >/proc/sysrq-trigger; dmesg | grep sched_clock_stable
-Observe the sched_clock_stable() value is 1.
-
-6) Reboot the VM
-
-7) Once the VM boots once again run inside it:
-# echo "t" >/proc/sysrq-trigger; dmesg | grep sched_clock_stable
-Observe the sched_clock_stable() value is now 0.
-
-Fixes: 5286c3662294 ("target/i386: properly reset TSC on reset")
-Signed-off-by: Maciej S. Szmigiero <maciej.szmigiero@oracle.com>
-Link: https://lore.kernel.org/r/5a605a88e9a231386dc803c60f5fed9b48108139.1734014926.git.maciej.szmigiero@oracle.com
-Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
-(cherry picked from commit 3f2a05b31ee9ce2ddb6c75a9bc3f5e7f7af9a76f)
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
----
- accel/kvm/kvm-all.c                | 11 +++++++++++
- configs/targets/i386-softmmu.mak   |  1 +
- configs/targets/x86_64-softmmu.mak |  1 +
- include/sysemu/kvm.h               |  8 ++++++++
- target/i386/kvm/kvm.c              | 15 +++++++++++++++
- 5 files changed, 36 insertions(+)
-
-diff --git a/accel/kvm/kvm-all.c b/accel/kvm/kvm-all.c
-index 801cff16a5..dec1d1c16a 100644
---- a/accel/kvm/kvm-all.c
-+++ b/accel/kvm/kvm-all.c
-@@ -437,6 +437,16 @@ int kvm_unpark_vcpu(KVMState *s, unsigned long vcpu_id)
-     return kvm_fd;
- }
- 
-+static void kvm_reset_parked_vcpus(void *param)
-+{
-+    KVMState *s = param;
-+    struct KVMParkedVcpu *cpu;
-+
-+    QLIST_FOREACH(cpu, &s->kvm_parked_vcpus, node) {
-+        kvm_arch_reset_parked_vcpu(cpu->vcpu_id, cpu->kvm_fd);
-+    }
-+}
-+
- int kvm_create_vcpu(CPUState *cpu)
- {
-     unsigned long vcpu_id = kvm_arch_vcpu_id(cpu);
-@@ -2728,6 +2738,7 @@ static int kvm_init(MachineState *ms)
-     }
- 
-     qemu_register_reset(kvm_unpoison_all, NULL);
-+    qemu_register_reset(kvm_reset_parked_vcpus, s);
- 
-     if (s->kernel_irqchip_allowed) {
-         kvm_irqchip_create(s);
-diff --git a/configs/targets/i386-softmmu.mak b/configs/targets/i386-softmmu.mak
-index 2ac69d5ba3..2eb0e86250 100644
---- a/configs/targets/i386-softmmu.mak
-+++ b/configs/targets/i386-softmmu.mak
-@@ -1,4 +1,5 @@
- TARGET_ARCH=i386
- TARGET_SUPPORTS_MTTCG=y
- TARGET_KVM_HAVE_GUEST_DEBUG=y
-+TARGET_KVM_HAVE_RESET_PARKED_VCPU=y
- TARGET_XML_FILES= gdb-xml/i386-32bit.xml
-diff --git a/configs/targets/x86_64-softmmu.mak b/configs/targets/x86_64-softmmu.mak
-index e12ac3dc59..920e9a4200 100644
---- a/configs/targets/x86_64-softmmu.mak
-+++ b/configs/targets/x86_64-softmmu.mak
-@@ -2,4 +2,5 @@ TARGET_ARCH=x86_64
- TARGET_BASE_ARCH=i386
- TARGET_SUPPORTS_MTTCG=y
- TARGET_KVM_HAVE_GUEST_DEBUG=y
-+TARGET_KVM_HAVE_RESET_PARKED_VCPU=y
- TARGET_XML_FILES= gdb-xml/i386-64bit.xml
-diff --git a/include/sysemu/kvm.h b/include/sysemu/kvm.h
-index c3a60b2890..ab17c09a55 100644
---- a/include/sysemu/kvm.h
-+++ b/include/sysemu/kvm.h
-@@ -377,6 +377,14 @@ int kvm_arch_init(MachineState *ms, KVMState *s);
- int kvm_arch_init_vcpu(CPUState *cpu);
- int kvm_arch_destroy_vcpu(CPUState *cpu);
- 
-+#ifdef TARGET_KVM_HAVE_RESET_PARKED_VCPU
-+void kvm_arch_reset_parked_vcpu(unsigned long vcpu_id, int kvm_fd);
-+#else
-+static inline void kvm_arch_reset_parked_vcpu(unsigned long vcpu_id, int kvm_fd)
-+{
-+}
-+#endif
-+
- bool kvm_vcpu_id_is_valid(int vcpu_id);
- 
- /* Returns VCPU ID to be used on KVM_CREATE_VCPU ioctl() */
-diff --git a/target/i386/kvm/kvm.c b/target/i386/kvm/kvm.c
-index 8e17942c3b..2ff618fbf1 100644
---- a/target/i386/kvm/kvm.c
-+++ b/target/i386/kvm/kvm.c
-@@ -2415,6 +2415,21 @@ void kvm_arch_after_reset_vcpu(X86CPU *cpu)
-     }
- }
- 
-+void kvm_arch_reset_parked_vcpu(unsigned long vcpu_id, int kvm_fd)
-+{
-+    g_autofree struct kvm_msrs *msrs = NULL;
-+
-+    msrs = g_malloc0(sizeof(*msrs) + sizeof(msrs->entries[0]));
-+    msrs->entries[0].index = MSR_IA32_TSC;
-+    msrs->entries[0].data = 1; /* match the value in x86_cpu_reset() */
-+    msrs->nmsrs++;
-+
-+    if (ioctl(kvm_fd, KVM_SET_MSRS, msrs) != 1) {
-+        warn_report("parked vCPU %lu TSC reset failed: %d",
-+                    vcpu_id, errno);
-+    }
-+}
-+
- void kvm_arch_do_init_vcpu(X86CPU *cpu)
- {
-     CPUX86State *env = &cpu->env;

debian/patches/extra/0005-hw-ide-reset-cancel-async-DMA-operation-before-reset.patch

@@ -0,0 +1,100 @@
+From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
+From: Fiona Ebner <f.ebner@proxmox.com>
+Date: Thu, 24 Aug 2023 11:22:21 +0200
+Subject: [PATCH] hw/ide: reset: cancel async DMA operation before reseting
+ state
+
+If there is a pending DMA operation during ide_bus_reset(), the fact
+that the IDEstate is already reset before the operation is canceled
+can be problematic. In particular, ide_dma_cb() might be called and
+then use the reset IDEstate which contains the signature after the
+reset. When used to construct the IO operation this leads to
+ide_get_sector() returning 0 and nsector being 1. This is particularly
+bad, because a write command will thus destroy the first sector which
+often contains a partition table or similar.
+
+Traces showing the unsolicited write happening with IDEstate
+0x5595af6949d0 being used after reset:
+
+> ahci_port_write ahci(0x5595af6923f0)[0]: port write [reg:PxSCTL] @ 0x2c: 0x00000300
+> ahci_reset_port ahci(0x5595af6923f0)[0]: reset port
+> ide_reset IDEstate 0x5595af6949d0
+> ide_reset IDEstate 0x5595af694da8
+> ide_bus_reset_aio aio_cancel
+> dma_aio_cancel dbs=0x7f64600089a0
+> dma_blk_cb dbs=0x7f64600089a0 ret=0
+> dma_complete dbs=0x7f64600089a0 ret=0 cb=0x5595acd40b30
+> ahci_populate_sglist ahci(0x5595af6923f0)[0]
+> ahci_dma_prepare_buf ahci(0x5595af6923f0)[0]: prepare buf limit=512 prepared=512
+> ide_dma_cb IDEState 0x5595af6949d0; sector_num=0 n=1 cmd=DMA WRITE
+> dma_blk_io dbs=0x7f6420802010 bs=0x5595ae2c6c30 offset=0 to_dev=1
+> dma_blk_cb dbs=0x7f6420802010 ret=0
+
+> (gdb) p *qiov
+> $11 = {iov = 0x7f647c76d840, niov = 1, {{nalloc = 1, local_iov = {iov_base = 0x0,
+>       iov_len = 512}}, {__pad = "\001\000\000\000\000\000\000\000\000\000\000",
+>       size = 512}}}
+> (gdb) bt
+> #0  blk_aio_pwritev (blk=0x5595ae2c6c30, offset=0, qiov=0x7f6420802070, flags=0,
+>     cb=0x5595ace6f0b0 <dma_blk_cb>, opaque=0x7f6420802010)
+>     at ../block/block-backend.c:1682
+> #1  0x00005595ace6f185 in dma_blk_cb (opaque=0x7f6420802010, ret=<optimized out>)
+>     at ../softmmu/dma-helpers.c:179
+> #2  0x00005595ace6f778 in dma_blk_io (ctx=0x5595ae0609f0,
+>     sg=sg@entry=0x5595af694d00, offset=offset@entry=0, align=align@entry=512,
+>     io_func=io_func@entry=0x5595ace6ee30 <dma_blk_write_io_func>,
+>     io_func_opaque=io_func_opaque@entry=0x5595ae2c6c30,
+>     cb=0x5595acd40b30 <ide_dma_cb>, opaque=0x5595af6949d0,
+>     dir=DMA_DIRECTION_TO_DEVICE) at ../softmmu/dma-helpers.c:244
+> #3  0x00005595ace6f90a in dma_blk_write (blk=0x5595ae2c6c30,
+>     sg=sg@entry=0x5595af694d00, offset=offset@entry=0, align=align@entry=512,
+>     cb=cb@entry=0x5595acd40b30 <ide_dma_cb>, opaque=opaque@entry=0x5595af6949d0)
+>     at ../softmmu/dma-helpers.c:280
+> #4  0x00005595acd40e18 in ide_dma_cb (opaque=0x5595af6949d0, ret=<optimized out>)
+>     at ../hw/ide/core.c:953
+> #5  0x00005595ace6f319 in dma_complete (ret=0, dbs=0x7f64600089a0)
+>     at ../softmmu/dma-helpers.c:107
+> #6  dma_blk_cb (opaque=0x7f64600089a0, ret=0) at ../softmmu/dma-helpers.c:127
+> #7  0x00005595ad12227d in blk_aio_complete (acb=0x7f6460005b10)
+>     at ../block/block-backend.c:1527
+> #8  blk_aio_complete (acb=0x7f6460005b10) at ../block/block-backend.c:1524
+> #9  blk_aio_write_entry (opaque=0x7f6460005b10) at ../block/block-backend.c:1594
+> #10 0x00005595ad258cfb in coroutine_trampoline (i0=<optimized out>,
+>     i1=<optimized out>) at ../util/coroutine-ucontext.c:177
+
+Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
+---
+ hw/ide/core.c | 14 +++++++-------
+ 1 file changed, 7 insertions(+), 7 deletions(-)
+
+diff --git a/hw/ide/core.c b/hw/ide/core.c
+index 6a74afe564..289347af58 100644
+--- a/hw/ide/core.c
++++ b/hw/ide/core.c
+@@ -2515,19 +2515,19 @@ static void ide_dummy_transfer_stop(IDEState *s)
+ 
+ void ide_bus_reset(IDEBus *bus)
+ {
+-    bus->unit = 0;
+-    bus->cmd = 0;
+-    ide_reset(&bus->ifs[0]);
+-    ide_reset(&bus->ifs[1]);
+-    ide_clear_hob(bus);
+-
+-    /* pending async DMA */
++    /* pending async DMA - needs the IDEState before it is reset */
+     if (bus->dma->aiocb) {
+         trace_ide_bus_reset_aio();
+         blk_aio_cancel(bus->dma->aiocb);
+         bus->dma->aiocb = NULL;
+     }
+ 
++    bus->unit = 0;
++    bus->cmd = 0;
++    ide_reset(&bus->ifs[0]);
++    ide_reset(&bus->ifs[1]);
++    ide_clear_hob(bus);
++
+     /* reset dma provider too */
+     if (bus->dma->ops->reset) {
+         bus->dma->ops->reset(bus->dma);

debian/patches/extra/0005-i386-cpu-Mark-avx10_version-filtered-when-prefix-is-.patch

@@ -1,41 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Zhao Liu <zhao1.liu@intel.com>
-Date: Wed, 6 Nov 2024 11:07:18 +0800
-Subject: [PATCH] i386/cpu: Mark avx10_version filtered when prefix is NULL
-
-In x86_cpu_filter_features(), if host doesn't support AVX10, the
-configured avx10_version should be marked as filtered regardless of
-whether prefix is NULL or not.
-
-Check prefix before warn_report() instead of checking for
-have_filtered_features.
-
-Cc: qemu-stable@nongnu.org
-Fixes: commit bccfb846fd52 ("target/i386: add AVX10 feature and AVX10 version property")
-Signed-off-by: Zhao Liu <zhao1.liu@intel.com>
-Reviewed-by: Tao Su <tao1.su@linux.intel.com>
-Link: https://lore.kernel.org/r/20241106030728.553238-2-zhao1.liu@intel.com
-Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
-(cherry picked from commit cf4c263551886964c5d58bd7b675b13fd497b402)
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
----
- target/i386/cpu.c | 6 ++++--
- 1 file changed, 4 insertions(+), 2 deletions(-)
-
-diff --git a/target/i386/cpu.c b/target/i386/cpu.c
-index 3725dbbc4b..1981aeaba5 100644
---- a/target/i386/cpu.c
-+++ b/target/i386/cpu.c
-@@ -7718,8 +7718,10 @@ static bool x86_cpu_filter_features(X86CPU *cpu, bool verbose)
-             env->avx10_version = version;
-             have_filtered_features = true;
-         }
--    } else if (env->avx10_version && prefix) {
--        warn_report("%s: avx10.%d.", prefix, env->avx10_version);
-+    } else if (env->avx10_version) {
-+        if (prefix) {
-+            warn_report("%s: avx10.%d.", prefix, env->avx10_version);
-+        }
-         have_filtered_features = true;
-     }
- 

debian/patches/extra/0006-Revert-Revert-graph-lock-Disable-locking-for-now.patch

@@ -0,0 +1,140 @@
+From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
+From: Fiona Ebner <f.ebner@proxmox.com>
+Date: Thu, 28 Sep 2023 10:07:03 +0200
+Subject: [PATCH] Revert "Revert "graph-lock: Disable locking for now""
+
+This reverts commit 3cce22defb4b0e47cf135444e30cc673cff5ebad.
+
+There are still some issues with graph locking, e.g. deadlocks during
+backup canceling [0]. Because the AioContext locks still exist, it
+should be safe to disable locking again.
+
+From the original 80fc5d2600 ("graph-lock: Disable locking for now"):
+
+> We don't currently rely on graph locking yet. It is supposed to replace
+> the AioContext lock eventually to enable multiqueue support, but as long
+> as we still have the AioContext lock, it is sufficient without the graph
+> lock. Once the AioContext lock goes away, the deadlock doesn't exist any
+> more either and this commit can be reverted. (Of course, it can also be
+> reverted while the AioContext lock still exists if the callers have been
+> fixed.)
+
+[0]: https://lists.nongnu.org/archive/html/qemu-devel/2023-09/msg00729.html
+
+Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
+---
+ block/graph-lock.c | 24 ++++++++++++++++++++++++
+ 1 file changed, 24 insertions(+)
+
+diff --git a/block/graph-lock.c b/block/graph-lock.c
+index 5e66f01ae8..5c2873262a 100644
+--- a/block/graph-lock.c
++++ b/block/graph-lock.c
+@@ -30,8 +30,10 @@ BdrvGraphLock graph_lock;
+ /* Protects the list of aiocontext and orphaned_reader_count */
+ static QemuMutex aio_context_list_lock;
+ 
++#if 0
+ /* Written and read with atomic operations. */
+ static int has_writer;
++#endif
+ 
+ /*
+  * A reader coroutine could move from an AioContext to another.
+@@ -88,6 +90,7 @@ void unregister_aiocontext(AioContext *ctx)
+     g_free(ctx->bdrv_graph);
+ }
+ 
++#if 0
+ static uint32_t reader_count(void)
+ {
+     BdrvGraphRWlock *brdv_graph;
+@@ -105,12 +108,19 @@ static uint32_t reader_count(void)
+     assert((int32_t)rd >= 0);
+     return rd;
+ }
++#endif
+ 
+ void bdrv_graph_wrlock(BlockDriverState *bs)
+ {
++#if 0
+     AioContext *ctx = NULL;
+ 
+     GLOBAL_STATE_CODE();
++    /*
++     * TODO Some callers hold an AioContext lock when this is called, which
++     * causes deadlocks. Reenable once the AioContext locking is cleaned up (or
++     * AioContext locks are gone).
++     */
+     assert(!qatomic_read(&has_writer));
+ 
+     /*
+@@ -158,11 +168,13 @@ void bdrv_graph_wrlock(BlockDriverState *bs)
+     if (ctx) {
+         aio_context_acquire(bdrv_get_aio_context(bs));
+     }
++#endif
+ }
+ 
+ void bdrv_graph_wrunlock(void)
+ {
+     GLOBAL_STATE_CODE();
++#if 0
+     QEMU_LOCK_GUARD(&aio_context_list_lock);
+     assert(qatomic_read(&has_writer));
+ 
+@@ -174,10 +186,13 @@ void bdrv_graph_wrunlock(void)
+ 
+     /* Wake up all coroutine that are waiting to read the graph */
+     qemu_co_enter_all(&reader_queue, &aio_context_list_lock);
++#endif
+ }
+ 
+ void coroutine_fn bdrv_graph_co_rdlock(void)
+ {
++    /* TODO Reenable when wrlock is reenabled */
++#if 0
+     BdrvGraphRWlock *bdrv_graph;
+     bdrv_graph = qemu_get_current_aio_context()->bdrv_graph;
+ 
+@@ -237,10 +252,12 @@ void coroutine_fn bdrv_graph_co_rdlock(void)
+             qemu_co_queue_wait(&reader_queue, &aio_context_list_lock);
+         }
+     }
++#endif
+ }
+ 
+ void coroutine_fn bdrv_graph_co_rdunlock(void)
+ {
++#if 0
+     BdrvGraphRWlock *bdrv_graph;
+     bdrv_graph = qemu_get_current_aio_context()->bdrv_graph;
+ 
+@@ -258,6 +275,7 @@ void coroutine_fn bdrv_graph_co_rdunlock(void)
+     if (qatomic_read(&has_writer)) {
+         aio_wait_kick();
+     }
++#endif
+ }
+ 
+ void bdrv_graph_rdlock_main_loop(void)
+@@ -275,13 +293,19 @@ void bdrv_graph_rdunlock_main_loop(void)
+ void assert_bdrv_graph_readable(void)
+ {
+     /* reader_count() is slow due to aio_context_list_lock lock contention */
++    /* TODO Reenable when wrlock is reenabled */
++#if 0
+ #ifdef CONFIG_DEBUG_GRAPH_LOCK
+     assert(qemu_in_main_thread() || reader_count());
+ #endif
++#endif
+ }
+ 
+ void assert_bdrv_graph_writable(void)
+ {
+     assert(qemu_in_main_thread());
++    /* TODO Reenable when wrlock is reenabled */
++#if 0
+     assert(qatomic_read(&has_writer));
++#endif
+ }

debian/patches/extra/0006-net-Fix-announce_self.patch

@@ -1,67 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Laurent Vivier <lvivier@redhat.com>
-Date: Fri, 17 Jan 2025 12:17:08 +0100
-Subject: [PATCH] net: Fix announce_self
-
-b9ad513e1876 ("net: Remove receive_raw()") adds an iovec entry
-in qemu_deliver_packet_iov() to add the virtio-net header
-in the data when QEMU_NET_PACKET_FLAG_RAW is set but forgets
-to increase the number of iovec entries in the array, so
-receive_iov() will only send the first entry (the virtio-net
-entry, full of 0) and no data. The packet will be discarded.
-
-The only user of QEMU_NET_PACKET_FLAG_RAW is announce_self.
-
-We can see the problem with tcpdump:
-
-- QEMU parameters:
-
-  .. -monitor stdio \
-     -netdev bridge,id=netdev0,br=virbr0 \
-     -device virtio-net,mac=9a:2b:2c:2d:2e:2f,netdev=netdev0 \
-
-- HMP command:
-
-  (qemu) announce_self
-
-- TCP dump:
-
-  $ sudo tcpdump -nxi virbr0
-
-  without the fix:
-
-    <nothing>
-
-  with the fix:
-
-   ARP, Reverse Request who-is 9a:2b:2c:2d:2e:2f tell 9a:2b:2c:2d:2e:2f, length 46
-        0x0000:  0001 0800 0604 0003 9a2b 2c2d 2e2f 0000
-        0x0010:  0000 9a2b 2c2d 2e2f 0000 0000 0000 0000
-        0x0020:  0000 0000 0000 0000 0000 0000 0000
-
-Reported-by: Xiaohui Li <xiaohli@redhat.com>
-Bug: https://issues.redhat.com/browse/RHEL-73891
-Fixes: b9ad513e1876 ("net: Remove receive_raw()")
-Cc: akihiko.odaki@daynix.com
-Signed-off-by: Laurent Vivier <lvivier@redhat.com>
-Reviewed-by: Akihiko Odaki <akihiko.odaki@daynix.com>
-Acked-by: Jason Wang <jasowang@redhat.com>
-Reviewed-by: Michael Tokarev <mjt@tls.msk.ru>
-(picked from https://lore.kernel.org/qemu-devel/20250117111709.970789-2-lvivier@redhat.com/)
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
----
- net/net.c | 1 +
- 1 file changed, 1 insertion(+)
-
-diff --git a/net/net.c b/net/net.c
-index 7ef6885876..fefa701bb2 100644
---- a/net/net.c
-+++ b/net/net.c
-@@ -822,6 +822,7 @@ static ssize_t qemu_deliver_packet_iov(NetClientState *sender,
-         iov_copy[0].iov_len =  nc->vnet_hdr_len;
-         memcpy(&iov_copy[1], iov, iovcnt * sizeof(*iov));
-         iov = iov_copy;
-+        iovcnt++;
-     }
- 
-     if (nc->info->receive_iov) {

debian/patches/extra/0007-migration-states-workaround-snapshot-performance-reg.patch

@@ -0,0 +1,57 @@
+From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
+From: Fiona Ebner <f.ebner@proxmox.com>
+Date: Thu, 28 Sep 2023 11:19:14 +0200
+Subject: [PATCH] migration states: workaround snapshot performance regression
+
+Commit 813cd616 ("migration: Use migration_transferred_bytes() to
+calculate rate_limit") introduced a prohibitive performance regression
+when taking a snapshot [0]. The reason turns out to be the flushing
+done by migration_transferred_bytes()
+
+Just use a _noflush version of the relevant function as a workaround
+until upstream fixes the issue. This is inspired by a not-applied
+upstream series [1], but doing the very minimum to avoid the
+regression.
+
+[0]: https://gitlab.com/qemu-project/qemu/-/issues/1821
+[1]: https://lists.nongnu.org/archive/html/qemu-devel/2023-05/msg07708.html
+
+Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
+---
+ migration/migration-stats.c | 16 +++++++++++++++-
+ 1 file changed, 15 insertions(+), 1 deletion(-)
+
+diff --git a/migration/migration-stats.c b/migration/migration-stats.c
+index 095d6d75bb..8073c8ebaa 100644
+--- a/migration/migration-stats.c
++++ b/migration/migration-stats.c
+@@ -18,6 +18,20 @@
+ 
+ MigrationAtomicStats mig_stats;
+ 
++/*
++ * Same as migration_transferred_bytes below, but using the _noflush
++ * variant of qemu_file_transferred() to avoid a performance
++ * regression in migration_rate_exceeded().
++ */
++static uint64_t migration_transferred_bytes_noflush(QEMUFile *f)
++{
++    uint64_t multifd = stat64_get(&mig_stats.multifd_bytes);
++    uint64_t qemu_file = qemu_file_transferred_noflush(f);
++
++    trace_migration_transferred_bytes(qemu_file, multifd);
++    return qemu_file + multifd;
++}
++
+ bool migration_rate_exceeded(QEMUFile *f)
+ {
+     if (qemu_file_get_error(f)) {
+@@ -25,7 +39,7 @@ bool migration_rate_exceeded(QEMUFile *f)
+     }
+ 
+     uint64_t rate_limit_start = stat64_get(&mig_stats.rate_limit_start);
+-    uint64_t rate_limit_current = migration_transferred_bytes(f);
++    uint64_t rate_limit_current = migration_transferred_bytes_noflush(f);
+     uint64_t rate_limit_used = rate_limit_current - rate_limit_start;
+     uint64_t rate_limit_max = stat64_get(&mig_stats.rate_limit_max);
+ 

debian/patches/extra/0007-net-dump-Correctly-compute-Ethernet-packet-offset.patch

@@ -1,67 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Laurent Vivier <lvivier@redhat.com>
-Date: Fri, 17 Jan 2025 12:17:09 +0100
-Subject: [PATCH] net/dump: Correctly compute Ethernet packet offset
-
-When a packet is sent with QEMU_NET_PACKET_FLAG_RAW by QEMU it
-never includes virtio-net header even if qemu_get_vnet_hdr_len()
-is not 0, and filter-dump is not managing this case.
-
-The only user of QEMU_NET_PACKET_FLAG_RAW is announce_self,
-we can show the problem using it and tcpddump:
-
-- QEMU parameters:
-
-  .. -monitor stdio \
-     -netdev bridge,id=netdev0,br=virbr0 \
-     -device virtio-net,mac=9a:2b:2c:2d:2e:2f,netdev=netdev0 \
-     -object filter-dump,netdev=netdev0,file=log.pcap,id=pcap0
-
-- HMP command:
-
-  (qemu) announce_self
-
-- TCP dump:
-
-  $ tcpdump -nxr log.pcap
-
-  without the fix:
-
-    08:00:06:04:00:03 > 2e:2f:80:35:00:01, ethertype Unknown (0x9a2b), length 50:
-         0x0000:  2c2d 2e2f 0000 0000 9a2b 2c2d 2e2f 0000
-         0x0010:  0000 0000 0000 0000 0000 0000 0000 0000
-         0x0020:  0000 0000
-
-  with the fix:
-
-    ARP, Reverse Request who-is 9a:2b:2c:2d:2e:2f tell 9a:2b:2c:2d:2e:2f, length 46
-         0x0000:  0001 0800 0604 0003 9a2b 2c2d 2e2f 0000
-         0x0010:  0000 9a2b 2c2d 2e2f 0000 0000 0000 0000
-         0x0020:  0000 0000 0000 0000 0000 0000 0000
-
-Fixes: 481c52320a26 ("net: Strip virtio-net header when dumping")
-Cc: akihiko.odaki@daynix.com
-Signed-off-by: Laurent Vivier <lvivier@redhat.com>
-Reviewed-by: Akihiko Odaki <akihiko.odaki@daynix.com>
-Acked-by: Jason Wang <jasowang@redhat.com>
-Reviewed-by: Michael Tokarev <mjt@tls.msk.ru>
-(picked from https://lore.kernel.org/qemu-devel/20250117111709.970789-3-lvivier@redhat.com/)
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
----
- net/dump.c | 3 ++-
- 1 file changed, 2 insertions(+), 1 deletion(-)
-
-diff --git a/net/dump.c b/net/dump.c
-index 956e34a123..42ab8d7716 100644
---- a/net/dump.c
-+++ b/net/dump.c
-@@ -155,7 +155,8 @@ static ssize_t filter_dump_receive_iov(NetFilterState *nf, NetClientState *sndr,
- {
-     NetFilterDumpState *nfds = FILTER_DUMP(nf);
- 
--    dump_receive_iov(&nfds->ds, iov, iovcnt, qemu_get_vnet_hdr_len(nf->netdev));
-+    dump_receive_iov(&nfds->ds, iov, iovcnt, flags & QEMU_NET_PACKET_FLAG_RAW ?
-+                     0 : qemu_get_vnet_hdr_len(nf->netdev));
-     return 0;
- }
- 

debian/patches/extra/0008-Revert-x86-acpi-workaround-Windows-not-handling-name.patch

@@ -0,0 +1,45 @@
+From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
+From: Fiona Ebner <f.ebner@proxmox.com>
+Date: Fri, 17 Nov 2023 11:18:06 +0100
+Subject: [PATCH] Revert "x86: acpi: workaround Windows not handling name
+ references in Package properly"
+
+This reverts commit 44d975ef340e2f21f236f9520c53e1b30d2213a4.
+
+As reported in the community forum [0] and reproduced locally this
+breaks VirtIO network adapters in (at least) the German ISO of Windows
+Server 2022. The fix itself was for
+
+> Issue is not fatal but as result acpi-index/"PCI Label ID" property
+> is either not shown in device details page or shows incorrect value.
+
+so revert and tolerate that as a stop-gap, rather than have the
+devices not working at all.
+
+[0]: https://forum.proxmox.com/threads/92094/post-605684
+
+Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
+---
+ hw/i386/acpi-build.c | 8 ++------
+ 1 file changed, 2 insertions(+), 6 deletions(-)
+
+diff --git a/hw/i386/acpi-build.c b/hw/i386/acpi-build.c
+index bb12b0ad43..de14d3c3da 100644
+--- a/hw/i386/acpi-build.c
++++ b/hw/i386/acpi-build.c
+@@ -362,13 +362,9 @@ Aml *aml_pci_device_dsm(void)
+     {
+         Aml *params = aml_local(0);
+         Aml *pkg = aml_package(2);
+-        aml_append(pkg, aml_int(0));
+-        aml_append(pkg, aml_int(0));
++        aml_append(pkg, aml_name("BSEL"));
++        aml_append(pkg, aml_name("ASUN"));
+         aml_append(method, aml_store(pkg, params));
+-        aml_append(method,
+-            aml_store(aml_name("BSEL"), aml_index(params, aml_int(0))));
+-        aml_append(method,
+-            aml_store(aml_name("ASUN"), aml_index(params, aml_int(1))));
+         aml_append(method,
+             aml_return(aml_call5("PDSM", aml_arg(0), aml_arg(1),
+                                  aml_arg(2), aml_arg(3), params))

debian/patches/extra/0008-pci-acpi-Windows-PCI-Label-Id-bug-workaround.patch

@@ -1,96 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Igor Mammedov <imammedo@redhat.com>
-Date: Wed, 15 Jan 2025 13:53:41 +0100
-Subject: [PATCH] pci: acpi: Windows 'PCI Label Id' bug workaround
-
-Current versions of Windows call _DSM(func=7) regardless
-of whether it is supported or not. It leads to NICs having bogus
-'PCI Label Id = 0', where none should be set at all.
-
-Also presence of 'PCI Label Id' triggers another Windows bug
-on localized versions that leads to hangs. The later bug is fixed
-in latest updates for 'Windows Server' but not in consumer
-versions of Windows (and there is no plans to fix it
-as far as I'm aware).
-
-Given it's easy, implement Microsoft suggested workaround
-(return invalid Package) so that affected Windows versions
-could boot on QEMU.
-This would effectvely remove bogus 'PCI Label Id's on NICs,
-but MS teem confirmed that flipping 'PCI Label Id' should not
-change 'Network Connection' ennumeration, so it should be safe
-for QEMU to change _DSM without any compat code.
-
-Smoke tested with WinXP and WS2022
-Resolves: https://gitlab.com/qemu-project/qemu/-/issues/774
-Signed-off-by: Igor Mammedov <imammedo@redhat.com>
-Message-Id: <20250115125342.3883374-3-imammedo@redhat.com>
-Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
-Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
-(cherry picked from commit 0b053391985abcc40b16ac8fc4a7f6588d1d95c1)
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
----
- hw/i386/acpi-build.c | 33 +++++++++++++++++++++++----------
- 1 file changed, 23 insertions(+), 10 deletions(-)
-
-diff --git a/hw/i386/acpi-build.c b/hw/i386/acpi-build.c
-index 9fcc2897b8..f7b961e04c 100644
---- a/hw/i386/acpi-build.c
-+++ b/hw/i386/acpi-build.c
-@@ -654,6 +654,7 @@ static Aml *aml_pci_pdsm(void)
-     Aml *acpi_index = aml_local(2);
-     Aml *zero = aml_int(0);
-     Aml *one = aml_int(1);
-+    Aml *not_supp = aml_int(0xFFFFFFFF);
-     Aml *func = aml_arg(2);
-     Aml *params = aml_arg(4);
-     Aml *bnum = aml_derefof(aml_index(params, aml_int(0)));
-@@ -678,7 +679,7 @@ static Aml *aml_pci_pdsm(void)
-          */
-         ifctx1 = aml_if(aml_lnot(
-                      aml_or(aml_equal(acpi_index, zero),
--                            aml_equal(acpi_index, aml_int(0xFFFFFFFF)), NULL)
-+                            aml_equal(acpi_index, not_supp), NULL)
-                  ));
-         {
-             /* have supported functions */
-@@ -704,18 +705,30 @@ static Aml *aml_pci_pdsm(void)
-     {
-        Aml *pkg = aml_package(2);
- 
--       aml_append(pkg, zero);
--       /*
--        * optional, if not impl. should return null string
--        */
--       aml_append(pkg, aml_string("%s", ""));
--       aml_append(ifctx, aml_store(pkg, ret));
--
-        aml_append(ifctx, aml_store(aml_call2("AIDX", bnum, sunum), acpi_index));
-+       aml_append(ifctx, aml_store(pkg, ret));
-        /*
--        * update acpi-index to actual value
-+        * Windows calls func=7 without checking if it's available,
-+        * as workaround Microsoft has suggested to return invalid for func7
-+        * Package, so return 2 elements package but only initialize elements
-+        * when acpi_index is supported and leave them uninitialized, which
-+        * leads elements to being Uninitialized ObjectType and should trip
-+        * Windows into discarding result as an unexpected and prevent setting
-+        * bogus 'PCI Label' on the device.
-         */
--       aml_append(ifctx, aml_store(acpi_index, aml_index(ret, zero)));
-+       ifctx1 = aml_if(aml_lnot(aml_lor(
-+                    aml_equal(acpi_index, zero), aml_equal(acpi_index, not_supp)
-+                )));
-+       {
-+           aml_append(ifctx1, aml_store(acpi_index, aml_index(ret, zero)));
-+           /*
-+            * optional, if not impl. should return null string
-+            */
-+           aml_append(ifctx1, aml_store(aml_string("%s", ""),
-+                                        aml_index(ret, one)));
-+       }
-+       aml_append(ifctx, ifctx1);
-+
-        aml_append(ifctx, aml_return(ret));
-     }
- 

debian/patches/extra/0009-hw-ide-ahci-fix-legacy-software-reset.patch

@@ -0,0 +1,107 @@
+From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
+From: Niklas Cassel <niklas.cassel@wdc.com>
+Date: Wed, 8 Nov 2023 23:26:57 +0100
+Subject: [PATCH] hw/ide/ahci: fix legacy software reset
+
+Legacy software contains a standard mechanism for generating a reset to a
+Serial ATA device - setting the SRST (software reset) bit in the Device
+Control register.
+
+Serial ATA has a more robust mechanism called COMRESET, also referred to
+as port reset. A port reset is the preferred mechanism for error
+recovery and should be used in place of software reset.
+
+Commit e2a5d9b3d9c3 ("hw/ide/ahci: simplify and document PxCI handling")
+improved the handling of PxCI, such that PxCI gets cleared after handling
+a non-NCQ, or NCQ command (instead of incorrectly clearing PxCI after
+receiving anything - even a FIS that failed to parse, which should NOT
+clear PxCI, so that you can see which command slot that caused an error).
+
+However, simply clearing PxCI after a non-NCQ, or NCQ command, is not
+enough, we also need to clear PxCI when receiving a SRST in the Device
+Control register.
+
+A legacy software reset is performed by the host sending two H2D FISes,
+the first H2D FIS asserts SRST, and the second H2D FIS deasserts SRST.
+
+The first H2D FIS will not get a D2H reply, and requires the FIS to have
+the C bit set to one, such that the HBA itself will clear the bit in PxCI.
+
+The second H2D FIS will get a D2H reply once the diagnostic is completed.
+The clearing of the bit in PxCI for this command should ideally be done
+in ahci_init_d2h() (if it was a legacy software reset that caused the
+reset (a COMRESET does not use a command slot)). However, since the reset
+value for PxCI is 0, modify ahci_reset_port() to actually clear PxCI to 0,
+that way we can avoid complex logic in ahci_init_d2h().
+
+This fixes an issue for FreeBSD where the device would fail to reset.
+The problem was not noticed in Linux, because Linux uses a COMRESET
+instead of a legacy software reset by default.
+
+Fixes: e2a5d9b3d9c3 ("hw/ide/ahci: simplify and document PxCI handling")
+Reported-by: Marcin Juszkiewicz <marcin.juszkiewicz@linaro.org>
+Signed-off-by: Niklas Cassel <niklas.cassel@wdc.com>
+(picked from https://lists.nongnu.org/archive/html/qemu-devel/2023-11/msg02277.html)
+Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
+---
+ hw/ide/ahci.c | 27 ++++++++++++++++++++++++++-
+ 1 file changed, 26 insertions(+), 1 deletion(-)
+
+diff --git a/hw/ide/ahci.c b/hw/ide/ahci.c
+index d0a774bc17..1718b7e902 100644
+--- a/hw/ide/ahci.c
++++ b/hw/ide/ahci.c
+@@ -623,9 +623,13 @@ static void ahci_init_d2h(AHCIDevice *ad)
+         return;
+     }
+ 
++    /*
++     * For simplicity, do not call ahci_clear_cmd_issue() for this
++     * ahci_write_fis_d2h(). (The reset value for PxCI is 0.)
++     */
+     if (ahci_write_fis_d2h(ad, true)) {
+         ad->init_d2h_sent = true;
+-        /* We're emulating receiving the first Reg H2D Fis from the device;
++        /* We're emulating receiving the first Reg D2H FIS from the device;
+          * Update the SIG register, but otherwise proceed as normal. */
+         pr->sig = ((uint32_t)ide_state->hcyl << 24) |
+             (ide_state->lcyl << 16) |
+@@ -663,6 +667,7 @@ static void ahci_reset_port(AHCIState *s, int port)
+     pr->scr_act = 0;
+     pr->tfdata = 0x7F;
+     pr->sig = 0xFFFFFFFF;
++    pr->cmd_issue = 0;
+     d->busy_slot = -1;
+     d->init_d2h_sent = false;
+ 
+@@ -1243,10 +1248,30 @@ static void handle_reg_h2d_fis(AHCIState *s, int port,
+         case STATE_RUN:
+             if (cmd_fis[15] & ATA_SRST) {
+                 s->dev[port].port_state = STATE_RESET;
++                /*
++                 * When setting SRST in the first H2D FIS in the reset sequence,
++                 * the device does not send a D2H FIS. Host software thus has to
++                 * set the "Clear Busy upon R_OK" bit such that PxCI (and BUSY)
++                 * gets cleared. See AHCI 1.3.1, section 10.4.1 Software Reset.
++                 */
++                if (opts & AHCI_CMD_CLR_BUSY) {
++                    ahci_clear_cmd_issue(ad, slot);
++                }
+             }
+             break;
+         case STATE_RESET:
+             if (!(cmd_fis[15] & ATA_SRST)) {
++                /*
++                 * When clearing SRST in the second H2D FIS in the reset
++                 * sequence, the device will execute diagnostics. When this is
++                 * done, the device will send a D2H FIS with the good status.
++                 * See SATA 3.5a Gold, section 11.4 Software reset protocol.
++                 *
++                 * This D2H FIS is the first D2H FIS received from the device,
++                 * and is received regardless if the reset was performed by a
++                 * COMRESET or by setting and clearing the SRST bit. Therefore,
++                 * the logic for this is found in ahci_init_d2h() and not here.
++                 */
+                 ahci_reset_port(s, port);
+             }
+             break;

debian/patches/extra/0009-hw-usb-hcd-xhci-pci-Use-modulo-to-select-MSI-vector-.patch

@@ -1,53 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Phil Dennis-Jordan <phil@philjordan.eu>
-Date: Fri, 13 Dec 2024 17:06:14 +0100
-Subject: [PATCH] hw/usb/hcd-xhci-pci: Use modulo to select MSI vector as per
- spec
-MIME-Version: 1.0
-Content-Type: text/plain; charset=UTF-8
-Content-Transfer-Encoding: 8bit
-
-QEMU would crash with a failed assertion if the XHCI controller
-attempted to raise the interrupt on an interrupter corresponding
-to a MSI vector with a higher index than the highest configured
-for the device by the guest driver.
-
-This behaviour is correct on the MSI/PCI side: per PCI 3.0 spec,
-devices must ensure they do not send MSI notifications for
-vectors beyond the range of those allocated by the system/driver
-software. Unlike MSI-X, there is no generic way for handling
-aliasing in the case of fewer allocated vectors than requested,
-so the specifics are up to device implementors. (Section
-6.8.3.4. "Sending Messages")
-
-It turns out the XHCI spec (Implementation Note in section 4.17,
-"Interrupters") requires that the host controller signal the MSI
-vector with the number computed by taking the interrupter number
-modulo the number of enabled MSI vectors.
-
-This change introduces that modulo calculation, fixing the
-failed assertion. This makes the device work correctly in MSI mode
-with macOS's XHCI driver, which only allocates a single vector.
-
-Signed-off-by: Phil Dennis-Jordan <phil@philjordan.eu>
-Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
-Message-ID: <20250112210056.16658-2-phil@philjordan.eu>
-Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
-(cherry picked from commit bb5b7fced6b5d3334ab20702fc846e47bb1fb731)
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
----
- hw/usb/hcd-xhci-pci.c | 1 +
- 1 file changed, 1 insertion(+)
-
-diff --git a/hw/usb/hcd-xhci-pci.c b/hw/usb/hcd-xhci-pci.c
-index a039f5778a..516e6909d2 100644
---- a/hw/usb/hcd-xhci-pci.c
-+++ b/hw/usb/hcd-xhci-pci.c
-@@ -74,6 +74,7 @@ static bool xhci_pci_intr_raise(XHCIState *xhci, int n, bool level)
-     }
- 
-     if (msi_enabled(pci_dev) && level) {
-+        n %= msi_nr_vectors_allocated(pci_dev);
-         msi_notify(pci_dev, n);
-         return true;
-     }

debian/patches/extra/0010-pci-ensure-valid-link-status-bits-for-downstream-por.patch

@@ -1,63 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Sebastian Ott <sebott@redhat.com>
-Date: Tue, 3 Dec 2024 13:19:28 +0100
-Subject: [PATCH] pci: ensure valid link status bits for downstream ports
-
-PCI hotplug for downstream endpoints on arm fails because Linux'
-PCIe hotplug driver doesn't like the QEMU provided LNKSTA:
-
-  pcieport 0000:08:01.0: pciehp: Slot(2): Card present
-  pcieport 0000:08:01.0: pciehp: Slot(2): Link Up
-  pcieport 0000:08:01.0: pciehp: Slot(2): Cannot train link: status 0x2000
-
-There's 2 cases where LNKSTA isn't setup properly:
-* the downstream device has no express capability
-* max link width of the bridge is 0
-
-Move the sanity checks added via 88c869198aa63
-("pci: Sanity test minimum downstream LNKSTA") outside of the
-branch to make sure downstream ports always have a valid LNKSTA.
-
-Signed-off-by: Sebastian Ott <sebott@redhat.com>
-Tested-by: Zhenyu Zhang <zhenyzha@redhat.com>
-Message-Id: <20241203121928.14861-1-sebott@redhat.com>
-Reviewed-by: Alex Williamson <alex.williamson@redhat.com>
-Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
-Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
-(cherry picked from commit 694632fd44987cc4618612a38ad151047524a590)
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
----
- hw/pci/pcie.c | 12 ++++++++----
- 1 file changed, 8 insertions(+), 4 deletions(-)
-
-diff --git a/hw/pci/pcie.c b/hw/pci/pcie.c
-index 0b455c8654..1b12db6fa2 100644
---- a/hw/pci/pcie.c
-+++ b/hw/pci/pcie.c
-@@ -1113,18 +1113,22 @@ void pcie_sync_bridge_lnk(PCIDevice *bridge_dev)
-         if ((lnksta & PCI_EXP_LNKSTA_NLW) > (lnkcap & PCI_EXP_LNKCAP_MLW)) {
-             lnksta &= ~PCI_EXP_LNKSTA_NLW;
-             lnksta |= lnkcap & PCI_EXP_LNKCAP_MLW;
--        } else if (!(lnksta & PCI_EXP_LNKSTA_NLW)) {
--            lnksta |= QEMU_PCI_EXP_LNKSTA_NLW(QEMU_PCI_EXP_LNK_X1);
-         }
- 
-         if ((lnksta & PCI_EXP_LNKSTA_CLS) > (lnkcap & PCI_EXP_LNKCAP_SLS)) {
-             lnksta &= ~PCI_EXP_LNKSTA_CLS;
-             lnksta |= lnkcap & PCI_EXP_LNKCAP_SLS;
--        } else if (!(lnksta & PCI_EXP_LNKSTA_CLS)) {
--            lnksta |= QEMU_PCI_EXP_LNKSTA_CLS(QEMU_PCI_EXP_LNK_2_5GT);
-         }
-     }
- 
-+    if (!(lnksta & PCI_EXP_LNKSTA_NLW)) {
-+        lnksta |= QEMU_PCI_EXP_LNKSTA_NLW(QEMU_PCI_EXP_LNK_X1);
-+    }
-+
-+    if (!(lnksta & PCI_EXP_LNKSTA_CLS)) {
-+        lnksta |= QEMU_PCI_EXP_LNKSTA_CLS(QEMU_PCI_EXP_LNK_2_5GT);
-+    }
-+
-     pci_word_test_and_clear_mask(exp_cap + PCI_EXP_LNKSTA,
-                                  PCI_EXP_LNKSTA_CLS | PCI_EXP_LNKSTA_NLW);
-     pci_word_test_and_set_mask(exp_cap + PCI_EXP_LNKSTA, lnksta &

debian/patches/extra/0010-ui-vnc-clipboard-fix-inflate_buffer.patch

@@ -0,0 +1,34 @@
+From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
+From: Fiona Ebner <f.ebner@proxmox.com>
+Date: Wed, 22 Nov 2023 13:17:25 +0100
+Subject: [PATCH] ui/vnc-clipboard: fix inflate_buffer
+
+Commit d921fea338 ("ui/vnc-clipboard: fix infinite loop in
+inflate_buffer (CVE-2023-3255)") removed this hunk, but it is still
+required, because it can happen that stream.avail_in becomes zero
+before coming across a return value of Z_STREAM_END.
+
+This fixes the host->guest direction with noNVC.
+
+Reported-by: Friedrich Weber <f.weber@proxmox.com>
+Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
+---
+ ui/vnc-clipboard.c | 5 +++++
+ 1 file changed, 5 insertions(+)
+
+diff --git a/ui/vnc-clipboard.c b/ui/vnc-clipboard.c
+index c759be3438..124b6fbd9c 100644
+--- a/ui/vnc-clipboard.c
++++ b/ui/vnc-clipboard.c
+@@ -69,6 +69,11 @@ static uint8_t *inflate_buffer(uint8_t *in, uint32_t in_len, uint32_t *size)
+         }
+     }
+ 
++    *size = stream.total_out;
++    inflateEnd(&stream);
++
++    return out;
++
+ err_end:
+     inflateEnd(&stream);
+ err:

debian/patches/extra/0011-pci-msix-Fix-msix-pba-read-vector-poll-end-calculati.patch

@@ -1,36 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Nicholas Piggin <npiggin@gmail.com>
-Date: Thu, 12 Dec 2024 22:04:02 +1000
-Subject: [PATCH] pci/msix: Fix msix pba read vector poll end calculation
-MIME-Version: 1.0
-Content-Type: text/plain; charset=UTF-8
-Content-Transfer-Encoding: 8bit
-
-The end vector calculation has a bug that results in polling fewer
-than required vectors when reading at a non-zero offset in PBA memory.
-
-Fixes: bbef882cc193 ("msi: add API to get notified about pending bit poll")
-Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
-Message-Id: <20241212120402.1475053-1-npiggin@gmail.com>
-Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
-Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
-Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
-(cherry picked from commit 42e2a7a0ab23784e44fcb18369e06067abc89305)
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
----
- hw/pci/msix.c | 2 +-
- 1 file changed, 1 insertion(+), 1 deletion(-)
-
-diff --git a/hw/pci/msix.c b/hw/pci/msix.c
-index 487e49834e..cc6e79ec67 100644
---- a/hw/pci/msix.c
-+++ b/hw/pci/msix.c
-@@ -250,7 +250,7 @@ static uint64_t msix_pba_mmio_read(void *opaque, hwaddr addr,
-     PCIDevice *dev = opaque;
-     if (dev->msix_vector_poll_notifier) {
-         unsigned vector_start = addr * 8;
--        unsigned vector_end = MIN(addr + size * 8, dev->msix_entries_nr);
-+        unsigned vector_end = MIN((addr + size) * 8, dev->msix_entries_nr);
-         dev->msix_vector_poll_notifier(dev, vector_start, vector_end);
-     }
- 

debian/patches/pve/0001-PVE-Config-block-file-change-locking-default-to-off.patch

@@ -14,7 +14,7 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  1 file changed, 2 insertions(+), 2 deletions(-)
 
 diff --git a/block/file-posix.c b/block/file-posix.c
-index 90fa54352c..e2ea071315 100644
+index aa89789737..0db366a851 100644
 --- a/block/file-posix.c
 +++ b/block/file-posix.c
 @@ -564,7 +564,7 @@ static QemuOptsList raw_runtime_opts = {

debian/patches/pve/0002-PVE-Config-Adjust-network-script-path-to-etc-kvm.patch

@@ -9,12 +9,12 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  1 file changed, 2 insertions(+), 2 deletions(-)
 
 diff --git a/include/net/net.h b/include/net/net.h
-index cdd5b109b0..653a37e9d1 100644
+index 1448d00afb..d1601d32c1 100644
 --- a/include/net/net.h
 +++ b/include/net/net.h
-@@ -305,8 +305,8 @@ void netdev_add(QemuOpts *opts, Error **errp);
- 
+@@ -258,8 +258,8 @@ void netdev_add(QemuOpts *opts, Error **errp);
  int net_hub_id_for_client(NetClientState *nc, int *id);
+ NetClientState *net_hub_port_find(int hub_id);
  
 -#define DEFAULT_NETWORK_SCRIPT CONFIG_SYSCONFDIR "/qemu-ifup"
 -#define DEFAULT_NETWORK_DOWN_SCRIPT CONFIG_SYSCONFDIR "/qemu-ifdown"

debian/patches/pve/0003-PVE-Config-set-the-CPU-model-to-kvm64-32-instead-of-.patch

@@ -10,10 +10,10 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  1 file changed, 2 insertions(+), 2 deletions(-)
 
 diff --git a/target/i386/cpu.h b/target/i386/cpu.h
-index 4c239a6970..be09263fb0 100644
+index e0771a1043..1018ccc0b8 100644
 --- a/target/i386/cpu.h
 +++ b/target/i386/cpu.h
-@@ -2475,9 +2475,9 @@ uint64_t cpu_get_tsc(CPUX86State *env);
+@@ -2243,9 +2243,9 @@ uint64_t cpu_get_tsc(CPUX86State *env);
  #define CPU_RESOLVING_TYPE TYPE_X86_CPU
  
  #ifdef TARGET_X86_64

debian/patches/pve/0004-PVE-Config-ui-spice-default-to-pve-certificates.patch

@@ -9,10 +9,10 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  1 file changed, 9 insertions(+), 6 deletions(-)
 
 diff --git a/ui/spice-core.c b/ui/spice-core.c
-index bd9dbe03f1..a7ecaad9c7 100644
+index 52a59386d7..b20c25aee0 100644
 --- a/ui/spice-core.c
 +++ b/ui/spice-core.c
-@@ -690,32 +690,35 @@ static void qemu_spice_init(void)
+@@ -691,32 +691,35 @@ static void qemu_spice_init(void)
  
      if (tls_port) {
          x509_dir = qemu_opt_get(opts, "x509-dir");

debian/patches/pve/0005-PVE-Config-glusterfs-no-default-logfile-if-daemonize.patch

@@ -9,10 +9,10 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  1 file changed, 11 insertions(+), 4 deletions(-)
 
 diff --git a/block/gluster.c b/block/gluster.c
-index e9c038042b..c8457a5014 100644
+index ad5fadbe79..d0011085c4 100644
 --- a/block/gluster.c
 +++ b/block/gluster.c
-@@ -42,7 +42,7 @@
+@@ -43,7 +43,7 @@
  #define GLUSTER_DEBUG_DEFAULT       4
  #define GLUSTER_DEBUG_MAX           9
  #define GLUSTER_OPT_LOGFILE         "logfile"
@@ -21,7 +21,7 @@ index e9c038042b..c8457a5014 100644
  /*
   * Several versions of GlusterFS (3.12? -> 6.0.1) fail when the transfer size
   * is greater or equal to 1024 MiB, so we are limiting the transfer size to 512
-@@ -421,6 +421,7 @@ static struct glfs *qemu_gluster_glfs_init(BlockdevOptionsGluster *gconf,
+@@ -425,6 +425,7 @@ static struct glfs *qemu_gluster_glfs_init(BlockdevOptionsGluster *gconf,
      int old_errno;
      SocketAddressList *server;
      uint64_t port;
@@ -29,7 +29,7 @@ index e9c038042b..c8457a5014 100644
  
      glfs = glfs_find_preopened(gconf->volume);
      if (glfs) {
-@@ -463,9 +464,15 @@ static struct glfs *qemu_gluster_glfs_init(BlockdevOptionsGluster *gconf,
+@@ -467,9 +468,15 @@ static struct glfs *qemu_gluster_glfs_init(BlockdevOptionsGluster *gconf,
          }
      }
  

debian/patches/pve/0006-PVE-Config-rbd-block-rbd-disable-rbd_cache_writethro.patch

@@ -18,7 +18,7 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  1 file changed, 2 insertions(+)
 
 diff --git a/block/rbd.c b/block/rbd.c
-index 04ed0e242e..728bce3b1e 100644
+index 978671411e..a4749f3b1b 100644
 --- a/block/rbd.c
 +++ b/block/rbd.c
 @@ -963,6 +963,8 @@ static int qemu_rbd_connect(rados_t *cluster, rados_ioctx_t *io_ctx,

debian/patches/pve/0007-PVE-Up-glusterfs-allow-partial-reads.patch

@@ -16,10 +16,10 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  1 file changed, 9 insertions(+), 1 deletion(-)
 
 diff --git a/block/gluster.c b/block/gluster.c
-index c8457a5014..c3a9555591 100644
+index d0011085c4..2df3d6e35d 100644
 --- a/block/gluster.c
 +++ b/block/gluster.c
-@@ -57,6 +57,7 @@ typedef struct GlusterAIOCB {
+@@ -58,6 +58,7 @@ typedef struct GlusterAIOCB {
      int ret;
      Coroutine *coroutine;
      AioContext *aio_context;
@@ -27,7 +27,7 @@ index c8457a5014..c3a9555591 100644
  } GlusterAIOCB;
  
  typedef struct BDRVGlusterState {
-@@ -746,8 +747,10 @@ static void gluster_finish_aiocb(struct glfs_fd *fd, ssize_t ret,
+@@ -753,8 +754,10 @@ static void gluster_finish_aiocb(struct glfs_fd *fd, ssize_t ret,
          acb->ret = 0; /* Success */
      } else if (ret < 0) {
          acb->ret = -errno; /* Read/Write failed */
@@ -39,7 +39,7 @@ index c8457a5014..c3a9555591 100644
      }
  
      aio_co_schedule(acb->aio_context, acb->coroutine);
-@@ -1018,6 +1021,7 @@ static coroutine_fn int qemu_gluster_co_pwrite_zeroes(BlockDriverState *bs,
+@@ -1021,6 +1024,7 @@ static coroutine_fn int qemu_gluster_co_pwrite_zeroes(BlockDriverState *bs,
      acb.ret = 0;
      acb.coroutine = qemu_coroutine_self();
      acb.aio_context = bdrv_get_aio_context(bs);
@@ -47,7 +47,7 @@ index c8457a5014..c3a9555591 100644
  
      ret = glfs_zerofill_async(s->fd, offset, bytes, gluster_finish_aiocb, &acb);
      if (ret < 0) {
-@@ -1198,9 +1202,11 @@ static coroutine_fn int qemu_gluster_co_rw(BlockDriverState *bs,
+@@ -1201,9 +1205,11 @@ static coroutine_fn int qemu_gluster_co_rw(BlockDriverState *bs,
      acb.aio_context = bdrv_get_aio_context(bs);
  
      if (write) {
@@ -59,7 +59,7 @@ index c8457a5014..c3a9555591 100644
          ret = glfs_preadv_async(s->fd, qiov->iov, qiov->niov, offset, 0,
                                  gluster_finish_aiocb, &acb);
      }
-@@ -1263,6 +1269,7 @@ static coroutine_fn int qemu_gluster_co_flush_to_disk(BlockDriverState *bs)
+@@ -1266,6 +1272,7 @@ static coroutine_fn int qemu_gluster_co_flush_to_disk(BlockDriverState *bs)
      acb.ret = 0;
      acb.coroutine = qemu_coroutine_self();
      acb.aio_context = bdrv_get_aio_context(bs);
@@ -67,7 +67,7 @@ index c8457a5014..c3a9555591 100644
  
      ret = glfs_fsync_async(s->fd, gluster_finish_aiocb, &acb);
      if (ret < 0) {
-@@ -1311,6 +1318,7 @@ static coroutine_fn int qemu_gluster_co_pdiscard(BlockDriverState *bs,
+@@ -1314,6 +1321,7 @@ static coroutine_fn int qemu_gluster_co_pdiscard(BlockDriverState *bs,
      acb.ret = 0;
      acb.coroutine = qemu_coroutine_self();
      acb.aio_context = bdrv_get_aio_context(bs);

debian/patches/pve/0008-PVE-Up-qemu-img-return-success-on-info-without-snaps.patch

@@ -9,10 +9,10 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  1 file changed, 2 insertions(+), 1 deletion(-)
 
 diff --git a/qemu-img.c b/qemu-img.c
-index 7668f86769..2575e97b43 100644
+index 27f48051b0..bb287d8538 100644
 --- a/qemu-img.c
 +++ b/qemu-img.c
-@@ -3075,7 +3075,8 @@ static int img_info(int argc, char **argv)
+@@ -3062,7 +3062,8 @@ static int img_info(int argc, char **argv)
      list = collect_image_info_list(image_opts, filename, fmt, chain,
                                     force_share);
      if (!list) {

debian/patches/pve/0009-PVE-Up-qemu-img-dd-add-osize-and-read-from-to-stdin-.patch

@@ -38,10 +38,10 @@ Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
  2 files changed, 133 insertions(+), 73 deletions(-)
 
 diff --git a/qemu-img-cmds.hx b/qemu-img-cmds.hx
-index c9dd70a892..048788b23d 100644
+index 1b1dab5b17..d1616c045a 100644
 --- a/qemu-img-cmds.hx
 +++ b/qemu-img-cmds.hx
-@@ -60,9 +60,9 @@ SRST
+@@ -58,9 +58,9 @@ SRST
  ERST
  
  DEF("dd", img_dd,
@@ -54,10 +54,10 @@ index c9dd70a892..048788b23d 100644
  
  DEF("info", img_info,
 diff --git a/qemu-img.c b/qemu-img.c
-index 2575e97b43..8ec68b346f 100644
+index bb287d8538..09c0340d16 100644
 --- a/qemu-img.c
 +++ b/qemu-img.c
-@@ -4993,10 +4993,12 @@ static int img_bitmap(int argc, char **argv)
+@@ -4888,10 +4888,12 @@ static int img_bitmap(int argc, char **argv)
  #define C_IF      04
  #define C_OF      010
  #define C_SKIP    020
@@ -70,7 +70,7 @@ index 2575e97b43..8ec68b346f 100644
  };
  
  struct DdIo {
-@@ -5072,6 +5074,19 @@ static int img_dd_skip(const char *arg,
+@@ -4967,6 +4969,19 @@ static int img_dd_skip(const char *arg,
      return 0;
  }
  
@@ -90,7 +90,7 @@ index 2575e97b43..8ec68b346f 100644
  static int img_dd(int argc, char **argv)
  {
      int ret = 0;
-@@ -5112,6 +5127,7 @@ static int img_dd(int argc, char **argv)
+@@ -5007,6 +5022,7 @@ static int img_dd(int argc, char **argv)
          { "if", img_dd_if, C_IF },
          { "of", img_dd_of, C_OF },
          { "skip", img_dd_skip, C_SKIP },
@@ -98,7 +98,7 @@ index 2575e97b43..8ec68b346f 100644
          { NULL, NULL, 0 }
      };
      const struct option long_options[] = {
-@@ -5187,91 +5203,112 @@ static int img_dd(int argc, char **argv)
+@@ -5082,91 +5098,112 @@ static int img_dd(int argc, char **argv)
          arg = NULL;
      }
  
@@ -275,7 +275,7 @@ index 2575e97b43..8ec68b346f 100644
      }
  
      if (dd.flags & C_SKIP && (in.offset > INT64_MAX / in.bsz ||
-@@ -5288,20 +5325,43 @@ static int img_dd(int argc, char **argv)
+@@ -5183,20 +5220,43 @@ static int img_dd(int argc, char **argv)
      in.buf = g_new(uint8_t, in.bsz);
  
      for (out_pos = 0; in_pos < size; ) {

debian/patches/pve/0010-PVE-Up-qemu-img-dd-add-isize-parameter.patch

@@ -16,10 +16,10 @@ Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
  1 file changed, 25 insertions(+), 3 deletions(-)
 
 diff --git a/qemu-img.c b/qemu-img.c
-index 8ec68b346f..b98184bba1 100644
+index 09c0340d16..556535d9d5 100644
 --- a/qemu-img.c
 +++ b/qemu-img.c
-@@ -4994,11 +4994,13 @@ static int img_bitmap(int argc, char **argv)
+@@ -4889,11 +4889,13 @@ static int img_bitmap(int argc, char **argv)
  #define C_OF      010
  #define C_SKIP    020
  #define C_OSIZE   040
@@ -33,7 +33,7 @@ index 8ec68b346f..b98184bba1 100644
  };
  
  struct DdIo {
-@@ -5087,6 +5089,19 @@ static int img_dd_osize(const char *arg,
+@@ -4982,6 +4984,19 @@ static int img_dd_osize(const char *arg,
      return 0;
  }
  
@@ -53,7 +53,7 @@ index 8ec68b346f..b98184bba1 100644
  static int img_dd(int argc, char **argv)
  {
      int ret = 0;
-@@ -5101,12 +5116,14 @@ static int img_dd(int argc, char **argv)
+@@ -4996,12 +5011,14 @@ static int img_dd(int argc, char **argv)
      int c, i;
      const char *out_fmt = "raw";
      const char *fmt = NULL;
@@ -69,7 +69,7 @@ index 8ec68b346f..b98184bba1 100644
      };
      struct DdIo in = {
          .bsz = 512, /* Block size is by default 512 bytes */
-@@ -5128,6 +5145,7 @@ static int img_dd(int argc, char **argv)
+@@ -5023,6 +5040,7 @@ static int img_dd(int argc, char **argv)
          { "of", img_dd_of, C_OF },
          { "skip", img_dd_skip, C_SKIP },
          { "osize", img_dd_osize, C_OSIZE },
@@ -77,7 +77,7 @@ index 8ec68b346f..b98184bba1 100644
          { NULL, NULL, 0 }
      };
      const struct option long_options[] = {
-@@ -5324,9 +5342,10 @@ static int img_dd(int argc, char **argv)
+@@ -5219,9 +5237,10 @@ static int img_dd(int argc, char **argv)
  
      in.buf = g_new(uint8_t, in.bsz);
  
@@ -90,7 +90,7 @@ index 8ec68b346f..b98184bba1 100644
          if (blk1) {
              in_ret = blk_pread(blk1, in_pos, bytes, in.buf, 0);
              if (in_ret == 0) {
-@@ -5335,6 +5354,9 @@ static int img_dd(int argc, char **argv)
+@@ -5230,6 +5249,9 @@ static int img_dd(int argc, char **argv)
          } else {
              in_ret = read(STDIN_FILENO, in.buf, bytes);
              if (in_ret == 0) {

debian/patches/pve/0011-PVE-Up-qemu-img-dd-add-n-skip_create.patch

@@ -13,10 +13,10 @@ Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
  3 files changed, 26 insertions(+), 12 deletions(-)
 
 diff --git a/docs/tools/qemu-img.rst b/docs/tools/qemu-img.rst
-index 3653adb963..d83e8fb3c0 100644
+index 15aeddc6d8..5e713e231d 100644
 --- a/docs/tools/qemu-img.rst
 +++ b/docs/tools/qemu-img.rst
-@@ -212,6 +212,10 @@ Parameters to convert subcommand:
+@@ -208,6 +208,10 @@ Parameters to convert subcommand:
  
  Parameters to dd subcommand:
  
@@ -27,7 +27,7 @@ index 3653adb963..d83e8fb3c0 100644
  .. program:: qemu-img-dd
  
  .. option:: bs=BLOCK_SIZE
-@@ -492,7 +496,7 @@ Command description:
+@@ -488,7 +492,7 @@ Command description:
    it doesn't need to be specified separately in this case.
  
  
@@ -36,7 +36,7 @@ index 3653adb963..d83e8fb3c0 100644
  
    dd copies from *INPUT* file to *OUTPUT* file converting it from
    *FMT* format to *OUTPUT_FMT* format.
-@@ -503,6 +507,11 @@ Command description:
+@@ -499,6 +503,11 @@ Command description:
  
    The size syntax is similar to :manpage:`dd(1)`'s size syntax.
  
@@ -49,10 +49,10 @@ index 3653adb963..d83e8fb3c0 100644
  
    Give information about the disk image *FILENAME*. Use it in
 diff --git a/qemu-img-cmds.hx b/qemu-img-cmds.hx
-index 048788b23d..0b29a67a06 100644
+index d1616c045a..b5b0bb4467 100644
 --- a/qemu-img-cmds.hx
 +++ b/qemu-img-cmds.hx
-@@ -60,9 +60,9 @@ SRST
+@@ -58,9 +58,9 @@ SRST
  ERST
  
  DEF("dd", img_dd,
@@ -65,10 +65,10 @@ index 048788b23d..0b29a67a06 100644
  
  DEF("info", img_info,
 diff --git a/qemu-img.c b/qemu-img.c
-index b98184bba1..6fc8384f64 100644
+index 556535d9d5..289c78febb 100644
 --- a/qemu-img.c
 +++ b/qemu-img.c
-@@ -5118,7 +5118,7 @@ static int img_dd(int argc, char **argv)
+@@ -5013,7 +5013,7 @@ static int img_dd(int argc, char **argv)
      const char *fmt = NULL;
      int64_t size = 0, readsize = 0;
      int64_t out_pos, in_pos;
@@ -77,7 +77,7 @@ index b98184bba1..6fc8384f64 100644
      struct DdInfo dd = {
          .flags = 0,
          .count = 0,
-@@ -5156,7 +5156,7 @@ static int img_dd(int argc, char **argv)
+@@ -5051,7 +5051,7 @@ static int img_dd(int argc, char **argv)
          { 0, 0, 0, 0 }
      };
  
@@ -86,7 +86,7 @@ index b98184bba1..6fc8384f64 100644
          if (c == EOF) {
              break;
          }
-@@ -5176,6 +5176,9 @@ static int img_dd(int argc, char **argv)
+@@ -5071,6 +5071,9 @@ static int img_dd(int argc, char **argv)
          case 'h':
              help();
              break;
@@ -96,7 +96,7 @@ index b98184bba1..6fc8384f64 100644
          case 'U':
              force_share = true;
              break;
-@@ -5306,13 +5309,15 @@ static int img_dd(int argc, char **argv)
+@@ -5201,13 +5204,15 @@ static int img_dd(int argc, char **argv)
                                  size - in.bsz * in.offset, &error_abort);
          }
  

debian/patches/pve/0012-qemu-img-dd-add-l-option-for-loading-a-snapshot.patch

@@ -12,10 +12,10 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  3 files changed, 36 insertions(+), 7 deletions(-)
 
 diff --git a/docs/tools/qemu-img.rst b/docs/tools/qemu-img.rst
-index d83e8fb3c0..61c6b21859 100644
+index 5e713e231d..9390d5e5cf 100644
 --- a/docs/tools/qemu-img.rst
 +++ b/docs/tools/qemu-img.rst
-@@ -496,10 +496,10 @@ Command description:
+@@ -492,10 +492,10 @@ Command description:
    it doesn't need to be specified separately in this case.
  
  
@@ -30,10 +30,10 @@ index d83e8fb3c0..61c6b21859 100644
    The data is by default read and written using blocks of 512 bytes but can be
    modified by specifying *BLOCK_SIZE*. If count=\ *BLOCKS* is specified
 diff --git a/qemu-img-cmds.hx b/qemu-img-cmds.hx
-index 0b29a67a06..758f397232 100644
+index b5b0bb4467..36f97e1f19 100644
 --- a/qemu-img-cmds.hx
 +++ b/qemu-img-cmds.hx
-@@ -60,9 +60,9 @@ SRST
+@@ -58,9 +58,9 @@ SRST
  ERST
  
  DEF("dd", img_dd,
@@ -46,10 +46,10 @@ index 0b29a67a06..758f397232 100644
  
  DEF("info", img_info,
 diff --git a/qemu-img.c b/qemu-img.c
-index 6fc8384f64..a6c88e0860 100644
+index 289c78febb..da543d05cb 100644
 --- a/qemu-img.c
 +++ b/qemu-img.c
-@@ -5110,6 +5110,7 @@ static int img_dd(int argc, char **argv)
+@@ -5005,6 +5005,7 @@ static int img_dd(int argc, char **argv)
      BlockDriver *drv = NULL, *proto_drv = NULL;
      BlockBackend *blk1 = NULL, *blk2 = NULL;
      QemuOpts *opts = NULL;
@@ -57,7 +57,7 @@ index 6fc8384f64..a6c88e0860 100644
      QemuOptsList *create_opts = NULL;
      Error *local_err = NULL;
      bool image_opts = false;
-@@ -5119,6 +5120,7 @@ static int img_dd(int argc, char **argv)
+@@ -5014,6 +5015,7 @@ static int img_dd(int argc, char **argv)
      int64_t size = 0, readsize = 0;
      int64_t out_pos, in_pos;
      bool force_share = false, skip_create = false;
@@ -65,7 +65,7 @@ index 6fc8384f64..a6c88e0860 100644
      struct DdInfo dd = {
          .flags = 0,
          .count = 0,
-@@ -5156,7 +5158,7 @@ static int img_dd(int argc, char **argv)
+@@ -5051,7 +5053,7 @@ static int img_dd(int argc, char **argv)
          { 0, 0, 0, 0 }
      };
  
@@ -74,7 +74,7 @@ index 6fc8384f64..a6c88e0860 100644
          if (c == EOF) {
              break;
          }
-@@ -5179,6 +5181,19 @@ static int img_dd(int argc, char **argv)
+@@ -5074,6 +5076,19 @@ static int img_dd(int argc, char **argv)
          case 'n':
              skip_create = true;
              break;
@@ -94,7 +94,7 @@ index 6fc8384f64..a6c88e0860 100644
          case 'U':
              force_share = true;
              break;
-@@ -5238,11 +5253,24 @@ static int img_dd(int argc, char **argv)
+@@ -5133,11 +5148,24 @@ static int img_dd(int argc, char **argv)
      if (dd.flags & C_IF) {
          blk1 = img_open(image_opts, in.filename, fmt, 0, false, false,
                          force_share);
@@ -120,7 +120,7 @@ index 6fc8384f64..a6c88e0860 100644
      }
  
      if (dd.flags & C_OSIZE) {
-@@ -5397,6 +5425,7 @@ static int img_dd(int argc, char **argv)
+@@ -5292,6 +5320,7 @@ static int img_dd(int argc, char **argv)
  out:
      g_free(arg);
      qemu_opts_del(opts);

debian/patches/pve/0013-PVE-virtio-balloon-improve-query-balloon.patch

@@ -18,10 +18,10 @@ Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
  4 files changed, 82 insertions(+), 4 deletions(-)
 
 diff --git a/hw/core/machine-hmp-cmds.c b/hw/core/machine-hmp-cmds.c
-index 8701f00cc7..3b4c5ef403 100644
+index c3e55ef9e9..0e32e6201f 100644
 --- a/hw/core/machine-hmp-cmds.c
 +++ b/hw/core/machine-hmp-cmds.c
-@@ -179,7 +179,35 @@ void hmp_info_balloon(Monitor *mon, const QDict *qdict)
+@@ -169,7 +169,35 @@ void hmp_info_balloon(Monitor *mon, const QDict *qdict)
          return;
      }
  
@@ -59,10 +59,10 @@ index 8701f00cc7..3b4c5ef403 100644
      qapi_free_BalloonInfo(info);
  }
 diff --git a/hw/virtio/virtio-balloon.c b/hw/virtio/virtio-balloon.c
-index afd2ad6dd6..c724218c17 100644
+index d004cf29d2..2660ed520b 100644
 --- a/hw/virtio/virtio-balloon.c
 +++ b/hw/virtio/virtio-balloon.c
-@@ -795,8 +795,37 @@ static uint64_t virtio_balloon_get_features(VirtIODevice *vdev, uint64_t f,
+@@ -782,8 +782,37 @@ static uint64_t virtio_balloon_get_features(VirtIODevice *vdev, uint64_t f,
  static void virtio_balloon_stat(void *opaque, BalloonInfo *info)
  {
      VirtIOBalloon *dev = opaque;
@@ -103,10 +103,10 @@ index afd2ad6dd6..c724218c17 100644
  
  static void virtio_balloon_to_target(void *opaque, ram_addr_t target)
 diff --git a/qapi/machine.json b/qapi/machine.json
-index a6b8795b09..9f7ed0eaa0 100644
+index a08b6576ca..5c9a4d55f4 100644
 --- a/qapi/machine.json
 +++ b/qapi/machine.json
-@@ -1163,9 +1163,29 @@
+@@ -1063,9 +1063,29 @@
  # @actual: the logical size of the VM in bytes Formula used:
  #     logical_vm_size = vm_ram_size - balloon_size
  #
@@ -138,10 +138,10 @@ index a6b8795b09..9f7ed0eaa0 100644
  ##
  # @query-balloon:
 diff --git a/qapi/pragma.json b/qapi/pragma.json
-index 023a2ef7bc..6aaa9cb975 100644
+index 7f810b0e97..325e684411 100644
 --- a/qapi/pragma.json
 +++ b/qapi/pragma.json
-@@ -81,6 +81,7 @@
+@@ -35,6 +35,7 @@
      'member-name-exceptions': [     # visible in:
          'ACPISlotType',             # query-acpi-ospm-status
          'AcpiTableOptions',         # -acpitable

debian/patches/pve/0014-PVE-qapi-modify-query-machines.patch

@@ -13,10 +13,10 @@ Signed-off-by: Dietmar Maurer <dietmar@proxmox.com>
  2 files changed, 9 insertions(+), 1 deletion(-)
 
 diff --git a/hw/core/machine-qmp-cmds.c b/hw/core/machine-qmp-cmds.c
-index 130217da8f..52a6d74820 100644
+index 3860a50c3b..40821e2317 100644
 --- a/hw/core/machine-qmp-cmds.c
 +++ b/hw/core/machine-qmp-cmds.c
-@@ -90,6 +90,12 @@ MachineInfoList *qmp_query_machines(bool has_compat_props, bool compat_props,
+@@ -91,6 +91,12 @@ MachineInfoList *qmp_query_machines(Error **errp)
          info->numa_mem_supported = mc->numa_mem_supported;
          info->deprecated = !!mc->deprecation_reason;
          info->acpi = !!object_class_property_find(OBJECT_CLASS(mc), "acpi");
@@ -30,10 +30,10 @@ index 130217da8f..52a6d74820 100644
              info->default_cpu_type = g_strdup(mc->default_cpu_type);
          }
 diff --git a/qapi/machine.json b/qapi/machine.json
-index 9f7ed0eaa0..16366b774a 100644
+index 5c9a4d55f4..fbb61f18e4 100644
 --- a/qapi/machine.json
 +++ b/qapi/machine.json
-@@ -167,6 +167,8 @@
+@@ -139,6 +139,8 @@
  #
  # @is-default: whether the machine is default
  #
@@ -42,7 +42,7 @@ index 9f7ed0eaa0..16366b774a 100644
  # @cpu-max: maximum number of CPUs supported by the machine type
  #     (since 1.5)
  #
-@@ -199,7 +201,7 @@
+@@ -163,7 +165,7 @@
  ##
  { 'struct': 'MachineInfo',
    'data': { 'name': 'str', '*alias': 'str',
@@ -50,4 +50,4 @@ index 9f7ed0eaa0..16366b774a 100644
 +            '*is-default': 'bool', '*is-current': 'bool', 'cpu-max': 'int',
              'hotpluggable-cpus': 'bool',  'numa-mem-supported': 'bool',
              'deprecated': 'bool', '*default-cpu-type': 'str',
-             '*default-ram-id': 'str', 'acpi': 'bool',
+             '*default-ram-id': 'str', 'acpi': 'bool' } }

debian/patches/pve/0015-PVE-qapi-modify-spice-query.patch

@@ -14,10 +14,10 @@ Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
  2 files changed, 7 insertions(+)
 
 diff --git a/qapi/ui.json b/qapi/ui.json
-index 460a26b981..42b911bda3 100644
+index 006616aa77..dfd1d3e36b 100644
 --- a/qapi/ui.json
 +++ b/qapi/ui.json
-@@ -312,11 +312,14 @@
+@@ -317,11 +317,14 @@
  #
  # @channels: a list of @SpiceChannel for each active spice channel
  #
@@ -33,7 +33,7 @@ index 460a26b981..42b911bda3 100644
    'if': 'CONFIG_SPICE' }
  
 diff --git a/ui/spice-core.c b/ui/spice-core.c
-index a7ecaad9c7..fecf002d50 100644
+index b20c25aee0..26baeb7846 100644
 --- a/ui/spice-core.c
 +++ b/ui/spice-core.c
 @@ -548,6 +548,10 @@ static SpiceInfo *qmp_query_spice_real(Error **errp)

debian/patches/pve/0016-PVE-add-IOChannel-implementation-for-savevm-async.patch

@@ -14,21 +14,20 @@ Additionally, allows tracking the current position from the outside
 (intended to be used for progress tracking).
 
 Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
-Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
 ---
- migration/channel-savevm-async.c | 184 +++++++++++++++++++++++++++++++
+ migration/channel-savevm-async.c | 183 +++++++++++++++++++++++++++++++
  migration/channel-savevm-async.h |  51 +++++++++
  migration/meson.build            |   1 +
- 3 files changed, 236 insertions(+)
+ 3 files changed, 235 insertions(+)
  create mode 100644 migration/channel-savevm-async.c
  create mode 100644 migration/channel-savevm-async.h
 
 diff --git a/migration/channel-savevm-async.c b/migration/channel-savevm-async.c
 new file mode 100644
-index 0000000000..081a192f49
+index 0000000000..aab081ce07
 --- /dev/null
 +++ b/migration/channel-savevm-async.c
-@@ -0,0 +1,184 @@
+@@ -0,0 +1,183 @@
 +/*
 + * QIO Channel implementation to be used by savevm-async QMP calls
 + */
@@ -175,9 +174,8 @@ index 0000000000..081a192f49
 +
 +static void
 +qio_channel_savevm_async_set_aio_fd_handler(QIOChannel *ioc,
-+                                            AioContext *read_ctx,
++                                            AioContext *ctx,
 +                                            IOHandler *io_read,
-+                                            AioContext *write_ctx,
 +                                            IOHandler *io_write,
 +                                            void *opaque)
 +{
@@ -271,7 +269,7 @@ index 0000000000..17ae2cb261
 +
 +#endif /* QIO_CHANNEL_SAVEVM_ASYNC_H */
 diff --git a/migration/meson.build b/migration/meson.build
-index d53cf3417a..b00d58064d 100644
+index 1ae28523a1..37ddcb5d60 100644
 --- a/migration/meson.build
 +++ b/migration/meson.build
 @@ -13,6 +13,7 @@ system_ss.add(files(
@@ -279,6 +277,6 @@ index d53cf3417a..b00d58064d 100644
    'channel.c',
    'channel-block.c',
 +  'channel-savevm-async.c',
-   'cpu-throttle.c',
    'dirtyrate.c',
    'exec.c',
+   'fd.c',

debian/patches/pve/0017-PVE-add-savevm-async-for-background-state-snapshots.patch

@@ -27,10 +27,7 @@ Signed-off-by: Stefan Reiter <s.reiter@proxmox.com>
 [FE: further improve aborting
      adapt to removal of QEMUFileOps
      improve condition for entering final stage
-     adapt to QAPI and other changes for 8.2
-     make sure to not call vm_start() from coroutine
-     stop CPU throttling after finishing
-     force raw format when loading state as suggested by Friedrich Weber]
+     adapt to QAPI and other changes for 8.0]
 Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
 ---
  hmp-commands-info.hx         |  13 +
@@ -38,20 +35,20 @@ Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
  include/migration/snapshot.h |   2 +
  include/monitor/hmp.h        |   3 +
  migration/meson.build        |   1 +
- migration/savevm-async.c     | 571 +++++++++++++++++++++++++++++++++++
+ migration/savevm-async.c     | 531 +++++++++++++++++++++++++++++++++++
  monitor/hmp-cmds.c           |  38 +++
  qapi/migration.json          |  34 +++
- qapi/misc.json               |  18 ++
+ qapi/misc.json               |  16 ++
  qemu-options.hx              |  12 +
- system/vl.c                  |  10 +
- 11 files changed, 719 insertions(+)
+ softmmu/vl.c                 |  10 +
+ 11 files changed, 677 insertions(+)
  create mode 100644 migration/savevm-async.c
 
 diff --git a/hmp-commands-info.hx b/hmp-commands-info.hx
-index c59cd6637b..d1a7b99add 100644
+index f5b37eb74a..10fdd822e0 100644
 --- a/hmp-commands-info.hx
 +++ b/hmp-commands-info.hx
-@@ -512,6 +512,19 @@ SRST
+@@ -525,6 +525,19 @@ SRST
      Show current migration parameters.
  ERST
  
@@ -72,10 +69,10 @@ index c59cd6637b..d1a7b99add 100644
          .name       = "balloon",
          .args_type  = "",
 diff --git a/hmp-commands.hx b/hmp-commands.hx
-index 06746f0afc..0c7c6f2c16 100644
+index 2cbd0f77a0..e352f86872 100644
 --- a/hmp-commands.hx
 +++ b/hmp-commands.hx
-@@ -1859,3 +1859,20 @@ SRST
+@@ -1865,3 +1865,20 @@ SRST
    List event channels in the guest
  ERST
  #endif
@@ -97,18 +94,18 @@ index 06746f0afc..0c7c6f2c16 100644
 +        .coroutine  = true,
 +    },
 diff --git a/include/migration/snapshot.h b/include/migration/snapshot.h
-index 9e4dcaaa75..2581730d74 100644
+index e72083b117..c846d37806 100644
 --- a/include/migration/snapshot.h
 +++ b/include/migration/snapshot.h
-@@ -68,4 +68,6 @@ bool delete_snapshot(const char *name,
-  */
- void load_snapshot_resume(RunState state);
+@@ -61,4 +61,6 @@ bool delete_snapshot(const char *name,
+                     bool has_devices, strList *devices,
+                     Error **errp);
  
 +int load_snapshot_from_blockdev(const char *filename, Error **errp);
 +
  #endif
 diff --git a/include/monitor/hmp.h b/include/monitor/hmp.h
-index ae116d9804..2596cc2426 100644
+index 13f9a2dedb..7a7def7530 100644
 --- a/include/monitor/hmp.h
 +++ b/include/monitor/hmp.h
 @@ -28,6 +28,7 @@ void hmp_info_status(Monitor *mon, const QDict *qdict);
@@ -119,7 +116,7 @@ index ae116d9804..2596cc2426 100644
  void hmp_info_migrate(Monitor *mon, const QDict *qdict);
  void hmp_info_migrate_capabilities(Monitor *mon, const QDict *qdict);
  void hmp_info_migrate_parameters(Monitor *mon, const QDict *qdict);
-@@ -92,6 +93,8 @@ void hmp_closefd(Monitor *mon, const QDict *qdict);
+@@ -94,6 +95,8 @@ void hmp_closefd(Monitor *mon, const QDict *qdict);
  void hmp_mouse_move(Monitor *mon, const QDict *qdict);
  void hmp_mouse_button(Monitor *mon, const QDict *qdict);
  void hmp_mouse_set(Monitor *mon, const QDict *qdict);
@@ -129,10 +126,10 @@ index ae116d9804..2596cc2426 100644
  void coroutine_fn hmp_screendump(Monitor *mon, const QDict *qdict);
  void hmp_chardev_add(Monitor *mon, const QDict *qdict);
 diff --git a/migration/meson.build b/migration/meson.build
-index b00d58064d..075b013971 100644
+index 37ddcb5d60..07f6057acc 100644
 --- a/migration/meson.build
 +++ b/migration/meson.build
-@@ -29,6 +29,7 @@ system_ss.add(files(
+@@ -26,6 +26,7 @@ system_ss.add(files(
    'options.c',
    'postcopy-ram.c',
    'savevm.c',
@@ -142,10 +139,10 @@ index b00d58064d..075b013971 100644
    'threadinfo.c',
 diff --git a/migration/savevm-async.c b/migration/savevm-async.c
 new file mode 100644
-index 0000000000..ee8ef316d0
+index 0000000000..e9fc18fb10
 --- /dev/null
 +++ b/migration/savevm-async.c
-@@ -0,0 +1,571 @@
+@@ -0,0 +1,531 @@
 +#include "qemu/osdep.h"
 +#include "migration/channel-savevm-async.h"
 +#include "migration/migration.h"
@@ -156,7 +153,6 @@ index 0000000000..ee8ef316d0
 +#include "migration/global_state.h"
 +#include "migration/ram.h"
 +#include "migration/qemu-file.h"
-+#include "sysemu/cpu-throttle.h"
 +#include "sysemu/sysemu.h"
 +#include "sysemu/runstate.h"
 +#include "block/block.h"
@@ -168,7 +164,6 @@ index 0000000000..ee8ef316d0
 +#include "qapi/qapi-commands-misc.h"
 +#include "qapi/qapi-commands-block.h"
 +#include "qemu/cutils.h"
-+#include "qemu/error-report.h"
 +#include "qemu/timer.h"
 +#include "qemu/main-loop.h"
 +#include "qemu/rcu.h"
@@ -262,7 +257,6 @@ index 0000000000..ee8ef316d0
 +    }
 +
 +    if (snap_state.target) {
-+        BlockDriverState *target_bs = blk_bs(snap_state.target);
 +        if (!savevm_aborted()) {
 +            /* try to truncate, but ignore errors (will fail on block devices).
 +            * note1: bdrv_read() need whole blocks, so we need to round up
@@ -271,9 +265,7 @@ index 0000000000..ee8ef316d0
 +            size_t size = QEMU_ALIGN_UP(snap_state.bs_pos, BDRV_SECTOR_SIZE*2);
 +            blk_truncate(snap_state.target, size, false, PREALLOC_MODE_OFF, 0, NULL);
 +        }
-+        if (target_bs) {
-+            bdrv_op_unblock_all(target_bs, snap_state.blocker);
-+        }
++        blk_op_unblock_all(snap_state.target, snap_state.blocker);
 +        error_free(snap_state.blocker);
 +        snap_state.blocker = NULL;
 +        blk_unref(snap_state.target);
@@ -297,7 +289,7 @@ index 0000000000..ee8ef316d0
 +    DPRINTF("save_snapshot_error: %s\n", msg);
 +
 +    if (!snap_state.error) {
-+        error_setg(&snap_state.error, "%s", msg);
++        error_set(&snap_state.error, ERROR_CLASS_GENERIC_ERROR, "%s", msg);
 +    }
 +
 +    g_free (msg);
@@ -308,6 +300,7 @@ index 0000000000..ee8ef316d0
 +static void process_savevm_finalize(void *opaque)
 +{
 +    int ret;
++    AioContext *iohandler_ctx = iohandler_get_aio_context();
 +    MigrationState *ms = migrate_get_current();
 +
 +    bool aborted = savevm_aborted();
@@ -324,7 +317,9 @@ index 0000000000..ee8ef316d0
 +     * so move it back. It can stay in the main context and live out its live
 +     * there, since we're done with it after this method ends anyway.
 +     */
++    aio_context_acquire(iohandler_ctx);
 +    blk_set_aio_context(snap_state.target, qemu_get_aio_context(), NULL);
++    aio_context_release(iohandler_ctx);
 +
 +    ret = vm_stop_force_state(RUN_STATE_FINISH_MIGRATE);
 +    if (ret < 0) {
@@ -349,12 +344,6 @@ index 0000000000..ee8ef316d0
 +        ret || aborted ? MIGRATION_STATUS_FAILED : MIGRATION_STATUS_COMPLETED);
 +    ms->to_dst_file = NULL;
 +
-+    /*
-+     * Same as in migration_iteration_finish(): saving RAM might've turned on CPU throttling for
-+     * auto-converge, make sure to disable it.
-+     */
-+    cpu_throttle_stop();
-+
 +    qemu_savevm_state_cleanup();
 +
 +    ret = save_snapshot_cleanup();
@@ -407,12 +396,12 @@ index 0000000000..ee8ef316d0
 +         * lock. Similar to what is done in migration.c, call the exact variant
 +         * only once pend_precopy in the estimate is below the threshold.
 +         */
-+        bql_unlock();
++        qemu_mutex_unlock_iothread();
 +        qemu_savevm_state_pending_estimate(&pend_precopy, &pend_postcopy);
 +        if (pend_precopy <= threshold) {
 +            qemu_savevm_state_pending_exact(&pend_precopy, &pend_postcopy);
 +        }
-+        bql_lock();
++        qemu_mutex_lock_iothread();
 +        pending_size = pend_precopy + pend_postcopy;
 +
 +        /*
@@ -452,25 +441,21 @@ index 0000000000..ee8ef316d0
 +     * so move there now and after every flush.
 +     */
 +    aio_co_reschedule_self(qemu_get_aio_context());
-+    bdrv_graph_co_rdlock();
-+    bs = bdrv_first(&it);
-+    bdrv_graph_co_rdunlock();
-+    while (bs) {
++    for (bs = bdrv_first(&it); bs; bs = bdrv_next(&it)) {
 +        /* target has BDRV_O_NO_FLUSH, no sense calling bdrv_flush on it */
-+        if (bs != blk_bs(snap_state.target)) {
-+            AioContext *bs_ctx = bdrv_get_aio_context(bs);
-+            if (bs_ctx != qemu_get_aio_context()) {
-+                DPRINTF("savevm: async flushing drive %s\n", bs->filename);
-+                aio_co_reschedule_self(bs_ctx);
-+                bdrv_graph_co_rdlock();
-+                bdrv_flush(bs);
-+                bdrv_graph_co_rdunlock();
-+                aio_co_reschedule_self(qemu_get_aio_context());
-+            }
++        if (bs == blk_bs(snap_state.target)) {
++            continue;
++        }
++
++        AioContext *bs_ctx = bdrv_get_aio_context(bs);
++        if (bs_ctx != qemu_get_aio_context()) {
++            DPRINTF("savevm: async flushing drive %s\n", bs->filename);
++            aio_co_reschedule_self(bs_ctx);
++            bdrv_graph_co_rdlock();
++            bdrv_flush(bs);
++            bdrv_graph_co_rdunlock();
++            aio_co_reschedule_self(qemu_get_aio_context());
 +        }
-+        bdrv_graph_co_rdlock();
-+        bs = bdrv_next(&it);
-+        bdrv_graph_co_rdunlock();
 +    }
 +
 +    DPRINTF("timing: async flushing took %ld ms\n",
@@ -484,18 +469,23 @@ index 0000000000..ee8ef316d0
 +    Error *local_err = NULL;
 +    MigrationState *ms = migrate_get_current();
 +    AioContext *iohandler_ctx = iohandler_get_aio_context();
-+    BlockDriverState *target_bs = NULL;
-+    int ret = 0;
 +
 +    int bdrv_oflags = BDRV_O_RDWR | BDRV_O_RESIZE | BDRV_O_NO_FLUSH;
 +
 +    if (snap_state.state != SAVE_STATE_DONE) {
-+        error_setg(errp, "VM snapshot already started\n");
++        error_set(errp, ERROR_CLASS_GENERIC_ERROR,
++                  "VM snapshot already started\n");
 +        return;
 +    }
 +
-+    if (migration_is_running()) {
-+        error_setg(errp, "There's a migration process in progress");
++    if (migration_is_running(ms->state)) {
++        error_set(errp, ERROR_CLASS_GENERIC_ERROR, QERR_MIGRATION_ACTIVE);
++        return;
++    }
++
++    if (migrate_block()) {
++        error_set(errp, ERROR_CLASS_GENERIC_ERROR,
++                  "Block migration and snapshots are incompatible");
 +        return;
 +    }
 +
@@ -527,12 +517,7 @@ index 0000000000..ee8ef316d0
 +    qdict_put_str(options, "driver", "raw");
 +    snap_state.target = blk_new_open(statefile, NULL, options, bdrv_oflags, &local_err);
 +    if (!snap_state.target) {
-+        error_setg(errp, "failed to open '%s'", statefile);
-+        goto restart;
-+    }
-+    target_bs = blk_bs(snap_state.target);
-+    if (!target_bs) {
-+        error_setg(errp, "failed to open '%s' - no block driver state", statefile);
++        error_set(errp, ERROR_CLASS_GENERIC_ERROR, "failed to open '%s'", statefile);
 +        goto restart;
 +    }
 +
@@ -541,7 +526,7 @@ index 0000000000..ee8ef316d0
 +    snap_state.file = qemu_file_new_output(ioc);
 +
 +    if (!snap_state.file) {
-+        error_setg(errp, "failed to open '%s'", statefile);
++        error_set(errp, ERROR_CLASS_GENERIC_ERROR, "failed to open '%s'", statefile);
 +        goto restart;
 +    }
 +
@@ -550,36 +535,27 @@ index 0000000000..ee8ef316d0
 +     * State is cleared in process_savevm_co, but has to be initialized
 +     * here (blocking main thread, from QMP) to avoid race conditions.
 +     */
-+    if (migrate_init(ms, errp)) {
-+        return;
-+    }
++    migrate_init(ms);
 +    memset(&mig_stats, 0, sizeof(mig_stats));
++    memset(&compression_counters, 0, sizeof(compression_counters));
 +    ms->to_dst_file = snap_state.file;
 +
 +    error_setg(&snap_state.blocker, "block device is in use by savevm");
-+    bdrv_op_block_all(target_bs, snap_state.blocker);
++    blk_op_block_all(snap_state.target, snap_state.blocker);
 +
 +    snap_state.state = SAVE_STATE_ACTIVE;
 +    snap_state.finalize_bh = qemu_bh_new(process_savevm_finalize, &snap_state);
++    snap_state.co = qemu_coroutine_create(&process_savevm_co, NULL);
++    qemu_mutex_unlock_iothread();
 +    qemu_savevm_state_header(snap_state.file);
-+    ret = qemu_savevm_state_setup(snap_state.file, &local_err);
-+    if (ret != 0) {
-+        error_setg_errno(errp, -ret, "savevm state setup failed: %s",
-+                         local_err ? error_get_pretty(local_err) : "unknown error");
-+        return;
-+    }
++    qemu_savevm_state_setup(snap_state.file);
++    qemu_mutex_lock_iothread();
 +
 +    /* Async processing from here on out happens in iohandler context, so let
 +     * the target bdrv have its home there.
 +     */
-+    ret = blk_set_aio_context(snap_state.target, iohandler_ctx, &local_err);
-+    if (ret != 0) {
-+        warn_report("failed to set iohandler context for VM state target: %s %s",
-+                    local_err ? error_get_pretty(local_err) : "unknown error",
-+                    strerror(-ret));
-+    }
++    blk_set_aio_context(snap_state.target, iohandler_ctx, &local_err);
 +
-+    snap_state.co = qemu_coroutine_create(&process_savevm_co, NULL);
 +    aio_co_schedule(iohandler_ctx, snap_state.co);
 +
 +    return;
@@ -594,10 +570,29 @@ index 0000000000..ee8ef316d0
 +    }
 +}
 +
-+static void coroutine_fn wait_for_close_co(void *opaque)
++void coroutine_fn qmp_savevm_end(Error **errp)
 +{
 +    int64_t timeout;
 +
++    if (snap_state.state == SAVE_STATE_DONE) {
++        error_set(errp, ERROR_CLASS_GENERIC_ERROR,
++                  "VM snapshot not started\n");
++        return;
++    }
++
++    if (snap_state.state == SAVE_STATE_ACTIVE) {
++        snap_state.state = SAVE_STATE_CANCELLED;
++        goto wait_for_close;
++    }
++
++    if (snap_state.saved_vm_running) {
++        vm_start();
++        snap_state.saved_vm_running = false;
++    }
++
++    snap_state.state = SAVE_STATE_DONE;
++
++wait_for_close:
 +    if (!snap_state.target) {
 +        DPRINTF("savevm-end: no target file open\n");
 +        return;
@@ -625,61 +620,25 @@ index 0000000000..ee8ef316d0
 +    DPRINTF("savevm-end: cleanup done\n");
 +}
 +
-+void qmp_savevm_end(Error **errp)
-+{
-+    if (snap_state.state == SAVE_STATE_DONE) {
-+        error_setg(errp, "VM snapshot not started\n");
-+        return;
-+    }
-+
-+    Coroutine *wait_for_close = qemu_coroutine_create(wait_for_close_co, NULL);
-+
-+    if (snap_state.state == SAVE_STATE_ACTIVE) {
-+        snap_state.state = SAVE_STATE_CANCELLED;
-+        qemu_coroutine_enter(wait_for_close);
-+        return;
-+    }
-+
-+    if (snap_state.saved_vm_running) {
-+        vm_start();
-+        snap_state.saved_vm_running = false;
-+    }
-+
-+    snap_state.state = SAVE_STATE_DONE;
-+
-+    qemu_coroutine_enter(wait_for_close);
-+}
-+
 +int load_snapshot_from_blockdev(const char *filename, Error **errp)
 +{
 +    BlockBackend *be;
-+    BlockDriverState *bs = NULL;
 +    Error *local_err = NULL;
 +    Error *blocker = NULL;
-+    QDict *options;
 +
 +    QEMUFile *f;
 +    size_t bs_pos = 0;
 +    int ret = -EINVAL;
 +
-+    options = qdict_new();
-+    qdict_put_str(options, "driver", "raw");
-+
-+    be = blk_new_open(filename, NULL, options, 0, &local_err);
++    be = blk_new_open(filename, NULL, NULL, 0, &local_err);
 +
 +    if (!be) {
 +        error_setg(errp, "Could not open VM state file");
 +        goto the_end;
 +    }
 +
-+    bs = blk_bs(be);
-+    if (!bs) {
-+        error_setg(errp, "Could not open VM state file - missing block driver state");
-+        goto the_end;
-+    }
-+
 +    error_setg(&blocker, "block device is in use by load state");
-+    bdrv_op_block_all(bs, blocker);
++    blk_op_block_all(be, blocker);
 +
 +    /* restore the VM state */
 +    f = qemu_file_new_input(QIO_CHANNEL(qio_channel_savevm_async_new(be, &bs_pos)));
@@ -709,30 +668,28 @@ index 0000000000..ee8ef316d0
 +
 + the_end:
 +    if (be) {
-+        if (bs) {
-+            bdrv_op_unblock_all(bs, blocker);
-+        }
++        blk_op_unblock_all(be, blocker);
 +        error_free(blocker);
 +        blk_unref(be);
 +    }
 +    return ret;
 +}
 diff --git a/monitor/hmp-cmds.c b/monitor/hmp-cmds.c
-index f601d06ab8..874084565f 100644
+index 6c559b48c8..91be698308 100644
 --- a/monitor/hmp-cmds.c
 +++ b/monitor/hmp-cmds.c
-@@ -24,6 +24,7 @@
+@@ -22,6 +22,7 @@
+ #include "monitor/monitor-internal.h"
  #include "qapi/error.h"
  #include "qapi/qapi-commands-control.h"
- #include "qapi/qapi-commands-machine.h"
 +#include "qapi/qapi-commands-migration.h"
  #include "qapi/qapi-commands-misc.h"
  #include "qapi/qmp/qdict.h"
- #include "qemu/cutils.h"
-@@ -434,3 +435,40 @@ void hmp_dumpdtb(Monitor *mon, const QDict *qdict)
-     monitor_printf(mon, "dtb dumped to %s", filename);
+ #include "qapi/qmp/qerror.h"
+@@ -443,3 +444,40 @@ void hmp_info_mtree(Monitor *mon, const QDict *qdict)
+ 
+     mtree_info(flatview, dispatch_tree, owner, disabled);
  }
- #endif
 +
 +void hmp_savevm_start(Monitor *mon, const QDict *qdict)
 +{
@@ -771,10 +728,10 @@ index f601d06ab8..874084565f 100644
 +    }
 +}
 diff --git a/qapi/migration.json b/qapi/migration.json
-index a605dc26db..927b1e1c7d 100644
+index 8843e74b59..aca0ca1ac1 100644
 --- a/qapi/migration.json
 +++ b/qapi/migration.json
-@@ -276,6 +276,40 @@
+@@ -291,6 +291,40 @@
             '*dirty-limit-throttle-time-per-round': 'uint64',
             '*dirty-limit-ring-full-time': 'uint64'} }
  
@@ -816,10 +773,10 @@ index a605dc26db..927b1e1c7d 100644
  # @query-migrate:
  #
 diff --git a/qapi/misc.json b/qapi/misc.json
-index 559b66f201..7959e89c1e 100644
+index cda2effa81..94a58bb0bf 100644
 --- a/qapi/misc.json
 +++ b/qapi/misc.json
-@@ -454,6 +454,24 @@
+@@ -456,6 +456,22 @@
  ##
  { 'command': 'query-fdsets', 'returns': ['FdsetInfo'] }
  
@@ -828,8 +785,6 @@ index 559b66f201..7959e89c1e 100644
 +#
 +# Prepare for snapshot and halt VM. Save VM state to statefile.
 +#
-+# @statefile: target file that state should be written to.
-+#
 +##
 +{ 'command': 'savevm-start', 'data': { '*statefile': 'str' } }
 +
@@ -839,16 +794,16 @@ index 559b66f201..7959e89c1e 100644
 +# Resume VM after a snapshot.
 +#
 +##
-+{ 'command': 'savevm-end' }
++{ 'command': 'savevm-end', 'coroutine': true }
 +
  ##
  # @CommandLineParameterType:
  #
 diff --git a/qemu-options.hx b/qemu-options.hx
-index dacc9790a4..c05f411599 100644
+index b56f6b2fb2..c8c78c92d4 100644
 --- a/qemu-options.hx
 +++ b/qemu-options.hx
-@@ -4764,6 +4764,18 @@ SRST
+@@ -4479,6 +4479,18 @@ SRST
      Start right away with a saved state (``loadvm`` in monitor)
  ERST
  
@@ -867,10 +822,10 @@ index dacc9790a4..c05f411599 100644
  #ifndef _WIN32
  DEF("daemonize", 0, QEMU_OPTION_daemonize, \
      "-daemonize      daemonize QEMU after initializing\n", QEMU_ARCH_ALL)
-diff --git a/system/vl.c b/system/vl.c
-index 2f855d83fb..39d451bb41 100644
---- a/system/vl.c
-+++ b/system/vl.c
+diff --git a/softmmu/vl.c b/softmmu/vl.c
+index b0b96f67fa..f3251de3e7 100644
+--- a/softmmu/vl.c
++++ b/softmmu/vl.c
 @@ -164,6 +164,7 @@ static const char *accelerators;
  static bool have_custom_ram_size;
  static const char *ram_memdev_id;
@@ -879,10 +834,10 @@ index 2f855d83fb..39d451bb41 100644
  static QTAILQ_HEAD(, ObjectOption) object_opts = QTAILQ_HEAD_INITIALIZER(object_opts);
  static QTAILQ_HEAD(, DeviceOption) device_opts = QTAILQ_HEAD_INITIALIZER(device_opts);
  static int display_remote;
-@@ -2725,6 +2726,12 @@ void qmp_x_exit_preconfig(Error **errp)
-         RunState state = autostart ? RUN_STATE_RUNNING : runstate_get();
+@@ -2643,6 +2644,12 @@ void qmp_x_exit_preconfig(Error **errp)
+ 
+     if (loadvm) {
          load_snapshot(loadvm, NULL, false, NULL, &error_fatal);
-         load_snapshot_resume(state);
 +    } else if (loadstate) {
 +        Error *local_err = NULL;
 +        if (load_snapshot_from_blockdev(loadstate, &local_err) < 0) {
@@ -892,7 +847,7 @@ index 2f855d83fb..39d451bb41 100644
      }
      if (replay_mode != REPLAY_MODE_NONE) {
          replay_vmstate_init();
-@@ -3262,6 +3269,9 @@ void qemu_init(int argc, char **argv)
+@@ -3190,6 +3197,9 @@ void qemu_init(int argc, char **argv)
              case QEMU_OPTION_loadvm:
                  loadvm = optarg;
                  break;

debian/patches/pve/0018-PVE-add-optional-buffer-size-to-QEMUFile.patch

@@ -13,18 +13,18 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
 [FE: adapt to removal of QEMUFileOps]
 Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
 ---
- migration/qemu-file.c    | 48 +++++++++++++++++++++++++++-------------
+ migration/qemu-file.c    | 49 +++++++++++++++++++++++++++-------------
  migration/qemu-file.h    |  2 ++
- migration/savevm-async.c |  5 +++--
- 3 files changed, 38 insertions(+), 17 deletions(-)
+ migration/savevm-async.c |  5 ++--
+ 3 files changed, 38 insertions(+), 18 deletions(-)
 
 diff --git a/migration/qemu-file.c b/migration/qemu-file.c
-index b6d2f588bd..754dc0b3f7 100644
+index 19c33c9985..e9ffff0f0a 100644
 --- a/migration/qemu-file.c
 +++ b/migration/qemu-file.c
-@@ -34,8 +34,8 @@
- #include "rdma.h"
- #include "io/channel-file.h"
+@@ -33,8 +33,8 @@
+ #include "options.h"
+ #include "qapi/error.h"
  
 -#define IO_BUF_SIZE 32768
 -#define MAX_IOV_SIZE MIN_CONST(IOV_MAX, 64)
@@ -32,8 +32,8 @@ index b6d2f588bd..754dc0b3f7 100644
 +#define MAX_IOV_SIZE MIN_CONST(IOV_MAX, 256)
  
  struct QEMUFile {
-     QIOChannel *ioc;
-@@ -43,7 +43,8 @@ struct QEMUFile {
+     const QEMUFileHooks *hooks;
+@@ -46,7 +46,8 @@ struct QEMUFile {
  
      int buf_index;
      int buf_size; /* 0 when writing */
@@ -93,8 +93,8 @@ index b6d2f588bd..754dc0b3f7 100644
 +    return qemu_file_new_impl(ioc, false, buffer_size);
  }
  
- /*
-@@ -327,7 +342,7 @@ static ssize_t coroutine_mixed_fn qemu_fill_buffer(QEMUFile *f)
+ void qemu_file_set_hooks(QEMUFile *f, const QEMUFileHooks *hooks)
+@@ -375,7 +390,7 @@ static ssize_t coroutine_mixed_fn qemu_fill_buffer(QEMUFile *f)
      do {
          len = qio_channel_read(f->ioc,
                                 (char *)f->buf + pending,
@@ -103,17 +103,16 @@ index b6d2f588bd..754dc0b3f7 100644
                                 &local_error);
          if (len == QIO_CHANNEL_ERR_BLOCK) {
              if (qemu_in_coroutine()) {
-@@ -367,6 +382,9 @@ int qemu_fclose(QEMUFile *f)
-         ret = ret2;
+@@ -425,6 +440,8 @@ int qemu_fclose(QEMUFile *f)
      }
      g_clear_pointer(&f->ioc, object_unref);
-+
+ 
 +    free(f->buf);
 +
-     error_free(f->last_error_obj);
-     g_free(f);
-     trace_qemu_file_fclose();
-@@ -415,7 +433,7 @@ static void add_buf_to_iovec(QEMUFile *f, size_t len)
+     /* If any error was spotted before closing, we should report it
+      * instead of the close() return value.
+      */
+@@ -479,7 +496,7 @@ static void add_buf_to_iovec(QEMUFile *f, size_t len)
  {
      if (!add_to_iovec(f, f->buf + f->buf_index, len, false)) {
          f->buf_index += len;
@@ -122,7 +121,7 @@ index b6d2f588bd..754dc0b3f7 100644
              qemu_fflush(f);
          }
      }
-@@ -440,7 +458,7 @@ void qemu_put_buffer(QEMUFile *f, const uint8_t *buf, size_t size)
+@@ -504,7 +521,7 @@ void qemu_put_buffer(QEMUFile *f, const uint8_t *buf, size_t size)
      }
  
      while (size > 0) {
@@ -131,7 +130,7 @@ index b6d2f588bd..754dc0b3f7 100644
          if (l > size) {
              l = size;
          }
-@@ -586,8 +604,8 @@ size_t coroutine_mixed_fn qemu_peek_buffer(QEMUFile *f, uint8_t **buf, size_t si
+@@ -549,8 +566,8 @@ size_t coroutine_mixed_fn qemu_peek_buffer(QEMUFile *f, uint8_t **buf, size_t si
      size_t index;
  
      assert(!qemu_file_is_writable(f));
@@ -142,7 +141,7 @@ index b6d2f588bd..754dc0b3f7 100644
  
      /* The 1st byte to read from */
      index = f->buf_index + offset;
-@@ -637,7 +655,7 @@ size_t coroutine_mixed_fn qemu_get_buffer(QEMUFile *f, uint8_t *buf, size_t size
+@@ -600,7 +617,7 @@ size_t coroutine_mixed_fn qemu_get_buffer(QEMUFile *f, uint8_t *buf, size_t size
          size_t res;
          uint8_t *src;
  
@@ -151,7 +150,7 @@ index b6d2f588bd..754dc0b3f7 100644
          if (res == 0) {
              return done;
          }
-@@ -671,7 +689,7 @@ size_t coroutine_mixed_fn qemu_get_buffer(QEMUFile *f, uint8_t *buf, size_t size
+@@ -634,7 +651,7 @@ size_t coroutine_mixed_fn qemu_get_buffer(QEMUFile *f, uint8_t *buf, size_t size
   */
  size_t coroutine_mixed_fn qemu_get_buffer_in_place(QEMUFile *f, uint8_t **buf, size_t size)
  {
@@ -160,7 +159,7 @@ index b6d2f588bd..754dc0b3f7 100644
          size_t res;
          uint8_t *src = NULL;
  
-@@ -696,7 +714,7 @@ int coroutine_mixed_fn qemu_peek_byte(QEMUFile *f, int offset)
+@@ -659,7 +676,7 @@ int coroutine_mixed_fn qemu_peek_byte(QEMUFile *f, int offset)
      int index = f->buf_index + offset;
  
      assert(!qemu_file_is_writable(f));
@@ -169,25 +168,34 @@ index b6d2f588bd..754dc0b3f7 100644
  
      if (index >= f->buf_size) {
          qemu_fill_buffer(f);
+@@ -777,7 +794,7 @@ static int qemu_compress_data(z_stream *stream, uint8_t *dest, size_t dest_len,
+ ssize_t qemu_put_compression_data(QEMUFile *f, z_stream *stream,
+                                   const uint8_t *p, size_t size)
+ {
+-    ssize_t blen = IO_BUF_SIZE - f->buf_index - sizeof(int32_t);
++    ssize_t blen = f->buf_allocated_size - f->buf_index - sizeof(int32_t);
+ 
+     if (blen < compressBound(size)) {
+         return -1;
 diff --git a/migration/qemu-file.h b/migration/qemu-file.h
-index 11c2120edd..edf3c5d147 100644
+index 47015f5201..1312b7c903 100644
 --- a/migration/qemu-file.h
 +++ b/migration/qemu-file.h
-@@ -30,7 +30,9 @@
- #include "io/channel.h"
+@@ -63,7 +63,9 @@ typedef struct QEMUFileHooks {
+ } QEMUFileHooks;
  
  QEMUFile *qemu_file_new_input(QIOChannel *ioc);
 +QEMUFile *qemu_file_new_input_sized(QIOChannel *ioc, size_t buffer_size);
  QEMUFile *qemu_file_new_output(QIOChannel *ioc);
 +QEMUFile *qemu_file_new_output_sized(QIOChannel *ioc, size_t buffer_size);
+ void qemu_file_set_hooks(QEMUFile *f, const QEMUFileHooks *hooks);
  int qemu_fclose(QEMUFile *f);
  
- /*
 diff --git a/migration/savevm-async.c b/migration/savevm-async.c
-index ee8ef316d0..1e79fce9ba 100644
+index e9fc18fb10..80624fada8 100644
 --- a/migration/savevm-async.c
 +++ b/migration/savevm-async.c
-@@ -390,7 +390,7 @@ void qmp_savevm_start(const char *statefile, Error **errp)
+@@ -378,7 +378,7 @@ void qmp_savevm_start(const char *statefile, Error **errp)
  
      QIOChannel *ioc = QIO_CHANNEL(qio_channel_savevm_async_new(snap_state.target,
                                                                 &snap_state.bs_pos));
@@ -195,9 +203,9 @@ index ee8ef316d0..1e79fce9ba 100644
 +    snap_state.file = qemu_file_new_output_sized(ioc, 4 * 1024 * 1024);
  
      if (!snap_state.file) {
-         error_setg(errp, "failed to open '%s'", statefile);
-@@ -534,7 +534,8 @@ int load_snapshot_from_blockdev(const char *filename, Error **errp)
-     bdrv_op_block_all(bs, blocker);
+         error_set(errp, ERROR_CLASS_GENERIC_ERROR, "failed to open '%s'", statefile);
+@@ -496,7 +496,8 @@ int load_snapshot_from_blockdev(const char *filename, Error **errp)
+     blk_op_block_all(be, blocker);
  
      /* restore the VM state */
 -    f = qemu_file_new_input(QIO_CHANNEL(qio_channel_savevm_async_new(be, &bs_pos)));

debian/patches/pve/0019-PVE-block-add-the-zeroinit-block-driver-filter.patch

@@ -4,34 +4,32 @@ Date: Mon, 6 Apr 2020 12:16:47 +0200
 Subject: [PATCH] PVE: block: add the zeroinit block driver filter
 
 Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
-[FE: adapt to changed function signatures
-     adhere to block graph lock requirements
-     use dedicated function to open file child]
+[FE: adapt to changed function signatures]
 Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
 ---
  block/meson.build |   1 +
- block/zeroinit.c  | 207 ++++++++++++++++++++++++++++++++++++++++++++++
- 2 files changed, 208 insertions(+)
+ block/zeroinit.c  | 200 ++++++++++++++++++++++++++++++++++++++++++++++
+ 2 files changed, 201 insertions(+)
  create mode 100644 block/zeroinit.c
 
 diff --git a/block/meson.build b/block/meson.build
-index f1262ec2ba..6a60b5d6b9 100644
+index 529fc172c6..1833c71ce9 100644
 --- a/block/meson.build
 +++ b/block/meson.build
-@@ -39,6 +39,7 @@ block_ss.add(files(
-   'throttle.c',
+@@ -40,6 +40,7 @@ block_ss.add(files(
    'throttle-groups.c',
+   'throttle.c',
    'write-threshold.c',
 +  'zeroinit.c',
- ), zstd, zlib)
+ ), zstd, zlib, gnutls)
  
  system_ss.add(when: 'CONFIG_TCG', if_true: files('blkreplay.c'))
 diff --git a/block/zeroinit.c b/block/zeroinit.c
 new file mode 100644
-index 0000000000..2b2b194ccf
+index 0000000000..1257342724
 --- /dev/null
 +++ b/block/zeroinit.c
-@@ -0,0 +1,207 @@
+@@ -0,0 +1,200 @@
 +/*
 + * Filter to fake a zero-initialized block device.
 + *
@@ -46,7 +44,6 @@ index 0000000000..2b2b194ccf
 +#include "qapi/error.h"
 +#include "block/block_int.h"
 +#include "block/block-io.h"
-+#include "block/graph-lock.h"
 +#include "qapi/qmp/qdict.h"
 +#include "qapi/qmp/qstring.h"
 +#include "qemu/cutils.h"
@@ -112,9 +109,12 @@ index 0000000000..2b2b194ccf
 +    }
 +
 +    /* Open the raw file */
-+    ret = bdrv_open_file_child(qemu_opt_get(opts, "x-next"), options, "next",
-+                               bs, &local_err);
-+    if (ret < 0) {
++    bs->file = bdrv_open_child(qemu_opt_get(opts, "x-next"), options, "next",
++                               bs, &child_of_bds,
++                               BDRV_CHILD_FILTERED | BDRV_CHILD_PRIMARY,
++                               false, &local_err);
++    if (local_err) {
++        ret = -EINVAL;
 +        error_propagate(errp, local_err);
 +        goto fail;
 +    }
@@ -125,9 +125,7 @@ index 0000000000..2b2b194ccf
 +    ret = 0;
 +fail:
 +    if (ret < 0) {
-+        bdrv_graph_wrlock();
 +        bdrv_unref_child(bs, bs->file);
-+        bdrv_graph_wrunlock();
 +    }
 +    qemu_opts_del(opts);
 +    return ret;
@@ -139,22 +137,19 @@ index 0000000000..2b2b194ccf
 +    (void)s;
 +}
 +
-+static coroutine_fn int64_t GRAPH_RDLOCK
-+zeroinit_co_getlength(BlockDriverState *bs)
++static coroutine_fn int64_t zeroinit_co_getlength(BlockDriverState *bs)
 +{
 +    return bdrv_co_getlength(bs->file->bs);
 +}
 +
-+static int coroutine_fn GRAPH_RDLOCK
-+zeroinit_co_preadv(BlockDriverState *bs, int64_t offset, int64_t bytes,
-+                   QEMUIOVector *qiov, BdrvRequestFlags flags)
++static int coroutine_fn zeroinit_co_preadv(BlockDriverState *bs,
++    int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags)
 +{
 +    return bdrv_co_preadv(bs->file, offset, bytes, qiov, flags);
 +}
 +
-+static int coroutine_fn GRAPH_RDLOCK
-+zeroinit_co_pwrite_zeroes(BlockDriverState *bs, int64_t offset, int64_t bytes,
-+                          BdrvRequestFlags flags)
++static int coroutine_fn zeroinit_co_pwrite_zeroes(BlockDriverState *bs, int64_t offset,
++                                                 int64_t bytes, BdrvRequestFlags flags)
 +{
 +    BDRVZeroinitState *s = bs->opaque;
 +    if (offset >= s->extents)
@@ -162,9 +157,8 @@ index 0000000000..2b2b194ccf
 +    return bdrv_pwrite_zeroes(bs->file, offset, bytes, flags);
 +}
 +
-+static int coroutine_fn GRAPH_RDLOCK
-+zeroinit_co_pwritev(BlockDriverState *bs, int64_t offset, int64_t bytes,
-+                    QEMUIOVector *qiov, BdrvRequestFlags flags)
++static int coroutine_fn zeroinit_co_pwritev(BlockDriverState *bs,
++    int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags)
 +{
 +    BDRVZeroinitState *s = bs->opaque;
 +    int64_t extents = offset + bytes;
@@ -173,35 +167,32 @@ index 0000000000..2b2b194ccf
 +    return bdrv_co_pwritev(bs->file, offset, bytes, qiov, flags);
 +}
 +
-+static coroutine_fn int GRAPH_RDLOCK
-+zeroinit_co_flush(BlockDriverState *bs)
++static coroutine_fn int zeroinit_co_flush(BlockDriverState *bs)
 +{
 +    return bdrv_co_flush(bs->file->bs);
 +}
 +
-+static int GRAPH_RDLOCK
-+zeroinit_has_zero_init(BlockDriverState *bs)
++static int zeroinit_has_zero_init(BlockDriverState *bs)
 +{
 +    BDRVZeroinitState *s = bs->opaque;
 +    return s->has_zero_init;
 +}
 +
-+static int coroutine_fn GRAPH_RDLOCK
-+zeroinit_co_pdiscard(BlockDriverState *bs, int64_t offset, int64_t bytes)
++static int coroutine_fn zeroinit_co_pdiscard(BlockDriverState *bs,
++                                             int64_t offset, int64_t bytes)
 +{
 +    return bdrv_co_pdiscard(bs->file, offset, bytes);
 +}
 +
-+static int GRAPH_RDLOCK
-+zeroinit_co_truncate(BlockDriverState *bs, int64_t offset, _Bool exact,
-+                     PreallocMode prealloc, BdrvRequestFlags req_flags,
-+                     Error **errp)
++static int zeroinit_co_truncate(BlockDriverState *bs, int64_t offset,
++                                _Bool exact, PreallocMode prealloc,
++                                BdrvRequestFlags req_flags, Error **errp)
 +{
 +    return bdrv_co_truncate(bs->file, offset, exact, prealloc, req_flags, errp);
 +}
 +
-+static coroutine_fn int GRAPH_RDLOCK
-+zeroinit_co_get_info(BlockDriverState *bs, BlockDriverInfo *bdi)
++static coroutine_fn int zeroinit_co_get_info(BlockDriverState *bs,
++                                             BlockDriverInfo *bdi)
 +{
 +    return bdrv_co_get_info(bs->file->bs, bdi);
 +}
@@ -212,7 +203,7 @@ index 0000000000..2b2b194ccf
 +    .instance_size                    = sizeof(BDRVZeroinitState),
 +
 +    .bdrv_parse_filename              = zeroinit_parse_filename,
-+    .bdrv_open                        = zeroinit_open,
++    .bdrv_file_open                   = zeroinit_open,
 +    .bdrv_close                       = zeroinit_close,
 +    .bdrv_co_getlength                = zeroinit_co_getlength,
 +    .bdrv_child_perm                  = bdrv_default_perms,

debian/patches/pve/0020-PVE-Add-dummy-id-command-line-parameter.patch

@@ -10,14 +10,14 @@ Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
 Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
 ---
  qemu-options.hx | 3 +++
- system/vl.c     | 8 ++++++++
+ softmmu/vl.c    | 8 ++++++++
  2 files changed, 11 insertions(+)
 
 diff --git a/qemu-options.hx b/qemu-options.hx
-index c05f411599..0732077a0e 100644
+index c8c78c92d4..20ca2cdba7 100644
 --- a/qemu-options.hx
 +++ b/qemu-options.hx
-@@ -1239,6 +1239,9 @@ legacy PC, they are not recommended for modern configurations.
+@@ -1197,6 +1197,9 @@ legacy PC, they are not recommended for modern configurations.
  
  ERST
  
@@ -27,11 +27,11 @@ index c05f411599..0732077a0e 100644
  DEF("fda", HAS_ARG, QEMU_OPTION_fda,
      "-fda/-fdb file  use 'file' as floppy disk 0/1 image\n", QEMU_ARCH_ALL)
  DEF("fdb", HAS_ARG, QEMU_OPTION_fdb, "", QEMU_ARCH_ALL)
-diff --git a/system/vl.c b/system/vl.c
-index 39d451bb41..e7cae51f13 100644
---- a/system/vl.c
-+++ b/system/vl.c
-@@ -2762,6 +2762,7 @@ void qemu_init(int argc, char **argv)
+diff --git a/softmmu/vl.c b/softmmu/vl.c
+index f3251de3e7..1b63ffd33d 100644
+--- a/softmmu/vl.c
++++ b/softmmu/vl.c
+@@ -2679,6 +2679,7 @@ void qemu_init(int argc, char **argv)
      MachineClass *machine_class;
      bool userconfig = true;
      FILE *vmstate_dump_file = NULL;
@@ -39,7 +39,7 @@ index 39d451bb41..e7cae51f13 100644
  
      qemu_add_opts(&qemu_drive_opts);
      qemu_add_drive_opts(&qemu_legacy_drive_opts);
-@@ -3374,6 +3375,13 @@ void qemu_init(int argc, char **argv)
+@@ -3302,6 +3303,13 @@ void qemu_init(int argc, char **argv)
                  machine_parse_property_opt(qemu_find_opts("smp-opts"),
                                             "smp", optarg);
                  break;
@@ -50,6 +50,6 @@ index 39d451bb41..e7cae51f13 100644
 +                    exit(1);
 +                }
 +                break;
- #ifdef CONFIG_VNC
              case QEMU_OPTION_vnc:
                  vnc_parse(optarg);
+                 break;

debian/patches/pve/0021-PVE-Config-Revert-target-i386-disable-LINT0-after-re.patch

@@ -11,10 +11,10 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  1 file changed, 9 insertions(+)
 
 diff --git a/hw/intc/apic_common.c b/hw/intc/apic_common.c
-index 62f3bbf203..89e0c7d995 100644
+index 4a34f03047..59b917e50c 100644
 --- a/hw/intc/apic_common.c
 +++ b/hw/intc/apic_common.c
-@@ -263,6 +263,15 @@ static void apic_reset_common(DeviceState *dev)
+@@ -252,6 +252,15 @@ static void apic_reset_common(DeviceState *dev)
      info->vapic_base_update(s);
  
      apic_init_reset(dev);

debian/patches/pve/0022-PVE-Up-Config-file-posix-make-locking-optiono-on-cre.patch

@@ -9,14 +9,14 @@ Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
 Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
 ---
  block/file-posix.c   | 59 ++++++++++++++++++++++++++++++--------------
- qapi/block-core.json |  7 +++++-
- 2 files changed, 46 insertions(+), 20 deletions(-)
+ qapi/block-core.json |  3 ++-
+ 2 files changed, 42 insertions(+), 20 deletions(-)
 
 diff --git a/block/file-posix.c b/block/file-posix.c
-index e2ea071315..4c3dc56c8e 100644
+index 0db366a851..46f1ee38ae 100644
 --- a/block/file-posix.c
 +++ b/block/file-posix.c
-@@ -2884,6 +2884,7 @@ raw_co_create(BlockdevCreateOptions *options, Error **errp)
+@@ -2870,6 +2870,7 @@ raw_co_create(BlockdevCreateOptions *options, Error **errp)
      int fd;
      uint64_t perm, shared;
      int result = 0;
@@ -24,7 +24,7 @@ index e2ea071315..4c3dc56c8e 100644
  
      /* Validate options and set default values */
      assert(options->driver == BLOCKDEV_DRIVER_FILE);
-@@ -2924,19 +2925,22 @@ raw_co_create(BlockdevCreateOptions *options, Error **errp)
+@@ -2910,19 +2911,22 @@ raw_co_create(BlockdevCreateOptions *options, Error **errp)
      perm = BLK_PERM_WRITE | BLK_PERM_RESIZE;
      shared = BLK_PERM_ALL & ~BLK_PERM_RESIZE;
  
@@ -59,7 +59,7 @@ index e2ea071315..4c3dc56c8e 100644
      }
  
      /* Clear the file by truncating it to 0 */
-@@ -2990,13 +2994,15 @@ raw_co_create(BlockdevCreateOptions *options, Error **errp)
+@@ -2976,13 +2980,15 @@ raw_co_create(BlockdevCreateOptions *options, Error **errp)
      }
  
  out_unlock:
@@ -82,7 +82,7 @@ index e2ea071315..4c3dc56c8e 100644
      }
  
  out_close:
-@@ -3020,6 +3026,7 @@ raw_co_create_opts(BlockDriver *drv, const char *filename,
+@@ -3006,6 +3012,7 @@ raw_co_create_opts(BlockDriver *drv, const char *filename,
      PreallocMode prealloc;
      char *buf = NULL;
      Error *local_err = NULL;
@@ -90,7 +90,7 @@ index e2ea071315..4c3dc56c8e 100644
  
      /* Skip file: protocol prefix */
      strstart(filename, "file:", &filename);
-@@ -3042,6 +3049,18 @@ raw_co_create_opts(BlockDriver *drv, const char *filename,
+@@ -3028,6 +3035,18 @@ raw_co_create_opts(BlockDriver *drv, const char *filename,
          return -EINVAL;
      }
  
@@ -109,7 +109,7 @@ index e2ea071315..4c3dc56c8e 100644
      options = (BlockdevCreateOptions) {
          .driver     = BLOCKDEV_DRIVER_FILE,
          .u.file     = {
-@@ -3053,6 +3072,8 @@ raw_co_create_opts(BlockDriver *drv, const char *filename,
+@@ -3039,6 +3058,8 @@ raw_co_create_opts(BlockDriver *drv, const char *filename,
              .nocow              = nocow,
              .has_extent_size_hint = has_extent_size_hint,
              .extent_size_hint   = extent_size_hint,
@@ -119,21 +119,10 @@ index e2ea071315..4c3dc56c8e 100644
      };
      return raw_co_create(&options, errp);
 diff --git a/qapi/block-core.json b/qapi/block-core.json
-index 48ba32049f..321d1fd0e1 100644
+index 903392cb8f..125aa89858 100644
 --- a/qapi/block-core.json
 +++ b/qapi/block-core.json
-@@ -4974,6 +4974,10 @@
- # @extent-size-hint: Extent size hint to add to the image file; 0 for
- #     not adding an extent size hint (default: 1 MB, since 5.1)
- #
-+# @locking: whether to enable file locking.  If set to 'auto', only
-+#     enable when Open File Descriptor (OFD) locking API is available
-+#     (default: auto).
-+#
- # Since: 2.12
- ##
- { 'struct': 'BlockdevCreateOptionsFile',
-@@ -4981,7 +4985,8 @@
+@@ -4876,7 +4876,8 @@
              'size':                 'size',
              '*preallocation':       'PreallocMode',
              '*nocow':               'bool',

debian/patches/pve/0023-PVE-monitor-disable-oob-capability.patch

@@ -18,10 +18,10 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  1 file changed, 1 insertion(+), 2 deletions(-)
 
 diff --git a/monitor/qmp.c b/monitor/qmp.c
-index eb181d5979..20fc0d20a6 100644
+index c15bf1e1fc..04fe25c62c 100644
 --- a/monitor/qmp.c
 +++ b/monitor/qmp.c
-@@ -534,8 +534,7 @@ void monitor_init_qmp(Chardev *chr, bool pretty, Error **errp)
+@@ -553,8 +553,7 @@ void monitor_init_qmp(Chardev *chr, bool pretty, Error **errp)
      qemu_chr_fe_set_echo(&mon->common.chr, true);
  
      /* Note: we run QMP monitor in I/O thread when @chr supports that */

debian/patches/pve/0024-PVE-Compat-4.0-used-balloon-qemu-4-0-config-size-fal.patch

@@ -26,10 +26,10 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  1 file changed, 2 insertions(+), 1 deletion(-)
 
 diff --git a/hw/core/machine.c b/hw/core/machine.c
-index f29fe95964..2c327fc36a 100644
+index f0d35c6401..1427983543 100644
 --- a/hw/core/machine.c
 +++ b/hw/core/machine.c
-@@ -180,7 +180,8 @@ GlobalProperty hw_compat_4_0[] = {
+@@ -148,7 +148,8 @@ GlobalProperty hw_compat_4_0[] = {
      { "virtio-vga",     "edid", "false" },
      { "virtio-gpu-device", "edid", "false" },
      { "virtio-device", "use-started", "false" },

debian/patches/pve/0025-PVE-Allow-version-code-in-machine-type.patch

@@ -16,15 +16,15 @@ Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
 ---
  hw/core/machine-qmp-cmds.c |  5 +++++
  include/hw/boards.h        |  2 ++
- qapi/machine.json          |  3 +++
- system/vl.c                | 24 ++++++++++++++++++++++++
- 4 files changed, 34 insertions(+)
+ qapi/machine.json          |  4 +++-
+ softmmu/vl.c               | 25 +++++++++++++++++++++++++
+ 4 files changed, 35 insertions(+), 1 deletion(-)
 
 diff --git a/hw/core/machine-qmp-cmds.c b/hw/core/machine-qmp-cmds.c
-index 52a6d74820..362128842d 100644
+index 40821e2317..ee93ddd69a 100644
 --- a/hw/core/machine-qmp-cmds.c
 +++ b/hw/core/machine-qmp-cmds.c
-@@ -94,6 +94,11 @@ MachineInfoList *qmp_query_machines(bool has_compat_props, bool compat_props,
+@@ -95,6 +95,11 @@ MachineInfoList *qmp_query_machines(Error **errp)
          if (strcmp(mc->name, MACHINE_GET_CLASS(current_machine)->name) == 0) {
              info->has_is_current = true;
              info->is_current = true;
@@ -37,64 +37,65 @@ index 52a6d74820..362128842d 100644
  
          if (mc->default_cpu_type) {
 diff --git a/include/hw/boards.h b/include/hw/boards.h
-index 36fbb9b59d..d1741ea121 100644
+index ed83360198..f8b88cd86a 100644
 --- a/include/hw/boards.h
 +++ b/include/hw/boards.h
-@@ -268,6 +268,8 @@ struct MachineClass {
+@@ -235,6 +235,8 @@ struct MachineClass {
      const char *desc;
      const char *deprecation_reason;
  
 +    const char *pve_version;
 +
      void (*init)(MachineState *state);
-     void (*reset)(MachineState *state, ResetType type);
+     void (*reset)(MachineState *state, ShutdownCause reason);
      void (*wakeup)(MachineState *state);
 diff --git a/qapi/machine.json b/qapi/machine.json
-index 16366b774a..12cfd3f260 100644
+index fbb61f18e4..7da3c519ba 100644
 --- a/qapi/machine.json
 +++ b/qapi/machine.json
-@@ -189,6 +189,8 @@
+@@ -161,6 +161,8 @@
  #
  # @acpi: machine type supports ACPI (since 8.0)
  #
 +# @pve-version: custom PVE version suffix specified as 'machine+pveN'
 +#
- # @compat-props: The machine type's compatibility properties.  Only
- #     present when query-machines argument @compat-props is true.
- #     (since 9.1)
-@@ -205,6 +207,7 @@
+ # Since: 1.2
+ ##
+ { 'struct': 'MachineInfo',
+@@ -168,7 +170,7 @@
+             '*is-default': 'bool', '*is-current': 'bool', 'cpu-max': 'int',
              'hotpluggable-cpus': 'bool',  'numa-mem-supported': 'bool',
              'deprecated': 'bool', '*default-cpu-type': 'str',
-             '*default-ram-id': 'str', 'acpi': 'bool',
-+            '*pve-version': 'str',
-             '*compat-props': { 'type': ['CompatProperty'],
-                                'features': ['unstable'] } } }
+-            '*default-ram-id': 'str', 'acpi': 'bool' } }
++            '*default-ram-id': 'str', 'acpi': 'bool', '*pve-version': 'str' } }
  
-diff --git a/system/vl.c b/system/vl.c
-index e7cae51f13..3f4916ac5a 100644
---- a/system/vl.c
-+++ b/system/vl.c
-@@ -1675,6 +1675,7 @@ static MachineClass *select_machine(QDict *qdict, Error **errp)
+ ##
+ # @query-machines:
+diff --git a/softmmu/vl.c b/softmmu/vl.c
+index 1b63ffd33d..20ba2c5c87 100644
+--- a/softmmu/vl.c
++++ b/softmmu/vl.c
+@@ -1597,6 +1597,7 @@ static const QEMUOption *lookup_opt(int argc, char **argv,
+ static MachineClass *select_machine(QDict *qdict, Error **errp)
  {
-     ERRP_GUARD();
-     const char *machine_type = qdict_get_try_str(qdict, "type");
+     const char *optarg = qdict_get_try_str(qdict, "type");
 +    const char *pvever = qdict_get_try_str(qdict, "pvever");
-     g_autoptr(GSList) machines = object_class_get_list(TYPE_MACHINE, false);
-     MachineClass *machine_class = NULL;
+     GSList *machines = object_class_get_list(TYPE_MACHINE, false);
+     MachineClass *machine_class;
+     Error *local_err = NULL;
+@@ -1614,6 +1615,11 @@ static MachineClass *select_machine(QDict *qdict, Error **errp)
+         }
+     }
  
-@@ -1694,7 +1695,11 @@ static MachineClass *select_machine(QDict *qdict, Error **errp)
-     if (!machine_class) {
-         error_append_hint(errp,
-                           "Use -machine help to list supported machines\n");
-+    } else {
++    if (machine_class) {
 +        machine_class->pve_version = g_strdup(pvever);
 +        qdict_del(qdict, "pvever");
-     }
++    }
 +
-     return machine_class;
- }
- 
-@@ -3316,12 +3321,31 @@ void qemu_init(int argc, char **argv)
+     g_slist_free(machines);
+     if (local_err) {
+         error_append_hint(&local_err, "Use -machine help to list supported machines\n");
+@@ -3244,12 +3250,31 @@ void qemu_init(int argc, char **argv)
              case QEMU_OPTION_machine:
                  {
                      bool help;

debian/patches/pve/0026-block-backup-move-bcs-bitmap-initialization-to-job-c.patch

@@ -25,7 +25,7 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  1 file changed, 4 insertions(+), 4 deletions(-)
 
 diff --git a/block/backup.c b/block/backup.c
-index a1292c01ec..2e38b30d67 100644
+index db3791f4d1..39410dcf8d 100644
 --- a/block/backup.c
 +++ b/block/backup.c
 @@ -237,8 +237,8 @@ static void backup_init_bcs_bitmap(BackupBlockJob *job)
@@ -48,9 +48,9 @@ index a1292c01ec..2e38b30d67 100644
      if (s->sync_mode == MIRROR_SYNC_MODE_TOP) {
          int64_t offset = 0;
          int64_t count;
-@@ -502,6 +500,8 @@ BlockJob *backup_job_create(const char *job_id, BlockDriverState *bs,
+@@ -495,6 +493,8 @@ BlockJob *backup_job_create(const char *job_id, BlockDriverState *bs,
+     block_job_add_bdrv(&job->common, "target", target, 0, BLK_PERM_ALL,
                         &error_abort);
-     bdrv_graph_wrunlock();
  
 +    backup_init_bcs_bitmap(job);
 +

debian/patches/pve/0027-PVE-Backup-add-vma-backup-format-code.patch

@@ -10,29 +10,28 @@ skip reads.
 
 Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
 [FE: improvements during create
-     allow partial restore
-     allow specifying disk formats for create operation]
+     allow partial restore]
 Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
 ---
  block/meson.build |   2 +
  meson.build       |   5 +
- vma-reader.c      | 867 ++++++++++++++++++++++++++++++++++++++++++
- vma-writer.c      | 816 ++++++++++++++++++++++++++++++++++++++++
- vma.c             | 941 ++++++++++++++++++++++++++++++++++++++++++++++
+ vma-reader.c      | 867 ++++++++++++++++++++++++++++++++++++++++++++
+ vma-writer.c      | 818 +++++++++++++++++++++++++++++++++++++++++
+ vma.c             | 900 ++++++++++++++++++++++++++++++++++++++++++++++
  vma.h             | 150 ++++++++
- 6 files changed, 2781 insertions(+)
+ 6 files changed, 2742 insertions(+)
  create mode 100644 vma-reader.c
  create mode 100644 vma-writer.c
  create mode 100644 vma.c
  create mode 100644 vma.h
 
 diff --git a/block/meson.build b/block/meson.build
-index 6a60b5d6b9..652c8cbdb7 100644
+index 1833c71ce9..59b71ba9f3 100644
 --- a/block/meson.build
 +++ b/block/meson.build
-@@ -42,6 +42,8 @@ block_ss.add(files(
+@@ -43,6 +43,8 @@ block_ss.add(files(
    'zeroinit.c',
- ), zstd, zlib)
+ ), zstd, zlib, gnutls)
  
 +block_ss.add(files('../vma-writer.c'), libuuid)
 +
@@ -40,10 +39,10 @@ index 6a60b5d6b9..652c8cbdb7 100644
  system_ss.add(files('block-ram-registrar.c'))
  
 diff --git a/meson.build b/meson.build
-index 147097c652..b9b673c271 100644
+index a9c4f28247..cd95530d3b 100644
 --- a/meson.build
 +++ b/meson.build
-@@ -2129,6 +2129,8 @@ endif
+@@ -1778,6 +1778,8 @@ endif
  
  has_gettid = cc.has_function('gettid')
  
@@ -52,19 +51,19 @@ index 147097c652..b9b673c271 100644
  # libselinux
  selinux = dependency('libselinux',
                       required: get_option('selinux'),
-@@ -4344,6 +4346,9 @@ if have_tools
-                dependencies: [blockdev, qemuutil, selinux],
+@@ -3908,6 +3910,9 @@ if have_tools
+                dependencies: [blockdev, qemuutil, gnutls, selinux],
                 install: true)
  
 +  vma = executable('vma', files('vma.c', 'vma-reader.c') + genh,
-+                   dependencies: [authz, block, crypto, io, qemuutil, qom], install: true)
++                   dependencies: [authz, block, crypto, io, qom], install: true)
 +
    subdir('storage-daemon')
- 
-   foreach exe: [ 'qemu-img', 'qemu-io', 'qemu-nbd', 'qemu-storage-daemon']
+   subdir('contrib/rdmacm-mux')
+   subdir('contrib/elf2dmp')
 diff --git a/vma-reader.c b/vma-reader.c
 new file mode 100644
-index 0000000000..bb65ad313c
+index 0000000000..81a891c6b1
 --- /dev/null
 +++ b/vma-reader.c
 @@ -0,0 +1,867 @@
@@ -89,7 +88,6 @@ index 0000000000..bb65ad313c
 +#include "qemu/ratelimit.h"
 +#include "vma.h"
 +#include "block/block.h"
-+#include "block/graph-lock.h"
 +#include "sysemu/block-backend.h"
 +
 +static unsigned char zero_vma_block[VMA_BLOCK_SIZE];
@@ -382,6 +380,7 @@ index 0000000000..bb65ad313c
 +    }
 +
 +
++    int count = 0;
 +    for (i = 1; i < 256; i++) {
 +        VmaDeviceInfoHeader *dih = &h->dev_info[i];
 +        uint32_t devname_ptr = GUINT32_FROM_BE(dih->devname_ptr);
@@ -389,6 +388,7 @@ index 0000000000..bb65ad313c
 +        const char *devname =  get_header_str(vmar, devname_ptr);
 +
 +        if (size && devname) {
++            count++;
 +            vmar->devinfo[i].size = size;
 +            vmar->devinfo[i].devname = devname;
 +
@@ -600,10 +600,8 @@ index 0000000000..bb65ad313c
 +    } else {
 +        int res = blk_pwrite(target, sector_num * BDRV_SECTOR_SIZE, nb_sectors * BDRV_SECTOR_SIZE, buf, 0);
 +        if (res < 0) {
-+            bdrv_graph_rdlock_main_loop();
 +            error_setg(errp, "blk_pwrite to %s failed (%d)",
 +                       bdrv_get_device_name(blk_bs(target)), res);
-+            bdrv_graph_rdunlock_main_loop();
 +            return -1;
 +        }
 +    }
@@ -883,7 +881,8 @@ index 0000000000..bb65ad313c
 +
 +            int64_t cluster_num, end;
 +
-+            end = DIV_ROUND_UP(vmar->devinfo[i].size, VMA_CLUSTER_SIZE);
++            end = (vmar->devinfo[i].size + VMA_CLUSTER_SIZE - 1) /
++                VMA_CLUSTER_SIZE;
 +
 +            for (cluster_num = 0; cluster_num < end; cluster_num++) {
 +                if (!vma_reader_get_bitmap(rstate, cluster_num)) {
@@ -937,10 +936,10 @@ index 0000000000..bb65ad313c
 +
 diff --git a/vma-writer.c b/vma-writer.c
 new file mode 100644
-index 0000000000..3f489092df
+index 0000000000..126b296647
 --- /dev/null
 +++ b/vma-writer.c
-@@ -0,0 +1,816 @@
+@@ -0,0 +1,818 @@
 +/*
 + * VMA: Virtual Machine Archive
 + *
@@ -1134,7 +1133,8 @@ index 0000000000..3f489092df
 +    vmaw->stream_info[n].devname = g_strdup(devname);
 +    vmaw->stream_info[n].size = size;
 +
-+    vmaw->stream_info[n].cluster_count = DIV_ROUND_UP(size, VMA_CLUSTER_SIZE);
++    vmaw->stream_info[n].cluster_count = (size + VMA_CLUSTER_SIZE - 1) /
++        VMA_CLUSTER_SIZE;
 +
 +    vmaw->stream_count = n;
 +
@@ -1514,16 +1514,17 @@ index 0000000000..3f489092df
 +    int i;
 +
 +    g_assert(vmaw != NULL);
-+    g_assert(status != NULL);
 +
-+    status->status = vmaw->status;
-+    g_strlcpy(status->errmsg, vmaw->errmsg, sizeof(status->errmsg));
-+    for (i = 0; i <= 255; i++) {
-+        status->stream_info[i] = vmaw->stream_info[i];
++    if (status) {
++        status->status = vmaw->status;
++        g_strlcpy(status->errmsg, vmaw->errmsg, sizeof(status->errmsg));
++        for (i = 0; i <= 255; i++) {
++            status->stream_info[i] = vmaw->stream_info[i];
++        }
++
++        uuid_unparse_lower(vmaw->uuid, status->uuid_str);
 +    }
 +
-+    uuid_unparse_lower(vmaw->uuid, status->uuid_str);
-+
 +    status->closed = vmaw->closed;
 +
 +    return vmaw->status;
@@ -1759,10 +1760,10 @@ index 0000000000..3f489092df
 +}
 diff --git a/vma.c b/vma.c
 new file mode 100644
-index 0000000000..8d4b4be414
+index 0000000000..347f6283ca
 --- /dev/null
 +++ b/vma.c
-@@ -0,0 +1,941 @@
+@@ -0,0 +1,900 @@
 +/*
 + * VMA: Virtual Machine Archive
 + *
@@ -1794,8 +1795,8 @@ index 0000000000..8d4b4be414
 +        "usage: vma command [command options]\n"
 +        "\n"
 +        "vma list <filename>\n"
-+        "vma config <filename> [-c <config>]\n"
-+        "vma create <filename> [-c <config>] [-d format=<format>:<device name>=<path> [-d ...]]\n"
++        "vma config <filename> [-c config]\n"
++        "vma create <filename> [-c config] pathname ...\n"
 +        "vma extract <filename> [-d <drive-list>] [-r <fifo>] <targetdir>\n"
 +        "vma verify <filename> [-v]\n"
 +        ;
@@ -2075,17 +2076,17 @@ index 0000000000..8d4b4be414
 +                        inbuf);
 +            }
 +
-+            RestoreMap *restore_map = g_new0(RestoreMap, 1);
-+            restore_map->devname = g_strdup(devname);
-+            restore_map->path = g_strdup(path);
-+            restore_map->format = format;
-+            restore_map->throttling_bps = bps_value;
-+            restore_map->throttling_group = group;
-+            restore_map->cache = cache;
-+            restore_map->write_zero = write_zero;
-+            restore_map->skip = skip;
++            RestoreMap *map = g_new0(RestoreMap, 1);
++            map->devname = g_strdup(devname);
++            map->path = g_strdup(path);
++            map->format = format;
++            map->throttling_bps = bps_value;
++            map->throttling_group = group;
++            map->cache = cache;
++            map->write_zero = write_zero;
++            map->skip = skip;
 +
-+            g_hash_table_insert(devmap, restore_map->devname, restore_map);
++            g_hash_table_insert(devmap, map->devname, map);
 +
 +        };
 +    }
@@ -2384,16 +2385,14 @@ index 0000000000..8d4b4be414
 +
 +static int create_archive(int argc, char **argv)
 +{
-+    int c;
++    int i, c;
 +    int verbose = 0;
-+    bool expect_format = true;
 +    const char *archivename;
 +    GList *backup_coroutines = NULL;
 +    GList *config_files = NULL;
-+    GList *disk_infos = NULL;
 +
 +    for (;;) {
-+        c = getopt(argc, argv, "hvc:d:");
++        c = getopt(argc, argv, "hvc:");
 +        if (c == -1) {
 +            break;
 +        }
@@ -2405,9 +2404,6 @@ index 0000000000..8d4b4be414
 +        case 'c':
 +            config_files = g_list_append(config_files, optarg);
 +            break;
-+        case 'd':
-+            disk_infos = g_list_append(disk_infos, optarg);
-+            break;
 +        case 'v':
 +            verbose = 1;
 +            break;
@@ -2453,48 +2449,16 @@ index 0000000000..8d4b4be414
 +        l = g_list_next(l);
 +    }
 +
-+    /*
-+     * Don't allow mixing new and old way to specifiy disks.
-+     * TODO PVE 9 drop old way and always require format.
-+     */
-+    if (optind < argc && g_list_first(disk_infos)) {
-+        unlink(archivename);
-+        g_error("Unexpected extra argument - specify all devices via '-d'");
-+    }
-+
-+    while (optind < argc) {
-+        expect_format = false;
-+        disk_infos = g_list_append(disk_infos, argv[optind++]);
-+    }
-+
 +    int devcount = 0;
-+    GList *disk_l = disk_infos;
-+    while (disk_l && disk_l->data) {
-+        char *disk_info = disk_l->data;
-+        const char *path = NULL;
++    while (optind < argc) {
++        const char *path = argv[optind++];
 +        char *devname = NULL;
-+        char *format = NULL;
-+        QDict *options = qdict_new();
-+
-+        if (try_parse_option(&disk_info, "format", &format, disk_info)) {
-+            qdict_put_str(options, "driver", format);
-+        } else {
-+            if (expect_format) {
-+                unlink(archivename);
-+                g_error("No format specified for device: '%s'", disk_info);
-+            } else {
-+                g_warning("Specifying a device without a format is deprecated"
-+                          " - use '-d format=<format>:%s'",
-+                          disk_info);
-+            }
-+        }
-+
-+        path = extract_devname(disk_info, &devname, devcount++);
++        path = extract_devname(path, &devname, devcount++);
 +
 +        Error *errp = NULL;
 +        BlockBackend *target;
 +
-+        target = blk_new_open(path, NULL, options, 0, &errp);
++        target = blk_new_open(path, NULL, NULL, 0, &errp);
 +        if (!target) {
 +            unlink(archivename);
 +            g_error("bdrv_open '%s' failed - %s", path, error_get_pretty(errp));
@@ -2516,8 +2480,6 @@ index 0000000000..8d4b4be414
 +        // Don't enter coroutine yet, because it might write the header before
 +        // all streams can be registered.
 +        backup_coroutines = g_list_append(backup_coroutines, co);
-+
-+        disk_l = g_list_next(disk_l);
 +    }
 +
 +    VmaStatus vmastat;
@@ -2581,7 +2543,6 @@ index 0000000000..8d4b4be414
 +    vma_writer_get_status(vmaw, &vmastat);
 +
 +    if (verbose) {
-+        int i;
 +        for (i = 0; i < 256; i++) {
 +            VmaStreamInfo *si = &vmastat.stream_info[i];
 +            if (si->size) {
@@ -2599,7 +2560,6 @@ index 0000000000..8d4b4be414
 +
 +    g_list_free(backup_coroutines);
 +    g_list_free(config_files);
-+    g_list_free(disk_infos);
 +    vma_writer_destroy(vmaw);
 +    return 0;
 +}

debian/patches/pve/0028-PVE-Backup-add-backup-dump-block-driver.patch

@@ -12,20 +12,20 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
 [FE: adapt to coroutine changes]
 Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
 ---
- block/backup-dump.c              | 172 +++++++++++++++++++++++++++++++
+ block/backup-dump.c              | 168 +++++++++++++++++++++++++++++++
  block/backup.c                   |  30 ++----
  block/meson.build                |   1 +
  include/block/block_int-common.h |  35 +++++++
  job.c                            |   3 +-
- 5 files changed, 218 insertions(+), 23 deletions(-)
+ 5 files changed, 214 insertions(+), 23 deletions(-)
  create mode 100644 block/backup-dump.c
 
 diff --git a/block/backup-dump.c b/block/backup-dump.c
 new file mode 100644
-index 0000000000..e46abf1070
+index 0000000000..232a094426
 --- /dev/null
 +++ b/block/backup-dump.c
-@@ -0,0 +1,172 @@
+@@ -0,0 +1,168 @@
 +/*
 + * BlockDriver to send backup data stream to a callback function
 + *
@@ -37,8 +37,6 @@ index 0000000000..e46abf1070
 + */
 +
 +#include "qemu/osdep.h"
-+
-+#include "qapi/qmp/qdict.h"
 +#include "qom/object_interfaces.h"
 +#include "block/block_int.h"
 +
@@ -171,7 +169,7 @@ index 0000000000..e46abf1070
 +block_init(bdrv_backup_dump_init);
 +
 +
-+BlockDriverState *coroutine_fn bdrv_co_backup_dump_create(
++BlockDriverState *bdrv_backup_dump_create(
 +    int dump_cb_block_size,
 +    uint64_t byte_size,
 +    BackupDumpFunc *dump_cb,
@@ -179,11 +177,9 @@ index 0000000000..e46abf1070
 +    Error **errp)
 +{
 +    BDRVBackupDumpState *state;
++    BlockDriverState *bs = bdrv_new_open_driver(
++        &bdrv_backup_dump_drive, NULL, BDRV_O_RDWR, errp);
 +
-+    QDict *options = qdict_new();
-+    qdict_put_str(options, "driver", "backup-dump-drive");
-+
-+    BlockDriverState *bs = bdrv_co_open(NULL, NULL, options, BDRV_O_RDWR, errp);
 +    if (!bs) {
 +        return NULL;
 +    }
@@ -199,7 +195,7 @@ index 0000000000..e46abf1070
 +    return bs;
 +}
 diff --git a/block/backup.c b/block/backup.c
-index 2e38b30d67..fe69723ada 100644
+index 39410dcf8d..af87fa6aa9 100644
 --- a/block/backup.c
 +++ b/block/backup.c
 @@ -29,28 +29,6 @@
@@ -231,7 +227,7 @@ index 2e38b30d67..fe69723ada 100644
  static const BlockJobDriver backup_job_driver;
  
  static void backup_cleanup_sync_bitmap(BackupBlockJob *job, int ret)
-@@ -462,6 +440,14 @@ BlockJob *backup_job_create(const char *job_id, BlockDriverState *bs,
+@@ -457,6 +435,14 @@ BlockJob *backup_job_create(const char *job_id, BlockDriverState *bs,
      }
  
      cluster_size = block_copy_cluster_size(bcs);
@@ -247,7 +243,7 @@ index 2e38b30d67..fe69723ada 100644
      if (perf->max_chunk && perf->max_chunk < cluster_size) {
          error_setg(errp, "Required max-chunk (%" PRIi64 ") is less than backup "
 diff --git a/block/meson.build b/block/meson.build
-index 652c8cbdb7..e1cf5a2e65 100644
+index 59b71ba9f3..6fde9f7dcd 100644
 --- a/block/meson.build
 +++ b/block/meson.build
 @@ -4,6 +4,7 @@ block_ss.add(files(
@@ -255,11 +251,11 @@ index 652c8cbdb7..e1cf5a2e65 100644
    'amend.c',
    'backup.c',
 +  'backup-dump.c',
+   'copy-before-write.c',
    'blkdebug.c',
    'blklogwrites.c',
-   'blkverify.c',
 diff --git a/include/block/block_int-common.h b/include/block/block_int-common.h
-index ebb4e56a50..e717a74e5f 100644
+index 74195c3004..0f2e1817ad 100644
 --- a/include/block/block_int-common.h
 +++ b/include/block/block_int-common.h
 @@ -26,6 +26,7 @@
@@ -276,7 +272,7 @@ index ebb4e56a50..e717a74e5f 100644
  
 +typedef int BackupDumpFunc(void *opaque, uint64_t offset, uint64_t bytes, const void *buf);
 +
-+BlockDriverState *coroutine_fn bdrv_co_backup_dump_create(
++BlockDriverState *bdrv_backup_dump_create(
 +    int dump_cb_block_size,
 +    uint64_t byte_size,
 +    BackupDumpFunc *dump_cb,
@@ -312,10 +308,10 @@ index ebb4e56a50..e717a74e5f 100644
      BDRV_TRACKED_READ,
      BDRV_TRACKED_WRITE,
 diff --git a/job.c b/job.c
-index 660ce22c56..baf54c8d60 100644
+index 72d57f0934..93e22d180b 100644
 --- a/job.c
 +++ b/job.c
-@@ -331,7 +331,8 @@ static bool job_started_locked(Job *job)
+@@ -330,7 +330,8 @@ static bool job_started_locked(Job *job)
  }
  
  /* Called with job_mutex held. */

debian/patches/pve/0029-PVE-Add-sequential-job-transaction-support.patch

@@ -11,10 +11,10 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  2 files changed, 46 insertions(+)
 
 diff --git a/include/qemu/job.h b/include/qemu/job.h
-index 2b873f2576..528cd6acb9 100644
+index e502787dd8..963cf2bef5 100644
 --- a/include/qemu/job.h
 +++ b/include/qemu/job.h
-@@ -362,6 +362,18 @@ void job_unlock(void);
+@@ -381,6 +381,18 @@ void job_unlock(void);
   */
  JobTxn *job_txn_new(void);
  
@@ -34,10 +34,10 @@ index 2b873f2576..528cd6acb9 100644
   * Release a reference that was previously acquired with job_txn_add_job or
   * job_txn_new. If it's the last reference to the object, it will be freed.
 diff --git a/job.c b/job.c
-index baf54c8d60..3ac5e5cde2 100644
+index 93e22d180b..2b31f1e14f 100644
 --- a/job.c
 +++ b/job.c
-@@ -94,6 +94,8 @@ struct JobTxn {
+@@ -93,6 +93,8 @@ struct JobTxn {
  
      /* Reference count */
      int refcnt;
@@ -46,7 +46,7 @@ index baf54c8d60..3ac5e5cde2 100644
  };
  
  void job_lock(void)
-@@ -119,6 +121,25 @@ JobTxn *job_txn_new(void)
+@@ -118,6 +120,25 @@ JobTxn *job_txn_new(void)
      return txn;
  }
  
@@ -72,7 +72,7 @@ index baf54c8d60..3ac5e5cde2 100644
  /* Called with job_mutex held. */
  static void job_txn_ref_locked(JobTxn *txn)
  {
-@@ -1042,6 +1063,12 @@ static void job_completed_txn_success_locked(Job *job)
+@@ -1057,6 +1078,12 @@ static void job_completed_txn_success_locked(Job *job)
       */
      QLIST_FOREACH(other_job, &txn->jobs, txn_list) {
          if (!job_is_completed_locked(other_job)) {
@@ -85,7 +85,7 @@ index baf54c8d60..3ac5e5cde2 100644
              return;
          }
          assert(other_job->ret == 0);
-@@ -1253,6 +1280,13 @@ int job_finish_sync_locked(Job *job,
+@@ -1268,6 +1295,13 @@ int job_finish_sync_locked(Job *job,
          return -EBUSY;
      }
  

debian/patches/pve/0030-PVE-Backup-Proxmox-backup-patches-for-QEMU.patch

@@ -84,31 +84,69 @@ Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
      create jobs in a drained section]
 Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
 ---
- block/meson.build              |    5 +
- block/monitor/block-hmp-cmds.c |   39 ++
- blockdev.c                     |    1 +
- hmp-commands-info.hx           |   14 +
- hmp-commands.hx                |   29 +
- include/monitor/hmp.h          |    3 +
- meson.build                    |    1 +
- monitor/hmp-cmds.c             |   72 +++
- proxmox-backup-client.c        |  146 +++++
- proxmox-backup-client.h        |   60 ++
- pve-backup.c                   | 1090 ++++++++++++++++++++++++++++++++
- qapi/block-core.json           |  233 +++++++
- qapi/common.json               |   14 +
- qapi/machine.json              |   16 +-
- 14 files changed, 1709 insertions(+), 14 deletions(-)
+ block/backup-dump.c              |   10 +-
+ block/meson.build                |    5 +
+ block/monitor/block-hmp-cmds.c   |   39 ++
+ blockdev.c                       |    1 +
+ hmp-commands-info.hx             |   14 +
+ hmp-commands.hx                  |   29 +
+ include/block/block_int-common.h |    2 +-
+ include/monitor/hmp.h            |    3 +
+ meson.build                      |    1 +
+ monitor/hmp-cmds.c               |   72 ++
+ proxmox-backup-client.c          |  146 ++++
+ proxmox-backup-client.h          |   60 ++
+ pve-backup.c                     | 1067 ++++++++++++++++++++++++++++++
+ qapi/block-core.json             |  229 +++++++
+ qapi/common.json                 |   14 +
+ qapi/machine.json                |   16 +-
+ 16 files changed, 1690 insertions(+), 18 deletions(-)
  create mode 100644 proxmox-backup-client.c
  create mode 100644 proxmox-backup-client.h
  create mode 100644 pve-backup.c
 
+diff --git a/block/backup-dump.c b/block/backup-dump.c
+index 232a094426..e46abf1070 100644
+--- a/block/backup-dump.c
++++ b/block/backup-dump.c
+@@ -9,6 +9,8 @@
+  */
+ 
+ #include "qemu/osdep.h"
++
++#include "qapi/qmp/qdict.h"
+ #include "qom/object_interfaces.h"
+ #include "block/block_int.h"
+ 
+@@ -141,7 +143,7 @@ static void bdrv_backup_dump_init(void)
+ block_init(bdrv_backup_dump_init);
+ 
+ 
+-BlockDriverState *bdrv_backup_dump_create(
++BlockDriverState *coroutine_fn bdrv_co_backup_dump_create(
+     int dump_cb_block_size,
+     uint64_t byte_size,
+     BackupDumpFunc *dump_cb,
+@@ -149,9 +151,11 @@ BlockDriverState *bdrv_backup_dump_create(
+     Error **errp)
+ {
+     BDRVBackupDumpState *state;
+-    BlockDriverState *bs = bdrv_new_open_driver(
+-        &bdrv_backup_dump_drive, NULL, BDRV_O_RDWR, errp);
+ 
++    QDict *options = qdict_new();
++    qdict_put_str(options, "driver", "backup-dump-drive");
++
++    BlockDriverState *bs = bdrv_co_open(NULL, NULL, options, BDRV_O_RDWR, errp);
+     if (!bs) {
+         return NULL;
+     }
 diff --git a/block/meson.build b/block/meson.build
-index e1cf5a2e65..2367e1ac1b 100644
+index 6fde9f7dcd..6d468f89e5 100644
 --- a/block/meson.build
 +++ b/block/meson.build
-@@ -44,6 +44,11 @@ block_ss.add(files(
- ), zstd, zlib)
+@@ -45,6 +45,11 @@ block_ss.add(files(
+ ), zstd, zlib, gnutls)
  
  block_ss.add(files('../vma-writer.c'), libuuid)
 +block_ss.add(files(
@@ -120,10 +158,10 @@ index e1cf5a2e65..2367e1ac1b 100644
  system_ss.add(when: 'CONFIG_TCG', if_true: files('blkreplay.c'))
  system_ss.add(files('block-ram-registrar.c'))
 diff --git a/block/monitor/block-hmp-cmds.c b/block/monitor/block-hmp-cmds.c
-index bdf2eb50b6..439a7a14c8 100644
+index ca2599de44..6efe28cef5 100644
 --- a/block/monitor/block-hmp-cmds.c
 +++ b/block/monitor/block-hmp-cmds.c
-@@ -1009,3 +1009,42 @@ void hmp_change_medium(Monitor *mon, const char *device, const char *target,
+@@ -1029,3 +1029,42 @@ void hmp_change_medium(Monitor *mon, const char *device, const char *target,
      qmp_blockdev_change_medium(device, NULL, target, arg, true, force,
                                 !!read_only, read_only_mode, errp);
  }
@@ -167,7 +205,7 @@ index bdf2eb50b6..439a7a14c8 100644
 +    hmp_handle_error(mon, error);
 +}
 diff --git a/blockdev.c b/blockdev.c
-index 79d47b1920..3f67eb413d 100644
+index 060d86a65f..79c3575612 100644
 --- a/blockdev.c
 +++ b/blockdev.c
 @@ -37,6 +37,7 @@
@@ -179,10 +217,10 @@ index 79d47b1920..3f67eb413d 100644
  #include "monitor/monitor.h"
  #include "qemu/error-report.h"
 diff --git a/hmp-commands-info.hx b/hmp-commands-info.hx
-index d1a7b99add..af588145ff 100644
+index 10fdd822e0..15937793c1 100644
 --- a/hmp-commands-info.hx
 +++ b/hmp-commands-info.hx
-@@ -458,6 +458,20 @@ SRST
+@@ -471,6 +471,20 @@ SRST
      Show the current VM UUID.
  ERST
  
@@ -204,7 +242,7 @@ index d1a7b99add..af588145ff 100644
      {
          .name       = "usernet",
 diff --git a/hmp-commands.hx b/hmp-commands.hx
-index 0c7c6f2c16..bf8315f226 100644
+index e352f86872..0c8b6725fb 100644
 --- a/hmp-commands.hx
 +++ b/hmp-commands.hx
 @@ -101,6 +101,35 @@ ERST
@@ -243,8 +281,21 @@ index 0c7c6f2c16..bf8315f226 100644
  ERST
  
      {
+diff --git a/include/block/block_int-common.h b/include/block/block_int-common.h
+index 0f2e1817ad..0a0339eee4 100644
+--- a/include/block/block_int-common.h
++++ b/include/block/block_int-common.h
+@@ -63,7 +63,7 @@
+ 
+ typedef int BackupDumpFunc(void *opaque, uint64_t offset, uint64_t bytes, const void *buf);
+ 
+-BlockDriverState *bdrv_backup_dump_create(
++BlockDriverState *coroutine_fn bdrv_co_backup_dump_create(
+     int dump_cb_block_size,
+     uint64_t byte_size,
+     BackupDumpFunc *dump_cb,
 diff --git a/include/monitor/hmp.h b/include/monitor/hmp.h
-index 2596cc2426..9dda91d65a 100644
+index 7a7def7530..cba7afe70c 100644
 --- a/include/monitor/hmp.h
 +++ b/include/monitor/hmp.h
 @@ -32,6 +32,7 @@ void hmp_info_savevm(Monitor *mon, const QDict *qdict);
@@ -255,7 +306,7 @@ index 2596cc2426..9dda91d65a 100644
  void hmp_info_cpus(Monitor *mon, const QDict *qdict);
  void hmp_info_vnc(Monitor *mon, const QDict *qdict);
  void hmp_info_spice(Monitor *mon, const QDict *qdict);
-@@ -82,6 +83,8 @@ void hmp_change_vnc(Monitor *mon, const char *device, const char *target,
+@@ -84,6 +85,8 @@ void hmp_change_vnc(Monitor *mon, const char *device, const char *target,
  void hmp_change_medium(Monitor *mon, const char *device, const char *target,
                         const char *arg, const char *read_only, bool force,
                         Error **errp);
@@ -265,10 +316,10 @@ index 2596cc2426..9dda91d65a 100644
  void hmp_device_add(Monitor *mon, const QDict *qdict);
  void hmp_device_del(Monitor *mon, const QDict *qdict);
 diff --git a/meson.build b/meson.build
-index b9b673c271..f6fb9b4fd8 100644
+index cd95530d3b..d53976d621 100644
 --- a/meson.build
 +++ b/meson.build
-@@ -2130,6 +2130,7 @@ endif
+@@ -1779,6 +1779,7 @@ endif
  has_gettid = cc.has_function('gettid')
  
  libuuid = cc.find_library('uuid', required: true)
@@ -277,18 +328,18 @@ index b9b673c271..f6fb9b4fd8 100644
  # libselinux
  selinux = dependency('libselinux',
 diff --git a/monitor/hmp-cmds.c b/monitor/hmp-cmds.c
-index 874084565f..bedeb81f8c 100644
+index 91be698308..5b9c231a4c 100644
 --- a/monitor/hmp-cmds.c
 +++ b/monitor/hmp-cmds.c
-@@ -22,6 +22,7 @@
+@@ -21,6 +21,7 @@
  #include "qemu/help_option.h"
  #include "monitor/monitor-internal.h"
  #include "qapi/error.h"
 +#include "qapi/qapi-commands-block-core.h"
  #include "qapi/qapi-commands-control.h"
- #include "qapi/qapi-commands-machine.h"
  #include "qapi/qapi-commands-migration.h"
-@@ -119,6 +120,77 @@ void hmp_sync_profile(Monitor *mon, const QDict *qdict)
+ #include "qapi/qapi-commands-misc.h"
+@@ -144,6 +145,77 @@ void hmp_sync_profile(Monitor *mon, const QDict *qdict)
      }
  }
  
@@ -586,10 +637,10 @@ index 0000000000..8cbf645b2c
 +#endif /* PROXMOX_BACKUP_CLIENT_H */
 diff --git a/pve-backup.c b/pve-backup.c
 new file mode 100644
-index 0000000000..fea0152de0
+index 0000000000..d84d807654
 --- /dev/null
 +++ b/pve-backup.c
-@@ -0,0 +1,1090 @@
+@@ -0,0 +1,1067 @@
 +#include "proxmox-backup-client.h"
 +#include "vma.h"
 +
@@ -600,7 +651,6 @@ index 0000000000..fea0152de0
 +#include "block/block_int-global-state.h"
 +#include "block/blockjob.h"
 +#include "block/dirty-bitmap.h"
-+#include "block/graph-lock.h"
 +#include "qapi/qapi-commands-block.h"
 +#include "qapi/qmp/qerror.h"
 +#include "qemu/cutils.h"
@@ -626,6 +676,7 @@ index 0000000000..fea0152de0
 + * ---end-bad-example--
 + *
 + * ==> Always use CoMutext inside coroutines.
++ * ==> Never acquire/release AioContext withing coroutines (because that use QemuRecMutex)
 + *
 + */
 +
@@ -678,6 +729,7 @@ index 0000000000..fea0152de0
 +    uint64_t block_size;
 +    uint8_t dev_id;
 +    int completed_ret; // INT_MAX if not completed
++    char targetfile[PATH_MAX];
 +    BdrvDirtyBitmap *bitmap;
 +    BlockDriverState *target;
 +    BlockJob *job;
@@ -898,12 +950,7 @@ index 0000000000..fea0152de0
 +
 +    qemu_co_mutex_lock(&backup_state.backup_mutex);
 +
-+    /*
-+     * All jobs in the transaction will be canceled when one receives an error.
-+     * The first error wins, so only set it for ECANCELED if it was the last
-+     * job. This allows more interesting errors from other jobs to win.
-+     */
-+    if (ret < 0 && (ret != -ECANCELED || !g_list_nth(backup_state.di_list, 1))) {
++    if (ret < 0) {
 +        Error *local_err = NULL;
 +        error_setg(&local_err, "job failed with err %d - %s", ret, strerror(-ret));
 +        pvebackup_propagate_error(local_err);
@@ -927,6 +974,13 @@ index 0000000000..fea0152de0
 +        }
 +    }
 +
++    if (di->job) {
++        WITH_JOB_LOCK_GUARD() {
++            job_unref_locked(&di->job->job);
++            di->job = NULL;
++        }
++    }
++
 +    // remove self from job list
 +    backup_state.di_list = g_list_remove(backup_state.di_list, di);
 +
@@ -946,16 +1000,6 @@ index 0000000000..fea0152de0
 +    di->completed_ret = ret;
 +
 +    /*
-+     * Needs to happen outside of coroutine, because it takes the graph write lock.
-+     */
-+    if (di->job) {
-+        WITH_JOB_LOCK_GUARD() {
-+            job_unref_locked(&di->job->job);
-+            di->job = NULL;
-+        }
-+    }
-+
-+    /*
 +     * Schedule stream cleanup in async coroutine. close_image and finish might
 +     * take a while, so we can't block on them here. This way it also doesn't
 +     * matter if we're already running in a coroutine or not.
@@ -1076,7 +1120,8 @@ index 0000000000..fea0152de0
 +}
 +
 +/*
-+ * backup_job_create can *not* be run from a coroutine, so this can't either.
++ * backup_job_create can *not* be run from a coroutine (and requires an
++ * acquired AioContext), so this can't either.
 + * The caller is responsible that backup_mutex is held nonetheless.
 + */
 +static void create_backup_jobs_bh(void *opaque) {
@@ -1109,6 +1154,9 @@ index 0000000000..fea0152de0
 +            sync_mode = MIRROR_SYNC_MODE_BITMAP;
 +            bitmap_mode = BITMAP_SYNC_MODE_ON_SUCCESS;
 +        }
++        AioContext *aio_context = bdrv_get_aio_context(di->bs);
++        aio_context_acquire(aio_context);
++
 +        bdrv_drained_begin(di->bs);
 +
 +        BlockJob *job = backup_job_create(
@@ -1119,6 +1167,8 @@ index 0000000000..fea0152de0
 +
 +        bdrv_drained_end(di->bs);
 +
++        aio_context_release(aio_context);
++
 +        di->job = job;
 +        if (job) {
 +            WITH_JOB_LOCK_GUARD() {
@@ -1169,66 +1219,6 @@ index 0000000000..fea0152de0
 +    aio_co_enter(data->ctx, data->co);
 +}
 +
-+/*
-+ * Returns a list of device infos, which needs to be freed by the caller. In
-+ * case of an error, errp will be set, but the returned value might still be a
-+ * list.
-+ */
-+static GList coroutine_fn GRAPH_RDLOCK *get_device_info(
-+    const char *devlist,
-+    Error **errp)
-+{
-+    gchar **devs = NULL;
-+    GList *di_list = NULL;
-+
-+    if (devlist) {
-+        devs = g_strsplit_set(devlist, ",;:", -1);
-+
-+        gchar **d = devs;
-+        while (d && *d) {
-+            BlockBackend *blk = blk_by_name(*d);
-+            if (!blk) {
-+                error_set(errp, ERROR_CLASS_DEVICE_NOT_FOUND,
-+                          "Device '%s' not found", *d);
-+                goto err;
-+            }
-+            BlockDriverState *bs = blk_bs(blk);
-+            if (!bdrv_co_is_inserted(bs)) {
-+                error_setg(errp, "Device '%s' has no medium", *d);
-+                goto err;
-+            }
-+            PVEBackupDevInfo *di = g_new0(PVEBackupDevInfo, 1);
-+            di->bs = bs;
-+            di_list = g_list_append(di_list, di);
-+            d++;
-+        }
-+    } else {
-+        BdrvNextIterator it;
-+
-+        for (BlockDriverState *bs = bdrv_first(&it); bs; bs = bdrv_next(&it)) {
-+            if (!bdrv_co_is_inserted(bs) || bdrv_is_read_only(bs)) {
-+                continue;
-+            }
-+
-+            PVEBackupDevInfo *di = g_new0(PVEBackupDevInfo, 1);
-+            di->bs = bs;
-+            di_list = g_list_append(di_list, di);
-+        }
-+    }
-+
-+    if (!di_list) {
-+        error_set(errp, ERROR_CLASS_GENERIC_ERROR, "empty device list");
-+        goto err;
-+    }
-+
-+err:
-+    if (devs) {
-+        g_strfreev(devs);
-+    }
-+
-+    return di_list;
-+}
-+
 +UuidInfo coroutine_fn *qmp_backup(
 +    const char *backup_file,
 +    const char *password,
@@ -1254,10 +1244,13 @@ index 0000000000..fea0152de0
 +
 +    qemu_co_mutex_lock(&backup_state.backup_mutex);
 +
++    BlockBackend *blk;
++    BlockDriverState *bs = NULL;
 +    Error *local_err = NULL;
 +    uuid_t uuid;
 +    VmaWriter *vmaw = NULL;
 +    ProxmoxBackupHandle *pbs = NULL;
++    gchar **devs = NULL;
 +    GList *di_list = NULL;
 +    GList *l;
 +    UuidInfo *uuid_info;
@@ -1275,14 +1268,48 @@ index 0000000000..fea0152de0
 +    /* Todo: try to auto-detect format based on file name */
 +    format = has_format ? format : BACKUP_FORMAT_VMA;
 +
-+    bdrv_graph_co_rdlock();
-+    di_list = get_device_info(devlist, &local_err);
-+    bdrv_graph_co_rdunlock();
-+    if (local_err) {
-+        error_propagate(errp, local_err);
++    if (devlist) {
++        devs = g_strsplit_set(devlist, ",;:", -1);
++
++        gchar **d = devs;
++        while (d && *d) {
++            blk = blk_by_name(*d);
++            if (blk) {
++                bs = blk_bs(blk);
++                if (!bdrv_co_is_inserted(bs)) {
++                    error_setg(errp, QERR_DEVICE_HAS_NO_MEDIUM, *d);
++                    goto err;
++                }
++                PVEBackupDevInfo *di = g_new0(PVEBackupDevInfo, 1);
++                di->bs = bs;
++                di_list = g_list_append(di_list, di);
++            } else {
++                error_set(errp, ERROR_CLASS_DEVICE_NOT_FOUND,
++                          "Device '%s' not found", *d);
++                goto err;
++            }
++            d++;
++        }
++
++    } else {
++        BdrvNextIterator it;
++
++        bs = NULL;
++        for (bs = bdrv_first(&it); bs; bs = bdrv_next(&it)) {
++            if (!bdrv_co_is_inserted(bs) || bdrv_is_read_only(bs)) {
++                continue;
++            }
++
++            PVEBackupDevInfo *di = g_new0(PVEBackupDevInfo, 1);
++            di->bs = bs;
++            di_list = g_list_append(di_list, di);
++        }
++    }
++
++    if (!di_list) {
++        error_set(errp, ERROR_CLASS_GENERIC_ERROR, "empty device list");
 +        goto err;
 +    }
-+    assert(di_list);
 +
 +    size_t total = 0;
 +
@@ -1290,11 +1317,7 @@ index 0000000000..fea0152de0
 +    while (l) {
 +        PVEBackupDevInfo *di = (PVEBackupDevInfo *)l->data;
 +        l = g_list_next(l);
-+
-+        bdrv_graph_co_rdlock();
-+        bool blocked = bdrv_op_is_blocked(di->bs, BLOCK_OP_TYPE_BACKUP_SOURCE, errp);
-+        bdrv_graph_co_rdunlock();
-+        if (blocked) {
++        if (bdrv_op_is_blocked(di->bs, BLOCK_OP_TYPE_BACKUP_SOURCE, errp)) {
 +            goto err;
 +        }
 +
@@ -1378,9 +1401,7 @@ index 0000000000..fea0152de0
 +
 +            di->block_size = dump_cb_block_size;
 +
-+            bdrv_graph_co_rdlock();
 +            const char *devname = bdrv_get_device_name(di->bs);
-+            bdrv_graph_co_rdunlock();
 +            PBSBitmapAction action = PBS_BITMAP_ACTION_NOT_USED;
 +            size_t dirty = di->size;
 +
@@ -1440,7 +1461,9 @@ index 0000000000..fea0152de0
 +    } else if (format == BACKUP_FORMAT_VMA) {
 +        vmaw = vma_writer_create(backup_file, uuid, &local_err);
 +        if (!vmaw) {
-+            error_propagate(errp, local_err);
++            if (local_err) {
++                error_propagate(errp, local_err);
++            }
 +            goto err_mutex;
 +        }
 +
@@ -1454,9 +1477,7 @@ index 0000000000..fea0152de0
 +                goto err_mutex;
 +            }
 +
-+            bdrv_graph_co_rdlock();
 +            const char *devname = bdrv_get_device_name(di->bs);
-+            bdrv_graph_co_rdunlock();
 +            di->dev_id = vma_writer_register_stream(vmaw, devname, di->size);
 +            if (di->dev_id <= 0) {
 +                error_set(errp, ERROR_CLASS_GENERIC_ERROR,
@@ -1568,11 +1589,18 @@ index 0000000000..fea0152de0
 +            bdrv_co_unref(di->target);
 +        }
 +
++        if (di->targetfile[0]) {
++            unlink(di->targetfile);
++        }
 +        g_free(di);
 +    }
 +    g_list_free(di_list);
 +    backup_state.di_list = NULL;
 +
++    if (devs) {
++        g_strfreev(devs);
++    }
++
 +    if (vmaw) {
 +        Error *err = NULL;
 +        vma_writer_close(vmaw, &err);
@@ -1681,10 +1709,10 @@ index 0000000000..fea0152de0
 +    return ret;
 +}
 diff --git a/qapi/block-core.json b/qapi/block-core.json
-index 321d1fd0e1..68caf30084 100644
+index 125aa89858..331c8336d1 100644
 --- a/qapi/block-core.json
 +++ b/qapi/block-core.json
-@@ -851,6 +851,239 @@
+@@ -839,6 +839,235 @@
  { 'command': 'query-block', 'returns': ['BlockInfo'],
    'allow-preconfig': true }
  
@@ -1753,9 +1781,6 @@ index 321d1fd0e1..68caf30084 100644
 +# @config-file: a configuration file to include into
 +#               the backup archive.
 +#
-+# @firewall-file: a firewall configuration file to include into the backup
-+#     archive.
-+#
 +# @speed: the maximum speed, in bytes per second
 +#
 +# @devlist: list of block device names (separated by ',', ';'
@@ -1823,7 +1848,9 @@ index 321d1fd0e1..68caf30084 100644
 +#
 +# Cancel the current executing backup process.
 +#
-+# .. note:: This command succeeds even if there is no backup process running.
++# Returns: nothing on success
++#
++# Notes: This command succeeds even if there is no backup process running.
 +#
 +##
 +{ 'command': 'backup-cancel', 'coroutine': true }
@@ -1846,9 +1873,6 @@ index 321d1fd0e1..68caf30084 100644
 +#
 +# @pbs-library-version: Running version of libproxmox-backup-qemu0 library.
 +#
-+# @backup-max-workers: Whether the 'max-workers' @BackupPerf setting is
-+#     supported or not.
-+#
 +##
 +{ 'struct': 'ProxmoxSupportStatus',
 +  'data': { 'pbs-dirty-bitmap': 'bool',
@@ -1925,10 +1949,10 @@ index 321d1fd0e1..68caf30084 100644
  # @BlockDeviceTimedStats:
  #
 diff --git a/qapi/common.json b/qapi/common.json
-index 6ffc7a3789..9c6c671ece 100644
+index 6fed9cde1a..630a2a8f9a 100644
 --- a/qapi/common.json
 +++ b/qapi/common.json
-@@ -212,3 +212,17 @@
+@@ -207,3 +207,17 @@
  ##
  { 'struct': 'HumanReadableText',
    'data': { 'human-readable-text': 'str' } }
@@ -1942,12 +1966,12 @@ index 6ffc7a3789..9c6c671ece 100644
 +#
 +# Since: 0.14.0
 +#
-+# .. note:: If no UUID was specified for the guest, a null UUID is
++# Notes: If no UUID was specified for the guest, a null UUID is
 +#     returned.
 +##
 +{ 'struct': 'UuidInfo', 'data': {'UUID': 'str'} }
 diff --git a/qapi/machine.json b/qapi/machine.json
-index 12cfd3f260..a8abdb42a3 100644
+index 7da3c519ba..888457f810 100644
 --- a/qapi/machine.json
 +++ b/qapi/machine.json
 @@ -4,6 +4,8 @@
@@ -1959,7 +1983,7 @@ index 12cfd3f260..a8abdb42a3 100644
  ##
  # = Machines
  ##
-@@ -302,20 +304,6 @@
+@@ -230,20 +232,6 @@
  ##
  { 'command': 'query-target', 'returns': 'TargetInfo' }
  
@@ -1972,8 +1996,8 @@ index 12cfd3f260..a8abdb42a3 100644
 -#
 -# Since: 0.14
 -#
--# .. note:: If no UUID was specified for the guest, the nil UUID (all
--#    zeroes) is returned.
+-# Notes: If no UUID was specified for the guest, a null UUID is
+-#     returned.
 -##
 -{ 'struct': 'UuidInfo', 'data': {'UUID': 'str'} }
 -

debian/patches/pve/0031-PVE-Backup-pbs-restore-new-command-to-restore-from-p.patch

@@ -14,20 +14,20 @@ Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
  create mode 100644 pbs-restore.c
 
 diff --git a/meson.build b/meson.build
-index f6fb9b4fd8..f666d0f028 100644
+index d53976d621..c3330310d9 100644
 --- a/meson.build
 +++ b/meson.build
-@@ -4350,6 +4350,10 @@ if have_tools
+@@ -3914,6 +3914,10 @@ if have_tools
    vma = executable('vma', files('vma.c', 'vma-reader.c') + genh,
-                    dependencies: [authz, block, crypto, io, qemuutil, qom], install: true)
+                    dependencies: [authz, block, crypto, io, qom], install: true)
  
 +  pbs_restore = executable('pbs-restore', files('pbs-restore.c') + genh,
-+                  dependencies: [authz, block, crypto, io, qemuutil, qom,
++                  dependencies: [authz, block, crypto, io, qom,
 +                    libproxmox_backup_qemu], install: true)
 +
    subdir('storage-daemon')
- 
-   foreach exe: [ 'qemu-img', 'qemu-io', 'qemu-nbd', 'qemu-storage-daemon']
+   subdir('contrib/rdmacm-mux')
+   subdir('contrib/elf2dmp')
 diff --git a/pbs-restore.c b/pbs-restore.c
 new file mode 100644
 index 0000000000..f03d9bab8d

debian/patches/pve/0032-PVE-Add-PBS-block-driver-to-map-backup-archives-into.patch

@@ -14,33 +14,35 @@ Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
      getlength is now a coroutine function]
 Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
 ---
- block/meson.build    |   2 +
- block/pbs.c          | 306 +++++++++++++++++++++++++++++++++++++++++++
+ block/meson.build    |   3 +
+ block/pbs.c          | 305 +++++++++++++++++++++++++++++++++++++++++++
+ configure            |   9 ++
  meson.build          |   2 +-
- qapi/block-core.json |  29 ++++
+ qapi/block-core.json |  13 ++
  qapi/pragma.json     |   1 +
- 5 files changed, 339 insertions(+), 1 deletion(-)
+ 6 files changed, 332 insertions(+), 1 deletion(-)
  create mode 100644 block/pbs.c
 
 diff --git a/block/meson.build b/block/meson.build
-index 2367e1ac1b..e178047ec9 100644
+index 6d468f89e5..becc99ac4e 100644
 --- a/block/meson.build
 +++ b/block/meson.build
-@@ -49,6 +49,8 @@ block_ss.add(files(
+@@ -50,6 +50,9 @@ block_ss.add(files(
    '../pve-backup.c',
  ), libproxmox_backup_qemu)
  
-+block_ss.add(files('pbs.c'), libproxmox_backup_qemu)
++block_ss.add(when: 'CONFIG_PBS_BDRV', if_true: files('pbs.c'))
++block_ss.add(when: 'CONFIG_PBS_BDRV', if_true: libproxmox_backup_qemu)
 +
  
  system_ss.add(when: 'CONFIG_TCG', if_true: files('blkreplay.c'))
  system_ss.add(files('block-ram-registrar.c'))
 diff --git a/block/pbs.c b/block/pbs.c
 new file mode 100644
-index 0000000000..2d5e28ce8f
+index 0000000000..a2211e0f3b
 --- /dev/null
 +++ b/block/pbs.c
-@@ -0,0 +1,306 @@
+@@ -0,0 +1,305 @@
 +/*
 + * Proxmox Backup Server read-only block driver
 + */
@@ -68,7 +70,7 @@ index 0000000000..2d5e28ce8f
 +
 +typedef struct {
 +    ProxmoxRestoreHandle *conn;
-+    uint8_t aid;
++    char aid;
 +    int64_t length;
 +
 +    char *repository;
@@ -201,18 +203,12 @@ index 0000000000..2d5e28ce8f
 +    }
 +
 +    /* acquire handle and length */
-+    ret = proxmox_restore_open_image(s->conn, s->archive, &pbs_error);
-+    if (ret < 0) {
++    s->aid = proxmox_restore_open_image(s->conn, s->archive, &pbs_error);
++    if (s->aid < 0) {
 +        if (pbs_error && errp) error_setg(errp, "PBS open_image failed: %s", pbs_error);
 +        if (pbs_error) proxmox_backup_free_error(pbs_error);
 +        return -ENODEV;
 +    }
-+    if (ret > UINT8_MAX) {
-+        error_setg(errp, "PBS open_image returned an ID larger than %u", UINT8_MAX);
-+        return -ENODEV;
-+    }
-+    s->aid = ret;
-+
 +    s->length = proxmox_restore_get_image_length(s->conn, s->aid, &pbs_error);
 +    if (s->length < 0) {
 +        if (pbs_error && errp) error_setg(errp, "PBS get_image_length failed: %s", pbs_error);
@@ -223,6 +219,12 @@ index 0000000000..2d5e28ce8f
 +    return 0;
 +}
 +
++static int pbs_file_open(BlockDriverState *bs, QDict *options, int flags,
++                         Error **errp)
++{
++    return pbs_open(bs, options, flags, errp);
++}
++
 +static void pbs_close(BlockDriverState *bs) {
 +    BDRVPBSState *s = bs->opaque;
 +    g_free(s->repository);
@@ -232,8 +234,7 @@ index 0000000000..2d5e28ce8f
 +    proxmox_restore_disconnect(s->conn);
 +}
 +
-+static coroutine_fn int64_t GRAPH_RDLOCK
-+pbs_co_getlength(BlockDriverState *bs)
++static coroutine_fn int64_t pbs_co_getlength(BlockDriverState *bs)
 +{
 +    BDRVPBSState *s = bs->opaque;
 +    return s->length;
@@ -250,9 +251,9 @@ index 0000000000..2d5e28ce8f
 +    aio_co_schedule(rcb->ctx, rcb->co);
 +}
 +
-+static coroutine_fn int GRAPH_RDLOCK
-+pbs_co_preadv(BlockDriverState *bs, int64_t offset, int64_t bytes,
-+              QEMUIOVector *qiov, BdrvRequestFlags flags)
++static coroutine_fn int pbs_co_preadv(BlockDriverState *bs,
++                                      int64_t offset, int64_t bytes,
++                                      QEMUIOVector *qiov, BdrvRequestFlags flags)
 +{
 +    BDRVPBSState *s = bs->opaque;
 +    int ret;
@@ -297,17 +298,16 @@ index 0000000000..2d5e28ce8f
 +    return 0;
 +}
 +
-+static coroutine_fn int GRAPH_RDLOCK
-+pbs_co_pwritev(BlockDriverState *bs, int64_t offset, int64_t bytes,
-+               QEMUIOVector *qiov, BdrvRequestFlags flags)
++static coroutine_fn int pbs_co_pwritev(BlockDriverState *bs,
++                                       int64_t offset, int64_t bytes,
++                                       QEMUIOVector *qiov, BdrvRequestFlags flags)
 +{
 +    fprintf(stderr, "pbs-bdrv: cannot write to backup file, make sure "
 +           "any attached disk devices are set to read-only!\n");
 +    return -EPERM;
 +}
 +
-+static void GRAPH_RDLOCK
-+pbs_refresh_filename(BlockDriverState *bs)
++static void pbs_refresh_filename(BlockDriverState *bs)
 +{
 +    BDRVPBSState *s = bs->opaque;
 +    if (s->namespace) {
@@ -330,6 +330,7 @@ index 0000000000..2d5e28ce8f
 +
 +    .bdrv_parse_filename    = pbs_parse_filename,
 +
++    .bdrv_file_open         = pbs_file_open,
 +    .bdrv_open              = pbs_open,
 +    .bdrv_close             = pbs_close,
 +    .bdrv_co_getlength      = pbs_co_getlength,
@@ -347,24 +348,65 @@ index 0000000000..2d5e28ce8f
 +}
 +
 +block_init(bdrv_pbs_init);
+diff --git a/configure b/configure
+index 133f4e3235..f5a830c1f3 100755
+--- a/configure
++++ b/configure
+@@ -256,6 +256,7 @@ qemu_suffix="qemu"
+ softmmu="yes"
+ linux_user=""
+ bsd_user=""
++pbs_bdrv="yes"
+ plugins="$default_feature"
+ ninja=""
+ python=
+@@ -809,6 +810,10 @@ for opt do
+   ;;
+   --enable-download) download="enabled"; git_submodules_action=update;
+   ;;
++  --disable-pbs-bdrv) pbs_bdrv="no"
++  ;;
++  --enable-pbs-bdrv) pbs_bdrv="yes"
++  ;;
+   --enable-plugins) if test "$mingw32" = "yes"; then
+                         error_exit "TCG plugins not currently supported on Windows platforms"
+                     else
+@@ -959,6 +964,7 @@ cat << EOF
+   bsd-user        all BSD usermode emulation targets
+   pie             Position Independent Executables
+   debug-tcg       TCG debugging (default is disabled)
++  pbs-bdrv        Proxmox backup server read-only block driver support
+ 
+ NOTE: The object files are built at the place where configure is launched
+ EOF
+@@ -1744,6 +1750,9 @@ if test "$solaris" = "yes" ; then
+ fi
+ echo "SRC_PATH=$source_path" >> $config_host_mak
+ echo "TARGET_DIRS=$target_list" >> $config_host_mak
++if test "$pbs_bdrv" = "yes" ; then
++  echo "CONFIG_PBS_BDRV=y" >> $config_host_mak
++fi
+ 
+ # XXX: suppress that
+ if [ "$bsd" = "yes" ] ; then
 diff --git a/meson.build b/meson.build
-index f666d0f028..4c85736ec3 100644
+index c3330310d9..cbfc9a43fb 100644
 --- a/meson.build
 +++ b/meson.build
-@@ -4815,7 +4815,7 @@ summary_info += {'Query Processing Library support': qpl}
- summary_info += {'UADK Library support': uadk}
- summary_info += {'qatzip support':    qatzip}
+@@ -4319,7 +4319,7 @@ summary_info += {'bzip2 support':     libbzip2}
+ summary_info += {'lzfse support':     liblzfse}
+ summary_info += {'zstd support':      zstd}
  summary_info += {'NUMA host support': numa}
 -summary_info += {'capstone':          capstone}
 +summary_info += {'PBS bdrv support':  config_host.has_key('CONFIG_PBS_BDRV')}
  summary_info += {'libpmem support':   libpmem}
  summary_info += {'libdaxctl support': libdaxctl}
- summary_info += {'libcbor support':   libcbor}
+ summary_info += {'libudev':           libudev}
 diff --git a/qapi/block-core.json b/qapi/block-core.json
-index 68caf30084..d45e8975a7 100644
+index 331c8336d1..a818d5f90f 100644
 --- a/qapi/block-core.json
 +++ b/qapi/block-core.json
-@@ -3466,6 +3466,7 @@
+@@ -3396,6 +3396,7 @@
              'parallels', 'preallocate', 'qcow', 'qcow2', 'qed', 'quorum',
              'raw', 'rbd',
              { 'name': 'replication', 'if': 'CONFIG_REPLICATION' },
@@ -372,7 +414,7 @@ index 68caf30084..d45e8975a7 100644
              'ssh', 'throttle', 'vdi', 'vhdx',
              { 'name': 'virtio-blk-vfio-pci', 'if': 'CONFIG_BLKIO' },
              { 'name': 'virtio-blk-vhost-user', 'if': 'CONFIG_BLKIO' },
-@@ -3552,6 +3553,33 @@
+@@ -3482,6 +3483,17 @@
  { 'struct': 'BlockdevOptionsNull',
    'data': { '*size': 'int', '*latency-ns': 'uint64', '*read-zeroes': 'bool' } }
  
@@ -381,22 +423,6 @@ index 68caf30084..d45e8975a7 100644
 +#
 +# Driver specific block device options for the PBS backend.
 +#
-+# @repository: Proxmox Backup Server repository.
-+#
-+# @snapshot: backup snapshots ID.
-+#
-+# @archive: archive name.
-+#
-+# @keyfile: keyfile to use for encryption.
-+#
-+# @password: password to use for connection.
-+#
-+# @fingerprint: backup server fingerprint.
-+#
-+# @key_password: password to unlock key.
-+#
-+# @namespace: namespace where backup snapshot lives.
-+#
 +##
 +{ 'struct': 'BlockdevOptionsPbs',
 +  'data': { 'repository': 'str', 'snapshot': 'str', 'archive': 'str',
@@ -406,7 +432,7 @@ index 68caf30084..d45e8975a7 100644
  ##
  # @BlockdevOptionsNVMe:
  #
-@@ -4993,6 +5021,7 @@
+@@ -4886,6 +4898,7 @@
        'nfs':        'BlockdevOptionsNfs',
        'null-aio':   'BlockdevOptionsNull',
        'null-co':    'BlockdevOptionsNull',
@@ -415,10 +441,10 @@ index 68caf30084..d45e8975a7 100644
        'nvme-io_uring': { 'type': 'BlockdevOptionsNvmeIoUring',
                           'if': 'CONFIG_BLKIO' },
 diff --git a/qapi/pragma.json b/qapi/pragma.json
-index 6aaa9cb975..e9c595c4ba 100644
+index 325e684411..b6079f6a0e 100644
 --- a/qapi/pragma.json
 +++ b/qapi/pragma.json
-@@ -91,6 +91,7 @@
+@@ -45,6 +45,7 @@
          'BlockInfo',                # query-block
          'BlockdevAioOptions',       # blockdev-add, -blockdev
          'BlockdevDriver',           # blockdev-add, query-blockstats, ...

debian/patches/pve/0033-PVE-redirect-stderr-to-journal-when-daemonized.patch

@@ -9,15 +9,15 @@ fitting.
 Signed-off-by: Stefan Reiter <s.reiter@proxmox.com>
 Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
 ---
- meson.build | 3 ++-
+ meson.build | 2 ++
  os-posix.c  | 7 +++++--
- 2 files changed, 7 insertions(+), 3 deletions(-)
+ 2 files changed, 7 insertions(+), 2 deletions(-)
 
 diff --git a/meson.build b/meson.build
-index 4c85736ec3..57f666d722 100644
+index cbfc9a43fb..8206270272 100644
 --- a/meson.build
 +++ b/meson.build
-@@ -2130,6 +2130,7 @@ endif
+@@ -1779,6 +1779,7 @@ endif
  has_gettid = cc.has_function('gettid')
  
  libuuid = cc.find_library('uuid', required: true)
@@ -25,29 +25,28 @@ index 4c85736ec3..57f666d722 100644
  libproxmox_backup_qemu = cc.find_library('proxmox_backup_qemu', required: true)
  
  # libselinux
-@@ -3744,7 +3745,7 @@ if have_block
-   if host_os == 'windows'
-     system_ss.add(files('os-win32.c'))
-   else
--    blockdev_ss.add(files('os-posix.c'))
-+    blockdev_ss.add(files('os-posix.c'), libsystemd)
-   endif
+@@ -3406,6 +3407,7 @@ if have_block
+   # os-posix.c contains POSIX-specific functions used by qemu-storage-daemon,
+   # os-win32.c does not
+   blockdev_ss.add(when: 'CONFIG_POSIX', if_true: files('os-posix.c'))
++  blockdev_ss.add(when: 'CONFIG_POSIX', if_true: libsystemd)
+   system_ss.add(when: 'CONFIG_WIN32', if_true: [files('os-win32.c')])
  endif
  
 diff --git a/os-posix.c b/os-posix.c
-index 43f9a43f3f..a47e46d1c2 100644
+index cfcb96533c..fb2ad87009 100644
 --- a/os-posix.c
 +++ b/os-posix.c
-@@ -29,6 +29,8 @@
+@@ -28,6 +28,8 @@
  #include <pwd.h>
  #include <grp.h>
  #include <libgen.h>
 +#include <systemd/sd-journal.h>
 +#include <syslog.h>
  
- #include "qemu/error-report.h"
- #include "qemu/log.h"
-@@ -306,9 +308,10 @@ void os_setup_post(void)
+ /* Needed early for CONFIG_BSD etc. */
+ #include "net/slirp.h"
+@@ -310,9 +312,10 @@ void os_setup_post(void)
  
          dup2(fd, 0);
          dup2(fd, 1);

debian/patches/pve/0034-PVE-Migrate-dirty-bitmap-state-via-savevm.patch

@@ -26,52 +26,46 @@ Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
  create mode 100644 migration/pbs-state.c
 
 diff --git a/include/migration/misc.h b/include/migration/misc.h
-index 804eb23c06..c75b146ae6 100644
+index 7dcc0b5c2c..4c940b2475 100644
 --- a/include/migration/misc.h
 +++ b/include/migration/misc.h
-@@ -106,4 +106,7 @@ bool migration_incoming_postcopy_advised(void);
- /* True if background snapshot is active */
- bool migration_in_bg_snapshot(void);
+@@ -77,4 +77,7 @@ bool migration_in_bg_snapshot(void);
+ /* migration/block-dirty-bitmap.c */
+ void dirty_bitmap_mig_init(void);
  
 +/* migration/pbs-state.c */
 +void pbs_state_mig_init(void);
 +
  #endif
 diff --git a/migration/meson.build b/migration/meson.build
-index 075b013971..eca57cb2a3 100644
+index 07f6057acc..343994d891 100644
 --- a/migration/meson.build
 +++ b/migration/meson.build
-@@ -8,6 +8,7 @@ migration_files = files(
+@@ -7,7 +7,9 @@ migration_files = files(
+   'vmstate.c',
    'qemu-file.c',
    'yank_functions.c',
++  'pbs-state.c',
  )
 +system_ss.add(libproxmox_backup_qemu)
  
  system_ss.add(files(
    'block-dirty-bitmap.c',
-@@ -27,6 +28,7 @@ system_ss.add(files(
-   'multifd-zlib.c',
-   'multifd-zero-page.c',
-   'options.c',
-+  'pbs-state.c',
-   'postcopy-ram.c',
-   'savevm.c',
-   'savevm-async.c',
 diff --git a/migration/migration.c b/migration/migration.c
-index 8c5bd0a75c..491d9aa017 100644
+index 7a4c8beb5d..0a955a2a18 100644
 --- a/migration/migration.c
 +++ b/migration/migration.c
-@@ -266,6 +266,7 @@ void migration_object_init(void)
- 
-     /* Initialize cpu throttle timers */
-     cpu_throttle_init();
+@@ -162,6 +162,7 @@ void migration_object_init(void)
+     blk_mig_init();
+     ram_mig_init();
+     dirty_bitmap_mig_init();
 +    pbs_state_mig_init();
  }
  
- typedef struct {
+ void migration_cancel(const Error *error)
 diff --git a/migration/pbs-state.c b/migration/pbs-state.c
 new file mode 100644
-index 0000000000..a97187e4d7
+index 0000000000..887e998b9e
 --- /dev/null
 +++ b/migration/pbs-state.c
 @@ -0,0 +1,104 @@
@@ -120,7 +114,7 @@ index 0000000000..a97187e4d7
 +}
 +
 +/* serialize PBS state and send to target via f, called on source */
-+static int pbs_state_save_setup(QEMUFile *f, void *opaque, Error **errp)
++static int pbs_state_save_setup(QEMUFile *f, void *opaque)
 +{
 +    size_t buf_size;
 +    uint8_t *buf = proxmox_export_state(&buf_size);
@@ -180,10 +174,10 @@ index 0000000000..a97187e4d7
 +                         NULL);
 +}
 diff --git a/pve-backup.c b/pve-backup.c
-index fea0152de0..faa6a9b93c 100644
+index d84d807654..9c8b88d075 100644
 --- a/pve-backup.c
 +++ b/pve-backup.c
-@@ -1083,6 +1083,7 @@ ProxmoxSupportStatus *qmp_query_proxmox_support(Error **errp)
+@@ -1060,6 +1060,7 @@ ProxmoxSupportStatus *qmp_query_proxmox_support(Error **errp)
      ret->pbs_library_version = g_strdup(proxmox_backup_qemu_version());
      ret->pbs_dirty_bitmap = true;
      ret->pbs_dirty_bitmap_savevm = true;
@@ -192,10 +186,10 @@ index fea0152de0..faa6a9b93c 100644
      ret->pbs_masterkey = true;
      ret->backup_max_workers = true;
 diff --git a/qapi/block-core.json b/qapi/block-core.json
-index d45e8975a7..9795247c1f 100644
+index a818d5f90f..48eb47c6ea 100644
 --- a/qapi/block-core.json
 +++ b/qapi/block-core.json
-@@ -1004,6 +1004,11 @@
+@@ -991,6 +991,11 @@
  # @pbs-dirty-bitmap-savevm: True if 'dirty-bitmaps' migration capability can
  #                           safely be set for savevm-async.
  #
@@ -207,7 +201,7 @@ index d45e8975a7..9795247c1f 100644
  # @pbs-masterkey: True if the QMP backup call supports the 'master_keyfile'
  #                 parameter.
  #
-@@ -1017,6 +1022,7 @@
+@@ -1001,6 +1006,7 @@
    'data': { 'pbs-dirty-bitmap': 'bool',
              'query-bitmap-info': 'bool',
              'pbs-dirty-bitmap-savevm': 'bool',

debian/patches/pve/0035-migration-block-dirty-bitmap-migrate-other-bitmaps-e.patch

@@ -15,22 +15,18 @@ transferred.
 Signed-off-by: Stefan Reiter <s.reiter@proxmox.com>
 Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
 ---
- migration/block-dirty-bitmap.c | 6 +++++-
- 1 file changed, 5 insertions(+), 1 deletion(-)
+ migration/block-dirty-bitmap.c | 2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
 
 diff --git a/migration/block-dirty-bitmap.c b/migration/block-dirty-bitmap.c
-index a7d55048c2..44078ea670 100644
+index e1ae3b7316..285dd1d148 100644
 --- a/migration/block-dirty-bitmap.c
 +++ b/migration/block-dirty-bitmap.c
-@@ -539,7 +539,11 @@ static int add_bitmaps_to_list(DBMSaveState *s, BlockDriverState *bs,
-         }
+@@ -540,7 +540,7 @@ static int add_bitmaps_to_list(DBMSaveState *s, BlockDriverState *bs,
  
-         if (bdrv_dirty_bitmap_check(bitmap, BDRV_BITMAP_DEFAULT, errp)) {
+         if (bdrv_dirty_bitmap_check(bitmap, BDRV_BITMAP_DEFAULT, &local_err)) {
+             error_report_err(local_err);
 -            return -1;
-+            if (errp != NULL) {
-+                error_report_err(*errp);
-+                *errp = NULL;
-+            }
 +            continue;
          }
  

debian/patches/pve/0036-PVE-fall-back-to-open-iscsi-initiatorname.patch

@@ -21,10 +21,10 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  1 file changed, 30 insertions(+)
 
 diff --git a/block/iscsi.c b/block/iscsi.c
-index 979bf90cb7..961714a4be 100644
+index 34f97ab646..398782963d 100644
 --- a/block/iscsi.c
 +++ b/block/iscsi.c
-@@ -1392,12 +1392,42 @@ static char *get_initiator_name(QemuOpts *opts)
+@@ -1391,12 +1391,42 @@ static char *get_initiator_name(QemuOpts *opts)
      const char *name;
      char *iscsi_name;
      UuidInfo *uuid_info;

debian/patches/pve/0037-PVE-block-stream-increase-chunk-size.patch

@@ -11,7 +11,7 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  1 file changed, 1 insertion(+), 1 deletion(-)
 
 diff --git a/block/stream.c b/block/stream.c
-index 9076203193..1d1c65f061 100644
+index e522bbdec5..afed72db55 100644
 --- a/block/stream.c
 +++ b/block/stream.c
 @@ -27,7 +27,7 @@ enum {

debian/patches/pve/0038-block-io-accept-NULL-qiov-in-bdrv_pad_request.patch

@@ -0,0 +1,33 @@
+From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
+From: Stefan Reiter <s.reiter@proxmox.com>
+Date: Tue, 2 Mar 2021 16:11:54 +0100
+Subject: [PATCH] block/io: accept NULL qiov in bdrv_pad_request
+
+Some operations, e.g. block-stream, perform reads while discarding the
+results (only copy-on-read matters). In this case they will pass NULL as
+the target QEMUIOVector, which will however trip bdrv_pad_request, since
+it wants to extend its passed vector.
+
+Simply check for NULL and do nothing, there's no reason to pad the
+target if it will be discarded anyway.
+
+Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
+---
+ block/io.c | 4 ++++
+ 1 file changed, 4 insertions(+)
+
+diff --git a/block/io.c b/block/io.c
+index 055fcf7438..63f7b3ad3e 100644
+--- a/block/io.c
++++ b/block/io.c
+@@ -1710,6 +1710,10 @@ static int bdrv_pad_request(BlockDriverState *bs,
+     int sliced_niov;
+     size_t sliced_head, sliced_tail;
+ 
++    if (!qiov) {
++        return 0;
++    }
++
+     /* Should have been checked by the caller already */
+     ret = bdrv_check_request32(*offset, *bytes, *qiov, *qiov_offset);
+     if (ret < 0) {

debian/patches/pve/0039-block-add-alloc-track-driver.patch

@@ -19,33 +19,27 @@ well.
 This only worked if the target supports backing images, so up until now
 only for qcow2, with alloc-track any driver for the target can be used.
 
-Replacing the node cannot be done in the
-track_co_change_backing_file() callback, because replacing a node
-cannot happen in a coroutine and requires the block graph lock
-exclusively. Could either become a special option for the stream job,
-or maybe the upcoming blockdev-replace QMP command can be used in the
-future.
+If 'auto-remove' is set, alloc-track will automatically detach itself
+once the backing image is removed. It will be replaced by 'file'.
 
 Signed-off-by: Stefan Reiter <s.reiter@proxmox.com>
 Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
 [FE: adapt to changed function signatures
      make error return value consistent with QEMU
-     avoid premature break during read
-     adhere to block graph lock requirements]
+     avoid premature break during read]
 Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
 ---
- block/alloc-track.c | 366 ++++++++++++++++++++++++++++++++++++++++++++
+ block/alloc-track.c | 352 ++++++++++++++++++++++++++++++++++++++++++++
  block/meson.build   |   1 +
- block/stream.c      |  34 ++++
- 3 files changed, 401 insertions(+)
+ 2 files changed, 353 insertions(+)
  create mode 100644 block/alloc-track.c
 
 diff --git a/block/alloc-track.c b/block/alloc-track.c
 new file mode 100644
-index 0000000000..b4a9851144
+index 0000000000..b75d7c6460
 --- /dev/null
 +++ b/block/alloc-track.c
-@@ -0,0 +1,366 @@
+@@ -0,0 +1,352 @@
 +/*
 + * Node to allow backing images to be applied to any node. Assumes a blank
 + * image to begin with, only new writes are tracked as allocated, thus this
@@ -62,11 +56,9 @@ index 0000000000..b4a9851144
 +#include "qapi/error.h"
 +#include "block/block_int.h"
 +#include "block/dirty-bitmap.h"
-+#include "block/graph-lock.h"
 +#include "qapi/qmp/qdict.h"
 +#include "qapi/qmp/qstring.h"
 +#include "qemu/cutils.h"
-+#include "qemu/error-report.h"
 +#include "qemu/option.h"
 +#include "qemu/module.h"
 +#include "sysemu/block-backend.h"
@@ -75,12 +67,12 @@ index 0000000000..b4a9851144
 +
 +typedef enum DropState {
 +    DropNone,
++    DropRequested,
 +    DropInProgress,
 +} DropState;
 +
 +typedef struct {
 +    BdrvDirtyBitmap *bitmap;
-+    uint64_t granularity;
 +    DropState drop_state;
 +    bool auto_remove;
 +} BDRVAllocTrackState;
@@ -99,29 +91,26 @@ index 0000000000..b4a9851144
 +    },
 +};
 +
-+static void GRAPH_RDLOCK
-+track_refresh_limits(BlockDriverState *bs, Error **errp)
++static void track_refresh_limits(BlockDriverState *bs, Error **errp)
 +{
-+    BDRVAllocTrackState *s = bs->opaque;
++    BlockDriverInfo bdi;
 +
 +    if (!bs->file) {
 +        return;
 +    }
 +
-+    /*
-+     * Always use alignment from underlying write device so RMW cycle for
-+     * bdrv_pwritev reads data from our backing via track_co_preadv. Also use at
-+     * least the bitmap granularity.
-+     */
++    /* always use alignment from underlying write device so RMW cycle for
++     * bdrv_pwritev reads data from our backing via track_co_preadv (no partial
++     * cluster allocation in 'file') */
++    bdrv_get_info(bs->file->bs, &bdi);
 +    bs->bl.request_alignment = MAX(bs->file->bs->bl.request_alignment,
-+                                   s->granularity);
++                                   MAX(bdi.cluster_size, BDRV_SECTOR_SIZE));
 +}
 +
 +static int track_open(BlockDriverState *bs, QDict *options, int flags,
 +                      Error **errp)
 +{
 +    BDRVAllocTrackState *s = bs->opaque;
-+    BdrvChild *file = NULL;
 +    QemuOpts *opts;
 +    Error *local_err = NULL;
 +    int ret = 0;
@@ -137,45 +126,18 @@ index 0000000000..b4a9851144
 +    s->auto_remove = qemu_opt_get_bool(opts, TRACK_OPT_AUTO_REMOVE, false);
 +
 +    /* open the target (write) node, backing will be attached by block layer */
-+    file = bdrv_open_child(NULL, options, "file", bs, &child_of_bds,
-+                           BDRV_CHILD_DATA | BDRV_CHILD_METADATA, false,
-+                           &local_err);
-+    bdrv_graph_wrlock();
-+    bs->file = file;
-+    bdrv_graph_wrunlock();
++    bs->file = bdrv_open_child(NULL, options, "file", bs, &child_of_bds,
++                               BDRV_CHILD_DATA | BDRV_CHILD_METADATA, false,
++                               &local_err);
 +    if (local_err) {
 +        ret = -EINVAL;
 +        error_propagate(errp, local_err);
 +        goto fail;
 +    }
 +
-+    bdrv_graph_rdlock_main_loop();
-+    BlockDriverInfo bdi = {0};
-+    ret = bdrv_get_info(bs->file->bs, &bdi);
-+    if (ret < 0) {
-+        /*
-+         * Not a hard failure. Worst that can happen is partial cluster
-+         * allocation in the write target. However, the driver here returns its
-+         * allocation status based on the dirty bitmap, so any other data that
-+         * maps to such a cluster will still be copied later by a stream job (or
-+         * during writes to that cluster).
-+         */
-+        warn_report("alloc-track: unable to query cluster size for write target: %s",
-+                    strerror(ret));
-+    }
-+    ret = 0;
-+    /*
-+     * Always consider alignment from underlying write device so RMW cycle for
-+     * bdrv_pwritev reads data from our backing via track_co_preadv. Also try to
-+     * avoid partial cluster allocation in the write target by considering the
-+     * cluster size.
-+     */
-+    s->granularity = MAX(bs->file->bs->bl.request_alignment,
-+                         MAX(bdi.cluster_size, BDRV_SECTOR_SIZE));
 +    track_refresh_limits(bs, errp);
-+    s->bitmap = bdrv_create_dirty_bitmap(bs->file->bs, s->granularity, NULL,
-+                                         &local_err);
-+    bdrv_graph_rdunlock_main_loop();
++    uint64_t gran = bs->bl.request_alignment;
++    s->bitmap = bdrv_create_dirty_bitmap(bs->file->bs, gran, NULL, &local_err);
 +    if (local_err) {
 +        ret = -EIO;
 +        error_propagate(errp, local_err);
@@ -186,9 +148,7 @@ index 0000000000..b4a9851144
 +
 +fail:
 +    if (ret < 0) {
-+        bdrv_graph_wrlock();
 +        bdrv_unref_child(bs, bs->file);
-+        bdrv_graph_wrunlock();
 +        if (s->bitmap) {
 +            bdrv_release_dirty_bitmap(s->bitmap);
 +        }
@@ -205,15 +165,13 @@ index 0000000000..b4a9851144
 +    }
 +}
 +
-+static coroutine_fn int64_t GRAPH_RDLOCK
-+track_co_getlength(BlockDriverState *bs)
++static coroutine_fn int64_t track_co_getlength(BlockDriverState *bs)
 +{
 +    return bdrv_co_getlength(bs->file->bs);
 +}
 +
-+static int coroutine_fn GRAPH_RDLOCK
-+track_co_preadv(BlockDriverState *bs, int64_t offset, int64_t bytes,
-+                QEMUIOVector *qiov, BdrvRequestFlags flags)
++static int coroutine_fn track_co_preadv(BlockDriverState *bs,
++    int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags)
 +{
 +    BDRVAllocTrackState *s = bs->opaque;
 +    QEMUIOVector local_qiov;
@@ -271,39 +229,36 @@ index 0000000000..b4a9851144
 +    return ret;
 +}
 +
-+static int coroutine_fn GRAPH_RDLOCK
-+track_co_pwritev(BlockDriverState *bs, int64_t offset, int64_t bytes,
-+                 QEMUIOVector *qiov, BdrvRequestFlags flags)
++static int coroutine_fn track_co_pwritev(BlockDriverState *bs,
++    int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags)
 +{
 +    return bdrv_co_pwritev(bs->file, offset, bytes, qiov, flags);
 +}
 +
-+static int coroutine_fn GRAPH_RDLOCK
-+track_co_pwrite_zeroes(BlockDriverState *bs, int64_t offset, int64_t bytes,
-+                       BdrvRequestFlags flags)
++static int coroutine_fn track_co_pwrite_zeroes(BlockDriverState *bs,
++    int64_t offset, int64_t bytes, BdrvRequestFlags flags)
 +{
 +    return bdrv_co_pwrite_zeroes(bs->file, offset, bytes, flags);
 +}
 +
-+static int coroutine_fn GRAPH_RDLOCK
-+track_co_pdiscard(BlockDriverState *bs, int64_t offset, int64_t bytes)
++static int coroutine_fn track_co_pdiscard(BlockDriverState *bs,
++    int64_t offset, int64_t bytes)
 +{
 +    return bdrv_co_pdiscard(bs->file, offset, bytes);
 +}
 +
-+static coroutine_fn int GRAPH_RDLOCK
-+track_co_flush(BlockDriverState *bs)
++static coroutine_fn int track_co_flush(BlockDriverState *bs)
 +{
 +    return bdrv_co_flush(bs->file->bs);
 +}
 +
-+static int coroutine_fn GRAPH_RDLOCK
-+track_co_block_status(BlockDriverState *bs, bool want_zero,
-+                                            int64_t offset,
-+                                            int64_t bytes,
-+                                            int64_t *pnum,
-+                                            int64_t *map,
-+                                            BlockDriverState **file)
++static int coroutine_fn track_co_block_status(BlockDriverState *bs,
++                                              bool want_zero,
++                                              int64_t offset,
++                                              int64_t bytes,
++                                              int64_t *pnum,
++                                              int64_t *map,
++                                              BlockDriverState **file)
 +{
 +    BDRVAllocTrackState *s = bs->opaque;
 +
@@ -329,10 +284,10 @@ index 0000000000..b4a9851144
 +    return 0;
 +}
 +
-+static void GRAPH_RDLOCK
-+track_child_perm(BlockDriverState *bs, BdrvChild *c, BdrvChildRole role,
-+                 BlockReopenQueue *reopen_queue, uint64_t perm, uint64_t shared,
-+                 uint64_t *nperm, uint64_t *nshared)
++static void track_child_perm(BlockDriverState *bs, BdrvChild *c,
++                             BdrvChildRole role, BlockReopenQueue *reopen_queue,
++                             uint64_t perm, uint64_t shared,
++                             uint64_t *nperm, uint64_t *nshared)
 +{
 +    BDRVAllocTrackState *s = bs->opaque;
 +
@@ -355,28 +310,53 @@ index 0000000000..b4a9851144
 +    }
 +}
 +
-+static int coroutine_fn GRAPH_RDLOCK
-+track_co_change_backing_file(BlockDriverState *bs, const char *backing_file,
-+                             const char *backing_fmt)
++static void track_drop(void *opaque)
 +{
-+    /*
-+     * Note that the actual backing file graph change is already done in the
-+     * stream job itself with bdrv_set_backing_hd_drained(), so no need to
-+     * actually do anything here. But still needs to be implemented, to make
-+     * our caller (i.e. bdrv_co_change_backing_file() do the right thing).
-+     *
-+     * FIXME
-+     * We'd like to auto-remove ourselves from the block graph, but it cannot
-+     * be done from a coroutine. Currently done in the stream job, where it
-+     * kinda fits better, but in the long-term, a special parameter would be
-+     * nice (or done via qemu-server via upcoming blockdev-replace QMP command).
-+     */
-+    if (backing_file == NULL) {
-+        BDRVAllocTrackState *s = bs->opaque;
-+        bdrv_drained_begin(bs);
-+        s->drop_state = DropInProgress;
-+        bdrv_child_refresh_perms(bs, bs->file, &error_abort);
-+        bdrv_drained_end(bs);
++    BlockDriverState *bs = (BlockDriverState*)opaque;
++    BlockDriverState *file = bs->file->bs;
++    BDRVAllocTrackState *s = bs->opaque;
++
++    assert(file);
++
++    /* we rely on the fact that we're not used anywhere else, so let's wait
++     * until we're only used once - in the drive connected to the guest (and one
++     * ref is held by bdrv_ref in track_change_backing_file) */
++    if (bs->refcnt > 2) {
++        aio_bh_schedule_oneshot(qemu_get_aio_context(), track_drop, opaque);
++        return;
++    }
++    AioContext *aio_context = bdrv_get_aio_context(bs);
++    aio_context_acquire(aio_context);
++
++    bdrv_drained_begin(bs);
++
++    /* now that we're drained, we can safely set 'DropInProgress' */
++    s->drop_state = DropInProgress;
++    bdrv_child_refresh_perms(bs, bs->file, &error_abort);
++
++    bdrv_replace_node(bs, file, &error_abort);
++    bdrv_set_backing_hd(bs, NULL, &error_abort);
++    bdrv_drained_end(bs);
++    bdrv_unref(bs);
++    aio_context_release(aio_context);
++}
++
++static int track_change_backing_file(BlockDriverState *bs,
++                                     const char *backing_file,
++                                     const char *backing_fmt)
++{
++    BDRVAllocTrackState *s = bs->opaque;
++    if (s->auto_remove && s->drop_state == DropNone &&
++        backing_file == NULL && backing_fmt == NULL)
++    {
++        /* backing file has been disconnected, there's no longer any use for
++         * this node, so let's remove ourselves from the block graph - we need
++         * to schedule this for later however, since when this function is
++         * called, the blockjob modifying us is probably not done yet and has a
++         * blocker on 'bs' */
++        s->drop_state = DropRequested;
++        bdrv_ref(bs);
++        aio_bh_schedule_oneshot(qemu_get_aio_context(), track_drop, (void*)bs);
 +    }
 +
 +    return 0;
@@ -386,7 +366,7 @@ index 0000000000..b4a9851144
 +    .format_name                      = "alloc-track",
 +    .instance_size                    = sizeof(BDRVAllocTrackState),
 +
-+    .bdrv_open                        = track_open,
++    .bdrv_file_open                   = track_open,
 +    .bdrv_close                       = track_close,
 +    .bdrv_co_getlength                = track_co_getlength,
 +    .bdrv_child_perm                  = track_child_perm,
@@ -403,7 +383,7 @@ index 0000000000..b4a9851144
 +    .supports_backing                 = true,
 +
 +    .bdrv_co_block_status             = track_co_block_status,
-+    .bdrv_co_change_backing_file      = track_co_change_backing_file,
++    .bdrv_change_backing_file         = track_change_backing_file,
 +};
 +
 +static void bdrv_alloc_track_init(void)
@@ -413,7 +393,7 @@ index 0000000000..b4a9851144
 +
 +block_init(bdrv_alloc_track_init);
 diff --git a/block/meson.build b/block/meson.build
-index e178047ec9..7ef7250d31 100644
+index becc99ac4e..0a69836593 100644
 --- a/block/meson.build
 +++ b/block/meson.build
 @@ -2,6 +2,7 @@ block_ss.add(genh)
@@ -424,48 +404,3 @@ index e178047ec9..7ef7250d31 100644
    'amend.c',
    'backup.c',
    'backup-dump.c',
-diff --git a/block/stream.c b/block/stream.c
-index 1d1c65f061..d499c8883f 100644
---- a/block/stream.c
-+++ b/block/stream.c
-@@ -120,6 +120,40 @@ static int stream_prepare(Job *job)
-             ret = -EPERM;
-             goto out;
-         }
-+
-+        /*
-+         * This cannot be done in the co_change_backing_file callback, because
-+         * bdrv_replace_node() cannot be done in a coroutine. The latter also
-+         * requires the graph lock exclusively. Only required for the
-+         * alloc-track driver.
-+         *
-+         * The long-term plan is to either have an explicit parameter for the
-+         * stream job or use the upcoming blockdev-replace QMP command.
-+         */
-+        if (base_id == NULL && strcmp(unfiltered_bs->drv->format_name, "alloc-track") == 0) {
-+            BlockDriverState *file_bs;
-+
-+            bdrv_graph_rdlock_main_loop();
-+            file_bs = unfiltered_bs->file->bs;
-+            bdrv_graph_rdunlock_main_loop();
-+
-+            bdrv_ref(unfiltered_bs); // unrefed by bdrv_replace_node()
-+            bdrv_drained_begin(file_bs);
-+            bdrv_graph_wrlock();
-+
-+            bdrv_replace_node(unfiltered_bs, file_bs, &local_err);
-+
-+            bdrv_graph_wrunlock();
-+            bdrv_drained_end(file_bs);
-+            bdrv_unref(unfiltered_bs);
-+
-+            if (local_err) {
-+                error_prepend(&local_err, "failed to replace alloc-track node: ");
-+                error_report_err(local_err);
-+                ret = -EPERM;
-+                goto out;
-+            }
-+        }
-     }
- 
- out:

debian/patches/pve/0040-Revert-block-rbd-workaround-for-ceph-issue-53784.patch

@@ -7,16 +7,15 @@ This reverts commit fc176116cdea816ceb8dd969080b2b95f58edbc0 in
 preparation to revert 0347a8fd4c3faaedf119be04c197804be40a384b.
 
 Signed-off-by: Fabian Ebner <f.ebner@proxmox.com>
-Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
 ---
  block/rbd.c | 42 ++----------------------------------------
  1 file changed, 2 insertions(+), 40 deletions(-)
 
 diff --git a/block/rbd.c b/block/rbd.c
-index 728bce3b1e..6c9a8e0add 100644
+index a4749f3b1b..53e0396b51 100644
 --- a/block/rbd.c
 +++ b/block/rbd.c
-@@ -1515,7 +1515,6 @@ static int coroutine_fn qemu_rbd_co_block_status(BlockDriverState *bs,
+@@ -1511,7 +1511,6 @@ static int coroutine_fn qemu_rbd_co_block_status(BlockDriverState *bs,
      int status, r;
      RBDDiffIterateReq req = { .offs = offset };
      uint64_t features, flags;
@@ -24,7 +23,7 @@ index 728bce3b1e..6c9a8e0add 100644
  
      assert(offset + bytes <= s->image_size);
  
-@@ -1543,43 +1542,7 @@ static int coroutine_fn qemu_rbd_co_block_status(BlockDriverState *bs,
+@@ -1539,43 +1538,7 @@ static int coroutine_fn qemu_rbd_co_block_status(BlockDriverState *bs,
          return status;
      }
  
@@ -69,7 +68,7 @@ index 728bce3b1e..6c9a8e0add 100644
                            qemu_rbd_diff_iterate_cb, &req);
      if (r < 0 && r != QEMU_RBD_EXIT_DIFF_ITERATE2) {
          return status;
-@@ -1598,8 +1561,7 @@ static int coroutine_fn qemu_rbd_co_block_status(BlockDriverState *bs,
+@@ -1594,8 +1557,7 @@ static int coroutine_fn qemu_rbd_co_block_status(BlockDriverState *bs,
          status = BDRV_BLOCK_ZERO | BDRV_BLOCK_OFFSET_VALID;
      }
  

debian/patches/pve/0041-Revert-block-rbd-fix-handling-of-holes-in-.bdrv_co_b.patch

@@ -8,16 +8,15 @@ This reverts commit 9e302f64bb407a9bb097b626da97228c2654cfee in
 preparation to revert 0347a8fd4c3faaedf119be04c197804be40a384b.
 
 Signed-off-by: Fabian Ebner <f.ebner@proxmox.com>
-Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
 ---
  block/rbd.c | 10 +++++-----
  1 file changed, 5 insertions(+), 5 deletions(-)
 
 diff --git a/block/rbd.c b/block/rbd.c
-index 6c9a8e0add..6f5fe90f3a 100644
+index 53e0396b51..0913a0af39 100644
 --- a/block/rbd.c
 +++ b/block/rbd.c
-@@ -1474,11 +1474,11 @@ static int qemu_rbd_diff_iterate_cb(uint64_t offs, size_t len,
+@@ -1470,11 +1470,11 @@ static int qemu_rbd_diff_iterate_cb(uint64_t offs, size_t len,
      RBDDiffIterateReq *req = opaque;
  
      assert(req->offs + req->bytes <= offs);

debian/patches/pve/0042-Revert-block-rbd-implement-bdrv_co_block_status.patch

@@ -18,13 +18,12 @@ Upstream bug report:
 https://gitlab.com/qemu-project/qemu/-/issues/1026
 
 Signed-off-by: Fabian Ebner <f.ebner@proxmox.com>
-Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
 ---
  block/rbd.c | 112 ----------------------------------------------------
  1 file changed, 112 deletions(-)
 
 diff --git a/block/rbd.c b/block/rbd.c
-index 6f5fe90f3a..24e820d056 100644
+index 0913a0af39..1dab254517 100644
 --- a/block/rbd.c
 +++ b/block/rbd.c
 @@ -108,12 +108,6 @@ typedef struct RBDTask {
@@ -40,7 +39,7 @@ index 6f5fe90f3a..24e820d056 100644
  static int qemu_rbd_connect(rados_t *cluster, rados_ioctx_t *io_ctx,
                              BlockdevOptionsRbd *opts, bool cache,
                              const char *keypairs, const char *secretid,
-@@ -1460,111 +1454,6 @@ static ImageInfoSpecific *qemu_rbd_get_specific_info(BlockDriverState *bs,
+@@ -1456,111 +1450,6 @@ static ImageInfoSpecific *qemu_rbd_get_specific_info(BlockDriverState *bs,
      return spec_info;
  }
  
@@ -152,7 +151,7 @@ index 6f5fe90f3a..24e820d056 100644
  static int64_t coroutine_fn qemu_rbd_co_getlength(BlockDriverState *bs)
  {
      BDRVRBDState *s = bs->opaque;
-@@ -1801,7 +1690,6 @@ static BlockDriver bdrv_rbd = {
+@@ -1796,7 +1685,6 @@ static BlockDriver bdrv_rbd = {
  #ifdef LIBRBD_SUPPORTS_WRITE_ZEROES
      .bdrv_co_pwrite_zeroes  = qemu_rbd_co_pwrite_zeroes,
  #endif

debian/patches/pve/0042-alloc-track-error-out-when-auto-remove-is-not-set.patch

@@ -1,43 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Fiona Ebner <f.ebner@proxmox.com>
-Date: Tue, 26 Mar 2024 14:57:51 +0100
-Subject: [PATCH] alloc-track: error out when auto-remove is not set
-
-Since replacing the node now happens in the stream job, where the
-option cannot be read from (it's internal to the driver), it will
-always be treated as on.
-
-qemu-server will always set it, make sure to have other users notice
-the change (should they even exist). The option can be fully dropped
-in the future while adding a version guard in qemu-server.
-
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
----
- block/alloc-track.c | 7 +++++--
- 1 file changed, 5 insertions(+), 2 deletions(-)
-
-diff --git a/block/alloc-track.c b/block/alloc-track.c
-index b4a9851144..fc7d58a5d0 100644
---- a/block/alloc-track.c
-+++ b/block/alloc-track.c
-@@ -34,7 +34,6 @@ typedef struct {
-     BdrvDirtyBitmap *bitmap;
-     uint64_t granularity;
-     DropState drop_state;
--    bool auto_remove;
- } BDRVAllocTrackState;
- 
- static QemuOptsList runtime_opts = {
-@@ -86,7 +85,11 @@ static int track_open(BlockDriverState *bs, QDict *options, int flags,
-         goto fail;
-     }
- 
--    s->auto_remove = qemu_opt_get_bool(opts, TRACK_OPT_AUTO_REMOVE, false);
-+    if (!qemu_opt_get_bool(opts, TRACK_OPT_AUTO_REMOVE, false)) {
-+        error_setg(errp, "alloc-track: requires auto-remove option to be set to on");
-+        ret = -EINVAL;
-+        goto fail;
-+    }
- 
-     /* open the target (write) node, backing will be attached by block layer */
-     file = bdrv_open_child(NULL, options, "file", bs, &child_of_bds,

debian/patches/pve/0043-alloc-track-avoid-seemingly-superfluous-child-permis.patch

@@ -1,84 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Fiona Ebner <f.ebner@proxmox.com>
-Date: Wed, 27 Mar 2024 11:15:39 +0100
-Subject: [PATCH] alloc-track: avoid seemingly superfluous child permission
- update
-
-Doesn't seem necessary nowadays (maybe after commit "alloc-track: fix
-deadlock during drop" where the dropping is not rescheduled and delayed
-anymore or some upstream change). Should there really be some issue,
-instead of having a drop state, this could also be just based off the
-fact whether there is still a backing child.
-
-Dumping the cumulative (shared) permissions for the BDS with a debug
-print yields the same values after this patch and with QEMU 8.1,
-namely 3 and 5.
-
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
----
- block/alloc-track.c | 26 --------------------------
- 1 file changed, 26 deletions(-)
-
-diff --git a/block/alloc-track.c b/block/alloc-track.c
-index fc7d58a5d0..b56425b7f0 100644
---- a/block/alloc-track.c
-+++ b/block/alloc-track.c
-@@ -25,15 +25,9 @@
- 
- #define TRACK_OPT_AUTO_REMOVE "auto-remove"
- 
--typedef enum DropState {
--    DropNone,
--    DropInProgress,
--} DropState;
--
- typedef struct {
-     BdrvDirtyBitmap *bitmap;
-     uint64_t granularity;
--    DropState drop_state;
- } BDRVAllocTrackState;
- 
- static QemuOptsList runtime_opts = {
-@@ -137,8 +131,6 @@ static int track_open(BlockDriverState *bs, QDict *options, int flags,
-         goto fail;
-     }
- 
--    s->drop_state = DropNone;
--
- fail:
-     if (ret < 0) {
-         bdrv_graph_wrlock();
-@@ -289,18 +281,8 @@ track_child_perm(BlockDriverState *bs, BdrvChild *c, BdrvChildRole role,
-                  BlockReopenQueue *reopen_queue, uint64_t perm, uint64_t shared,
-                  uint64_t *nperm, uint64_t *nshared)
- {
--    BDRVAllocTrackState *s = bs->opaque;
--
-     *nshared = BLK_PERM_ALL;
- 
--    /* in case we're currently dropping ourselves, claim to not use any
--     * permissions at all - which is fine, since from this point on we will
--     * never issue a read or write anymore */
--    if (s->drop_state == DropInProgress) {
--        *nperm = 0;
--        return;
--    }
--
-     if (role & BDRV_CHILD_DATA) {
-         *nperm = perm & DEFAULT_PERM_PASSTHROUGH;
-     } else {
-@@ -326,14 +308,6 @@ track_co_change_backing_file(BlockDriverState *bs, const char *backing_file,
-      * kinda fits better, but in the long-term, a special parameter would be
-      * nice (or done via qemu-server via upcoming blockdev-replace QMP command).
-      */
--    if (backing_file == NULL) {
--        BDRVAllocTrackState *s = bs->opaque;
--        bdrv_drained_begin(bs);
--        s->drop_state = DropInProgress;
--        bdrv_child_refresh_perms(bs, bs->file, &error_abort);
--        bdrv_drained_end(bs);
--    }
--
-     return 0;
- }
- 

debian/patches/pve/0043-alloc-track-fix-deadlock-during-drop.patch

@@ -0,0 +1,153 @@
+From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
+From: Fiona Ebner <f.ebner@proxmox.com>
+Date: Thu, 6 Apr 2023 14:59:31 +0200
+Subject: [PATCH] alloc-track: fix deadlock during drop
+
+by replacing the block node directly after changing the backing file
+instead of rescheduling it.
+
+With changes in QEMU 8.0, calling bdrv_get_info (and bdrv_unref)
+during drop can lead to a deadlock when using iothread (only triggered
+with multiple disks, except during debugging where it also triggered
+with one disk sometimes):
+1. job_unref_locked acquires the AioContext and calls job->driver->free
+2. track_drop gets scheduled
+3. bdrv_graph_wrlock is called and polls which leads to track_drop being
+   called
+4. track_drop acquires the AioContext recursively
+5. bdrv_get_info is a wrapped coroutine (since 8.0) and thus polls for
+   bdrv_co_get_info. This releases the AioContext, but only once! The
+   documentation for the AIO_WAIT_WHILE macro states that the
+   AioContext lock needs to be acquired exactly once, but there does
+   not seem to be a way for track_drop to know if it acquired the lock
+   recursively or not (without adding further hacks).
+6. Because the AioContext is still held by the main thread once, it can't
+   be acquired before entering bdrv_co_get_info in co_schedule_bh_cb
+   which happens in the iothread
+
+When doing the operation in change_backing_file, the AioContext has
+already been acquired by the caller, so the issue with the recursive
+lock goes away.
+
+The comment explaining why delaying the replace is necessary is
+> we need to schedule this for later however, since when this function
+> is called, the blockjob modifying us is probably not done yet and
+> has a blocker on 'bs'
+
+However, there is no check for blockers in bdrv_replace_node. It would
+need to be done by us, the caller, with check_to_replace_node.
+Furthermore, the mirror job also does its call to bdrv_replace_node
+while there is an active blocker (inserted by mirror itself) and they
+use a specialized version to check for blockers instead of
+check_to_replace_node there. Alloc-track could also do something
+similar to check for other blockers, but it should be fine to rely on
+Proxmox VE that no other operation with the blockdev is going on.
+
+Mirror also drains the target before replacing the node, but the
+target can have other users. In case of alloc-track the file child
+should not be accessible by anybody else and so there can't be an
+in-flight operation for the file child when alloc-track is drained.
+
+The rescheduling based on refcounting is a hack and it doesn't seem to
+be necessary anymore. It's not clear what the original issue from the
+comment was. Testing with older builds with track_drop done directly
+without rescheduling also didn't lead to any noticable issue for me.
+
+One issue it might have been is the one fixed by b1e1af394d
+("block/stream: Drain subtree around graph change"), where
+block-stream had a use-after-free if the base node changed at an
+inconvenient time (which alloc-track's auto-drop does).
+
+It's also not possible to just not auto-replace the alloc-track. Not
+replacing it at all leads to other operations like block resize
+hanging, and there is no good way to replace it manually via QMP
+(there is x-blockdev-change, but it is experimental and doesn't
+implement the required operation yet). Also, it's just cleaner in
+general to not leave unnecessary block nodes lying around.
+
+Suggested-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
+Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
+---
+ block/alloc-track.c | 54 ++++++++++++++-------------------------------
+ 1 file changed, 16 insertions(+), 38 deletions(-)
+
+diff --git a/block/alloc-track.c b/block/alloc-track.c
+index b75d7c6460..76da140a68 100644
+--- a/block/alloc-track.c
++++ b/block/alloc-track.c
+@@ -25,7 +25,6 @@
+ 
+ typedef enum DropState {
+     DropNone,
+-    DropRequested,
+     DropInProgress,
+ } DropState;
+ 
+@@ -268,37 +267,6 @@ static void track_child_perm(BlockDriverState *bs, BdrvChild *c,
+     }
+ }
+ 
+-static void track_drop(void *opaque)
+-{
+-    BlockDriverState *bs = (BlockDriverState*)opaque;
+-    BlockDriverState *file = bs->file->bs;
+-    BDRVAllocTrackState *s = bs->opaque;
+-
+-    assert(file);
+-
+-    /* we rely on the fact that we're not used anywhere else, so let's wait
+-     * until we're only used once - in the drive connected to the guest (and one
+-     * ref is held by bdrv_ref in track_change_backing_file) */
+-    if (bs->refcnt > 2) {
+-        aio_bh_schedule_oneshot(qemu_get_aio_context(), track_drop, opaque);
+-        return;
+-    }
+-    AioContext *aio_context = bdrv_get_aio_context(bs);
+-    aio_context_acquire(aio_context);
+-
+-    bdrv_drained_begin(bs);
+-
+-    /* now that we're drained, we can safely set 'DropInProgress' */
+-    s->drop_state = DropInProgress;
+-    bdrv_child_refresh_perms(bs, bs->file, &error_abort);
+-
+-    bdrv_replace_node(bs, file, &error_abort);
+-    bdrv_set_backing_hd(bs, NULL, &error_abort);
+-    bdrv_drained_end(bs);
+-    bdrv_unref(bs);
+-    aio_context_release(aio_context);
+-}
+-
+ static int track_change_backing_file(BlockDriverState *bs,
+                                      const char *backing_file,
+                                      const char *backing_fmt)
+@@ -308,13 +276,23 @@ static int track_change_backing_file(BlockDriverState *bs,
+         backing_file == NULL && backing_fmt == NULL)
+     {
+         /* backing file has been disconnected, there's no longer any use for
+-         * this node, so let's remove ourselves from the block graph - we need
+-         * to schedule this for later however, since when this function is
+-         * called, the blockjob modifying us is probably not done yet and has a
+-         * blocker on 'bs' */
+-        s->drop_state = DropRequested;
++         * this node, so let's remove ourselves from the block graph */
++        BlockDriverState *file = bs->file->bs;
++
++        /* Just to be sure, because bdrv_replace_node unrefs it */
+         bdrv_ref(bs);
+-        aio_bh_schedule_oneshot(qemu_get_aio_context(), track_drop, (void*)bs);
++        bdrv_drained_begin(bs);
++
++        /* now that we're drained, we can safely set 'DropInProgress' */
++        s->drop_state = DropInProgress;
++
++        bdrv_child_refresh_perms(bs, bs->file, &error_abort);
++
++        bdrv_replace_node(bs, file, &error_abort);
++        bdrv_set_backing_hd(bs, NULL, &error_abort);
++
++        bdrv_drained_end(bs);
++        bdrv_unref(bs);
+     }
+ 
+     return 0;

debian/patches/pve/0044-PVE-backup-add-fleecing-option.patch

@@ -1,337 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Fiona Ebner <f.ebner@proxmox.com>
-Date: Thu, 11 Apr 2024 11:29:28 +0200
-Subject: [PATCH] PVE backup: add fleecing option
-
-When a fleecing option is given, it is expected that each device has
-a corresponding "-fleecing" block device already attached, except for
-EFI disk and TPM state, where fleecing is never used.
-
-The following graph was adapted from [0] which also contains more
-details about fleecing.
-
-[guest]
-   |
-   | root
-   v                 file
-[copy-before-write]<------[snapshot-access]
-   |           |
-   | file      | target
-   v           v
-[source] [fleecing]
-
-For fleecing, a copy-before-write filter is inserted on top of the
-source node, as well as a snapshot-access node pointing to the filter
-node which allows to read the consistent state of the image at the
-time it was inserted. New guest writes are passed through the
-copy-before-write filter which will first copy over old data to the
-fleecing image in case that old data is still needed by the
-snapshot-access node.
-
-The backup process will sequentially read from the snapshot access,
-which has a bitmap and knows whether to read from the original image
-or the fleecing image to get the "snapshot" state, i.e. data from the
-source image at the time when the copy-before-write filter was
-inserted. After reading, the copied sections are discarded from the
-fleecing image to reduce space usage.
-
-All of this can be restricted by an initial dirty bitmap to parts of
-the source image that are required for an incremental backup.
-
-For discard to work, it is necessary that the fleecing image does not
-have a larger cluster size than the backup job granularity. Since
-querying that size does not always work, e.g. for RBD with krbd, the
-cluster size will not be reported, a minimum of 4 MiB is used. A job
-with PBS target already has at least this granularity, so it's just
-relevant for other targets. I.e. edge cases where this minimum is not
-enough should be very rare in practice. If ever necessary in the
-future, can still add a passed-in value for the backup QMP command to
-override.
-
-Additionally, the cbw-timeout and on-cbw-error=break-snapshot options
-are set when installing the copy-before-write filter and
-snapshot-access. When an error or timeout occurs, the problematic (and
-each further) snapshot operation will fail and thus cancel the backup
-instead of breaking the guest write.
-
-Note that job_id cannot be inferred from the snapshot-access bs because
-it has no parent, so just pass the one from the original bs.
-
-[0]: https://www.mail-archive.com/qemu-devel@nongnu.org/msg876056.html
-
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
-Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
----
- block/monitor/block-hmp-cmds.c |   1 +
- pve-backup.c                   | 135 ++++++++++++++++++++++++++++++++-
- qapi/block-core.json           |  10 ++-
- 3 files changed, 142 insertions(+), 4 deletions(-)
-
-diff --git a/block/monitor/block-hmp-cmds.c b/block/monitor/block-hmp-cmds.c
-index 439a7a14c8..d0e7771dcc 100644
---- a/block/monitor/block-hmp-cmds.c
-+++ b/block/monitor/block-hmp-cmds.c
-@@ -1044,6 +1044,7 @@ void coroutine_fn hmp_backup(Monitor *mon, const QDict *qdict)
-         NULL, NULL,
-         devlist, qdict_haskey(qdict, "speed"), speed,
-         false, 0, // BackupPerf max-workers
-+        false, false, // fleecing
-         &error);
- 
-     hmp_handle_error(mon, error);
-diff --git a/pve-backup.c b/pve-backup.c
-index faa6a9b93c..4b0820c8a7 100644
---- a/pve-backup.c
-+++ b/pve-backup.c
-@@ -7,9 +7,11 @@
- #include "sysemu/blockdev.h"
- #include "block/block_int-global-state.h"
- #include "block/blockjob.h"
-+#include "block/copy-before-write.h"
- #include "block/dirty-bitmap.h"
- #include "block/graph-lock.h"
- #include "qapi/qapi-commands-block.h"
-+#include "qapi/qmp/qdict.h"
- #include "qapi/qmp/qerror.h"
- #include "qemu/cutils.h"
- 
-@@ -80,8 +82,15 @@ static void pvebackup_init(void)
- // initialize PVEBackupState at startup
- opts_init(pvebackup_init);
- 
-+typedef struct PVEBackupFleecingInfo {
-+    BlockDriverState *bs;
-+    BlockDriverState *cbw;
-+    BlockDriverState *snapshot_access;
-+} PVEBackupFleecingInfo;
-+
- typedef struct PVEBackupDevInfo {
-     BlockDriverState *bs;
-+    PVEBackupFleecingInfo fleecing;
-     size_t size;
-     uint64_t block_size;
-     uint8_t dev_id;
-@@ -353,6 +362,22 @@ static void pvebackup_complete_cb(void *opaque, int ret)
-     PVEBackupDevInfo *di = opaque;
-     di->completed_ret = ret;
- 
-+    /*
-+     * Handle block-graph specific cleanup (for fleecing) outside of the coroutine, because the work
-+     * won't be done as a coroutine anyways:
-+     * - For snapshot_access, allows doing bdrv_unref() directly. Doing it via bdrv_co_unref() would
-+     *   just spawn a BH calling bdrv_unref().
-+     * - For cbw, draining would need to spawn a BH.
-+     */
-+    if (di->fleecing.snapshot_access) {
-+        bdrv_unref(di->fleecing.snapshot_access);
-+        di->fleecing.snapshot_access = NULL;
-+    }
-+    if (di->fleecing.cbw) {
-+        bdrv_cbw_drop(di->fleecing.cbw);
-+        di->fleecing.cbw = NULL;
-+    }
-+
-     /*
-      * Needs to happen outside of coroutine, because it takes the graph write lock.
-      */
-@@ -519,9 +544,77 @@ static void create_backup_jobs_bh(void *opaque) {
-         }
-         bdrv_drained_begin(di->bs);
- 
-+        BackupPerf perf = (BackupPerf){ .max_workers = backup_state.perf.max_workers };
-+
-+        BlockDriverState *source_bs = di->bs;
-+        bool discard_source = false;
-+        bdrv_graph_co_rdlock();
-+        const char *job_id = bdrv_get_device_name(di->bs);
-+        bdrv_graph_co_rdunlock();
-+        if (di->fleecing.bs) {
-+            QDict *cbw_opts = qdict_new();
-+            qdict_put_str(cbw_opts, "driver", "copy-before-write");
-+            qdict_put_str(cbw_opts, "file", bdrv_get_node_name(di->bs));
-+            qdict_put_str(cbw_opts, "target", bdrv_get_node_name(di->fleecing.bs));
-+
-+            if (di->bitmap) {
-+                /*
-+                 * Only guest writes to parts relevant for the backup need to be intercepted with
-+                 * old data being copied to the fleecing image.
-+                 */
-+                qdict_put_str(cbw_opts, "bitmap.node", bdrv_get_node_name(di->bs));
-+                qdict_put_str(cbw_opts, "bitmap.name", bdrv_dirty_bitmap_name(di->bitmap));
-+            }
-+            /*
-+             * Fleecing storage is supposed to be fast and it's better to break backup than guest
-+             * writes. Certain guest drivers like VirtIO-win have 60 seconds timeout by default, so
-+             * abort a bit before that.
-+             */
-+            qdict_put_str(cbw_opts, "on-cbw-error", "break-snapshot");
-+            qdict_put_int(cbw_opts, "cbw-timeout", 45);
-+
-+            di->fleecing.cbw = bdrv_insert_node(di->bs, cbw_opts, BDRV_O_RDWR, &local_err);
-+
-+            if (!di->fleecing.cbw) {
-+                error_setg(errp, "appending cbw node for fleecing failed: %s",
-+                           local_err ? error_get_pretty(local_err) : "unknown error");
-+                break;
-+            }
-+
-+            QDict *snapshot_access_opts = qdict_new();
-+            qdict_put_str(snapshot_access_opts, "driver", "snapshot-access");
-+            qdict_put_str(snapshot_access_opts, "file", bdrv_get_node_name(di->fleecing.cbw));
-+
-+            di->fleecing.snapshot_access =
-+                bdrv_open(NULL, NULL, snapshot_access_opts, BDRV_O_RDWR | BDRV_O_UNMAP, &local_err);
-+            if (!di->fleecing.snapshot_access) {
-+                error_setg(errp, "setting up snapshot access for fleecing failed: %s",
-+                           local_err ? error_get_pretty(local_err) : "unknown error");
-+                break;
-+            }
-+            source_bs = di->fleecing.snapshot_access;
-+            discard_source = true;
-+
-+            /*
-+             * bdrv_get_info() just retuns 0 (= doesn't matter) for RBD when using krbd. But discard
-+             * on the fleecing image won't work if the backup job's granularity is less than the RBD
-+             * object size (default 4 MiB), so it does matter. Always use at least 4 MiB. With a PBS
-+             * target, the backup job granularity would already be at least this much.
-+             */
-+            perf.min_cluster_size = 4 * 1024 * 1024;
-+            /*
-+             * For discard to work, cluster size for the backup job must be at least the same as for
-+             * the fleecing image.
-+             */
-+            BlockDriverInfo bdi;
-+            if (bdrv_get_info(di->fleecing.bs, &bdi) >= 0) {
-+                perf.min_cluster_size = MAX(perf.min_cluster_size, bdi.cluster_size);
-+            }
-+        }
-+
-         BlockJob *job = backup_job_create(
--            NULL, di->bs, di->target, backup_state.speed, sync_mode, di->bitmap,
--            bitmap_mode, false, NULL, &backup_state.perf, BLOCKDEV_ON_ERROR_REPORT,
-+            job_id, source_bs, di->target, backup_state.speed, sync_mode, di->bitmap,
-+            bitmap_mode, false, discard_source, NULL, &perf, BLOCKDEV_ON_ERROR_REPORT,
-             BLOCKDEV_ON_ERROR_REPORT, JOB_DEFAULT, pvebackup_complete_cb, di, backup_state.txn,
-             &local_err);
- 
-@@ -577,6 +670,14 @@ static void create_backup_jobs_bh(void *opaque) {
-     aio_co_enter(data->ctx, data->co);
- }
- 
-+/*
-+ * EFI disk and TPM state are small and it's just not worth setting up fleecing for them.
-+ */
-+static bool device_uses_fleecing(const char *device_id)
-+{
-+    return strncmp(device_id, "drive-efidisk", 13) && strncmp(device_id, "drive-tpmstate", 14);
-+}
-+
- /*
-  * Returns a list of device infos, which needs to be freed by the caller. In
-  * case of an error, errp will be set, but the returned value might still be a
-@@ -584,6 +685,7 @@ static void create_backup_jobs_bh(void *opaque) {
-  */
- static GList coroutine_fn GRAPH_RDLOCK *get_device_info(
-     const char *devlist,
-+    bool fleecing,
-     Error **errp)
- {
-     gchar **devs = NULL;
-@@ -607,6 +709,31 @@ static GList coroutine_fn GRAPH_RDLOCK *get_device_info(
-             }
-             PVEBackupDevInfo *di = g_new0(PVEBackupDevInfo, 1);
-             di->bs = bs;
-+
-+            if (fleecing && device_uses_fleecing(*d)) {
-+                g_autofree gchar *fleecing_devid = g_strconcat(*d, "-fleecing", NULL);
-+                BlockBackend *fleecing_blk = blk_by_name(fleecing_devid);
-+                if (!fleecing_blk) {
-+                    error_set(errp, ERROR_CLASS_DEVICE_NOT_FOUND,
-+                              "Device '%s' not found", fleecing_devid);
-+                    goto err;
-+                }
-+                BlockDriverState *fleecing_bs = blk_bs(fleecing_blk);
-+                if (!bdrv_co_is_inserted(fleecing_bs)) {
-+                    error_setg(errp, "Device '%s' has no medium", fleecing_devid);
-+                    goto err;
-+                }
-+                /*
-+                 * Fleecing image needs to be the same size to act as a cbw target.
-+                 */
-+                if (bs->total_sectors != fleecing_bs->total_sectors) {
-+                    error_setg(errp, "Size mismatch for '%s' - sector count %ld != %ld",
-+                               fleecing_devid, fleecing_bs->total_sectors, bs->total_sectors);
-+                    goto err;
-+                }
-+                di->fleecing.bs = fleecing_bs;
-+            }
-+
-             di_list = g_list_append(di_list, di);
-             d++;
-         }
-@@ -656,6 +783,7 @@ UuidInfo coroutine_fn *qmp_backup(
-     const char *devlist,
-     bool has_speed, int64_t speed,
-     bool has_max_workers, int64_t max_workers,
-+    bool has_fleecing, bool fleecing,
-     Error **errp)
- {
-     assert(qemu_in_coroutine());
-@@ -684,7 +812,7 @@ UuidInfo coroutine_fn *qmp_backup(
-     format = has_format ? format : BACKUP_FORMAT_VMA;
- 
-     bdrv_graph_co_rdlock();
--    di_list = get_device_info(devlist, &local_err);
-+    di_list = get_device_info(devlist, has_fleecing && fleecing, &local_err);
-     bdrv_graph_co_rdunlock();
-     if (local_err) {
-         error_propagate(errp, local_err);
-@@ -1087,5 +1215,6 @@ ProxmoxSupportStatus *qmp_query_proxmox_support(Error **errp)
-     ret->query_bitmap_info = true;
-     ret->pbs_masterkey = true;
-     ret->backup_max_workers = true;
-+    ret->backup_fleecing = true;
-     return ret;
- }
-diff --git a/qapi/block-core.json b/qapi/block-core.json
-index 9795247c1f..c581f1f238 100644
---- a/qapi/block-core.json
-+++ b/qapi/block-core.json
-@@ -948,6 +948,10 @@
- #
- # @max-workers: see @BackupPerf for details. Default 16.
- #
-+# @fleecing: perform a backup with fleecing. For each device in @devlist, a
-+#            corresponing '-fleecing' device with the same size already needs to
-+#            be present.
-+#
- # Returns: the uuid of the backup job
- #
- ##
-@@ -968,7 +972,8 @@
-                                     '*firewall-file': 'str',
-                                     '*devlist': 'str',
-                                     '*speed': 'int',
--                                    '*max-workers': 'int' },
-+                                    '*max-workers': 'int',
-+                                    '*fleecing': 'bool' },
-   'returns': 'UuidInfo', 'coroutine': true }
- 
- ##
-@@ -1014,6 +1019,8 @@
- #
- # @pbs-library-version: Running version of libproxmox-backup-qemu0 library.
- #
-+# @backup-fleecing: Whether backup fleecing is supported or not.
-+#
- # @backup-max-workers: Whether the 'max-workers' @BackupPerf setting is
- #     supported or not.
- #
-@@ -1025,6 +1032,7 @@
-             'pbs-dirty-bitmap-migration': 'bool',
-             'pbs-masterkey': 'bool',
-             'pbs-library-version': 'str',
-+            'backup-fleecing': 'bool',
-             'backup-max-workers': 'bool' } }
- 
- ##

debian/patches/pve/0044-migration-for-snapshots-hold-the-BQL-during-setup-ca.patch

@@ -0,0 +1,190 @@
+From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
+From: Fiona Ebner <f.ebner@proxmox.com>
+Date: Fri, 5 May 2023 13:39:53 +0200
+Subject: [PATCH] migration: for snapshots, hold the BQL during setup callbacks
+
+In spirit, this is a partial revert of commit 9b09503752 ("migration:
+run setup callbacks out of big lock"), but only for the snapshot case.
+
+For snapshots, the bdrv_writev_vmstate() function is used during setup
+(in QIOChannelBlock backing the QEMUFile), but not holding the BQL
+while calling it could lead to an assertion failure. To understand
+how, first note the following:
+
+1. Generated coroutine wrappers for block layer functions spawn the
+coroutine and use AIO_WAIT_WHILE()/aio_poll() to wait for it.
+2. If the host OS switches threads at an inconvenient time, it can
+happen that a bottom half scheduled for the main thread's AioContext
+is executed as part of a vCPU thread's aio_poll().
+
+An example leading to the assertion failure is as follows:
+
+main thread:
+1. A snapshot-save QMP command gets issued.
+2. snapshot_save_job_bh() is scheduled.
+
+vCPU thread:
+3. aio_poll() for the main thread's AioContext is called (e.g. when
+the guest writes to a pflash device, as part of blk_pwrite which is a
+generated coroutine wrapper).
+4. snapshot_save_job_bh() is executed as part of aio_poll().
+3. qemu_savevm_state() is called.
+4. qemu_mutex_unlock_iothread() is called. Now
+qemu_get_current_aio_context() returns 0x0.
+5. bdrv_writev_vmstate() is executed during the usual savevm setup.
+But this function is a generated coroutine wrapper, so it uses
+AIO_WAIT_WHILE. There, the assertion
+assert(qemu_get_current_aio_context() == qemu_get_aio_context());
+will fail.
+
+To fix it, ensure that the BQL is held during setup. To avoid changing
+the behavior for migration too, introduce conditionals for the setup
+callbacks that need the BQL and only take the lock if it's not already
+held.
+
+Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
+---
+ include/migration/register.h   |  2 +-
+ migration/block-dirty-bitmap.c | 15 ++++++++++++---
+ migration/block.c              | 15 ++++++++++++---
+ migration/ram.c                | 16 +++++++++++++---
+ migration/savevm.c             |  2 --
+ 5 files changed, 38 insertions(+), 12 deletions(-)
+
+diff --git a/include/migration/register.h b/include/migration/register.h
+index 90914f32f5..c728fd9120 100644
+--- a/include/migration/register.h
++++ b/include/migration/register.h
+@@ -43,9 +43,9 @@ typedef struct SaveVMHandlers {
+      * by other locks.
+      */
+     int (*save_live_iterate)(QEMUFile *f, void *opaque);
++    int (*save_setup)(QEMUFile *f, void *opaque);
+ 
+     /* This runs outside the iothread lock!  */
+-    int (*save_setup)(QEMUFile *f, void *opaque);
+     /* Note for save_live_pending:
+      * must_precopy:
+      * - must be migrated in precopy or in stopped state
+diff --git a/migration/block-dirty-bitmap.c b/migration/block-dirty-bitmap.c
+index 285dd1d148..f7ee5a74d9 100644
+--- a/migration/block-dirty-bitmap.c
++++ b/migration/block-dirty-bitmap.c
+@@ -1219,10 +1219,17 @@ static int dirty_bitmap_save_setup(QEMUFile *f, void *opaque)
+ {
+     DBMSaveState *s = &((DBMState *)opaque)->save;
+     SaveBitmapState *dbms = NULL;
++    bool release_lock = false;
+ 
+-    qemu_mutex_lock_iothread();
++    /* For snapshots, the BQL is held during setup. */
++    if (!qemu_mutex_iothread_locked()) {
++        qemu_mutex_lock_iothread();
++        release_lock = true;
++    }
+     if (init_dirty_bitmap_migration(s) < 0) {
+-        qemu_mutex_unlock_iothread();
++        if (release_lock) {
++            qemu_mutex_unlock_iothread();
++        }
+         return -1;
+     }
+ 
+@@ -1230,7 +1237,9 @@ static int dirty_bitmap_save_setup(QEMUFile *f, void *opaque)
+         send_bitmap_start(f, s, dbms);
+     }
+     qemu_put_bitmap_flags(f, DIRTY_BITMAP_MIG_FLAG_EOS);
+-    qemu_mutex_unlock_iothread();
++    if (release_lock) {
++        qemu_mutex_unlock_iothread();
++    }
+     return 0;
+ }
+ 
+diff --git a/migration/block.c b/migration/block.c
+index 86c2256a2b..8423e0c9f9 100644
+--- a/migration/block.c
++++ b/migration/block.c
+@@ -725,21 +725,30 @@ static void block_migration_cleanup(void *opaque)
+ static int block_save_setup(QEMUFile *f, void *opaque)
+ {
+     int ret;
++    bool release_lock = false;
+ 
+     trace_migration_block_save("setup", block_mig_state.submitted,
+                                block_mig_state.transferred);
+ 
+-    qemu_mutex_lock_iothread();
++    /* For snapshots, the BQL is held during setup. */
++    if (!qemu_mutex_iothread_locked()) {
++        qemu_mutex_lock_iothread();
++        release_lock = true;
++    }
+     ret = init_blk_migration(f);
+     if (ret < 0) {
+-        qemu_mutex_unlock_iothread();
++        if (release_lock) {
++            qemu_mutex_unlock_iothread();
++        }
+         return ret;
+     }
+ 
+     /* start track dirty blocks */
+     ret = set_dirty_tracking();
+ 
+-    qemu_mutex_unlock_iothread();
++    if (release_lock) {
++        qemu_mutex_unlock_iothread();
++    }
+ 
+     if (ret) {
+         return ret;
+diff --git a/migration/ram.c b/migration/ram.c
+index 9040d66e61..01532c9fc9 100644
+--- a/migration/ram.c
++++ b/migration/ram.c
+@@ -2895,8 +2895,16 @@ static void migration_bitmap_clear_discarded_pages(RAMState *rs)
+ 
+ static void ram_init_bitmaps(RAMState *rs)
+ {
+-    /* For memory_global_dirty_log_start below.  */
+-    qemu_mutex_lock_iothread();
++    bool release_lock = false;
++
++    /*
++     * For memory_global_dirty_log_start below.
++     * For snapshots, the BQL is held during setup.
++     */
++    if (!qemu_mutex_iothread_locked()) {
++        qemu_mutex_lock_iothread();
++        release_lock = true;
++    }
+     qemu_mutex_lock_ramlist();
+ 
+     WITH_RCU_READ_LOCK_GUARD() {
+@@ -2908,7 +2916,9 @@ static void ram_init_bitmaps(RAMState *rs)
+         }
+     }
+     qemu_mutex_unlock_ramlist();
+-    qemu_mutex_unlock_iothread();
++    if (release_lock) {
++        qemu_mutex_unlock_iothread();
++    }
+ 
+     /*
+      * After an eventual first bitmap sync, fixup the initial bitmap
+diff --git a/migration/savevm.c b/migration/savevm.c
+index a2cb8855e2..ea8b30a630 100644
+--- a/migration/savevm.c
++++ b/migration/savevm.c
+@@ -1625,10 +1625,8 @@ static int qemu_savevm_state(QEMUFile *f, Error **errp)
+     reset_vfio_bytes_transferred();
+     ms->to_dst_file = f;
+ 
+-    qemu_mutex_unlock_iothread();
+     qemu_savevm_state_header(f);
+     qemu_savevm_state_setup(f);
+-    qemu_mutex_lock_iothread();
+ 
+     while (qemu_file_get_error(f) == 0) {
+         if (qemu_savevm_state_iterate(f, false) > 0) {

debian/patches/pve/0045-PVE-backup-improve-error-when-copy-before-write-fail.patch

@@ -1,117 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Fiona Ebner <f.ebner@proxmox.com>
-Date: Mon, 29 Apr 2024 14:43:58 +0200
-Subject: [PATCH] PVE backup: improve error when copy-before-write fails for
- fleecing
-
-With fleecing, failure for copy-before-write does not fail the guest
-write, but only sets the snapshot error that is associated to the
-copy-before-write filter, making further requests to the snapshot
-access fail with EACCES, which then also fails the job. But that error
-code is not the root cause of why the backup failed, so bubble up the
-original snapshot error instead.
-
-Reported-by: Friedrich Weber <f.weber@proxmox.com>
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
-Tested-by: Friedrich Weber <f.weber@proxmox.com>
----
- block/copy-before-write.c | 18 ++++++++++++------
- block/copy-before-write.h |  1 +
- pve-backup.c              |  9 +++++++++
- 3 files changed, 22 insertions(+), 6 deletions(-)
-
-diff --git a/block/copy-before-write.c b/block/copy-before-write.c
-index 81afeff1c7..fdf9cdc0cd 100644
---- a/block/copy-before-write.c
-+++ b/block/copy-before-write.c
-@@ -27,6 +27,7 @@
- #include "qapi/qmp/qjson.h"
- 
- #include "sysemu/block-backend.h"
-+#include "qemu/atomic.h"
- #include "qemu/cutils.h"
- #include "qapi/error.h"
- #include "block/block_int.h"
-@@ -75,7 +76,8 @@ typedef struct BDRVCopyBeforeWriteState {
-      * @snapshot_error is normally zero. But on first copy-before-write failure
-      * when @on_cbw_error == ON_CBW_ERROR_BREAK_SNAPSHOT, @snapshot_error takes
-      * value of this error (<0). After that all in-flight and further
--     * snapshot-API requests will fail with that error.
-+     * snapshot-API requests will fail with that error. To be accessed with
-+     * atomics.
-      */
-     int snapshot_error;
- } BDRVCopyBeforeWriteState;
-@@ -115,7 +117,7 @@ static coroutine_fn int cbw_do_copy_before_write(BlockDriverState *bs,
-         return 0;
-     }
- 
--    if (s->snapshot_error) {
-+    if (qatomic_read(&s->snapshot_error)) {
-         return 0;
-     }
- 
-@@ -139,9 +141,7 @@ static coroutine_fn int cbw_do_copy_before_write(BlockDriverState *bs,
-     WITH_QEMU_LOCK_GUARD(&s->lock) {
-         if (ret < 0) {
-             assert(s->on_cbw_error == ON_CBW_ERROR_BREAK_SNAPSHOT);
--            if (!s->snapshot_error) {
--                s->snapshot_error = ret;
--            }
-+            qatomic_cmpxchg(&s->snapshot_error, 0, ret);
-         } else {
-             bdrv_set_dirty_bitmap(s->done_bitmap, off, end - off);
-         }
-@@ -215,7 +215,7 @@ cbw_snapshot_read_lock(BlockDriverState *bs, int64_t offset, int64_t bytes,
- 
-     QEMU_LOCK_GUARD(&s->lock);
- 
--    if (s->snapshot_error) {
-+    if (qatomic_read(&s->snapshot_error)) {
-         g_free(req);
-         return NULL;
-     }
-@@ -595,6 +595,12 @@ void bdrv_cbw_drop(BlockDriverState *bs)
-     bdrv_unref(bs);
- }
- 
-+int bdrv_cbw_snapshot_error(BlockDriverState *bs)
-+{
-+    BDRVCopyBeforeWriteState *s = bs->opaque;
-+    return qatomic_read(&s->snapshot_error);
-+}
-+
- static void cbw_init(void)
- {
-     bdrv_register(&bdrv_cbw_filter);
-diff --git a/block/copy-before-write.h b/block/copy-before-write.h
-index 2a5d4ba693..969da3620f 100644
---- a/block/copy-before-write.h
-+++ b/block/copy-before-write.h
-@@ -44,5 +44,6 @@ BlockDriverState *bdrv_cbw_append(BlockDriverState *source,
-                                   BlockCopyState **bcs,
-                                   Error **errp);
- void bdrv_cbw_drop(BlockDriverState *bs);
-+int bdrv_cbw_snapshot_error(BlockDriverState *bs);
- 
- #endif /* COPY_BEFORE_WRITE_H */
-diff --git a/pve-backup.c b/pve-backup.c
-index 4b0820c8a7..81697d9bf9 100644
---- a/pve-backup.c
-+++ b/pve-backup.c
-@@ -374,6 +374,15 @@ static void pvebackup_complete_cb(void *opaque, int ret)
-         di->fleecing.snapshot_access = NULL;
-     }
-     if (di->fleecing.cbw) {
-+        /*
-+         * With fleecing, failure for cbw does not fail the guest write, but only sets the snapshot
-+         * error, making further requests to the snapshot fail with EACCES, which then also fail the
-+         * job. But that code is not the root cause and just confusing, so update it.
-+         */
-+        int snapshot_error = bdrv_cbw_snapshot_error(di->fleecing.cbw);
-+        if (di->completed_ret == -EACCES && snapshot_error) {
-+            di->completed_ret = snapshot_error;
-+        }
-         bdrv_cbw_drop(di->fleecing.cbw);
-         di->fleecing.cbw = NULL;
-     }

debian/patches/pve/0045-savevm-async-don-t-hold-BQL-during-setup.patch

@@ -0,0 +1,29 @@
+From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
+From: Fiona Ebner <f.ebner@proxmox.com>
+Date: Fri, 5 May 2023 15:30:16 +0200
+Subject: [PATCH] savevm-async: don't hold BQL during setup
+
+See commit "migration: for snapshots, hold the BQL during setup
+callbacks" for why. This is separate, because a version of that one
+will hopefully land upstream.
+
+Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
+---
+ migration/savevm-async.c | 2 --
+ 1 file changed, 2 deletions(-)
+
+diff --git a/migration/savevm-async.c b/migration/savevm-async.c
+index 80624fada8..b1d85a4b41 100644
+--- a/migration/savevm-async.c
++++ b/migration/savevm-async.c
+@@ -401,10 +401,8 @@ void qmp_savevm_start(const char *statefile, Error **errp)
+     snap_state.state = SAVE_STATE_ACTIVE;
+     snap_state.finalize_bh = qemu_bh_new(process_savevm_finalize, &snap_state);
+     snap_state.co = qemu_coroutine_create(&process_savevm_co, NULL);
+-    qemu_mutex_unlock_iothread();
+     qemu_savevm_state_header(snap_state.file);
+     qemu_savevm_state_setup(snap_state.file);
+-    qemu_mutex_lock_iothread();
+ 
+     /* Async processing from here on out happens in iohandler context, so let
+      * the target bdrv have its home there.

debian/patches/pve/0046-PVE-backup-fixup-error-handling-for-fleecing.patch

@@ -1,103 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Fiona Ebner <f.ebner@proxmox.com>
-Date: Thu, 7 Nov 2024 17:51:14 +0100
-Subject: [PATCH] PVE backup: fixup error handling for fleecing
-
-The drained section needs to be terminated before breaking out of the
-loop in the error scenarios. Otherwise, guest IO on the drive would
-become stuck.
-
-If the job is created successfully, then the job completion callback
-will clean up the snapshot access block nodes. In case failure
-happened before the job is created, there was no cleanup for the
-snapshot access block nodes yet. Add it.
-
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
----
- pve-backup.c | 38 +++++++++++++++++++++++++-------------
- 1 file changed, 25 insertions(+), 13 deletions(-)
-
-diff --git a/pve-backup.c b/pve-backup.c
-index 81697d9bf9..320c660589 100644
---- a/pve-backup.c
-+++ b/pve-backup.c
-@@ -357,22 +357,23 @@ static void coroutine_fn pvebackup_co_complete_stream(void *opaque)
-     qemu_co_mutex_unlock(&backup_state.backup_mutex);
- }
- 
--static void pvebackup_complete_cb(void *opaque, int ret)
-+static void cleanup_snapshot_access(PVEBackupDevInfo *di)
- {
--    PVEBackupDevInfo *di = opaque;
--    di->completed_ret = ret;
--
--    /*
--     * Handle block-graph specific cleanup (for fleecing) outside of the coroutine, because the work
--     * won't be done as a coroutine anyways:
--     * - For snapshot_access, allows doing bdrv_unref() directly. Doing it via bdrv_co_unref() would
--     *   just spawn a BH calling bdrv_unref().
--     * - For cbw, draining would need to spawn a BH.
--     */
-     if (di->fleecing.snapshot_access) {
-         bdrv_unref(di->fleecing.snapshot_access);
-         di->fleecing.snapshot_access = NULL;
-     }
-+    if (di->fleecing.cbw) {
-+        bdrv_cbw_drop(di->fleecing.cbw);
-+        di->fleecing.cbw = NULL;
-+    }
-+}
-+
-+static void pvebackup_complete_cb(void *opaque, int ret)
-+{
-+    PVEBackupDevInfo *di = opaque;
-+    di->completed_ret = ret;
-+
-     if (di->fleecing.cbw) {
-         /*
-          * With fleecing, failure for cbw does not fail the guest write, but only sets the snapshot
-@@ -383,10 +384,17 @@ static void pvebackup_complete_cb(void *opaque, int ret)
-         if (di->completed_ret == -EACCES && snapshot_error) {
-             di->completed_ret = snapshot_error;
-         }
--        bdrv_cbw_drop(di->fleecing.cbw);
--        di->fleecing.cbw = NULL;
-     }
- 
-+    /*
-+     * Handle block-graph specific cleanup (for fleecing) outside of the coroutine, because the work
-+     * won't be done as a coroutine anyways:
-+     * - For snapshot_access, allows doing bdrv_unref() directly. Doing it via bdrv_co_unref() would
-+     *   just spawn a BH calling bdrv_unref().
-+     * - For cbw, draining would need to spawn a BH.
-+     */
-+    cleanup_snapshot_access(di);
-+
-     /*
-      * Needs to happen outside of coroutine, because it takes the graph write lock.
-      */
-@@ -587,6 +595,7 @@ static void create_backup_jobs_bh(void *opaque) {
-             if (!di->fleecing.cbw) {
-                 error_setg(errp, "appending cbw node for fleecing failed: %s",
-                            local_err ? error_get_pretty(local_err) : "unknown error");
-+                bdrv_drained_end(di->bs);
-                 break;
-             }
- 
-@@ -599,6 +608,8 @@ static void create_backup_jobs_bh(void *opaque) {
-             if (!di->fleecing.snapshot_access) {
-                 error_setg(errp, "setting up snapshot access for fleecing failed: %s",
-                            local_err ? error_get_pretty(local_err) : "unknown error");
-+                cleanup_snapshot_access(di);
-+                bdrv_drained_end(di->bs);
-                 break;
-             }
-             source_bs = di->fleecing.snapshot_access;
-@@ -637,6 +648,7 @@ static void create_backup_jobs_bh(void *opaque) {
-         }
- 
-         if (!job || local_err) {
-+            cleanup_snapshot_access(di);
-             error_setg(errp, "backup_job_create failed: %s",
-                        local_err ? error_get_pretty(local_err) : "null");
-             break;

debian/patches/pve/0047-PVE-backup-factor-out-setting-up-snapshot-access-for.patch

@@ -1,135 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Fiona Ebner <f.ebner@proxmox.com>
-Date: Thu, 7 Nov 2024 17:51:15 +0100
-Subject: [PATCH] PVE backup: factor out setting up snapshot access for
- fleecing
-
-Avoids some line bloat in the create_backup_jobs_bh() function and is
-in preparation for setting up the snapshot access independently of
-fleecing, in particular that will be useful for providing access to
-the snapshot via NBD.
-
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
----
- pve-backup.c | 95 ++++++++++++++++++++++++++++++++--------------------
- 1 file changed, 58 insertions(+), 37 deletions(-)
-
-diff --git a/pve-backup.c b/pve-backup.c
-index 320c660589..d8d0c04b0f 100644
---- a/pve-backup.c
-+++ b/pve-backup.c
-@@ -525,6 +525,62 @@ static int coroutine_fn pvebackup_co_add_config(
-     goto out;
- }
- 
-+/*
-+ * Setup a snapshot-access block node for a device with associated fleecing image.
-+ */
-+static int setup_snapshot_access(PVEBackupDevInfo *di, Error **errp)
-+{
-+    Error *local_err = NULL;
-+
-+    if (!di->fleecing.bs) {
-+        error_setg(errp, "no associated fleecing image");
-+        return -1;
-+    }
-+
-+    QDict *cbw_opts = qdict_new();
-+    qdict_put_str(cbw_opts, "driver", "copy-before-write");
-+    qdict_put_str(cbw_opts, "file", bdrv_get_node_name(di->bs));
-+    qdict_put_str(cbw_opts, "target", bdrv_get_node_name(di->fleecing.bs));
-+
-+    if (di->bitmap) {
-+        /*
-+         * Only guest writes to parts relevant for the backup need to be intercepted with
-+         * old data being copied to the fleecing image.
-+         */
-+        qdict_put_str(cbw_opts, "bitmap.node", bdrv_get_node_name(di->bs));
-+        qdict_put_str(cbw_opts, "bitmap.name", bdrv_dirty_bitmap_name(di->bitmap));
-+    }
-+    /*
-+     * Fleecing storage is supposed to be fast and it's better to break backup than guest
-+     * writes. Certain guest drivers like VirtIO-win have 60 seconds timeout by default, so
-+     * abort a bit before that.
-+     */
-+    qdict_put_str(cbw_opts, "on-cbw-error", "break-snapshot");
-+    qdict_put_int(cbw_opts, "cbw-timeout", 45);
-+
-+    di->fleecing.cbw = bdrv_insert_node(di->bs, cbw_opts, BDRV_O_RDWR, &local_err);
-+
-+    if (!di->fleecing.cbw) {
-+        error_setg(errp, "appending cbw node for fleecing failed: %s",
-+                   local_err ? error_get_pretty(local_err) : "unknown error");
-+        return -1;
-+    }
-+
-+    QDict *snapshot_access_opts = qdict_new();
-+    qdict_put_str(snapshot_access_opts, "driver", "snapshot-access");
-+    qdict_put_str(snapshot_access_opts, "file", bdrv_get_node_name(di->fleecing.cbw));
-+
-+    di->fleecing.snapshot_access =
-+        bdrv_open(NULL, NULL, snapshot_access_opts, BDRV_O_RDWR | BDRV_O_UNMAP, &local_err);
-+    if (!di->fleecing.snapshot_access) {
-+        error_setg(errp, "setting up snapshot access for fleecing failed: %s",
-+                   local_err ? error_get_pretty(local_err) : "unknown error");
-+        return -1;
-+    }
-+
-+    return 0;
-+}
-+
- /*
-  * backup_job_create can *not* be run from a coroutine, so this can't either.
-  * The caller is responsible that backup_mutex is held nonetheless.
-@@ -569,49 +625,14 @@ static void create_backup_jobs_bh(void *opaque) {
-         const char *job_id = bdrv_get_device_name(di->bs);
-         bdrv_graph_co_rdunlock();
-         if (di->fleecing.bs) {
--            QDict *cbw_opts = qdict_new();
--            qdict_put_str(cbw_opts, "driver", "copy-before-write");
--            qdict_put_str(cbw_opts, "file", bdrv_get_node_name(di->bs));
--            qdict_put_str(cbw_opts, "target", bdrv_get_node_name(di->fleecing.bs));
--
--            if (di->bitmap) {
--                /*
--                 * Only guest writes to parts relevant for the backup need to be intercepted with
--                 * old data being copied to the fleecing image.
--                 */
--                qdict_put_str(cbw_opts, "bitmap.node", bdrv_get_node_name(di->bs));
--                qdict_put_str(cbw_opts, "bitmap.name", bdrv_dirty_bitmap_name(di->bitmap));
--            }
--            /*
--             * Fleecing storage is supposed to be fast and it's better to break backup than guest
--             * writes. Certain guest drivers like VirtIO-win have 60 seconds timeout by default, so
--             * abort a bit before that.
--             */
--            qdict_put_str(cbw_opts, "on-cbw-error", "break-snapshot");
--            qdict_put_int(cbw_opts, "cbw-timeout", 45);
--
--            di->fleecing.cbw = bdrv_insert_node(di->bs, cbw_opts, BDRV_O_RDWR, &local_err);
--
--            if (!di->fleecing.cbw) {
--                error_setg(errp, "appending cbw node for fleecing failed: %s",
--                           local_err ? error_get_pretty(local_err) : "unknown error");
--                bdrv_drained_end(di->bs);
--                break;
--            }
--
--            QDict *snapshot_access_opts = qdict_new();
--            qdict_put_str(snapshot_access_opts, "driver", "snapshot-access");
--            qdict_put_str(snapshot_access_opts, "file", bdrv_get_node_name(di->fleecing.cbw));
--
--            di->fleecing.snapshot_access =
--                bdrv_open(NULL, NULL, snapshot_access_opts, BDRV_O_RDWR | BDRV_O_UNMAP, &local_err);
--            if (!di->fleecing.snapshot_access) {
-+            if (setup_snapshot_access(di, &local_err) < 0) {
-                 error_setg(errp, "setting up snapshot access for fleecing failed: %s",
-                            local_err ? error_get_pretty(local_err) : "unknown error");
-                 cleanup_snapshot_access(di);
-                 bdrv_drained_end(di->bs);
-                 break;
-             }
-+
-             source_bs = di->fleecing.snapshot_access;
-             discard_source = true;
- 

debian/patches/pve/0048-PVE-backup-save-device-name-in-device-info-structure.patch

@@ -1,135 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Fiona Ebner <f.ebner@proxmox.com>
-Date: Thu, 7 Nov 2024 17:51:16 +0100
-Subject: [PATCH] PVE backup: save device name in device info structure
-
-The device name needs to be queried while holding the graph read lock
-and since it doesn't change during the whole operation, just get it
-once during setup and avoid the need to query it again in different
-places.
-
-Also in preparation to use it more often in error messages and for the
-upcoming external backup access API.
-
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
----
- pve-backup.c | 29 +++++++++++++++--------------
- 1 file changed, 15 insertions(+), 14 deletions(-)
-
-diff --git a/pve-backup.c b/pve-backup.c
-index d8d0c04b0f..e2110ce0db 100644
---- a/pve-backup.c
-+++ b/pve-backup.c
-@@ -94,6 +94,7 @@ typedef struct PVEBackupDevInfo {
-     size_t size;
-     uint64_t block_size;
-     uint8_t dev_id;
-+    char* device_name;
-     int completed_ret; // INT_MAX if not completed
-     BdrvDirtyBitmap *bitmap;
-     BlockDriverState *target;
-@@ -327,6 +328,8 @@ static void coroutine_fn pvebackup_co_complete_stream(void *opaque)
-     }
- 
-     di->bs = NULL;
-+    g_free(di->device_name);
-+    di->device_name = NULL;
- 
-     assert(di->target == NULL);
- 
-@@ -621,9 +624,6 @@ static void create_backup_jobs_bh(void *opaque) {
- 
-         BlockDriverState *source_bs = di->bs;
-         bool discard_source = false;
--        bdrv_graph_co_rdlock();
--        const char *job_id = bdrv_get_device_name(di->bs);
--        bdrv_graph_co_rdunlock();
-         if (di->fleecing.bs) {
-             if (setup_snapshot_access(di, &local_err) < 0) {
-                 error_setg(errp, "setting up snapshot access for fleecing failed: %s",
-@@ -654,7 +654,7 @@ static void create_backup_jobs_bh(void *opaque) {
-         }
- 
-         BlockJob *job = backup_job_create(
--            job_id, source_bs, di->target, backup_state.speed, sync_mode, di->bitmap,
-+            di->device_name, source_bs, di->target, backup_state.speed, sync_mode, di->bitmap,
-             bitmap_mode, false, discard_source, NULL, &perf, BLOCKDEV_ON_ERROR_REPORT,
-             BLOCKDEV_ON_ERROR_REPORT, JOB_DEFAULT, pvebackup_complete_cb, di, backup_state.txn,
-             &local_err);
-@@ -751,6 +751,7 @@ static GList coroutine_fn GRAPH_RDLOCK *get_device_info(
-             }
-             PVEBackupDevInfo *di = g_new0(PVEBackupDevInfo, 1);
-             di->bs = bs;
-+            di->device_name = g_strdup(bdrv_get_device_name(bs));
- 
-             if (fleecing && device_uses_fleecing(*d)) {
-                 g_autofree gchar *fleecing_devid = g_strconcat(*d, "-fleecing", NULL);
-@@ -789,6 +790,7 @@ static GList coroutine_fn GRAPH_RDLOCK *get_device_info(
- 
-             PVEBackupDevInfo *di = g_new0(PVEBackupDevInfo, 1);
-             di->bs = bs;
-+            di->device_name = g_strdup(bdrv_get_device_name(bs));
-             di_list = g_list_append(di_list, di);
-         }
-     }
-@@ -956,9 +958,6 @@ UuidInfo coroutine_fn *qmp_backup(
- 
-             di->block_size = dump_cb_block_size;
- 
--            bdrv_graph_co_rdlock();
--            const char *devname = bdrv_get_device_name(di->bs);
--            bdrv_graph_co_rdunlock();
-             PBSBitmapAction action = PBS_BITMAP_ACTION_NOT_USED;
-             size_t dirty = di->size;
- 
-@@ -973,7 +972,8 @@ UuidInfo coroutine_fn *qmp_backup(
-                     }
-                     action = PBS_BITMAP_ACTION_NEW;
-                 } else {
--                    expect_only_dirty = proxmox_backup_check_incremental(pbs, devname, di->size) != 0;
-+                    expect_only_dirty =
-+                        proxmox_backup_check_incremental(pbs, di->device_name, di->size) != 0;
-                 }
- 
-                 if (expect_only_dirty) {
-@@ -997,7 +997,8 @@ UuidInfo coroutine_fn *qmp_backup(
-                 }
-             }
- 
--            int dev_id = proxmox_backup_co_register_image(pbs, devname, di->size, expect_only_dirty, errp);
-+            int dev_id = proxmox_backup_co_register_image(pbs, di->device_name, di->size,
-+                                                          expect_only_dirty, errp);
-             if (dev_id < 0) {
-                 goto err_mutex;
-             }
-@@ -1009,7 +1010,7 @@ UuidInfo coroutine_fn *qmp_backup(
-             di->dev_id = dev_id;
- 
-             PBSBitmapInfo *info = g_malloc(sizeof(*info));
--            info->drive = g_strdup(devname);
-+            info->drive = g_strdup(di->device_name);
-             info->action = action;
-             info->size = di->size;
-             info->dirty = dirty;
-@@ -1032,10 +1033,7 @@ UuidInfo coroutine_fn *qmp_backup(
-                 goto err_mutex;
-             }
- 
--            bdrv_graph_co_rdlock();
--            const char *devname = bdrv_get_device_name(di->bs);
--            bdrv_graph_co_rdunlock();
--            di->dev_id = vma_writer_register_stream(vmaw, devname, di->size);
-+            di->dev_id = vma_writer_register_stream(vmaw, di->device_name, di->size);
-             if (di->dev_id <= 0) {
-                 error_set(errp, ERROR_CLASS_GENERIC_ERROR,
-                           "register_stream failed");
-@@ -1146,6 +1144,9 @@ err:
-             bdrv_co_unref(di->target);
-         }
- 
-+        g_free(di->device_name);
-+        di->device_name = NULL;
-+
-         g_free(di);
-     }
-     g_list_free(di_list);

debian/patches/pve/0049-PVE-backup-include-device-name-in-error-when-setting.patch

@@ -1,25 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Fiona Ebner <f.ebner@proxmox.com>
-Date: Thu, 7 Nov 2024 17:51:17 +0100
-Subject: [PATCH] PVE backup: include device name in error when setting up
- snapshot access fails
-
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
----
- pve-backup.c | 3 ++-
- 1 file changed, 2 insertions(+), 1 deletion(-)
-
-diff --git a/pve-backup.c b/pve-backup.c
-index e2110ce0db..32352fb5ec 100644
---- a/pve-backup.c
-+++ b/pve-backup.c
-@@ -626,7 +626,8 @@ static void create_backup_jobs_bh(void *opaque) {
-         bool discard_source = false;
-         if (di->fleecing.bs) {
-             if (setup_snapshot_access(di, &local_err) < 0) {
--                error_setg(errp, "setting up snapshot access for fleecing failed: %s",
-+                error_setg(errp, "%s - setting up snapshot access for fleecing failed: %s",
-+                           di->device_name,
-                            local_err ? error_get_pretty(local_err) : "unknown error");
-                 cleanup_snapshot_access(di);
-                 bdrv_drained_end(di->bs);

debian/patches/pve/0050-adapt-machine-version-deprecation-for-Proxmox-VE.patch

@@ -1,137 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Fiona Ebner <f.ebner@proxmox.com>
-Date: Fri, 3 Jan 2025 14:03:12 +0100
-Subject: [PATCH] adapt machine version deprecation for Proxmox VE
-
-In commit a35f8577a0 ("include/hw: add macros for deprecation &
-removal of versioned machines"), a new machine version deprecation and
-removal policy was introduced. After only 3 years a machine version
-will be deprecated while being removed after 6 years.
-
-The deprecation is a bit early considering major PVE releases are
-approximately every 2 years. This means that a deprecation warning can
-already happen for a machine version that was introduced during the
-previous major release. This would scare users for no good reason, so
-avoid deprecating machine versions in PVE too early and define a
-baseline of machine versions that will be supported throughout a
-single major PVE release.
-
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
----
- include/hw/boards.h | 78 +++++++++++++++++++++++++++++----------------
- 1 file changed, 51 insertions(+), 27 deletions(-)
-
-diff --git a/include/hw/boards.h b/include/hw/boards.h
-index d1741ea121..3f9befda14 100644
---- a/include/hw/boards.h
-+++ b/include/hw/boards.h
-@@ -631,42 +631,66 @@ struct MachineState {
- 
- 
- /*
-- * How many years/major releases for each phase
-- * of the life cycle. Assumes use of versioning
-- * scheme where major is bumped each year
-+ * Baseline of machine versions that are still considered supported throughout
-+ * current major Proxmox VE release. Machine versions older than this are
-+ * considered to be deprecated in Proxmox VE.
-+ *
-+ * Machine versions older than 6 years are removed just like in upstream QEMU.
-+ * (policy takes effect with QEMU 10.1). Assumes yearly major QEMU release.
-+ *
-+ * QEMU release cylce N.0 in ~April, N.1 in ~August, N.2 in ~December
-+ * Debian/PVE release cylce ~every two years in summer
-+ *
-+ * PVE - last QEMU - machine versions dropped - baseline
-+ *   8         9.2              2.3 and older        2.4
-+ *   9        11.2              5.2 and older        6.0
-+ *  10        13.2              7.2 and older        8.0
-+ */
-+#define MACHINE_VER_BASELINE_PVE_MAJOR 2
-+#define MACHINE_VER_BASELINE_PVE_MINOR 4
-+#define MACHINE_VER_DELETION_MAJOR (QEMU_VERSION_MAJOR - 6)
-+#define MACHINE_VER_DELETION_MINOR QEMU_VERSION_MINOR
-+
-+/*
-+ * Proxmox VE needs to support the baseline throughout a major PVE release. So
-+ * a QEMU release where the baseline is already deleted cannot be used.
-+ * Removal policy after 6 years takes effect with QEMU 10.1.
-  */
--#define MACHINE_VER_DELETION_MAJOR 6
--#define MACHINE_VER_DEPRECATION_MAJOR 3
-+#if ((QEMU_VERSION_MAJOR > 10) || ((QEMU_VERSION_MAJOR == 10) && (QEMU_VERSION_MINOR >= 1)))
-+#if ((MACHINE_VER_BASELINE_PVE_MAJOR < MACHINE_VER_DELETION_MAJOR) || \
-+     ((MACHINE_VER_BASELINE_PVE_MAJOR == MACHINE_VER_DELETION_MAJOR) && \
-+      (MACHINE_VER_BASELINE_PVE_MINOR < MACHINE_VER_DELETION_MINOR)))
-+#error "Baseline machine version needed by Proxmox VE not supported anymore by this QEMU release"
-+#endif
-+#endif
- 
- /*
-  * Expands to a static string containing a deprecation
-  * message for a versioned machine type
-  */
- #define MACHINE_VER_DEPRECATION_MSG \
--    "machines more than " stringify(MACHINE_VER_DEPRECATION_MAJOR) \
--    " years old are subject to deletion after " \
--    stringify(MACHINE_VER_DELETION_MAJOR) " years"
--
--#define _MACHINE_VER_IS_EXPIRED_IMPL(cutoff, major, minor) \
--    (((QEMU_VERSION_MAJOR - major) > cutoff) || \
--     (((QEMU_VERSION_MAJOR - major) == cutoff) && \
--      (QEMU_VERSION_MINOR - minor) >= 0))
--
--#define _MACHINE_VER_IS_EXPIRED2(cutoff, major, minor) \
--    _MACHINE_VER_IS_EXPIRED_IMPL(cutoff, major, minor)
--#define _MACHINE_VER_IS_EXPIRED3(cutoff, major, minor, micro) \
--    _MACHINE_VER_IS_EXPIRED_IMPL(cutoff, major, minor)
--#define _MACHINE_VER_IS_EXPIRED4(cutoff, major, minor, _unused, tag) \
--    _MACHINE_VER_IS_EXPIRED_IMPL(cutoff, major, minor)
--#define _MACHINE_VER_IS_EXPIRED5(cutoff, major, minor, micro, _unused, tag)   \
--    _MACHINE_VER_IS_EXPIRED_IMPL(cutoff, major, minor)
--
--#define _MACHINE_IS_EXPIRED(cutoff, ...) \
-+    "old machine version is subject to deletion during current major Proxmox VE release"
-+
-+#define _MACHINE_VER_IS_EXPIRED_IMPL(baseline_major, baseline_minor, major, minor) \
-+    ((major < baseline_major) || \
-+     ((major == baseline_major) && \
-+      (minor < baseline_minor)))
-+
-+#define _MACHINE_VER_IS_EXPIRED2(baseline_major, baseline_minor, major, minor) \
-+    _MACHINE_VER_IS_EXPIRED_IMPL(baseline_major, baseline_minor, major, minor)
-+#define _MACHINE_VER_IS_EXPIRED3(baseline_major, baseline_minor, major, minor, micro) \
-+    _MACHINE_VER_IS_EXPIRED_IMPL(baseline_major, baseline_minor, major, minor)
-+#define _MACHINE_VER_IS_EXPIRED4(baseline_major, baseline_minor, major, minor, _unused, tag) \
-+    _MACHINE_VER_IS_EXPIRED_IMPL(baseline_major, baseline_minor, major, minor)
-+#define _MACHINE_VER_IS_EXPIRED5(baseline_major, baseline_minor, major, minor, micro, _unused, tag)   \
-+    _MACHINE_VER_IS_EXPIRED_IMPL(baseline_major, baseline_minor, major, minor)
-+
-+#define _MACHINE_IS_EXPIRED(baseline_major, baseline_minor, ...) \
-     _MACHINE_VER_PICK(__VA_ARGS__, \
-                       _MACHINE_VER_IS_EXPIRED5, \
-                       _MACHINE_VER_IS_EXPIRED4, \
-                       _MACHINE_VER_IS_EXPIRED3, \
--                      _MACHINE_VER_IS_EXPIRED2) (cutoff, __VA_ARGS__)
-+                      _MACHINE_VER_IS_EXPIRED2) (baseline_major, baseline_minor, __VA_ARGS__)
- 
- /*
-  * Evaluates true when a machine type with (major, minor)
-@@ -675,7 +699,7 @@ struct MachineState {
-  * lifecycle rules
-  */
- #define MACHINE_VER_IS_DEPRECATED(...) \
--    _MACHINE_IS_EXPIRED(MACHINE_VER_DEPRECATION_MAJOR, __VA_ARGS__)
-+    _MACHINE_IS_EXPIRED(MACHINE_VER_BASELINE_PVE_MAJOR, MACHINE_VER_BASELINE_PVE_MINOR, __VA_ARGS__)
- 
- /*
-  * Evaluates true when a machine type with (major, minor)
-@@ -684,7 +708,7 @@ struct MachineState {
-  * lifecycle rules
-  */
- #define MACHINE_VER_SHOULD_DELETE(...) \
--    _MACHINE_IS_EXPIRED(MACHINE_VER_DELETION_MAJOR, __VA_ARGS__)
-+    _MACHINE_IS_EXPIRED(MACHINE_VER_DELETION_MAJOR, MACHINE_VER_DELETION_MINOR, __VA_ARGS__)
- 
- /*
-  * Sets the deprecation reason for a versioned machine based

debian/patches/pve/0051-Revert-hpet-avoid-timer-storms-on-periodic-timers.patch

@@ -1,50 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Fiona Ebner <f.ebner@proxmox.com>
-Date: Wed, 19 Mar 2025 17:31:05 +0100
-Subject: [PATCH] Revert "hpet: avoid timer storms on periodic timers"
-
-This reverts commit 7c912ffb59e8137091894d767433e65c3df8b0bf.
-
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
----
- hw/timer/hpet.c | 13 ++-----------
- 1 file changed, 2 insertions(+), 11 deletions(-)
-
-diff --git a/hw/timer/hpet.c b/hw/timer/hpet.c
-index 5399f1b2a3..8ccc421cbb 100644
---- a/hw/timer/hpet.c
-+++ b/hw/timer/hpet.c
-@@ -59,7 +59,6 @@ typedef struct HPETTimer {  /* timers */
-     uint8_t wrap_flag;      /* timer pop will indicate wrap for one-shot 32-bit
-                              * mode. Next pop will be actual timer expiration.
-                              */
--    uint64_t last;          /* last value armed, to avoid timer storms */
- } HPETTimer;
- 
- struct HPETState {
-@@ -267,7 +266,6 @@ static int hpet_post_load(void *opaque, int version_id)
-     for (i = 0; i < s->num_timers; i++) {
-         HPETTimer *t = &s->timer[i];
-         t->cmp64 = hpet_calculate_cmp64(t, s->hpet_counter, t->cmp);
--        t->last = qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL) - NANOSECONDS_PER_SECOND;
-     }
-     /* Recalculate the offset between the main counter and guest time */
-     if (!s->hpet_offset_saved) {
-@@ -366,15 +364,8 @@ static const VMStateDescription vmstate_hpet = {
- 
- static void hpet_arm(HPETTimer *t, uint64_t tick)
- {
--    uint64_t ns = hpet_get_ns(t->state, tick);
--
--    /* Clamp period to reasonable min value (1 us) */
--    if (timer_is_periodic(t) && ns - t->last < 1000) {
--        ns = t->last + 1000;
--    }
--
--    t->last = ns;
--    timer_mod(t->qemu_timer, ns);
-+    /* FIXME: Clamp period to reasonable min value? */
-+    timer_mod(t->qemu_timer, hpet_get_ns(t->state, tick));
- }
- 
- /*

debian/patches/pve/0052-Revert-hpet-store-full-64-bit-target-value-of-the-co.patch

@@ -1,203 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Fiona Ebner <f.ebner@proxmox.com>
-Date: Wed, 19 Mar 2025 17:31:08 +0100
-Subject: [PATCH] Revert "hpet: store full 64-bit target value of the counter"
-
-This reverts commit 242d665396407f83a6acbffc804882eeb21cfdad.
-
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
----
- hw/timer/hpet.c | 111 +++++++++++++++++++++++++++---------------------
- 1 file changed, 62 insertions(+), 49 deletions(-)
-
-diff --git a/hw/timer/hpet.c b/hw/timer/hpet.c
-index 8ccc421cbb..415a9433f1 100644
---- a/hw/timer/hpet.c
-+++ b/hw/timer/hpet.c
-@@ -54,7 +54,6 @@ typedef struct HPETTimer {  /* timers */
-     uint64_t cmp;           /* comparator */
-     uint64_t fsb;           /* FSB route */
-     /* Hidden register state */
--    uint64_t cmp64;         /* comparator (extended to counter width) */
-     uint64_t period;        /* Last value written to comparator */
-     uint8_t wrap_flag;      /* timer pop will indicate wrap for one-shot 32-bit
-                              * mode. Next pop will be actual timer expiration.
-@@ -116,6 +115,11 @@ static uint32_t timer_enabled(HPETTimer *t)
- }
- 
- static uint32_t hpet_time_after(uint64_t a, uint64_t b)
-+{
-+    return ((int32_t)(b - a) < 0);
-+}
-+
-+static uint32_t hpet_time_after64(uint64_t a, uint64_t b)
- {
-     return ((int64_t)(b - a) < 0);
- }
-@@ -152,32 +156,27 @@ static uint64_t hpet_get_ticks(HPETState *s)
-     return ns_to_ticks(qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL) + s->hpet_offset);
- }
- 
--static uint64_t hpet_get_ns(HPETState *s, uint64_t tick)
--{
--    return ticks_to_ns(tick) - s->hpet_offset;
--}
--
- /*
-- * calculate next value of the general counter that matches the
-- * target (either entirely, or the low 32-bit only depending on
-- * the timer mode).
-+ * calculate diff between comparator value and current ticks
-  */
--static uint64_t hpet_calculate_cmp64(HPETTimer *t, uint64_t cur_tick, uint64_t target)
-+static inline uint64_t hpet_calculate_diff(HPETTimer *t, uint64_t current)
- {
-+
-     if (t->config & HPET_TN_32BIT) {
--        uint64_t result = deposit64(cur_tick, 0, 32, target);
--        if (result < cur_tick) {
--            result += 0x100000000ULL;
--        }
--        return result;
-+        uint32_t diff, cmp;
-+
-+        cmp = (uint32_t)t->cmp;
-+        diff = cmp - (uint32_t)current;
-+        diff = (int32_t)diff > 0 ? diff : (uint32_t)1;
-+        return (uint64_t)diff;
-     } else {
--        return target;
--    }
--}
-+        uint64_t diff, cmp;
- 
--static uint64_t hpet_next_wrap(uint64_t cur_tick)
--{
--    return (cur_tick | 0xffffffffU) + 1;
-+        cmp = t->cmp;
-+        diff = cmp - current;
-+        diff = (int64_t)diff > 0 ? diff : (uint64_t)1;
-+        return diff;
-+    }
- }
- 
- static void update_irq(struct HPETTimer *timer, int set)
-@@ -261,12 +260,7 @@ static bool hpet_validate_num_timers(void *opaque, int version_id)
- static int hpet_post_load(void *opaque, int version_id)
- {
-     HPETState *s = opaque;
--    int i;
- 
--    for (i = 0; i < s->num_timers; i++) {
--        HPETTimer *t = &s->timer[i];
--        t->cmp64 = hpet_calculate_cmp64(t, s->hpet_counter, t->cmp);
--    }
-     /* Recalculate the offset between the main counter and guest time */
-     if (!s->hpet_offset_saved) {
-         s->hpet_offset = ticks_to_ns(s->hpet_counter)
-@@ -362,10 +356,14 @@ static const VMStateDescription vmstate_hpet = {
-     }
- };
- 
--static void hpet_arm(HPETTimer *t, uint64_t tick)
-+static void hpet_arm(HPETTimer *t, uint64_t ticks)
- {
--    /* FIXME: Clamp period to reasonable min value? */
--    timer_mod(t->qemu_timer, hpet_get_ns(t->state, tick));
-+    if (ticks < ns_to_ticks(INT64_MAX / 2)) {
-+        timer_mod(t->qemu_timer,
-+                  qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL) + ticks_to_ns(ticks));
-+    } else {
-+        timer_del(t->qemu_timer);
-+    }
- }
- 
- /*
-@@ -374,44 +372,54 @@ static void hpet_arm(HPETTimer *t, uint64_t tick)
- static void hpet_timer(void *opaque)
- {
-     HPETTimer *t = opaque;
-+    uint64_t diff;
-+
-     uint64_t period = t->period;
-     uint64_t cur_tick = hpet_get_ticks(t->state);
- 
-     if (timer_is_periodic(t) && period != 0) {
--        while (hpet_time_after(cur_tick, t->cmp64)) {
--            t->cmp64 += period;
--        }
-         if (t->config & HPET_TN_32BIT) {
--            t->cmp = (uint32_t)t->cmp64;
-+            while (hpet_time_after(cur_tick, t->cmp)) {
-+                t->cmp = (uint32_t)(t->cmp + t->period);
-+            }
-         } else {
--            t->cmp = t->cmp64;
-+            while (hpet_time_after64(cur_tick, t->cmp)) {
-+                t->cmp += period;
-+            }
-+        }
-+        diff = hpet_calculate_diff(t, cur_tick);
-+        hpet_arm(t, diff);
-+    } else if (t->config & HPET_TN_32BIT && !timer_is_periodic(t)) {
-+        if (t->wrap_flag) {
-+            diff = hpet_calculate_diff(t, cur_tick);
-+            hpet_arm(t, diff);
-+            t->wrap_flag = 0;
-         }
--        hpet_arm(t, t->cmp64);
--    } else if (t->wrap_flag) {
--        t->wrap_flag = 0;
--        hpet_arm(t, t->cmp64);
-     }
-     update_irq(t, 1);
- }
- 
- static void hpet_set_timer(HPETTimer *t)
- {
-+    uint64_t diff;
-+    uint32_t wrap_diff;  /* how many ticks until we wrap? */
-     uint64_t cur_tick = hpet_get_ticks(t->state);
- 
-+    /* whenever new timer is being set up, make sure wrap_flag is 0 */
-     t->wrap_flag = 0;
--    t->cmp64 = hpet_calculate_cmp64(t, cur_tick, t->cmp);
--    if (t->config & HPET_TN_32BIT) {
--
--        /* hpet spec says in one-shot 32-bit mode, generate an interrupt when
--         * counter wraps in addition to an interrupt with comparator match.
--         */
--        if (!timer_is_periodic(t) && t->cmp64 > hpet_next_wrap(cur_tick)) {
-+    diff = hpet_calculate_diff(t, cur_tick);
-+
-+    /* hpet spec says in one-shot 32-bit mode, generate an interrupt when
-+     * counter wraps in addition to an interrupt with comparator match.
-+     */
-+    if (t->config & HPET_TN_32BIT && !timer_is_periodic(t)) {
-+        wrap_diff = 0xffffffff - (uint32_t)cur_tick;
-+        if (wrap_diff < (uint32_t)diff) {
-+            diff = wrap_diff;
-             t->wrap_flag = 1;
--            hpet_arm(t, hpet_next_wrap(cur_tick));
--            return;
-         }
-     }
--    hpet_arm(t, t->cmp64);
-+    hpet_arm(t, diff);
- }
- 
- static void hpet_del_timer(HPETTimer *t)
-@@ -542,7 +550,12 @@ static void hpet_ram_write(void *opaque, hwaddr addr,
-                 timer->cmp = deposit64(timer->cmp, shift, len, value);
-             }
-             if (timer_is_periodic(timer)) {
--                timer->period = deposit64(timer->period, shift, len, value);
-+                /*
-+                 * FIXME: Clamp period to reasonable min value?
-+                 * Clamp period to reasonable max value
-+                 */
-+                new_val = deposit64(timer->period, shift, len, value);
-+                timer->period = MIN(new_val, (timer->config & HPET_TN_32BIT ? ~0u : ~0ull) >> 1);
-             }
-             timer->config &= ~HPET_TN_SETVAL;
-             if (hpet_enabled(s)) {

debian/patches/pve/0053-Revert-hpet-accept-64-bit-reads-and-writes.patch

@@ -1,281 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Fiona Ebner <f.ebner@proxmox.com>
-Date: Wed, 19 Mar 2025 17:31:09 +0100
-Subject: [PATCH] Revert "hpet: accept 64-bit reads and writes"
-
-This reverts commit c2366567378dd8fb89329816003801f54e30e6f3.
-
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
----
- hw/timer/hpet.c       | 137 +++++++++++++++++++++++++++++-------------
- hw/timer/trace-events |   3 +-
- 2 files changed, 96 insertions(+), 44 deletions(-)
-
-diff --git a/hw/timer/hpet.c b/hw/timer/hpet.c
-index 415a9433f1..e1ac877759 100644
---- a/hw/timer/hpet.c
-+++ b/hw/timer/hpet.c
-@@ -437,7 +437,6 @@ static uint64_t hpet_ram_read(void *opaque, hwaddr addr,
-                               unsigned size)
- {
-     HPETState *s = opaque;
--    int shift = (addr & 4) * 8;
-     uint64_t cur_tick;
- 
-     trace_hpet_ram_read(addr);
-@@ -452,33 +451,52 @@ static uint64_t hpet_ram_read(void *opaque, hwaddr addr,
-             return 0;
-         }
- 
--        switch (addr & 0x18) {
--        case HPET_TN_CFG: // including interrupt capabilities
--            return timer->config >> shift;
-+        switch ((addr - 0x100) % 0x20) {
-+        case HPET_TN_CFG:
-+            return timer->config;
-+        case HPET_TN_CFG + 4: // Interrupt capabilities
-+            return timer->config >> 32;
-         case HPET_TN_CMP: // comparator register
--            return timer->cmp >> shift;
-+            return timer->cmp;
-+        case HPET_TN_CMP + 4:
-+            return timer->cmp >> 32;
-         case HPET_TN_ROUTE:
--            return timer->fsb >> shift;
-+            return timer->fsb;
-+        case HPET_TN_ROUTE + 4:
-+            return timer->fsb >> 32;
-         default:
-             trace_hpet_ram_read_invalid();
-             break;
-         }
-     } else {
--        switch (addr & ~4) {
--        case HPET_ID: // including HPET_PERIOD
--            return s->capability >> shift;
-+        switch (addr) {
-+        case HPET_ID:
-+            return s->capability;
-+        case HPET_PERIOD:
-+            return s->capability >> 32;
-         case HPET_CFG:
--            return s->config >> shift;
-+            return s->config;
-+        case HPET_CFG + 4:
-+            trace_hpet_invalid_hpet_cfg(4);
-+            return 0;
-         case HPET_COUNTER:
-             if (hpet_enabled(s)) {
-                 cur_tick = hpet_get_ticks(s);
-             } else {
-                 cur_tick = s->hpet_counter;
-             }
--            trace_hpet_ram_read_reading_counter(addr & 4, cur_tick);
--            return cur_tick >> shift;
-+            trace_hpet_ram_read_reading_counter(0, cur_tick);
-+            return cur_tick;
-+        case HPET_COUNTER + 4:
-+            if (hpet_enabled(s)) {
-+                cur_tick = hpet_get_ticks(s);
-+            } else {
-+                cur_tick = s->hpet_counter;
-+            }
-+            trace_hpet_ram_read_reading_counter(4, cur_tick);
-+            return cur_tick >> 32;
-         case HPET_STATUS:
--            return s->isr >> shift;
-+            return s->isr;
-         default:
-             trace_hpet_ram_read_invalid();
-             break;
-@@ -492,11 +510,11 @@ static void hpet_ram_write(void *opaque, hwaddr addr,
- {
-     int i;
-     HPETState *s = opaque;
--    int shift = (addr & 4) * 8;
--    int len = MIN(size * 8, 64 - shift);
-     uint64_t old_val, new_val, cleared;
- 
-     trace_hpet_ram_write(addr, value);
-+    old_val = hpet_ram_read(opaque, addr, 4);
-+    new_val = value;
- 
-     /*address range of all TN regs*/
-     if (addr >= 0x100 && addr <= 0x3ff) {
-@@ -508,12 +526,9 @@ static void hpet_ram_write(void *opaque, hwaddr addr,
-             trace_hpet_timer_id_out_of_range(timer_id);
-             return;
-         }
--        switch (addr & 0x18) {
-+        switch ((addr - 0x100) % 0x20) {
-         case HPET_TN_CFG:
--            trace_hpet_ram_write_tn_cfg(addr & 4);
--            old_val = timer->config;
--            new_val = deposit64(old_val, shift, len, value);
--            new_val = hpet_fixup_reg(new_val, old_val, HPET_TN_CFG_WRITE_MASK);
-+            trace_hpet_ram_write_tn_cfg();
-             if (deactivating_bit(old_val, new_val, HPET_TN_TYPE_LEVEL)) {
-                 /*
-                  * Do this before changing timer->config; otherwise, if
-@@ -521,7 +536,8 @@ static void hpet_ram_write(void *opaque, hwaddr addr,
-                  */
-                 update_irq(timer, 0);
-             }
--            timer->config = new_val;
-+            new_val = hpet_fixup_reg(new_val, old_val, HPET_TN_CFG_WRITE_MASK);
-+            timer->config = (timer->config & 0xffffffff00000000ULL) | new_val;
-             if (activating_bit(old_val, new_val, HPET_TN_ENABLE)
-                 && (s->isr & (1 << timer_id))) {
-                 update_irq(timer, 1);
-@@ -534,28 +550,56 @@ static void hpet_ram_write(void *opaque, hwaddr addr,
-                 hpet_set_timer(timer);
-             }
-             break;
-+        case HPET_TN_CFG + 4: // Interrupt capabilities
-+            trace_hpet_ram_write_invalid_tn_cfg(4);
-+            break;
-         case HPET_TN_CMP: // comparator register
-+            trace_hpet_ram_write_tn_cmp(0);
-             if (timer->config & HPET_TN_32BIT) {
--                /* High 32-bits are zero, leave them untouched.  */
--                if (shift) {
--                    trace_hpet_ram_write_invalid_tn_cmp();
--                    break;
-+                new_val = (uint32_t)new_val;
-+            }
-+            if (!timer_is_periodic(timer)
-+                || (timer->config & HPET_TN_SETVAL)) {
-+                timer->cmp = (timer->cmp & 0xffffffff00000000ULL) | new_val;
-+            }
-+            if (timer_is_periodic(timer)) {
-+                /*
-+                 * FIXME: Clamp period to reasonable min value?
-+                 * Clamp period to reasonable max value
-+                 */
-+                if (timer->config & HPET_TN_32BIT) {
-+                    new_val = MIN(new_val, ~0u >> 1);
-                 }
--                len = 64;
--                value = (uint32_t) value;
-+                timer->period =
-+                    (timer->period & 0xffffffff00000000ULL) | new_val;
-+            }
-+            /*
-+             * FIXME: on a 64-bit write, HPET_TN_SETVAL should apply to the
-+             * high bits part as well.
-+             */
-+            timer->config &= ~HPET_TN_SETVAL;
-+            if (hpet_enabled(s)) {
-+                hpet_set_timer(timer);
-             }
--            trace_hpet_ram_write_tn_cmp(addr & 4);
-+            break;
-+        case HPET_TN_CMP + 4: // comparator register high order
-+            if (timer->config & HPET_TN_32BIT) {
-+                trace_hpet_ram_write_invalid_tn_cmp();
-+                break;
-+            }
-+            trace_hpet_ram_write_tn_cmp(4);
-             if (!timer_is_periodic(timer)
-                 || (timer->config & HPET_TN_SETVAL)) {
--                timer->cmp = deposit64(timer->cmp, shift, len, value);
-+                timer->cmp = (timer->cmp & 0xffffffffULL) | new_val << 32;
-             }
-             if (timer_is_periodic(timer)) {
-                 /*
-                  * FIXME: Clamp period to reasonable min value?
-                  * Clamp period to reasonable max value
-                  */
--                new_val = deposit64(timer->period, shift, len, value);
--                timer->period = MIN(new_val, (timer->config & HPET_TN_32BIT ? ~0u : ~0ull) >> 1);
-+                new_val = MIN(new_val, ~0u >> 1);
-+                timer->period =
-+                    (timer->period & 0xffffffffULL) | new_val << 32;
-             }
-             timer->config &= ~HPET_TN_SETVAL;
-             if (hpet_enabled(s)) {
-@@ -563,7 +607,10 @@ static void hpet_ram_write(void *opaque, hwaddr addr,
-             }
-             break;
-         case HPET_TN_ROUTE:
--            timer->fsb = deposit64(timer->fsb, shift, len, value);
-+            timer->fsb = (timer->fsb & 0xffffffff00000000ULL) | new_val;
-+            break;
-+        case HPET_TN_ROUTE + 4:
-+            timer->fsb = (new_val << 32) | (timer->fsb & 0xffffffff);
-             break;
-         default:
-             trace_hpet_ram_write_invalid();
-@@ -571,14 +618,12 @@ static void hpet_ram_write(void *opaque, hwaddr addr,
-         }
-         return;
-     } else {
--        switch (addr & ~4) {
-+        switch (addr) {
-         case HPET_ID:
-             return;
-         case HPET_CFG:
--            old_val = s->config;
--            new_val = deposit64(old_val, shift, len, value);
-             new_val = hpet_fixup_reg(new_val, old_val, HPET_CFG_WRITE_MASK);
--            s->config = new_val;
-+            s->config = (s->config & 0xffffffff00000000ULL) | new_val;
-             if (activating_bit(old_val, new_val, HPET_CFG_ENABLE)) {
-                 /* Enable main counter and interrupt generation. */
-                 s->hpet_offset =
-@@ -608,8 +653,10 @@ static void hpet_ram_write(void *opaque, hwaddr addr,
-                 qemu_set_irq(s->irqs[RTC_ISA_IRQ], s->rtc_irq_level);
-             }
-             break;
-+        case HPET_CFG + 4:
-+            trace_hpet_invalid_hpet_cfg(4);
-+            break;
-         case HPET_STATUS:
--            new_val = value << shift;
-             cleared = new_val & s->isr;
-             for (i = 0; i < s->num_timers; i++) {
-                 if (cleared & (1 << i)) {
-@@ -621,7 +668,15 @@ static void hpet_ram_write(void *opaque, hwaddr addr,
-             if (hpet_enabled(s)) {
-                 trace_hpet_ram_write_counter_write_while_enabled();
-             }
--            s->hpet_counter = deposit64(s->hpet_counter, shift, len, value);
-+            s->hpet_counter =
-+                (s->hpet_counter & 0xffffffff00000000ULL) | value;
-+            trace_hpet_ram_write_counter_written(0, value, s->hpet_counter);
-+            break;
-+        case HPET_COUNTER + 4:
-+            trace_hpet_ram_write_counter_write_while_enabled();
-+            s->hpet_counter =
-+                (s->hpet_counter & 0xffffffffULL) | (((uint64_t)value) << 32);
-+            trace_hpet_ram_write_counter_written(4, value, s->hpet_counter);
-             break;
-         default:
-             trace_hpet_ram_write_invalid();
-@@ -635,11 +690,7 @@ static const MemoryRegionOps hpet_ram_ops = {
-     .write = hpet_ram_write,
-     .valid = {
-         .min_access_size = 4,
--        .max_access_size = 8,
--    },
--    .impl = {
--        .min_access_size = 4,
--        .max_access_size = 8,
-+        .max_access_size = 4,
-     },
-     .endianness = DEVICE_NATIVE_ENDIAN,
- };
-diff --git a/hw/timer/trace-events b/hw/timer/trace-events
-index 5cfc369fba..219747df2f 100644
---- a/hw/timer/trace-events
-+++ b/hw/timer/trace-events
-@@ -114,7 +114,8 @@ hpet_ram_read_reading_counter(uint8_t reg_off, uint64_t cur_tick) "reading count
- hpet_ram_read_invalid(void) "invalid hpet_ram_readl"
- hpet_ram_write(uint64_t addr, uint64_t value) "enter hpet_ram_writel at 0x%" PRIx64 " = 0x%" PRIx64
- hpet_ram_write_timer_id(uint64_t timer_id) "hpet_ram_writel timer_id = 0x%" PRIx64
--hpet_ram_write_tn_cfg(uint8_t reg_off) "hpet_ram_writel HPET_TN_CFG + %" PRIu8
-+hpet_ram_write_tn_cfg(void) "hpet_ram_writel HPET_TN_CFG"
-+hpet_ram_write_invalid_tn_cfg(uint8_t reg_off) "invalid HPET_TN_CFG + %" PRIu8 " write"
- hpet_ram_write_tn_cmp(uint8_t reg_off) "hpet_ram_writel HPET_TN_CMP + %" PRIu8
- hpet_ram_write_invalid_tn_cmp(void) "invalid HPET_TN_CMP + 4 write"
- hpet_ram_write_invalid(void) "invalid hpet_ram_writel"

debian/patches/pve/0054-Revert-hpet-place-read-only-bits-directly-in-new_val.patch

@@ -1,64 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Fiona Ebner <f.ebner@proxmox.com>
-Date: Wed, 19 Mar 2025 17:31:10 +0100
-Subject: [PATCH] Revert "hpet: place read-only bits directly in "new_val""
-
-This reverts commit ba88935b0fac2588b0a739f810b58dfabf7f92c8.
-
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
----
- hw/timer/hpet.c | 15 ++++++++-------
- 1 file changed, 8 insertions(+), 7 deletions(-)
-
-diff --git a/hw/timer/hpet.c b/hw/timer/hpet.c
-index e1ac877759..b12bbaf10d 100644
---- a/hw/timer/hpet.c
-+++ b/hw/timer/hpet.c
-@@ -510,7 +510,7 @@ static void hpet_ram_write(void *opaque, hwaddr addr,
- {
-     int i;
-     HPETState *s = opaque;
--    uint64_t old_val, new_val, cleared;
-+    uint64_t old_val, new_val, val;
- 
-     trace_hpet_ram_write(addr, value);
-     old_val = hpet_ram_read(opaque, addr, 4);
-@@ -536,12 +536,13 @@ static void hpet_ram_write(void *opaque, hwaddr addr,
-                  */
-                 update_irq(timer, 0);
-             }
--            new_val = hpet_fixup_reg(new_val, old_val, HPET_TN_CFG_WRITE_MASK);
--            timer->config = (timer->config & 0xffffffff00000000ULL) | new_val;
-+            val = hpet_fixup_reg(new_val, old_val, HPET_TN_CFG_WRITE_MASK);
-+            timer->config = (timer->config & 0xffffffff00000000ULL) | val;
-             if (activating_bit(old_val, new_val, HPET_TN_ENABLE)
-                 && (s->isr & (1 << timer_id))) {
-                 update_irq(timer, 1);
-             }
-+
-             if (new_val & HPET_TN_32BIT) {
-                 timer->cmp = (uint32_t)timer->cmp;
-                 timer->period = (uint32_t)timer->period;
-@@ -622,8 +623,8 @@ static void hpet_ram_write(void *opaque, hwaddr addr,
-         case HPET_ID:
-             return;
-         case HPET_CFG:
--            new_val = hpet_fixup_reg(new_val, old_val, HPET_CFG_WRITE_MASK);
--            s->config = (s->config & 0xffffffff00000000ULL) | new_val;
-+            val = hpet_fixup_reg(new_val, old_val, HPET_CFG_WRITE_MASK);
-+            s->config = (s->config & 0xffffffff00000000ULL) | val;
-             if (activating_bit(old_val, new_val, HPET_CFG_ENABLE)) {
-                 /* Enable main counter and interrupt generation. */
-                 s->hpet_offset =
-@@ -657,9 +658,9 @@ static void hpet_ram_write(void *opaque, hwaddr addr,
-             trace_hpet_invalid_hpet_cfg(4);
-             break;
-         case HPET_STATUS:
--            cleared = new_val & s->isr;
-+            val = new_val & s->isr;
-             for (i = 0; i < s->num_timers; i++) {
--                if (cleared & (1 << i)) {
-+                if (val & (1 << i)) {
-                     update_irq(&s->timer[i], 0);
-                 }
-             }

debian/patches/pve/0055-Revert-hpet-remove-unnecessary-variable-index.patch

@@ -1,68 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Fiona Ebner <f.ebner@proxmox.com>
-Date: Wed, 19 Mar 2025 17:31:11 +0100
-Subject: [PATCH] Revert "hpet: remove unnecessary variable "index""
-
-This reverts commit 5895879aca252f4ebb2d1078eaf836c61ec54e9b.
-
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
----
- hw/timer/hpet.c | 15 ++++++++-------
- 1 file changed, 8 insertions(+), 7 deletions(-)
-
-diff --git a/hw/timer/hpet.c b/hw/timer/hpet.c
-index b12bbaf10d..6f83d88516 100644
---- a/hw/timer/hpet.c
-+++ b/hw/timer/hpet.c
-@@ -437,12 +437,12 @@ static uint64_t hpet_ram_read(void *opaque, hwaddr addr,
-                               unsigned size)
- {
-     HPETState *s = opaque;
--    uint64_t cur_tick;
-+    uint64_t cur_tick, index;
- 
-     trace_hpet_ram_read(addr);
--
-+    index = addr;
-     /*address range of all TN regs*/
--    if (addr >= 0x100 && addr <= 0x3ff) {
-+    if (index >= 0x100 && index <= 0x3ff) {
-         uint8_t timer_id = (addr - 0x100) / 0x20;
-         HPETTimer *timer = &s->timer[timer_id];
- 
-@@ -469,7 +469,7 @@ static uint64_t hpet_ram_read(void *opaque, hwaddr addr,
-             break;
-         }
-     } else {
--        switch (addr) {
-+        switch (index) {
-         case HPET_ID:
-             return s->capability;
-         case HPET_PERIOD:
-@@ -510,14 +510,15 @@ static void hpet_ram_write(void *opaque, hwaddr addr,
- {
-     int i;
-     HPETState *s = opaque;
--    uint64_t old_val, new_val, val;
-+    uint64_t old_val, new_val, val, index;
- 
-     trace_hpet_ram_write(addr, value);
-+    index = addr;
-     old_val = hpet_ram_read(opaque, addr, 4);
-     new_val = value;
- 
-     /*address range of all TN regs*/
--    if (addr >= 0x100 && addr <= 0x3ff) {
-+    if (index >= 0x100 && index <= 0x3ff) {
-         uint8_t timer_id = (addr - 0x100) / 0x20;
-         HPETTimer *timer = &s->timer[timer_id];
- 
-@@ -619,7 +620,7 @@ static void hpet_ram_write(void *opaque, hwaddr addr,
-         }
-         return;
-     } else {
--        switch (addr) {
-+        switch (index) {
-         case HPET_ID:
-             return;
-         case HPET_CFG:

debian/patches/pve/0056-Revert-hpet-ignore-high-bits-of-comparator-in-32-bit.patch

@@ -1,40 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Fiona Ebner <f.ebner@proxmox.com>
-Date: Wed, 19 Mar 2025 17:31:12 +0100
-Subject: [PATCH] Revert "hpet: ignore high bits of comparator in 32-bit mode"
-
-This reverts commit 9eb7fad3546a89ee7cf0e90f5b1daccf89725cea.
-
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
----
- hw/timer/hpet.c       | 4 ----
- hw/timer/trace-events | 1 -
- 2 files changed, 5 deletions(-)
-
-diff --git a/hw/timer/hpet.c b/hw/timer/hpet.c
-index 6f83d88516..509986c0a9 100644
---- a/hw/timer/hpet.c
-+++ b/hw/timer/hpet.c
-@@ -585,10 +585,6 @@ static void hpet_ram_write(void *opaque, hwaddr addr,
-             }
-             break;
-         case HPET_TN_CMP + 4: // comparator register high order
--            if (timer->config & HPET_TN_32BIT) {
--                trace_hpet_ram_write_invalid_tn_cmp();
--                break;
--            }
-             trace_hpet_ram_write_tn_cmp(4);
-             if (!timer_is_periodic(timer)
-                 || (timer->config & HPET_TN_SETVAL)) {
-diff --git a/hw/timer/trace-events b/hw/timer/trace-events
-index 219747df2f..b86870fb22 100644
---- a/hw/timer/trace-events
-+++ b/hw/timer/trace-events
-@@ -117,7 +117,6 @@ hpet_ram_write_timer_id(uint64_t timer_id) "hpet_ram_writel timer_id = 0x%" PRIx
- hpet_ram_write_tn_cfg(void) "hpet_ram_writel HPET_TN_CFG"
- hpet_ram_write_invalid_tn_cfg(uint8_t reg_off) "invalid HPET_TN_CFG + %" PRIu8 " write"
- hpet_ram_write_tn_cmp(uint8_t reg_off) "hpet_ram_writel HPET_TN_CMP + %" PRIu8
--hpet_ram_write_invalid_tn_cmp(void) "invalid HPET_TN_CMP + 4 write"
- hpet_ram_write_invalid(void) "invalid hpet_ram_writel"
- hpet_ram_write_counter_write_while_enabled(void) "Writing counter while HPET enabled!"
- hpet_ram_write_counter_written(uint8_t reg_off, uint64_t value, uint64_t counter) "HPET counter + %" PRIu8 "written. crt = 0x%" PRIx64 " -> 0x%" PRIx64

debian/patches/pve/0057-Revert-hpet-fix-and-cleanup-persistence-of-interrupt.patch

@@ -1,120 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Fiona Ebner <f.ebner@proxmox.com>
-Date: Wed, 19 Mar 2025 17:31:13 +0100
-Subject: [PATCH] Revert "hpet: fix and cleanup persistence of interrupt
- status"
-
-This reverts commit f0ccf770789e48b7a73497b465fdc892d28c1339.
-
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
----
- hw/timer/hpet.c | 60 ++++++++++++++++---------------------------------
- 1 file changed, 19 insertions(+), 41 deletions(-)
-
-diff --git a/hw/timer/hpet.c b/hw/timer/hpet.c
-index 509986c0a9..402cc960f0 100644
---- a/hw/timer/hpet.c
-+++ b/hw/timer/hpet.c
-@@ -196,31 +196,21 @@ static void update_irq(struct HPETTimer *timer, int set)
-     }
-     s = timer->state;
-     mask = 1 << timer->tn;
--
--    if (set && (timer->config & HPET_TN_TYPE_LEVEL)) {
--        /*
--         * If HPET_TN_ENABLE bit is 0, "the timer will still operate and
--         * generate appropriate status bits, but will not cause an interrupt"
--         */
--        s->isr |= mask;
--    } else {
-+    if (!set || !timer_enabled(timer) || !hpet_enabled(timer->state)) {
-         s->isr &= ~mask;
--    }
--
--    if (set && timer_enabled(timer) && hpet_enabled(s)) {
--        if (timer_fsb_route(timer)) {
--            address_space_stl_le(&address_space_memory, timer->fsb >> 32,
--                                 timer->fsb & 0xffffffff, MEMTXATTRS_UNSPECIFIED,
--                                 NULL);
--        } else if (timer->config & HPET_TN_TYPE_LEVEL) {
--            qemu_irq_raise(s->irqs[route]);
--        } else {
--            qemu_irq_pulse(s->irqs[route]);
--        }
--    } else {
-         if (!timer_fsb_route(timer)) {
-             qemu_irq_lower(s->irqs[route]);
-         }
-+    } else if (timer_fsb_route(timer)) {
-+        address_space_stl_le(&address_space_memory, timer->fsb >> 32,
-+                             timer->fsb & 0xffffffff, MEMTXATTRS_UNSPECIFIED,
-+                             NULL);
-+    } else if (timer->config & HPET_TN_TYPE_LEVEL) {
-+        s->isr |= mask;
-+        qemu_irq_raise(s->irqs[route]);
-+    } else {
-+        s->isr &= ~mask;
-+        qemu_irq_pulse(s->irqs[route]);
-     }
- }
- 
-@@ -424,13 +414,8 @@ static void hpet_set_timer(HPETTimer *t)
- 
- static void hpet_del_timer(HPETTimer *t)
- {
--    HPETState *s = t->state;
-     timer_del(t->qemu_timer);
--
--    if (s->isr & (1 << t->tn)) {
--        /* For level-triggered interrupt, this leaves ISR set but lowers irq.  */
--        update_irq(t, 1);
--    }
-+    update_irq(t, 0);
- }
- 
- static uint64_t hpet_ram_read(void *opaque, hwaddr addr,
-@@ -530,26 +515,20 @@ static void hpet_ram_write(void *opaque, hwaddr addr,
-         switch ((addr - 0x100) % 0x20) {
-         case HPET_TN_CFG:
-             trace_hpet_ram_write_tn_cfg();
--            if (deactivating_bit(old_val, new_val, HPET_TN_TYPE_LEVEL)) {
--                /*
--                 * Do this before changing timer->config; otherwise, if
--                 * HPET_TN_FSB is set, update_irq will not lower the qemu_irq.
--                 */
-+            if (activating_bit(old_val, new_val, HPET_TN_FSB_ENABLE)) {
-                 update_irq(timer, 0);
-             }
-             val = hpet_fixup_reg(new_val, old_val, HPET_TN_CFG_WRITE_MASK);
-             timer->config = (timer->config & 0xffffffff00000000ULL) | val;
--            if (activating_bit(old_val, new_val, HPET_TN_ENABLE)
--                && (s->isr & (1 << timer_id))) {
--                update_irq(timer, 1);
--            }
--
-             if (new_val & HPET_TN_32BIT) {
-                 timer->cmp = (uint32_t)timer->cmp;
-                 timer->period = (uint32_t)timer->period;
-             }
--            if (hpet_enabled(s)) {
-+            if (activating_bit(old_val, new_val, HPET_TN_ENABLE) &&
-+                hpet_enabled(s)) {
-                 hpet_set_timer(timer);
-+            } else if (deactivating_bit(old_val, new_val, HPET_TN_ENABLE)) {
-+                hpet_del_timer(timer);
-             }
-             break;
-         case HPET_TN_CFG + 4: // Interrupt capabilities
-@@ -627,10 +606,9 @@ static void hpet_ram_write(void *opaque, hwaddr addr,
-                 s->hpet_offset =
-                     ticks_to_ns(s->hpet_counter) - qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL);
-                 for (i = 0; i < s->num_timers; i++) {
--                    if (timer_enabled(&s->timer[i]) && (s->isr & (1 << i))) {
--                        update_irq(&s->timer[i], 1);
-+                    if ((&s->timer[i])->cmp != ~0ULL) {
-+                        hpet_set_timer(&s->timer[i]);
-                     }
--                    hpet_set_timer(&s->timer[i]);
-                 }
-             } else if (deactivating_bit(old_val, new_val, HPET_CFG_ENABLE)) {
-                 /* Halt main counter and disable interrupt generation. */

debian/patches/pve/0058-savevm-async-improve-setting-state-of-snapshot-opera.patch

@@ -1,81 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Fiona Ebner <f.ebner@proxmox.com>
-Date: Mon, 31 Mar 2025 16:55:02 +0200
-Subject: [PATCH] savevm-async: improve setting state of snapshot operation in
- savevm-end handler
-
-One of the callers of wait_for_close_co() already sets the state to
-SAVE_STATE_DONE before, but that is not fully correct, because at that
-moment, the operation is not fully done. In particular, if closing the
-target later fails, the state would even be set to SAVE_STATE_ERROR
-afterwards. DONE -> ERROR is not a valid state transition. Although,
-it should not matter in practice as long as the relevant QMP commands
-are sequential.
-
-The other caller does not set the state and so there seems to be a
-race that could lead to the state not getting set at all. This is
-because before this commit, the wait_for_close_co() function could
-return early when there is no target file anymore. This can only
-happen when canceling and needs to happen right around the time when
-the snapshot is already finishing and closing the target.
-
-Simply avoid the early return and always set the state within the
-wait_for_close_co() function rather than at the call site.
-
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
----
- migration/savevm-async.c | 33 +++++++++++++++------------------
- 1 file changed, 15 insertions(+), 18 deletions(-)
-
-diff --git a/migration/savevm-async.c b/migration/savevm-async.c
-index 1e79fce9ba..e63dc6d8a3 100644
---- a/migration/savevm-async.c
-+++ b/migration/savevm-async.c
-@@ -450,23 +450,22 @@ static void coroutine_fn wait_for_close_co(void *opaque)
- {
-     int64_t timeout;
- 
--    if (!snap_state.target) {
--        DPRINTF("savevm-end: no target file open\n");
--        return;
--    }
--
--    /* wait until cleanup is done before returning, this ensures that after this
--     * call exits the statefile will be closed and can be removed immediately */
--    DPRINTF("savevm-end: waiting for cleanup\n");
--    timeout = 30L * 1000 * 1000 * 1000;
--    qemu_co_sleep_ns_wakeable(&snap_state.target_close_wait,
--                              QEMU_CLOCK_REALTIME, timeout);
-     if (snap_state.target) {
--        save_snapshot_error("timeout waiting for target file close in "
--                            "qmp_savevm_end");
--        /* we cannot assume the snapshot finished in this case, so leave the
--         * state alone - caller has to figure something out */
--        return;
-+        /* wait until cleanup is done before returning, this ensures that after this
-+         * call exits the statefile will be closed and can be removed immediately */
-+        DPRINTF("savevm-end: waiting for cleanup\n");
-+        timeout = 30L * 1000 * 1000 * 1000;
-+        qemu_co_sleep_ns_wakeable(&snap_state.target_close_wait,
-+                                  QEMU_CLOCK_REALTIME, timeout);
-+        if (snap_state.target) {
-+            save_snapshot_error("timeout waiting for target file close in "
-+                                "qmp_savevm_end");
-+            /* we cannot assume the snapshot finished in this case, so leave the
-+             * state alone - caller has to figure something out */
-+            return;
-+        }
-+    } else {
-+        DPRINTF("savevm-end: no target file open\n");
-     }
- 
-     // File closed and no other error, so ensure next snapshot can be started.
-@@ -497,8 +496,6 @@ void qmp_savevm_end(Error **errp)
-         snap_state.saved_vm_running = false;
-     }
- 
--    snap_state.state = SAVE_STATE_DONE;
--
-     qemu_coroutine_enter(wait_for_close);
- }
- 

debian/patches/pve/0059-savevm-async-rename-saved_vm_running-to-vm_needs_sta.patch

@@ -1,71 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Fiona Ebner <f.ebner@proxmox.com>
-Date: Mon, 31 Mar 2025 16:55:03 +0200
-Subject: [PATCH] savevm-async: rename saved_vm_running to vm_needs_start
-
-This is what the variable actually expresses. Otherwise, setting it
-to false after starting the VM doesn't make sense.
-
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
----
- migration/savevm-async.c | 16 ++++++++--------
- 1 file changed, 8 insertions(+), 8 deletions(-)
-
-diff --git a/migration/savevm-async.c b/migration/savevm-async.c
-index e63dc6d8a3..1e34c31e8b 100644
---- a/migration/savevm-async.c
-+++ b/migration/savevm-async.c
-@@ -51,7 +51,7 @@ static struct SnapshotState {
-     int state;
-     Error *error;
-     Error *blocker;
--    int saved_vm_running;
-+    int vm_needs_start;
-     QEMUFile *file;
-     int64_t total_time;
-     QEMUBH *finalize_bh;
-@@ -224,9 +224,9 @@ static void process_savevm_finalize(void *opaque)
-         save_snapshot_error("process_savevm_cleanup: invalid state: %d",
-                             snap_state.state);
-     }
--    if (snap_state.saved_vm_running) {
-+    if (snap_state.vm_needs_start) {
-         vm_start();
--        snap_state.saved_vm_running = false;
-+        snap_state.vm_needs_start = false;
-     }
- 
-     DPRINTF("timing: process_savevm_finalize (full) took %ld ms\n",
-@@ -352,7 +352,7 @@ void qmp_savevm_start(const char *statefile, Error **errp)
-     }
- 
-     /* initialize snapshot info */
--    snap_state.saved_vm_running = runstate_is_running();
-+    snap_state.vm_needs_start = runstate_is_running();
-     snap_state.bs_pos = 0;
-     snap_state.total_time = qemu_clock_get_ms(QEMU_CLOCK_REALTIME);
-     snap_state.blocker = NULL;
-@@ -440,9 +440,9 @@ restart:
- 
-     save_snapshot_error("setup failed");
- 
--    if (snap_state.saved_vm_running) {
-+    if (snap_state.vm_needs_start) {
-         vm_start();
--        snap_state.saved_vm_running = false;
-+        snap_state.vm_needs_start = false;
-     }
- }
- 
-@@ -491,9 +491,9 @@ void qmp_savevm_end(Error **errp)
-         return;
-     }
- 
--    if (snap_state.saved_vm_running) {
-+    if (snap_state.vm_needs_start) {
-         vm_start();
--        snap_state.saved_vm_running = false;
-+        snap_state.vm_needs_start = false;
-     }
- 
-     qemu_coroutine_enter(wait_for_close);

debian/patches/pve/0060-savevm-async-improve-runstate-preservation-cleanup-e.patch

@@ -1,120 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Fiona Ebner <f.ebner@proxmox.com>
-Date: Mon, 31 Mar 2025 16:55:04 +0200
-Subject: [PATCH] savevm-async: improve runstate preservation, cleanup error
- handling
-
-Determine if VM needs to be started after finishing right before
-actually stopping the VM instead of at the beginning.
-
-In qmp_savevm_start(), the only path stopping the VM returns right
-aftwards, so there is no need for the vm_start() handling after
-errors.
-
-Lastly, improve the code style for checking whether migrate_init()
-failed by explicitly comparing against 0.
-
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
-[WB: squashed error handling commits, rename goto branch instead of
-     inlining it]
-Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
----
- migration/savevm-async.c | 25 ++++++++++---------------
- 1 file changed, 10 insertions(+), 15 deletions(-)
-
-diff --git a/migration/savevm-async.c b/migration/savevm-async.c
-index 1e34c31e8b..d8d2c80475 100644
---- a/migration/savevm-async.c
-+++ b/migration/savevm-async.c
-@@ -178,6 +178,7 @@ static void process_savevm_finalize(void *opaque)
-      */
-     blk_set_aio_context(snap_state.target, qemu_get_aio_context(), NULL);
- 
-+    snap_state.vm_needs_start = runstate_is_running();
-     ret = vm_stop_force_state(RUN_STATE_FINISH_MIGRATE);
-     if (ret < 0) {
-         save_snapshot_error("vm_stop_force_state error %d", ret);
-@@ -352,7 +353,6 @@ void qmp_savevm_start(const char *statefile, Error **errp)
-     }
- 
-     /* initialize snapshot info */
--    snap_state.vm_needs_start = runstate_is_running();
-     snap_state.bs_pos = 0;
-     snap_state.total_time = qemu_clock_get_ms(QEMU_CLOCK_REALTIME);
-     snap_state.blocker = NULL;
-@@ -364,13 +364,14 @@ void qmp_savevm_start(const char *statefile, Error **errp)
-     }
- 
-     if (!statefile) {
-+        snap_state.vm_needs_start = runstate_is_running();
-         vm_stop(RUN_STATE_SAVE_VM);
-         snap_state.state = SAVE_STATE_COMPLETED;
-         return;
-     }
- 
-     if (qemu_savevm_state_blocked(errp)) {
--        return;
-+        goto fail;
-     }
- 
-     /* Open the image */
-@@ -380,12 +381,12 @@ void qmp_savevm_start(const char *statefile, Error **errp)
-     snap_state.target = blk_new_open(statefile, NULL, options, bdrv_oflags, &local_err);
-     if (!snap_state.target) {
-         error_setg(errp, "failed to open '%s'", statefile);
--        goto restart;
-+        goto fail;
-     }
-     target_bs = blk_bs(snap_state.target);
-     if (!target_bs) {
-         error_setg(errp, "failed to open '%s' - no block driver state", statefile);
--        goto restart;
-+        goto fail;
-     }
- 
-     QIOChannel *ioc = QIO_CHANNEL(qio_channel_savevm_async_new(snap_state.target,
-@@ -394,7 +395,7 @@ void qmp_savevm_start(const char *statefile, Error **errp)
- 
-     if (!snap_state.file) {
-         error_setg(errp, "failed to open '%s'", statefile);
--        goto restart;
-+        goto fail;
-     }
- 
-     /*
-@@ -402,8 +403,8 @@ void qmp_savevm_start(const char *statefile, Error **errp)
-      * State is cleared in process_savevm_co, but has to be initialized
-      * here (blocking main thread, from QMP) to avoid race conditions.
-      */
--    if (migrate_init(ms, errp)) {
--        return;
-+    if (migrate_init(ms, errp) != 0) {
-+        goto fail;
-     }
-     memset(&mig_stats, 0, sizeof(mig_stats));
-     ms->to_dst_file = snap_state.file;
-@@ -418,7 +419,7 @@ void qmp_savevm_start(const char *statefile, Error **errp)
-     if (ret != 0) {
-         error_setg_errno(errp, -ret, "savevm state setup failed: %s",
-                          local_err ? error_get_pretty(local_err) : "unknown error");
--        return;
-+        goto fail;
-     }
- 
-     /* Async processing from here on out happens in iohandler context, so let
-@@ -436,14 +437,8 @@ void qmp_savevm_start(const char *statefile, Error **errp)
- 
-     return;
- 
--restart:
--
-+fail:
-     save_snapshot_error("setup failed");
--
--    if (snap_state.vm_needs_start) {
--        vm_start();
--        snap_state.vm_needs_start = false;
--    }
- }
- 
- static void coroutine_fn wait_for_close_co(void *opaque)

debian/patches/pve/0061-savevm-async-use-dedicated-iothread-for-state-file.patch

@@ -1,185 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Fiona Ebner <f.ebner@proxmox.com>
-Date: Mon, 31 Mar 2025 16:55:06 +0200
-Subject: [PATCH] savevm-async: use dedicated iothread for state file
-
-Having the state file be in the iohandler context means that a
-blk_drain_all() call in the main thread or vCPU thread that happens
-while the snapshot is running will result in a deadlock.
-
-For example, the main thread might be stuck in:
-
-> 0  0x00007300ac9552d6 in __ppoll (fds=0x64bd5a411a50, nfds=2, timeout=<optimized out>, timeout@entry=0x0, sigmask=sigmask@entry=0x0) at ../sysdeps/unix/sysv/linux/ppoll.c:42
-> 1  0x000064bd51af3cad in ppoll (__ss=0x0, __timeout=0x0, __nfds=<optimized out>, __fds=<optimized out>) at /usr/include/x86_64-linux-gnu/bits/poll2.h:64
-> 2  0x000064bd51ad8799 in fdmon_poll_wait (ctx=0x64bd58d968a0, ready_list=0x7ffcfcc15558, timeout=-1) at ../util/fdmon-poll.c:79
-> 3  0x000064bd51ad7c3d in aio_poll (ctx=0x64bd58d968a0, blocking=blocking@entry=true) at ../util/aio-posix.c:671
-> 4  0x000064bd519a0b5d in bdrv_drain_all_begin () at ../block/io.c:531
-> 5  bdrv_drain_all_begin () at ../block/io.c:510
-> 6  0x000064bd519943c4 in blk_drain_all () at ../block/block-backend.c:2085
-> 7  0x000064bd5160fc5a in virtio_scsi_dataplane_stop (vdev=0x64bd5a215190) at ../hw/scsi/virtio-scsi-dataplane.c:213
-> 8  0x000064bd51664e90 in virtio_bus_stop_ioeventfd (bus=0x64bd5a215110) at ../hw/virtio/virtio-bus.c:259
-> 9  0x000064bd5166511b in virtio_bus_stop_ioeventfd (bus=<optimized out>) at ../hw/virtio/virtio-bus.c:251
-> 10 virtio_bus_reset (bus=<optimized out>) at ../hw/virtio/virtio-bus.c:107
-> 11 0x000064bd51667431 in virtio_pci_reset (qdev=<optimized out>) at ../hw/virtio/virtio-pci.c:2296
-...
-> 34 0x000064bd517aa951 in pc_machine_reset (machine=<optimized out>, type=<optimized out>) at ../hw/i386/pc.c:1722
-> 35 0x000064bd516aa4c4 in qemu_system_reset (reason=reason@entry=SHUTDOWN_CAUSE_GUEST_RESET) at ../system/runstate.c:525
-> 36 0x000064bd516aaeb9 in main_loop_should_exit (status=<synthetic pointer>) at ../system/runstate.c:801
-> 37 qemu_main_loop () at ../system/runstate.c:834
-
-which is in block/io.c:
-
-> /* Now poll the in-flight requests */
-> AIO_WAIT_WHILE_UNLOCKED(NULL, bdrv_drain_all_poll());
-
-The working theory is: The deadlock happens because the IO is issued
-from the process_savevm_co() coroutine, which doesn't get scheduled
-again to complete in-flight requests when the main thread is stuck
-there polling. The main thread itself is the one that would need to
-schedule it. In case of a vCPU triggering the VirtIO SCSI dataplane
-stop, which happens during (Linux) boot, the vCPU thread will hold the
-big QEMU lock (BQL) blocking the main thread from making progress
-scheduling the process_savevm_co() coroutine.
-
-This change should also help in general to reduce load on the main
-thread and for it to get stuck on IO, i.e. same benefits as using a
-dedicated IO thread for regular drives. This is particularly
-interesting when the VM state storage is a network storage like NFS.
-
-With some luck, it could also help with bug #6262 [0]. The failure
-there happens while issuing/right after the savevm-start QMP command,
-so the most likely coroutine is the process_savevm_co() that was
-previously scheduled to the iohandler context. Likely someone polls
-the iohandler context and wants to enter the already scheduled
-coroutine leading to the abort():
-> qemu_aio_coroutine_enter: Co-routine was already scheduled in 'aio_co_schedule'
-With a dedicated iothread, there hopefully is no such race.
-
-The comment above querying the pending bytes wrongly talked about the
-"iothread lock", but should've been "iohandler lock". This was even
-renamed to BQL (big QEMU lock) a few releases ago. Even if that was
-not a typo to begin with, there are no AioContext locks anymore.
-
-[0]: https://bugzilla.proxmox.com/show_bug.cgi?id=6262
-
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
-[WB: update to the changed error handling in the previous commit]
-Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
----
- migration/savevm-async.c | 42 ++++++++++++++++++++++++++++------------
- 1 file changed, 30 insertions(+), 12 deletions(-)
-
-diff --git a/migration/savevm-async.c b/migration/savevm-async.c
-index d8d2c80475..11ea4c601d 100644
---- a/migration/savevm-async.c
-+++ b/migration/savevm-async.c
-@@ -25,6 +25,7 @@
- #include "qemu/main-loop.h"
- #include "qemu/rcu.h"
- #include "qemu/yank.h"
-+#include "sysemu/iothread.h"
- 
- /* #define DEBUG_SAVEVM_STATE */
- 
-@@ -57,6 +58,7 @@ static struct SnapshotState {
-     QEMUBH *finalize_bh;
-     Coroutine *co;
-     QemuCoSleep target_close_wait;
-+    IOThread *iothread;
- } snap_state;
- 
- static bool savevm_aborted(void)
-@@ -256,16 +258,13 @@ static void coroutine_fn process_savevm_co(void *opaque)
-         uint64_t threshold = 400 * 1000;
- 
-         /*
--         * pending_{estimate,exact} are expected to be called without iothread
--         * lock. Similar to what is done in migration.c, call the exact variant
--         * only once pend_precopy in the estimate is below the threshold.
-+         * Similar to what is done in migration.c, call the exact variant only
-+         * once pend_precopy in the estimate is below the threshold.
-          */
--        bql_unlock();
-         qemu_savevm_state_pending_estimate(&pend_precopy, &pend_postcopy);
-         if (pend_precopy <= threshold) {
-             qemu_savevm_state_pending_exact(&pend_precopy, &pend_postcopy);
-         }
--        bql_lock();
-         pending_size = pend_precopy + pend_postcopy;
- 
-         /*
-@@ -332,11 +331,17 @@ static void coroutine_fn process_savevm_co(void *opaque)
-     qemu_bh_schedule(snap_state.finalize_bh);
- }
- 
-+static void savevm_cleanup_iothread(void) {
-+    if (snap_state.iothread) {
-+        iothread_destroy(snap_state.iothread);
-+        snap_state.iothread = NULL;
-+    }
-+}
-+
- void qmp_savevm_start(const char *statefile, Error **errp)
- {
-     Error *local_err = NULL;
-     MigrationState *ms = migrate_get_current();
--    AioContext *iohandler_ctx = iohandler_get_aio_context();
-     BlockDriverState *target_bs = NULL;
-     int ret = 0;
- 
-@@ -374,6 +379,19 @@ void qmp_savevm_start(const char *statefile, Error **errp)
-         goto fail;
-     }
- 
-+    if (snap_state.iothread) {
-+        /* This is not expected, so warn about it, but no point in re-creating a new iothread. */
-+        warn_report("iothread for snapshot already exists - re-using");
-+    } else {
-+        snap_state.iothread =
-+            iothread_create("__proxmox_savevm_async_iothread__", &local_err);
-+        if (!snap_state.iothread) {
-+            error_setg(errp, "creating iothread failed: %s",
-+                       local_err ? error_get_pretty(local_err) : "unknown error");
-+            goto fail;
-+        }
-+    }
-+
-     /* Open the image */
-     QDict *options = NULL;
-     options = qdict_new();
-@@ -422,22 +440,20 @@ void qmp_savevm_start(const char *statefile, Error **errp)
-         goto fail;
-     }
- 
--    /* Async processing from here on out happens in iohandler context, so let
--     * the target bdrv have its home there.
--     */
--    ret = blk_set_aio_context(snap_state.target, iohandler_ctx, &local_err);
-+    ret = blk_set_aio_context(snap_state.target, snap_state.iothread->ctx, &local_err);
-     if (ret != 0) {
--        warn_report("failed to set iohandler context for VM state target: %s %s",
-+        warn_report("failed to set iothread context for VM state target: %s %s",
-                     local_err ? error_get_pretty(local_err) : "unknown error",
-                     strerror(-ret));
-     }
- 
-     snap_state.co = qemu_coroutine_create(&process_savevm_co, NULL);
--    aio_co_schedule(iohandler_ctx, snap_state.co);
-+    aio_co_schedule(snap_state.iothread->ctx, snap_state.co);
- 
-     return;
- 
- fail:
-+    savevm_cleanup_iothread();
-     save_snapshot_error("setup failed");
- }
- 
-@@ -463,6 +479,8 @@ static void coroutine_fn wait_for_close_co(void *opaque)
-         DPRINTF("savevm-end: no target file open\n");
-     }
- 
-+    savevm_cleanup_iothread();
-+
-     // File closed and no other error, so ensure next snapshot can be started.
-     if (snap_state.state != SAVE_STATE_ERROR) {
-         snap_state.state = SAVE_STATE_DONE;

debian/patches/pve/0062-savevm-async-treat-failure-to-set-iothread-context-a.patch

@@ -1,33 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Fiona Ebner <f.ebner@proxmox.com>
-Date: Mon, 31 Mar 2025 16:55:07 +0200
-Subject: [PATCH] savevm-async: treat failure to set iothread context as a hard
- failure
-
-This is not expected to ever fail and there might be assumptions about
-having the expected context down the line.
-
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
-[WB: update to changed error handling]
-Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
----
- migration/savevm-async.c | 6 +++---
- 1 file changed, 3 insertions(+), 3 deletions(-)
-
-diff --git a/migration/savevm-async.c b/migration/savevm-async.c
-index 11ea4c601d..f2b10b5519 100644
---- a/migration/savevm-async.c
-+++ b/migration/savevm-async.c
-@@ -442,9 +442,9 @@ void qmp_savevm_start(const char *statefile, Error **errp)
- 
-     ret = blk_set_aio_context(snap_state.target, snap_state.iothread->ctx, &local_err);
-     if (ret != 0) {
--        warn_report("failed to set iothread context for VM state target: %s %s",
--                    local_err ? error_get_pretty(local_err) : "unknown error",
--                    strerror(-ret));
-+        error_setg_errno(errp, -ret, "failed to set iothread context for VM state target: %s",
-+                         local_err ? error_get_pretty(local_err) : "unknown error");
-+        goto fail;
-     }
- 
-     snap_state.co = qemu_coroutine_create(&process_savevm_co, NULL);

debian/patches/pve/0063-PVE-backup-clean-up-directly-in-setup_snapshot_acces.patch

@@ -1,41 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Fiona Ebner <f.ebner@proxmox.com>
-Date: Thu, 3 Apr 2025 14:30:41 +0200
-Subject: [PATCH] PVE backup: clean up directly in setup_snapshot_access() when
- it fails
-
-The only thing that might need to be cleaned up after
-setup_snapshot_access() failed is dropping the cbw filter. Do so in
-the single branch it matters inside setup_snapshot_access() itself.
-This avoids the need that callers of setup_snapshot_access() use
-cleanup_snapshot_access() when the call failed.
-
-Suggested-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
-Reviewed-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
----
- pve-backup.c | 4 +++-
- 1 file changed, 3 insertions(+), 1 deletion(-)
-
-diff --git a/pve-backup.c b/pve-backup.c
-index 32352fb5ec..2408f182bc 100644
---- a/pve-backup.c
-+++ b/pve-backup.c
-@@ -576,6 +576,9 @@ static int setup_snapshot_access(PVEBackupDevInfo *di, Error **errp)
-     di->fleecing.snapshot_access =
-         bdrv_open(NULL, NULL, snapshot_access_opts, BDRV_O_RDWR | BDRV_O_UNMAP, &local_err);
-     if (!di->fleecing.snapshot_access) {
-+        bdrv_cbw_drop(di->fleecing.cbw);
-+        di->fleecing.cbw = NULL;
-+
-         error_setg(errp, "setting up snapshot access for fleecing failed: %s",
-                    local_err ? error_get_pretty(local_err) : "unknown error");
-         return -1;
-@@ -629,7 +632,6 @@ static void create_backup_jobs_bh(void *opaque) {
-                 error_setg(errp, "%s - setting up snapshot access for fleecing failed: %s",
-                            di->device_name,
-                            local_err ? error_get_pretty(local_err) : "unknown error");
--                cleanup_snapshot_access(di);
-                 bdrv_drained_end(di->bs);
-                 break;
-             }

debian/patches/pve/0064-PVE-backup-factor-out-helper-to-clear-backup-state-s.patch

@@ -1,59 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Fiona Ebner <f.ebner@proxmox.com>
-Date: Thu, 3 Apr 2025 14:30:42 +0200
-Subject: [PATCH] PVE backup: factor out helper to clear backup state's bitmap
- list
-
-Suggested-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
-Reviewed-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
----
- pve-backup.c | 28 ++++++++++++++++++----------
- 1 file changed, 18 insertions(+), 10 deletions(-)
-
-diff --git a/pve-backup.c b/pve-backup.c
-index 2408f182bc..915649b5f9 100644
---- a/pve-backup.c
-+++ b/pve-backup.c
-@@ -811,6 +811,23 @@ err:
-     return di_list;
- }
- 
-+/*
-+ * To be called with the backup_state.stat mutex held.
-+ */
-+static void clear_backup_state_bitmap_list(void) {
-+
-+    if (backup_state.stat.bitmap_list) {
-+        GList *bl = backup_state.stat.bitmap_list;
-+        while (bl) {
-+            g_free(((PBSBitmapInfo *)bl->data)->drive);
-+            g_free(bl->data);
-+            bl = g_list_next(bl);
-+        }
-+        g_list_free(backup_state.stat.bitmap_list);
-+        backup_state.stat.bitmap_list = NULL;
-+    }
-+}
-+
- UuidInfo coroutine_fn *qmp_backup(
-     const char *backup_file,
-     const char *password,
-@@ -898,16 +915,7 @@ UuidInfo coroutine_fn *qmp_backup(
-     backup_state.stat.reused = 0;
- 
-     /* clear previous backup's bitmap_list */
--    if (backup_state.stat.bitmap_list) {
--        GList *bl = backup_state.stat.bitmap_list;
--        while (bl) {
--            g_free(((PBSBitmapInfo *)bl->data)->drive);
--            g_free(bl->data);
--            bl = g_list_next(bl);
--        }
--        g_list_free(backup_state.stat.bitmap_list);
--        backup_state.stat.bitmap_list = NULL;
--    }
-+    clear_backup_state_bitmap_list();
- 
-     if (format == BACKUP_FORMAT_PBS) {
-         if (!password) {

debian/patches/pve/0065-PVE-backup-factor-out-helper-to-initialize-backup-st.patch

@@ -1,95 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Fiona Ebner <f.ebner@proxmox.com>
-Date: Thu, 3 Apr 2025 14:30:43 +0200
-Subject: [PATCH] PVE backup: factor out helper to initialize backup state stat
- struct
-
-Suggested-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
-Reviewed-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
----
- pve-backup.c | 62 ++++++++++++++++++++++++++++++++--------------------
- 1 file changed, 38 insertions(+), 24 deletions(-)
-
-diff --git a/pve-backup.c b/pve-backup.c
-index 915649b5f9..88a981f81c 100644
---- a/pve-backup.c
-+++ b/pve-backup.c
-@@ -828,6 +828,43 @@ static void clear_backup_state_bitmap_list(void) {
-     }
- }
- 
-+/*
-+ * Initializes most of the backup state 'stat' struct. Note that 'reused' and
-+ * 'bitmap_list' are not changed by this function and need to be handled by
-+ * the caller. In particular, 'reused' needs to be set before calling this
-+ * function.
-+ *
-+ * To be called with the backup_state.stat mutex held.
-+ */
-+static void initialize_backup_state_stat(
-+    const char *backup_file,
-+    uuid_t uuid,
-+    size_t total)
-+{
-+    if (backup_state.stat.error) {
-+        error_free(backup_state.stat.error);
-+        backup_state.stat.error = NULL;
-+    }
-+
-+    backup_state.stat.start_time = time(NULL);
-+    backup_state.stat.end_time = 0;
-+
-+    if (backup_state.stat.backup_file) {
-+        g_free(backup_state.stat.backup_file);
-+    }
-+    backup_state.stat.backup_file = g_strdup(backup_file);
-+
-+    uuid_copy(backup_state.stat.uuid, uuid);
-+    uuid_unparse_lower(uuid, backup_state.stat.uuid_str);
-+
-+    backup_state.stat.total = total;
-+    backup_state.stat.dirty = total - backup_state.stat.reused;
-+    backup_state.stat.transferred = 0;
-+    backup_state.stat.zero_bytes = 0;
-+    backup_state.stat.finishing = false;
-+    backup_state.stat.starting = true;
-+}
-+
- UuidInfo coroutine_fn *qmp_backup(
-     const char *backup_file,
-     const char *password,
-@@ -1070,32 +1107,9 @@ UuidInfo coroutine_fn *qmp_backup(
-         }
-     }
-     /* initialize global backup_state now */
--    /* note: 'reused' and 'bitmap_list' are initialized earlier */
--
--    if (backup_state.stat.error) {
--        error_free(backup_state.stat.error);
--        backup_state.stat.error = NULL;
--    }
--
--    backup_state.stat.start_time = time(NULL);
--    backup_state.stat.end_time = 0;
--
--    if (backup_state.stat.backup_file) {
--        g_free(backup_state.stat.backup_file);
--    }
--    backup_state.stat.backup_file = g_strdup(backup_file);
--
--    uuid_copy(backup_state.stat.uuid, uuid);
--    uuid_unparse_lower(uuid, backup_state.stat.uuid_str);
-+    initialize_backup_state_stat(backup_file, uuid, total);
-     char *uuid_str = g_strdup(backup_state.stat.uuid_str);
- 
--    backup_state.stat.total = total;
--    backup_state.stat.dirty = total - backup_state.stat.reused;
--    backup_state.stat.transferred = 0;
--    backup_state.stat.zero_bytes = 0;
--    backup_state.stat.finishing = false;
--    backup_state.stat.starting = true;
--
-     qemu_mutex_unlock(&backup_state.stat.lock);
- 
-     backup_state.speed = (has_speed && speed > 0) ? speed : 0;

debian/patches/pve/0066-PVE-backup-add-target-ID-in-backup-state.patch

@@ -1,63 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Fiona Ebner <f.ebner@proxmox.com>
-Date: Thu, 3 Apr 2025 14:30:44 +0200
-Subject: [PATCH] PVE backup: add target ID in backup state
-
-In preparation for allowing multiple backup providers and potentially
-multiple targets for a given provider. Each backup target can then
-have its own dirty bitmap and there can be additional checks that the
-current backup state is actually associated to the expected target.
-
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
-Reviewed-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
----
- pve-backup.c | 15 ++++++++++++++-
- 1 file changed, 14 insertions(+), 1 deletion(-)
-
-diff --git a/pve-backup.c b/pve-backup.c
-index 88a981f81c..8789a0667a 100644
---- a/pve-backup.c
-+++ b/pve-backup.c
-@@ -70,6 +70,7 @@ static struct PVEBackupState {
-     JobTxn *txn;
-     CoMutex backup_mutex;
-     CoMutex dump_callback_mutex;
-+    char *target_id;
- } backup_state;
- 
- static void pvebackup_init(void)
-@@ -865,6 +866,16 @@ static void initialize_backup_state_stat(
-     backup_state.stat.starting = true;
- }
- 
-+/*
-+ * To be called with the backup_state mutex held.
-+ */
-+static void backup_state_set_target_id(const char *target_id) {
-+    if (backup_state.target_id) {
-+        g_free(backup_state.target_id);
-+    }
-+    backup_state.target_id = g_strdup(target_id);
-+}
-+
- UuidInfo coroutine_fn *qmp_backup(
-     const char *backup_file,
-     const char *password,
-@@ -904,7 +915,7 @@ UuidInfo coroutine_fn *qmp_backup(
- 
-     if (backup_state.di_list) {
-         error_set(errp, ERROR_CLASS_GENERIC_ERROR,
--                  "previous backup not finished");
-+                  "previous backup for target '%s' not finished", backup_state.target_id);
-         qemu_co_mutex_unlock(&backup_state.backup_mutex);
-         return NULL;
-     }
-@@ -1122,6 +1133,8 @@ UuidInfo coroutine_fn *qmp_backup(
-     backup_state.vmaw = vmaw;
-     backup_state.pbs = pbs;
- 
-+    backup_state_set_target_id("Proxmox");
-+
-     backup_state.di_list = di_list;
- 
-     uuid_info = g_malloc0(sizeof(*uuid_info));

debian/patches/pve/0067-PVE-backup-get-device-info-allow-caller-to-specify-f.patch

@@ -1,57 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Fiona Ebner <f.ebner@proxmox.com>
-Date: Thu, 3 Apr 2025 14:30:45 +0200
-Subject: [PATCH] PVE backup: get device info: allow caller to specify filter
- for which devices use fleecing
-
-For providing snapshot-access to external backup providers, EFI and
-TPM also need an associated fleecing image. The new caller will thus
-need a different filter.
-
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
-Reviewed-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
----
- pve-backup.c | 9 +++++----
- 1 file changed, 5 insertions(+), 4 deletions(-)
-
-diff --git a/pve-backup.c b/pve-backup.c
-index 8789a0667a..755f1abcf1 100644
---- a/pve-backup.c
-+++ b/pve-backup.c
-@@ -719,7 +719,7 @@ static void create_backup_jobs_bh(void *opaque) {
- /*
-  * EFI disk and TPM state are small and it's just not worth setting up fleecing for them.
-  */
--static bool device_uses_fleecing(const char *device_id)
-+static bool fleecing_no_efi_tpm(const char *device_id)
- {
-     return strncmp(device_id, "drive-efidisk", 13) && strncmp(device_id, "drive-tpmstate", 14);
- }
-@@ -731,7 +731,7 @@ static bool device_uses_fleecing(const char *device_id)
-  */
- static GList coroutine_fn GRAPH_RDLOCK *get_device_info(
-     const char *devlist,
--    bool fleecing,
-+    bool (*device_uses_fleecing)(const char*),
-     Error **errp)
- {
-     gchar **devs = NULL;
-@@ -757,7 +757,7 @@ static GList coroutine_fn GRAPH_RDLOCK *get_device_info(
-             di->bs = bs;
-             di->device_name = g_strdup(bdrv_get_device_name(bs));
- 
--            if (fleecing && device_uses_fleecing(*d)) {
-+            if (device_uses_fleecing && device_uses_fleecing(*d)) {
-                 g_autofree gchar *fleecing_devid = g_strconcat(*d, "-fleecing", NULL);
-                 BlockBackend *fleecing_blk = blk_by_name(fleecing_devid);
-                 if (!fleecing_blk) {
-@@ -924,7 +924,8 @@ UuidInfo coroutine_fn *qmp_backup(
-     format = has_format ? format : BACKUP_FORMAT_VMA;
- 
-     bdrv_graph_co_rdlock();
--    di_list = get_device_info(devlist, has_fleecing && fleecing, &local_err);
-+    di_list = get_device_info(devlist, (has_fleecing && fleecing) ? fleecing_no_efi_tpm : NULL,
-+                              &local_err);
-     bdrv_graph_co_rdunlock();
-     if (local_err) {
-         error_propagate(errp, local_err);

debian/patches/pve/0068-PVE-backup-implement-backup-access-setup-and-teardow.patch

@@ -1,495 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Fiona Ebner <f.ebner@proxmox.com>
-Date: Thu, 3 Apr 2025 14:30:46 +0200
-Subject: [PATCH] PVE backup: implement backup access setup and teardown API
- for external providers
-
-For external backup providers, the state of the VM's disk images at
-the time the backup is started is preserved via a snapshot-access
-block node. Old data is moved to the fleecing image when new guest
-writes come in. The snapshot-access block node, as well as the
-associated bitmap in case of incremental backup, will be exported via
-NBD to the external provider. The NBD export will be done by the
-management layer, the missing functionality is setting up and tearing
-down the snapshot-access block nodes, which this patch adds.
-
-It is necessary to also set up fleecing for EFI and TPM disks, so that
-old data can be moved out of the way when a new guest write comes in.
-
-There can only be one regular backup or one active backup access at
-a time, because both require replacing the original block node of the
-drive. Thus the backup state is re-used, and checks are added to
-prohibit regular backup while snapshot access is active and vice
-versa.
-
-The block nodes added by the backup-access-setup QMP call are not
-tracked anywhere else (there is no job they are associated to like for
-regular backup). This requires adding a callback for teardown when
-QEMU exits, i.e. in qemu_cleanup(). Otherwise, there will be an
-assertion failure that the block graph is not empty when QEMU exits
-before the backup-access-teardown QMP command is called.
-
-The code for the qmp_backup_access_setup() was based on the existing
-qmp_backup() routine.
-
-The return value for the setup QMP command contains information about
-the snapshot-access block nodes that can be used by the management
-layer to set up the NBD exports.
-
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
-Reviewed-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
----
- pve-backup.c         | 264 ++++++++++++++++++++++++++++++++++++++++++-
- pve-backup.h         |  16 +++
- qapi/block-core.json |  49 ++++++++
- system/runstate.c    |   6 +
- 4 files changed, 329 insertions(+), 6 deletions(-)
- create mode 100644 pve-backup.h
-
-diff --git a/pve-backup.c b/pve-backup.c
-index 755f1abcf1..091b5bd231 100644
---- a/pve-backup.c
-+++ b/pve-backup.c
-@@ -1,4 +1,5 @@
- #include "proxmox-backup-client.h"
-+#include "pve-backup.h"
- #include "vma.h"
- 
- #include "qemu/osdep.h"
-@@ -588,6 +589,36 @@ static int setup_snapshot_access(PVEBackupDevInfo *di, Error **errp)
-     return 0;
- }
- 
-+static void setup_all_snapshot_access_bh(void *opaque)
-+{
-+    assert(!qemu_in_coroutine());
-+
-+    CoCtxData *data = (CoCtxData*)opaque;
-+    Error **errp = (Error**)data->data;
-+
-+    Error *local_err = NULL;
-+
-+    GList *l =  backup_state.di_list;
-+    while (l) {
-+        PVEBackupDevInfo *di = (PVEBackupDevInfo *)l->data;
-+        l = g_list_next(l);
-+
-+        bdrv_drained_begin(di->bs);
-+
-+        if (setup_snapshot_access(di, &local_err) < 0) {
-+            bdrv_drained_end(di->bs);
-+            error_setg(errp, "%s - setting up snapshot access failed: %s", di->device_name,
-+                       local_err ? error_get_pretty(local_err) : "unknown error");
-+            break;
-+        }
-+
-+        bdrv_drained_end(di->bs);
-+    }
-+
-+    /* return */
-+    aio_co_enter(data->ctx, data->co);
-+}
-+
- /*
-  * backup_job_create can *not* be run from a coroutine, so this can't either.
-  * The caller is responsible that backup_mutex is held nonetheless.
-@@ -724,6 +755,11 @@ static bool fleecing_no_efi_tpm(const char *device_id)
-     return strncmp(device_id, "drive-efidisk", 13) && strncmp(device_id, "drive-tpmstate", 14);
- }
- 
-+static bool fleecing_all(const char *device_id)
-+{
-+    return true;
-+}
-+
- /*
-  * Returns a list of device infos, which needs to be freed by the caller. In
-  * case of an error, errp will be set, but the returned value might still be a
-@@ -839,8 +875,9 @@ static void clear_backup_state_bitmap_list(void) {
-  */
- static void initialize_backup_state_stat(
-     const char *backup_file,
--    uuid_t uuid,
--    size_t total)
-+    uuid_t *uuid,
-+    size_t total,
-+    bool starting)
- {
-     if (backup_state.stat.error) {
-         error_free(backup_state.stat.error);
-@@ -855,15 +892,19 @@ static void initialize_backup_state_stat(
-     }
-     backup_state.stat.backup_file = g_strdup(backup_file);
- 
--    uuid_copy(backup_state.stat.uuid, uuid);
--    uuid_unparse_lower(uuid, backup_state.stat.uuid_str);
-+    if (uuid) {
-+        uuid_copy(backup_state.stat.uuid, *uuid);
-+        uuid_unparse_lower(*uuid, backup_state.stat.uuid_str);
-+    } else {
-+        backup_state.stat.uuid_str[0] = '\0';
-+    }
- 
-     backup_state.stat.total = total;
-     backup_state.stat.dirty = total - backup_state.stat.reused;
-     backup_state.stat.transferred = 0;
-     backup_state.stat.zero_bytes = 0;
-     backup_state.stat.finishing = false;
--    backup_state.stat.starting = true;
-+    backup_state.stat.starting = starting;
- }
- 
- /*
-@@ -876,6 +917,216 @@ static void backup_state_set_target_id(const char *target_id) {
-     backup_state.target_id = g_strdup(target_id);
- }
- 
-+BackupAccessInfoList *coroutine_fn qmp_backup_access_setup(
-+    const char *target_id,
-+    const char *devlist,
-+    Error **errp)
-+{
-+    assert(qemu_in_coroutine());
-+
-+    qemu_co_mutex_lock(&backup_state.backup_mutex);
-+
-+    Error *local_err = NULL;
-+    GList *di_list = NULL;
-+    GList *l;
-+
-+    if (backup_state.di_list) {
-+        error_set(errp, ERROR_CLASS_GENERIC_ERROR,
-+                  "previous backup for target '%s' not finished", backup_state.target_id);
-+        qemu_co_mutex_unlock(&backup_state.backup_mutex);
-+        return NULL;
-+    }
-+
-+    bdrv_graph_co_rdlock();
-+    di_list = get_device_info(devlist, fleecing_all, &local_err);
-+    bdrv_graph_co_rdunlock();
-+    if (local_err) {
-+        error_propagate(errp, local_err);
-+        goto err;
-+    }
-+    assert(di_list);
-+
-+    size_t total = 0;
-+
-+    l = di_list;
-+    while (l) {
-+        PVEBackupDevInfo *di = (PVEBackupDevInfo *)l->data;
-+        l = g_list_next(l);
-+
-+        ssize_t size = bdrv_getlength(di->bs);
-+        if (size < 0) {
-+            error_setg_errno(errp, -size, "bdrv_getlength failed");
-+            goto err;
-+        }
-+        di->size = size;
-+        total += size;
-+
-+        di->completed_ret = INT_MAX;
-+    }
-+
-+    qemu_mutex_lock(&backup_state.stat.lock);
-+    backup_state.stat.reused = 0;
-+
-+    /* clear previous backup's bitmap_list */
-+    clear_backup_state_bitmap_list();
-+
-+    /* starting=false, because there is no associated QEMU job */
-+    initialize_backup_state_stat(NULL, NULL, total, false);
-+
-+    qemu_mutex_unlock(&backup_state.stat.lock);
-+
-+    backup_state_set_target_id(target_id);
-+
-+    backup_state.vmaw = NULL;
-+    backup_state.pbs = NULL;
-+
-+    backup_state.di_list = di_list;
-+
-+    /* Run setup_all_snapshot_access_bh outside of coroutine (in BH) but keep
-+    * backup_mutex locked. This is fine, a CoMutex can be held across yield
-+    * points, and we'll release it as soon as the BH reschedules us.
-+    */
-+    CoCtxData waker = {
-+        .co = qemu_coroutine_self(),
-+        .ctx = qemu_get_current_aio_context(),
-+        .data = &local_err,
-+    };
-+    aio_bh_schedule_oneshot(waker.ctx, setup_all_snapshot_access_bh, &waker);
-+    qemu_coroutine_yield();
-+
-+    if (local_err) {
-+        error_propagate(errp, local_err);
-+        goto err;
-+    }
-+
-+    qemu_co_mutex_unlock(&backup_state.backup_mutex);
-+
-+    BackupAccessInfoList *bai_head = NULL, **p_bai_next = &bai_head;
-+
-+    l = di_list;
-+    while (l) {
-+        PVEBackupDevInfo *di = (PVEBackupDevInfo *)l->data;
-+        l = g_list_next(l);
-+
-+        BackupAccessInfoList *info = g_malloc0(sizeof(*info));
-+        info->value = g_malloc0(sizeof(*info->value));
-+        info->value->node_name = g_strdup(bdrv_get_node_name(di->fleecing.snapshot_access));
-+        info->value->device = g_strdup(di->device_name);
-+        info->value->size = di->size;
-+
-+        *p_bai_next = info;
-+        p_bai_next = &info->next;
-+    }
-+
-+    return bai_head;
-+
-+err:
-+
-+    l = di_list;
-+    while (l) {
-+        PVEBackupDevInfo *di = (PVEBackupDevInfo *)l->data;
-+        l = g_list_next(l);
-+
-+        g_free(di->device_name);
-+        di->device_name = NULL;
-+
-+        g_free(di);
-+    }
-+    g_list_free(di_list);
-+    backup_state.di_list = NULL;
-+
-+    qemu_co_mutex_unlock(&backup_state.backup_mutex);
-+    return NULL;
-+}
-+
-+/*
-+ * Caller needs to hold the backup mutex or the BQL.
-+ */
-+void backup_access_teardown(void)
-+{
-+    GList *l = backup_state.di_list;
-+
-+    qemu_mutex_lock(&backup_state.stat.lock);
-+    backup_state.stat.finishing = true;
-+    qemu_mutex_unlock(&backup_state.stat.lock);
-+
-+    while (l) {
-+        PVEBackupDevInfo *di = (PVEBackupDevInfo *)l->data;
-+        l = g_list_next(l);
-+
-+        if (di->fleecing.snapshot_access) {
-+            bdrv_unref(di->fleecing.snapshot_access);
-+            di->fleecing.snapshot_access = NULL;
-+        }
-+        if (di->fleecing.cbw) {
-+            bdrv_cbw_drop(di->fleecing.cbw);
-+            di->fleecing.cbw = NULL;
-+        }
-+
-+        g_free(di->device_name);
-+        di->device_name = NULL;
-+
-+        g_free(di);
-+    }
-+    g_list_free(backup_state.di_list);
-+    backup_state.di_list = NULL;
-+
-+    qemu_mutex_lock(&backup_state.stat.lock);
-+    backup_state.stat.end_time = time(NULL);
-+    backup_state.stat.finishing = false;
-+    qemu_mutex_unlock(&backup_state.stat.lock);
-+}
-+
-+// Not done in a coroutine, because bdrv_co_unref() and cbw_drop() would just spawn BHs anyways.
-+// Caller needs to hold the backup_state.backup_mutex lock
-+static void backup_access_teardown_bh(void *opaque)
-+{
-+    CoCtxData *data = (CoCtxData*)opaque;
-+
-+    backup_access_teardown();
-+
-+    /* return */
-+    aio_co_enter(data->ctx, data->co);
-+}
-+
-+void coroutine_fn qmp_backup_access_teardown(const char *target_id, Error **errp)
-+{
-+    assert(qemu_in_coroutine());
-+
-+    qemu_co_mutex_lock(&backup_state.backup_mutex);
-+
-+    if (!backup_state.target_id) { // nothing to do
-+        qemu_co_mutex_unlock(&backup_state.backup_mutex);
-+        return;
-+    }
-+
-+    /*
-+     * Continue with target_id == NULL, used by the callback registered for qemu_cleanup()
-+     */
-+    if (target_id && strcmp(backup_state.target_id, target_id)) {
-+        error_setg(errp, "cannot teardown backup access - got target %s instead of %s",
-+                   target_id, backup_state.target_id);
-+        qemu_co_mutex_unlock(&backup_state.backup_mutex);
-+        return;
-+    }
-+
-+    if (!strcmp(backup_state.target_id, "Proxmox VE")) {
-+        error_setg(errp, "cannot teardown backup access for PVE - use backup-cancel instead");
-+        qemu_co_mutex_unlock(&backup_state.backup_mutex);
-+        return;
-+    }
-+
-+    CoCtxData waker = {
-+        .co = qemu_coroutine_self(),
-+        .ctx = qemu_get_current_aio_context(),
-+    };
-+    aio_bh_schedule_oneshot(waker.ctx, backup_access_teardown_bh, &waker);
-+    qemu_coroutine_yield();
-+
-+    qemu_co_mutex_unlock(&backup_state.backup_mutex);
-+    return;
-+}
-+
- UuidInfo coroutine_fn *qmp_backup(
-     const char *backup_file,
-     const char *password,
-@@ -1119,7 +1370,7 @@ UuidInfo coroutine_fn *qmp_backup(
-         }
-     }
-     /* initialize global backup_state now */
--    initialize_backup_state_stat(backup_file, uuid, total);
-+    initialize_backup_state_stat(backup_file, &uuid, total, true);
-     char *uuid_str = g_strdup(backup_state.stat.uuid_str);
- 
-     qemu_mutex_unlock(&backup_state.stat.lock);
-@@ -1298,5 +1549,6 @@ ProxmoxSupportStatus *qmp_query_proxmox_support(Error **errp)
-     ret->pbs_masterkey = true;
-     ret->backup_max_workers = true;
-     ret->backup_fleecing = true;
-+    ret->backup_access_api = true;
-     return ret;
- }
-diff --git a/pve-backup.h b/pve-backup.h
-new file mode 100644
-index 0000000000..4033bc848f
---- /dev/null
-+++ b/pve-backup.h
-@@ -0,0 +1,16 @@
-+/*
-+ * Bacup code used by Proxmox VE
-+ *
-+ * Copyright (C) Proxmox Server Solutions
-+ *
-+ * This work is licensed under the terms of the GNU GPL, version 2 or later.
-+ * See the COPYING file in the top-level directory.
-+ *
-+ */
-+
-+#ifndef PVE_BACKUP_H
-+#define PVE_BACKUP_H
-+
-+void backup_access_teardown(void);
-+
-+#endif /* PVE_BACKUP_H */
-diff --git a/qapi/block-core.json b/qapi/block-core.json
-index c581f1f238..3f092221ce 100644
---- a/qapi/block-core.json
-+++ b/qapi/block-core.json
-@@ -1019,6 +1019,9 @@
- #
- # @pbs-library-version: Running version of libproxmox-backup-qemu0 library.
- #
-+# @backup-access-api: Whether backup access API for external providers is
-+#     supported or not.
-+#
- # @backup-fleecing: Whether backup fleecing is supported or not.
- #
- # @backup-max-workers: Whether the 'max-workers' @BackupPerf setting is
-@@ -1032,6 +1035,7 @@
-             'pbs-dirty-bitmap-migration': 'bool',
-             'pbs-masterkey': 'bool',
-             'pbs-library-version': 'str',
-+            'backup-access-api': 'bool',
-             'backup-fleecing': 'bool',
-             'backup-max-workers': 'bool' } }
- 
-@@ -1098,6 +1102,51 @@
- ##
- { 'command': 'query-pbs-bitmap-info', 'returns': ['PBSBitmapInfo'] }
- 
-+##
-+# @BackupAccessInfo:
-+#
-+# Info associated to a snapshot access for backup.  For more information about
-+# the bitmap see @BackupAccessBitmapMode.
-+#
-+# @node-name: the block node name of the snapshot-access node.
-+#
-+# @device: the device on top of which the snapshot access was created.
-+#
-+# @size: the size of the block device in bytes.
-+#
-+##
-+{ 'struct': 'BackupAccessInfo',
-+  'data': { 'node-name': 'str', 'device': 'str', 'size': 'size' } }
-+
-+##
-+# @backup-access-setup:
-+#
-+# Set up snapshot access to VM drives for an external backup provider.  No other
-+# backup or backup access can be done before tearing down the backup access.
-+#
-+# @target-id: the unique ID of the backup target.
-+#
-+# @devlist: list of block device names (separated by ',', ';' or ':'). By
-+#     default the backup includes all writable block devices.
-+#
-+# Returns: a list of @BackupAccessInfo, one for each device.
-+#
-+##
-+{ 'command': 'backup-access-setup',
-+  'data': { 'target-id': 'str', '*devlist': 'str' },
-+  'returns': [ 'BackupAccessInfo' ], 'coroutine': true }
-+
-+##
-+# @backup-access-teardown:
-+#
-+# Tear down previously setup snapshot access for the same target.
-+#
-+# @target-id: the ID of the backup target.
-+#
-+##
-+{ 'command': 'backup-access-teardown', 'data': { 'target-id': 'str' },
-+  'coroutine': true }
-+
- ##
- # @BlockDeviceTimedStats:
- #
-diff --git a/system/runstate.c b/system/runstate.c
-index c2c9afa905..6f93d7c2fb 100644
---- a/system/runstate.c
-+++ b/system/runstate.c
-@@ -60,6 +60,7 @@
- #include "sysemu/sysemu.h"
- #include "sysemu/tpm.h"
- #include "trace.h"
-+#include "pve-backup.h"
- 
- static NotifierList exit_notifiers =
-     NOTIFIER_LIST_INITIALIZER(exit_notifiers);
-@@ -920,6 +921,11 @@ void qemu_cleanup(int status)
-      * requests happening from here on anyway.
-      */
-     bdrv_drain_all_begin();
-+    /*
-+     * The backup access is set up by a QMP command, but is neither owned by a monitor nor
-+     * associated to a BlockBackend. Need to tear it down manually here.
-+     */
-+    backup_access_teardown();
-     job_cancel_sync_all();
-     bdrv_close_all();
-