bump version to 10.0.2-4

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>

Makefile

@@ -24,6 +24,7 @@ endif
 
 PC_BIOS_FW_PURGE_LIST_IN = \
 	hppa-firmware.img \
+	hppa-firmware64.img \
 	openbios-ppc \
 	openbios-sparc32 \
 	openbios-sparc64 \
@@ -31,7 +32,8 @@ PC_BIOS_FW_PURGE_LIST_IN = \
 	s390-ccw.img \
 	s390-netboot.img \
 	u-boot.e500 \
-	.*\.dtb \
+	.*[a-zA-Z0-9]\.dtb \
+	.*[a-zA-Z0-9]\.dts \
 	qemu_vga.ndrv \
 	slof.bin \
 	opensbi-riscv.*-generic-fw_dynamic.bin \
@@ -54,9 +56,8 @@ $(BUILDDIR): submodule
 
 .PHONY: deb kvm
 deb kvm: $(DEBS)
-$(DEB_DBG): $(DEB)
-$(DEB): $(BUILDDIR)
-	cd $(BUILDDIR); dpkg-buildpackage -b -us -uc -j
+$(DEBS) &: $(BUILDDIR)
+	cd $(BUILDDIR); dpkg-buildpackage -b -us -uc
 	lintian $(DEBS)
 
 sbuild: $(DSC)

debian/changelog

@@ -1,3 +1,221 @@
+pve-qemu-kvm (10.0.2-4) trixie; urgency=medium
+
+  * savevm-async: reuse migration blocker check for snapshots/hibernation to
+    avoid crashing a VM when on these actions if its configuration does not
+    support them.
+
+ -- Proxmox Support Team <support@proxmox.com>  Thu, 17 Jul 2025 23:08:46 +0200
+
+pve-qemu-kvm (10.0.2-3) trixie; urgency=medium
+
+  * add backup/zeroinit/track-alloc blockdev patches
+
+ -- Proxmox Support Team <support@proxmox.com>  Thu, 03 Jul 2025 12:02:03 +0200
+
+pve-qemu-kvm (10.0.2-2) trixie; urgency=medium
+
+  * drop support for accessing Gluster based storage directly due to its
+    effective end of support and maintenance. The last upstream release
+    happened over 2.5 years ago and there's currently no one providing
+    enterprise support or security updates. Further, upstream QEMU will remove
+    the integration in one of the next releases, so use the upcomming PVE 9
+    major release to provide a clean cut.
+    User can either stay on Proxmox VE 8 until its end-of-life (probably end
+    of June 2026), or mount GlusterFS "manually" (e.g., /etc/fstab) and add it
+    as directory storage to Proxmox VE.
+    We recommend moving to other actively maintained storage technology
+    altogether though.
+
+ -- Proxmox Support Team <support@proxmox.com>  Wed, 25 Jun 2025 19:23:30 +0200
+
+pve-qemu-kvm (10.0.2-1) trixie; urgency=medium
+
+  * update QEMU and downstream patches for 10.0.2 release.
+
+ -- Proxmox Support Team <support@proxmox.com>  Tue, 17 Jun 2025 08:38:51 +0200
+
+pve-qemu-kvm (9.2.0-5) bookworm; urgency=medium
+
+  * pve backup: backup-access api: simplify bitmap logic
+
+ -- 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

debian/control

@@ -12,7 +12,6 @@ Build-Depends: debhelper-compat (= 13),
                libepoxy-dev,
                libfdt-dev,
                libgbm-dev,
-               libglusterfs-dev (>= 5.2-2),
                libgnutls28-dev,
                libiscsi-dev (>= 1.12.0),
                libjpeg-dev,
@@ -47,13 +46,6 @@ Package: pve-qemu-kvm
 Architecture: any
 Depends: ceph-common (>= 0.48),
          iproute2,
-         libgfapi0 | glusterfs-common (>= 5.6),
-         libgfchangelog0 | glusterfs-common (>= 5.6),
-         libgfdb0 | glusterfs-common (>= 5.6),
-         libgfrpc0 | glusterfs-common (>= 5.6),
-         libgfxdr0 | glusterfs-common (>= 5.6),
-         libglusterfs-dev | glusterfs-common (>= 5.6),
-         libglusterfs0 | glusterfs-common (>= 5.6),
          libiscsi4 (>= 1.12.0) | libiscsi7,
          libjpeg62-turbo,
          libspice-server1 (>= 0.14.0~),

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 8.1.1]
+[FE: rebased for 9.1.2]
 Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
 ---
- block/mirror.c                         | 98 +++++++++++++++++++++-----
+ block/mirror.c                         | 99 ++++++++++++++++++++------
  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(+), 27 deletions(-)
+ 5 files changed, 142 insertions(+), 28 deletions(-)
 
 diff --git a/block/mirror.c b/block/mirror.c
-index d3cacd1708..1ff42c8af1 100644
+index a53582f17b..fafca1360e 100644
 --- a/block/mirror.c
 +++ b/block/mirror.c
 @@ -51,7 +51,7 @@ typedef struct MirrorBlockJob {
@@ -50,7 +50,7 @@ index d3cacd1708..1ff42c8af1 100644
      BlockMirrorBackingMode backing_mode;
      /* Whether the target image requires explicit zero-initialization */
      bool zero_target;
-@@ -65,6 +65,8 @@ typedef struct MirrorBlockJob {
+@@ -73,6 +73,8 @@ typedef struct MirrorBlockJob {
      size_t buf_size;
      int64_t bdev_length;
      unsigned long *cow_bitmap;
@@ -59,9 +59,9 @@ index d3cacd1708..1ff42c8af1 100644
      BdrvDirtyBitmap *dirty_bitmap;
      BdrvDirtyBitmapIter *dbi;
      uint8_t *buf;
-@@ -705,7 +707,8 @@ static int mirror_exit_common(Job *job)
-     bdrv_child_refresh_perms(mirror_top_bs, mirror_top_bs->backing,
+@@ -723,7 +725,8 @@ static int mirror_exit_common(Job *job)
                               &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 d3cacd1708..1ff42c8af1 100644
          BlockDriverState *unfiltered_target = bdrv_skip_filters(target_bs);
  
          if (bdrv_cow_bs(unfiltered_target) != backing) {
-@@ -809,6 +812,16 @@ static void mirror_abort(Job *job)
+@@ -824,6 +827,16 @@ static void mirror_abort(Job *job)
      assert(ret == 0);
  }
  
@@ -86,7 +86,7 @@ index d3cacd1708..1ff42c8af1 100644
  static void coroutine_fn mirror_throttle(MirrorBlockJob *s)
  {
      int64_t now = qemu_clock_get_ns(QEMU_CLOCK_REALTIME);
-@@ -997,7 +1010,8 @@ static int coroutine_fn mirror_run(Job *job, Error **errp)
+@@ -1020,7 +1033,8 @@ 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 d3cacd1708..1ff42c8af1 100644
          ret = mirror_dirty_init(s);
          if (ret < 0 || job_is_cancelled(&s->common.job)) {
              goto immediate_exit;
-@@ -1251,6 +1265,7 @@ static const BlockJobDriver mirror_job_driver = {
+@@ -1309,6 +1323,7 @@ static const BlockJobDriver mirror_job_driver = {
          .run                    = mirror_run,
          .prepare                = mirror_prepare,
          .abort                  = mirror_abort,
@@ -104,7 +104,7 @@ index d3cacd1708..1ff42c8af1 100644
          .pause                  = mirror_pause,
          .complete               = mirror_complete,
          .cancel                 = mirror_cancel,
-@@ -1267,6 +1282,7 @@ static const BlockJobDriver commit_active_job_driver = {
+@@ -1327,6 +1342,7 @@ static const BlockJobDriver commit_active_job_driver = {
          .run                    = mirror_run,
          .prepare                = mirror_prepare,
          .abort                  = mirror_abort,
@@ -112,7 +112,7 @@ index d3cacd1708..1ff42c8af1 100644
          .pause                  = mirror_pause,
          .complete               = mirror_complete,
          .cancel                 = commit_active_cancel,
-@@ -1658,7 +1674,10 @@ static BlockJob *mirror_start_job(
+@@ -1719,7 +1735,10 @@ static BlockJob *mirror_start_job(
                               BlockCompletionFunc *cb,
                               void *opaque,
                               const BlockJobDriver *driver,
@@ -123,10 +123,10 @@ index d3cacd1708..1ff42c8af1 100644
 +                             BlockDriverState *base,
                               bool auto_complete, const char *filter_node_name,
                               bool is_mirror, MirrorCopyMode copy_mode,
-                              Error **errp)
-@@ -1670,10 +1689,39 @@ static BlockJob *mirror_start_job(
-     uint64_t target_perms, target_shared_perms;
-     int ret;
+                              bool base_ro,
+@@ -1734,10 +1753,39 @@ static BlockJob *mirror_start_job(
+ 
+     GLOBAL_STATE_CODE();
  
 -    if (granularity == 0) {
 -        granularity = bdrv_get_default_bitmap_granularity(target);
@@ -166,7 +166,7 @@ index d3cacd1708..1ff42c8af1 100644
      assert(is_power_of_2(granularity));
  
      if (buf_size < 0) {
-@@ -1804,7 +1852,9 @@ static BlockJob *mirror_start_job(
+@@ -1878,7 +1926,9 @@ 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 d3cacd1708..1ff42c8af1 100644
 +    s->bitmap_mode = bitmap_mode;
      s->backing_mode = backing_mode;
      s->zero_target = zero_target;
-     s->copy_mode = copy_mode;
-@@ -1825,6 +1875,18 @@ static BlockJob *mirror_start_job(
-         bdrv_disable_dirty_bitmap(s->dirty_bitmap);
-     }
+     qatomic_set(&s->copy_mode, copy_mode);
+@@ -1904,6 +1954,18 @@ 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 d3cacd1708..1ff42c8af1 100644
 +        }
 +    }
 +
+     bdrv_graph_wrlock();
      ret = block_job_add_bdrv(&s->common, "source", bs, 0,
                               BLK_PERM_WRITE_UNCHANGED | BLK_PERM_WRITE |
-                              BLK_PERM_CONSISTENT_READ,
-@@ -1902,6 +1964,9 @@ fail:
+@@ -1986,6 +2048,9 @@ fail:
          if (s->dirty_bitmap) {
              bdrv_release_dirty_bitmap(s->dirty_bitmap);
          }
@@ -206,7 +206,7 @@ index d3cacd1708..1ff42c8af1 100644
          job_early_fail(&s->common.job);
      }
  
-@@ -1919,31 +1984,25 @@ void mirror_start(const char *job_id, BlockDriverState *bs,
+@@ -2008,35 +2073,28 @@ 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,
@@ -231,19 +231,23 @@ index d3cacd1708..1ff42c8af1 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, errp);
+-                     filter_node_name, true, copy_mode, false, errp);
 +                     &mirror_job_driver, mode, bitmap, bitmap_mode, base,
-+                     false, filter_node_name, true, copy_mode, errp);
++                     false, filter_node_name, true, copy_mode, false, errp);
  }
  
  BlockJob *commit_active_start(const char *job_id, BlockDriverState *bs,
-@@ -1970,7 +2029,8 @@ BlockJob *commit_active_start(const char *job_id, BlockDriverState *bs,
+@@ -2063,7 +2121,8 @@ 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,
@@ -251,13 +255,13 @@ index d3cacd1708..1ff42c8af1 100644
 +                     &commit_active_job_driver, MIRROR_SYNC_MODE_FULL,
 +                     NULL, 0, base, auto_complete,
                       filter_node_name, false, MIRROR_COPY_MODE_BACKGROUND,
-                      errp);
+                      base_read_only, errp);
      if (!job) {
 diff --git a/blockdev.c b/blockdev.c
-index c28462a633..a402fa4bf7 100644
+index 1d1f27cfff..ec45bbaa52 100644
 --- a/blockdev.c
 +++ b/blockdev.c
-@@ -2849,6 +2849,9 @@ static void blockdev_mirror_common(const char *job_id, BlockDriverState *bs,
+@@ -2797,6 +2797,9 @@ static void blockdev_mirror_common(const char *job_id, BlockDriverState *bs,
                                     BlockDriverState *target,
                                     const char *replaces,
                                     enum MirrorSyncMode sync,
@@ -267,7 +271,7 @@ index c28462a633..a402fa4bf7 100644
                                     BlockMirrorBackingMode backing_mode,
                                     bool zero_target,
                                     bool has_speed, int64_t speed,
-@@ -2867,6 +2870,7 @@ static void blockdev_mirror_common(const char *job_id, BlockDriverState *bs,
+@@ -2815,6 +2818,7 @@ static void blockdev_mirror_common(const char *job_id, BlockDriverState *bs,
  {
      BlockDriverState *unfiltered_bs;
      int job_flags = JOB_DEFAULT;
@@ -275,7 +279,7 @@ index c28462a633..a402fa4bf7 100644
  
      GLOBAL_STATE_CODE();
      GRAPH_RDLOCK_GUARD_MAINLOOP();
-@@ -2921,6 +2925,29 @@ static void blockdev_mirror_common(const char *job_id, BlockDriverState *bs,
+@@ -2869,6 +2873,29 @@ static void blockdev_mirror_common(const char *job_id, BlockDriverState *bs,
          sync = MIRROR_SYNC_MODE_FULL;
      }
  
@@ -305,7 +309,7 @@ index c28462a633..a402fa4bf7 100644
      if (!replaces) {
          /* We want to mirror from @bs, but keep implicit filters on top */
          unfiltered_bs = bdrv_skip_implicit_filters(bs);
-@@ -2966,8 +2993,8 @@ static void blockdev_mirror_common(const char *job_id, BlockDriverState *bs,
+@@ -2910,8 +2937,8 @@ 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,
@@ -316,7 +320,7 @@ index c28462a633..a402fa4bf7 100644
                   on_source_error, on_target_error, unmap, filter_node_name,
                   copy_mode, errp);
  }
-@@ -3115,6 +3142,8 @@ void qmp_drive_mirror(DriveMirror *arg, Error **errp)
+@@ -3055,6 +3082,8 @@ void qmp_drive_mirror(DriveMirror *arg, Error **errp)
  
      blockdev_mirror_common(arg->job_id, bs, target_bs,
                             arg->replaces, arg->sync,
@@ -325,7 +329,7 @@ index c28462a633..a402fa4bf7 100644
                             backing_mode, zero_target,
                             arg->has_speed, arg->speed,
                             arg->has_granularity, arg->granularity,
-@@ -3136,6 +3165,8 @@ void qmp_blockdev_mirror(const char *job_id,
+@@ -3074,6 +3103,8 @@ void qmp_blockdev_mirror(const char *job_id,
                           const char *device, const char *target,
                           const char *replaces,
                           MirrorSyncMode sync,
@@ -334,7 +338,7 @@ index c28462a633..a402fa4bf7 100644
                           bool has_speed, int64_t speed,
                           bool has_granularity, uint32_t granularity,
                           bool has_buf_size, int64_t buf_size,
-@@ -3184,7 +3215,8 @@ void qmp_blockdev_mirror(const char *job_id,
+@@ -3114,7 +3145,8 @@ void qmp_blockdev_mirror(const char *job_id,
      }
  
      blockdev_mirror_common(job_id, bs, target_bs,
@@ -345,10 +349,10 @@ index c28462a633..a402fa4bf7 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 da5fb31089..32f0f9858a 100644
+index eb2d92a226..f0c642b194 100644
 --- a/include/block/block_int-global-state.h
 +++ b/include/block/block_int-global-state.h
-@@ -152,7 +152,9 @@ void mirror_start(const char *job_id, BlockDriverState *bs,
+@@ -158,7 +158,9 @@ 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,
@@ -360,10 +364,10 @@ index da5fb31089..32f0f9858a 100644
                    BlockdevOnError on_source_error,
                    BlockdevOnError on_target_error,
 diff --git a/qapi/block-core.json b/qapi/block-core.json
-index bca1a0c372..a5cea82139 100644
+index b1937780e1..0e5f148d30 100644
 --- a/qapi/block-core.json
 +++ b/qapi/block-core.json
-@@ -2145,6 +2145,15 @@
+@@ -2182,6 +2182,15 @@
  #     destination (all the disk, only the sectors allocated in the
  #     topmost image, or only new I/O).
  #
@@ -379,7 +383,7 @@ index bca1a0c372..a5cea82139 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
-@@ -2187,7 +2196,9 @@
+@@ -2224,7 +2233,9 @@
  { 'struct': 'DriveMirror',
    'data': { '*job-id': 'str', 'device': 'str', 'target': 'str',
              '*format': 'str', '*node-name': 'str', '*replaces': 'str',
@@ -390,7 +394,7 @@ index bca1a0c372..a5cea82139 100644
              '*speed': 'int', '*granularity': 'uint32',
              '*buf-size': 'int', '*on-source-error': 'BlockdevOnError',
              '*on-target-error': 'BlockdevOnError',
-@@ -2471,6 +2482,15 @@
+@@ -2503,6 +2514,15 @@
  #     destination (all the disk, only the sectors allocated in the
  #     topmost image, or only new I/O).
  #
@@ -406,7 +410,7 @@ index bca1a0c372..a5cea82139 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
-@@ -2521,7 +2541,8 @@
+@@ -2551,7 +2571,8 @@
  { 'command': 'blockdev-mirror',
    'data': { '*job-id': 'str', 'device': 'str', 'target': 'str',
              '*replaces': 'str',
@@ -417,10 +421,10 @@ index bca1a0c372..a5cea82139 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 d727a5fee8..8a34aa2328 100644
+index 2b358eaaa8..2a149fe021 100644
 --- a/tests/unit/test-block-iothread.c
 +++ b/tests/unit/test-block-iothread.c
-@@ -757,8 +757,8 @@ static void test_propagate_mirror(void)
+@@ -755,8 +755,8 @@ static void test_propagate_mirror(void)
  
      /* Start a mirror job */
      mirror_start("job0", src, target, NULL, JOB_DEFAULT, 0, 0, 0,
@@ -430,4 +434,4 @@ index d727a5fee8..8a34aa2328 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 1ff42c8af1..11b8a8e959 100644
+index fafca1360e..05e738bcce 100644
 --- a/block/mirror.c
 +++ b/block/mirror.c
-@@ -682,8 +682,6 @@ static int mirror_exit_common(Job *job)
+@@ -694,8 +694,6 @@ static int mirror_exit_common(Job *job)
          bdrv_unfreeze_backing_chain(mirror_top_bs, target_bs);
      }
  
@@ -36,9 +36,9 @@ index 1ff42c8af1..11b8a8e959 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);
-@@ -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);
+@@ -805,6 +803,18 @@ static int mirror_exit_common(Job *job)
+     bdrv_drained_end(target_bs);
+     bdrv_unref(target_bs);
  
 +    if (s->sync_bitmap) {
 +        if (s->bitmap_mode == BITMAP_SYNC_MODE_ALWAYS ||
@@ -55,7 +55,7 @@ index 1ff42c8af1..11b8a8e959 100644
      bs_opaque->job = NULL;
  
      bdrv_drained_end(src);
-@@ -1699,10 +1709,6 @@ static BlockJob *mirror_start_job(
+@@ -1763,10 +1773,6 @@ static BlockJob *mirror_start_job(
                         " sync mode",
                         MirrorSyncMode_str(sync_mode));
              return NULL;
@@ -66,7 +66,7 @@ index 1ff42c8af1..11b8a8e959 100644
          }
      } else if (bitmap) {
          error_setg(errp,
-@@ -1719,6 +1725,12 @@ static BlockJob *mirror_start_job(
+@@ -1783,6 +1789,12 @@ 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 a402fa4bf7..01b0ab0549 100644
+index ec45bbaa52..9fab7ec554 100644
 --- a/blockdev.c
 +++ b/blockdev.c
-@@ -2946,6 +2946,9 @@ static void blockdev_mirror_common(const char *job_id, BlockDriverState *bs,
+@@ -2894,6 +2894,9 @@ 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 11b8a8e959..00f2665ca4 100644
+index 05e738bcce..2a2a227f3b 100644
 --- a/block/mirror.c
 +++ b/block/mirror.c
-@@ -792,8 +792,8 @@ static int mirror_exit_common(Job *job)
+@@ -809,8 +809,8 @@ 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 11b8a8e959..00f2665ca4 100644
          }
      }
      bdrv_release_dirty_bitmap(s->dirty_bitmap);
-@@ -1892,11 +1892,8 @@ static BlockJob *mirror_start_job(
+@@ -1971,11 +1971,8 @@ static BlockJob *mirror_start_job(
      }
  
      if (s->sync_mode == MIRROR_SYNC_MODE_BITMAP) {
@@ -43,4 +43,4 @@ index 11b8a8e959..00f2665ca4 100644
 +                                         NULL, true);
      }
  
-     ret = block_job_add_bdrv(&s->common, "source", bs, 0,
+     bdrv_graph_wrlock();

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 8.0]
+[FE: rebase for 8.2.2]
 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 00f2665ca4..60cf574de5 100644
+index 2a2a227f3b..87c0856979 100644
 --- a/block/mirror.c
 +++ b/block/mirror.c
-@@ -1699,31 +1699,13 @@ static BlockJob *mirror_start_job(
-     uint64_t target_perms, target_shared_perms;
-     int ret;
+@@ -1763,31 +1763,13 @@ static BlockJob *mirror_start_job(
+ 
+     GLOBAL_STATE_CODE();
  
 -    if (sync_mode == MIRROR_SYNC_MODE_INCREMENTAL) {
 -        error_setg(errp, "Sync mode '%s' not supported",
@@ -62,10 +62,10 @@ index 00f2665ca4..60cf574de5 100644
  
          if (bitmap_mode != BITMAP_SYNC_MODE_NEVER) {
 diff --git a/blockdev.c b/blockdev.c
-index 01b0ab0549..cd5f205ad1 100644
+index 9fab7ec554..158ac9314b 100644
 --- a/blockdev.c
 +++ b/blockdev.c
-@@ -2925,7 +2925,36 @@ static void blockdev_mirror_common(const char *job_id, BlockDriverState *bs,
+@@ -2873,7 +2873,36 @@ 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 965f5d5450..e04bd059b6 100644
+index c3740ec616..7f38ce6b8b 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 965f5d5450..e04bd059b6 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 252de85681..8db28f9272 100644
+index 5676eb334e..4c452a6aeb 100644
 --- a/monitor/monitor-internal.h
 +++ b/monitor/monitor-internal.h
 @@ -151,6 +151,13 @@ typedef struct {
@@ -78,10 +78,10 @@ index 252de85681..8db28f9272 100644
  
  /**
 diff --git a/monitor/monitor.c b/monitor/monitor.c
-index dc352f9e9d..56e1307014 100644
+index c5a5d30877..07775784d4 100644
 --- a/monitor/monitor.c
 +++ b/monitor/monitor.c
-@@ -117,6 +117,21 @@ bool monitor_cur_is_qmp(void)
+@@ -116,6 +116,21 @@ bool monitor_cur_is_qmp(void)
      return cur_mon && monitor_is_qmp(cur_mon);
  }
  
@@ -104,7 +104,7 @@ index dc352f9e9d..56e1307014 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 a239945e8d..589c9524f8 100644
+index 2f46cf9e49..f093e256e9 100644
 --- a/monitor/qmp.c
 +++ b/monitor/qmp.c
 @@ -165,6 +165,8 @@ static void monitor_qmp_dispatch(MonitorQMP *mon, QObject *req)
@@ -144,7 +144,7 @@ index a239945e8d..589c9524f8 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 e569224eae..eb03782e91 100644
 --- a/qapi/qmp-dispatch.c
 +++ b/qapi/qmp-dispatch.c
 @@ -117,16 +117,28 @@ typedef struct QmpDispatchBH {
@@ -189,7 +189,7 @@ index 176b549473..790bb7d1da 100644
          aio_bh_schedule_oneshot(iohandler_get_aio_context(), do_qmp_dispatch_bh,
                                  &data);
 diff --git a/stubs/monitor-core.c b/stubs/monitor-core.c
-index afa477aae6..d3ff124bf3 100644
+index 1894cdfe1f..d74d0459f0 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 afa477aae6..d3ff124bf3 100644
 +    return -1;
 +}
 +
- void monitor_init_qmp(Chardev *chr, bool pretty, Error **errp)
+ void qapi_event_emit(QAPIEvent event, QDict *qdict)
  {
  }

debian/patches/extra/0002-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 c3508acbb1..289347af58 100644
+index b14983ec54..41c543e627 100644
 --- a/hw/ide/core.c
 +++ b/hw/ide/core.c
-@@ -444,7 +444,7 @@ static void ide_trim_bh_cb(void *opaque)
+@@ -456,7 +456,7 @@ static void ide_trim_bh_cb(void *opaque)
      iocb->bh = NULL;
      qemu_aio_unref(iocb);
  
@@ -67,7 +67,7 @@ index c3508acbb1..289347af58 100644
      blk_dec_in_flight(blk);
  }
  
-@@ -504,6 +504,8 @@ static void ide_issue_trim_cb(void *opaque, int ret)
+@@ -516,6 +516,8 @@ static void ide_issue_trim_cb(void *opaque, int ret)
  done:
      iocb->aiocb = NULL;
      if (iocb->bh) {
@@ -76,7 +76,7 @@ index c3508acbb1..289347af58 100644
          replay_bh_schedule_event(iocb->bh);
      }
  }
-@@ -516,9 +518,6 @@ BlockAIOCB *ide_issue_trim(
+@@ -528,9 +530,6 @@ BlockAIOCB *ide_issue_trim(
      IDEDevice *dev = s->unit ? s->bus->slave : s->bus->master;
      TrimAIOCB *iocb;
  
@@ -86,7 +86,7 @@ index c3508acbb1..289347af58 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,
-@@ -742,8 +741,9 @@ void ide_cancel_dma_sync(IDEState *s)
+@@ -754,8 +753,9 @@ void ide_cancel_dma_sync(IDEState *s)
       */
      if (s->bus->dma->aiocb) {
          trace_ide_cancel_dma_sync_remaining();

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

@@ -1,69 +0,0 @@
-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/0004-migration-block-dirty-bitmap-fix-loading-bitmap-when.patch

@@ -1,48 +0,0 @@
-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/0005-Revert-Revert-graph-lock-Disable-locking-for-now.patch

@@ -1,140 +0,0 @@
-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-migration-states-workaround-snapshot-performance-reg.patch

@@ -1,57 +0,0 @@
-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-Revert-x86-acpi-workaround-Windows-not-handling-name.patch

@@ -1,45 +0,0 @@
-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-target-i386-the-sgx_epc_get_section-stub-is-reachabl.patch

@@ -1,34 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Paolo Bonzini <pbonzini@redhat.com>
-Date: Tue, 1 Feb 2022 20:09:41 +0100
-Subject: [PATCH] target/i386: the sgx_epc_get_section stub is reachable
-MIME-Version: 1.0
-Content-Type: text/plain; charset=UTF-8
-Content-Transfer-Encoding: 8bit
-
-The sgx_epc_get_section stub is reachable from cpu_x86_cpuid.  It
-should not assert, instead it should just return true just like
-the "real" sgx_epc_get_section does when SGX is disabled.
-
-Reported-by: Vladimír Beneš <vbenes@redhat.com>
-Cc: qemu-stable@nongnu.org
-Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
-Message-ID: <20220201190941.106001-1-pbonzini@redhat.com>
-Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
-(cherry-picked from commit 219615740425d9683588207b40a365e6741691a6)
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
----
- hw/i386/sgx-stub.c | 2 +-
- 1 file changed, 1 insertion(+), 1 deletion(-)
-
-diff --git a/hw/i386/sgx-stub.c b/hw/i386/sgx-stub.c
-index 26833eb233..16b1dfd90b 100644
---- a/hw/i386/sgx-stub.c
-+++ b/hw/i386/sgx-stub.c
-@@ -34,5 +34,5 @@ void pc_machine_init_sgx_epc(PCMachineState *pcms)
- 
- bool sgx_epc_get_section(int section_nr, uint64_t *addr, uint64_t *size)
- {
--    g_assert_not_reached();
-+    return true;
- }

debian/patches/extra/0009-ui-clipboard-mark-type-as-not-available-when-there-i.patch

@@ -1,86 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Fiona Ebner <f.ebner@proxmox.com>
-Date: Wed, 24 Jan 2024 11:57:48 +0100
-Subject: [PATCH] ui/clipboard: mark type as not available when there is no
- data
-MIME-Version: 1.0
-Content-Type: text/plain; charset=UTF-8
-Content-Transfer-Encoding: 8bit
-
-With VNC, a client can send a non-extended VNC_MSG_CLIENT_CUT_TEXT
-message with len=0. In qemu_clipboard_set_data(), the clipboard info
-will be updated setting data to NULL (because g_memdup(data, size)
-returns NULL when size is 0). If the client does not set the
-VNC_ENCODING_CLIPBOARD_EXT feature when setting up the encodings, then
-the 'request' callback for the clipboard peer is not initialized.
-Later, because data is NULL, qemu_clipboard_request() can be reached
-via vdagent_chr_write() and vdagent_clipboard_recv_request() and
-there, the clipboard owner's 'request' callback will be attempted to
-be called, but that is a NULL pointer.
-
-In particular, this can happen when using the KRDC (22.12.3) VNC
-client.
-
-Another scenario leading to the same issue is with two clients (say
-noVNC and KRDC):
-
-The noVNC client sets the extension VNC_FEATURE_CLIPBOARD_EXT and
-initializes its cbpeer.
-
-The KRDC client does not, but triggers a vnc_client_cut_text() (note
-it's not the _ext variant)). There, a new clipboard info with it as
-the 'owner' is created and via qemu_clipboard_set_data() is called,
-which in turn calls qemu_clipboard_update() with that info.
-
-In qemu_clipboard_update(), the notifier for the noVNC client will be
-called, i.e. vnc_clipboard_notify() and also set vs->cbinfo for the
-noVNC client. The 'owner' in that clipboard info is the clipboard peer
-for the KRDC client, which did not initialize the 'request' function.
-That sounds correct to me, it is the owner of that clipboard info.
-
-Then when noVNC sends a VNC_MSG_CLIENT_CUT_TEXT message (it did set
-the VNC_FEATURE_CLIPBOARD_EXT feature correctly, so a check for it
-passes), that clipboard info is passed to qemu_clipboard_request() and
-the original segfault still happens.
-
-Fix the issue by handling updates with size 0 differently. In
-particular, mark in the clipboard info that the type is not available.
-
-While at it, switch to g_memdup2(), because g_memdup() is deprecated.
-
-Cc: qemu-stable@nongnu.org
-Fixes: CVE-2023-6683
-Reported-by: Markus Frank <m.frank@proxmox.com>
-Suggested-by: Marc-André Lureau <marcandre.lureau@redhat.com>
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
-Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
-Tested-by: Markus Frank <m.frank@proxmox.com>
-(picked from https://lists.nongnu.org/archive/html/qemu-stable/2024-01/msg00228.html)
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
----
- ui/clipboard.c | 12 +++++++++---
- 1 file changed, 9 insertions(+), 3 deletions(-)
-
-diff --git a/ui/clipboard.c b/ui/clipboard.c
-index 3d14bffaf8..b3f6fa3c9e 100644
---- a/ui/clipboard.c
-+++ b/ui/clipboard.c
-@@ -163,9 +163,15 @@ void qemu_clipboard_set_data(QemuClipboardPeer *peer,
-     }
- 
-     g_free(info->types[type].data);
--    info->types[type].data = g_memdup(data, size);
--    info->types[type].size = size;
--    info->types[type].available = true;
-+    if (size) {
-+        info->types[type].data = g_memdup2(data, size);
-+        info->types[type].size = size;
-+        info->types[type].available = true;
-+    } else {
-+        info->types[type].data = NULL;
-+        info->types[type].size = 0;
-+        info->types[type].available = false;
-+    }
- 
-     if (update) {
-         qemu_clipboard_update(info);

debian/patches/extra/0010-virtio-scsi-Attach-event-vq-notifier-with-no_poll.patch

@@ -1,65 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Hanna Czenczek <hreitz@redhat.com>
-Date: Fri, 2 Feb 2024 16:31:56 +0100
-Subject: [PATCH] virtio-scsi: Attach event vq notifier with no_poll
-
-As of commit 38738f7dbbda90fbc161757b7f4be35b52205552 ("virtio-scsi:
-don't waste CPU polling the event virtqueue"), we only attach an io_read
-notifier for the virtio-scsi event virtqueue instead, and no polling
-notifiers.  During operation, the event virtqueue is typically
-non-empty, but none of the buffers are intended to be used immediately.
-Instead, they only get used when certain events occur.  Therefore, it
-makes no sense to continuously poll it when non-empty, because it is
-supposed to be and stay non-empty.
-
-We do this by using virtio_queue_aio_attach_host_notifier_no_poll()
-instead of virtio_queue_aio_attach_host_notifier() for the event
-virtqueue.
-
-Commit 766aa2de0f29b657148e04599320d771c36fd126 ("virtio-scsi: implement
-BlockDevOps->drained_begin()") however has virtio_scsi_drained_end() use
-virtio_queue_aio_attach_host_notifier() for all virtqueues, including
-the event virtqueue.  This can lead to it being polled again, undoing
-the benefit of commit 38738f7dbbda90fbc161757b7f4be35b52205552.
-
-Fix it by using virtio_queue_aio_attach_host_notifier_no_poll() for the
-event virtqueue.
-
-       ("virtio-scsi: implement BlockDevOps->drained_begin()")
-
-Reported-by: Fiona Ebner <f.ebner@proxmox.com>
-Fixes: 766aa2de0f29b657148e04599320d771c36fd126
-Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Fiona Ebner <f.ebner@proxmox.com>
-Reviewed-by: Fiona Ebner <f.ebner@proxmox.com>
-Signed-off-by: Hanna Czenczek <hreitz@redhat.com>
-Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
----
- hw/scsi/virtio-scsi.c | 7 ++++++-
- 1 file changed, 6 insertions(+), 1 deletion(-)
-
-diff --git a/hw/scsi/virtio-scsi.c b/hw/scsi/virtio-scsi.c
-index 45b95ea070..ad24a882fd 100644
---- a/hw/scsi/virtio-scsi.c
-+++ b/hw/scsi/virtio-scsi.c
-@@ -1148,6 +1148,7 @@ static void virtio_scsi_drained_begin(SCSIBus *bus)
- static void virtio_scsi_drained_end(SCSIBus *bus)
- {
-     VirtIOSCSI *s = container_of(bus, VirtIOSCSI, bus);
-+    VirtIOSCSICommon *vs = VIRTIO_SCSI_COMMON(s);
-     VirtIODevice *vdev = VIRTIO_DEVICE(s);
-     uint32_t total_queues = VIRTIO_SCSI_VQ_NUM_FIXED +
-                             s->parent_obj.conf.num_queues;
-@@ -1165,7 +1166,11 @@ static void virtio_scsi_drained_end(SCSIBus *bus)
- 
-     for (uint32_t i = 0; i < total_queues; i++) {
-         VirtQueue *vq = virtio_get_queue(vdev, i);
--        virtio_queue_aio_attach_host_notifier(vq, s->ctx);
-+        if (vq == vs->event_vq) {
-+            virtio_queue_aio_attach_host_notifier_no_poll(vq, s->ctx);
-+        } else {
-+            virtio_queue_aio_attach_host_notifier(vq, s->ctx);
-+        }
-     }
- }
- 

debian/patches/extra/0011-virtio-Re-enable-notifications-after-drain.patch

@@ -1,125 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Hanna Czenczek <hreitz@redhat.com>
-Date: Fri, 2 Feb 2024 16:31:57 +0100
-Subject: [PATCH] virtio: Re-enable notifications after drain
-
-During drain, we do not care about virtqueue notifications, which is why
-we remove the handlers on it.  When removing those handlers, whether vq
-notifications are enabled or not depends on whether we were in polling
-mode or not; if not, they are enabled (by default); if so, they have
-been disabled by the io_poll_start callback.
-
-Because we do not care about those notifications after removing the
-handlers, this is fine.  However, we have to explicitly ensure they are
-enabled when re-attaching the handlers, so we will resume receiving
-notifications.  We do this in virtio_queue_aio_attach_host_notifier*().
-If such a function is called while we are in a polling section,
-attaching the notifiers will then invoke the io_poll_start callback,
-re-disabling notifications.
-
-Because we will always miss virtqueue updates in the drained section, we
-also need to poll the virtqueue once after attaching the notifiers.
-
-Buglink: https://issues.redhat.com/browse/RHEL-3934
-Signed-off-by: Hanna Czenczek <hreitz@redhat.com>
-Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
----
- hw/virtio/virtio.c  | 42 ++++++++++++++++++++++++++++++++++++++++++
- include/block/aio.h |  7 ++++++-
- 2 files changed, 48 insertions(+), 1 deletion(-)
-
-diff --git a/hw/virtio/virtio.c b/hw/virtio/virtio.c
-index 969c25f4cf..02cce83111 100644
---- a/hw/virtio/virtio.c
-+++ b/hw/virtio/virtio.c
-@@ -3526,6 +3526,17 @@ static void virtio_queue_host_notifier_aio_poll_end(EventNotifier *n)
- 
- void virtio_queue_aio_attach_host_notifier(VirtQueue *vq, AioContext *ctx)
- {
-+    /*
-+     * virtio_queue_aio_detach_host_notifier() can leave notifications disabled.
-+     * Re-enable them.  (And if detach has not been used before, notifications
-+     * being enabled is still the default state while a notifier is attached;
-+     * see virtio_queue_host_notifier_aio_poll_end(), which will always leave
-+     * notifications enabled once the polling section is left.)
-+     */
-+    if (!virtio_queue_get_notification(vq)) {
-+        virtio_queue_set_notification(vq, 1);
-+    }
-+
-     aio_set_event_notifier(ctx, &vq->host_notifier,
-                            virtio_queue_host_notifier_read,
-                            virtio_queue_host_notifier_aio_poll,
-@@ -3533,6 +3544,13 @@ void virtio_queue_aio_attach_host_notifier(VirtQueue *vq, AioContext *ctx)
-     aio_set_event_notifier_poll(ctx, &vq->host_notifier,
-                                 virtio_queue_host_notifier_aio_poll_begin,
-                                 virtio_queue_host_notifier_aio_poll_end);
-+
-+    /*
-+     * We will have ignored notifications about new requests from the guest
-+     * while no notifiers were attached, so "kick" the virt queue to process
-+     * those requests now.
-+     */
-+    event_notifier_set(&vq->host_notifier);
- }
- 
- /*
-@@ -3543,14 +3561,38 @@ void virtio_queue_aio_attach_host_notifier(VirtQueue *vq, AioContext *ctx)
-  */
- void virtio_queue_aio_attach_host_notifier_no_poll(VirtQueue *vq, AioContext *ctx)
- {
-+    /* See virtio_queue_aio_attach_host_notifier() */
-+    if (!virtio_queue_get_notification(vq)) {
-+        virtio_queue_set_notification(vq, 1);
-+    }
-+
-     aio_set_event_notifier(ctx, &vq->host_notifier,
-                            virtio_queue_host_notifier_read,
-                            NULL, NULL);
-+
-+    /*
-+     * See virtio_queue_aio_attach_host_notifier().
-+     * Note that this may be unnecessary for the type of virtqueues this
-+     * function is used for.  Still, it will not hurt to have a quick look into
-+     * whether we can/should process any of the virtqueue elements.
-+     */
-+    event_notifier_set(&vq->host_notifier);
- }
- 
- void virtio_queue_aio_detach_host_notifier(VirtQueue *vq, AioContext *ctx)
- {
-     aio_set_event_notifier(ctx, &vq->host_notifier, NULL, NULL, NULL);
-+
-+    /*
-+     * aio_set_event_notifier_poll() does not guarantee whether io_poll_end()
-+     * will run after io_poll_begin(), so by removing the notifier, we do not
-+     * know whether virtio_queue_host_notifier_aio_poll_end() has run after a
-+     * previous virtio_queue_host_notifier_aio_poll_begin(), i.e. whether
-+     * notifications are enabled or disabled.  It does not really matter anyway;
-+     * we just removed the notifier, so we do not care about notifications until
-+     * we potentially re-attach it.  The attach_host_notifier functions will
-+     * ensure that notifications are enabled again when they are needed.
-+     */
- }
- 
- void virtio_queue_host_notifier_read(EventNotifier *n)
-diff --git a/include/block/aio.h b/include/block/aio.h
-index 32042e8905..79efadfa48 100644
---- a/include/block/aio.h
-+++ b/include/block/aio.h
-@@ -498,9 +498,14 @@ void aio_set_event_notifier(AioContext *ctx,
-                             AioPollFn *io_poll,
-                             EventNotifierHandler *io_poll_ready);
- 
--/* Set polling begin/end callbacks for an event notifier that has already been
-+/*
-+ * Set polling begin/end callbacks for an event notifier that has already been
-  * registered with aio_set_event_notifier.  Do nothing if the event notifier is
-  * not registered.
-+ *
-+ * Note that if the io_poll_end() callback (or the entire notifier) is removed
-+ * during polling, it will not be called, so an io_poll_begin() is not
-+ * necessarily always followed by an io_poll_end().
-  */
- void aio_set_event_notifier_poll(AioContext *ctx,
-                                  EventNotifier *notifier,

debian/patches/extra/0012-qemu_init-increase-NOFILE-soft-limit-on-POSIX.patch

@@ -1,119 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Fiona Ebner <f.ebner@proxmox.com>
-Date: Mon, 18 Dec 2023 11:13:40 +0100
-Subject: [PATCH] qemu_init: increase NOFILE soft limit on POSIX
-MIME-Version: 1.0
-Content-Type: text/plain; charset=UTF-8
-Content-Transfer-Encoding: 8bit
-
-In many configurations, e.g. multiple vNICs with multiple queues or
-with many Ceph OSDs, the default soft limit of 1024 is not enough.
-QEMU is supposed to work fine with file descriptors >= 1024 and does
-not use select() on POSIX. Bump the soft limit to the allowed hard
-limit to avoid issues with the aforementioned configurations.
-
-Of course the limit could be raised from the outside, but the man page
-of systemd.exec states about 'LimitNOFILE=':
-
-> Don't use.
-> [...]
-> Typically applications should increase their soft limit to the hard
-> limit on their own, if they are OK with working with file
-> descriptors above 1023,
-
-If the soft limit is already the same as the hard limit, avoid the
-superfluous setrlimit call. This can avoid a warning with a strict
-seccomp filter blocking setrlimit if NOFILE was already raised before
-executing QEMU.
-
-Buglink: https://bugzilla.proxmox.com/show_bug.cgi?id=4507
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
-Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
----
- include/sysemu/os-posix.h |  1 +
- include/sysemu/os-win32.h |  5 +++++
- os-posix.c                | 22 ++++++++++++++++++++++
- softmmu/vl.c              |  2 ++
- 4 files changed, 30 insertions(+)
-
-diff --git a/include/sysemu/os-posix.h b/include/sysemu/os-posix.h
-index 1030d39904..edc415aff5 100644
---- a/include/sysemu/os-posix.h
-+++ b/include/sysemu/os-posix.h
-@@ -48,6 +48,7 @@ void os_setup_early_signal_handling(void);
- void os_set_proc_name(const char *s);
- void os_setup_signal_handling(void);
- void os_daemonize(void);
-+void os_setup_limits(void);
- void os_setup_post(void);
- int os_mlock(void);
- 
-diff --git a/include/sysemu/os-win32.h b/include/sysemu/os-win32.h
-index 91aa0d7ec0..f6e23fe01e 100644
---- a/include/sysemu/os-win32.h
-+++ b/include/sysemu/os-win32.h
-@@ -129,6 +129,11 @@ static inline int os_mlock(void)
-     return -ENOSYS;
- }
- 
-+void os_setup_limits(void)
-+{
-+    return;
-+}
-+
- #define fsync _commit
- 
- #if !defined(lseek)
-diff --git a/os-posix.c b/os-posix.c
-index cfcb96533c..0cc1d991b1 100644
---- a/os-posix.c
-+++ b/os-posix.c
-@@ -24,6 +24,7 @@
-  */
- 
- #include "qemu/osdep.h"
-+#include <sys/resource.h>
- #include <sys/wait.h>
- #include <pwd.h>
- #include <grp.h>
-@@ -286,6 +287,27 @@ void os_daemonize(void)
-     }
- }
- 
-+void os_setup_limits(void)
-+{
-+    struct rlimit nofile;
-+
-+    if (getrlimit(RLIMIT_NOFILE, &nofile) < 0) {
-+        warn_report("unable to query NOFILE limit: %s", strerror(errno));
-+        return;
-+    }
-+
-+    if (nofile.rlim_cur == nofile.rlim_max) {
-+        return;
-+    }
-+
-+    nofile.rlim_cur = nofile.rlim_max;
-+
-+    if (setrlimit(RLIMIT_NOFILE, &nofile) < 0) {
-+        warn_report("unable to set NOFILE limit: %s", strerror(errno));
-+        return;
-+    }
-+}
-+
- void os_setup_post(void)
- {
-     int fd = 0;
-diff --git a/softmmu/vl.c b/softmmu/vl.c
-index c9e9ede237..ba6ad8a8df 100644
---- a/softmmu/vl.c
-+++ b/softmmu/vl.c
-@@ -2713,6 +2713,8 @@ void qemu_init(int argc, char **argv)
-     error_init(argv[0]);
-     qemu_init_exec_dir(argv[0]);
- 
-+    os_setup_limits();
-+
-     qemu_init_arch_modules();
- 
-     qemu_init_subsystems();

debian/patches/extra/0013-virtio-blk-avoid-using-ioeventfd-state-in-irqfd-cond.patch

@@ -1,61 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Stefan Hajnoczi <stefanha@redhat.com>
-Date: Mon, 22 Jan 2024 12:26:25 -0500
-Subject: [PATCH] virtio-blk: avoid using ioeventfd state in irqfd conditional
-
-Requests that complete in an IOThread use irqfd to notify the guest
-while requests that complete in the main loop thread use the traditional
-qdev irq code path. The reason for this conditional is that the irq code
-path requires the BQL:
-
-  if (s->ioeventfd_started && !s->ioeventfd_disabled) {
-      virtio_notify_irqfd(vdev, req->vq);
-  } else {
-      virtio_notify(vdev, req->vq);
-  }
-
-There is a corner case where the conditional invokes the irq code path
-instead of the irqfd code path:
-
-  static void virtio_blk_stop_ioeventfd(VirtIODevice *vdev)
-  {
-      ...
-      /*
-       * Set ->ioeventfd_started to false before draining so that host notifiers
-       * are not detached/attached anymore.
-       */
-      s->ioeventfd_started = false;
-
-      /* Wait for virtio_blk_dma_restart_bh() and in flight I/O to complete */
-      blk_drain(s->conf.conf.blk);
-
-During blk_drain() the conditional produces the wrong result because
-ioeventfd_started is false.
-
-Use qemu_in_iothread() instead of checking the ioeventfd state.
-
-Cc: qemu-stable@nongnu.org
-Buglink: https://issues.redhat.com/browse/RHEL-15394
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
-Message-ID: <20240122172625.415386-1-stefanha@redhat.com>
-Reviewed-by: Kevin Wolf <kwolf@redhat.com>
-Signed-off-by: Kevin Wolf <kwolf@redhat.com>
-[FE: backport: dataplane -> ioeventfd rework didn't happen yet]
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
----
- hw/block/virtio-blk.c | 2 +-
- 1 file changed, 1 insertion(+), 1 deletion(-)
-
-diff --git a/hw/block/virtio-blk.c b/hw/block/virtio-blk.c
-index 39e7f23fab..61bd1f6859 100644
---- a/hw/block/virtio-blk.c
-+++ b/hw/block/virtio-blk.c
-@@ -64,7 +64,7 @@ static void virtio_blk_req_complete(VirtIOBlockReq *req, unsigned char status)
-     iov_discard_undo(&req->inhdr_undo);
-     iov_discard_undo(&req->outhdr_undo);
-     virtqueue_push(req->vq, &req->elem, req->in_len);
--    if (s->dataplane_started && !s->dataplane_disabled) {
-+    if (qemu_in_iothread()) {
-         virtio_blk_data_plane_notify(s->dataplane, req->vq);
-     } else {
-         virtio_notify(vdev, req->vq);

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

@@ -14,10 +14,10 @@ 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 7f540b03ed..ca551baa42 100644
+index 56d1972d15..cfa0b832ba 100644
 --- a/block/file-posix.c
 +++ b/block/file-posix.c
-@@ -563,7 +563,7 @@ static QemuOptsList raw_runtime_opts = {
+@@ -565,7 +565,7 @@ static QemuOptsList raw_runtime_opts = {
          {
              .name = "locking",
              .type = QEMU_OPT_STRING,
@@ -26,7 +26,7 @@ index 7f540b03ed..ca551baa42 100644
          },
          {
              .name = "pr-manager",
-@@ -663,7 +663,7 @@ static int raw_open_common(BlockDriverState *bs, QDict *options,
+@@ -665,7 +665,7 @@ static int raw_open_common(BlockDriverState *bs, QDict *options,
          s->use_lock = false;
          break;
      case ON_OFF_AUTO_AUTO:

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 685ec58318..22edf4ee96 100644
+index cdd5b109b0..653a37e9d1 100644
 --- a/include/net/net.h
 +++ b/include/net/net.h
-@@ -260,8 +260,8 @@ void netdev_add(QemuOpts *opts, Error **errp);
+@@ -305,8 +305,8 @@ 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 0893b794e9..6d650a58b9 100644
+index 76f24446a5..2a47d79b49 100644
 --- a/target/i386/cpu.h
 +++ b/target/i386/cpu.h
-@@ -2243,9 +2243,9 @@ uint64_t cpu_get_tsc(CPUX86State *env);
+@@ -2556,9 +2556,9 @@ 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 52a59386d7..b20c25aee0 100644
+index 0326c63bec..d523d00200 100644
 --- a/ui/spice-core.c
 +++ b/ui/spice-core.c
-@@ -691,32 +691,35 @@ static void qemu_spice_init(void)
+@@ -690,32 +690,35 @@ 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 ad5fadbe79..d0011085c4 100644
+index c6d25ae733..ccca125c3a 100644
 --- a/block/gluster.c
 +++ b/block/gluster.c
-@@ -43,7 +43,7 @@
+@@ -42,7 +42,7 @@
  #define GLUSTER_DEBUG_DEFAULT       4
  #define GLUSTER_DEBUG_MAX           9
  #define GLUSTER_OPT_LOGFILE         "logfile"
@@ -21,7 +21,7 @@ index ad5fadbe79..d0011085c4 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
-@@ -425,6 +425,7 @@ static struct glfs *qemu_gluster_glfs_init(BlockdevOptionsGluster *gconf,
+@@ -421,6 +421,7 @@ static struct glfs *qemu_gluster_glfs_init(BlockdevOptionsGluster *gconf,
      int old_errno;
      SocketAddressList *server;
      uint64_t port;
@@ -29,7 +29,7 @@ index ad5fadbe79..d0011085c4 100644
  
      glfs = glfs_find_preopened(gconf->volume);
      if (glfs) {
-@@ -467,9 +468,15 @@ static struct glfs *qemu_gluster_glfs_init(BlockdevOptionsGluster *gconf,
+@@ -463,9 +464,15 @@ 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 978671411e..a4749f3b1b 100644
+index af984fb7db..bf143fac00 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 d0011085c4..2df3d6e35d 100644
+index ccca125c3a..301a653ea7 100644
 --- a/block/gluster.c
 +++ b/block/gluster.c
-@@ -58,6 +58,7 @@ typedef struct GlusterAIOCB {
+@@ -57,6 +57,7 @@ typedef struct GlusterAIOCB {
      int ret;
      Coroutine *coroutine;
      AioContext *aio_context;
@@ -27,7 +27,7 @@ index d0011085c4..2df3d6e35d 100644
  } GlusterAIOCB;
  
  typedef struct BDRVGlusterState {
-@@ -753,8 +754,10 @@ static void gluster_finish_aiocb(struct glfs_fd *fd, ssize_t ret,
+@@ -746,8 +747,10 @@ 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 d0011085c4..2df3d6e35d 100644
      }
  
      aio_co_schedule(acb->aio_context, acb->coroutine);
-@@ -1021,6 +1024,7 @@ static coroutine_fn int qemu_gluster_co_pwrite_zeroes(BlockDriverState *bs,
+@@ -1018,6 +1021,7 @@ 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 d0011085c4..2df3d6e35d 100644
  
      ret = glfs_zerofill_async(s->fd, offset, bytes, gluster_finish_aiocb, &acb);
      if (ret < 0) {
-@@ -1201,9 +1205,11 @@ static coroutine_fn int qemu_gluster_co_rw(BlockDriverState *bs,
+@@ -1198,9 +1202,11 @@ 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 d0011085c4..2df3d6e35d 100644
          ret = glfs_preadv_async(s->fd, qiov->iov, qiov->niov, offset, 0,
                                  gluster_finish_aiocb, &acb);
      }
-@@ -1266,6 +1272,7 @@ static coroutine_fn int qemu_gluster_co_flush_to_disk(BlockDriverState *bs)
+@@ -1263,6 +1269,7 @@ 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 d0011085c4..2df3d6e35d 100644
  
      ret = glfs_fsync_async(s->fd, gluster_finish_aiocb, &acb);
      if (ret < 0) {
-@@ -1314,6 +1321,7 @@ static coroutine_fn int qemu_gluster_co_pdiscard(BlockDriverState *bs,
+@@ -1311,6 +1318,7 @@ 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 78433f3746..25d427edd1 100644
+index 2044c22a4c..4c8b5412c6 100644
 --- a/qemu-img.c
 +++ b/qemu-img.c
-@@ -3062,7 +3062,8 @@ static int img_info(int argc, char **argv)
+@@ -3075,7 +3075,8 @@ 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 1b1dab5b17..d1616c045a 100644
+index c9dd70a892..048788b23d 100644
 --- a/qemu-img-cmds.hx
 +++ b/qemu-img-cmds.hx
-@@ -58,9 +58,9 @@ SRST
+@@ -60,9 +60,9 @@ SRST
  ERST
  
  DEF("dd", img_dd,
@@ -54,10 +54,10 @@ index 1b1dab5b17..d1616c045a 100644
  
  DEF("info", img_info,
 diff --git a/qemu-img.c b/qemu-img.c
-index 25d427edd1..220e6ec577 100644
+index 4c8b5412c6..d5fa89a204 100644
 --- a/qemu-img.c
 +++ b/qemu-img.c
-@@ -4899,10 +4899,12 @@ static int img_bitmap(int argc, char **argv)
+@@ -4997,10 +4997,12 @@ 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 25d427edd1..220e6ec577 100644
  };
  
  struct DdIo {
-@@ -4978,6 +4980,19 @@ static int img_dd_skip(const char *arg,
+@@ -5076,6 +5078,19 @@ static int img_dd_skip(const char *arg,
      return 0;
  }
  
@@ -90,7 +90,7 @@ index 25d427edd1..220e6ec577 100644
  static int img_dd(int argc, char **argv)
  {
      int ret = 0;
-@@ -5018,6 +5033,7 @@ static int img_dd(int argc, char **argv)
+@@ -5116,6 +5131,7 @@ 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 25d427edd1..220e6ec577 100644
          { NULL, NULL, 0 }
      };
      const struct option long_options[] = {
-@@ -5093,91 +5109,112 @@ static int img_dd(int argc, char **argv)
+@@ -5191,91 +5207,112 @@ static int img_dd(int argc, char **argv)
          arg = NULL;
      }
  
@@ -275,7 +275,7 @@ index 25d427edd1..220e6ec577 100644
      }
  
      if (dd.flags & C_SKIP && (in.offset > INT64_MAX / in.bsz ||
-@@ -5194,20 +5231,43 @@ static int img_dd(int argc, char **argv)
+@@ -5292,20 +5329,43 @@ 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 220e6ec577..58bf9b43d1 100644
+index d5fa89a204..d458e85af2 100644
 --- a/qemu-img.c
 +++ b/qemu-img.c
-@@ -4900,11 +4900,13 @@ static int img_bitmap(int argc, char **argv)
+@@ -4998,11 +4998,13 @@ 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 220e6ec577..58bf9b43d1 100644
  };
  
  struct DdIo {
-@@ -4993,6 +4995,19 @@ static int img_dd_osize(const char *arg,
+@@ -5091,6 +5093,19 @@ static int img_dd_osize(const char *arg,
      return 0;
  }
  
@@ -53,7 +53,7 @@ index 220e6ec577..58bf9b43d1 100644
  static int img_dd(int argc, char **argv)
  {
      int ret = 0;
-@@ -5007,12 +5022,14 @@ static int img_dd(int argc, char **argv)
+@@ -5105,12 +5120,14 @@ 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 220e6ec577..58bf9b43d1 100644
      };
      struct DdIo in = {
          .bsz = 512, /* Block size is by default 512 bytes */
-@@ -5034,6 +5051,7 @@ static int img_dd(int argc, char **argv)
+@@ -5132,6 +5149,7 @@ 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 220e6ec577..58bf9b43d1 100644
          { NULL, NULL, 0 }
      };
      const struct option long_options[] = {
-@@ -5230,9 +5248,10 @@ static int img_dd(int argc, char **argv)
+@@ -5328,9 +5346,10 @@ static int img_dd(int argc, char **argv)
  
      in.buf = g_new(uint8_t, in.bsz);
  
@@ -90,7 +90,7 @@ index 220e6ec577..58bf9b43d1 100644
          if (blk1) {
              in_ret = blk_pread(blk1, in_pos, bytes, in.buf, 0);
              if (in_ret == 0) {
-@@ -5241,6 +5260,9 @@ static int img_dd(int argc, char **argv)
+@@ -5339,6 +5358,9 @@ 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 15aeddc6d8..5e713e231d 100644
+index 3653adb963..d83e8fb3c0 100644
 --- a/docs/tools/qemu-img.rst
 +++ b/docs/tools/qemu-img.rst
-@@ -208,6 +208,10 @@ Parameters to convert subcommand:
+@@ -212,6 +212,10 @@ Parameters to convert subcommand:
  
  Parameters to dd subcommand:
  
@@ -27,7 +27,7 @@ index 15aeddc6d8..5e713e231d 100644
  .. program:: qemu-img-dd
  
  .. option:: bs=BLOCK_SIZE
-@@ -488,7 +492,7 @@ Command description:
+@@ -492,7 +496,7 @@ Command description:
    it doesn't need to be specified separately in this case.
  
  
@@ -36,7 +36,7 @@ index 15aeddc6d8..5e713e231d 100644
  
    dd copies from *INPUT* file to *OUTPUT* file converting it from
    *FMT* format to *OUTPUT_FMT* format.
-@@ -499,6 +503,11 @@ Command description:
+@@ -503,6 +507,11 @@ Command description:
  
    The size syntax is similar to :manpage:`dd(1)`'s size syntax.
  
@@ -49,10 +49,10 @@ index 15aeddc6d8..5e713e231d 100644
  
    Give information about the disk image *FILENAME*. Use it in
 diff --git a/qemu-img-cmds.hx b/qemu-img-cmds.hx
-index d1616c045a..b5b0bb4467 100644
+index 048788b23d..0b29a67a06 100644
 --- a/qemu-img-cmds.hx
 +++ b/qemu-img-cmds.hx
-@@ -58,9 +58,9 @@ SRST
+@@ -60,9 +60,9 @@ SRST
  ERST
  
  DEF("dd", img_dd,
@@ -65,10 +65,10 @@ index d1616c045a..b5b0bb4467 100644
  
  DEF("info", img_info,
 diff --git a/qemu-img.c b/qemu-img.c
-index 58bf9b43d1..9d414d639b 100644
+index d458e85af2..dc13efba8b 100644
 --- a/qemu-img.c
 +++ b/qemu-img.c
-@@ -5024,7 +5024,7 @@ static int img_dd(int argc, char **argv)
+@@ -5122,7 +5122,7 @@ 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 58bf9b43d1..9d414d639b 100644
      struct DdInfo dd = {
          .flags = 0,
          .count = 0,
-@@ -5062,7 +5062,7 @@ static int img_dd(int argc, char **argv)
+@@ -5160,7 +5160,7 @@ static int img_dd(int argc, char **argv)
          { 0, 0, 0, 0 }
      };
  
@@ -86,7 +86,7 @@ index 58bf9b43d1..9d414d639b 100644
          if (c == EOF) {
              break;
          }
-@@ -5082,6 +5082,9 @@ static int img_dd(int argc, char **argv)
+@@ -5180,6 +5180,9 @@ static int img_dd(int argc, char **argv)
          case 'h':
              help();
              break;
@@ -96,7 +96,7 @@ index 58bf9b43d1..9d414d639b 100644
          case 'U':
              force_share = true;
              break;
-@@ -5212,13 +5215,15 @@ static int img_dd(int argc, char **argv)
+@@ -5310,13 +5313,15 @@ 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 5e713e231d..9390d5e5cf 100644
+index d83e8fb3c0..61c6b21859 100644
 --- a/docs/tools/qemu-img.rst
 +++ b/docs/tools/qemu-img.rst
-@@ -492,10 +492,10 @@ Command description:
+@@ -496,10 +496,10 @@ Command description:
    it doesn't need to be specified separately in this case.
  
  
@@ -30,10 +30,10 @@ index 5e713e231d..9390d5e5cf 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 b5b0bb4467..36f97e1f19 100644
+index 0b29a67a06..758f397232 100644
 --- a/qemu-img-cmds.hx
 +++ b/qemu-img-cmds.hx
-@@ -58,9 +58,9 @@ SRST
+@@ -60,9 +60,9 @@ SRST
  ERST
  
  DEF("dd", img_dd,
@@ -46,10 +46,10 @@ index b5b0bb4467..36f97e1f19 100644
  
  DEF("info", img_info,
 diff --git a/qemu-img.c b/qemu-img.c
-index 9d414d639b..e13a12137b 100644
+index dc13efba8b..02f2e0aa45 100644
 --- a/qemu-img.c
 +++ b/qemu-img.c
-@@ -5016,6 +5016,7 @@ static int img_dd(int argc, char **argv)
+@@ -5114,6 +5114,7 @@ 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 9d414d639b..e13a12137b 100644
      QemuOptsList *create_opts = NULL;
      Error *local_err = NULL;
      bool image_opts = false;
-@@ -5025,6 +5026,7 @@ static int img_dd(int argc, char **argv)
+@@ -5123,6 +5124,7 @@ 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 9d414d639b..e13a12137b 100644
      struct DdInfo dd = {
          .flags = 0,
          .count = 0,
-@@ -5062,7 +5064,7 @@ static int img_dd(int argc, char **argv)
+@@ -5160,7 +5162,7 @@ static int img_dd(int argc, char **argv)
          { 0, 0, 0, 0 }
      };
  
@@ -74,7 +74,7 @@ index 9d414d639b..e13a12137b 100644
          if (c == EOF) {
              break;
          }
-@@ -5085,6 +5087,19 @@ static int img_dd(int argc, char **argv)
+@@ -5183,6 +5185,19 @@ static int img_dd(int argc, char **argv)
          case 'n':
              skip_create = true;
              break;
@@ -94,7 +94,7 @@ index 9d414d639b..e13a12137b 100644
          case 'U':
              force_share = true;
              break;
-@@ -5144,11 +5159,24 @@ static int img_dd(int argc, char **argv)
+@@ -5242,11 +5257,24 @@ 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 9d414d639b..e13a12137b 100644
      }
  
      if (dd.flags & C_OSIZE) {
-@@ -5303,6 +5331,7 @@ static int img_dd(int argc, char **argv)
+@@ -5401,6 +5429,7 @@ 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 c3e55ef9e9..0e32e6201f 100644
+index c6325cdcaa..7f817d622d 100644
 --- a/hw/core/machine-hmp-cmds.c
 +++ b/hw/core/machine-hmp-cmds.c
-@@ -169,7 +169,35 @@ void hmp_info_balloon(Monitor *mon, const QDict *qdict)
+@@ -179,7 +179,35 @@ void hmp_info_balloon(Monitor *mon, const QDict *qdict)
          return;
      }
  
@@ -59,10 +59,10 @@ index c3e55ef9e9..0e32e6201f 100644
      qapi_free_BalloonInfo(info);
  }
 diff --git a/hw/virtio/virtio-balloon.c b/hw/virtio/virtio-balloon.c
-index d004cf29d2..2660ed520b 100644
+index 2eb5a14fa2..aa2fd6c32f 100644
 --- a/hw/virtio/virtio-balloon.c
 +++ b/hw/virtio/virtio-balloon.c
-@@ -782,8 +782,37 @@ static uint64_t virtio_balloon_get_features(VirtIODevice *vdev, uint64_t f,
+@@ -795,8 +795,37 @@ 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 d004cf29d2..2660ed520b 100644
  
  static void virtio_balloon_to_target(void *opaque, ram_addr_t target)
 diff --git a/qapi/machine.json b/qapi/machine.json
-index a08b6576ca..5c9a4d55f4 100644
+index a6b8795b09..9f7ed0eaa0 100644
 --- a/qapi/machine.json
 +++ b/qapi/machine.json
-@@ -1063,9 +1063,29 @@
+@@ -1163,9 +1163,29 @@
  # @actual: the logical size of the VM in bytes Formula used:
  #     logical_vm_size = vm_ram_size - balloon_size
  #
@@ -138,10 +138,10 @@ index a08b6576ca..5c9a4d55f4 100644
  ##
  # @query-balloon:
 diff --git a/qapi/pragma.json b/qapi/pragma.json
-index 7f810b0e97..325e684411 100644
+index 023a2ef7bc..6aaa9cb975 100644
 --- a/qapi/pragma.json
 +++ b/qapi/pragma.json
-@@ -35,6 +35,7 @@
+@@ -81,6 +81,7 @@
      '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 3860a50c3b..40821e2317 100644
+index 1bc21b84a4..93fb4bc24a 100644
 --- a/hw/core/machine-qmp-cmds.c
 +++ b/hw/core/machine-qmp-cmds.c
-@@ -91,6 +91,12 @@ MachineInfoList *qmp_query_machines(Error **errp)
+@@ -91,6 +91,12 @@ MachineInfoList *qmp_query_machines(bool has_compat_props, bool compat_props,
          info->numa_mem_supported = mc->numa_mem_supported;
          info->deprecated = !!mc->deprecation_reason;
          info->acpi = !!object_class_property_find(OBJECT_CLASS(mc), "acpi");
@@ -26,14 +26,14 @@ index 3860a50c3b..40821e2317 100644
 +            info->is_current = true;
 +        }
 +
-         if (mc->default_cpu_type) {
-             info->default_cpu_type = g_strdup(mc->default_cpu_type);
+         if (default_cpu_type) {
+             info->default_cpu_type = g_strdup(default_cpu_type);
          }
 diff --git a/qapi/machine.json b/qapi/machine.json
-index 5c9a4d55f4..fbb61f18e4 100644
+index 9f7ed0eaa0..16366b774a 100644
 --- a/qapi/machine.json
 +++ b/qapi/machine.json
-@@ -139,6 +139,8 @@
+@@ -167,6 +167,8 @@
  #
  # @is-default: whether the machine is default
  #
@@ -42,7 +42,7 @@ index 5c9a4d55f4..fbb61f18e4 100644
  # @cpu-max: maximum number of CPUs supported by the machine type
  #     (since 1.5)
  #
-@@ -163,7 +165,7 @@
+@@ -199,7 +201,7 @@
  ##
  { 'struct': 'MachineInfo',
    'data': { 'name': 'str', '*alias': 'str',
@@ -50,4 +50,4 @@ index 5c9a4d55f4..fbb61f18e4 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 006616aa77..dfd1d3e36b 100644
+index c536d4e524..c2df48959b 100644
 --- a/qapi/ui.json
 +++ b/qapi/ui.json
-@@ -317,11 +317,14 @@
+@@ -312,11 +312,14 @@
  #
  # @channels: a list of @SpiceChannel for each active spice channel
  #
@@ -33,7 +33,7 @@ index 006616aa77..dfd1d3e36b 100644
    'if': 'CONFIG_SPICE' }
  
 diff --git a/ui/spice-core.c b/ui/spice-core.c
-index b20c25aee0..26baeb7846 100644
+index d523d00200..c76c224706 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

@@ -16,26 +16,26 @@ Additionally, allows tracking the current position from the outside
 Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
 Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
 ---
- migration/channel-savevm-async.c | 183 +++++++++++++++++++++++++++++++
+ migration/channel-savevm-async.c | 184 +++++++++++++++++++++++++++++++
  migration/channel-savevm-async.h |  51 +++++++++
  migration/meson.build            |   1 +
- 3 files changed, 235 insertions(+)
+ 3 files changed, 236 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..aab081ce07
+index 0000000000..e57ab2ae40
 --- /dev/null
 +++ b/migration/channel-savevm-async.c
-@@ -0,0 +1,183 @@
+@@ -0,0 +1,184 @@
 +/*
 + * QIO Channel implementation to be used by savevm-async QMP calls
 + */
 +#include "qemu/osdep.h"
 +#include "migration/channel-savevm-async.h"
 +#include "qapi/error.h"
-+#include "sysemu/block-backend.h"
++#include "system/block-backend.h"
 +#include "trace.h"
 +
 +QIOChannelSavevmAsync *
@@ -175,8 +175,9 @@ index 0000000000..aab081ce07
 +
 +static void
 +qio_channel_savevm_async_set_aio_fd_handler(QIOChannel *ioc,
-+                                            AioContext *ctx,
++                                            AioContext *read_ctx,
 +                                            IOHandler *io_read,
++                                            AioContext *write_ctx,
 +                                            IOHandler *io_write,
 +                                            void *opaque)
 +{
@@ -270,14 +271,14 @@ index 0000000000..17ae2cb261
 +
 +#endif /* QIO_CHANNEL_SAVEVM_ASYNC_H */
 diff --git a/migration/meson.build b/migration/meson.build
-index 1ae28523a1..37ddcb5d60 100644
+index 9aa48b290e..cf66c78681 100644
 --- a/migration/meson.build
 +++ b/migration/meson.build
-@@ -13,6 +13,7 @@ system_ss.add(files(
-   'block-dirty-bitmap.c',
+@@ -14,6 +14,7 @@ system_ss.add(files(
+   'block-active.c',
    'channel.c',
    'channel-block.c',
 +  'channel-savevm-async.c',
-   'dirtyrate.c',
-   'exec.c',
-   'fd.c',
+   'cpr.c',
+   'cpr-transfer.c',
+   'cpu-throttle.c',

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

@@ -27,28 +27,34 @@ 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 8.0]
+     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
+     improve setting state in savevm-end handler
+     improve runstate preservation
+     use dedicated iothread for state file to avoid deadlock, bug #6262]
 Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
 ---
  hmp-commands-info.hx         |  13 +
- hmp-commands.hx              |  17 ++
+ hmp-commands.hx              |  17 +
  include/migration/snapshot.h |   2 +
  include/monitor/hmp.h        |   3 +
  migration/meson.build        |   1 +
- migration/savevm-async.c     | 531 +++++++++++++++++++++++++++++++++++
+ migration/savevm-async.c     | 581 +++++++++++++++++++++++++++++++++++
  monitor/hmp-cmds.c           |  38 +++
- qapi/migration.json          |  34 +++
- qapi/misc.json               |  16 ++
+ qapi/migration.json          |  34 ++
+ qapi/misc.json               |  18 ++
  qemu-options.hx              |  12 +
- softmmu/vl.c                 |  10 +
- 11 files changed, 677 insertions(+)
+ system/vl.c                  |  10 +
+ 11 files changed, 729 insertions(+)
  create mode 100644 migration/savevm-async.c
 
 diff --git a/hmp-commands-info.hx b/hmp-commands-info.hx
-index f5b37eb74a..10fdd822e0 100644
+index c59cd6637b..d1a7b99add 100644
 --- a/hmp-commands-info.hx
 +++ b/hmp-commands-info.hx
-@@ -525,6 +525,19 @@ SRST
+@@ -512,6 +512,19 @@ SRST
      Show current migration parameters.
  ERST
  
@@ -69,10 +75,10 @@ index f5b37eb74a..10fdd822e0 100644
          .name       = "balloon",
          .args_type  = "",
 diff --git a/hmp-commands.hx b/hmp-commands.hx
-index 2cbd0f77a0..e352f86872 100644
+index 06746f0afc..0c7c6f2c16 100644
 --- a/hmp-commands.hx
 +++ b/hmp-commands.hx
-@@ -1865,3 +1865,20 @@ SRST
+@@ -1859,3 +1859,20 @@ SRST
    List event channels in the guest
  ERST
  #endif
@@ -94,18 +100,18 @@ index 2cbd0f77a0..e352f86872 100644
 +        .coroutine  = true,
 +    },
 diff --git a/include/migration/snapshot.h b/include/migration/snapshot.h
-index e72083b117..c846d37806 100644
+index 9e4dcaaa75..2581730d74 100644
 --- a/include/migration/snapshot.h
 +++ b/include/migration/snapshot.h
-@@ -61,4 +61,6 @@ bool delete_snapshot(const char *name,
-                     bool has_devices, strList *devices,
-                     Error **errp);
+@@ -68,4 +68,6 @@ bool delete_snapshot(const char *name,
+  */
+ void load_snapshot_resume(RunState state);
  
 +int load_snapshot_from_blockdev(const char *filename, Error **errp);
 +
  #endif
 diff --git a/include/monitor/hmp.h b/include/monitor/hmp.h
-index 13f9a2dedb..7a7def7530 100644
+index ae116d9804..2596cc2426 100644
 --- a/include/monitor/hmp.h
 +++ b/include/monitor/hmp.h
 @@ -28,6 +28,7 @@ void hmp_info_status(Monitor *mon, const QDict *qdict);
@@ -116,7 +122,7 @@ index 13f9a2dedb..7a7def7530 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);
-@@ -94,6 +95,8 @@ void hmp_closefd(Monitor *mon, const QDict *qdict);
+@@ -92,6 +93,8 @@ 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);
@@ -126,10 +132,10 @@ index 13f9a2dedb..7a7def7530 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 37ddcb5d60..07f6057acc 100644
+index cf66c78681..46e92249a1 100644
 --- a/migration/meson.build
 +++ b/migration/meson.build
-@@ -26,6 +26,7 @@ system_ss.add(files(
+@@ -33,6 +33,7 @@ system_ss.add(files(
    'options.c',
    'postcopy-ram.c',
    'savevm.c',
@@ -139,10 +145,10 @@ index 37ddcb5d60..07f6057acc 100644
    'threadinfo.c',
 diff --git a/migration/savevm-async.c b/migration/savevm-async.c
 new file mode 100644
-index 0000000000..e9fc18fb10
+index 0000000000..56e0fa6c69
 --- /dev/null
 +++ b/migration/savevm-async.c
-@@ -0,0 +1,531 @@
+@@ -0,0 +1,581 @@
 +#include "qemu/osdep.h"
 +#include "migration/channel-savevm-async.h"
 +#include "migration/migration.h"
@@ -153,21 +159,24 @@ index 0000000000..e9fc18fb10
 +#include "migration/global_state.h"
 +#include "migration/ram.h"
 +#include "migration/qemu-file.h"
-+#include "sysemu/sysemu.h"
-+#include "sysemu/runstate.h"
++#include "system/cpu-throttle.h"
++#include "system/system.h"
++#include "system/runstate.h"
 +#include "block/block.h"
-+#include "sysemu/block-backend.h"
++#include "system/block-backend.h"
 +#include "qapi/error.h"
 +#include "qapi/qmp/qerror.h"
-+#include "qapi/qmp/qdict.h"
++#include "qobject/qdict.h"
 +#include "qapi/qapi-commands-migration.h"
 +#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"
 +#include "qemu/yank.h"
++#include "system/iothread.h"
 +
 +/* #define DEBUG_SAVEVM_STATE */
 +
@@ -194,12 +203,13 @@ index 0000000000..e9fc18fb10
 +    int state;
 +    Error *error;
 +    Error *blocker;
-+    int saved_vm_running;
++    int vm_needs_start;
 +    QEMUFile *file;
 +    int64_t total_time;
 +    QEMUBH *finalize_bh;
 +    Coroutine *co;
 +    QemuCoSleep target_close_wait;
++    IOThread *iothread;
 +} snap_state;
 +
 +static bool savevm_aborted(void)
@@ -257,6 +267,7 @@ index 0000000000..e9fc18fb10
 +    }
 +
 +    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
@@ -265,7 +276,9 @@ index 0000000000..e9fc18fb10
 +            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);
 +        }
-+        blk_op_unblock_all(snap_state.target, snap_state.blocker);
++        if (target_bs) {
++            bdrv_op_unblock_all(target_bs, snap_state.blocker);
++        }
 +        error_free(snap_state.blocker);
 +        snap_state.blocker = NULL;
 +        blk_unref(snap_state.target);
@@ -289,7 +302,7 @@ index 0000000000..e9fc18fb10
 +    DPRINTF("save_snapshot_error: %s\n", msg);
 +
 +    if (!snap_state.error) {
-+        error_set(&snap_state.error, ERROR_CLASS_GENERIC_ERROR, "%s", msg);
++        error_setg(&snap_state.error, "%s", msg);
 +    }
 +
 +    g_free (msg);
@@ -300,7 +313,6 @@ index 0000000000..e9fc18fb10
 +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();
@@ -317,10 +329,9 @@ index 0000000000..e9fc18fb10
 +     * 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);
 +
++    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);
@@ -328,7 +339,7 @@ index 0000000000..e9fc18fb10
 +
 +    if (!aborted) {
 +        /* skip state saving if we aborted, snapshot will be invalid anyway */
-+        (void)qemu_savevm_state_complete_precopy(snap_state.file, false, false);
++        (void)qemu_savevm_state_complete_precopy(snap_state.file, false);
 +        ret = qemu_file_get_error(snap_state.file);
 +        if (ret < 0) {
 +            save_snapshot_error("qemu_savevm_state_complete_precopy error %d", ret);
@@ -344,6 +355,12 @@ index 0000000000..e9fc18fb10
 +        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();
@@ -361,9 +378,9 @@ index 0000000000..e9fc18fb10
 +        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",
@@ -392,16 +409,13 @@ index 0000000000..e9fc18fb10
 +        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.
 +         */
-+        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);
 +        }
-+        qemu_mutex_lock_iothread();
 +        pending_size = pend_precopy + pend_postcopy;
 +
 +        /*
@@ -441,21 +455,25 @@ index 0000000000..e9fc18fb10
 +     * so move there now and after every flush.
 +     */
 +    aio_co_reschedule_self(qemu_get_aio_context());
-+    for (bs = bdrv_first(&it); bs; bs = bdrv_next(&it)) {
++    bdrv_graph_co_rdlock();
++    bs = bdrv_first(&it);
++    bdrv_graph_co_rdunlock();
++    while (bs) {
 +        /* target has BDRV_O_NO_FLUSH, no sense calling bdrv_flush on it */
-+        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());
++        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());
++            }
 +        }
++        bdrv_graph_co_rdlock();
++        bs = bdrv_next(&it);
++        bdrv_graph_co_rdunlock();
 +    }
 +
 +    DPRINTF("timing: async flushing took %ld ms\n",
@@ -464,33 +482,33 @@ index 0000000000..e9fc18fb10
 +    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;
 +
 +    int bdrv_oflags = BDRV_O_RDWR | BDRV_O_RESIZE | BDRV_O_NO_FLUSH;
 +
 +    if (snap_state.state != SAVE_STATE_DONE) {
-+        error_set(errp, ERROR_CLASS_GENERIC_ERROR,
-+                  "VM snapshot already started\n");
++        error_setg(errp, "VM snapshot already started\n");
 +        return;
 +    }
 +
-+    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");
++    if (migration_is_running()) {
++        error_setg(errp, "There's a migration process in progress");
 +        return;
 +    }
 +
 +    /* initialize snapshot info */
-+    snap_state.saved_vm_running = runstate_is_running();
 +    snap_state.bs_pos = 0;
 +    snap_state.total_time = qemu_clock_get_ms(QEMU_CLOCK_REALTIME);
 +    snap_state.blocker = NULL;
@@ -502,13 +520,27 @@ index 0000000000..e9fc18fb10
 +    }
 +
 +    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;
++    }
++
++    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 */
@@ -517,8 +549,13 @@ index 0000000000..e9fc18fb10
 +    qdict_put_str(options, "driver", "raw");
 +    snap_state.target = blk_new_open(statefile, NULL, options, bdrv_oflags, &local_err);
 +    if (!snap_state.target) {
-+        error_set(errp, ERROR_CLASS_GENERIC_ERROR, "failed to open '%s'", statefile);
-+        goto restart;
++        error_setg(errp, "failed to open '%s'", statefile);
++        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 fail;
 +    }
 +
 +    QIOChannel *ioc = QIO_CHANNEL(qio_channel_savevm_async_new(snap_state.target,
@@ -526,8 +563,8 @@ index 0000000000..e9fc18fb10
 +    snap_state.file = qemu_file_new_output(ioc);
 +
 +    if (!snap_state.file) {
-+        error_set(errp, ERROR_CLASS_GENERIC_ERROR, "failed to open '%s'", statefile);
-+        goto restart;
++        error_setg(errp, "failed to open '%s'", statefile);
++        goto fail;
 +    }
 +
 +    /*
@@ -535,83 +572,66 @@ index 0000000000..e9fc18fb10
 +     * State is cleared in process_savevm_co, but has to be initialized
 +     * here (blocking main thread, from QMP) to avoid race conditions.
 +     */
-+    migrate_init(ms);
++    if (migrate_init(ms, errp) != 0) {
++        goto fail;
++    }
 +    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");
-+    blk_op_block_all(snap_state.target, snap_state.blocker);
++    bdrv_op_block_all(target_bs, 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);
-+    qemu_savevm_state_setup(snap_state.file);
-+    qemu_mutex_lock_iothread();
++    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");
++        goto fail;
++    }
 +
-+    /* Async processing from here on out happens in iohandler context, so let
-+     * the target bdrv have its home there.
-+     */
-+    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) {
++        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;
++    }
 +
-+    aio_co_schedule(iohandler_ctx, snap_state.co);
++    snap_state.co = qemu_coroutine_create(&process_savevm_co, NULL);
++    aio_co_schedule(snap_state.iothread->ctx, snap_state.co);
 +
 +    return;
 +
-+restart:
-+
++fail:
++    savevm_cleanup_iothread();
 +    save_snapshot_error("setup failed");
-+
-+    if (snap_state.saved_vm_running) {
-+        vm_start();
-+        snap_state.saved_vm_running = false;
-+    }
 +}
 +
-+void coroutine_fn qmp_savevm_end(Error **errp)
++static void coroutine_fn wait_for_close_co(void *opaque)
 +{
 +    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;
-+    }
-+
-+    /* 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");
 +    }
 +
++    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;
@@ -620,25 +640,59 @@ index 0000000000..e9fc18fb10
 +    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.vm_needs_start) {
++        vm_start();
++        snap_state.vm_needs_start = false;
++    }
++
++    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;
 +
-+    be = blk_new_open(filename, NULL, NULL, 0, &local_err);
++    options = qdict_new();
++    qdict_put_str(options, "driver", "raw");
++
++    be = blk_new_open(filename, NULL, options, 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");
-+    blk_op_block_all(be, blocker);
++    bdrv_op_block_all(bs, blocker);
 +
 +    /* restore the VM state */
 +    f = qemu_file_new_input(QIO_CHANNEL(qio_channel_savevm_async_new(be, &bs_pos)));
@@ -668,28 +722,30 @@ index 0000000000..e9fc18fb10
 +
 + the_end:
 +    if (be) {
-+        blk_op_unblock_all(be, blocker);
++        if (bs) {
++            bdrv_op_unblock_all(bs, blocker);
++        }
 +        error_free(blocker);
 +        blk_unref(be);
 +    }
 +    return ret;
 +}
 diff --git a/monitor/hmp-cmds.c b/monitor/hmp-cmds.c
-index 6c559b48c8..91be698308 100644
+index 7ded3378cf..bade2a4b92 100644
 --- a/monitor/hmp-cmds.c
 +++ b/monitor/hmp-cmds.c
-@@ -22,6 +22,7 @@
- #include "monitor/monitor-internal.h"
+@@ -24,6 +24,7 @@
  #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 "qapi/qmp/qerror.h"
-@@ -443,3 +444,40 @@ void hmp_info_mtree(Monitor *mon, const QDict *qdict)
- 
-     mtree_info(flatview, dispatch_tree, owner, disabled);
+ #include "qobject/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'\n", filename);
  }
+ #endif
 +
 +void hmp_savevm_start(Monitor *mon, const QDict *qdict)
 +{
@@ -728,10 +784,10 @@ index 6c559b48c8..91be698308 100644
 +    }
 +}
 diff --git a/qapi/migration.json b/qapi/migration.json
-index 8843e74b59..aca0ca1ac1 100644
+index 8b9c53595c..ff3479da65 100644
 --- a/qapi/migration.json
 +++ b/qapi/migration.json
-@@ -291,6 +291,40 @@
+@@ -279,6 +279,40 @@
             '*dirty-limit-throttle-time-per-round': 'uint64',
             '*dirty-limit-ring-full-time': 'uint64'} }
  
@@ -773,10 +829,10 @@ index 8843e74b59..aca0ca1ac1 100644
  # @query-migrate:
  #
 diff --git a/qapi/misc.json b/qapi/misc.json
-index cda2effa81..94a58bb0bf 100644
+index 559b66f201..7959e89c1e 100644
 --- a/qapi/misc.json
 +++ b/qapi/misc.json
-@@ -456,6 +456,22 @@
+@@ -454,6 +454,24 @@
  ##
  { 'command': 'query-fdsets', 'returns': ['FdsetInfo'] }
  
@@ -785,6 +841,8 @@ index cda2effa81..94a58bb0bf 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' } }
 +
@@ -794,16 +852,16 @@ index cda2effa81..94a58bb0bf 100644
 +# Resume VM after a snapshot.
 +#
 +##
-+{ 'command': 'savevm-end', 'coroutine': true }
++{ 'command': 'savevm-end' }
 +
  ##
  # @CommandLineParameterType:
  #
 diff --git a/qemu-options.hx b/qemu-options.hx
-index 8073f5edf5..dc1ececc9c 100644
+index dc694a99a3..defee0c06a 100644
 --- a/qemu-options.hx
 +++ b/qemu-options.hx
-@@ -4483,6 +4483,18 @@ SRST
+@@ -4862,6 +4862,18 @@ SRST
      Start right away with a saved state (``loadvm`` in monitor)
  ERST
  
@@ -822,11 +880,11 @@ index 8073f5edf5..dc1ececc9c 100644
  #ifndef _WIN32
  DEF("daemonize", 0, QEMU_OPTION_daemonize, \
      "-daemonize      daemonize QEMU after initializing\n", QEMU_ARCH_ALL)
-diff --git a/softmmu/vl.c b/softmmu/vl.c
-index ba6ad8a8df..ddeace306e 100644
---- a/softmmu/vl.c
-+++ b/softmmu/vl.c
-@@ -164,6 +164,7 @@ static const char *accelerators;
+diff --git a/system/vl.c b/system/vl.c
+index ec93988a03..9b36ace6b4 100644
+--- a/system/vl.c
++++ b/system/vl.c
+@@ -171,6 +171,7 @@ static const char *accelerators;
  static bool have_custom_ram_size;
  static const char *ram_memdev_id;
  static QDict *machine_opts_dict;
@@ -834,10 +892,10 @@ index ba6ad8a8df..ddeace306e 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;
-@@ -2647,6 +2648,12 @@ void qmp_x_exit_preconfig(Error **errp)
- 
-     if (loadvm) {
+@@ -2814,6 +2815,12 @@ void qmp_x_exit_preconfig(Error **errp)
+         RunState state = autostart ? RUN_STATE_RUNNING : runstate_get();
          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) {
@@ -847,7 +905,7 @@ index ba6ad8a8df..ddeace306e 100644
      }
      if (replay_mode != REPLAY_MODE_NONE) {
          replay_vmstate_init();
-@@ -3196,6 +3203,9 @@ void qemu_init(int argc, char **argv)
+@@ -3360,6 +3367,9 @@ 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

@@ -13,27 +13,27 @@ 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    | 49 +++++++++++++++++++++++++++-------------
+ migration/qemu-file.c    | 48 +++++++++++++++++++++++++++-------------
  migration/qemu-file.h    |  2 ++
- migration/savevm-async.c |  5 ++--
- 3 files changed, 38 insertions(+), 18 deletions(-)
+ migration/savevm-async.c |  5 +++--
+ 3 files changed, 38 insertions(+), 17 deletions(-)
 
 diff --git a/migration/qemu-file.c b/migration/qemu-file.c
-index 19c33c9985..e9ffff0f0a 100644
+index 1303a5bf58..6e2d58d5c0 100644
 --- a/migration/qemu-file.c
 +++ b/migration/qemu-file.c
-@@ -33,8 +33,8 @@
- #include "options.h"
- #include "qapi/error.h"
+@@ -34,8 +34,8 @@
+ #include "rdma.h"
+ #include "io/channel-file.h"
  
 -#define IO_BUF_SIZE 32768
 -#define MAX_IOV_SIZE MIN_CONST(IOV_MAX, 64)
 +#define DEFAULT_IO_BUF_SIZE 32768
 +#define MAX_IOV_SIZE MIN_CONST(IOV_MAX, 256)
  
- struct QEMUFile {
-     const QEMUFileHooks *hooks;
-@@ -46,7 +46,8 @@ struct QEMUFile {
+ typedef struct FdEntry {
+     QTAILQ_ENTRY(FdEntry) entry;
+@@ -48,7 +48,8 @@ struct QEMUFile {
  
      int buf_index;
      int buf_size; /* 0 when writing */
@@ -43,7 +43,7 @@ index 19c33c9985..e9ffff0f0a 100644
  
      DECLARE_BITMAP(may_free, MAX_IOV_SIZE);
      struct iovec iov[MAX_IOV_SIZE];
-@@ -100,7 +101,9 @@ int qemu_file_shutdown(QEMUFile *f)
+@@ -108,7 +109,9 @@ int qemu_file_shutdown(QEMUFile *f)
      return 0;
  }
  
@@ -54,16 +54,16 @@ index 19c33c9985..e9ffff0f0a 100644
  {
      QEMUFile *f;
  
-@@ -109,6 +112,8 @@ static QEMUFile *qemu_file_new_impl(QIOChannel *ioc, bool is_writable)
-     object_ref(ioc);
-     f->ioc = ioc;
+@@ -119,6 +122,8 @@ static QEMUFile *qemu_file_new_impl(QIOChannel *ioc, bool is_writable)
      f->is_writable = is_writable;
+     f->can_pass_fd = qio_channel_has_feature(ioc, QIO_CHANNEL_FEATURE_FD_PASS);
+     QTAILQ_INIT(&f->fds);
 +    f->buf_allocated_size = buffer_size;
 +    f->buf = malloc(buffer_size);
  
      return f;
  }
-@@ -119,17 +124,27 @@ static QEMUFile *qemu_file_new_impl(QIOChannel *ioc, bool is_writable)
+@@ -129,17 +134,27 @@ static QEMUFile *qemu_file_new_impl(QIOChannel *ioc, bool is_writable)
   */
  QEMUFile *qemu_file_get_return_path(QEMUFile *f)
  {
@@ -93,26 +93,27 @@ index 19c33c9985..e9ffff0f0a 100644
 +    return qemu_file_new_impl(ioc, false, buffer_size);
  }
  
- 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)
+ /*
+@@ -339,7 +354,7 @@ static ssize_t coroutine_mixed_fn qemu_fill_buffer(QEMUFile *f)
+     }
+ 
      do {
-         len = qio_channel_read(f->ioc,
-                                (char *)f->buf + pending,
--                               IO_BUF_SIZE - pending,
-+                               f->buf_allocated_size - pending,
-                                &local_error);
+-        struct iovec iov = { f->buf + pending, IO_BUF_SIZE - pending };
++        struct iovec iov = { f->buf + pending, f->buf_allocated_size - pending };
+         len = qio_channel_readv_full(f->ioc, &iov, 1, pfds, pnfd, 0,
+                                      &local_error);
          if (len == QIO_CHANNEL_ERR_BLOCK) {
-             if (qemu_in_coroutine()) {
-@@ -425,6 +440,8 @@ int qemu_fclose(QEMUFile *f)
+@@ -443,6 +458,9 @@ int qemu_fclose(QEMUFile *f)
+         g_free(fde);
      }
      g_clear_pointer(&f->ioc, object_unref);
- 
++
 +    free(f->buf);
 +
-     /* 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)
+     error_free(f->last_error_obj);
+     g_free(f);
+     trace_qemu_file_fclose();
+@@ -491,7 +509,7 @@ 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;
@@ -121,7 +122,7 @@ index 19c33c9985..e9ffff0f0a 100644
              qemu_fflush(f);
          }
      }
-@@ -504,7 +521,7 @@ void qemu_put_buffer(QEMUFile *f, const uint8_t *buf, size_t size)
+@@ -516,7 +534,7 @@ void qemu_put_buffer(QEMUFile *f, const uint8_t *buf, size_t size)
      }
  
      while (size > 0) {
@@ -130,7 +131,7 @@ index 19c33c9985..e9ffff0f0a 100644
          if (l > size) {
              l = size;
          }
-@@ -549,8 +566,8 @@ size_t coroutine_mixed_fn qemu_peek_buffer(QEMUFile *f, uint8_t **buf, size_t si
+@@ -662,8 +680,8 @@ 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));
@@ -141,7 +142,7 @@ index 19c33c9985..e9ffff0f0a 100644
  
      /* The 1st byte to read from */
      index = f->buf_index + offset;
-@@ -600,7 +617,7 @@ size_t coroutine_mixed_fn qemu_get_buffer(QEMUFile *f, uint8_t *buf, size_t size
+@@ -713,7 +731,7 @@ size_t coroutine_mixed_fn qemu_get_buffer(QEMUFile *f, uint8_t *buf, size_t size
          size_t res;
          uint8_t *src;
  
@@ -150,7 +151,7 @@ index 19c33c9985..e9ffff0f0a 100644
          if (res == 0) {
              return done;
          }
-@@ -634,7 +651,7 @@ size_t coroutine_mixed_fn qemu_get_buffer(QEMUFile *f, uint8_t *buf, size_t size
+@@ -747,7 +765,7 @@ 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)
  {
@@ -159,7 +160,7 @@ index 19c33c9985..e9ffff0f0a 100644
          size_t res;
          uint8_t *src = NULL;
  
-@@ -659,7 +676,7 @@ int coroutine_mixed_fn qemu_peek_byte(QEMUFile *f, int offset)
+@@ -772,7 +790,7 @@ int coroutine_mixed_fn qemu_peek_byte(QEMUFile *f, int offset)
      int index = f->buf_index + offset;
  
      assert(!qemu_file_is_writable(f));
@@ -168,34 +169,25 @@ index 19c33c9985..e9ffff0f0a 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 47015f5201..1312b7c903 100644
+index f5b9f430e0..0179b90698 100644
 --- a/migration/qemu-file.h
 +++ b/migration/qemu-file.h
-@@ -63,7 +63,9 @@ typedef struct QEMUFileHooks {
- } QEMUFileHooks;
+@@ -30,7 +30,9 @@
+ #include "io/channel.h"
  
  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);
  
+ G_DEFINE_AUTOPTR_CLEANUP_FUNC(QEMUFile, qemu_fclose)
 diff --git a/migration/savevm-async.c b/migration/savevm-async.c
-index e9fc18fb10..80624fada8 100644
+index 56e0fa6c69..730b815494 100644
 --- a/migration/savevm-async.c
 +++ b/migration/savevm-async.c
-@@ -378,7 +378,7 @@ void qmp_savevm_start(const char *statefile, Error **errp)
+@@ -409,7 +409,7 @@ 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));
@@ -203,9 +195,9 @@ index e9fc18fb10..80624fada8 100644
 +    snap_state.file = qemu_file_new_output_sized(ioc, 4 * 1024 * 1024);
  
      if (!snap_state.file) {
-         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);
+         error_setg(errp, "failed to open '%s'", statefile);
+@@ -544,7 +544,8 @@ int load_snapshot_from_blockdev(const char *filename, Error **errp)
+     bdrv_op_block_all(bs, 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

@@ -4,32 +4,34 @@ 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]
+[FE: adapt to changed function signatures
+     adhere to block graph lock requirements
+     use dedicated function to open file child]
 Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
 ---
  block/meson.build |   1 +
- block/zeroinit.c  | 200 ++++++++++++++++++++++++++++++++++++++++++++++
- 2 files changed, 201 insertions(+)
+ block/zeroinit.c  | 207 ++++++++++++++++++++++++++++++++++++++++++++++
+ 2 files changed, 208 insertions(+)
  create mode 100644 block/zeroinit.c
 
 diff --git a/block/meson.build b/block/meson.build
-index 529fc172c6..1833c71ce9 100644
+index 34b1b2a306..a21d9a5411 100644
 --- a/block/meson.build
 +++ b/block/meson.build
-@@ -40,6 +40,7 @@ block_ss.add(files(
-   'throttle-groups.c',
+@@ -39,6 +39,7 @@ block_ss.add(files(
    'throttle.c',
+   'throttle-groups.c',
    'write-threshold.c',
 +  'zeroinit.c',
- ), zstd, zlib, gnutls)
+ ), zstd, zlib)
  
  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..1257342724
+index 0000000000..f9d513db15
 --- /dev/null
 +++ b/block/zeroinit.c
-@@ -0,0 +1,200 @@
+@@ -0,0 +1,207 @@
 +/*
 + * Filter to fake a zero-initialized block device.
 + *
@@ -44,8 +46,9 @@ index 0000000000..1257342724
 +#include "qapi/error.h"
 +#include "block/block_int.h"
 +#include "block/block-io.h"
-+#include "qapi/qmp/qdict.h"
-+#include "qapi/qmp/qstring.h"
++#include "block/graph-lock.h"
++#include "qobject/qdict.h"
++#include "qobject/qstring.h"
 +#include "qemu/cutils.h"
 +#include "qemu/option.h"
 +#include "qemu/module.h"
@@ -109,12 +112,9 @@ index 0000000000..1257342724
 +    }
 +
 +    /* Open the raw file */
-+    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;
++    ret = bdrv_open_file_child(qemu_opt_get(opts, "x-next"), options, "next",
++                               bs, &local_err);
++    if (ret < 0) {
 +        error_propagate(errp, local_err);
 +        goto fail;
 +    }
@@ -125,7 +125,9 @@ index 0000000000..1257342724
 +    ret = 0;
 +fail:
 +    if (ret < 0) {
++        bdrv_graph_wrlock();
 +        bdrv_unref_child(bs, bs->file);
++        bdrv_graph_wrunlock();
 +    }
 +    qemu_opts_del(opts);
 +    return ret;
@@ -137,19 +139,22 @@ index 0000000000..1257342724
 +    (void)s;
 +}
 +
-+static coroutine_fn int64_t zeroinit_co_getlength(BlockDriverState *bs)
++static coroutine_fn int64_t GRAPH_RDLOCK
++zeroinit_co_getlength(BlockDriverState *bs)
 +{
 +    return bdrv_co_getlength(bs->file->bs);
 +}
 +
-+static int coroutine_fn zeroinit_co_preadv(BlockDriverState *bs,
-+    int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags)
++static int coroutine_fn GRAPH_RDLOCK
++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 zeroinit_co_pwrite_zeroes(BlockDriverState *bs, int64_t offset,
-+                                                 int64_t bytes, BdrvRequestFlags flags)
++static int coroutine_fn GRAPH_RDLOCK
++zeroinit_co_pwrite_zeroes(BlockDriverState *bs, int64_t offset, int64_t bytes,
++                          BdrvRequestFlags flags)
 +{
 +    BDRVZeroinitState *s = bs->opaque;
 +    if (offset >= s->extents)
@@ -157,8 +162,9 @@ index 0000000000..1257342724
 +    return bdrv_pwrite_zeroes(bs->file, offset, bytes, flags);
 +}
 +
-+static int coroutine_fn zeroinit_co_pwritev(BlockDriverState *bs,
-+    int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags)
++static int coroutine_fn GRAPH_RDLOCK
++zeroinit_co_pwritev(BlockDriverState *bs, int64_t offset, int64_t bytes,
++                    QEMUIOVector *qiov, BdrvRequestFlags flags)
 +{
 +    BDRVZeroinitState *s = bs->opaque;
 +    int64_t extents = offset + bytes;
@@ -167,32 +173,35 @@ index 0000000000..1257342724
 +    return bdrv_co_pwritev(bs->file, offset, bytes, qiov, flags);
 +}
 +
-+static coroutine_fn int zeroinit_co_flush(BlockDriverState *bs)
++static coroutine_fn int GRAPH_RDLOCK
++zeroinit_co_flush(BlockDriverState *bs)
 +{
 +    return bdrv_co_flush(bs->file->bs);
 +}
 +
-+static int zeroinit_has_zero_init(BlockDriverState *bs)
++static int GRAPH_RDLOCK
++zeroinit_has_zero_init(BlockDriverState *bs)
 +{
 +    BDRVZeroinitState *s = bs->opaque;
 +    return s->has_zero_init;
 +}
 +
-+static int coroutine_fn zeroinit_co_pdiscard(BlockDriverState *bs,
-+                                             int64_t offset, int64_t bytes)
++static int coroutine_fn GRAPH_RDLOCK
++zeroinit_co_pdiscard(BlockDriverState *bs, int64_t offset, int64_t bytes)
 +{
 +    return bdrv_co_pdiscard(bs->file, offset, bytes);
 +}
 +
-+static int zeroinit_co_truncate(BlockDriverState *bs, int64_t offset,
-+                                _Bool exact, PreallocMode prealloc,
-+                                BdrvRequestFlags req_flags, Error **errp)
++static int GRAPH_RDLOCK
++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 zeroinit_co_get_info(BlockDriverState *bs,
-+                                             BlockDriverInfo *bdi)
++static coroutine_fn int GRAPH_RDLOCK
++zeroinit_co_get_info(BlockDriverState *bs, BlockDriverInfo *bdi)
 +{
 +    return bdrv_co_get_info(bs->file->bs, bdi);
 +}
@@ -203,7 +212,7 @@ index 0000000000..1257342724
 +    .instance_size                    = sizeof(BDRVZeroinitState),
 +
 +    .bdrv_parse_filename              = zeroinit_parse_filename,
-+    .bdrv_file_open                   = zeroinit_open,
++    .bdrv_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 +++
- softmmu/vl.c    | 8 ++++++++
+ system/vl.c     | 8 ++++++++
  2 files changed, 11 insertions(+)
 
 diff --git a/qemu-options.hx b/qemu-options.hx
-index dc1ececc9c..848d2dfdd1 100644
+index defee0c06a..fb980a05cf 100644
 --- a/qemu-options.hx
 +++ b/qemu-options.hx
-@@ -1197,6 +1197,9 @@ legacy PC, they are not recommended for modern configurations.
+@@ -1280,6 +1280,9 @@ legacy PC, they are not recommended for modern configurations.
  
  ERST
  
@@ -27,11 +27,11 @@ index dc1ececc9c..848d2dfdd1 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/softmmu/vl.c b/softmmu/vl.c
-index ddeace306e..3ee90b3b94 100644
---- a/softmmu/vl.c
-+++ b/softmmu/vl.c
-@@ -2683,6 +2683,7 @@ void qemu_init(int argc, char **argv)
+diff --git a/system/vl.c b/system/vl.c
+index 9b36ace6b4..452742ab58 100644
+--- a/system/vl.c
++++ b/system/vl.c
+@@ -2854,6 +2854,7 @@ void qemu_init(int argc, char **argv)
      MachineClass *machine_class;
      bool userconfig = true;
      FILE *vmstate_dump_file = NULL;
@@ -39,7 +39,7 @@ index ddeace306e..3ee90b3b94 100644
  
      qemu_add_opts(&qemu_drive_opts);
      qemu_add_drive_opts(&qemu_legacy_drive_opts);
-@@ -3308,6 +3309,13 @@ void qemu_init(int argc, char **argv)
+@@ -3472,6 +3473,13 @@ void qemu_init(int argc, char **argv)
                  machine_parse_property_opt(qemu_find_opts("smp-opts"),
                                             "smp", optarg);
                  break;
@@ -50,6 +50,6 @@ index ddeace306e..3ee90b3b94 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 4a34f03047..59b917e50c 100644
+index 2a3e878c4d..efbed1aea3 100644
 --- a/hw/intc/apic_common.c
 +++ b/hw/intc/apic_common.c
-@@ -252,6 +252,15 @@ static void apic_reset_common(DeviceState *dev)
+@@ -263,6 +263,15 @@ 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 |  3 ++-
- 2 files changed, 42 insertions(+), 20 deletions(-)
+ qapi/block-core.json |  7 +++++-
+ 2 files changed, 46 insertions(+), 20 deletions(-)
 
 diff --git a/block/file-posix.c b/block/file-posix.c
-index ca551baa42..8b3b83e9d4 100644
+index cfa0b832ba..d5c28cccc9 100644
 --- a/block/file-posix.c
 +++ b/block/file-posix.c
-@@ -2873,6 +2873,7 @@ raw_co_create(BlockdevCreateOptions *options, Error **errp)
+@@ -2897,6 +2897,7 @@ raw_co_create(BlockdevCreateOptions *options, Error **errp)
      int fd;
      uint64_t perm, shared;
      int result = 0;
@@ -24,7 +24,7 @@ index ca551baa42..8b3b83e9d4 100644
  
      /* Validate options and set default values */
      assert(options->driver == BLOCKDEV_DRIVER_FILE);
-@@ -2913,19 +2914,22 @@ raw_co_create(BlockdevCreateOptions *options, Error **errp)
+@@ -2937,19 +2938,22 @@ 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 ca551baa42..8b3b83e9d4 100644
      }
  
      /* Clear the file by truncating it to 0 */
-@@ -2979,13 +2983,15 @@ raw_co_create(BlockdevCreateOptions *options, Error **errp)
+@@ -3003,13 +3007,15 @@ raw_co_create(BlockdevCreateOptions *options, Error **errp)
      }
  
  out_unlock:
@@ -82,7 +82,7 @@ index ca551baa42..8b3b83e9d4 100644
      }
  
  out_close:
-@@ -3009,6 +3015,7 @@ raw_co_create_opts(BlockDriver *drv, const char *filename,
+@@ -3033,6 +3039,7 @@ raw_co_create_opts(BlockDriver *drv, const char *filename,
      PreallocMode prealloc;
      char *buf = NULL;
      Error *local_err = NULL;
@@ -90,7 +90,7 @@ index ca551baa42..8b3b83e9d4 100644
  
      /* Skip file: protocol prefix */
      strstart(filename, "file:", &filename);
-@@ -3031,6 +3038,18 @@ raw_co_create_opts(BlockDriver *drv, const char *filename,
+@@ -3055,6 +3062,18 @@ raw_co_create_opts(BlockDriver *drv, const char *filename,
          return -EINVAL;
      }
  
@@ -109,7 +109,7 @@ index ca551baa42..8b3b83e9d4 100644
      options = (BlockdevCreateOptions) {
          .driver     = BLOCKDEV_DRIVER_FILE,
          .u.file     = {
-@@ -3042,6 +3061,8 @@ raw_co_create_opts(BlockDriver *drv, const char *filename,
+@@ -3066,6 +3085,8 @@ 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,
@@ -119,10 +119,21 @@ index ca551baa42..8b3b83e9d4 100644
      };
      return raw_co_create(&options, errp);
 diff --git a/qapi/block-core.json b/qapi/block-core.json
-index a5cea82139..bb471c078d 100644
+index 0e5f148d30..1c05413916 100644
 --- a/qapi/block-core.json
 +++ b/qapi/block-core.json
-@@ -4880,7 +4880,8 @@
+@@ -5016,6 +5016,10 @@
+ # @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',
+@@ -5023,7 +5027,8 @@
              '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 589c9524f8..2505dd658a 100644
+index f093e256e9..78f1c8e3c8 100644
 --- a/monitor/qmp.c
 +++ b/monitor/qmp.c
-@@ -536,8 +536,7 @@ void monitor_init_qmp(Chardev *chr, bool pretty, Error **errp)
+@@ -534,8 +534,7 @@ 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 f0d35c6401..1427983543 100644
+index 63c6ef93d2..9a34017e5a 100644
 --- a/hw/core/machine.c
 +++ b/hw/core/machine.c
-@@ -148,7 +148,8 @@ GlobalProperty hw_compat_4_0[] = {
+@@ -193,7 +193,8 @@ 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          |  4 +++-
- softmmu/vl.c               | 25 +++++++++++++++++++++++++
- 4 files changed, 35 insertions(+), 1 deletion(-)
+ qapi/machine.json          |  3 +++
+ system/vl.c                | 24 ++++++++++++++++++++++++
+ 4 files changed, 34 insertions(+)
 
 diff --git a/hw/core/machine-qmp-cmds.c b/hw/core/machine-qmp-cmds.c
-index 40821e2317..ee93ddd69a 100644
+index 93fb4bc24a..b9999423b4 100644
 --- a/hw/core/machine-qmp-cmds.c
 +++ b/hw/core/machine-qmp-cmds.c
-@@ -95,6 +95,11 @@ MachineInfoList *qmp_query_machines(Error **errp)
+@@ -95,6 +95,11 @@ MachineInfoList *qmp_query_machines(bool has_compat_props, bool compat_props,
          if (strcmp(mc->name, MACHINE_GET_CLASS(current_machine)->name) == 0) {
              info->has_is_current = true;
              info->is_current = true;
@@ -35,67 +35,66 @@ index 40821e2317..ee93ddd69a 100644
 +            }
          }
  
-         if (mc->default_cpu_type) {
+         if (default_cpu_type) {
 diff --git a/include/hw/boards.h b/include/hw/boards.h
-index ed83360198..f8b88cd86a 100644
+index f22b2e7fc7..8ada4d5832 100644
 --- a/include/hw/boards.h
 +++ b/include/hw/boards.h
-@@ -235,6 +235,8 @@ struct MachineClass {
+@@ -271,6 +271,8 @@ struct MachineClass {
      const char *desc;
      const char *deprecation_reason;
  
 +    const char *pve_version;
 +
      void (*init)(MachineState *state);
-     void (*reset)(MachineState *state, ShutdownCause reason);
+     void (*reset)(MachineState *state, ResetType type);
      void (*wakeup)(MachineState *state);
 diff --git a/qapi/machine.json b/qapi/machine.json
-index fbb61f18e4..7da3c519ba 100644
+index 16366b774a..12cfd3f260 100644
 --- a/qapi/machine.json
 +++ b/qapi/machine.json
-@@ -161,6 +161,8 @@
+@@ -189,6 +189,8 @@
  #
  # @acpi: machine type supports ACPI (since 8.0)
  #
 +# @pve-version: custom PVE version suffix specified as 'machine+pveN'
 +#
- # Since: 1.2
- ##
- { 'struct': 'MachineInfo',
-@@ -168,7 +170,7 @@
-             '*is-default': 'bool', '*is-current': 'bool', 'cpu-max': 'int',
+ # @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 @@
              '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', '*pve-version': 'str' } }
+             '*default-ram-id': 'str', 'acpi': 'bool',
++            '*pve-version': 'str',
+             '*compat-props': { 'type': ['CompatProperty'],
+                                'features': ['unstable'] } } }
  
- ##
- # @query-machines:
-diff --git a/softmmu/vl.c b/softmmu/vl.c
-index 3ee90b3b94..4b6d0b82fd 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)
+diff --git a/system/vl.c b/system/vl.c
+index 452742ab58..c3707b2412 100644
+--- a/system/vl.c
++++ b/system/vl.c
+@@ -1674,6 +1674,7 @@ static MachineClass *select_machine(QDict *qdict, Error **errp)
  {
-     const char *optarg = qdict_get_try_str(qdict, "type");
+     ERRP_GUARD();
+     const char *machine_type = qdict_get_try_str(qdict, "type");
 +    const char *pvever = qdict_get_try_str(qdict, "pvever");
-     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)
-         }
-     }
+     g_autoptr(GSList) machines = object_class_get_list(TYPE_MACHINE, false);
+     MachineClass *machine_class = NULL;
  
-+    if (machine_class) {
+@@ -1693,7 +1694,11 @@ static MachineClass *select_machine(QDict *qdict, Error **errp)
+     if (!machine_class) {
+         error_append_hint(errp,
+                           "Use -machine help to list supported machines\n");
++    } else {
 +        machine_class->pve_version = g_strdup(pvever);
 +        qdict_del(qdict, "pvever");
-+    }
+     }
 +
-     g_slist_free(machines);
-     if (local_err) {
-         error_append_hint(&local_err, "Use -machine help to list supported machines\n");
-@@ -3250,12 +3256,31 @@ void qemu_init(int argc, char **argv)
+     return machine_class;
+ }
+ 
+@@ -3414,12 +3419,31 @@ 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 db3791f4d1..39410dcf8d 100644
+index 79652bf57b..cc747e9163 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 db3791f4d1..39410dcf8d 100644
      if (s->sync_mode == MIRROR_SYNC_MODE_TOP) {
          int64_t offset = 0;
          int64_t count;
-@@ -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,
+@@ -502,6 +500,8 @@ BlockJob *backup_job_create(const char *job_id, BlockDriverState *bs,
                         &error_abort);
+     bdrv_graph_wrunlock();
  
 +    backup_init_bcs_bitmap(job);
 +

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

@@ -10,28 +10,29 @@ skip reads.
 
 Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
 [FE: improvements during create
-     allow partial restore]
+     allow partial restore
+     allow specifying disk formats for create operation]
 Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
 ---
  block/meson.build |   2 +
  meson.build       |   5 +
- vma-reader.c      | 867 ++++++++++++++++++++++++++++++++++++++++++++
- vma-writer.c      | 818 +++++++++++++++++++++++++++++++++++++++++
- vma.c             | 900 ++++++++++++++++++++++++++++++++++++++++++++++
+ vma-reader.c      | 867 ++++++++++++++++++++++++++++++++++++++++++
+ vma-writer.c      | 816 ++++++++++++++++++++++++++++++++++++++++
+ vma.c             | 941 ++++++++++++++++++++++++++++++++++++++++++++++
  vma.h             | 150 ++++++++
- 6 files changed, 2742 insertions(+)
+ 6 files changed, 2781 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 1833c71ce9..59b71ba9f3 100644
+index a21d9a5411..1373612c10 100644
 --- a/block/meson.build
 +++ b/block/meson.build
-@@ -43,6 +43,8 @@ block_ss.add(files(
+@@ -42,6 +42,8 @@ block_ss.add(files(
    'zeroinit.c',
- ), zstd, zlib, gnutls)
+ ), zstd, zlib)
  
 +block_ss.add(files('../vma-writer.c'), libuuid)
 +
@@ -39,10 +40,10 @@ index 1833c71ce9..59b71ba9f3 100644
  system_ss.add(files('block-ram-registrar.c'))
  
 diff --git a/meson.build b/meson.build
-index a9c4f28247..cd95530d3b 100644
+index 8ec796d835..680ab48b9b 100644
 --- a/meson.build
 +++ b/meson.build
-@@ -1778,6 +1778,8 @@ endif
+@@ -2161,6 +2161,8 @@ endif
  
  has_gettid = cc.has_function('gettid')
  
@@ -51,19 +52,19 @@ index a9c4f28247..cd95530d3b 100644
  # libselinux
  selinux = dependency('libselinux',
                       required: get_option('selinux'),
-@@ -3908,6 +3910,9 @@ if have_tools
-                dependencies: [blockdev, qemuutil, gnutls, selinux],
+@@ -4367,6 +4369,9 @@ if have_tools
+                dependencies: [blockdev, qemuutil, selinux],
                 install: true)
  
 +  vma = executable('vma', files('vma.c', 'vma-reader.c') + genh,
-+                   dependencies: [authz, block, crypto, io, qom], install: true)
++                   dependencies: [authz, block, crypto, io, qemuutil, qom], install: true)
 +
    subdir('storage-daemon')
-   subdir('contrib/rdmacm-mux')
-   subdir('contrib/elf2dmp')
+ 
+   foreach exe: [ 'qemu-img', 'qemu-io', 'qemu-nbd', 'qemu-storage-daemon']
 diff --git a/vma-reader.c b/vma-reader.c
 new file mode 100644
-index 0000000000..81a891c6b1
+index 0000000000..1888b21851
 --- /dev/null
 +++ b/vma-reader.c
 @@ -0,0 +1,867 @@
@@ -88,7 +89,8 @@ index 0000000000..81a891c6b1
 +#include "qemu/ratelimit.h"
 +#include "vma.h"
 +#include "block/block.h"
-+#include "sysemu/block-backend.h"
++#include "block/graph-lock.h"
++#include "system/block-backend.h"
 +
 +static unsigned char zero_vma_block[VMA_BLOCK_SIZE];
 +
@@ -380,7 +382,6 @@ index 0000000000..81a891c6b1
 +    }
 +
 +
-+    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);
@@ -388,7 +389,6 @@ index 0000000000..81a891c6b1
 +        const char *devname =  get_header_str(vmar, devname_ptr);
 +
 +        if (size && devname) {
-+            count++;
 +            vmar->devinfo[i].size = size;
 +            vmar->devinfo[i].devname = devname;
 +
@@ -600,8 +600,10 @@ index 0000000000..81a891c6b1
 +    } 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;
 +        }
 +    }
@@ -881,8 +883,7 @@ index 0000000000..81a891c6b1
 +
 +            int64_t cluster_num, end;
 +
-+            end = (vmar->devinfo[i].size + VMA_CLUSTER_SIZE - 1) /
-+                VMA_CLUSTER_SIZE;
++            end = DIV_ROUND_UP(vmar->devinfo[i].size, VMA_CLUSTER_SIZE);
 +
 +            for (cluster_num = 0; cluster_num < end; cluster_num++) {
 +                if (!vma_reader_get_bitmap(rstate, cluster_num)) {
@@ -936,10 +937,10 @@ index 0000000000..81a891c6b1
 +
 diff --git a/vma-writer.c b/vma-writer.c
 new file mode 100644
-index 0000000000..126b296647
+index 0000000000..3f489092df
 --- /dev/null
 +++ b/vma-writer.c
-@@ -0,0 +1,818 @@
+@@ -0,0 +1,816 @@
 +/*
 + * VMA: Virtual Machine Archive
 + *
@@ -1133,8 +1134,7 @@ index 0000000000..126b296647
 +    vmaw->stream_info[n].devname = g_strdup(devname);
 +    vmaw->stream_info[n].size = size;
 +
-+    vmaw->stream_info[n].cluster_count = (size + VMA_CLUSTER_SIZE - 1) /
-+        VMA_CLUSTER_SIZE;
++    vmaw->stream_info[n].cluster_count = DIV_ROUND_UP(size, VMA_CLUSTER_SIZE);
 +
 +    vmaw->stream_count = n;
 +
@@ -1514,17 +1514,16 @@ index 0000000000..126b296647
 +    int i;
 +
 +    g_assert(vmaw != NULL);
++    g_assert(status != NULL);
 +
-+    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);
++    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);
++
 +    status->closed = vmaw->closed;
 +
 +    return vmaw->status;
@@ -1760,10 +1759,10 @@ index 0000000000..126b296647
 +}
 diff --git a/vma.c b/vma.c
 new file mode 100644
-index 0000000000..347f6283ca
+index 0000000000..0e990b5e30
 --- /dev/null
 +++ b/vma.c
-@@ -0,0 +1,900 @@
+@@ -0,0 +1,941 @@
 +/*
 + * VMA: Virtual Machine Archive
 + *
@@ -1786,8 +1785,8 @@ index 0000000000..347f6283ca
 +#include "qemu/main-loop.h"
 +#include "qemu/cutils.h"
 +#include "qemu/memalign.h"
-+#include "qapi/qmp/qdict.h"
-+#include "sysemu/block-backend.h"
++#include "qobject/qdict.h"
++#include "system/block-backend.h"
 +
 +static void help(void)
 +{
@@ -1795,8 +1794,8 @@ index 0000000000..347f6283ca
 +        "usage: vma command [command options]\n"
 +        "\n"
 +        "vma list <filename>\n"
-+        "vma config <filename> [-c config]\n"
-+        "vma create <filename> [-c config] pathname ...\n"
++        "vma config <filename> [-c <config>]\n"
++        "vma create <filename> [-c <config>] [-d format=<format>:<device name>=<path> [-d ...]]\n"
 +        "vma extract <filename> [-d <drive-list>] [-r <fifo>] <targetdir>\n"
 +        "vma verify <filename> [-v]\n"
 +        ;
@@ -2076,17 +2075,17 @@ index 0000000000..347f6283ca
 +                        inbuf);
 +            }
 +
-+            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;
++            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;
 +
-+            g_hash_table_insert(devmap, map->devname, map);
++            g_hash_table_insert(devmap, restore_map->devname, restore_map);
 +
 +        };
 +    }
@@ -2385,14 +2384,16 @@ index 0000000000..347f6283ca
 +
 +static int create_archive(int argc, char **argv)
 +{
-+    int i, c;
++    int 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:");
++        c = getopt(argc, argv, "hvc:d:");
 +        if (c == -1) {
 +            break;
 +        }
@@ -2404,6 +2405,9 @@ index 0000000000..347f6283ca
 +        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;
@@ -2449,16 +2453,48 @@ index 0000000000..347f6283ca
 +        l = g_list_next(l);
 +    }
 +
-+    int devcount = 0;
++    /*
++     * 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) {
-+        const char *path = argv[optind++];
++        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;
 +        char *devname = NULL;
-+        path = extract_devname(path, &devname, devcount++);
++        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++);
 +
 +        Error *errp = NULL;
 +        BlockBackend *target;
 +
-+        target = blk_new_open(path, NULL, NULL, 0, &errp);
++        target = blk_new_open(path, NULL, options, 0, &errp);
 +        if (!target) {
 +            unlink(archivename);
 +            g_error("bdrv_open '%s' failed - %s", path, error_get_pretty(errp));
@@ -2480,6 +2516,8 @@ index 0000000000..347f6283ca
 +        // 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;
@@ -2543,6 +2581,7 @@ index 0000000000..347f6283ca
 +    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) {
@@ -2560,6 +2599,7 @@ index 0000000000..347f6283ca
 +
 +    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

@@ -22,7 +22,7 @@ Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
 
 diff --git a/block/backup-dump.c b/block/backup-dump.c
 new file mode 100644
-index 0000000000..e46abf1070
+index 0000000000..354593bc10
 --- /dev/null
 +++ b/block/backup-dump.c
 @@ -0,0 +1,172 @@
@@ -38,7 +38,7 @@ index 0000000000..e46abf1070
 +
 +#include "qemu/osdep.h"
 +
-+#include "qapi/qmp/qdict.h"
++#include "qobject/qdict.h"
 +#include "qom/object_interfaces.h"
 +#include "block/block_int.h"
 +
@@ -199,7 +199,7 @@ index 0000000000..e46abf1070
 +    return bs;
 +}
 diff --git a/block/backup.c b/block/backup.c
-index 39410dcf8d..af87fa6aa9 100644
+index cc747e9163..6f7c45f922 100644
 --- a/block/backup.c
 +++ b/block/backup.c
 @@ -29,28 +29,6 @@
@@ -231,7 +231,7 @@ index 39410dcf8d..af87fa6aa9 100644
  static const BlockJobDriver backup_job_driver;
  
  static void backup_cleanup_sync_bitmap(BackupBlockJob *job, int ret)
-@@ -457,6 +435,14 @@ BlockJob *backup_job_create(const char *job_id, BlockDriverState *bs,
+@@ -462,6 +440,14 @@ BlockJob *backup_job_create(const char *job_id, BlockDriverState *bs,
      }
  
      cluster_size = block_copy_cluster_size(bcs);
@@ -247,7 +247,7 @@ index 39410dcf8d..af87fa6aa9 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 59b71ba9f3..6fde9f7dcd 100644
+index 1373612c10..6278c4af0f 100644
 --- a/block/meson.build
 +++ b/block/meson.build
 @@ -4,6 +4,7 @@ block_ss.add(files(
@@ -255,11 +255,11 @@ index 59b71ba9f3..6fde9f7dcd 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 74195c3004..0a0339eee4 100644
+index ebb4e56a50..e717a74e5f 100644
 --- a/include/block/block_int-common.h
 +++ b/include/block/block_int-common.h
 @@ -26,6 +26,7 @@
@@ -312,10 +312,10 @@ index 74195c3004..0a0339eee4 100644
      BDRV_TRACKED_READ,
      BDRV_TRACKED_WRITE,
 diff --git a/job.c b/job.c
-index 72d57f0934..93e22d180b 100644
+index 0653bc2ba6..b981070ee8 100644
 --- a/job.c
 +++ b/job.c
-@@ -330,7 +330,8 @@ static bool job_started_locked(Job *job)
+@@ -337,7 +337,8 @@ 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 e502787dd8..963cf2bef5 100644
+index a5a04155ea..562cc7eaec 100644
 --- a/include/qemu/job.h
 +++ b/include/qemu/job.h
-@@ -381,6 +381,18 @@ void job_unlock(void);
+@@ -362,6 +362,18 @@ void job_unlock(void);
   */
  JobTxn *job_txn_new(void);
  
@@ -34,10 +34,10 @@ index e502787dd8..963cf2bef5 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 93e22d180b..2b31f1e14f 100644
+index b981070ee8..f4646866ec 100644
 --- a/job.c
 +++ b/job.c
-@@ -93,6 +93,8 @@ struct JobTxn {
+@@ -94,6 +94,8 @@ struct JobTxn {
  
      /* Reference count */
      int refcnt;
@@ -46,7 +46,7 @@ index 93e22d180b..2b31f1e14f 100644
  };
  
  void job_lock(void)
-@@ -118,6 +120,25 @@ JobTxn *job_txn_new(void)
+@@ -119,6 +121,25 @@ JobTxn *job_txn_new(void)
      return txn;
  }
  
@@ -72,7 +72,7 @@ index 93e22d180b..2b31f1e14f 100644
  /* Called with job_mutex held. */
  static void job_txn_ref_locked(JobTxn *txn)
  {
-@@ -1057,6 +1078,12 @@ static void job_completed_txn_success_locked(Job *job)
+@@ -1048,6 +1069,12 @@ 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 93e22d180b..2b31f1e14f 100644
              return;
          }
          assert(other_job->ret == 0);
-@@ -1268,6 +1295,13 @@ int job_finish_sync_locked(Job *job,
+@@ -1259,6 +1286,13 @@ int job_finish_sync_locked(Job *job,
          return -EBUSY;
      }
  

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

@@ -94,21 +94,21 @@ Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
  monitor/hmp-cmds.c             |   72 +++
  proxmox-backup-client.c        |  146 +++++
  proxmox-backup-client.h        |   60 ++
- pve-backup.c                   | 1089 ++++++++++++++++++++++++++++++++
- qapi/block-core.json           |  229 +++++++
+ pve-backup.c                   | 1096 ++++++++++++++++++++++++++++++++
+ qapi/block-core.json           |  233 +++++++
  qapi/common.json               |   14 +
  qapi/machine.json              |   16 +-
- 14 files changed, 1704 insertions(+), 14 deletions(-)
+ 14 files changed, 1715 insertions(+), 14 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/meson.build b/block/meson.build
-index 6fde9f7dcd..6d468f89e5 100644
+index 6278c4af0f..d1b16e40e9 100644
 --- a/block/meson.build
 +++ b/block/meson.build
-@@ -45,6 +45,11 @@ block_ss.add(files(
- ), zstd, zlib, gnutls)
+@@ -44,6 +44,11 @@ block_ss.add(files(
+ ), zstd, zlib)
  
  block_ss.add(files('../vma-writer.c'), libuuid)
 +block_ss.add(files(
@@ -120,10 +120,10 @@ index 6fde9f7dcd..6d468f89e5 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 ca2599de44..6efe28cef5 100644
+index 6919a49bf5..4f30f99644 100644
 --- a/block/monitor/block-hmp-cmds.c
 +++ b/block/monitor/block-hmp-cmds.c
-@@ -1029,3 +1029,42 @@ void hmp_change_medium(Monitor *mon, const char *device, const char *target,
+@@ -1010,3 +1010,42 @@ 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 +167,7 @@ index ca2599de44..6efe28cef5 100644
 +    hmp_handle_error(mon, error);
 +}
 diff --git a/blockdev.c b/blockdev.c
-index cd5f205ad1..7793143d76 100644
+index 158ac9314b..17de5d2ae4 100644
 --- a/blockdev.c
 +++ b/blockdev.c
 @@ -37,6 +37,7 @@
@@ -179,10 +179,10 @@ index cd5f205ad1..7793143d76 100644
  #include "monitor/monitor.h"
  #include "qemu/error-report.h"
 diff --git a/hmp-commands-info.hx b/hmp-commands-info.hx
-index 10fdd822e0..15937793c1 100644
+index d1a7b99add..af588145ff 100644
 --- a/hmp-commands-info.hx
 +++ b/hmp-commands-info.hx
-@@ -471,6 +471,20 @@ SRST
+@@ -458,6 +458,20 @@ SRST
      Show the current VM UUID.
  ERST
  
@@ -204,7 +204,7 @@ index 10fdd822e0..15937793c1 100644
      {
          .name       = "usernet",
 diff --git a/hmp-commands.hx b/hmp-commands.hx
-index e352f86872..0c8b6725fb 100644
+index 0c7c6f2c16..bf8315f226 100644
 --- a/hmp-commands.hx
 +++ b/hmp-commands.hx
 @@ -101,6 +101,35 @@ ERST
@@ -244,7 +244,7 @@ index e352f86872..0c8b6725fb 100644
  
      {
 diff --git a/include/monitor/hmp.h b/include/monitor/hmp.h
-index 7a7def7530..cba7afe70c 100644
+index 2596cc2426..9dda91d65a 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 +255,7 @@ index 7a7def7530..cba7afe70c 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);
-@@ -84,6 +85,8 @@ void hmp_change_vnc(Monitor *mon, const char *device, const char *target,
+@@ -82,6 +83,8 @@ 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 +265,10 @@ index 7a7def7530..cba7afe70c 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 cd95530d3b..d53976d621 100644
+index 680ab48b9b..1f74de1d93 100644
 --- a/meson.build
 +++ b/meson.build
-@@ -1779,6 +1779,7 @@ endif
+@@ -2162,6 +2162,7 @@ endif
  has_gettid = cc.has_function('gettid')
  
  libuuid = cc.find_library('uuid', required: true)
@@ -277,18 +277,18 @@ index cd95530d3b..d53976d621 100644
  # libselinux
  selinux = dependency('libselinux',
 diff --git a/monitor/hmp-cmds.c b/monitor/hmp-cmds.c
-index 91be698308..5b9c231a4c 100644
+index bade2a4b92..77a2fa409a 100644
 --- a/monitor/hmp-cmds.c
 +++ b/monitor/hmp-cmds.c
-@@ -21,6 +21,7 @@
+@@ -22,6 +22,7 @@
  #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"
- #include "qapi/qapi-commands-misc.h"
-@@ -144,6 +145,77 @@ void hmp_sync_profile(Monitor *mon, const QDict *qdict)
+@@ -119,6 +120,77 @@ void hmp_sync_profile(Monitor *mon, const QDict *qdict)
      }
  }
  
@@ -586,21 +586,23 @@ index 0000000000..8cbf645b2c
 +#endif /* PROXMOX_BACKUP_CLIENT_H */
 diff --git a/pve-backup.c b/pve-backup.c
 new file mode 100644
-index 0000000000..ae3d137e12
+index 0000000000..e931cb9203
 --- /dev/null
 +++ b/pve-backup.c
-@@ -0,0 +1,1089 @@
+@@ -0,0 +1,1096 @@
 +#include "proxmox-backup-client.h"
 +#include "vma.h"
 +
 +#include "qemu/osdep.h"
 +#include "qemu/module.h"
-+#include "sysemu/block-backend.h"
-+#include "sysemu/blockdev.h"
++#include "system/block-backend.h"
++#include "system/blockdev.h"
 +#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 "qobject/qdict.h"
 +#include "qapi/qmp/qerror.h"
 +#include "qemu/cutils.h"
 +
@@ -625,7 +627,6 @@ index 0000000000..ae3d137e12
 + * ---end-bad-example--
 + *
 + * ==> Always use CoMutext inside coroutines.
-+ * ==> Never acquire/release AioContext withing coroutines (because that use QemuRecMutex)
 + *
 + */
 +
@@ -677,8 +678,8 @@ index 0000000000..ae3d137e12
 +    size_t size;
 +    uint64_t block_size;
 +    uint8_t dev_id;
++    char* device_name;
 +    int completed_ret; // INT_MAX if not completed
-+    char targetfile[PATH_MAX];
 +    BdrvDirtyBitmap *bitmap;
 +    BlockDriverState *target;
 +    BlockJob *job;
@@ -911,6 +912,8 @@ index 0000000000..ae3d137e12
 +    }
 +
 +    di->bs = NULL;
++    g_free(di->device_name);
++    di->device_name = NULL;
 +
 +    assert(di->target == NULL);
 +
@@ -928,13 +931,6 @@ index 0000000000..ae3d137e12
 +        }
 +    }
 +
-+    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);
 +
@@ -954,6 +950,16 @@ index 0000000000..ae3d137e12
 +    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.
@@ -1074,8 +1080,7 @@ index 0000000000..ae3d137e12
 +}
 +
 +/*
-+ * backup_job_create can *not* be run from a coroutine (and requires an
-+ * acquired AioContext), so this can't either.
++ * 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.
 + */
 +static void create_backup_jobs_bh(void *opaque) {
@@ -1108,9 +1113,6 @@ index 0000000000..ae3d137e12
 +            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(
@@ -1121,8 +1123,6 @@ index 0000000000..ae3d137e12
 +
 +        bdrv_drained_end(di->bs);
 +
-+        aio_context_release(aio_context);
-+
 +        di->job = job;
 +        if (job) {
 +            WITH_JOB_LOCK_GUARD() {
@@ -1178,7 +1178,7 @@ index 0000000000..ae3d137e12
 + * case of an error, errp will be set, but the returned value might still be a
 + * list.
 + */
-+static GList coroutine_fn *get_device_info(
++static GList coroutine_fn GRAPH_RDLOCK *get_device_info(
 +    const char *devlist,
 +    Error **errp)
 +{
@@ -1198,11 +1198,13 @@ index 0000000000..ae3d137e12
 +            }
 +            BlockDriverState *bs = blk_bs(blk);
 +            if (!bdrv_co_is_inserted(bs)) {
-+                error_setg(errp, QERR_DEVICE_HAS_NO_MEDIUM, *d);
++                error_setg(errp, "Device '%s' has no medium", *d);
 +                goto err;
 +            }
 +            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);
 +            d++;
 +        }
@@ -1216,6 +1218,7 @@ index 0000000000..ae3d137e12
 +
 +            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);
 +        }
 +    }
@@ -1279,7 +1282,9 @@ index 0000000000..ae3d137e12
 +    /* 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);
 +        goto err;
@@ -1292,7 +1297,11 @@ index 0000000000..ae3d137e12
 +    while (l) {
 +        PVEBackupDevInfo *di = (PVEBackupDevInfo *)l->data;
 +        l = g_list_next(l);
-+        if (bdrv_op_is_blocked(di->bs, BLOCK_OP_TYPE_BACKUP_SOURCE, errp)) {
++
++        bdrv_graph_co_rdlock();
++        bool blocked = bdrv_op_is_blocked(di->bs, BLOCK_OP_TYPE_BACKUP_SOURCE, errp);
++        bdrv_graph_co_rdunlock();
++        if (blocked) {
 +            goto err;
 +        }
 +
@@ -1376,7 +1385,6 @@ index 0000000000..ae3d137e12
 +
 +            di->block_size = dump_cb_block_size;
 +
-+            const char *devname = bdrv_get_device_name(di->bs);
 +            PBSBitmapAction action = PBS_BITMAP_ACTION_NOT_USED;
 +            size_t dirty = di->size;
 +
@@ -1391,7 +1399,8 @@ index 0000000000..ae3d137e12
 +                    }
 +                    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) {
@@ -1415,7 +1424,8 @@ index 0000000000..ae3d137e12
 +                }
 +            }
 +
-+            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;
 +            }
@@ -1427,7 +1437,7 @@ index 0000000000..ae3d137e12
 +            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;
@@ -1436,9 +1446,7 @@ index 0000000000..ae3d137e12
 +    } else if (format == BACKUP_FORMAT_VMA) {
 +        vmaw = vma_writer_create(backup_file, uuid, &local_err);
 +        if (!vmaw) {
-+            if (local_err) {
-+                error_propagate(errp, local_err);
-+            }
++            error_propagate(errp, local_err);
 +            goto err_mutex;
 +        }
 +
@@ -1452,8 +1460,7 @@ index 0000000000..ae3d137e12
 +                goto err_mutex;
 +            }
 +
-+            const char *devname = bdrv_get_device_name(di->bs);
-+            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");
@@ -1564,9 +1571,9 @@ index 0000000000..ae3d137e12
 +            bdrv_co_unref(di->target);
 +        }
 +
-+        if (di->targetfile[0]) {
-+            unlink(di->targetfile);
-+        }
++        g_free(di->device_name);
++        di->device_name = NULL;
++
 +        g_free(di);
 +    }
 +    g_list_free(di_list);
@@ -1680,10 +1687,10 @@ index 0000000000..ae3d137e12
 +    return ret;
 +}
 diff --git a/qapi/block-core.json b/qapi/block-core.json
-index bb471c078d..1b8462a51b 100644
+index 1c05413916..dd98e03bf1 100644
 --- a/qapi/block-core.json
 +++ b/qapi/block-core.json
-@@ -839,6 +839,235 @@
+@@ -855,6 +855,239 @@
  { 'command': 'query-block', 'returns': ['BlockInfo'],
    'allow-preconfig': true }
  
@@ -1752,6 +1759,9 @@ index bb471c078d..1b8462a51b 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 ',', ';'
@@ -1819,9 +1829,7 @@ index bb471c078d..1b8462a51b 100644
 +#
 +# Cancel the current executing backup process.
 +#
-+# Returns: nothing on success
-+#
-+# Notes: This command succeeds even if there is no backup process running.
++# .. note:: This command succeeds even if there is no backup process running.
 +#
 +##
 +{ 'command': 'backup-cancel', 'coroutine': true }
@@ -1844,6 +1852,9 @@ index bb471c078d..1b8462a51b 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',
@@ -1920,13 +1931,13 @@ index bb471c078d..1b8462a51b 100644
  # @BlockDeviceTimedStats:
  #
 diff --git a/qapi/common.json b/qapi/common.json
-index 6fed9cde1a..630a2a8f9a 100644
+index 0e3a0bbbfb..554680c716 100644
 --- a/qapi/common.json
 +++ b/qapi/common.json
-@@ -207,3 +207,17 @@
+@@ -226,3 +226,17 @@
  ##
- { 'struct': 'HumanReadableText',
-   'data': { 'human-readable-text': 'str' } }
+ { 'enum': 'EndianMode',
+   'data': [ 'unspecified', 'little', 'big' ] }
 +
 +##
 +# @UuidInfo:
@@ -1937,12 +1948,12 @@ index 6fed9cde1a..630a2a8f9a 100644
 +#
 +# Since: 0.14.0
 +#
-+# Notes: If no UUID was specified for the guest, a null UUID is
++# .. note:: 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 7da3c519ba..888457f810 100644
+index 12cfd3f260..a8abdb42a3 100644
 --- a/qapi/machine.json
 +++ b/qapi/machine.json
 @@ -4,6 +4,8 @@
@@ -1954,7 +1965,7 @@ index 7da3c519ba..888457f810 100644
  ##
  # = Machines
  ##
-@@ -230,20 +232,6 @@
+@@ -302,20 +304,6 @@
  ##
  { 'command': 'query-target', 'returns': 'TargetInfo' }
  
@@ -1967,8 +1978,8 @@ index 7da3c519ba..888457f810 100644
 -#
 -# Since: 0.14
 -#
--# Notes: If no UUID was specified for the guest, a null UUID is
--#     returned.
+-# .. note:: If no UUID was specified for the guest, the nil UUID (all
+-#    zeroes) is returned.
 -##
 -{ 'struct': 'UuidInfo', 'data': {'UUID': 'str'} }
 -

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

@@ -14,23 +14,23 @@ Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
  create mode 100644 pbs-restore.c
 
 diff --git a/meson.build b/meson.build
-index d53976d621..c3330310d9 100644
+index 1f74de1d93..8508aab9c9 100644
 --- a/meson.build
 +++ b/meson.build
-@@ -3914,6 +3914,10 @@ if have_tools
+@@ -4373,6 +4373,10 @@ if have_tools
    vma = executable('vma', files('vma.c', 'vma-reader.c') + genh,
-                    dependencies: [authz, block, crypto, io, qom], install: true)
+                    dependencies: [authz, block, crypto, io, qemuutil, qom], install: true)
  
 +  pbs_restore = executable('pbs-restore', files('pbs-restore.c') + genh,
-+                  dependencies: [authz, block, crypto, io, qom,
++                  dependencies: [authz, block, crypto, io, qemuutil, qom,
 +                    libproxmox_backup_qemu], install: true)
 +
    subdir('storage-daemon')
-   subdir('contrib/rdmacm-mux')
-   subdir('contrib/elf2dmp')
+ 
+   foreach exe: [ 'qemu-img', 'qemu-io', 'qemu-nbd', 'qemu-storage-daemon']
 diff --git a/pbs-restore.c b/pbs-restore.c
 new file mode 100644
-index 0000000000..f03d9bab8d
+index 0000000000..f165f418af
 --- /dev/null
 +++ b/pbs-restore.c
 @@ -0,0 +1,236 @@
@@ -57,8 +57,8 @@ index 0000000000..f03d9bab8d
 +#include "qemu/main-loop.h"
 +#include "qemu/cutils.h"
 +#include "qapi/error.h"
-+#include "qapi/qmp/qdict.h"
-+#include "sysemu/block-backend.h"
++#include "qobject/qdict.h"
++#include "system/block-backend.h"
 +
 +#include <proxmox-backup-qemu.h>
 +

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

@@ -14,43 +14,41 @@ 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    |   3 +
- block/pbs.c          | 305 +++++++++++++++++++++++++++++++++++++++++++
- configure            |   9 ++
+ block/meson.build    |   2 +
+ block/pbs.c          | 306 +++++++++++++++++++++++++++++++++++++++++++
  meson.build          |   2 +-
- qapi/block-core.json |  13 ++
+ qapi/block-core.json |  29 ++++
  qapi/pragma.json     |   1 +
- 6 files changed, 332 insertions(+), 1 deletion(-)
+ 5 files changed, 339 insertions(+), 1 deletion(-)
  create mode 100644 block/pbs.c
 
 diff --git a/block/meson.build b/block/meson.build
-index 6d468f89e5..becc99ac4e 100644
+index d1b16e40e9..d243372c41 100644
 --- a/block/meson.build
 +++ b/block/meson.build
-@@ -50,6 +50,9 @@ block_ss.add(files(
+@@ -49,6 +49,8 @@ block_ss.add(files(
    '../pve-backup.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)
++block_ss.add(files('pbs.c'), 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..a2211e0f3b
+index 0000000000..3e41421716
 --- /dev/null
 +++ b/block/pbs.c
-@@ -0,0 +1,305 @@
+@@ -0,0 +1,306 @@
 +/*
 + * Proxmox Backup Server read-only block driver
 + */
 +
 +#include "qemu/osdep.h"
 +#include "qapi/error.h"
-+#include "qapi/qmp/qdict.h"
-+#include "qapi/qmp/qstring.h"
++#include "qobject/qdict.h"
++#include "qobject/qstring.h"
 +#include "qemu/module.h"
 +#include "qemu/option.h"
 +#include "qemu/cutils.h"
@@ -70,7 +68,7 @@ index 0000000000..a2211e0f3b
 +
 +typedef struct {
 +    ProxmoxRestoreHandle *conn;
-+    char aid;
++    uint8_t aid;
 +    int64_t length;
 +
 +    char *repository;
@@ -203,12 +201,18 @@ index 0000000000..a2211e0f3b
 +    }
 +
 +    /* acquire handle and length */
-+    s->aid = proxmox_restore_open_image(s->conn, s->archive, &pbs_error);
-+    if (s->aid < 0) {
++    ret = proxmox_restore_open_image(s->conn, s->archive, &pbs_error);
++    if (ret < 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);
@@ -219,12 +223,6 @@ index 0000000000..a2211e0f3b
 +    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);
@@ -234,7 +232,8 @@ index 0000000000..a2211e0f3b
 +    proxmox_restore_disconnect(s->conn);
 +}
 +
-+static coroutine_fn int64_t pbs_co_getlength(BlockDriverState *bs)
++static coroutine_fn int64_t GRAPH_RDLOCK
++pbs_co_getlength(BlockDriverState *bs)
 +{
 +    BDRVPBSState *s = bs->opaque;
 +    return s->length;
@@ -251,9 +250,9 @@ index 0000000000..a2211e0f3b
 +    aio_co_schedule(rcb->ctx, rcb->co);
 +}
 +
-+static coroutine_fn int pbs_co_preadv(BlockDriverState *bs,
-+                                      int64_t offset, int64_t bytes,
-+                                      QEMUIOVector *qiov, BdrvRequestFlags flags)
++static coroutine_fn int GRAPH_RDLOCK
++pbs_co_preadv(BlockDriverState *bs, int64_t offset, int64_t bytes,
++              QEMUIOVector *qiov, BdrvRequestFlags flags)
 +{
 +    BDRVPBSState *s = bs->opaque;
 +    int ret;
@@ -298,16 +297,17 @@ index 0000000000..a2211e0f3b
 +    return 0;
 +}
 +
-+static coroutine_fn int pbs_co_pwritev(BlockDriverState *bs,
-+                                       int64_t offset, int64_t bytes,
-+                                       QEMUIOVector *qiov, BdrvRequestFlags flags)
++static coroutine_fn int GRAPH_RDLOCK
++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 pbs_refresh_filename(BlockDriverState *bs)
++static void GRAPH_RDLOCK
++pbs_refresh_filename(BlockDriverState *bs)
 +{
 +    BDRVPBSState *s = bs->opaque;
 +    if (s->namespace) {
@@ -330,7 +330,6 @@ index 0000000000..a2211e0f3b
 +
 +    .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,
@@ -348,65 +347,24 @@ index 0000000000..a2211e0f3b
 +}
 +
 +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 c3330310d9..cbfc9a43fb 100644
+index 8508aab9c9..9c39f54f86 100644
 --- a/meson.build
 +++ b/meson.build
-@@ -4319,7 +4319,7 @@ summary_info += {'bzip2 support':     libbzip2}
- summary_info += {'lzfse support':     liblzfse}
- summary_info += {'zstd support':      zstd}
+@@ -4838,7 +4838,7 @@ summary_info += {'Query Processing Library support': qpl}
+ summary_info += {'UADK Library support': uadk}
+ summary_info += {'qatzip support':    qatzip}
  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 += {'libudev':           libudev}
+ summary_info += {'libcbor support':   libcbor}
 diff --git a/qapi/block-core.json b/qapi/block-core.json
-index 1b8462a51b..d67a6d448a 100644
+index dd98e03bf1..0c3ebfa74e 100644
 --- a/qapi/block-core.json
 +++ b/qapi/block-core.json
-@@ -3396,6 +3396,7 @@
+@@ -3470,6 +3470,7 @@
              'parallels', 'preallocate', 'qcow', 'qcow2', 'qed', 'quorum',
              'raw', 'rbd',
              { 'name': 'replication', 'if': 'CONFIG_REPLICATION' },
@@ -414,7 +372,7 @@ index 1b8462a51b..d67a6d448a 100644
              'ssh', 'throttle', 'vdi', 'vhdx',
              { 'name': 'virtio-blk-vfio-pci', 'if': 'CONFIG_BLKIO' },
              { 'name': 'virtio-blk-vhost-user', 'if': 'CONFIG_BLKIO' },
-@@ -3482,6 +3483,17 @@
+@@ -3556,6 +3557,33 @@
  { 'struct': 'BlockdevOptionsNull',
    'data': { '*size': 'int', '*latency-ns': 'uint64', '*read-zeroes': 'bool' } }
  
@@ -423,6 +381,22 @@ index 1b8462a51b..d67a6d448a 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',
@@ -432,7 +406,7 @@ index 1b8462a51b..d67a6d448a 100644
  ##
  # @BlockdevOptionsNVMe:
  #
-@@ -4890,6 +4902,7 @@
+@@ -5003,6 +5031,7 @@
        'nfs':        'BlockdevOptionsNfs',
        'null-aio':   'BlockdevOptionsNull',
        'null-co':    'BlockdevOptionsNull',
@@ -441,10 +415,10 @@ index 1b8462a51b..d67a6d448a 100644
        'nvme-io_uring': { 'type': 'BlockdevOptionsNvmeIoUring',
                           'if': 'CONFIG_BLKIO' },
 diff --git a/qapi/pragma.json b/qapi/pragma.json
-index 325e684411..b6079f6a0e 100644
+index 6aaa9cb975..e9c595c4ba 100644
 --- a/qapi/pragma.json
 +++ b/qapi/pragma.json
-@@ -45,6 +45,7 @@
+@@ -91,6 +91,7 @@
          '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 | 2 ++
+ meson.build | 3 ++-
  os-posix.c  | 7 +++++--
- 2 files changed, 7 insertions(+), 2 deletions(-)
+ 2 files changed, 7 insertions(+), 3 deletions(-)
 
 diff --git a/meson.build b/meson.build
-index cbfc9a43fb..8206270272 100644
+index 9c39f54f86..60af7fa723 100644
 --- a/meson.build
 +++ b/meson.build
-@@ -1779,6 +1779,7 @@ endif
+@@ -2162,6 +2162,7 @@ endif
  has_gettid = cc.has_function('gettid')
  
  libuuid = cc.find_library('uuid', required: true)
@@ -25,16 +25,17 @@ index cbfc9a43fb..8206270272 100644
  libproxmox_backup_qemu = cc.find_library('proxmox_backup_qemu', required: true)
  
  # libselinux
-@@ -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')])
+@@ -3766,7 +3767,7 @@ 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
  endif
  
 diff --git a/os-posix.c b/os-posix.c
-index 0cc1d991b1..f33d9901cf 100644
+index 52925c23d3..84b96d3da9 100644
 --- a/os-posix.c
 +++ b/os-posix.c
 @@ -29,6 +29,8 @@
@@ -44,9 +45,9 @@ index 0cc1d991b1..f33d9901cf 100644
 +#include <systemd/sd-journal.h>
 +#include <syslog.h>
  
- /* Needed early for CONFIG_BSD etc. */
- #include "net/slirp.h"
-@@ -332,9 +334,10 @@ void os_setup_post(void)
+ #include "qemu/error-report.h"
+ #include "qemu/log.h"
+@@ -306,9 +308,10 @@ void os_setup_post(void)
  
          dup2(fd, 0);
          dup2(fd, 1);

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

@@ -26,46 +26,52 @@ 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 7dcc0b5c2c..4c940b2475 100644
+index 8fd36eba1d..e963e93e71 100644
 --- a/include/migration/misc.h
 +++ b/include/migration/misc.h
-@@ -77,4 +77,7 @@ bool migration_in_bg_snapshot(void);
- /* migration/block-dirty-bitmap.c */
- void dirty_bitmap_mig_init(void);
+@@ -140,4 +140,7 @@ bool multifd_device_state_save_thread_should_exit(void);
+ void multifd_abort_device_state_save_threads(void);
+ bool multifd_join_device_state_save_threads(void);
  
 +/* migration/pbs-state.c */
 +void pbs_state_mig_init(void);
 +
  #endif
 diff --git a/migration/meson.build b/migration/meson.build
-index 07f6057acc..343994d891 100644
+index 46e92249a1..fb3fd7d7d0 100644
 --- a/migration/meson.build
 +++ b/migration/meson.build
-@@ -7,7 +7,9 @@ migration_files = files(
-   'vmstate.c',
+@@ -8,6 +8,7 @@ migration_files = files(
    'qemu-file.c',
    'yank_functions.c',
-+  'pbs-state.c',
  )
 +system_ss.add(libproxmox_backup_qemu)
  
  system_ss.add(files(
    'block-dirty-bitmap.c',
+@@ -31,6 +32,7 @@ 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 7a4c8beb5d..0a955a2a18 100644
+index d46e776e24..2f3430f440 100644
 --- a/migration/migration.c
 +++ b/migration/migration.c
-@@ -162,6 +162,7 @@ void migration_object_init(void)
-     blk_mig_init();
-     ram_mig_init();
-     dirty_bitmap_mig_init();
+@@ -319,6 +319,7 @@ void migration_object_init(void)
+ 
+     /* Initialize cpu throttle timers */
+     cpu_throttle_init();
 +    pbs_state_mig_init();
  }
  
- void migration_cancel(const Error *error)
+ typedef struct {
 diff --git a/migration/pbs-state.c b/migration/pbs-state.c
 new file mode 100644
-index 0000000000..887e998b9e
+index 0000000000..a97187e4d7
 --- /dev/null
 +++ b/migration/pbs-state.c
 @@ -0,0 +1,104 @@
@@ -114,7 +120,7 @@ index 0000000000..887e998b9e
 +}
 +
 +/* serialize PBS state and send to target via f, called on source */
-+static int pbs_state_save_setup(QEMUFile *f, void *opaque)
++static int pbs_state_save_setup(QEMUFile *f, void *opaque, Error **errp)
 +{
 +    size_t buf_size;
 +    uint8_t *buf = proxmox_export_state(&buf_size);
@@ -174,10 +180,10 @@ index 0000000000..887e998b9e
 +                         NULL);
 +}
 diff --git a/pve-backup.c b/pve-backup.c
-index ae3d137e12..e6b17b797e 100644
+index e931cb9203..366b015589 100644
 --- a/pve-backup.c
 +++ b/pve-backup.c
-@@ -1082,6 +1082,7 @@ ProxmoxSupportStatus *qmp_query_proxmox_support(Error **errp)
+@@ -1089,6 +1089,7 @@ 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;
@@ -186,10 +192,10 @@ index ae3d137e12..e6b17b797e 100644
      ret->pbs_masterkey = true;
      ret->backup_max_workers = true;
 diff --git a/qapi/block-core.json b/qapi/block-core.json
-index d67a6d448a..09de550c95 100644
+index 0c3ebfa74e..6838187607 100644
 --- a/qapi/block-core.json
 +++ b/qapi/block-core.json
-@@ -991,6 +991,11 @@
+@@ -1008,6 +1008,11 @@
  # @pbs-dirty-bitmap-savevm: True if 'dirty-bitmaps' migration capability can
  #                           safely be set for savevm-async.
  #
@@ -201,7 +207,7 @@ index d67a6d448a..09de550c95 100644
  # @pbs-masterkey: True if the QMP backup call supports the 'master_keyfile'
  #                 parameter.
  #
-@@ -1001,6 +1006,7 @@
+@@ -1021,6 +1026,7 @@
    '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

@@ -15,18 +15,22 @@ transferred.
 Signed-off-by: Stefan Reiter <s.reiter@proxmox.com>
 Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
 ---
- migration/block-dirty-bitmap.c | 2 +-
- 1 file changed, 1 insertion(+), 1 deletion(-)
+ migration/block-dirty-bitmap.c | 6 +++++-
+ 1 file changed, 5 insertions(+), 1 deletion(-)
 
 diff --git a/migration/block-dirty-bitmap.c b/migration/block-dirty-bitmap.c
-index e1ae3b7316..285dd1d148 100644
+index f2c352d4a7..931a8481e9 100644
 --- a/migration/block-dirty-bitmap.c
 +++ b/migration/block-dirty-bitmap.c
-@@ -540,7 +540,7 @@ static int add_bitmaps_to_list(DBMSaveState *s, BlockDriverState *bs,
+@@ -539,7 +539,11 @@ static int add_bitmaps_to_list(DBMSaveState *s, BlockDriverState *bs,
+         }
  
-         if (bdrv_dirty_bitmap_check(bitmap, BDRV_BITMAP_DEFAULT, &local_err)) {
-             error_report_err(local_err);
+         if (bdrv_dirty_bitmap_check(bitmap, BDRV_BITMAP_DEFAULT, errp)) {
 -            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 34f97ab646..398782963d 100644
+index 2f0f4dac09..b523137cff 100644
 --- a/block/iscsi.c
 +++ b/block/iscsi.c
-@@ -1391,12 +1391,42 @@ static char *get_initiator_name(QemuOpts *opts)
+@@ -1392,12 +1392,42 @@ 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 e522bbdec5..afed72db55 100644
+index 999d9e56d4..e187cd1262 100644
 --- a/block/stream.c
 +++ b/block/stream.c
 @@ -27,7 +27,7 @@ enum {

debian/patches/pve/0038-block-add-alloc-track-driver.patch

@@ -19,27 +19,34 @@ 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.
 
-If 'auto-remove' is set, alloc-track will automatically detach itself
-once the backing image is removed. It will be replaced by 'file'.
+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.
 
 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]
+     avoid premature break during read
+     adhere to block graph lock requirements
+     avoid superfluous child permission update]
 Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
 ---
- block/alloc-track.c | 352 ++++++++++++++++++++++++++++++++++++++++++++
+ block/alloc-track.c | 343 ++++++++++++++++++++++++++++++++++++++++++++
  block/meson.build   |   1 +
- 2 files changed, 353 insertions(+)
+ block/stream.c      |  34 +++++
+ 3 files changed, 378 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..b75d7c6460
+index 0000000000..718aaabf2a
 --- /dev/null
 +++ b/block/alloc-track.c
-@@ -0,0 +1,352 @@
+@@ -0,0 +1,343 @@
 +/*
 + * 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
@@ -56,25 +63,20 @@ index 0000000000..b75d7c6460
 +#include "qapi/error.h"
 +#include "block/block_int.h"
 +#include "block/dirty-bitmap.h"
-+#include "qapi/qmp/qdict.h"
-+#include "qapi/qmp/qstring.h"
++#include "block/graph-lock.h"
++#include "qobject/qdict.h"
++#include "qobject/qstring.h"
 +#include "qemu/cutils.h"
++#include "qemu/error-report.h"
 +#include "qemu/option.h"
 +#include "qemu/module.h"
-+#include "sysemu/block-backend.h"
++#include "system/block-backend.h"
 +
 +#define TRACK_OPT_AUTO_REMOVE "auto-remove"
 +
-+typedef enum DropState {
-+    DropNone,
-+    DropRequested,
-+    DropInProgress,
-+} DropState;
-+
 +typedef struct {
 +    BdrvDirtyBitmap *bitmap;
-+    DropState drop_state;
-+    bool auto_remove;
++    uint64_t granularity;
 +} BDRVAllocTrackState;
 +
 +static QemuOptsList runtime_opts = {
@@ -91,26 +93,29 @@ index 0000000000..b75d7c6460
 +    },
 +};
 +
-+static void track_refresh_limits(BlockDriverState *bs, Error **errp)
++static void GRAPH_RDLOCK
++track_refresh_limits(BlockDriverState *bs, Error **errp)
 +{
-+    BlockDriverInfo bdi;
++    BDRVAllocTrackState *s = bs->opaque;
 +
 +    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 (no partial
-+     * cluster allocation in 'file') */
-+    bdrv_get_info(bs->file->bs, &bdi);
++    /*
++     * 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.
++     */
 +    bs->bl.request_alignment = MAX(bs->file->bs->bl.request_alignment,
-+                                   MAX(bdi.cluster_size, BDRV_SECTOR_SIZE));
++                                   s->granularity);
 +}
 +
 +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;
@@ -123,32 +128,63 @@ index 0000000000..b75d7c6460
 +        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 */
-+    bs->file = bdrv_open_child(NULL, options, "file", bs, &child_of_bds,
-+                               BDRV_CHILD_DATA | BDRV_CHILD_METADATA, false,
-+                               &local_err);
++    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();
 +    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);
-+    uint64_t gran = bs->bl.request_alignment;
-+    s->bitmap = bdrv_create_dirty_bitmap(bs->file->bs, gran, NULL, &local_err);
++    s->bitmap = bdrv_create_dirty_bitmap(bs->file->bs, s->granularity, NULL,
++                                         &local_err);
++    bdrv_graph_rdunlock_main_loop();
 +    if (local_err) {
 +        ret = -EIO;
 +        error_propagate(errp, local_err);
 +        goto fail;
 +    }
 +
-+    s->drop_state = DropNone;
-+
 +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);
 +        }
@@ -165,13 +201,15 @@ index 0000000000..b75d7c6460
 +    }
 +}
 +
-+static coroutine_fn int64_t track_co_getlength(BlockDriverState *bs)
++static coroutine_fn int64_t GRAPH_RDLOCK
++track_co_getlength(BlockDriverState *bs)
 +{
 +    return bdrv_co_getlength(bs->file->bs);
 +}
 +
-+static int coroutine_fn track_co_preadv(BlockDriverState *bs,
-+    int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags)
++static int coroutine_fn GRAPH_RDLOCK
++track_co_preadv(BlockDriverState *bs, int64_t offset, int64_t bytes,
++                QEMUIOVector *qiov, BdrvRequestFlags flags)
 +{
 +    BDRVAllocTrackState *s = bs->opaque;
 +    QEMUIOVector local_qiov;
@@ -229,36 +267,39 @@ index 0000000000..b75d7c6460
 +    return ret;
 +}
 +
-+static int coroutine_fn track_co_pwritev(BlockDriverState *bs,
-+    int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags)
++static int coroutine_fn GRAPH_RDLOCK
++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 track_co_pwrite_zeroes(BlockDriverState *bs,
-+    int64_t offset, int64_t bytes, BdrvRequestFlags flags)
++static int coroutine_fn GRAPH_RDLOCK
++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 track_co_pdiscard(BlockDriverState *bs,
-+    int64_t offset, int64_t bytes)
++static int coroutine_fn GRAPH_RDLOCK
++track_co_pdiscard(BlockDriverState *bs, int64_t offset, int64_t bytes)
 +{
 +    return bdrv_co_pdiscard(bs->file, offset, bytes);
 +}
 +
-+static coroutine_fn int track_co_flush(BlockDriverState *bs)
++static coroutine_fn int GRAPH_RDLOCK
++track_co_flush(BlockDriverState *bs)
 +{
 +    return bdrv_co_flush(bs->file->bs);
 +}
 +
-+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)
++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)
 +{
 +    BDRVAllocTrackState *s = bs->opaque;
 +
@@ -284,23 +325,13 @@ index 0000000000..b75d7c6460
 +    return 0;
 +}
 +
-+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)
++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)
 +{
-+    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 {
@@ -310,55 +341,22 @@ index 0000000000..b75d7c6460
 +    }
 +}
 +
-+static void track_drop(void *opaque)
++static int coroutine_fn GRAPH_RDLOCK
++track_co_change_backing_file(BlockDriverState *bs, const char *backing_file,
++                             const char *backing_fmt)
 +{
-+    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);
-+    }
-+
++    /*
++     * 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).
++     */
 +    return 0;
 +}
 +
@@ -366,7 +364,7 @@ index 0000000000..b75d7c6460
 +    .format_name                      = "alloc-track",
 +    .instance_size                    = sizeof(BDRVAllocTrackState),
 +
-+    .bdrv_file_open                   = track_open,
++    .bdrv_open                        = track_open,
 +    .bdrv_close                       = track_close,
 +    .bdrv_co_getlength                = track_co_getlength,
 +    .bdrv_child_perm                  = track_child_perm,
@@ -383,7 +381,7 @@ index 0000000000..b75d7c6460
 +    .supports_backing                 = true,
 +
 +    .bdrv_co_block_status             = track_co_block_status,
-+    .bdrv_change_backing_file         = track_change_backing_file,
++    .bdrv_co_change_backing_file      = track_co_change_backing_file,
 +};
 +
 +static void bdrv_alloc_track_init(void)
@@ -393,7 +391,7 @@ index 0000000000..b75d7c6460
 +
 +block_init(bdrv_alloc_track_init);
 diff --git a/block/meson.build b/block/meson.build
-index becc99ac4e..0a69836593 100644
+index d243372c41..9b45b5256d 100644
 --- a/block/meson.build
 +++ b/block/meson.build
 @@ -2,6 +2,7 @@ block_ss.add(genh)
@@ -404,3 +402,48 @@ index becc99ac4e..0a69836593 100644
    'amend.c',
    'backup.c',
    'backup-dump.c',
+diff --git a/block/stream.c b/block/stream.c
+index e187cd1262..0b61029399 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/0038-block-io-accept-NULL-qiov-in-bdrv_pad_request.patch

@@ -1,64 +0,0 @@
-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.
-
-If there is no qiov, no operation can be done with it, but the bytes
-and offset still need to be updated, so the subsequent aligned read
-will actually be aligned and not run into an assertion failure.
-
-Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
-[FE: do update bytes and offset in any case]
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
----
- block/io.c | 29 ++++++++++++++++-------------
- 1 file changed, 16 insertions(+), 13 deletions(-)
-
-diff --git a/block/io.c b/block/io.c
-index 83d1b1dfdc..e927881e40 100644
---- a/block/io.c
-+++ b/block/io.c
-@@ -1723,22 +1723,25 @@ static int bdrv_pad_request(BlockDriverState *bs,
-         return 0;
-     }
- 
--    sliced_iov = qemu_iovec_slice(*qiov, *qiov_offset, *bytes,
--                                  &sliced_head, &sliced_tail,
--                                  &sliced_niov);
--
--    /* Guaranteed by bdrv_check_request32() */
--    assert(*bytes <= SIZE_MAX);
--    ret = bdrv_create_padded_qiov(bs, pad, sliced_iov, sliced_niov,
--                                  sliced_head, *bytes);
--    if (ret < 0) {
--        bdrv_padding_finalize(pad);
--        return ret;
-+    if (qiov && *qiov) {
-+        sliced_iov = qemu_iovec_slice(*qiov, *qiov_offset, *bytes,
-+                                      &sliced_head, &sliced_tail,
-+                                      &sliced_niov);
-+
-+        /* Guaranteed by bdrv_check_request32() */
-+        assert(*bytes <= SIZE_MAX);
-+        ret = bdrv_create_padded_qiov(bs, pad, sliced_iov, sliced_niov,
-+                                      sliced_head, *bytes);
-+        if (ret < 0) {
-+            bdrv_padding_finalize(pad);
-+            return ret;
-+        }
-+        *qiov = &pad->local_qiov;
-+        *qiov_offset = 0;
-     }
-+
-     *bytes += pad->head + pad->tail;
-     *offset -= pad->head;
--    *qiov = &pad->local_qiov;
--    *qiov_offset = 0;
-     if (padded) {
-         *padded = true;
-     }

debian/patches/pve/0039-Revert-block-rbd-workaround-for-ceph-issue-53784.patch

@@ -13,10 +13,10 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  1 file changed, 2 insertions(+), 40 deletions(-)
 
 diff --git a/block/rbd.c b/block/rbd.c
-index a4749f3b1b..53e0396b51 100644
+index bf143fac00..70d92966f7 100644
 --- a/block/rbd.c
 +++ b/block/rbd.c
-@@ -1511,7 +1511,6 @@ static int coroutine_fn qemu_rbd_co_block_status(BlockDriverState *bs,
+@@ -1515,7 +1515,6 @@ static int coroutine_fn qemu_rbd_co_block_status(BlockDriverState *bs,
      int status, r;
      RBDDiffIterateReq req = { .offs = offset };
      uint64_t features, flags;
@@ -24,7 +24,7 @@ index a4749f3b1b..53e0396b51 100644
  
      assert(offset + bytes <= s->image_size);
  
-@@ -1539,43 +1538,7 @@ static int coroutine_fn qemu_rbd_co_block_status(BlockDriverState *bs,
+@@ -1543,43 +1542,7 @@ static int coroutine_fn qemu_rbd_co_block_status(BlockDriverState *bs,
          return status;
      }
  
@@ -69,7 +69,7 @@ index a4749f3b1b..53e0396b51 100644
                            qemu_rbd_diff_iterate_cb, &req);
      if (r < 0 && r != QEMU_RBD_EXIT_DIFF_ITERATE2) {
          return status;
-@@ -1594,8 +1557,7 @@ static int coroutine_fn qemu_rbd_co_block_status(BlockDriverState *bs,
+@@ -1598,8 +1561,7 @@ static int coroutine_fn qemu_rbd_co_block_status(BlockDriverState *bs,
          status = BDRV_BLOCK_ZERO | BDRV_BLOCK_OFFSET_VALID;
      }
  

debian/patches/pve/0040-Revert-block-rbd-fix-handling-of-holes-in-.bdrv_co_b.patch

@@ -14,10 +14,10 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  1 file changed, 5 insertions(+), 5 deletions(-)
 
 diff --git a/block/rbd.c b/block/rbd.c
-index 53e0396b51..0913a0af39 100644
+index 70d92966f7..931b513828 100644
 --- a/block/rbd.c
 +++ b/block/rbd.c
-@@ -1470,11 +1470,11 @@ static int qemu_rbd_diff_iterate_cb(uint64_t offs, size_t len,
+@@ -1474,11 +1474,11 @@ 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/0041-Revert-block-rbd-implement-bdrv_co_block_status.patch

@@ -24,7 +24,7 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  1 file changed, 112 deletions(-)
 
 diff --git a/block/rbd.c b/block/rbd.c
-index 0913a0af39..1dab254517 100644
+index 931b513828..4ab9bb5e02 100644
 --- a/block/rbd.c
 +++ b/block/rbd.c
 @@ -108,12 +108,6 @@ typedef struct RBDTask {
@@ -40,7 +40,7 @@ index 0913a0af39..1dab254517 100644
  static int qemu_rbd_connect(rados_t *cluster, rados_ioctx_t *io_ctx,
                              BlockdevOptionsRbd *opts, bool cache,
                              const char *keypairs, const char *secretid,
-@@ -1456,111 +1450,6 @@ static ImageInfoSpecific *qemu_rbd_get_specific_info(BlockDriverState *bs,
+@@ -1460,111 +1454,6 @@ static ImageInfoSpecific *qemu_rbd_get_specific_info(BlockDriverState *bs,
      return spec_info;
  }
  
@@ -152,7 +152,7 @@ index 0913a0af39..1dab254517 100644
  static int64_t coroutine_fn qemu_rbd_co_getlength(BlockDriverState *bs)
  {
      BDRVRBDState *s = bs->opaque;
-@@ -1796,7 +1685,6 @@ static BlockDriver bdrv_rbd = {
+@@ -1801,7 +1690,6 @@ static BlockDriver bdrv_rbd = {
  #ifdef LIBRBD_SUPPORTS_WRITE_ZEROES
      .bdrv_co_pwrite_zeroes  = qemu_rbd_co_pwrite_zeroes,
  #endif

debian/patches/pve/0042-PVE-backup-add-fleecing-option.patch

@@ -0,0 +1,472 @@
+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>
+[FE: improve error when cbw fails as reported by Friedrich Weber]
+Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
+---
+ block/copy-before-write.c      |  18 ++--
+ block/copy-before-write.h      |   1 +
+ block/monitor/block-hmp-cmds.c |   1 +
+ pve-backup.c                   | 175 ++++++++++++++++++++++++++++++++-
+ qapi/block-core.json           |  10 +-
+ 5 files changed, 195 insertions(+), 10 deletions(-)
+
+diff --git a/block/copy-before-write.c b/block/copy-before-write.c
+index fd470f5f92..5c23b578ef 100644
+--- a/block/copy-before-write.c
++++ b/block/copy-before-write.c
+@@ -27,6 +27,7 @@
+ #include "qobject/qjson.h"
+ 
+ #include "system/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/block/monitor/block-hmp-cmds.c b/block/monitor/block-hmp-cmds.c
+index 4f30f99644..66d16d342f 100644
+--- a/block/monitor/block-hmp-cmds.c
++++ b/block/monitor/block-hmp-cmds.c
+@@ -1045,6 +1045,7 @@ 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 366b015589..9b66788ab5 100644
+--- a/pve-backup.c
++++ b/pve-backup.c
+@@ -7,6 +7,7 @@
+ #include "system/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"
+@@ -81,8 +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;
+@@ -352,11 +360,44 @@ static void coroutine_fn pvebackup_co_complete_stream(void *opaque)
+     qemu_co_mutex_unlock(&backup_state.backup_mutex);
+ }
+ 
++static void cleanup_snapshot_access(PVEBackupDevInfo *di)
++{
++    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
++         * 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;
++        }
++    }
++
++    /*
++     * 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.
+      */
+@@ -487,6 +528,65 @@ 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) {
++        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;
++    }
++
++    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.
+@@ -523,9 +623,42 @@ 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;
++        if (di->fleecing.bs) {
++            if (setup_snapshot_access(di, &local_err) < 0) {
++                error_setg(errp, "%s - setting up snapshot access for fleecing failed: %s",
++                           di->device_name,
++                           local_err ? error_get_pretty(local_err) : "unknown error");
++                bdrv_drained_end(di->bs);
++                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,
++            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);
+ 
+@@ -539,6 +672,7 @@ 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;
+@@ -581,6 +715,14 @@ 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
+@@ -588,6 +730,7 @@ 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;
+@@ -613,6 +756,30 @@ 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)) {
++                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++;
+         }
+@@ -663,6 +830,7 @@ 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());
+@@ -691,7 +859,7 @@ 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);
+@@ -1093,5 +1261,6 @@ 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 6838187607..9bdcfa31ea 100644
+--- a/qapi/block-core.json
++++ b/qapi/block-core.json
+@@ -952,6 +952,10 @@
+ #
+ # @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
+ #
+ ##
+@@ -972,7 +976,8 @@
+                                     '*firewall-file': 'str',
+                                     '*devlist': 'str',
+                                     '*speed': 'int',
+-                                    '*max-workers': 'int' },
++                                    '*max-workers': 'int',
++                                    '*fleecing': 'bool' },
+   'returns': 'UuidInfo', 'coroutine': true }
+ 
+ ##
+@@ -1018,6 +1023,8 @@
+ #
+ # @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.
+ #
+@@ -1029,6 +1036,7 @@
+             'pbs-dirty-bitmap-migration': 'bool',
+             'pbs-masterkey': 'bool',
+             'pbs-library-version': 'str',
++            'backup-fleecing': 'bool',
+             'backup-max-workers': 'bool' } }
+ 
+ ##

debian/patches/pve/0043-adapt-machine-version-deprecation-for-Proxmox-VE.patch

@@ -0,0 +1,137 @@
+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 8ada4d5832..f9f3b75284 100644
+--- a/include/hw/boards.h
++++ b/include/hw/boards.h
+@@ -636,42 +636,66 @@ 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)
+@@ -680,7 +704,7 @@ 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)
+@@ -689,7 +713,7 @@ 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/0043-alloc-track-fix-deadlock-during-drop.patch

@@ -1,154 +0,0 @@
-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>
-Signed-off-by: Thomas Lamprecht <t.lamprecht@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-Revert-hpet-avoid-timer-storms-on-periodic-timers.patch

@@ -0,0 +1,50 @@
+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 ccb97b6806..0f45af8bbe 100644
+--- a/hw/timer/hpet.c
++++ b/hw/timer/hpet.c
+@@ -61,7 +61,6 @@ 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 {
+@@ -262,7 +261,6 @@ 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) {
+@@ -350,15 +348,8 @@ 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/0044-migration-for-snapshots-hold-the-BQL-during-setup-ca.patch

@@ -1,191 +0,0 @@
-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>
-Signed-off-by: Thomas Lamprecht <t.lamprecht@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 6e1514f69f..6a1aec7031 100644
---- a/migration/ram.c
-+++ b/migration/ram.c
-@@ -2896,8 +2896,16 @@ 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() {
-@@ -2909,7 +2917,9 @@ 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 d60c4f487a..3c015722f7 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-Revert-hpet-store-full-64-bit-target-value-of-the-co.patch

@@ -0,0 +1,202 @@
+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 | 109 +++++++++++++++++++++++++++---------------------
+ 1 file changed, 61 insertions(+), 48 deletions(-)
+
+diff --git a/hw/timer/hpet.c b/hw/timer/hpet.c
+index 0f45af8bbe..635a060d38 100644
+--- a/hw/timer/hpet.c
++++ b/hw/timer/hpet.c
+@@ -56,7 +56,6 @@ 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.
+@@ -119,6 +118,11 @@ 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);
+ }
+@@ -155,32 +159,27 @@ 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)
+@@ -256,12 +255,7 @@ 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)
+@@ -346,10 +340,14 @@ 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);
++    }
+ }
+ 
+ /*
+@@ -358,44 +356,54 @@ 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) {
++    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 (!timer_is_periodic(t) && t->cmp64 > hpet_next_wrap(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)
+@@ -526,7 +534,12 @@ 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/0045-savevm-async-don-t-hold-BQL-during-setup.patch

@@ -1,30 +0,0 @@
-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>
-Signed-off-by: Thomas Lamprecht <t.lamprecht@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-Revert-hpet-accept-64-bit-reads-and-writes.patch

@@ -0,0 +1,281 @@
+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 635a060d38..5f4bb5667d 100644
+--- a/hw/timer/hpet.c
++++ b/hw/timer/hpet.c
+@@ -421,7 +421,6 @@ 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);
+@@ -436,33 +435,52 @@ 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;
+@@ -476,11 +494,11 @@ 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) {
+@@ -492,12 +510,9 @@ 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
+@@ -505,7 +520,8 @@ 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);
+@@ -518,28 +534,56 @@ 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)) {
+@@ -547,7 +591,10 @@ 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();
+@@ -555,14 +602,12 @@ 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 =
+@@ -592,8 +637,10 @@ 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)) {
+@@ -605,7 +652,15 @@ 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();
+@@ -619,11 +674,7 @@ 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 c5b6db49f5..dd8a53c690 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/0047-Revert-hpet-place-read-only-bits-directly-in-new_val.patch

@@ -0,0 +1,64 @@
+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 5f4bb5667d..5e3bf1f153 100644
+--- a/hw/timer/hpet.c
++++ b/hw/timer/hpet.c
+@@ -494,7 +494,7 @@ 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);
+@@ -520,12 +520,13 @@ 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;
+@@ -606,8 +607,8 @@ 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 =
+@@ -641,9 +642,9 @@ 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/0048-Revert-hpet-remove-unnecessary-variable-index.patch

@@ -0,0 +1,68 @@
+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 5e3bf1f153..daef12c8cf 100644
+--- a/hw/timer/hpet.c
++++ b/hw/timer/hpet.c
+@@ -421,12 +421,12 @@ 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];
+ 
+@@ -453,7 +453,7 @@ 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:
+@@ -494,14 +494,15 @@ 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];
+ 
+@@ -603,7 +604,7 @@ 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/0049-Revert-hpet-ignore-high-bits-of-comparator-in-32-bit.patch

@@ -0,0 +1,40 @@
+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 daef12c8cf..927263e2ff 100644
+--- a/hw/timer/hpet.c
++++ b/hw/timer/hpet.c
+@@ -569,10 +569,6 @@ 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 dd8a53c690..2b81ee0812 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/0050-Revert-hpet-fix-and-cleanup-persistence-of-interrupt.patch

@@ -0,0 +1,120 @@
+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 927263e2ff..5aae09f166 100644
+--- a/hw/timer/hpet.c
++++ b/hw/timer/hpet.c
+@@ -199,31 +199,21 @@ 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]);
+     }
+ }
+ 
+@@ -408,13 +398,8 @@ 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,
+@@ -514,26 +499,20 @@ 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
+@@ -611,10 +590,9 @@ 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/0051-PVE-backup-factor-out-helper-to-clear-backup-state-s.patch

@@ -0,0 +1,59 @@
+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 9b66788ab5..588ee98ffc 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/0052-PVE-backup-factor-out-helper-to-initialize-backup-st.patch

@@ -0,0 +1,95 @@
+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 588ee98ffc..3be9930ad3 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/0053-PVE-backup-add-target-ID-in-backup-state.patch

@@ -0,0 +1,63 @@
+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 3be9930ad3..87778f7e76 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/0054-PVE-backup-get-device-info-allow-caller-to-specify-f.patch

@@ -0,0 +1,57 @@
+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 87778f7e76..bd81621d51 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/0055-PVE-backup-implement-backup-access-setup-and-teardow.patch

@@ -0,0 +1,898 @@
+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.
+
+There can be one dirty bitmap for each backup target ID for each
+device (which are tracked in the backup_access_bitmaps hash table).
+The QMP user can specify the ID of the bitmap it likes to use. This ID
+is then compared to the current one for the given target and device.
+If they match, the bitmap is re-used (should it still exist on the
+drive, otherwise re-created). If there is a mismatch, the old bitmap
+is removed and a new one is created.
+
+The return value of the QMP command includes information about what
+bitmap action was taken. Similar to what the query-backup QMP command
+returns for regular backup. It also includes the bitmap name and
+associated block node, so the management layer can then set up an NBD
+export with the bitmap.
+
+While the backup access is active, a background bitmap is also
+required. This is necessary to implement bitmap handling according to
+the original reference [0]. In particular:
+
+- in the error case, new writes since the backup access was set up are
+  in the background bitmap. Because of failure, the previously tracked
+  writes from the backup access bitmap are still required too. Thus,
+  the bitmap is merged with the background bitmap to get all new
+  writes since the last backup.
+
+- in the success case, continue tracking for the next incremental
+  backup in the backup access bitmap. New writes since the backup
+  access was set up are in the background bitmap. Because the backup
+  was successfully, clear the backup access bitmap and merge back the
+  background bitmap to get only the new writes.
+
+Since QEMU cannot know if the backup was successful or not (except if
+failure already happens during the setup QMP command), the management
+layer needs to tell it via the teardown QMP command.
+
+The bitmap action is also recorded in the device info now.
+
+The backup-access api keeps track of what bitmap names got used for
+which devices and thus knows when a bitmap went missing. Propagate
+this information to the QMP user with a new 'missing-recreated'
+variant for the taken bitmap action.
+
+[0]: https://lore.kernel.org/qemu-devel/b68833dd-8864-4d72-7c61-c134a9835036@ya.ru/
+
+Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
+Reviewed-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
+---
+ pve-backup.c         | 519 +++++++++++++++++++++++++++++++++++++++----
+ pve-backup.h         |  16 ++
+ qapi/block-core.json |  99 ++++++++-
+ system/runstate.c    |   6 +
+ 4 files changed, 596 insertions(+), 44 deletions(-)
+ create mode 100644 pve-backup.h
+
+diff --git a/pve-backup.c b/pve-backup.c
+index bd81621d51..0450303017 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"
+@@ -14,6 +15,7 @@
+ #include "qobject/qdict.h"
+ #include "qapi/qmp/qerror.h"
+ #include "qemu/cutils.h"
++#include "qemu/error-report.h"
+ 
+ #if defined(CONFIG_MALLOC_TRIM)
+ #include <malloc.h>
+@@ -40,6 +42,7 @@
+  */
+ 
+ const char *PBS_BITMAP_NAME = "pbs-incremental-dirty-bitmap";
++const char *BACKGROUND_BITMAP_NAME = "backup-access-background-bitmap";
+ 
+ static struct PVEBackupState {
+     struct {
+@@ -98,8 +101,11 @@ typedef struct PVEBackupDevInfo {
+     char* device_name;
+     int completed_ret; // INT_MAX if not completed
+     BdrvDirtyBitmap *bitmap;
++    BdrvDirtyBitmap *background_bitmap; // used for external backup access
++    PBSBitmapAction bitmap_action;
+     BlockDriverState *target;
+     BlockJob *job;
++    BackupAccessSetupBitmapMode requested_bitmap_mode;
+ } PVEBackupDevInfo;
+ 
+ static void pvebackup_propagate_error(Error *err)
+@@ -361,6 +367,67 @@ static void coroutine_fn pvebackup_co_complete_stream(void *opaque)
+     qemu_co_mutex_unlock(&backup_state.backup_mutex);
+ }
+ 
++/*
++ * New writes since the backup access was set up are in the background bitmap. Because of failure,
++ * the previously tracked writes in di->bitmap are still required too. Thus, merge with the
++ * background bitmap to get all new writes since the last backup.
++ */
++static void handle_backup_access_bitmaps_in_error_case(PVEBackupDevInfo *di)
++{
++    Error *local_err = NULL;
++
++    if (di->bs && di->background_bitmap) {
++        bdrv_drained_begin(di->bs);
++        if (di->bitmap) {
++            bdrv_enable_dirty_bitmap(di->bitmap);
++            if (!bdrv_merge_dirty_bitmap(di->bitmap, di->background_bitmap, NULL, &local_err)) {
++                warn_report("backup access: %s - could not merge bitmaps in error path - %s",
++                            di->device_name,
++                            local_err ? error_get_pretty(local_err) : "unknown error");
++                /*
++                 * Could not merge, drop original bitmap too.
++                 */
++                bdrv_release_dirty_bitmap(di->bitmap);
++            }
++        } else {
++            warn_report("backup access: %s - expected bitmap not present", di->device_name);
++        }
++        bdrv_release_dirty_bitmap(di->background_bitmap);
++        bdrv_drained_end(di->bs);
++    }
++}
++
++/*
++ * Continue tracking for next incremental backup in di->bitmap. New writes since the backup access
++ * was set up are in the background bitmap. Because the backup was successful, clear di->bitmap and
++ * merge back the background bitmap to get only the new writes.
++ */
++static void handle_backup_access_bitmaps_after_success(PVEBackupDevInfo *di)
++{
++    Error *local_err = NULL;
++
++    if (di->bs && di->background_bitmap) {
++        bdrv_drained_begin(di->bs);
++        if (di->bitmap) {
++            bdrv_enable_dirty_bitmap(di->bitmap);
++            bdrv_clear_dirty_bitmap(di->bitmap, NULL);
++            if (!bdrv_merge_dirty_bitmap(di->bitmap, di->background_bitmap, NULL, &local_err)) {
++                warn_report("backup access: %s - could not merge bitmaps after backup - %s",
++                            di->device_name,
++                            local_err ? error_get_pretty(local_err) : "unknown error");
++                /*
++                 * Could not merge, drop original bitmap too.
++                 */
++                bdrv_release_dirty_bitmap(di->bitmap);
++            }
++        } else {
++            warn_report("backup access: %s - expected bitmap not present", di->device_name);
++        }
++        bdrv_release_dirty_bitmap(di->background_bitmap);
++        bdrv_drained_end(di->bs);
++    }
++}
++
+ static void cleanup_snapshot_access(PVEBackupDevInfo *di)
+ {
+     if (di->fleecing.snapshot_access) {
+@@ -588,6 +655,51 @@ 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 (di->bitmap) {
++            BdrvDirtyBitmap *background_bitmap =
++                bdrv_create_dirty_bitmap(di->bs, PROXMOX_BACKUP_DEFAULT_CHUNK_SIZE,
++                                         BACKGROUND_BITMAP_NAME, &local_err);
++            if (!background_bitmap) {
++                error_setg(errp, "%s - creating background bitmap for backup access failed: %s",
++                           di->device_name,
++                           local_err ? error_get_pretty(local_err) : "unknown error");
++                bdrv_drained_end(di->bs);
++                break;
++            }
++            di->background_bitmap = background_bitmap;
++            bdrv_disable_dirty_bitmap(di->bitmap);
++        }
++
++        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 +836,62 @@ 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;
++}
++
++static PVEBackupDevInfo coroutine_fn GRAPH_RDLOCK *get_single_device_info(
++    const char *device,
++    bool (*device_uses_fleecing)(const char*),
++    Error **errp)
++{
++    BlockBackend *blk = blk_by_name(device);
++    if (!blk) {
++        error_set(errp, ERROR_CLASS_DEVICE_NOT_FOUND,
++                  "Device '%s' not found", device);
++        return NULL;
++    }
++    BlockDriverState *bs = blk_bs(blk);
++    if (!bdrv_co_is_inserted(bs)) {
++        error_setg(errp, "Device '%s' has no medium", device);
++        return NULL;
++    }
++    PVEBackupDevInfo *di = g_new0(PVEBackupDevInfo, 1);
++    di->bs = bs;
++    di->device_name = g_strdup(bdrv_get_device_name(bs));
++
++    if (device_uses_fleecing && device_uses_fleecing(device)) {
++        g_autofree gchar *fleecing_devid = g_strconcat(device, "-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 fail;
++        }
++        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 fail;
++        }
++        /*
++         * 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 fail;
++        }
++        di->fleecing.bs = fleecing_bs;
++    }
++
++    return di;
++fail:
++    g_free(di->device_name);
++    g_free(di);
++    return NULL;
++}
++
+ /*
+  * 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
+@@ -742,45 +910,10 @@ static GList coroutine_fn GRAPH_RDLOCK *get_device_info(
+ 
+         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);
++            PVEBackupDevInfo *di = get_single_device_info(*d, device_uses_fleecing, errp);
++            if (!di) {
+                 goto err;
+             }
+-            PVEBackupDevInfo *di = g_new0(PVEBackupDevInfo, 1);
+-            di->bs = bs;
+-            di->device_name = g_strdup(bdrv_get_device_name(bs));
+-
+-            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) {
+-                    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++;
+         }
+@@ -839,8 +972,9 @@ 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 +989,19 @@ 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 +1014,299 @@ 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,
++    BackupAccessSourceDeviceList *devices,
++    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();
++    for (BackupAccessSourceDeviceList *it = devices; it; it = it->next) {
++        PVEBackupDevInfo *di = get_single_device_info(it->value->device, fleecing_all, &local_err);
++        if (!di) {
++            bdrv_graph_co_rdunlock();
++            error_propagate(errp, local_err);
++            goto err;
++        }
++        di->requested_bitmap_mode = it->value->bitmap_mode;
++        di_list = g_list_append(di_list, di);
++    }
++    bdrv_graph_co_rdunlock();
++    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();
++
++    const char *bitmap_name = target_id;
++
++    /* create bitmaps if requested */
++    l = di_list;
++    while (l) {
++        PVEBackupDevInfo *di = (PVEBackupDevInfo *)l->data;
++        l = g_list_next(l);
++
++        di->block_size = PROXMOX_BACKUP_DEFAULT_CHUNK_SIZE;
++
++        PBSBitmapAction action = PBS_BITMAP_ACTION_NOT_USED;
++        size_t dirty = di->size;
++
++        if (di->requested_bitmap_mode == BACKUP_ACCESS_SETUP_BITMAP_MODE_NONE ||
++            di->requested_bitmap_mode == BACKUP_ACCESS_SETUP_BITMAP_MODE_NEW) {
++            BdrvDirtyBitmap *old_bitmap = bdrv_find_dirty_bitmap(di->bs, bitmap_name);
++            if (old_bitmap) {
++                bdrv_release_dirty_bitmap(old_bitmap);
++                action = PBS_BITMAP_ACTION_NOT_USED_REMOVED; // set below for new
++            }
++        }
++
++        BdrvDirtyBitmap *bitmap = NULL;
++        if (di->requested_bitmap_mode == BACKUP_ACCESS_SETUP_BITMAP_MODE_NEW ||
++            di->requested_bitmap_mode == BACKUP_ACCESS_SETUP_BITMAP_MODE_USE) {
++            bitmap = bdrv_find_dirty_bitmap(di->bs, bitmap_name);
++            if (!bitmap) {
++                bitmap = bdrv_create_dirty_bitmap(di->bs, PROXMOX_BACKUP_DEFAULT_CHUNK_SIZE,
++                                                  bitmap_name, errp);
++                if (!bitmap) {
++                    qemu_mutex_unlock(&backup_state.stat.lock);
++                    goto err;
++                }
++                bdrv_set_dirty_bitmap(bitmap, 0, di->size);
++                if (di->requested_bitmap_mode == BACKUP_ACCESS_SETUP_BITMAP_MODE_USE) {
++                    action = PBS_BITMAP_ACTION_MISSING_RECREATED;
++                } else {
++                    action = PBS_BITMAP_ACTION_NEW;
++                }
++            } else {
++                if (di->requested_bitmap_mode == BACKUP_ACCESS_SETUP_BITMAP_MODE_NEW) {
++                    qemu_mutex_unlock(&backup_state.stat.lock);
++                    error_setg(errp, "internal error - removed old bitmap still present");
++                    goto err;
++                }
++                /* track clean chunks as reused */
++                dirty = MIN(bdrv_get_dirty_count(bitmap), di->size);
++                backup_state.stat.reused += di->size - dirty;
++                action = PBS_BITMAP_ACTION_USED;
++            }
++        }
++
++        PBSBitmapInfo *info = g_malloc(sizeof(*info));
++        info->drive = g_strdup(di->device_name);
++        info->action = action;
++        info->size = di->size;
++        info->dirty = dirty;
++        backup_state.stat.bitmap_list = g_list_append(backup_state.stat.bitmap_list, info);
++
++        di->bitmap = bitmap;
++        di->bitmap_action = action;
++    }
++
++    /* 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;
++        if (di->bitmap) {
++            info->value->bitmap_node_name = g_strdup(bdrv_get_node_name(di->bs));
++            info->value->bitmap_name = g_strdup(bitmap_name);
++            info->value->bitmap_action = di->bitmap_action;
++            info->value->has_bitmap_action = true;
++        }
++
++        *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);
++
++        handle_backup_access_bitmaps_in_error_case(di);
++
++        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(bool success)
++{
++    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;
++        }
++
++        if (success) {
++            handle_backup_access_bitmaps_after_success(di);
++        } else {
++            handle_backup_access_bitmaps_in_error_case(di);
++        }
++
++        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(*((bool*)data->data));
++
++    /* return */
++    aio_co_enter(data->ctx, data->co);
++}
++
++void coroutine_fn qmp_backup_access_teardown(const char *target_id, bool success, 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(),
++        .data = &success,
++    };
++    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,
+@@ -1068,6 +1499,7 @@ UuidInfo coroutine_fn *qmp_backup(
+             }
+ 
+             di->dev_id = dev_id;
++            di->bitmap_action = action;
+ 
+             PBSBitmapInfo *info = g_malloc(sizeof(*info));
+             info->drive = g_strdup(di->device_name);
+@@ -1119,7 +1551,7 @@ 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 +1730,6 @@ 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..9ebeef7c8f
+--- /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(bool success);
++
++#endif /* PVE_BACKUP_H */
+diff --git a/qapi/block-core.json b/qapi/block-core.json
+index 9bdcfa31ea..2fb51215f2 100644
+--- a/qapi/block-core.json
++++ b/qapi/block-core.json
+@@ -1023,6 +1023,9 @@
+ #
+ # @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
+@@ -1036,6 +1039,7 @@
+             'pbs-dirty-bitmap-migration': 'bool',
+             'pbs-masterkey': 'bool',
+             'pbs-library-version': 'str',
++            'backup-access-api': 'bool',
+             'backup-fleecing': 'bool',
+             'backup-max-workers': 'bool' } }
+ 
+@@ -1067,9 +1071,16 @@
+ #           base snapshot did not match the base given for the current job or
+ #           the crypt mode has changed.
+ #
++# @missing-recreated: A bitmap for incremental backup was expected to be
++#     present, but was missing and thus got recreated. For example, this can
++#     happen if the drive was re-attached or if the bitmap was deleted for some
++#     other reason. PBS does not currently keep track of this; the backup-access
++#     mechanism does.
++#
+ ##
+ { 'enum': 'PBSBitmapAction',
+-  'data': ['not-used', 'not-used-removed', 'new', 'used', 'invalid'] }
++  'data': ['not-used', 'not-used-removed', 'new', 'used', 'invalid',
++           'missing-recreated'] }
+ 
+ ##
+ # @PBSBitmapInfo:
+@@ -1102,6 +1113,92 @@
+ ##
+ { '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.
++#
++# @bitmap-node-name: the block node name the dirty bitmap is associated to.
++#
++# @bitmap-name: the name of the dirty bitmap associated to the backup access.
++#
++# @bitmap-action: the action taken on the dirty bitmap.
++#
++##
++{ 'struct': 'BackupAccessInfo',
++  'data': { 'node-name': 'str', 'device': 'str', 'size': 'size',
++            '*bitmap-node-name': 'str', '*bitmap-name': 'str',
++            '*bitmap-action': 'PBSBitmapAction' } }
++
++##
++# @BackupAccessSourceDevice:
++#
++# Source block device information for creating a backup access.
++#
++# @device: the block device name.
++#
++# @bitmap-mode: used to control whether the bitmap should be reused or
++#     recreated or not used. Default is not using a bitmap.
++#
++##
++{ 'struct': 'BackupAccessSourceDevice',
++  'data': { 'device': 'str', '*bitmap-mode': 'BackupAccessSetupBitmapMode' } }
++
++##
++# @BackupAccessSetupBitmapMode:
++#
++# How to setup a bitmap for a device for @backup-access-setup.
++#
++# @none: do not use a bitmap. Removes an existing bitmap if present.
++#
++# @new: create and use a new bitmap.
++#
++# @use: try to re-use an existing bitmap. Create a new one if it doesn't exist.
++##
++{ 'enum': 'BackupAccessSetupBitmapMode',
++  'data': ['none', 'new', 'use' ] }
++
++##
++# @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.
++#
++# @devices: list of devices for which to create the backup access.  Also
++#     controls whether to use/create a bitmap for the device.  Check the
++#     @bitmap-action in the result to see what action was actually taken for the
++#     bitmap.  Each target controls its own bitmaps.
++#
++# Returns: a list of @BackupAccessInfo, one for each device.
++#
++##
++{ 'command': 'backup-access-setup',
++  'data': { 'target-id': 'str', 'devices': [ 'BackupAccessSourceDevice' ] },
++  '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.
++#
++# @success: whether the backup done by the external provider was successful.
++#
++##
++{ 'command': 'backup-access-teardown',
++  'data': { 'target-id': 'str', 'success': 'bool' },
++  'coroutine': true }
++
+ ##
+ # @BlockDeviceTimedStats:
+ #
+diff --git a/system/runstate.c b/system/runstate.c
+index 272801d307..cf775213bd 100644
+--- a/system/runstate.c
++++ b/system/runstate.c
+@@ -60,6 +60,7 @@
+ #include "system/system.h"
+ #include "system/tpm.h"
+ #include "trace.h"
++#include "pve-backup.h"
+ 
+ static NotifierList exit_notifiers =
+     NOTIFIER_LIST_INITIALIZER(exit_notifiers);
+@@ -921,6 +922,11 @@ 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(false);
+     job_cancel_sync_all();
+     bdrv_close_all();
+ 

debian/patches/pve/0056-PVE-backup-prepare-for-the-switch-to-using-blockdev-.patch

@@ -0,0 +1,106 @@
+From 5a8cf9e98ba1668a6a20c2fcda1704de4103ff58 Mon Sep 17 00:00:00 2001
+From: Fiona Ebner <f.ebner@proxmox.com>
+Date: Wed, 2 Jul 2025 18:27:34 +0200
+Subject: [PATCH 56/59] PVE backup: prepare for the switch to using blockdev
+ rather than drive
+MIME-Version: 1.0
+Content-Type: text/plain; charset=UTF-8
+Content-Transfer-Encoding: 8bit
+
+Also allow finding block nodes by their node name rather than just via
+an associated block backend, which might not exist for block nodes.
+
+For regular drives, it is essential to not use the throttle group,
+because otherwise the limits intended only for the guest would also
+apply to the backup job.
+
+Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
+Signed-off-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
+---
+ pve-backup.c | 51 +++++++++++++++++++++++++++++++++++++++------------
+ 1 file changed, 39 insertions(+), 12 deletions(-)
+
+diff --git a/pve-backup.c b/pve-backup.c
+index 0450303017..457fcb7e5c 100644
+--- a/pve-backup.c
++++ b/pve-backup.c
+@@ -847,29 +847,56 @@ static PVEBackupDevInfo coroutine_fn GRAPH_RDLOCK *get_single_device_info(
+     Error **errp)
+ {
+     BlockBackend *blk = blk_by_name(device);
+-    if (!blk) {
+-        error_set(errp, ERROR_CLASS_DEVICE_NOT_FOUND,
+-                  "Device '%s' not found", device);
+-        return NULL;
++    BlockDriverState *root_bs, *bs;
++
++    if (blk) {
++        root_bs = bs = blk_bs(blk);
++    } else {
++        /* TODO PVE 10 - fleecing will always be attached without blk */
++        root_bs = bs = bdrv_find_node(device);
++        if (!bs) {
++            error_set(errp, ERROR_CLASS_DEVICE_NOT_FOUND,
++                      "Device '%s' not found", device);
++            return NULL;
++        }
++        /* For TPM, bs is already correct, otherwise need the file child. */
++        if (!strncmp(bs->drv->format_name, "throttle", 8)) {
++            if (!bs->file || !bs->file->bs) {
++                error_set(errp, ERROR_CLASS_DEVICE_NOT_FOUND,
++                          "Device '%s' not found (no file child)", device);
++                return NULL;
++            }
++            bs = bs->file->bs;
++        }
+     }
+-    BlockDriverState *bs = blk_bs(blk);
++
+     if (!bdrv_co_is_inserted(bs)) {
+         error_setg(errp, "Device '%s' has no medium", device);
+         return NULL;
+     }
++
+     PVEBackupDevInfo *di = g_new0(PVEBackupDevInfo, 1);
+     di->bs = bs;
+-    di->device_name = g_strdup(bdrv_get_device_name(bs));
++    /* Need the name of the root node, e.g. drive-scsi0 */
++    di->device_name = g_strdup(bdrv_get_device_or_node_name(root_bs));
+ 
+     if (device_uses_fleecing && device_uses_fleecing(device)) {
+         g_autofree gchar *fleecing_devid = g_strconcat(device, "-fleecing", NULL);
++        BlockDriverState *fleecing_bs;
++
+         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 fail;
++        if (fleecing_blk) {
++            fleecing_bs = blk_bs(fleecing_blk);
++        } else {
++            /* TODO PVE 10 - fleecing will always be attached without blk */
++            fleecing_bs = bdrv_find_node(fleecing_devid);
++            if (!fleecing_bs) {
++                error_set(errp, ERROR_CLASS_DEVICE_NOT_FOUND,
++                          "Device '%s' not found", fleecing_devid);
++                goto fail;
++            }
+         }
+-        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 fail;
+@@ -927,7 +954,7 @@ 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->device_name = g_strdup(bdrv_get_device_or_node_name(bs));
+             di_list = g_list_append(di_list, di);
+         }
+     }
+-- 
+2.39.5
+

debian/patches/pve/0057-block-zeroinit-support-using-as-blockdev-driver.patch

@@ -0,0 +1,71 @@
+From 5beb1f48555d74f468b6c0ca657d3be44c8ea8e3 Mon Sep 17 00:00:00 2001
+From: Fiona Ebner <f.ebner@proxmox.com>
+Date: Wed, 2 Jul 2025 18:27:35 +0200
+Subject: [PATCH 57/59] block/zeroinit: support using as blockdev driver
+MIME-Version: 1.0
+Content-Type: text/plain; charset=UTF-8
+Content-Transfer-Encoding: 8bit
+
+Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
+Signed-off-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
+---
+ block/zeroinit.c     | 12 +++++++++---
+ qapi/block-core.json |  5 +++--
+ 2 files changed, 12 insertions(+), 5 deletions(-)
+
+diff --git a/block/zeroinit.c b/block/zeroinit.c
+index f9d513db15..036edb17f5 100644
+--- a/block/zeroinit.c
++++ b/block/zeroinit.c
+@@ -66,6 +66,7 @@ static int zeroinit_open(BlockDriverState *bs, QDict *options, int flags,
+     QemuOpts *opts;
+     Error *local_err = NULL;
+     int ret;
++    const char *next = NULL;
+ 
+     s->extents = 0;
+ 
+@@ -77,9 +78,14 @@ static int zeroinit_open(BlockDriverState *bs, QDict *options, int flags,
+         goto fail;
+     }
+ 
+-    /* Open the raw file */
+-    ret = bdrv_open_file_child(qemu_opt_get(opts, "x-next"), options, "next",
+-                               bs, &local_err);
++
++    next = qemu_opt_get(opts, "x-next");
++
++    if (next) {
++        ret = bdrv_open_file_child(next, options, "next", bs, &local_err);
++    } else { /* when opened as a blockdev, there is no 'next' option */
++        ret = bdrv_open_file_child(NULL, options, "file", bs, &local_err);
++    }
+     if (ret < 0) {
+         error_propagate(errp, local_err);
+         goto fail;
+diff --git a/qapi/block-core.json b/qapi/block-core.json
+index 2fb51215f2..f8ed564cf0 100644
+--- a/qapi/block-core.json
++++ b/qapi/block-core.json
+@@ -3586,7 +3586,7 @@
+             { 'name': 'virtio-blk-vfio-pci', 'if': 'CONFIG_BLKIO' },
+             { 'name': 'virtio-blk-vhost-user', 'if': 'CONFIG_BLKIO' },
+             { 'name': 'virtio-blk-vhost-vdpa', 'if': 'CONFIG_BLKIO' },
+-            'vmdk', 'vpc', 'vvfat' ] }
++            'vmdk', 'vpc', 'vvfat', 'zeroinit' ] }
+ 
+ ##
+ # @BlockdevOptionsFile:
+@@ -5172,7 +5172,8 @@
+                       'if': 'CONFIG_BLKIO' },
+       'vmdk':       'BlockdevOptionsGenericCOWFormat',
+       'vpc':        'BlockdevOptionsGenericFormat',
+-      'vvfat':      'BlockdevOptionsVVFAT'
++      'vvfat':      'BlockdevOptionsVVFAT',
++      'zeroinit':   'BlockdevOptionsGenericFormat'
+   } }
+ 
+ ##
+-- 
+2.39.5
+

debian/patches/pve/0058-block-alloc-track-support-using-as-blockdev-driver.patch

@@ -0,0 +1,61 @@
+From d180b059731818ae34e43e11495c8ac081ab89b9 Mon Sep 17 00:00:00 2001
+From: Fiona Ebner <f.ebner@proxmox.com>
+Date: Wed, 2 Jul 2025 18:27:36 +0200
+Subject: [PATCH 58/59] block/alloc-track: support using as blockdev driver
+MIME-Version: 1.0
+Content-Type: text/plain; charset=UTF-8
+Content-Transfer-Encoding: 8bit
+
+Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
+Signed-off-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
+---
+ qapi/block-core.json | 19 ++++++++++++++++++-
+ 1 file changed, 18 insertions(+), 1 deletion(-)
+
+diff --git a/qapi/block-core.json b/qapi/block-core.json
+index f8ed564cf0..07c5773717 100644
+--- a/qapi/block-core.json
++++ b/qapi/block-core.json
+@@ -3567,7 +3567,8 @@
+ # Since: 2.9
+ ##
+ { 'enum': 'BlockdevDriver',
+-  'data': [ 'blkdebug', 'blklogwrites', 'blkreplay', 'blkverify', 'bochs',
++  'data': [ 'alloc-track',
++            'blkdebug', 'blklogwrites', 'blkreplay', 'blkverify', 'bochs',
+             'cloop', 'compress', 'copy-before-write', 'copy-on-read', 'dmg',
+             'file', 'snapshot-access', 'ftp', 'ftps',
+             {'name': 'gluster', 'features': [ 'deprecated' ] },
+@@ -3668,6 +3669,21 @@
+ { 'struct': 'BlockdevOptionsNull',
+   'data': { '*size': 'int', '*latency-ns': 'uint64', '*read-zeroes': 'bool' } }
+ 
++##
++# @BlockdevOptionsAllocTrack:
++#
++# Driver specific block device options for the alloc-track backend.
++#
++# @backing: backing file with the data.
++#
++# @auto-remove: whether the alloc-track driver should drop itself
++#     after completing the stream.
++#
++##
++{ 'struct': 'BlockdevOptionsAllocTrack',
++  'base': 'BlockdevOptionsGenericFormat',
++  'data': { 'auto-remove': 'bool', 'backing': 'BlockdevRefOrNull' } }
++
+ ##
+ # @BlockdevOptionsPbs:
+ #
+@@ -5114,6 +5130,7 @@
+             '*detect-zeroes': 'BlockdevDetectZeroesOptions' },
+   'discriminator': 'driver',
+   'data': {
++      'alloc-track':'BlockdevOptionsAllocTrack',
+       'blkdebug':   'BlockdevOptionsBlkdebug',
+       'blklogwrites':'BlockdevOptionsBlklogwrites',
+       'blkverify':  'BlockdevOptionsBlkverify',
+-- 
+2.39.5
+

debian/patches/pve/0059-block-qapi-include-child-references-in-block-device-.patch

@@ -0,0 +1,137 @@
+From 76442f3eafa8cbe647fe2d39e78e817ec681143c Mon Sep 17 00:00:00 2001
+From: Fiona Ebner <f.ebner@proxmox.com>
+Date: Wed, 2 Jul 2025 18:27:37 +0200
+Subject: [PATCH 59/59] block/qapi: include child references in block device
+ info
+MIME-Version: 1.0
+Content-Type: text/plain; charset=UTF-8
+Content-Transfer-Encoding: 8bit
+
+In combination with using a throttle filter to enforce IO limits for
+a guest device, knowing the 'file' child of a block device can be
+useful. If the throttle filter is only intended for guest IO, block
+jobs should not also be limited by the throttle filter, so the
+block operations need to be done with the 'file' child of the top
+throttle node as the target. In combination with mirroring, the name
+of that child is not fixed.
+
+Another scenario is when unplugging a guest device after mirroring
+below a top throttle node, where the mirror target is added explicitly
+via blockdev-add. After mirroring, the target becomes the new 'file'
+child of the throttle node. For unplugging, both the top throttle node
+and the mirror target need to be deleted, because only implicitly
+added child nodes are deleted automatically, and the current 'file'
+child of the throttle node was explicitly added (as the mirror
+target).
+
+In other scenarios, it could be useful to follow the backing chain.
+
+Note that iotests 191 and 273 use _filter_img_info, so the 'children'
+information is filtered out there.
+
+Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
+Signed-off-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
+---
+ block/qapi.c               | 10 ++++++++++
+ qapi/block-core.json       | 16 ++++++++++++++++
+ tests/qemu-iotests/184.out |  8 ++++++++
+ 3 files changed, 34 insertions(+)
+
+diff --git a/block/qapi.c b/block/qapi.c
+index 2c50a6bf3b..e08a1e970f 100644
+--- a/block/qapi.c
++++ b/block/qapi.c
+@@ -51,6 +51,8 @@ BlockDeviceInfo *bdrv_block_device_info(BlockBackend *blk,
+     ImageInfo *backing_info;
+     BlockDriverState *backing;
+     BlockDeviceInfo *info;
++    BlockdevChildList **children_list_tail;
++    BdrvChild *child;
+ 
+     if (!bs->drv) {
+         error_setg(errp, "Block device %s is ejected", bs->node_name);
+@@ -77,6 +79,14 @@ BlockDeviceInfo *bdrv_block_device_info(BlockBackend *blk,
+         info->node_name = g_strdup(bs->node_name);
+     }
+ 
++    children_list_tail = &info->children;
++    QLIST_FOREACH(child, &bs->children, next) {
++        BlockdevChild *child_ref = g_new0(BlockdevChild, 1);
++        child_ref->child = g_strdup(child->name);
++        child_ref->node_name = g_strdup(child->bs->node_name);
++        QAPI_LIST_APPEND(children_list_tail, child_ref);
++    }
++
+     backing = bdrv_cow_bs(bs);
+     if (backing) {
+         info->backing_file = g_strdup(backing->filename);
+diff --git a/qapi/block-core.json b/qapi/block-core.json
+index 07c5773717..4db27f5819 100644
+--- a/qapi/block-core.json
++++ b/qapi/block-core.json
+@@ -461,6 +461,19 @@
+             'direct': 'bool',
+             'no-flush': 'bool' } }
+ 
++##
++# @BlockdevChild:
++#
++# @child: The name of the child, for example 'file' or 'backing'.
++#
++# @node-name: The name of the child's block driver node.
++#
++# Since: 10.1
++##
++{ 'struct': 'BlockdevChild',
++  'data': { 'child': 'str',
++            'node-name': 'str' } }
++
+ ##
+ # @BlockDeviceInfo:
+ #
+@@ -486,6 +499,8 @@
+ # @backing_file_depth: number of files in the backing file chain
+ #     (since: 1.2)
+ #
++# @children: Information about child block nodes. (since: 10.1)
++#
+ # @active: true if the backend is active; typical cases for inactive backends
+ #     are on the migration source instance after migration completes and on the
+ #     destination before it completes. (since: 10.0)
+@@ -560,6 +575,7 @@
+ { 'struct': 'BlockDeviceInfo',
+   'data': { 'file': 'str', '*node-name': 'str', 'ro': 'bool', 'drv': 'str',
+             '*backing_file': 'str', 'backing_file_depth': 'int',
++            'children': ['BlockdevChild'],
+             'active': 'bool', 'encrypted': 'bool',
+             'detect_zeroes': 'BlockdevDetectZeroesOptions',
+             'bps': 'int', 'bps_rd': 'int', 'bps_wr': 'int',
+diff --git a/tests/qemu-iotests/184.out b/tests/qemu-iotests/184.out
+index 52692b6b3b..ef99bb2e9a 100644
+--- a/tests/qemu-iotests/184.out
++++ b/tests/qemu-iotests/184.out
+@@ -41,6 +41,12 @@ Testing:
+             },
+             "iops_wr": 0,
+             "ro": false,
++            "children": [
++                {
++                    "node-name": "disk0",
++                    "child": "file"
++                }
++            ],
+             "node-name": "throttle0",
+             "backing_file_depth": 1,
+             "drv": "throttle",
+@@ -69,6 +75,8 @@ Testing:
+             },
+             "iops_wr": 0,
+             "ro": false,
++            "children": [
++            ],
+             "node-name": "disk0",
+             "backing_file_depth": 0,
+             "drv": "null-co",
+-- 
+2.39.5
+

debian/patches/pve/0060-savevm-async-reuse-migration-blocker-check-for-snaps.patch

@@ -0,0 +1,150 @@
+From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
+From: Fiona Ebner <f.ebner@proxmox.com>
+Date: Wed, 18 Jun 2025 12:25:31 +0200
+Subject: [PATCH] savevm-async: reuse migration blocker check for
+ snapshots/hibernation
+
+Same rationale as with upstream QEMU commit 5aaac46793 ("migration:
+savevm: consult migration blockers"), migration and (async) snapshot
+are essentially the same operation and thus snapshot also needs to
+check for migration blockers. For example, this catches passed-through
+PCI devices, where the driver does not support migration and VirtIO-GL
+display, which also does not support migration yet.
+
+In the case of VirtIO-GL, there were crashes [0].
+
+However, the commit notes:
+
+> There is really no difference between live migration and savevm, except
+> that savevm does not require bdrv_invalidate_cache to be implemented
+> by all disks.  However, it is unlikely that savevm is used with anything
+> except qcow2 disks, so the penalty is small and worth the improvement
+> in catching bad usage of savevm.
+
+and for Proxmox VE, suspend-to-disk with VMDK does use savevm-async
+and would be broken by simply using migration_is_blocked(). To keep
+this working, introduce a new helper that filters blockers with the
+prefix used by the VMDK migration blocker.
+
+The function qemu_savevm_state_blocked() is called as part of
+savevm_async_is_blocked() so no check is lost with this
+patch. The helper is declared in migration/migration.c to be able to
+access the 'migration_blockers'.
+
+The VMDK blocker message is declared via a '#define', because using a
+'const char*' led to the linker to complain about multiple
+declarations. The message does not include the reference to the block
+node anymore, but users can still easily find a VMDK disk in the VM
+configuration.
+
+Note, this also "breaks" snapshot and hibernate with VNC clipboard by
+preventing it. Previously, this would "work", because the Proxmox VE
+API has no check yet, but the clipboard will be broken after rollback,
+in the sense that it cannot be used anymore, not just lost contents.
+So some users might consider adding the check here a breaking change
+even if it's technically correct to prevent snapshot and hibernate
+with VNC clipboard. But other users might rightfully complain about
+broken clipboard. And again, the check also prevents blockers from
+passed-through PCI devices, etc. so it seems worth tolerating that
+breakage.
+
+[0]: https://forum.proxmox.com/threads/136976/
+
+Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
+Message-ID: <20250618102531.57444-1-f.ebner@proxmox.com>
+---
+ block/vmdk.c                |  4 +---
+ include/migration/blocker.h |  2 ++
+ migration/migration.c       | 24 ++++++++++++++++++++++++
+ migration/migration.h       |  1 +
+ migration/savevm-async.c    |  2 +-
+ 5 files changed, 29 insertions(+), 4 deletions(-)
+
+diff --git a/block/vmdk.c b/block/vmdk.c
+index 2adec49912..80696a8d27 100644
+--- a/block/vmdk.c
++++ b/block/vmdk.c
+@@ -1402,9 +1402,7 @@ static int vmdk_open(BlockDriverState *bs, QDict *options, int flags,
+     qemu_co_mutex_init(&s->lock);
+ 
+     /* Disable migration when VMDK images are used */
+-    error_setg(&s->migration_blocker, "The vmdk format used by node '%s' "
+-               "does not support live migration",
+-               bdrv_get_device_or_node_name(bs));
++    error_setg(&s->migration_blocker, "%s", MIGRATION_BLOCKER_VMDK);
+     ret = migrate_add_blocker_normal(&s->migration_blocker, errp);
+     if (ret < 0) {
+         goto fail;
+diff --git a/include/migration/blocker.h b/include/migration/blocker.h
+index a687ac0efe..f36bfb2df1 100644
+--- a/include/migration/blocker.h
++++ b/include/migration/blocker.h
+@@ -18,6 +18,8 @@
+ 
+ #define MIG_MODE_ALL MIG_MODE__MAX
+ 
++#define MIGRATION_BLOCKER_VMDK "The vmdk format used by a disk does not support live migration"
++
+ /**
+  * @migrate_add_blocker - prevent all modes of migration from proceeding
+  *
+diff --git a/migration/migration.c b/migration/migration.c
+index 2f3430f440..ecad1aca32 100644
+--- a/migration/migration.c
++++ b/migration/migration.c
+@@ -2030,6 +2030,30 @@ bool migration_is_blocked(Error **errp)
+     return false;
+ }
+ 
++bool savevm_async_is_blocked(Error **errp)
++{
++    GSList *blockers = migration_blockers[migrate_mode()];
++
++    if (qemu_savevm_state_blocked(errp)) {
++        return true;
++    }
++
++    /*
++     * The limitation for VMDK images only applies to live-migration, not
++     * snapshots, see commit 5aaac46793 ("migration: savevm: consult migration
++     * blockers").
++     */
++    while (blockers) {
++        if (strcmp(error_get_pretty(blockers->data), MIGRATION_BLOCKER_VMDK)) {
++            error_propagate(errp, error_copy(blockers->data));
++            return true;
++        }
++        blockers = g_slist_next(blockers);
++    }
++
++    return false;
++}
++
+ /* Returns true if continue to migrate, or false if error detected */
+ static bool migrate_prepare(MigrationState *s, bool resume, Error **errp)
+ {
+diff --git a/migration/migration.h b/migration/migration.h
+index d53f7cad84..b772073572 100644
+--- a/migration/migration.h
++++ b/migration/migration.h
+@@ -531,6 +531,7 @@ int migration_call_notifiers(MigrationState *s, MigrationEventType type,
+ 
+ int migrate_init(MigrationState *s, Error **errp);
+ bool migration_is_blocked(Error **errp);
++bool savevm_async_is_blocked(Error **errp);
+ /* True if outgoing migration has entered postcopy phase */
+ bool migration_in_postcopy(void);
+ bool migration_postcopy_is_alive(MigrationStatus state);
+diff --git a/migration/savevm-async.c b/migration/savevm-async.c
+index 730b815494..6cb91dca27 100644
+--- a/migration/savevm-async.c
++++ b/migration/savevm-async.c
+@@ -375,7 +375,7 @@ void qmp_savevm_start(const char *statefile, Error **errp)
+         return;
+     }
+ 
+-    if (qemu_savevm_state_blocked(errp)) {
++    if (savevm_async_is_blocked(errp)) {
+         goto fail;
+     }
+ 

debian/patches/series

@@ -1,16 +1,5 @@
 extra/0001-monitor-qmp-fix-race-with-clients-disconnecting-earl.patch
-extra/0002-scsi-megasas-Internal-cdbs-have-16-byte-length.patch
-extra/0003-ide-avoid-potential-deadlock-when-draining-during-tr.patch
-extra/0004-migration-block-dirty-bitmap-fix-loading-bitmap-when.patch
-extra/0005-Revert-Revert-graph-lock-Disable-locking-for-now.patch
-extra/0006-migration-states-workaround-snapshot-performance-reg.patch
-extra/0007-Revert-x86-acpi-workaround-Windows-not-handling-name.patch
-extra/0008-target-i386-the-sgx_epc_get_section-stub-is-reachabl.patch
-extra/0009-ui-clipboard-mark-type-as-not-available-when-there-i.patch
-extra/0010-virtio-scsi-Attach-event-vq-notifier-with-no_poll.patch
-extra/0011-virtio-Re-enable-notifications-after-drain.patch
-extra/0012-qemu_init-increase-NOFILE-soft-limit-on-POSIX.patch
-extra/0013-virtio-blk-avoid-using-ioeventfd-state-in-irqfd-cond.patch
+extra/0002-ide-avoid-potential-deadlock-when-draining-during-tr.patch
 bitmap-mirror/0001-drive-mirror-add-support-for-sync-bitmap-mode-never.patch
 bitmap-mirror/0002-drive-mirror-add-support-for-conditional-and-always-.patch
 bitmap-mirror/0003-mirror-add-check-for-bitmap-mode-without-bitmap.patch
@@ -54,11 +43,26 @@ pve/0034-PVE-Migrate-dirty-bitmap-state-via-savevm.patch
 pve/0035-migration-block-dirty-bitmap-migrate-other-bitmaps-e.patch
 pve/0036-PVE-fall-back-to-open-iscsi-initiatorname.patch
 pve/0037-PVE-block-stream-increase-chunk-size.patch
-pve/0038-block-io-accept-NULL-qiov-in-bdrv_pad_request.patch
-pve/0039-block-add-alloc-track-driver.patch
-pve/0040-Revert-block-rbd-workaround-for-ceph-issue-53784.patch
-pve/0041-Revert-block-rbd-fix-handling-of-holes-in-.bdrv_co_b.patch
-pve/0042-Revert-block-rbd-implement-bdrv_co_block_status.patch
-pve/0043-alloc-track-fix-deadlock-during-drop.patch
-pve/0044-migration-for-snapshots-hold-the-BQL-during-setup-ca.patch
-pve/0045-savevm-async-don-t-hold-BQL-during-setup.patch
+pve/0038-block-add-alloc-track-driver.patch
+pve/0039-Revert-block-rbd-workaround-for-ceph-issue-53784.patch
+pve/0040-Revert-block-rbd-fix-handling-of-holes-in-.bdrv_co_b.patch
+pve/0041-Revert-block-rbd-implement-bdrv_co_block_status.patch
+pve/0042-PVE-backup-add-fleecing-option.patch
+pve/0043-adapt-machine-version-deprecation-for-Proxmox-VE.patch
+pve/0044-Revert-hpet-avoid-timer-storms-on-periodic-timers.patch
+pve/0045-Revert-hpet-store-full-64-bit-target-value-of-the-co.patch
+pve/0046-Revert-hpet-accept-64-bit-reads-and-writes.patch
+pve/0047-Revert-hpet-place-read-only-bits-directly-in-new_val.patch
+pve/0048-Revert-hpet-remove-unnecessary-variable-index.patch
+pve/0049-Revert-hpet-ignore-high-bits-of-comparator-in-32-bit.patch
+pve/0050-Revert-hpet-fix-and-cleanup-persistence-of-interrupt.patch
+pve/0051-PVE-backup-factor-out-helper-to-clear-backup-state-s.patch
+pve/0052-PVE-backup-factor-out-helper-to-initialize-backup-st.patch
+pve/0053-PVE-backup-add-target-ID-in-backup-state.patch
+pve/0054-PVE-backup-get-device-info-allow-caller-to-specify-f.patch
+pve/0055-PVE-backup-implement-backup-access-setup-and-teardow.patch
+pve/0056-PVE-backup-prepare-for-the-switch-to-using-blockdev-.patch
+pve/0057-block-zeroinit-support-using-as-blockdev-driver.patch
+pve/0058-block-alloc-track-support-using-as-blockdev-driver.patch
+pve/0059-block-qapi-include-child-references-in-block-device-.patch
+pve/0060-savevm-async-reuse-migration-blocker-check-for-snaps.patch

debian/rules

@@ -61,7 +61,6 @@ endif
 	    --disable-xen \
 	    --enable-curl \
 	    --enable-docs \
-	    --enable-glusterfs \
 	    --enable-gnutls \
 	    --enable-libiscsi \
 	    --enable-libusb \
@@ -131,8 +130,7 @@ binary-indep: build install
 binary-arch: build install
 	dh_testdir
 	dh_testroot
-	# exclude historic Changelog file, which stops at release 0.14
-	dh_installchangelogs --exclude=Changelog
+	dh_installchangelogs
 	dh_installdocs
 	dh_installexamples
 	dh_install

debian/source/lintian-overrides

@@ -1 +1,2 @@
 source-is-missing [roms/SLOF/*.oco]
+source-is-missing [linux-user/*/vdso-*.so]