bump version to 10.0.2-4

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

Makefile

@@ -56,8 +56,7 @@ $(BUILDDIR): submodule
 
 .PHONY: deb kvm
 deb kvm: $(DEBS)
-$(DEB_DBG): $(DEB)
-$(DEB): $(BUILDDIR)
+$(DEBS) &: $(BUILDDIR)
 	cd $(BUILDDIR); dpkg-buildpackage -b -us -uc
 	lintian $(DEBS)
 

debian/changelog

@@ -1,3 +1,73 @@
+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

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

@@ -38,7 +38,7 @@ Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
  5 files changed, 142 insertions(+), 28 deletions(-)
 
 diff --git a/block/mirror.c b/block/mirror.c
-index 2afe700b4d..c3d4be9b15 100644
+index a53582f17b..fafca1360e 100644
 --- a/block/mirror.c
 +++ b/block/mirror.c
 @@ -51,7 +51,7 @@ typedef struct MirrorBlockJob {
@@ -258,10 +258,10 @@ index 2afe700b4d..c3d4be9b15 100644
                       base_read_only, errp);
      if (!job) {
 diff --git a/blockdev.c b/blockdev.c
-index 6740663fda..38fa63155c 100644
+index 1d1f27cfff..ec45bbaa52 100644
 --- a/blockdev.c
 +++ b/blockdev.c
-@@ -2781,6 +2781,9 @@ 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,
@@ -271,7 +271,7 @@ index 6740663fda..38fa63155c 100644
                                     BlockMirrorBackingMode backing_mode,
                                     bool zero_target,
                                     bool has_speed, int64_t speed,
-@@ -2799,6 +2802,7 @@ static void blockdev_mirror_common(const char *job_id, BlockDriverState *bs,
+@@ -2815,6 +2818,7 @@ static void blockdev_mirror_common(const char *job_id, BlockDriverState *bs,
  {
      BlockDriverState *unfiltered_bs;
      int job_flags = JOB_DEFAULT;
@@ -279,7 +279,7 @@ index 6740663fda..38fa63155c 100644
  
      GLOBAL_STATE_CODE();
      GRAPH_RDLOCK_GUARD_MAINLOOP();
-@@ -2853,6 +2857,29 @@ static void blockdev_mirror_common(const char *job_id, BlockDriverState *bs,
+@@ -2869,6 +2873,29 @@ static void blockdev_mirror_common(const char *job_id, BlockDriverState *bs,
          sync = MIRROR_SYNC_MODE_FULL;
      }
  
@@ -309,7 +309,7 @@ index 6740663fda..38fa63155c 100644
      if (!replaces) {
          /* We want to mirror from @bs, but keep implicit filters on top */
          unfiltered_bs = bdrv_skip_implicit_filters(bs);
-@@ -2894,8 +2921,8 @@ static void blockdev_mirror_common(const char *job_id, BlockDriverState *bs,
+@@ -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,
@@ -320,7 +320,7 @@ index 6740663fda..38fa63155c 100644
                   on_source_error, on_target_error, unmap, filter_node_name,
                   copy_mode, errp);
  }
-@@ -3039,6 +3066,8 @@ void qmp_drive_mirror(DriveMirror *arg, Error **errp)
+@@ -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,
@@ -329,7 +329,7 @@ index 6740663fda..38fa63155c 100644
                             backing_mode, zero_target,
                             arg->has_speed, arg->speed,
                             arg->has_granularity, arg->granularity,
-@@ -3058,6 +3087,8 @@ void qmp_blockdev_mirror(const char *job_id,
+@@ -3074,6 +3103,8 @@ void qmp_blockdev_mirror(const char *job_id,
                           const char *device, const char *target,
                           const char *replaces,
                           MirrorSyncMode sync,
@@ -338,7 +338,7 @@ index 6740663fda..38fa63155c 100644
                           bool has_speed, int64_t speed,
                           bool has_granularity, uint32_t granularity,
                           bool has_buf_size, int64_t buf_size,
-@@ -3098,7 +3129,8 @@ void qmp_blockdev_mirror(const char *job_id,
+@@ -3114,7 +3145,8 @@ void qmp_blockdev_mirror(const char *job_id,
      }
  
      blockdev_mirror_common(job_id, bs, target_bs,
@@ -364,10 +364,10 @@ index eb2d92a226..f0c642b194 100644
                    BlockdevOnError on_source_error,
                    BlockdevOnError on_target_error,
 diff --git a/qapi/block-core.json b/qapi/block-core.json
-index fd3bcc1c17..48ba32049f 100644
+index b1937780e1..0e5f148d30 100644
 --- a/qapi/block-core.json
 +++ b/qapi/block-core.json
-@@ -2178,6 +2178,15 @@
+@@ -2182,6 +2182,15 @@
  #     destination (all the disk, only the sectors allocated in the
  #     topmost image, or only new I/O).
  #
@@ -383,7 +383,7 @@ index fd3bcc1c17..48ba32049f 100644
  # @granularity: granularity of the dirty bitmap, default is 64K if the
  #     image format doesn't have clusters, 4K if the clusters are
  #     smaller than that, else the cluster size.  Must be a power of 2
-@@ -2220,7 +2229,9 @@
+@@ -2224,7 +2233,9 @@
  { 'struct': 'DriveMirror',
    'data': { '*job-id': 'str', 'device': 'str', 'target': 'str',
              '*format': 'str', '*node-name': 'str', '*replaces': 'str',
@@ -394,7 +394,7 @@ index fd3bcc1c17..48ba32049f 100644
              '*speed': 'int', '*granularity': 'uint32',
              '*buf-size': 'int', '*on-source-error': 'BlockdevOnError',
              '*on-target-error': 'BlockdevOnError',
-@@ -2499,6 +2510,15 @@
+@@ -2503,6 +2514,15 @@
  #     destination (all the disk, only the sectors allocated in the
  #     topmost image, or only new I/O).
  #
@@ -410,7 +410,7 @@ index fd3bcc1c17..48ba32049f 100644
  # @granularity: granularity of the dirty bitmap, default is 64K if the
  #     image format doesn't have clusters, 4K if the clusters are
  #     smaller than that, else the cluster size.  Must be a power of 2
-@@ -2547,7 +2567,8 @@
+@@ -2551,7 +2571,8 @@
  { 'command': 'blockdev-mirror',
    'data': { '*job-id': 'str', 'device': 'str', 'target': 'str',
              '*replaces': 'str',
@@ -421,7 +421,7 @@ index fd3bcc1c17..48ba32049f 100644
              '*buf-size': 'int', '*on-source-error': 'BlockdevOnError',
              '*on-target-error': 'BlockdevOnError',
 diff --git a/tests/unit/test-block-iothread.c b/tests/unit/test-block-iothread.c
-index 20ed54f570..4f50a99334 100644
+index 2b358eaaa8..2a149fe021 100644
 --- a/tests/unit/test-block-iothread.c
 +++ b/tests/unit/test-block-iothread.c
 @@ -755,8 +755,8 @@ static void test_propagate_mirror(void)

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

@@ -24,7 +24,7 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  1 file changed, 18 insertions(+), 6 deletions(-)
 
 diff --git a/block/mirror.c b/block/mirror.c
-index c3d4be9b15..7b6f7c0068 100644
+index fafca1360e..05e738bcce 100644
 --- a/block/mirror.c
 +++ b/block/mirror.c
 @@ -694,8 +694,6 @@ static int mirror_exit_common(Job *job)

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

@@ -16,10 +16,10 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  1 file changed, 3 insertions(+)
 
 diff --git a/blockdev.c b/blockdev.c
-index 38fa63155c..204cf6fad1 100644
+index ec45bbaa52..9fab7ec554 100644
 --- a/blockdev.c
 +++ b/blockdev.c
-@@ -2878,6 +2878,9 @@ 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

@@ -16,7 +16,7 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  1 file changed, 4 insertions(+), 7 deletions(-)
 
 diff --git a/block/mirror.c b/block/mirror.c
-index 7b6f7c0068..2b1c07095d 100644
+index 05e738bcce..2a2a227f3b 100644
 --- a/block/mirror.c
 +++ b/block/mirror.c
 @@ -809,8 +809,8 @@ static int mirror_exit_common(Job *job)

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

@@ -21,7 +21,7 @@ Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
  3 files changed, 70 insertions(+), 59 deletions(-)
 
 diff --git a/block/mirror.c b/block/mirror.c
-index 2b1c07095d..f5787b380c 100644
+index 2a2a227f3b..87c0856979 100644
 --- a/block/mirror.c
 +++ b/block/mirror.c
 @@ -1763,31 +1763,13 @@ static BlockJob *mirror_start_job(
@@ -62,10 +62,10 @@ index 2b1c07095d..f5787b380c 100644
  
          if (bitmap_mode != BITMAP_SYNC_MODE_NEVER) {
 diff --git a/blockdev.c b/blockdev.c
-index 204cf6fad1..79d47b1920 100644
+index 9fab7ec554..158ac9314b 100644
 --- a/blockdev.c
 +++ b/blockdev.c
-@@ -2857,7 +2857,36 @@ 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

@@ -60,7 +60,7 @@ index c3740ec616..7f38ce6b8b 100644
  void monitor_init_globals(void);
  void monitor_init_globals_core(void);
 diff --git a/monitor/monitor-internal.h b/monitor/monitor-internal.h
-index cb628f681d..93dbd62fc2 100644
+index 5676eb334e..4c452a6aeb 100644
 --- a/monitor/monitor-internal.h
 +++ b/monitor/monitor-internal.h
 @@ -151,6 +151,13 @@ typedef struct {
@@ -78,7 +78,7 @@ index cb628f681d..93dbd62fc2 100644
  
  /**
 diff --git a/monitor/monitor.c b/monitor/monitor.c
-index 56786c0ccc..30071d0c8a 100644
+index c5a5d30877..07775784d4 100644
 --- a/monitor/monitor.c
 +++ b/monitor/monitor.c
 @@ -116,6 +116,21 @@ bool monitor_cur_is_qmp(void)
@@ -104,7 +104,7 @@ index 56786c0ccc..30071d0c8a 100644
   * Is @mon is using readline?
   * Note: not all HMP monitors use readline, e.g., gdbserver has a
 diff --git a/monitor/qmp.c b/monitor/qmp.c
-index 5e538f34c0..eb181d5979 100644
+index 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 5e538f34c0..eb181d5979 100644
          monitor_qmp_caps_reset(mon);
          data = qmp_greeting(mon);
 diff --git a/qapi/qmp-dispatch.c b/qapi/qmp-dispatch.c
-index 176b549473..790bb7d1da 100644
+index e569224eae..eb03782e91 100644
 --- a/qapi/qmp-dispatch.c
 +++ b/qapi/qmp-dispatch.c
 @@ -117,16 +117,28 @@ typedef struct QmpDispatchBH {

debian/patches/extra/0002-ide-avoid-potential-deadlock-when-draining-during-tr.patch

@@ -55,7 +55,7 @@ Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
  1 file changed, 6 insertions(+), 6 deletions(-)
 
 diff --git a/hw/ide/core.c b/hw/ide/core.c
-index 08d9218455..20d8c0cf66 100644
+index b14983ec54..41c543e627 100644
 --- a/hw/ide/core.c
 +++ b/hw/ide/core.c
 @@ -456,7 +456,7 @@ static void ide_trim_bh_cb(void *opaque)

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

@@ -14,10 +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 90fa54352c..e2ea071315 100644
+index 56d1972d15..cfa0b832ba 100644
 --- a/block/file-posix.c
 +++ b/block/file-posix.c
-@@ -564,7 +564,7 @@ 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 90fa54352c..e2ea071315 100644
          },
          {
              .name = "pr-manager",
-@@ -664,7 +664,7 @@ 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/0003-PVE-Config-set-the-CPU-model-to-kvm64-32-instead-of-.patch

@@ -10,10 +10,10 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  1 file changed, 2 insertions(+), 2 deletions(-)
 
 diff --git a/target/i386/cpu.h b/target/i386/cpu.h
-index 4c239a6970..be09263fb0 100644
+index 76f24446a5..2a47d79b49 100644
 --- a/target/i386/cpu.h
 +++ b/target/i386/cpu.h
-@@ -2475,9 +2475,9 @@ 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

@@ -9,7 +9,7 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  1 file changed, 9 insertions(+), 6 deletions(-)
 
 diff --git a/ui/spice-core.c b/ui/spice-core.c
-index bd9dbe03f1..a7ecaad9c7 100644
+index 0326c63bec..d523d00200 100644
 --- a/ui/spice-core.c
 +++ b/ui/spice-core.c
 @@ -690,32 +690,35 @@ static void qemu_spice_init(void)

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

@@ -9,7 +9,7 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  1 file changed, 11 insertions(+), 4 deletions(-)
 
 diff --git a/block/gluster.c b/block/gluster.c
-index e9c038042b..c8457a5014 100644
+index c6d25ae733..ccca125c3a 100644
 --- a/block/gluster.c
 +++ b/block/gluster.c
 @@ -42,7 +42,7 @@

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

@@ -18,7 +18,7 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  1 file changed, 2 insertions(+)
 
 diff --git a/block/rbd.c b/block/rbd.c
-index 04ed0e242e..728bce3b1e 100644
+index 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

@@ -16,7 +16,7 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  1 file changed, 9 insertions(+), 1 deletion(-)
 
 diff --git a/block/gluster.c b/block/gluster.c
-index c8457a5014..c3a9555591 100644
+index ccca125c3a..301a653ea7 100644
 --- a/block/gluster.c
 +++ b/block/gluster.c
 @@ -57,6 +57,7 @@ typedef struct GlusterAIOCB {

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

@@ -9,7 +9,7 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  1 file changed, 2 insertions(+), 1 deletion(-)
 
 diff --git a/qemu-img.c b/qemu-img.c
-index 7668f86769..2575e97b43 100644
+index 2044c22a4c..4c8b5412c6 100644
 --- a/qemu-img.c
 +++ b/qemu-img.c
 @@ -3075,7 +3075,8 @@ static int img_info(int argc, char **argv)

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

@@ -54,10 +54,10 @@ index c9dd70a892..048788b23d 100644
  
  DEF("info", img_info,
 diff --git a/qemu-img.c b/qemu-img.c
-index 2575e97b43..8ec68b346f 100644
+index 4c8b5412c6..d5fa89a204 100644
 --- a/qemu-img.c
 +++ b/qemu-img.c
-@@ -4993,10 +4993,12 @@ 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 2575e97b43..8ec68b346f 100644
  };
  
  struct DdIo {
-@@ -5072,6 +5074,19 @@ 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 2575e97b43..8ec68b346f 100644
  static int img_dd(int argc, char **argv)
  {
      int ret = 0;
-@@ -5112,6 +5127,7 @@ static int img_dd(int argc, char **argv)
+@@ -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 2575e97b43..8ec68b346f 100644
          { NULL, NULL, 0 }
      };
      const struct option long_options[] = {
-@@ -5187,91 +5203,112 @@ 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 2575e97b43..8ec68b346f 100644
      }
  
      if (dd.flags & C_SKIP && (in.offset > INT64_MAX / in.bsz ||
-@@ -5288,20 +5325,43 @@ static int img_dd(int argc, char **argv)
+@@ -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 8ec68b346f..b98184bba1 100644
+index d5fa89a204..d458e85af2 100644
 --- a/qemu-img.c
 +++ b/qemu-img.c
-@@ -4994,11 +4994,13 @@ 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 8ec68b346f..b98184bba1 100644
  };
  
  struct DdIo {
-@@ -5087,6 +5089,19 @@ 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 8ec68b346f..b98184bba1 100644
  static int img_dd(int argc, char **argv)
  {
      int ret = 0;
-@@ -5101,12 +5116,14 @@ static int img_dd(int argc, char **argv)
+@@ -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 8ec68b346f..b98184bba1 100644
      };
      struct DdIo in = {
          .bsz = 512, /* Block size is by default 512 bytes */
-@@ -5128,6 +5145,7 @@ static int img_dd(int argc, char **argv)
+@@ -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 8ec68b346f..b98184bba1 100644
          { NULL, NULL, 0 }
      };
      const struct option long_options[] = {
-@@ -5324,9 +5342,10 @@ 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 8ec68b346f..b98184bba1 100644
          if (blk1) {
              in_ret = blk_pread(blk1, in_pos, bytes, in.buf, 0);
              if (in_ret == 0) {
-@@ -5335,6 +5354,9 @@ static int img_dd(int argc, char **argv)
+@@ -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

@@ -65,10 +65,10 @@ index 048788b23d..0b29a67a06 100644
  
  DEF("info", img_info,
 diff --git a/qemu-img.c b/qemu-img.c
-index b98184bba1..6fc8384f64 100644
+index d458e85af2..dc13efba8b 100644
 --- a/qemu-img.c
 +++ b/qemu-img.c
-@@ -5118,7 +5118,7 @@ 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 b98184bba1..6fc8384f64 100644
      struct DdInfo dd = {
          .flags = 0,
          .count = 0,
-@@ -5156,7 +5156,7 @@ 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 b98184bba1..6fc8384f64 100644
          if (c == EOF) {
              break;
          }
-@@ -5176,6 +5176,9 @@ 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 b98184bba1..6fc8384f64 100644
          case 'U':
              force_share = true;
              break;
-@@ -5306,13 +5309,15 @@ 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

@@ -46,10 +46,10 @@ index 0b29a67a06..758f397232 100644
  
  DEF("info", img_info,
 diff --git a/qemu-img.c b/qemu-img.c
-index 6fc8384f64..a6c88e0860 100644
+index dc13efba8b..02f2e0aa45 100644
 --- a/qemu-img.c
 +++ b/qemu-img.c
-@@ -5110,6 +5110,7 @@ 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 6fc8384f64..a6c88e0860 100644
      QemuOptsList *create_opts = NULL;
      Error *local_err = NULL;
      bool image_opts = false;
-@@ -5119,6 +5120,7 @@ static int img_dd(int argc, char **argv)
+@@ -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 6fc8384f64..a6c88e0860 100644
      struct DdInfo dd = {
          .flags = 0,
          .count = 0,
-@@ -5156,7 +5158,7 @@ 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 6fc8384f64..a6c88e0860 100644
          if (c == EOF) {
              break;
          }
-@@ -5179,6 +5181,19 @@ 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 6fc8384f64..a6c88e0860 100644
          case 'U':
              force_share = true;
              break;
-@@ -5238,11 +5253,24 @@ 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 6fc8384f64..a6c88e0860 100644
      }
  
      if (dd.flags & C_OSIZE) {
-@@ -5397,6 +5425,7 @@ 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

@@ -18,7 +18,7 @@ Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
  4 files changed, 82 insertions(+), 4 deletions(-)
 
 diff --git a/hw/core/machine-hmp-cmds.c b/hw/core/machine-hmp-cmds.c
-index 8701f00cc7..3b4c5ef403 100644
+index c6325cdcaa..7f817d622d 100644
 --- a/hw/core/machine-hmp-cmds.c
 +++ b/hw/core/machine-hmp-cmds.c
 @@ -179,7 +179,35 @@ void hmp_info_balloon(Monitor *mon, const QDict *qdict)
@@ -59,7 +59,7 @@ index 8701f00cc7..3b4c5ef403 100644
      qapi_free_BalloonInfo(info);
  }
 diff --git a/hw/virtio/virtio-balloon.c b/hw/virtio/virtio-balloon.c
-index afd2ad6dd6..c724218c17 100644
+index 2eb5a14fa2..aa2fd6c32f 100644
 --- a/hw/virtio/virtio-balloon.c
 +++ b/hw/virtio/virtio-balloon.c
 @@ -795,8 +795,37 @@ static uint64_t virtio_balloon_get_features(VirtIODevice *vdev, uint64_t f,

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

@@ -13,10 +13,10 @@ Signed-off-by: Dietmar Maurer <dietmar@proxmox.com>
  2 files changed, 9 insertions(+), 1 deletion(-)
 
 diff --git a/hw/core/machine-qmp-cmds.c b/hw/core/machine-qmp-cmds.c
-index 130217da8f..52a6d74820 100644
+index 1bc21b84a4..93fb4bc24a 100644
 --- a/hw/core/machine-qmp-cmds.c
 +++ b/hw/core/machine-qmp-cmds.c
-@@ -90,6 +90,12 @@ MachineInfoList *qmp_query_machines(bool has_compat_props, bool compat_props,
+@@ -91,6 +91,12 @@ MachineInfoList *qmp_query_machines(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");
@@ -26,8 +26,8 @@ index 130217da8f..52a6d74820 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 9f7ed0eaa0..16366b774a 100644

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

@@ -14,7 +14,7 @@ Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
  2 files changed, 7 insertions(+)
 
 diff --git a/qapi/ui.json b/qapi/ui.json
-index 460a26b981..42b911bda3 100644
+index c536d4e524..c2df48959b 100644
 --- a/qapi/ui.json
 +++ b/qapi/ui.json
 @@ -312,11 +312,14 @@
@@ -33,7 +33,7 @@ index 460a26b981..42b911bda3 100644
    'if': 'CONFIG_SPICE' }
  
 diff --git a/ui/spice-core.c b/ui/spice-core.c
-index a7ecaad9c7..fecf002d50 100644
+index 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

@@ -25,7 +25,7 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
 
 diff --git a/migration/channel-savevm-async.c b/migration/channel-savevm-async.c
 new file mode 100644
-index 0000000000..081a192f49
+index 0000000000..e57ab2ae40
 --- /dev/null
 +++ b/migration/channel-savevm-async.c
 @@ -0,0 +1,184 @@
@@ -35,7 +35,7 @@ index 0000000000..081a192f49
 +#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 *
@@ -271,14 +271,14 @@ index 0000000000..17ae2cb261
 +
 +#endif /* QIO_CHANNEL_SAVEVM_ASYNC_H */
 diff --git a/migration/meson.build b/migration/meson.build
-index d53cf3417a..b00d58064d 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',
+   'cpr.c',
+   'cpr-transfer.c',
    'cpu-throttle.c',
-   'dirtyrate.c',
-   'exec.c',

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

@@ -30,21 +30,24 @@ Signed-off-by: Stefan Reiter <s.reiter@proxmox.com>
      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]
+     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     | 571 +++++++++++++++++++++++++++++++++++
+ migration/savevm-async.c     | 581 +++++++++++++++++++++++++++++++++++
  monitor/hmp-cmds.c           |  38 +++
- qapi/migration.json          |  34 +++
+ qapi/migration.json          |  34 ++
  qapi/misc.json               |  18 ++
  qemu-options.hx              |  12 +
  system/vl.c                  |  10 +
- 11 files changed, 719 insertions(+)
+ 11 files changed, 729 insertions(+)
  create mode 100644 migration/savevm-async.c
 
 diff --git a/hmp-commands-info.hx b/hmp-commands-info.hx
@@ -129,10 +132,10 @@ index ae116d9804..2596cc2426 100644
  void coroutine_fn hmp_screendump(Monitor *mon, const QDict *qdict);
  void hmp_chardev_add(Monitor *mon, const QDict *qdict);
 diff --git a/migration/meson.build b/migration/meson.build
-index b00d58064d..075b013971 100644
+index cf66c78681..46e92249a1 100644
 --- a/migration/meson.build
 +++ b/migration/meson.build
-@@ -29,6 +29,7 @@ system_ss.add(files(
+@@ -33,6 +33,7 @@ system_ss.add(files(
    'options.c',
    'postcopy-ram.c',
    'savevm.c',
@@ -142,10 +145,10 @@ index b00d58064d..075b013971 100644
    'threadinfo.c',
 diff --git a/migration/savevm-async.c b/migration/savevm-async.c
 new file mode 100644
-index 0000000000..ee8ef316d0
+index 0000000000..56e0fa6c69
 --- /dev/null
 +++ b/migration/savevm-async.c
-@@ -0,0 +1,571 @@
+@@ -0,0 +1,581 @@
 +#include "qemu/osdep.h"
 +#include "migration/channel-savevm-async.h"
 +#include "migration/migration.h"
@@ -156,14 +159,14 @@ index 0000000000..ee8ef316d0
 +#include "migration/global_state.h"
 +#include "migration/ram.h"
 +#include "migration/qemu-file.h"
-+#include "sysemu/cpu-throttle.h"
-+#include "sysemu/sysemu.h"
-+#include "sysemu/runstate.h"
++#include "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"
@@ -173,6 +176,7 @@ index 0000000000..ee8ef316d0
 +#include "qemu/main-loop.h"
 +#include "qemu/rcu.h"
 +#include "qemu/yank.h"
++#include "system/iothread.h"
 +
 +/* #define DEBUG_SAVEVM_STATE */
 +
@@ -199,12 +203,13 @@ index 0000000000..ee8ef316d0
 +    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)
@@ -326,6 +331,7 @@ index 0000000000..ee8ef316d0
 +     */
 +    blk_set_aio_context(snap_state.target, qemu_get_aio_context(), NULL);
 +
++    snap_state.vm_needs_start = runstate_is_running();
 +    ret = vm_stop_force_state(RUN_STATE_FINISH_MIGRATE);
 +    if (ret < 0) {
 +        save_snapshot_error("vm_stop_force_state error %d", ret);
@@ -333,7 +339,7 @@ index 0000000000..ee8ef316d0
 +
 +    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);
@@ -372,9 +378,9 @@ index 0000000000..ee8ef316d0
 +        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",
@@ -403,16 +409,13 @@ index 0000000000..ee8ef316d0
 +        uint64_t threshold = 400 * 1000;
 +
 +        /*
-+         * pending_{estimate,exact} are expected to be called without iothread
-+         * lock. Similar to what is done in migration.c, call the exact variant
-+         * only once pend_precopy in the estimate is below the threshold.
++         * Similar to what is done in migration.c, call the exact variant only
++         * once pend_precopy in the estimate is below the threshold.
 +         */
-+        bql_unlock();
 +        qemu_savevm_state_pending_estimate(&pend_precopy, &pend_postcopy);
 +        if (pend_precopy <= threshold) {
 +            qemu_savevm_state_pending_exact(&pend_precopy, &pend_postcopy);
 +        }
-+        bql_lock();
 +        pending_size = pend_precopy + pend_postcopy;
 +
 +        /*
@@ -479,11 +482,17 @@ index 0000000000..ee8ef316d0
 +    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;
 +
@@ -500,7 +509,6 @@ index 0000000000..ee8ef316d0
 +    }
 +
 +    /* 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;
@@ -512,13 +520,27 @@ index 0000000000..ee8ef316d0
 +    }
 +
 +    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 */
@@ -528,12 +550,12 @@ index 0000000000..ee8ef316d0
 +    snap_state.target = blk_new_open(statefile, NULL, options, bdrv_oflags, &local_err);
 +    if (!snap_state.target) {
 +        error_setg(errp, "failed to open '%s'", statefile);
-+        goto restart;
++        goto fail;
 +    }
 +    target_bs = blk_bs(snap_state.target);
 +    if (!target_bs) {
 +        error_setg(errp, "failed to open '%s' - no block driver state", statefile);
-+        goto restart;
++        goto fail;
 +    }
 +
 +    QIOChannel *ioc = QIO_CHANNEL(qio_channel_savevm_async_new(snap_state.target,
@@ -542,7 +564,7 @@ index 0000000000..ee8ef316d0
 +
 +    if (!snap_state.file) {
 +        error_setg(errp, "failed to open '%s'", statefile);
-+        goto restart;
++        goto fail;
 +    }
 +
 +    /*
@@ -550,8 +572,8 @@ index 0000000000..ee8ef316d0
 +     * State is cleared in process_savevm_co, but has to be initialized
 +     * here (blocking main thread, from QMP) to avoid race conditions.
 +     */
-+    if (migrate_init(ms, errp)) {
-+        return;
++    if (migrate_init(ms, errp) != 0) {
++        goto fail;
 +    }
 +    memset(&mig_stats, 0, sizeof(mig_stats));
 +    ms->to_dst_file = snap_state.file;
@@ -566,56 +588,49 @@ index 0000000000..ee8ef316d0
 +    if (ret != 0) {
 +        error_setg_errno(errp, -ret, "savevm state setup failed: %s",
 +                         local_err ? error_get_pretty(local_err) : "unknown error");
-+        return;
++        goto fail;
 +    }
 +
-+    /* Async processing from here on out happens in iohandler context, so let
-+     * the target bdrv have its home there.
-+     */
-+    ret = blk_set_aio_context(snap_state.target, iohandler_ctx, &local_err);
++    ret = blk_set_aio_context(snap_state.target, snap_state.iothread->ctx, &local_err);
 +    if (ret != 0) {
-+        warn_report("failed to set iohandler context for VM state target: %s %s",
-+                    local_err ? error_get_pretty(local_err) : "unknown error",
-+                    strerror(-ret));
++        error_setg_errno(errp, -ret, "failed to set iothread context for VM state target: %s",
++                         local_err ? error_get_pretty(local_err) : "unknown error");
++        goto fail;
 +    }
 +
 +    snap_state.co = qemu_coroutine_create(&process_savevm_co, NULL);
-+    aio_co_schedule(iohandler_ctx, snap_state.co);
++    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;
-+    }
 +}
 +
 +static void coroutine_fn wait_for_close_co(void *opaque)
 +{
 +    int64_t timeout;
 +
-+    if (!snap_state.target) {
++    if (snap_state.target) {
++        /* 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");
-+        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;
-+    }
++    savevm_cleanup_iothread();
 +
 +    // File closed and no other error, so ensure next snapshot can be started.
 +    if (snap_state.state != SAVE_STATE_ERROR) {
@@ -640,13 +655,11 @@ index 0000000000..ee8ef316d0
 +        return;
 +    }
 +
-+    if (snap_state.saved_vm_running) {
++    if (snap_state.vm_needs_start) {
 +        vm_start();
-+        snap_state.saved_vm_running = false;
++        snap_state.vm_needs_start = false;
 +    }
 +
-+    snap_state.state = SAVE_STATE_DONE;
-+
 +    qemu_coroutine_enter(wait_for_close);
 +}
 +
@@ -718,7 +731,7 @@ index 0000000000..ee8ef316d0
 +    return ret;
 +}
 diff --git a/monitor/hmp-cmds.c b/monitor/hmp-cmds.c
-index f601d06ab8..874084565f 100644
+index 7ded3378cf..bade2a4b92 100644
 --- a/monitor/hmp-cmds.c
 +++ b/monitor/hmp-cmds.c
 @@ -24,6 +24,7 @@
@@ -727,10 +740,10 @@ index f601d06ab8..874084565f 100644
  #include "qapi/qapi-commands-machine.h"
 +#include "qapi/qapi-commands-migration.h"
  #include "qapi/qapi-commands-misc.h"
- #include "qapi/qmp/qdict.h"
+ #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", filename);
+     monitor_printf(mon, "DTB dumped to '%s'\n", filename);
  }
  #endif
 +
@@ -771,10 +784,10 @@ index f601d06ab8..874084565f 100644
 +    }
 +}
 diff --git a/qapi/migration.json b/qapi/migration.json
-index a605dc26db..927b1e1c7d 100644
+index 8b9c53595c..ff3479da65 100644
 --- a/qapi/migration.json
 +++ b/qapi/migration.json
-@@ -276,6 +276,40 @@
+@@ -279,6 +279,40 @@
             '*dirty-limit-throttle-time-per-round': 'uint64',
             '*dirty-limit-ring-full-time': 'uint64'} }
  
@@ -845,10 +858,10 @@ index 559b66f201..7959e89c1e 100644
  # @CommandLineParameterType:
  #
 diff --git a/qemu-options.hx b/qemu-options.hx
-index dacc9790a4..c05f411599 100644
+index dc694a99a3..defee0c06a 100644
 --- a/qemu-options.hx
 +++ b/qemu-options.hx
-@@ -4764,6 +4764,18 @@ SRST
+@@ -4862,6 +4862,18 @@ SRST
      Start right away with a saved state (``loadvm`` in monitor)
  ERST
  
@@ -868,10 +881,10 @@ index dacc9790a4..c05f411599 100644
  DEF("daemonize", 0, QEMU_OPTION_daemonize, \
      "-daemonize      daemonize QEMU after initializing\n", QEMU_ARCH_ALL)
 diff --git a/system/vl.c b/system/vl.c
-index 2f855d83fb..39d451bb41 100644
+index ec93988a03..9b36ace6b4 100644
 --- a/system/vl.c
 +++ b/system/vl.c
-@@ -164,6 +164,7 @@ static const char *accelerators;
+@@ -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;
@@ -879,7 +892,7 @@ index 2f855d83fb..39d451bb41 100644
  static QTAILQ_HEAD(, ObjectOption) object_opts = QTAILQ_HEAD_INITIALIZER(object_opts);
  static QTAILQ_HEAD(, DeviceOption) device_opts = QTAILQ_HEAD_INITIALIZER(device_opts);
  static int display_remote;
-@@ -2725,6 +2726,12 @@ void qmp_x_exit_preconfig(Error **errp)
+@@ -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);
@@ -892,7 +905,7 @@ index 2f855d83fb..39d451bb41 100644
      }
      if (replay_mode != REPLAY_MODE_NONE) {
          replay_vmstate_init();
-@@ -3262,6 +3269,9 @@ void qemu_init(int argc, char **argv)
+@@ -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

@@ -19,7 +19,7 @@ Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
  3 files changed, 38 insertions(+), 17 deletions(-)
 
 diff --git a/migration/qemu-file.c b/migration/qemu-file.c
-index b6d2f588bd..754dc0b3f7 100644
+index 1303a5bf58..6e2d58d5c0 100644
 --- a/migration/qemu-file.c
 +++ b/migration/qemu-file.c
 @@ -34,8 +34,8 @@
@@ -31,9 +31,9 @@ index b6d2f588bd..754dc0b3f7 100644
 +#define DEFAULT_IO_BUF_SIZE 32768
 +#define MAX_IOV_SIZE MIN_CONST(IOV_MAX, 256)
  
- struct QEMUFile {
-     QIOChannel *ioc;
-@@ -43,7 +43,8 @@ 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 b6d2f588bd..754dc0b3f7 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 b6d2f588bd..754dc0b3f7 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)
  {
@@ -94,17 +94,17 @@ index b6d2f588bd..754dc0b3f7 100644
  }
  
  /*
-@@ -327,7 +342,7 @@ 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()) {
-@@ -367,6 +382,9 @@ int qemu_fclose(QEMUFile *f)
-         ret = ret2;
+@@ -443,6 +458,9 @@ int qemu_fclose(QEMUFile *f)
+         g_free(fde);
      }
      g_clear_pointer(&f->ioc, object_unref);
 +
@@ -113,7 +113,7 @@ index b6d2f588bd..754dc0b3f7 100644
      error_free(f->last_error_obj);
      g_free(f);
      trace_qemu_file_fclose();
-@@ -415,7 +433,7 @@ static void add_buf_to_iovec(QEMUFile *f, size_t len)
+@@ -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;
@@ -122,7 +122,7 @@ index b6d2f588bd..754dc0b3f7 100644
              qemu_fflush(f);
          }
      }
-@@ -440,7 +458,7 @@ void qemu_put_buffer(QEMUFile *f, const uint8_t *buf, size_t size)
+@@ -516,7 +534,7 @@ void qemu_put_buffer(QEMUFile *f, const uint8_t *buf, size_t size)
      }
  
      while (size > 0) {
@@ -131,7 +131,7 @@ index b6d2f588bd..754dc0b3f7 100644
          if (l > size) {
              l = size;
          }
-@@ -586,8 +604,8 @@ size_t coroutine_mixed_fn qemu_peek_buffer(QEMUFile *f, uint8_t **buf, size_t si
+@@ -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));
@@ -142,7 +142,7 @@ index b6d2f588bd..754dc0b3f7 100644
  
      /* The 1st byte to read from */
      index = f->buf_index + offset;
-@@ -637,7 +655,7 @@ size_t coroutine_mixed_fn qemu_get_buffer(QEMUFile *f, uint8_t *buf, size_t size
+@@ -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;
  
@@ -151,7 +151,7 @@ index b6d2f588bd..754dc0b3f7 100644
          if (res == 0) {
              return done;
          }
-@@ -671,7 +689,7 @@ size_t coroutine_mixed_fn qemu_get_buffer(QEMUFile *f, uint8_t *buf, size_t size
+@@ -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)
  {
@@ -160,7 +160,7 @@ index b6d2f588bd..754dc0b3f7 100644
          size_t res;
          uint8_t *src = NULL;
  
-@@ -696,7 +714,7 @@ int coroutine_mixed_fn qemu_peek_byte(QEMUFile *f, int offset)
+@@ -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));
@@ -170,7 +170,7 @@ index b6d2f588bd..754dc0b3f7 100644
      if (index >= f->buf_size) {
          qemu_fill_buffer(f);
 diff --git a/migration/qemu-file.h b/migration/qemu-file.h
-index 11c2120edd..edf3c5d147 100644
+index f5b9f430e0..0179b90698 100644
 --- a/migration/qemu-file.h
 +++ b/migration/qemu-file.h
 @@ -30,7 +30,9 @@
@@ -182,12 +182,12 @@ index 11c2120edd..edf3c5d147 100644
 +QEMUFile *qemu_file_new_output_sized(QIOChannel *ioc, size_t buffer_size);
  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 ee8ef316d0..1e79fce9ba 100644
+index 56e0fa6c69..730b815494 100644
 --- a/migration/savevm-async.c
 +++ b/migration/savevm-async.c
-@@ -390,7 +390,7 @@ 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));
@@ -196,7 +196,7 @@ index ee8ef316d0..1e79fce9ba 100644
  
      if (!snap_state.file) {
          error_setg(errp, "failed to open '%s'", statefile);
-@@ -534,7 +534,8 @@ int load_snapshot_from_blockdev(const char *filename, Error **errp)
+@@ -544,7 +544,8 @@ int load_snapshot_from_blockdev(const char *filename, Error **errp)
      bdrv_op_block_all(bs, blocker);
  
      /* restore the VM state */

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

@@ -15,7 +15,7 @@ Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
  create mode 100644 block/zeroinit.c
 
 diff --git a/block/meson.build b/block/meson.build
-index f1262ec2ba..6a60b5d6b9 100644
+index 34b1b2a306..a21d9a5411 100644
 --- a/block/meson.build
 +++ b/block/meson.build
 @@ -39,6 +39,7 @@ block_ss.add(files(
@@ -28,7 +28,7 @@ index f1262ec2ba..6a60b5d6b9 100644
  system_ss.add(when: 'CONFIG_TCG', if_true: files('blkreplay.c'))
 diff --git a/block/zeroinit.c b/block/zeroinit.c
 new file mode 100644
-index 0000000000..2b2b194ccf
+index 0000000000..f9d513db15
 --- /dev/null
 +++ b/block/zeroinit.c
 @@ -0,0 +1,207 @@
@@ -47,8 +47,8 @@ index 0000000000..2b2b194ccf
 +#include "block/block_int.h"
 +#include "block/block-io.h"
 +#include "block/graph-lock.h"
-+#include "qapi/qmp/qdict.h"
-+#include "qapi/qmp/qstring.h"
++#include "qobject/qdict.h"
++#include "qobject/qstring.h"
 +#include "qemu/cutils.h"
 +#include "qemu/option.h"
 +#include "qemu/module.h"

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

@@ -14,10 +14,10 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  2 files changed, 11 insertions(+)
 
 diff --git a/qemu-options.hx b/qemu-options.hx
-index c05f411599..0732077a0e 100644
+index defee0c06a..fb980a05cf 100644
 --- a/qemu-options.hx
 +++ b/qemu-options.hx
-@@ -1239,6 +1239,9 @@ legacy PC, they are not recommended for modern configurations.
+@@ -1280,6 +1280,9 @@ legacy PC, they are not recommended for modern configurations.
  
  ERST
  
@@ -28,10 +28,10 @@ index c05f411599..0732077a0e 100644
      "-fda/-fdb file  use 'file' as floppy disk 0/1 image\n", QEMU_ARCH_ALL)
  DEF("fdb", HAS_ARG, QEMU_OPTION_fdb, "", QEMU_ARCH_ALL)
 diff --git a/system/vl.c b/system/vl.c
-index 39d451bb41..e7cae51f13 100644
+index 9b36ace6b4..452742ab58 100644
 --- a/system/vl.c
 +++ b/system/vl.c
-@@ -2762,6 +2762,7 @@ void qemu_init(int argc, char **argv)
+@@ -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 39d451bb41..e7cae51f13 100644
  
      qemu_add_opts(&qemu_drive_opts);
      qemu_add_drive_opts(&qemu_legacy_drive_opts);
-@@ -3374,6 +3375,13 @@ void qemu_init(int argc, char **argv)
+@@ -3472,6 +3473,13 @@ void qemu_init(int argc, char **argv)
                  machine_parse_property_opt(qemu_find_opts("smp-opts"),
                                             "smp", optarg);
                  break;

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

@@ -11,7 +11,7 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  1 file changed, 9 insertions(+)
 
 diff --git a/hw/intc/apic_common.c b/hw/intc/apic_common.c
-index 62f3bbf203..89e0c7d995 100644
+index 2a3e878c4d..efbed1aea3 100644
 --- a/hw/intc/apic_common.c
 +++ b/hw/intc/apic_common.c
 @@ -263,6 +263,15 @@ static void apic_reset_common(DeviceState *dev)

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

@@ -13,10 +13,10 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  2 files changed, 46 insertions(+), 20 deletions(-)
 
 diff --git a/block/file-posix.c b/block/file-posix.c
-index e2ea071315..4c3dc56c8e 100644
+index cfa0b832ba..d5c28cccc9 100644
 --- a/block/file-posix.c
 +++ b/block/file-posix.c
-@@ -2884,6 +2884,7 @@ 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 e2ea071315..4c3dc56c8e 100644
  
      /* Validate options and set default values */
      assert(options->driver == BLOCKDEV_DRIVER_FILE);
-@@ -2924,19 +2925,22 @@ raw_co_create(BlockdevCreateOptions *options, Error **errp)
+@@ -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 e2ea071315..4c3dc56c8e 100644
      }
  
      /* Clear the file by truncating it to 0 */
-@@ -2990,13 +2994,15 @@ 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 e2ea071315..4c3dc56c8e 100644
      }
  
  out_close:
-@@ -3020,6 +3026,7 @@ 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 e2ea071315..4c3dc56c8e 100644
  
      /* Skip file: protocol prefix */
      strstart(filename, "file:", &filename);
-@@ -3042,6 +3049,18 @@ 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 e2ea071315..4c3dc56c8e 100644
      options = (BlockdevCreateOptions) {
          .driver     = BLOCKDEV_DRIVER_FILE,
          .u.file     = {
-@@ -3053,6 +3072,8 @@ 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 +119,10 @@ index e2ea071315..4c3dc56c8e 100644
      };
      return raw_co_create(&options, errp);
 diff --git a/qapi/block-core.json b/qapi/block-core.json
-index 48ba32049f..321d1fd0e1 100644
+index 0e5f148d30..1c05413916 100644
 --- a/qapi/block-core.json
 +++ b/qapi/block-core.json
-@@ -4974,6 +4974,10 @@
+@@ -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)
  #
@@ -133,7 +133,7 @@ index 48ba32049f..321d1fd0e1 100644
  # Since: 2.12
  ##
  { 'struct': 'BlockdevCreateOptionsFile',
-@@ -4981,7 +4985,8 @@
+@@ -5023,7 +5027,8 @@
              'size':                 'size',
              '*preallocation':       'PreallocMode',
              '*nocow':               'bool',

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

@@ -18,7 +18,7 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  1 file changed, 1 insertion(+), 2 deletions(-)
 
 diff --git a/monitor/qmp.c b/monitor/qmp.c
-index eb181d5979..20fc0d20a6 100644
+index f093e256e9..78f1c8e3c8 100644
 --- a/monitor/qmp.c
 +++ b/monitor/qmp.c
 @@ -534,8 +534,7 @@ void monitor_init_qmp(Chardev *chr, bool pretty, Error **errp)

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

@@ -26,10 +26,10 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  1 file changed, 2 insertions(+), 1 deletion(-)
 
 diff --git a/hw/core/machine.c b/hw/core/machine.c
-index f29fe95964..2c327fc36a 100644
+index 63c6ef93d2..9a34017e5a 100644
 --- a/hw/core/machine.c
 +++ b/hw/core/machine.c
-@@ -180,7 +180,8 @@ 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

@@ -21,10 +21,10 @@ Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
  4 files changed, 34 insertions(+)
 
 diff --git a/hw/core/machine-qmp-cmds.c b/hw/core/machine-qmp-cmds.c
-index 52a6d74820..362128842d 100644
+index 93fb4bc24a..b9999423b4 100644
 --- a/hw/core/machine-qmp-cmds.c
 +++ b/hw/core/machine-qmp-cmds.c
-@@ -94,6 +94,11 @@ MachineInfoList *qmp_query_machines(bool has_compat_props, bool compat_props,
+@@ -95,6 +95,11 @@ MachineInfoList *qmp_query_machines(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;
@@ -35,12 +35,12 @@ index 52a6d74820..362128842d 100644
 +            }
          }
  
-         if (mc->default_cpu_type) {
+         if (default_cpu_type) {
 diff --git a/include/hw/boards.h b/include/hw/boards.h
-index 36fbb9b59d..d1741ea121 100644
+index f22b2e7fc7..8ada4d5832 100644
 --- a/include/hw/boards.h
 +++ b/include/hw/boards.h
-@@ -268,6 +268,8 @@ struct MachineClass {
+@@ -271,6 +271,8 @@ struct MachineClass {
      const char *desc;
      const char *deprecation_reason;
  
@@ -71,10 +71,10 @@ index 16366b774a..12cfd3f260 100644
                                 'features': ['unstable'] } } }
  
 diff --git a/system/vl.c b/system/vl.c
-index e7cae51f13..3f4916ac5a 100644
+index 452742ab58..c3707b2412 100644
 --- a/system/vl.c
 +++ b/system/vl.c
-@@ -1675,6 +1675,7 @@ static MachineClass *select_machine(QDict *qdict, Error **errp)
+@@ -1674,6 +1674,7 @@ static MachineClass *select_machine(QDict *qdict, Error **errp)
  {
      ERRP_GUARD();
      const char *machine_type = qdict_get_try_str(qdict, "type");
@@ -82,7 +82,7 @@ index e7cae51f13..3f4916ac5a 100644
      g_autoptr(GSList) machines = object_class_get_list(TYPE_MACHINE, false);
      MachineClass *machine_class = NULL;
  
-@@ -1694,7 +1695,11 @@ static MachineClass *select_machine(QDict *qdict, Error **errp)
+@@ -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");
@@ -94,7 +94,7 @@ index e7cae51f13..3f4916ac5a 100644
      return machine_class;
  }
  
-@@ -3316,12 +3321,31 @@ void qemu_init(int argc, char **argv)
+@@ -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 a1292c01ec..2e38b30d67 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)

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

@@ -16,18 +16,18 @@ Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
 ---
  block/meson.build |   2 +
  meson.build       |   5 +
- vma-reader.c      | 868 ++++++++++++++++++++++++++++++++++++++++++
- vma-writer.c      | 817 ++++++++++++++++++++++++++++++++++++++++
+ vma-reader.c      | 867 ++++++++++++++++++++++++++++++++++++++++++
+ vma-writer.c      | 816 ++++++++++++++++++++++++++++++++++++++++
  vma.c             | 941 ++++++++++++++++++++++++++++++++++++++++++++++
  vma.h             | 150 ++++++++
- 6 files changed, 2783 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 6a60b5d6b9..652c8cbdb7 100644
+index a21d9a5411..1373612c10 100644
 --- a/block/meson.build
 +++ b/block/meson.build
 @@ -42,6 +42,8 @@ block_ss.add(files(
@@ -40,10 +40,10 @@ index 6a60b5d6b9..652c8cbdb7 100644
  system_ss.add(files('block-ram-registrar.c'))
  
 diff --git a/meson.build b/meson.build
-index 147097c652..b9b673c271 100644
+index 8ec796d835..680ab48b9b 100644
 --- a/meson.build
 +++ b/meson.build
-@@ -2129,6 +2129,8 @@ endif
+@@ -2161,6 +2161,8 @@ endif
  
  has_gettid = cc.has_function('gettid')
  
@@ -52,7 +52,7 @@ index 147097c652..b9b673c271 100644
  # libselinux
  selinux = dependency('libselinux',
                       required: get_option('selinux'),
-@@ -4344,6 +4346,9 @@ if have_tools
+@@ -4367,6 +4369,9 @@ if have_tools
                 dependencies: [blockdev, qemuutil, selinux],
                 install: true)
  
@@ -64,10 +64,10 @@ index 147097c652..b9b673c271 100644
    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..65015d2e1e
+index 0000000000..1888b21851
 --- /dev/null
 +++ b/vma-reader.c
-@@ -0,0 +1,868 @@
+@@ -0,0 +1,867 @@
 +/*
 + * VMA: Virtual Machine Archive
 + *
@@ -90,7 +90,7 @@ index 0000000000..65015d2e1e
 +#include "vma.h"
 +#include "block/block.h"
 +#include "block/graph-lock.h"
-+#include "sysemu/block-backend.h"
++#include "system/block-backend.h"
 +
 +static unsigned char zero_vma_block[VMA_BLOCK_SIZE];
 +
@@ -883,8 +883,7 @@ index 0000000000..65015d2e1e
 +
 +            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)) {
@@ -938,10 +937,10 @@ index 0000000000..65015d2e1e
 +
 diff --git a/vma-writer.c b/vma-writer.c
 new file mode 100644
-index 0000000000..a466652a5d
+index 0000000000..3f489092df
 --- /dev/null
 +++ b/vma-writer.c
-@@ -0,0 +1,817 @@
+@@ -0,0 +1,816 @@
 +/*
 + * VMA: Virtual Machine Archive
 + *
@@ -1135,8 +1134,7 @@ index 0000000000..a466652a5d
 +    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;
 +
@@ -1761,7 +1759,7 @@ index 0000000000..a466652a5d
 +}
 diff --git a/vma.c b/vma.c
 new file mode 100644
-index 0000000000..8d4b4be414
+index 0000000000..0e990b5e30
 --- /dev/null
 +++ b/vma.c
 @@ -0,0 +1,941 @@
@@ -1787,8 +1785,8 @@ index 0000000000..8d4b4be414
 +#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)
 +{

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 2e38b30d67..fe69723ada 100644
+index cc747e9163..6f7c45f922 100644
 --- a/block/backup.c
 +++ b/block/backup.c
 @@ -29,28 +29,6 @@
@@ -247,7 +247,7 @@ index 2e38b30d67..fe69723ada 100644
      if (perf->max_chunk && perf->max_chunk < cluster_size) {
          error_setg(errp, "Required max-chunk (%" PRIi64 ") is less than backup "
 diff --git a/block/meson.build b/block/meson.build
-index 652c8cbdb7..e1cf5a2e65 100644
+index 1373612c10..6278c4af0f 100644
 --- a/block/meson.build
 +++ b/block/meson.build
 @@ -4,6 +4,7 @@ block_ss.add(files(
@@ -312,10 +312,10 @@ index ebb4e56a50..e717a74e5f 100644
      BDRV_TRACKED_READ,
      BDRV_TRACKED_WRITE,
 diff --git a/job.c b/job.c
-index 660ce22c56..baf54c8d60 100644
+index 0653bc2ba6..b981070ee8 100644
 --- a/job.c
 +++ b/job.c
-@@ -331,7 +331,8 @@ 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

@@ -11,7 +11,7 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  2 files changed, 46 insertions(+)
 
 diff --git a/include/qemu/job.h b/include/qemu/job.h
-index 2b873f2576..528cd6acb9 100644
+index a5a04155ea..562cc7eaec 100644
 --- a/include/qemu/job.h
 +++ b/include/qemu/job.h
 @@ -362,6 +362,18 @@ void job_unlock(void);
@@ -34,7 +34,7 @@ index 2b873f2576..528cd6acb9 100644
   * Release a reference that was previously acquired with job_txn_add_job or
   * job_txn_new. If it's the last reference to the object, it will be freed.
 diff --git a/job.c b/job.c
-index baf54c8d60..3ac5e5cde2 100644
+index b981070ee8..f4646866ec 100644
 --- a/job.c
 +++ b/job.c
 @@ -94,6 +94,8 @@ struct JobTxn {
@@ -72,7 +72,7 @@ index baf54c8d60..3ac5e5cde2 100644
  /* Called with job_mutex held. */
  static void job_txn_ref_locked(JobTxn *txn)
  {
-@@ -1042,6 +1063,12 @@ static void job_completed_txn_success_locked(Job *job)
+@@ -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 baf54c8d60..3ac5e5cde2 100644
              return;
          }
          assert(other_job->ret == 0);
-@@ -1253,6 +1280,13 @@ 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

@@ -94,17 +94,17 @@ 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                   | 1092 ++++++++++++++++++++++++++++++++
+ pve-backup.c                   | 1096 ++++++++++++++++++++++++++++++++
  qapi/block-core.json           |  233 +++++++
  qapi/common.json               |   14 +
  qapi/machine.json              |   16 +-
- 14 files changed, 1711 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 e1cf5a2e65..2367e1ac1b 100644
+index 6278c4af0f..d1b16e40e9 100644
 --- a/block/meson.build
 +++ b/block/meson.build
 @@ -44,6 +44,11 @@ block_ss.add(files(
@@ -120,10 +120,10 @@ index e1cf5a2e65..2367e1ac1b 100644
  system_ss.add(when: 'CONFIG_TCG', if_true: files('blkreplay.c'))
  system_ss.add(files('block-ram-registrar.c'))
 diff --git a/block/monitor/block-hmp-cmds.c b/block/monitor/block-hmp-cmds.c
-index bdf2eb50b6..439a7a14c8 100644
+index 6919a49bf5..4f30f99644 100644
 --- a/block/monitor/block-hmp-cmds.c
 +++ b/block/monitor/block-hmp-cmds.c
-@@ -1009,3 +1009,42 @@ void hmp_change_medium(Monitor *mon, const char *device, const char *target,
+@@ -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 bdf2eb50b6..439a7a14c8 100644
 +    hmp_handle_error(mon, error);
 +}
 diff --git a/blockdev.c b/blockdev.c
-index 79d47b1920..3f67eb413d 100644
+index 158ac9314b..17de5d2ae4 100644
 --- a/blockdev.c
 +++ b/blockdev.c
 @@ -37,6 +37,7 @@
@@ -265,10 +265,10 @@ index 2596cc2426..9dda91d65a 100644
  void hmp_device_add(Monitor *mon, const QDict *qdict);
  void hmp_device_del(Monitor *mon, const QDict *qdict);
 diff --git a/meson.build b/meson.build
-index b9b673c271..f6fb9b4fd8 100644
+index 680ab48b9b..1f74de1d93 100644
 --- a/meson.build
 +++ b/meson.build
-@@ -2130,6 +2130,7 @@ endif
+@@ -2162,6 +2162,7 @@ endif
  has_gettid = cc.has_function('gettid')
  
  libuuid = cc.find_library('uuid', required: true)
@@ -277,7 +277,7 @@ index b9b673c271..f6fb9b4fd8 100644
  # libselinux
  selinux = dependency('libselinux',
 diff --git a/monitor/hmp-cmds.c b/monitor/hmp-cmds.c
-index 874084565f..bedeb81f8c 100644
+index bade2a4b92..77a2fa409a 100644
 --- a/monitor/hmp-cmds.c
 +++ b/monitor/hmp-cmds.c
 @@ -22,6 +22,7 @@
@@ -586,22 +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..9f83ecb310
+index 0000000000..e931cb9203
 --- /dev/null
 +++ b/pve-backup.c
-@@ -0,0 +1,1092 @@
+@@ -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"
 +
@@ -677,6 +678,7 @@ index 0000000000..9f83ecb310
 +    size_t size;
 +    uint64_t block_size;
 +    uint8_t dev_id;
++    char* device_name;
 +    int completed_ret; // INT_MAX if not completed
 +    BdrvDirtyBitmap *bitmap;
 +    BlockDriverState *target;
@@ -910,6 +912,8 @@ index 0000000000..9f83ecb310
 +    }
 +
 +    di->bs = NULL;
++    g_free(di->device_name);
++    di->device_name = NULL;
 +
 +    assert(di->target == NULL);
 +
@@ -1199,6 +1203,8 @@ index 0000000000..9f83ecb310
 +            }
 +            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++;
 +        }
@@ -1212,6 +1218,7 @@ index 0000000000..9f83ecb310
 +
 +            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);
 +        }
 +    }
@@ -1378,9 +1385,6 @@ index 0000000000..9f83ecb310
 +
 +            di->block_size = dump_cb_block_size;
 +
-+            bdrv_graph_co_rdlock();
-+            const char *devname = bdrv_get_device_name(di->bs);
-+            bdrv_graph_co_rdunlock();
 +            PBSBitmapAction action = PBS_BITMAP_ACTION_NOT_USED;
 +            size_t dirty = di->size;
 +
@@ -1395,7 +1399,8 @@ index 0000000000..9f83ecb310
 +                    }
 +                    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) {
@@ -1419,7 +1424,8 @@ index 0000000000..9f83ecb310
 +                }
 +            }
 +
-+            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;
 +            }
@@ -1431,7 +1437,7 @@ index 0000000000..9f83ecb310
 +            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;
@@ -1440,9 +1446,7 @@ index 0000000000..9f83ecb310
 +    } 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;
 +        }
 +
@@ -1456,10 +1460,7 @@ index 0000000000..9f83ecb310
 +                goto err_mutex;
 +            }
 +
-+            bdrv_graph_co_rdlock();
-+            const char *devname = bdrv_get_device_name(di->bs);
-+            bdrv_graph_co_rdunlock();
-+            di->dev_id = vma_writer_register_stream(vmaw, devname, di->size);
++            di->dev_id = vma_writer_register_stream(vmaw, di->device_name, di->size);
 +            if (di->dev_id <= 0) {
 +                error_set(errp, ERROR_CLASS_GENERIC_ERROR,
 +                          "register_stream failed");
@@ -1570,6 +1571,9 @@ index 0000000000..9f83ecb310
 +            bdrv_co_unref(di->target);
 +        }
 +
++        g_free(di->device_name);
++        di->device_name = NULL;
++
 +        g_free(di);
 +    }
 +    g_list_free(di_list);
@@ -1683,10 +1687,10 @@ index 0000000000..9f83ecb310
 +    return ret;
 +}
 diff --git a/qapi/block-core.json b/qapi/block-core.json
-index 321d1fd0e1..68caf30084 100644
+index 1c05413916..dd98e03bf1 100644
 --- a/qapi/block-core.json
 +++ b/qapi/block-core.json
-@@ -851,6 +851,239 @@
+@@ -855,6 +855,239 @@
  { 'command': 'query-block', 'returns': ['BlockInfo'],
    'allow-preconfig': true }
  
@@ -1927,13 +1931,13 @@ index 321d1fd0e1..68caf30084 100644
  # @BlockDeviceTimedStats:
  #
 diff --git a/qapi/common.json b/qapi/common.json
-index 6ffc7a3789..9c6c671ece 100644
+index 0e3a0bbbfb..554680c716 100644
 --- a/qapi/common.json
 +++ b/qapi/common.json
-@@ -212,3 +212,17 @@
+@@ -226,3 +226,17 @@
  ##
- { 'struct': 'HumanReadableText',
-   'data': { 'human-readable-text': 'str' } }
+ { 'enum': 'EndianMode',
+   'data': [ 'unspecified', 'little', 'big' ] }
 +
 +##
 +# @UuidInfo:

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

@@ -14,10 +14,10 @@ Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
  create mode 100644 pbs-restore.c
 
 diff --git a/meson.build b/meson.build
-index f6fb9b4fd8..f666d0f028 100644
+index 1f74de1d93..8508aab9c9 100644
 --- a/meson.build
 +++ b/meson.build
-@@ -4350,6 +4350,10 @@ 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, qemuutil, qom], install: true)
  
@@ -30,7 +30,7 @@ index f6fb9b4fd8..f666d0f028 100644
    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

@@ -23,7 +23,7 @@ Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
  create mode 100644 block/pbs.c
 
 diff --git a/block/meson.build b/block/meson.build
-index 2367e1ac1b..e178047ec9 100644
+index d1b16e40e9..d243372c41 100644
 --- a/block/meson.build
 +++ b/block/meson.build
 @@ -49,6 +49,8 @@ block_ss.add(files(
@@ -37,7 +37,7 @@ index 2367e1ac1b..e178047ec9 100644
  system_ss.add(files('block-ram-registrar.c'))
 diff --git a/block/pbs.c b/block/pbs.c
 new file mode 100644
-index 0000000000..2d5e28ce8f
+index 0000000000..3e41421716
 --- /dev/null
 +++ b/block/pbs.c
 @@ -0,0 +1,306 @@
@@ -47,8 +47,8 @@ index 0000000000..2d5e28ce8f
 +
 +#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"
@@ -348,10 +348,10 @@ index 0000000000..2d5e28ce8f
 +
 +block_init(bdrv_pbs_init);
 diff --git a/meson.build b/meson.build
-index f666d0f028..4c85736ec3 100644
+index 8508aab9c9..9c39f54f86 100644
 --- a/meson.build
 +++ b/meson.build
-@@ -4815,7 +4815,7 @@ summary_info += {'Query Processing Library support': qpl}
+@@ -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}
@@ -361,10 +361,10 @@ index f666d0f028..4c85736ec3 100644
  summary_info += {'libdaxctl support': libdaxctl}
  summary_info += {'libcbor support':   libcbor}
 diff --git a/qapi/block-core.json b/qapi/block-core.json
-index 68caf30084..d45e8975a7 100644
+index dd98e03bf1..0c3ebfa74e 100644
 --- a/qapi/block-core.json
 +++ b/qapi/block-core.json
-@@ -3466,6 +3466,7 @@
+@@ -3470,6 +3470,7 @@
              'parallels', 'preallocate', 'qcow', 'qcow2', 'qed', 'quorum',
              'raw', 'rbd',
              { 'name': 'replication', 'if': 'CONFIG_REPLICATION' },
@@ -372,7 +372,7 @@ index 68caf30084..d45e8975a7 100644
              'ssh', 'throttle', 'vdi', 'vhdx',
              { 'name': 'virtio-blk-vfio-pci', 'if': 'CONFIG_BLKIO' },
              { 'name': 'virtio-blk-vhost-user', 'if': 'CONFIG_BLKIO' },
-@@ -3552,6 +3553,33 @@
+@@ -3556,6 +3557,33 @@
  { 'struct': 'BlockdevOptionsNull',
    'data': { '*size': 'int', '*latency-ns': 'uint64', '*read-zeroes': 'bool' } }
  
@@ -406,7 +406,7 @@ index 68caf30084..d45e8975a7 100644
  ##
  # @BlockdevOptionsNVMe:
  #
-@@ -4993,6 +5021,7 @@
+@@ -5003,6 +5031,7 @@
        'nfs':        'BlockdevOptionsNfs',
        'null-aio':   'BlockdevOptionsNull',
        'null-co':    'BlockdevOptionsNull',

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

@@ -14,10 +14,10 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  2 files changed, 7 insertions(+), 3 deletions(-)
 
 diff --git a/meson.build b/meson.build
-index 4c85736ec3..57f666d722 100644
+index 9c39f54f86..60af7fa723 100644
 --- a/meson.build
 +++ b/meson.build
-@@ -2130,6 +2130,7 @@ endif
+@@ -2162,6 +2162,7 @@ endif
  has_gettid = cc.has_function('gettid')
  
  libuuid = cc.find_library('uuid', required: true)
@@ -25,7 +25,7 @@ index 4c85736ec3..57f666d722 100644
  libproxmox_backup_qemu = cc.find_library('proxmox_backup_qemu', required: true)
  
  # libselinux
-@@ -3744,7 +3745,7 @@ if have_block
+@@ -3766,7 +3767,7 @@ if have_block
    if host_os == 'windows'
      system_ss.add(files('os-win32.c'))
    else
@@ -35,7 +35,7 @@ index 4c85736ec3..57f666d722 100644
  endif
  
 diff --git a/os-posix.c b/os-posix.c
-index 43f9a43f3f..a47e46d1c2 100644
+index 52925c23d3..84b96d3da9 100644
 --- a/os-posix.c
 +++ b/os-posix.c
 @@ -29,6 +29,8 @@

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

@@ -26,19 +26,19 @@ Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
  create mode 100644 migration/pbs-state.c
 
 diff --git a/include/migration/misc.h b/include/migration/misc.h
-index 804eb23c06..c75b146ae6 100644
+index 8fd36eba1d..e963e93e71 100644
 --- a/include/migration/misc.h
 +++ b/include/migration/misc.h
-@@ -106,4 +106,7 @@ bool migration_incoming_postcopy_advised(void);
- /* True if background snapshot is active */
- bool migration_in_bg_snapshot(void);
+@@ -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 075b013971..eca57cb2a3 100644
+index 46e92249a1..fb3fd7d7d0 100644
 --- a/migration/meson.build
 +++ b/migration/meson.build
 @@ -8,6 +8,7 @@ migration_files = files(
@@ -49,7 +49,7 @@ index 075b013971..eca57cb2a3 100644
  
  system_ss.add(files(
    'block-dirty-bitmap.c',
-@@ -27,6 +28,7 @@ system_ss.add(files(
+@@ -31,6 +32,7 @@ system_ss.add(files(
    'multifd-zlib.c',
    'multifd-zero-page.c',
    'options.c',
@@ -58,10 +58,10 @@ index 075b013971..eca57cb2a3 100644
    'savevm.c',
    'savevm-async.c',
 diff --git a/migration/migration.c b/migration/migration.c
-index 8c5bd0a75c..491d9aa017 100644
+index d46e776e24..2f3430f440 100644
 --- a/migration/migration.c
 +++ b/migration/migration.c
-@@ -266,6 +266,7 @@ void migration_object_init(void)
+@@ -319,6 +319,7 @@ void migration_object_init(void)
  
      /* Initialize cpu throttle timers */
      cpu_throttle_init();
@@ -180,10 +180,10 @@ index 0000000000..a97187e4d7
 +                         NULL);
 +}
 diff --git a/pve-backup.c b/pve-backup.c
-index 9f83ecb310..57477f7f2a 100644
+index e931cb9203..366b015589 100644
 --- a/pve-backup.c
 +++ b/pve-backup.c
-@@ -1085,6 +1085,7 @@ 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;
@@ -192,10 +192,10 @@ index 9f83ecb310..57477f7f2a 100644
      ret->pbs_masterkey = true;
      ret->backup_max_workers = true;
 diff --git a/qapi/block-core.json b/qapi/block-core.json
-index d45e8975a7..9795247c1f 100644
+index 0c3ebfa74e..6838187607 100644
 --- a/qapi/block-core.json
 +++ b/qapi/block-core.json
-@@ -1004,6 +1004,11 @@
+@@ -1008,6 +1008,11 @@
  # @pbs-dirty-bitmap-savevm: True if 'dirty-bitmaps' migration capability can
  #                           safely be set for savevm-async.
  #
@@ -207,7 +207,7 @@ index d45e8975a7..9795247c1f 100644
  # @pbs-masterkey: True if the QMP backup call supports the 'master_keyfile'
  #                 parameter.
  #
-@@ -1017,6 +1022,7 @@
+@@ -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

@@ -15,20 +15,21 @@ transferred.
 Signed-off-by: Stefan Reiter <s.reiter@proxmox.com>
 Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
 ---
- migration/block-dirty-bitmap.c | 5 ++++-
- 1 file changed, 4 insertions(+), 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 a7d55048c2..77346a5fa2 100644
+index f2c352d4a7..931a8481e9 100644
 --- a/migration/block-dirty-bitmap.c
 +++ b/migration/block-dirty-bitmap.c
-@@ -539,7 +539,10 @@ 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, 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

@@ -21,7 +21,7 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  1 file changed, 30 insertions(+)
 
 diff --git a/block/iscsi.c b/block/iscsi.c
-index 979bf90cb7..961714a4be 100644
+index 2f0f4dac09..b523137cff 100644
 --- a/block/iscsi.c
 +++ b/block/iscsi.c
 @@ -1392,12 +1392,42 @@ static char *get_initiator_name(QemuOpts *opts)

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

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

@@ -31,21 +31,22 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
 [FE: adapt to changed function signatures
      make error return value consistent with QEMU
      avoid premature break during read
-     adhere to block graph lock requirements]
+     adhere to block graph lock requirements
+     avoid superfluous child permission update]
 Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
 ---
- block/alloc-track.c | 366 ++++++++++++++++++++++++++++++++++++++++++++
+ block/alloc-track.c | 343 ++++++++++++++++++++++++++++++++++++++++++++
  block/meson.build   |   1 +
- block/stream.c      |  34 ++++
- 3 files changed, 401 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..b4a9851144
+index 0000000000..718aaabf2a
 --- /dev/null
 +++ b/block/alloc-track.c
-@@ -0,0 +1,366 @@
+@@ -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
@@ -63,26 +64,19 @@ index 0000000000..b4a9851144
 +#include "block/block_int.h"
 +#include "block/dirty-bitmap.h"
 +#include "block/graph-lock.h"
-+#include "qapi/qmp/qdict.h"
-+#include "qapi/qmp/qstring.h"
++#include "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,
-+    DropInProgress,
-+} DropState;
-+
 +typedef struct {
 +    BdrvDirtyBitmap *bitmap;
 +    uint64_t granularity;
-+    DropState drop_state;
-+    bool auto_remove;
 +} BDRVAllocTrackState;
 +
 +static QemuOptsList runtime_opts = {
@@ -134,7 +128,11 @@ index 0000000000..b4a9851144
 +        goto fail;
 +    }
 +
-+    s->auto_remove = qemu_opt_get_bool(opts, TRACK_OPT_AUTO_REMOVE, false);
++    if (!qemu_opt_get_bool(opts, TRACK_OPT_AUTO_REMOVE, false)) {
++        error_setg(errp, "alloc-track: requires auto-remove option to be set to on");
++        ret = -EINVAL;
++        goto fail;
++    }
 +
 +    /* open the target (write) node, backing will be attached by block layer */
 +    file = bdrv_open_child(NULL, options, "file", bs, &child_of_bds,
@@ -182,8 +180,6 @@ index 0000000000..b4a9851144
 +        goto fail;
 +    }
 +
-+    s->drop_state = DropNone;
-+
 +fail:
 +    if (ret < 0) {
 +        bdrv_graph_wrlock();
@@ -334,18 +330,8 @@ index 0000000000..b4a9851144
 +                 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 {
@@ -371,14 +357,6 @@ index 0000000000..b4a9851144
 +     * kinda fits better, but in the long-term, a special parameter would be
 +     * nice (or done via qemu-server via upcoming blockdev-replace QMP command).
 +     */
-+    if (backing_file == NULL) {
-+        BDRVAllocTrackState *s = bs->opaque;
-+        bdrv_drained_begin(bs);
-+        s->drop_state = DropInProgress;
-+        bdrv_child_refresh_perms(bs, bs->file, &error_abort);
-+        bdrv_drained_end(bs);
-+    }
-+
 +    return 0;
 +}
 +
@@ -413,7 +391,7 @@ index 0000000000..b4a9851144
 +
 +block_init(bdrv_alloc_track_init);
 diff --git a/block/meson.build b/block/meson.build
-index e178047ec9..7ef7250d31 100644
+index d243372c41..9b45b5256d 100644
 --- a/block/meson.build
 +++ b/block/meson.build
 @@ -2,6 +2,7 @@ block_ss.add(genh)
@@ -425,7 +403,7 @@ index e178047ec9..7ef7250d31 100644
    'backup.c',
    'backup-dump.c',
 diff --git a/block/stream.c b/block/stream.c
-index 1d1c65f061..d499c8883f 100644
+index e187cd1262..0b61029399 100644
 --- a/block/stream.c
 +++ b/block/stream.c
 @@ -120,6 +120,40 @@ static int stream_prepare(Job *job)

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

@@ -13,7 +13,7 @@ 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 728bce3b1e..6c9a8e0add 100644
+index bf143fac00..70d92966f7 100644
 --- a/block/rbd.c
 +++ b/block/rbd.c
 @@ -1515,7 +1515,6 @@ static int coroutine_fn qemu_rbd_co_block_status(BlockDriverState *bs,

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

@@ -14,7 +14,7 @@ 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 6c9a8e0add..6f5fe90f3a 100644
+index 70d92966f7..931b513828 100644
 --- a/block/rbd.c
 +++ b/block/rbd.c
 @@ -1474,11 +1474,11 @@ static int qemu_rbd_diff_iterate_cb(uint64_t offs, size_t len,

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 6f5fe90f3a..24e820d056 100644
+index 931b513828..4ab9bb5e02 100644
 --- a/block/rbd.c
 +++ b/block/rbd.c
 @@ -108,12 +108,6 @@ typedef struct RBDTask {

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

@@ -61,17 +61,96 @@ it has no parent, so just pass the one from the original bs.
 
 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                   | 135 ++++++++++++++++++++++++++++++++-
- qapi/block-core.json           |  10 ++-
- 3 files changed, 142 insertions(+), 4 deletions(-)
+ 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 439a7a14c8..d0e7771dcc 100644
+index 4f30f99644..66d16d342f 100644
 --- a/block/monitor/block-hmp-cmds.c
 +++ b/block/monitor/block-hmp-cmds.c
-@@ -1044,6 +1044,7 @@ void coroutine_fn hmp_backup(Monitor *mon, const QDict *qdict)
+@@ -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
@@ -80,22 +159,18 @@ index 439a7a14c8..d0e7771dcc 100644
  
      hmp_handle_error(mon, error);
 diff --git a/pve-backup.c b/pve-backup.c
-index 57477f7f2a..0f098000dd 100644
+index 366b015589..9b66788ab5 100644
 --- a/pve-backup.c
 +++ b/pve-backup.c
-@@ -7,9 +7,11 @@
- #include "sysemu/blockdev.h"
+@@ -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"
-+#include "qapi/qmp/qdict.h"
- #include "qapi/qmp/qerror.h"
- #include "qemu/cutils.h"
- 
-@@ -80,8 +82,15 @@ static void pvebackup_init(void)
+@@ -81,8 +82,15 @@ static void pvebackup_init(void)
  // initialize PVEBackupState at startup
  opts_init(pvebackup_init);
  
@@ -111,17 +186,12 @@ index 57477f7f2a..0f098000dd 100644
      size_t size;
      uint64_t block_size;
      uint8_t dev_id;
-@@ -353,6 +362,22 @@ static void pvebackup_complete_cb(void *opaque, int ret)
-     PVEBackupDevInfo *di = opaque;
-     di->completed_ret = ret;
+@@ -352,11 +360,44 @@ static void coroutine_fn pvebackup_co_complete_stream(void *opaque)
+     qemu_co_mutex_unlock(&backup_state.backup_mutex);
+ }
  
-+    /*
-+     * 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.
-+     */
++static void cleanup_snapshot_access(PVEBackupDevInfo *di)
++{
 +    if (di->fleecing.snapshot_access) {
 +        bdrv_unref(di->fleecing.snapshot_access);
 +        di->fleecing.snapshot_access = NULL;
@@ -130,11 +200,104 @@ index 57477f7f2a..0f098000dd 100644
 +        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.
       */
-@@ -519,9 +544,77 @@ static void create_backup_jobs_bh(void *opaque) {
+@@ -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);
  
@@ -142,50 +305,15 @@ index 57477f7f2a..0f098000dd 100644
 +
 +        BlockDriverState *source_bs = di->bs;
 +        bool discard_source = false;
-+        bdrv_graph_co_rdlock();
-+        const char *job_id = bdrv_get_device_name(di->bs);
-+        bdrv_graph_co_rdunlock();
 +        if (di->fleecing.bs) {
-+            QDict *cbw_opts = qdict_new();
-+            qdict_put_str(cbw_opts, "driver", "copy-before-write");
-+            qdict_put_str(cbw_opts, "file", bdrv_get_node_name(di->bs));
-+            qdict_put_str(cbw_opts, "target", bdrv_get_node_name(di->fleecing.bs));
-+
-+            if (di->bitmap) {
-+                /*
-+                 * Only guest writes to parts relevant for the backup need to be intercepted with
-+                 * old data being copied to the fleecing image.
-+                 */
-+                qdict_put_str(cbw_opts, "bitmap.node", bdrv_get_node_name(di->bs));
-+                qdict_put_str(cbw_opts, "bitmap.name", bdrv_dirty_bitmap_name(di->bitmap));
-+            }
-+            /*
-+             * Fleecing storage is supposed to be fast and it's better to break backup than guest
-+             * writes. Certain guest drivers like VirtIO-win have 60 seconds timeout by default, so
-+             * abort a bit before that.
-+             */
-+            qdict_put_str(cbw_opts, "on-cbw-error", "break-snapshot");
-+            qdict_put_int(cbw_opts, "cbw-timeout", 45);
-+
-+            di->fleecing.cbw = bdrv_insert_node(di->bs, cbw_opts, BDRV_O_RDWR, &local_err);
-+
-+            if (!di->fleecing.cbw) {
-+                error_setg(errp, "appending cbw node for fleecing failed: %s",
++            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;
 +            }
 +
-+            QDict *snapshot_access_opts = qdict_new();
-+            qdict_put_str(snapshot_access_opts, "driver", "snapshot-access");
-+            qdict_put_str(snapshot_access_opts, "file", bdrv_get_node_name(di->fleecing.cbw));
-+
-+            di->fleecing.snapshot_access =
-+                bdrv_open(NULL, NULL, snapshot_access_opts, BDRV_O_RDWR | BDRV_O_UNMAP, &local_err);
-+            if (!di->fleecing.snapshot_access) {
-+                error_setg(errp, "setting up snapshot access for fleecing failed: %s",
-+                           local_err ? error_get_pretty(local_err) : "unknown error");
-+                break;
-+            }
 +            source_bs = di->fleecing.snapshot_access;
 +            discard_source = true;
 +
@@ -209,12 +337,20 @@ index 57477f7f2a..0f098000dd 100644
          BlockJob *job = backup_job_create(
 -            NULL, di->bs, di->target, backup_state.speed, sync_mode, di->bitmap,
 -            bitmap_mode, false, NULL, &backup_state.perf, BLOCKDEV_ON_ERROR_REPORT,
-+            job_id, source_bs, di->target, backup_state.speed, sync_mode, di->bitmap,
++            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);
  
-@@ -577,6 +670,14 @@ static void create_backup_jobs_bh(void *opaque) {
+@@ -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);
  }
  
@@ -229,7 +365,7 @@ index 57477f7f2a..0f098000dd 100644
  /*
   * Returns a list of device infos, which needs to be freed by the caller. In
   * case of an error, errp will be set, but the returned value might still be a
-@@ -584,6 +685,7 @@ static void create_backup_jobs_bh(void *opaque) {
+@@ -588,6 +730,7 @@ static void create_backup_jobs_bh(void *opaque) {
   */
  static GList coroutine_fn GRAPH_RDLOCK *get_device_info(
      const char *devlist,
@@ -237,11 +373,10 @@ index 57477f7f2a..0f098000dd 100644
      Error **errp)
  {
      gchar **devs = NULL;
-@@ -607,6 +709,31 @@ static GList coroutine_fn GRAPH_RDLOCK *get_device_info(
-             }
-             PVEBackupDevInfo *di = g_new0(PVEBackupDevInfo, 1);
+@@ -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);
@@ -269,7 +404,7 @@ index 57477f7f2a..0f098000dd 100644
              di_list = g_list_append(di_list, di);
              d++;
          }
-@@ -656,6 +783,7 @@ UuidInfo coroutine_fn *qmp_backup(
+@@ -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,
@@ -277,7 +412,7 @@ index 57477f7f2a..0f098000dd 100644
      Error **errp)
  {
      assert(qemu_in_coroutine());
-@@ -684,7 +812,7 @@ UuidInfo coroutine_fn *qmp_backup(
+@@ -691,7 +859,7 @@ UuidInfo coroutine_fn *qmp_backup(
      format = has_format ? format : BACKUP_FORMAT_VMA;
  
      bdrv_graph_co_rdlock();
@@ -286,7 +421,7 @@ index 57477f7f2a..0f098000dd 100644
      bdrv_graph_co_rdunlock();
      if (local_err) {
          error_propagate(errp, local_err);
-@@ -1089,5 +1217,6 @@ ProxmoxSupportStatus *qmp_query_proxmox_support(Error **errp)
+@@ -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;
@@ -294,10 +429,10 @@ index 57477f7f2a..0f098000dd 100644
      return ret;
  }
 diff --git a/qapi/block-core.json b/qapi/block-core.json
-index 9795247c1f..c581f1f238 100644
+index 6838187607..9bdcfa31ea 100644
 --- a/qapi/block-core.json
 +++ b/qapi/block-core.json
-@@ -948,6 +948,10 @@
+@@ -952,6 +952,10 @@
  #
  # @max-workers: see @BackupPerf for details. Default 16.
  #
@@ -308,7 +443,7 @@ index 9795247c1f..c581f1f238 100644
  # Returns: the uuid of the backup job
  #
  ##
-@@ -968,7 +972,8 @@
+@@ -972,7 +976,8 @@
                                      '*firewall-file': 'str',
                                      '*devlist': 'str',
                                      '*speed': 'int',
@@ -318,7 +453,7 @@ index 9795247c1f..c581f1f238 100644
    'returns': 'UuidInfo', 'coroutine': true }
  
  ##
-@@ -1014,6 +1019,8 @@
+@@ -1018,6 +1023,8 @@
  #
  # @pbs-library-version: Running version of libproxmox-backup-qemu0 library.
  #
@@ -327,7 +462,7 @@ index 9795247c1f..c581f1f238 100644
  # @backup-max-workers: Whether the 'max-workers' @BackupPerf setting is
  #     supported or not.
  #
-@@ -1025,6 +1032,7 @@
+@@ -1029,6 +1036,7 @@
              'pbs-dirty-bitmap-migration': 'bool',
              'pbs-masterkey': 'bool',
              'pbs-library-version': 'str',

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

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

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

@@ -22,10 +22,10 @@ Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
  1 file changed, 51 insertions(+), 27 deletions(-)
 
 diff --git a/include/hw/boards.h b/include/hw/boards.h
-index d1741ea121..3f9befda14 100644
+index 8ada4d5832..f9f3b75284 100644
 --- a/include/hw/boards.h
 +++ b/include/hw/boards.h
-@@ -631,42 +631,66 @@ struct MachineState {
+@@ -636,42 +636,66 @@ struct MachineState {
  
  
  /*
@@ -117,7 +117,7 @@ index d1741ea121..3f9befda14 100644
  
  /*
   * Evaluates true when a machine type with (major, minor)
-@@ -675,7 +699,7 @@ struct MachineState {
+@@ -680,7 +704,7 @@ struct MachineState {
   * lifecycle rules
   */
  #define MACHINE_VER_IS_DEPRECATED(...) \
@@ -126,7 +126,7 @@ index d1741ea121..3f9befda14 100644
  
  /*
   * Evaluates true when a machine type with (major, minor)
-@@ -684,7 +708,7 @@ struct MachineState {
+@@ -689,7 +713,7 @@ struct MachineState {
   * lifecycle rules
   */
  #define MACHINE_VER_SHOULD_DELETE(...) \

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

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

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

@@ -1,117 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Fiona Ebner <f.ebner@proxmox.com>
-Date: Mon, 29 Apr 2024 14:43:58 +0200
-Subject: [PATCH] PVE backup: improve error when copy-before-write fails for
- fleecing
-
-With fleecing, failure for copy-before-write does not fail the guest
-write, but only sets the snapshot error that is associated to the
-copy-before-write filter, making further requests to the snapshot
-access fail with EACCES, which then also fails the job. But that error
-code is not the root cause of why the backup failed, so bubble up the
-original snapshot error instead.
-
-Reported-by: Friedrich Weber <f.weber@proxmox.com>
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
-Tested-by: Friedrich Weber <f.weber@proxmox.com>
----
- block/copy-before-write.c | 18 ++++++++++++------
- block/copy-before-write.h |  1 +
- pve-backup.c              |  9 +++++++++
- 3 files changed, 22 insertions(+), 6 deletions(-)
-
-diff --git a/block/copy-before-write.c b/block/copy-before-write.c
-index 81afeff1c7..fdf9cdc0cd 100644
---- a/block/copy-before-write.c
-+++ b/block/copy-before-write.c
-@@ -27,6 +27,7 @@
- #include "qapi/qmp/qjson.h"
- 
- #include "sysemu/block-backend.h"
-+#include "qemu/atomic.h"
- #include "qemu/cutils.h"
- #include "qapi/error.h"
- #include "block/block_int.h"
-@@ -75,7 +76,8 @@ typedef struct BDRVCopyBeforeWriteState {
-      * @snapshot_error is normally zero. But on first copy-before-write failure
-      * when @on_cbw_error == ON_CBW_ERROR_BREAK_SNAPSHOT, @snapshot_error takes
-      * value of this error (<0). After that all in-flight and further
--     * snapshot-API requests will fail with that error.
-+     * snapshot-API requests will fail with that error. To be accessed with
-+     * atomics.
-      */
-     int snapshot_error;
- } BDRVCopyBeforeWriteState;
-@@ -115,7 +117,7 @@ static coroutine_fn int cbw_do_copy_before_write(BlockDriverState *bs,
-         return 0;
-     }
- 
--    if (s->snapshot_error) {
-+    if (qatomic_read(&s->snapshot_error)) {
-         return 0;
-     }
- 
-@@ -139,9 +141,7 @@ static coroutine_fn int cbw_do_copy_before_write(BlockDriverState *bs,
-     WITH_QEMU_LOCK_GUARD(&s->lock) {
-         if (ret < 0) {
-             assert(s->on_cbw_error == ON_CBW_ERROR_BREAK_SNAPSHOT);
--            if (!s->snapshot_error) {
--                s->snapshot_error = ret;
--            }
-+            qatomic_cmpxchg(&s->snapshot_error, 0, ret);
-         } else {
-             bdrv_set_dirty_bitmap(s->done_bitmap, off, end - off);
-         }
-@@ -215,7 +215,7 @@ cbw_snapshot_read_lock(BlockDriverState *bs, int64_t offset, int64_t bytes,
- 
-     QEMU_LOCK_GUARD(&s->lock);
- 
--    if (s->snapshot_error) {
-+    if (qatomic_read(&s->snapshot_error)) {
-         g_free(req);
-         return NULL;
-     }
-@@ -595,6 +595,12 @@ void bdrv_cbw_drop(BlockDriverState *bs)
-     bdrv_unref(bs);
- }
- 
-+int bdrv_cbw_snapshot_error(BlockDriverState *bs)
-+{
-+    BDRVCopyBeforeWriteState *s = bs->opaque;
-+    return qatomic_read(&s->snapshot_error);
-+}
-+
- static void cbw_init(void)
- {
-     bdrv_register(&bdrv_cbw_filter);
-diff --git a/block/copy-before-write.h b/block/copy-before-write.h
-index 2a5d4ba693..969da3620f 100644
---- a/block/copy-before-write.h
-+++ b/block/copy-before-write.h
-@@ -44,5 +44,6 @@ BlockDriverState *bdrv_cbw_append(BlockDriverState *source,
-                                   BlockCopyState **bcs,
-                                   Error **errp);
- void bdrv_cbw_drop(BlockDriverState *bs);
-+int bdrv_cbw_snapshot_error(BlockDriverState *bs);
- 
- #endif /* COPY_BEFORE_WRITE_H */
-diff --git a/pve-backup.c b/pve-backup.c
-index 0f098000dd..75da1dc051 100644
---- a/pve-backup.c
-+++ b/pve-backup.c
-@@ -374,6 +374,15 @@ static void pvebackup_complete_cb(void *opaque, int ret)
-         di->fleecing.snapshot_access = NULL;
-     }
-     if (di->fleecing.cbw) {
-+        /*
-+         * With fleecing, failure for cbw does not fail the guest write, but only sets the snapshot
-+         * error, making further requests to the snapshot fail with EACCES, which then also fail the
-+         * job. But that code is not the root cause and just confusing, so update it.
-+         */
-+        int snapshot_error = bdrv_cbw_snapshot_error(di->fleecing.cbw);
-+        if (di->completed_ret == -EACCES && snapshot_error) {
-+            di->completed_ret = snapshot_error;
-+        }
-         bdrv_cbw_drop(di->fleecing.cbw);
-         di->fleecing.cbw = NULL;
-     }

debian/patches/pve/0045-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/0046-PVE-backup-fixup-error-handling-for-fleecing.patch

@@ -1,103 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Fiona Ebner <f.ebner@proxmox.com>
-Date: Thu, 7 Nov 2024 17:51:14 +0100
-Subject: [PATCH] PVE backup: fixup error handling for fleecing
-
-The drained section needs to be terminated before breaking out of the
-loop in the error scenarios. Otherwise, guest IO on the drive would
-become stuck.
-
-If the job is created successfully, then the job completion callback
-will clean up the snapshot access block nodes. In case failure
-happened before the job is created, there was no cleanup for the
-snapshot access block nodes yet. Add it.
-
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
----
- pve-backup.c | 38 +++++++++++++++++++++++++-------------
- 1 file changed, 25 insertions(+), 13 deletions(-)
-
-diff --git a/pve-backup.c b/pve-backup.c
-index 75da1dc051..167f0b5c3f 100644
---- a/pve-backup.c
-+++ b/pve-backup.c
-@@ -357,22 +357,23 @@ static void coroutine_fn pvebackup_co_complete_stream(void *opaque)
-     qemu_co_mutex_unlock(&backup_state.backup_mutex);
- }
- 
--static void pvebackup_complete_cb(void *opaque, int ret)
-+static void cleanup_snapshot_access(PVEBackupDevInfo *di)
- {
--    PVEBackupDevInfo *di = opaque;
--    di->completed_ret = ret;
--
--    /*
--     * Handle block-graph specific cleanup (for fleecing) outside of the coroutine, because the work
--     * won't be done as a coroutine anyways:
--     * - For snapshot_access, allows doing bdrv_unref() directly. Doing it via bdrv_co_unref() would
--     *   just spawn a BH calling bdrv_unref().
--     * - For cbw, draining would need to spawn a BH.
--     */
-     if (di->fleecing.snapshot_access) {
-         bdrv_unref(di->fleecing.snapshot_access);
-         di->fleecing.snapshot_access = NULL;
-     }
-+    if (di->fleecing.cbw) {
-+        bdrv_cbw_drop(di->fleecing.cbw);
-+        di->fleecing.cbw = NULL;
-+    }
-+}
-+
-+static void pvebackup_complete_cb(void *opaque, int ret)
-+{
-+    PVEBackupDevInfo *di = opaque;
-+    di->completed_ret = ret;
-+
-     if (di->fleecing.cbw) {
-         /*
-          * With fleecing, failure for cbw does not fail the guest write, but only sets the snapshot
-@@ -383,10 +384,17 @@ static void pvebackup_complete_cb(void *opaque, int ret)
-         if (di->completed_ret == -EACCES && snapshot_error) {
-             di->completed_ret = snapshot_error;
-         }
--        bdrv_cbw_drop(di->fleecing.cbw);
--        di->fleecing.cbw = NULL;
-     }
- 
-+    /*
-+     * Handle block-graph specific cleanup (for fleecing) outside of the coroutine, because the work
-+     * won't be done as a coroutine anyways:
-+     * - For snapshot_access, allows doing bdrv_unref() directly. Doing it via bdrv_co_unref() would
-+     *   just spawn a BH calling bdrv_unref().
-+     * - For cbw, draining would need to spawn a BH.
-+     */
-+    cleanup_snapshot_access(di);
-+
-     /*
-      * Needs to happen outside of coroutine, because it takes the graph write lock.
-      */
-@@ -587,6 +595,7 @@ static void create_backup_jobs_bh(void *opaque) {
-             if (!di->fleecing.cbw) {
-                 error_setg(errp, "appending cbw node for fleecing failed: %s",
-                            local_err ? error_get_pretty(local_err) : "unknown error");
-+                bdrv_drained_end(di->bs);
-                 break;
-             }
- 
-@@ -599,6 +608,8 @@ static void create_backup_jobs_bh(void *opaque) {
-             if (!di->fleecing.snapshot_access) {
-                 error_setg(errp, "setting up snapshot access for fleecing failed: %s",
-                            local_err ? error_get_pretty(local_err) : "unknown error");
-+                cleanup_snapshot_access(di);
-+                bdrv_drained_end(di->bs);
-                 break;
-             }
-             source_bs = di->fleecing.snapshot_access;
-@@ -637,6 +648,7 @@ static void create_backup_jobs_bh(void *opaque) {
-         }
- 
-         if (!job || local_err) {
-+            cleanup_snapshot_access(di);
-             error_setg(errp, "backup_job_create failed: %s",
-                        local_err ? error_get_pretty(local_err) : "null");
-             break;

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

@@ -1,135 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Fiona Ebner <f.ebner@proxmox.com>
-Date: Thu, 7 Nov 2024 17:51:15 +0100
-Subject: [PATCH] PVE backup: factor out setting up snapshot access for
- fleecing
-
-Avoids some line bloat in the create_backup_jobs_bh() function and is
-in preparation for setting up the snapshot access independently of
-fleecing, in particular that will be useful for providing access to
-the snapshot via NBD.
-
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
----
- pve-backup.c | 95 ++++++++++++++++++++++++++++++++--------------------
- 1 file changed, 58 insertions(+), 37 deletions(-)
-
-diff --git a/pve-backup.c b/pve-backup.c
-index 167f0b5c3f..f136d004c4 100644
---- a/pve-backup.c
-+++ b/pve-backup.c
-@@ -525,6 +525,62 @@ static int coroutine_fn pvebackup_co_add_config(
-     goto out;
- }
- 
-+/*
-+ * Setup a snapshot-access block node for a device with associated fleecing image.
-+ */
-+static int setup_snapshot_access(PVEBackupDevInfo *di, Error **errp)
-+{
-+    Error *local_err = NULL;
-+
-+    if (!di->fleecing.bs) {
-+        error_setg(errp, "no associated fleecing image");
-+        return -1;
-+    }
-+
-+    QDict *cbw_opts = qdict_new();
-+    qdict_put_str(cbw_opts, "driver", "copy-before-write");
-+    qdict_put_str(cbw_opts, "file", bdrv_get_node_name(di->bs));
-+    qdict_put_str(cbw_opts, "target", bdrv_get_node_name(di->fleecing.bs));
-+
-+    if (di->bitmap) {
-+        /*
-+         * Only guest writes to parts relevant for the backup need to be intercepted with
-+         * old data being copied to the fleecing image.
-+         */
-+        qdict_put_str(cbw_opts, "bitmap.node", bdrv_get_node_name(di->bs));
-+        qdict_put_str(cbw_opts, "bitmap.name", bdrv_dirty_bitmap_name(di->bitmap));
-+    }
-+    /*
-+     * Fleecing storage is supposed to be fast and it's better to break backup than guest
-+     * writes. Certain guest drivers like VirtIO-win have 60 seconds timeout by default, so
-+     * abort a bit before that.
-+     */
-+    qdict_put_str(cbw_opts, "on-cbw-error", "break-snapshot");
-+    qdict_put_int(cbw_opts, "cbw-timeout", 45);
-+
-+    di->fleecing.cbw = bdrv_insert_node(di->bs, cbw_opts, BDRV_O_RDWR, &local_err);
-+
-+    if (!di->fleecing.cbw) {
-+        error_setg(errp, "appending cbw node for fleecing failed: %s",
-+                   local_err ? error_get_pretty(local_err) : "unknown error");
-+        return -1;
-+    }
-+
-+    QDict *snapshot_access_opts = qdict_new();
-+    qdict_put_str(snapshot_access_opts, "driver", "snapshot-access");
-+    qdict_put_str(snapshot_access_opts, "file", bdrv_get_node_name(di->fleecing.cbw));
-+
-+    di->fleecing.snapshot_access =
-+        bdrv_open(NULL, NULL, snapshot_access_opts, BDRV_O_RDWR | BDRV_O_UNMAP, &local_err);
-+    if (!di->fleecing.snapshot_access) {
-+        error_setg(errp, "setting up snapshot access for fleecing failed: %s",
-+                   local_err ? error_get_pretty(local_err) : "unknown error");
-+        return -1;
-+    }
-+
-+    return 0;
-+}
-+
- /*
-  * backup_job_create can *not* be run from a coroutine, so this can't either.
-  * The caller is responsible that backup_mutex is held nonetheless.
-@@ -569,49 +625,14 @@ static void create_backup_jobs_bh(void *opaque) {
-         const char *job_id = bdrv_get_device_name(di->bs);
-         bdrv_graph_co_rdunlock();
-         if (di->fleecing.bs) {
--            QDict *cbw_opts = qdict_new();
--            qdict_put_str(cbw_opts, "driver", "copy-before-write");
--            qdict_put_str(cbw_opts, "file", bdrv_get_node_name(di->bs));
--            qdict_put_str(cbw_opts, "target", bdrv_get_node_name(di->fleecing.bs));
--
--            if (di->bitmap) {
--                /*
--                 * Only guest writes to parts relevant for the backup need to be intercepted with
--                 * old data being copied to the fleecing image.
--                 */
--                qdict_put_str(cbw_opts, "bitmap.node", bdrv_get_node_name(di->bs));
--                qdict_put_str(cbw_opts, "bitmap.name", bdrv_dirty_bitmap_name(di->bitmap));
--            }
--            /*
--             * Fleecing storage is supposed to be fast and it's better to break backup than guest
--             * writes. Certain guest drivers like VirtIO-win have 60 seconds timeout by default, so
--             * abort a bit before that.
--             */
--            qdict_put_str(cbw_opts, "on-cbw-error", "break-snapshot");
--            qdict_put_int(cbw_opts, "cbw-timeout", 45);
--
--            di->fleecing.cbw = bdrv_insert_node(di->bs, cbw_opts, BDRV_O_RDWR, &local_err);
--
--            if (!di->fleecing.cbw) {
--                error_setg(errp, "appending cbw node for fleecing failed: %s",
--                           local_err ? error_get_pretty(local_err) : "unknown error");
--                bdrv_drained_end(di->bs);
--                break;
--            }
--
--            QDict *snapshot_access_opts = qdict_new();
--            qdict_put_str(snapshot_access_opts, "driver", "snapshot-access");
--            qdict_put_str(snapshot_access_opts, "file", bdrv_get_node_name(di->fleecing.cbw));
--
--            di->fleecing.snapshot_access =
--                bdrv_open(NULL, NULL, snapshot_access_opts, BDRV_O_RDWR | BDRV_O_UNMAP, &local_err);
--            if (!di->fleecing.snapshot_access) {
-+            if (setup_snapshot_access(di, &local_err) < 0) {
-                 error_setg(errp, "setting up snapshot access for fleecing failed: %s",
-                            local_err ? error_get_pretty(local_err) : "unknown error");
-                 cleanup_snapshot_access(di);
-                 bdrv_drained_end(di->bs);
-                 break;
-             }
-+
-             source_bs = di->fleecing.snapshot_access;
-             discard_source = true;
- 

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

@@ -1,135 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Fiona Ebner <f.ebner@proxmox.com>
-Date: Thu, 7 Nov 2024 17:51:16 +0100
-Subject: [PATCH] PVE backup: save device name in device info structure
-
-The device name needs to be queried while holding the graph read lock
-and since it doesn't change during the whole operation, just get it
-once during setup and avoid the need to query it again in different
-places.
-
-Also in preparation to use it more often in error messages and for the
-upcoming external backup access API.
-
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
----
- pve-backup.c | 29 +++++++++++++++--------------
- 1 file changed, 15 insertions(+), 14 deletions(-)
-
-diff --git a/pve-backup.c b/pve-backup.c
-index f136d004c4..8ccb281c8c 100644
---- a/pve-backup.c
-+++ b/pve-backup.c
-@@ -94,6 +94,7 @@ typedef struct PVEBackupDevInfo {
-     size_t size;
-     uint64_t block_size;
-     uint8_t dev_id;
-+    char* device_name;
-     int completed_ret; // INT_MAX if not completed
-     BdrvDirtyBitmap *bitmap;
-     BlockDriverState *target;
-@@ -327,6 +328,8 @@ static void coroutine_fn pvebackup_co_complete_stream(void *opaque)
-     }
- 
-     di->bs = NULL;
-+    g_free(di->device_name);
-+    di->device_name = NULL;
- 
-     assert(di->target == NULL);
- 
-@@ -621,9 +624,6 @@ static void create_backup_jobs_bh(void *opaque) {
- 
-         BlockDriverState *source_bs = di->bs;
-         bool discard_source = false;
--        bdrv_graph_co_rdlock();
--        const char *job_id = bdrv_get_device_name(di->bs);
--        bdrv_graph_co_rdunlock();
-         if (di->fleecing.bs) {
-             if (setup_snapshot_access(di, &local_err) < 0) {
-                 error_setg(errp, "setting up snapshot access for fleecing failed: %s",
-@@ -654,7 +654,7 @@ static void create_backup_jobs_bh(void *opaque) {
-         }
- 
-         BlockJob *job = backup_job_create(
--            job_id, source_bs, di->target, backup_state.speed, sync_mode, di->bitmap,
-+            di->device_name, source_bs, di->target, backup_state.speed, sync_mode, di->bitmap,
-             bitmap_mode, false, discard_source, NULL, &perf, BLOCKDEV_ON_ERROR_REPORT,
-             BLOCKDEV_ON_ERROR_REPORT, JOB_DEFAULT, pvebackup_complete_cb, di, backup_state.txn,
-             &local_err);
-@@ -751,6 +751,7 @@ static GList coroutine_fn GRAPH_RDLOCK *get_device_info(
-             }
-             PVEBackupDevInfo *di = g_new0(PVEBackupDevInfo, 1);
-             di->bs = bs;
-+            di->device_name = g_strdup(bdrv_get_device_name(bs));
- 
-             if (fleecing && device_uses_fleecing(*d)) {
-                 g_autofree gchar *fleecing_devid = g_strconcat(*d, "-fleecing", NULL);
-@@ -789,6 +790,7 @@ static GList coroutine_fn GRAPH_RDLOCK *get_device_info(
- 
-             PVEBackupDevInfo *di = g_new0(PVEBackupDevInfo, 1);
-             di->bs = bs;
-+            di->device_name = g_strdup(bdrv_get_device_name(bs));
-             di_list = g_list_append(di_list, di);
-         }
-     }
-@@ -956,9 +958,6 @@ UuidInfo coroutine_fn *qmp_backup(
- 
-             di->block_size = dump_cb_block_size;
- 
--            bdrv_graph_co_rdlock();
--            const char *devname = bdrv_get_device_name(di->bs);
--            bdrv_graph_co_rdunlock();
-             PBSBitmapAction action = PBS_BITMAP_ACTION_NOT_USED;
-             size_t dirty = di->size;
- 
-@@ -973,7 +972,8 @@ UuidInfo coroutine_fn *qmp_backup(
-                     }
-                     action = PBS_BITMAP_ACTION_NEW;
-                 } else {
--                    expect_only_dirty = proxmox_backup_check_incremental(pbs, devname, di->size) != 0;
-+                    expect_only_dirty =
-+                        proxmox_backup_check_incremental(pbs, di->device_name, di->size) != 0;
-                 }
- 
-                 if (expect_only_dirty) {
-@@ -997,7 +997,8 @@ UuidInfo coroutine_fn *qmp_backup(
-                 }
-             }
- 
--            int dev_id = proxmox_backup_co_register_image(pbs, devname, di->size, expect_only_dirty, errp);
-+            int dev_id = proxmox_backup_co_register_image(pbs, di->device_name, di->size,
-+                                                          expect_only_dirty, errp);
-             if (dev_id < 0) {
-                 goto err_mutex;
-             }
-@@ -1009,7 +1010,7 @@ UuidInfo coroutine_fn *qmp_backup(
-             di->dev_id = dev_id;
- 
-             PBSBitmapInfo *info = g_malloc(sizeof(*info));
--            info->drive = g_strdup(devname);
-+            info->drive = g_strdup(di->device_name);
-             info->action = action;
-             info->size = di->size;
-             info->dirty = dirty;
-@@ -1034,10 +1035,7 @@ UuidInfo coroutine_fn *qmp_backup(
-                 goto err_mutex;
-             }
- 
--            bdrv_graph_co_rdlock();
--            const char *devname = bdrv_get_device_name(di->bs);
--            bdrv_graph_co_rdunlock();
--            di->dev_id = vma_writer_register_stream(vmaw, devname, di->size);
-+            di->dev_id = vma_writer_register_stream(vmaw, di->device_name, di->size);
-             if (di->dev_id <= 0) {
-                 error_set(errp, ERROR_CLASS_GENERIC_ERROR,
-                           "register_stream failed");
-@@ -1148,6 +1146,9 @@ err:
-             bdrv_co_unref(di->target);
-         }
- 
-+        g_free(di->device_name);
-+        di->device_name = NULL;
-+
-         g_free(di);
-     }
-     g_list_free(di_list);

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

@@ -1,25 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Fiona Ebner <f.ebner@proxmox.com>
-Date: Thu, 7 Nov 2024 17:51:17 +0100
-Subject: [PATCH] PVE backup: include device name in error when setting up
- snapshot access fails
-
-Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
----
- pve-backup.c | 3 ++-
- 1 file changed, 2 insertions(+), 1 deletion(-)
-
-diff --git a/pve-backup.c b/pve-backup.c
-index 8ccb281c8c..255465676c 100644
---- a/pve-backup.c
-+++ b/pve-backup.c
-@@ -626,7 +626,8 @@ static void create_backup_jobs_bh(void *opaque) {
-         bool discard_source = false;
-         if (di->fleecing.bs) {
-             if (setup_snapshot_access(di, &local_err) < 0) {
--                error_setg(errp, "setting up snapshot access for fleecing failed: %s",
-+                error_setg(errp, "%s - setting up snapshot access for fleecing failed: %s",
-+                           di->device_name,
-                            local_err ? error_get_pretty(local_err) : "unknown error");
-                 cleanup_snapshot_access(di);
-                 bdrv_drained_end(di->bs);

debian/patches/pve/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

@@ -1,14 +1,5 @@
 extra/0001-monitor-qmp-fix-race-with-clients-disconnecting-earl.patch
 extra/0002-ide-avoid-potential-deadlock-when-draining-during-tr.patch
-extra/0003-tcg-Reset-free_temps-before-tcg_optimize.patch
-extra/0004-target-i386-Reset-TSCs-of-parked-vCPUs-too-on-VM-res.patch
-extra/0005-i386-cpu-Mark-avx10_version-filtered-when-prefix-is-.patch
-extra/0006-net-Fix-announce_self.patch
-extra/0007-net-dump-Correctly-compute-Ethernet-packet-offset.patch
-extra/0008-pci-acpi-Windows-PCI-Label-Id-bug-workaround.patch
-extra/0009-hw-usb-hcd-xhci-pci-Use-modulo-to-select-MSI-vector-.patch
-extra/0010-pci-ensure-valid-link-status-bits-for-downstream-por.patch
-extra/0011-pci-msix-Fix-msix-pba-read-vector-poll-end-calculati.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
@@ -56,12 +47,22 @@ 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-alloc-track-error-out-when-auto-remove-is-not-set.patch
-pve/0043-alloc-track-avoid-seemingly-superfluous-child-permis.patch
-pve/0044-PVE-backup-add-fleecing-option.patch
-pve/0045-PVE-backup-improve-error-when-copy-before-write-fail.patch
-pve/0046-PVE-backup-fixup-error-handling-for-fleecing.patch
-pve/0047-PVE-backup-factor-out-setting-up-snapshot-access-for.patch
-pve/0048-PVE-backup-save-device-name-in-device-info-structure.patch
-pve/0049-PVE-backup-include-device-name-in-error-when-setting.patch
-pve/0050-adapt-machine-version-deprecation-for-Proxmox-VE.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