Useful for retrieving flags from default CPU models.
Signed-off-by: Stefan Reiter
---
...E-Add-list-flags-command-line-option.patch | 213 ++
debian/patches/series | 1 +
2 files changed, 214 insertions(+)
create mode 100644
debian/patches/pve/0049
For supporting a GUI to easily create custom CPU models based on cluster
CPU availability.
Signed-off-by: Stefan Reiter
---
PVE/API2/Qemu/CPU.pm | 86
1 file changed, 86 insertions(+)
diff --git a/PVE/API2/Qemu/CPU.pm b/PVE/API2/Qemu/CPU.pm
index
"originalValue" is set to "null" in case it's the default, but
getValue() returns an empty string. This means that when editing a VM's
CPU config when the model is "default", the form would always be marked
dirty.
Signed-off-by: Stefan Reiter
---
w
and would have
to stay up-to-date with supported CPU models in QEMU.
pve-qemu: Stefan Reiter (1):
Add -list-flags command line option
...E-Add-list-flags-command-line-option.patch | 213 ++
debian/patches/series | 1 +
2 files changed, 214 insertions(
onfig
fails.
Signed-off-by: Stefan Reiter
---
PVE/API2/Qemu/CPU.pm| 174 +++-
PVE/QemuServer/CPUConfig.pm | 17 +++-
2 files changed, 189 insertions(+), 2 deletions(-)
diff --git a/PVE/API2/Qemu/CPU.pm b/PVE/API2/Qemu/CPU.pm
index 9cc1d77..8a3dcfd 100644
sidebar, so
add a new class that correctly aligns pve-itype-icon-* elements.
Signed-off-by: Stefan Reiter
---
www/css/ext6-pve.css | 4 +
www/manager6/Makefile | 3 +
www/manager6/dc/CPUTypeEdit.js| 81 ++
www/manager6/dc/CPUTypeView.js
Signed-off-by: Stefan Reiter
---
I believe this one is the most outdated, it's original title in my working
branch was "tmp", dating back to more than a year ago ;)
www/manager6/dc/CPUTypeEdit.js | 8 ++
www/manager6/form/VMCPUFlagSelector.js | 125 +++
case someone retrieved it to modify it with a PUT call afterwards.
Signed-off-by: Stefan Reiter
---
PVE/API2/Qemu/CPU.pm| 76 -
PVE/QemuServer/CPUConfig.pm | 7
2 files changed, 82 insertions(+), 1 deletion(-)
diff --git a/PVE/API2/Qemu/CPU.pm b
...and typo in name.
Signed-off-by: Stefan Reiter
---
www/manager6/form/VMCPUFlagSelector.js | 12 ++--
1 file changed, 6 insertions(+), 6 deletions(-)
diff --git a/www/manager6/form/VMCPUFlagSelector.js
b/www/manager6/form/VMCPUFlagSelector.js
index ace3c531..a6d17930 100644
--- a
When set to false, only default CPU models will be shown.
Signed-off-by: Stefan Reiter
---
www/manager6/form/CPUModelSelector.js | 16
1 file changed, 16 insertions(+)
diff --git a/www/manager6/form/CPUModelSelector.js
b/www/manager6/form/CPUModelSelector.js
index a4d883a4
CONFIG_SPICE)
in the QAPI JSON as that semantic was renamed in the meantime.
Signed-off-by: Stefan Reiter
---
See: https://lists.nongnu.org/archive/html/qemu-devel/2021-10/msg06676.html
...atch => 0003-qxl-fix-pre-save-logic.patch} | 0
...upport-for-flag-argument-with-value.patch} |
auf-7-0.97388/
https://forum.proxmox.com/threads/error-with-backup-when-backing-up-qmp-command-query-backup-failed-got-wrong-command-id.88017/page-3#post-416339
Signed-off-by: Stefan Reiter
---
PVE/VZDump/QemuServer.pm | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/PVE/VZDump/QemuSe
Works as intended, both i440fx and Q35 run with a secure boot efidisk.
Tested-by: Stefan Reiter
Reviewed-by: Stefan Reiter
On 10/11/21 2:10 PM, Dominik Csapak wrote:
ovmf with SMM enabled will not boot on i440fx (hangs on graphics
initialization), so load the non SMM variant.
should be no
On 10/21/21 11:47 AM, Thomas Lamprecht wrote:
On 21.10.21 11:34, Stefan Reiter wrote:
On 10/21/21 10:36 AM, Thomas Lamprecht wrote:
First add a new meta property that is currently exclusively set on new
VM creation and then read-only, initially add the creation time as UNIX
epoch and the QEMU
On 10/21/21 10:36 AM, Thomas Lamprecht wrote:
See commit 17858a1695 (hw/acpi/ich9: Set ACPI PCI hot-plug as default
on Q35)[0] in upstream QEMU repository for details about why the
change was made.
As that change affects systemds predictable interface naming[1],
e.g., by going from a previously
On 10/21/21 10:36 AM, Thomas Lamprecht wrote:
Signed-off-by: Thomas Lamprecht
---
test/cfg2cmd/q35-simple-6.0.conf| 13 ++
test/cfg2cmd/q35-simple-6.0.conf.cmd| 28 +
test/cfg2cmd/q35-simple-6.1.conf| 14 +++
test/cfg2cmd/q
On 10/21/21 10:36 AM, Thomas Lamprecht wrote:
First add a new meta property that is currently exclusively set on new
VM creation and then read-only, initially add the creation time as UNIX
epoch and the QEMU version that was installed during installation
(thought about using the one on first star
the timeout duration should be more than enough.
Also add a missing 'usleep' import, 'usleep' was used before but never
imported, apparently the other case never got triggered...
Signed-off-by: Stefan Reiter
---
Triggered during rollback testing, potentially due to extra lo
QEMU doesn't know about the tpmstate, so 'do_snapshots_with_qemu' should
never return true in that case. Note that inconsistencies related to
snapshot timing do not matter much, as the actual TPM data is exported
together with other device state by QEMU anyway.
Signed-off-by
Signed-off-by: Stefan Reiter
---
Applies on top of the 6.1 rebase.
.../extra/0006-qxl-fix-pre-save-logic.patch | 34 +++
debian/patches/series | 1 +
2 files changed, 35 insertions(+)
create mode 100644 debian/patches/extra/0006-qxl-fix-pre-save
and reorder the panel a bit, as otherwise I found it *very* cluttered.
Signed-off-by: Stefan Reiter
---
www/manager6/qemu/SystemEdit.js | 50 +++--
1 file changed, 22 insertions(+), 28 deletions(-)
diff --git a/www/manager6/qemu/SystemEdit.js b/www/manager6/qemu
Signed-off-by: Stefan Reiter
---
PVE/QemuServer.pm | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/PVE/QemuServer.pm b/PVE/QemuServer.pm
index eb29fc2..6fc28e8 100644
--- a/PVE/QemuServer.pm
+++ b/PVE/QemuServer.pm
@@ -386,7 +386,7 @@ my $confdesc = {
ostype
Add a specific 'win11' ostype for Windows 11 and Server 2022, with currently no
change on the backend, but enable q35, secure boot and a TPM in the wizard by
default if selected.
qemu-server: Stefan Reiter (1):
ostype: support Windows 11/Server 2022
PVE/QemuServer.pm | 3 ++-
1 fi
Windows 11 requires a TPM and secure boot, so try to auto-enable both.
Signed-off-by: Stefan Reiter
---
www/manager6/Utils.js | 1 +
www/manager6/qemu/SystemEdit.js | 16
2 files changed, 17 insertions(+)
diff --git a/www/manager6/Utils.js b/www/manager6/Utils.js
Signed-off-by: Stefan Reiter
---
qm.adoc | 30 --
1 file changed, 24 insertions(+), 6 deletions(-)
diff --git a/qm.adoc b/qm.adoc
index 80734e3..93576c7 100644
--- a/qm.adoc
+++ b/qm.adoc
@@ -733,10 +733,13 @@ the operating system. By default QEMU uses *SeaBIOS* for
Signed-off-by: Stefan Reiter
---
qm.adoc | 30 ++
1 file changed, 30 insertions(+)
diff --git a/qm.adoc b/qm.adoc
index 93576c7..b9f4269 100644
--- a/qm.adoc
+++ b/qm.adoc
@@ -775,6 +775,36 @@ you need to set the client resolution in the OVMF menu
(which you can
Provide support for secure boot by using the new "4m" and "4m-ms"
variants of the OVMF code/vars templates. This is specified on the
efidisk via the 'efitype' and 'ms-keys' parameters.
Signed-off-by: Stefan Reiter
---
Should depend on updated pve-edk2-
similar to an efidisk - this eliminates any concerns about
storing on pmxcfs
* always allocate the state as 4 MiB (on directory storage it might auto-shrink)
* fixes migration, since source and destination are now different
* add GUI patch
storage: Stefan Reiter (1):
import: don't check
to avoid it ever being treated as a regular drive
from the rest of the stack in case it gets left over after a backup for
some reason (shouldn't happen).
Signed-off-by: Stefan Reiter
---
PVE/API2/Qemu.pm | 5 ++
PVE/QemuMigrate.pm | 14 +++-
PVE/QemuServer.pm|
TPM state disks on directory storages may have completely unaligned
sizes, this check doesn't make sense for them.
This appears to just be a (weak) safeguard and not serve an actual
functional purpose, so simply get rid of it to allow migration of TPM
state.
Signed-off-by: Stefan R
Inspired by HDEfi for efidisks. Extends the DiskStorageSelector to allow
hiding the format, since tpmstate can only be stored in 'raw' format
(even on directory storages).
Signed-off-by: Stefan Reiter
---
www/manager6/Makefile| 1 +
www/manager
27;t work. swtpm itself has no method of
moving a backing storage at runtime.
Signed-off-by: Stefan Reiter
---
PVE/API2/Qemu.pm| 5 ++
PVE/QemuMigrate.pm | 14 -
PVE/QemuServer.pm | 135 +---
PVE/QemuServer/Drive.pm | 63 +++
Necessary for an OS to use a TPM attached to a OVMF VM.
Signed-off-by: Stefan Reiter
---
debian/rules | 4
1 file changed, 4 insertions(+)
diff --git a/debian/rules b/debian/rules
index fb85b29..ece77be 100755
--- a/debian/rules
+++ b/debian/rules
@@ -54,6 +54,8 @@ build-ovmf: setup-build
Inspired by HDEfi for efidisks. Extends the DiskStorageSelector to allow
hiding the format, since tpmstate can only be stored in 'raw' format
(even on directory storages).
Signed-off-by: Stefan Reiter
---
www/manager6/Makefile| 1 +
www/manager
ate the state as 4 MiB (on directory storage it might auto-shrink)
* fixes migration, since source and destination are now different
* add GUI patch
edk2-firmware: Stefan Reiter (1):
enable TPM and TPM2 support
debian/rules | 4 ++++
1 file changed, 4 insertions(+)
storage: Stefan Reiter (1
TPM state disks on directory storages may have completely unaligned
sizes, this check doesn't make sense for them.
This appears to just be a (weak) safeguard and not serve an actual
functional purpose, so simply get rid of it to allow migration of TPM
state.
Signed-off-by: Stefan R
FreeBSD 13 VMs will bug out and consume loads of idle CPU time if the
ratelimit is 1024 B/s or lower. Increasing to 2048 fixes the issue,
confirmed experimentally, and doesn't seem much worse as a default.
Signed-off-by: Stefan Reiter
---
On the one hand this *really* sounds like a FreeBS
igration should always work.
Tested-by: Stefan Reiter
If it breaks that we'd need to fallback to the old "false" with a QEMU
machine version check and we could only drop it then in PVE 8.0, so hopefully
it Just Works™ to safe that hassle..
__
On 8/19/21 3:21 PM, Radosław Korzeniewski wrote:
Hello,
I'm working on implementing VM incremental backup for a 3-party solution
based on Qemu dirty bitmap which is available in Proxmox. Unfortunately I'm
unable to make it working, so I decided to ask for some help.
First I'm adding a dirty bitm
On 09/08/2021 20:17, Nick Chevsky wrote:
Hi Stefan,
Thank you for your work on this; I've been testing it locally for a few
weeks and have since contributed improved Debian packaging and other fixes
upstream [3]. Please see my comment below the quoted code:
Thanks for your upstream work! I've
On 7/23/21 11:02 AM, Thomas Lamprecht wrote:
On 08.07.21 13:46, Stefan Reiter wrote:
If it doesn't exist, there's no need to keep it around at all. Such a
backup might easily be created if an efidisk is configured with BIOS
set to anything but 'OVMF'.
Unbreaks live-restor
7;tpm' config option is removed or the
version changed, effectively clearing any stored keys/data.
Signed-off-by: Stefan Reiter
---
PVE/QemuServer.pm | 134 +-
1 file changed, 133 insertions(+), 1 deletion(-)
diff --git a/PVE/QemuServer.pm b/PVE/Qe
pshots don't include TPM data
* migration may hickup, since destination and target are effectively "the same"
* needs improved clearing logic, or potentially none at all (would be a benefit
for the "efidisk-like" variant)
edk2-firmware: Stefan Reiter (1):
enable TPM
Necessary for an OS to use a TPM attached to a OVMF VM.
Signed-off-by: Stefan Reiter
---
debian/rules | 4
1 file changed, 4 insertions(+)
diff --git a/debian/rules b/debian/rules
index fb85b29..ece77be 100755
--- a/debian/rules
+++ b/debian/rules
@@ -54,6 +54,8 @@ build-ovmf: setup-build
If it doesn't exist, there's no need to keep it around at all. Such a
backup might easily be created if an efidisk is configured with BIOS
set to anything but 'OVMF'.
Unbreaks live-restore for such cases too.
Signed-off-by: Stefan Reiter
---
v1->v2:
* Actually remove e
...instead of hard failing. Requires a RPCEnvironment to issue the
warnings, so init a dummy one in run_config2command_tests.
Signed-off-by: Stefan Reiter
---
PVE/QemuServer.pm| 10 ++
test/run_config2command_tests.pl | 2 ++
2 files changed, 12 insertions(+)
diff
Signed-off-by: Stefan Reiter
---
PVE/QemuServer.pm | 8
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/PVE/QemuServer.pm b/PVE/QemuServer.pm
index b0fe257..0d8affa 100644
--- a/PVE/QemuServer.pm
+++ b/PVE/QemuServer.pm
@@ -6381,13 +6381,13 @@ sub
up in general.
Signed-off-by: Stefan Reiter
---
PVE/QemuServer.pm | 7 +++
1 file changed, 7 insertions(+)
diff --git a/PVE/QemuServer.pm b/PVE/QemuServer.pm
index 4082e69..a86a2a4 100644
--- a/PVE/QemuServer.pm
+++ b/PVE/QemuServer.pm
@@ -6471,6 +6471,13 @@ sub restore_proxmox_backup_ar
Firefox doesn't correctly break to the next line if the line-height is
set as is (with certain fonts, including the default on debian it
seems). Simply remove it, as it isn't necessary.
Suggested-by: Dominik Csapak
Signed-off-by: Stefan Reiter
---
www/manager6/dc/Guests.js | 1
: Stefan Reiter
---
PVE/API2/Qemu.pm | 8
1 file changed, 8 insertions(+)
diff --git a/PVE/API2/Qemu.pm b/PVE/API2/Qemu.pm
index 1e540f5..b750db0 100644
--- a/PVE/API2/Qemu.pm
+++ b/PVE/API2/Qemu.pm
@@ -1357,6 +1357,14 @@ my $update_vm_api = sub {
if defined($conf
On 7/1/21 12:57 PM, Thomas Lamprecht wrote:
On 01.07.21 11:37, Stefan Reiter wrote:
The efidisk never got restored correctly before, since we don't use the
generic print_drive_commandline_full for it, and as such it didn't get a
backing image attached. This not only causes the efidi
t from the live
restore entirely.
Signed-off-by: Stefan Reiter
---
best viewed with -w
I guess in QEMU < 6.0 this didn't cause an error, since I definitely tested with
EFI VMs before, however, it would have never actually restored the data and
instead always created a new one.
=threads as it used to.
Signed-off-by: Stefan Reiter
---
I tried my usual QEMU debugging routine, but couldn't figure anything specific
out - it's not related to EFI, happens on SeaBIOS too, but looking at GDB just
appears that it's waiting for the io_uring fd to become ready (in
zfs_mirror_size_check is already called in get_zfs_raid_setup when
necessary, so don't call it unconditionally, as this will cause a
false-positive error on RAID0 devices (where size mismatches are fine).
Signed-off-by: Stefan Reiter
---
proxinstall | 2 --
1 file changed, 2 deletions(-)
On 22/06/2021 08:02, Thomas Lamprecht wrote:
On 21.06.21 18:35, Stefan Reiter wrote:
KillMode 'none' is deprecated, and systemd loudly complains about that
in the journal. To avoid the warning, but keep the behaviour the same,
use KillMode 'process'.
This mode does tw
nt/101.scope:3: Unit configured to use KillMode=none. [...]
I still get the same warning about a 'ceph-volume@.service' in my logs, after
upgrading to pacific, though I'm not sure if that's something we need to/can
fix, or something left over from the upgrade - any ceph expertise we
lse, or
TimeoutStopUSec to infinity - for safety, we do both
In my testing, this replicated the previous behaviour exactly, but
without using the deprecated 'none' mode.
Signed-off-by: Stefan Reiter
---
Depends on updated pve-common from patch 2.
PVE/QemuServer.pm | 4 +++-
1 file
Used in qemu-server for avoiding KillMode 'none'. SendSIGKILL is a
boolean, so we need to use dbus_boolean to serialize it.
Signed-off-by: Stefan Reiter
---
src/PVE/Systemd.pm | 6 --
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/src/PVE/Systemd.pm b/src/PVE/System
similar for a version that'll probably never leave staging seemed
unnecessary.
For testing, build pve-qemu-kvm from git, as io_uring is already enabled
there.
On 6/21/21 5:33 PM, Stefan Reiter wrote:
The 'aio' setting is not visible to the guest, and so can be changed
during migratio
The 'aio' setting is not visible to the guest, and so can be changed
during migrations or snapshots without issue. It is thus only
dependendent on the actual QEMU version being >= 6.0, not machine
version.
Signed-off-by: Stefan Reiter
---
new in v2
tested migration from and to old
Note that the value in this enum directly represents the value passed to
QEMU, so we need to use the underscore.
Signed-off-by: Stefan Reiter
---
same as v1, par the commit message
PVE/QemuServer/Drive.pm | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/PVE/QemuServer
On 6/18/21 5:17 PM, Thomas Lamprecht wrote:
On 10.06.21 13:28, Dominik Csapak wrote:
On 6/10/21 11:29, Stefan Reiter wrote:
On 6/10/21 10:40 AM, Dominik Csapak wrote:
On 6/10/21 10:23, Stefan Reiter wrote:
On 6/10/21 9:37 AM, Dominik Csapak wrote:
and not the node where the browser connects
On 16/06/2021 20:27, Thomas Lamprecht wrote:
On 13.04.21 10:24, Fabian Ebner wrote:
Reported in the community forum[0].
In QEMU's hw/scsi/vmw_pvscsi.c in the SCSIBusInfo struct, the max_lun property
is set to 0. This means that in our stack, one cannot have multiple disks and
use 'scsihw: pvscs
On 6/15/21 10:04 AM, Fabian Ebner wrote:
Am 14.06.21 um 11:15 schrieb Stefan Reiter:
On 6/14/21 8:29 AM, Fabian Ebner wrote:
Any feedback for this?
Looks good to me and works as advertised:
Reviewed-by: Stefan Reiter
Tested-by: Stefan Reiter
Thanks for the review!
The `$unit
On 6/14/21 8:29 AM, Fabian Ebner wrote:
Any feedback for this?
Looks good to me and works as advertised:
Reviewed-by: Stefan Reiter
Tested-by: Stefan Reiter
The `$unit = $drive->{index} % maxdev` logic doesn't seem useful here though, I
think it's a copy/paste from the IDE/SA
On 6/10/21 10:40 AM, Dominik Csapak wrote:
On 6/10/21 10:23, Stefan Reiter wrote:
On 6/10/21 9:37 AM, Dominik Csapak wrote:
and not the node where the browser connects.
there are at least two good reasons for this:
* it is confusing, since the user would expect it to start where
the ui is
On 6/10/21 9:37 AM, Dominik Csapak wrote:
and not the node where the browser connects.
there are at least two good reasons for this:
* it is confusing, since the user would expect it to start where
the ui is pointint to
* the storage may not be available on the node the browser connects
to,
superseded by CachedChunkReader, with less code and more speed
Signed-off-by: Stefan Reiter
---
src/backup.rs| 3 -
src/backup/async_index_reader.rs | 215 ---
2 files changed, 218 deletions(-)
delete mode 100644 src/backup
admin/datastore reads linearly only, so no need for cache (capacity of 1
basically means no cache except for the currently active chunk).
mount can do random access too, so cache last 8 chunks for possibly a
mild performance improvement.
Signed-off-by: Stefan Reiter
---
src/api2/admin
Supports concurrent 'access' calls to the same key via a
BroadcastFuture. These are stored in a seperate HashMap, the LruCache
underneath is only modified once a valid value has been retrieved.
Signed-off-by: Stefan Reiter
---
src/tools.rs | 1 +
src/tools/async_lr
Provides a fast arbitrary read implementation with full async and
concurrency support.
Signed-off-by: Stefan Reiter
---
This is technically all that's needed for proxmox-backup-qemu to build and
function as intended, but I decided to also use this IMHO cleaner implementation
to replac
read_image_at, providing real
concurrency without lock contention
Signed-off-by: Stefan Reiter
---
src/restore.rs | 30 --
1 file changed, 8 insertions(+), 22 deletions(-)
diff --git a/src/restore.rs b/src/restore.rs
index 0790d7f..33959d9 100644
--- a/src/restore.rs
+++ b
Provides a shared AsyncLruCache of 256MB (w/ 4MB chunks) that can be
used by multiple readers at the same time. It is dropped once no more
readers exist, so the memory gets freed if all QEMU block/pbs instances
disappear.
Signed-off-by: Stefan Reiter
---
v2:
* adapt to dropping ChunkCache type
25 30 35 40
iteration
1.0.3-1A
patchedB
[0] see: https://lists.proxmox.com/pipermail/pbs-devel/2021-April/002932.html
proxmox-backup: Stefan Reiter (7):
tools/BroadcastFuture: add testcase for better understand
actual request concurrency is not supported in these traits).
Signed-off-by: Stefan Reiter
---
v2:
* drop 'seek_to_pos' from struct and implement error handling directly in
start_seek
* simplify future handling in poll_read with Option::get_or_insert_with
src/backup/cached_chunk
Setting this to 0 is not just useless, but breaks the logic horribly
enough to cause random segfaults - better forbid this, to avoid someone
else having to debug it again ;)
Signed-off-by: Stefan Reiter
---
src/tools/lru_cache.rs | 1 +
1 file changed, 1 insertion(+)
diff --git a/src/tools
Explicitly test that data will stay available and can be retrieved
immediately via listen(), even if the future producing the data and
notifying the consumers was already run in the past.
Wasn't broken or anything, but helps with understanding IMO.
Signed-off-by: Stefan Reiter
---
W
Note that the value in this enum directly represents the value passed to
QEMU, so we need to use the underscore.
Off by default, updated QEMU required, but no compatibility issues, as
this is a new value that wasn't valid before, and setting it is the
users choice.
Signed-off-by: Stefan R
in Debian bullseye, but there seems to be some dependency issues with our
pve-kernel-libc-dev package. For now, testing was conducted by building liburing
from source, i.e. clone 'https://github.com/axboe/liburing' and run
'make-debs.sh'.
pve-qemu: Stefan Reiter (1):
enable i
Signed-off-by: Stefan Reiter
---
Oh if only all QEMU changes were this simple...
debian/control | 1 +
debian/rules | 1 +
2 files changed, 2 insertions(+)
diff --git a/debian/control b/debian/control
index 38c45ce..4da3409 100644
--- a/debian/control
+++ b/debian/control
@@ -27,6 +27,7
On 6/7/21 12:23 PM, Fabian Grünbichler wrote:
On June 7, 2021 11:29 am, Stefan Reiter wrote:
On 6/4/21 11:47 AM, Fabian Grünbichler wrote:
we also need it for efidisks.
Signed-off-by: Fabian Grünbichler
---
PVE/QemuServer.pm | 8 ++--
PVE/QemuServer/Drive.pm | 10
High-level: If you make the $read_only_str contain the entire
",readonly=on" stanza, you can just leave it blank by default and avoid
updating all the test cases.
On 6/4/21 11:47 AM, Fabian Grünbichler wrote:
otherwise backups of templates using UEFI fail with storages like LVM
thin, where the
On 6/4/21 11:47 AM, Fabian Grünbichler wrote:
since using 'vma create ..' no longer works with immutable/read-only
base volumes.
Why that? It shouldn't matter to vma if the base is read-only? Though I
do see that in vma.c we don't tell QEMU that it should open the block
backend read-only...
On 6/4/21 11:47 AM, Fabian Grünbichler wrote:
we also need it for efidisks.
Signed-off-by: Fabian Grünbichler
---
PVE/QemuServer.pm | 8 ++--
PVE/QemuServer/Drive.pm | 10 ++
2 files changed, 12 insertions(+), 6 deletions(-)
diff --git a/PVE/QemuServer.pm b/PVE/QemuServe
For the three qemu-server changes:
Reviewed-by: Stefan Reiter
I'll also see if I can find some more stuff to cleanup in the code I
often traverse :)
On 6/4/21 3:49 PM, Fabian Ebner wrote:
This reverts commit ff09c795edd12b1cc4604ee28a7cdd99cdd1afa8. We wanted to wait
until PVE 7.0 fo
On 6/4/21 2:16 PM, Wolfgang Bumiller wrote:
On Wed, Jun 02, 2021 at 04:38:32PM +0200, Stefan Reiter wrote:
Provides a shared AsyncLruCache of 256MB (w/ 4MB chunks) that can be
used by multiple readers at the same time. It is dropped once no more
readers exist, so the memory gets freed if all
Provides a fast cache read implementation with full async and
concurrency support.
Signed-off-by: Stefan Reiter
---
This is technically all that's needed for proxmox-backup-qemu to build and
function as intended, but I decided to also use this IMHO cleaner implementation
to replac
read_image_at, providing real
concurrency without lock contention
Signed-off-by: Stefan Reiter
---
src/restore.rs | 30 --
1 file changed, 8 insertions(+), 22 deletions(-)
diff --git a/src/restore.rs b/src/restore.rs
index 0790d7f..33959d9 100644
--- a/src/restore.rs
+++ b
actual request concurrency is not supported in these traits).
Signed-off-by: Stefan Reiter
---
src/backup/cached_chunk_reader.rs | 116 +-
1 file changed, 114 insertions(+), 2 deletions(-)
diff --git a/src/backup/cached_chunk_reader.rs
b/src/backup
Setting this to 0 is not just useless, but breaks the logic horribly
enough to cause random segfaults - better forbid this, to avoid someone
else having to debug it again ;)
Signed-off-by: Stefan Reiter
---
src/tools/lru_cache.rs | 1 +
1 file changed, 1 insertion(+)
diff --git a/src/tools
superseded by CachedChunkReader, with less code and more speed
Signed-off-by: Stefan Reiter
---
src/backup.rs| 3 -
src/backup/async_index_reader.rs | 215 ---
2 files changed, 218 deletions(-)
delete mode 100644 src/backup
iteration
1.0.3-1A
patchedB
[0] see: https://lists.proxmox.com/pipermail/pbs-devel/2021-April/002932.html
proxmox-backup: Stefan Reiter (7):
tools/BroadcastFuture: add testcase for better understanding
tools: add AsyncLruCache as a wrapper around sync LruCache
backup: a
admin/datastore reads linearly only, so no need for cache (capacity of 1
basically means no cache except for the currently active chunk).
mount can do random access too, so cache last 8 chunks for possibly a
mild performance improvement.
Signed-off-by: Stefan Reiter
---
src/api2/admin
Provides a shared AsyncLruCache of 256MB (w/ 4MB chunks) that can be
used by multiple readers at the same time. It is dropped once no more
readers exist, so the memory gets freed if all QEMU block/pbs instances
disappear.
Signed-off-by: Stefan Reiter
---
src/lib.rs | 7 ++-
src
Supports concurrent 'access' calls to the same key via a
BroadcastFuture. These are stored in a seperate HashMap, the LruCache
underneath is only modified once a valid value has been retrieved.
Signed-off-by: Stefan Reiter
---
src/tools.rs | 1 +
src/tools/async_lr
Explicitly test that data will stay available and can be retrieved
immediately via listen(), even if the future producing the data and
notifying the consumers was already run in the past.
Signed-off-by: Stefan Reiter
---
Wasn't broken or anything, but helps with understanding IMO.
src/
QEMU 6.0 requires the updated version to build correctly, as the
keymap-gen tool gained some new parameters.
Signed-off-by: Stefan Reiter
---
keycodemapdb/data/keymaps.csv | 16 +-
keycodemapdb/meson.build | 1 +
keycodemapdb/tests/Makefile | 84 +++
keycodemapdb/tools/keymap
, but oh well):
https://lists.proxmox.com/pipermail/pve-devel/2021-May/048327.html
pve-qemu: Stefan Reiter (3):
udpate and rebase to QEMU v6.0.0
debian: ignore submodule checks in QEMU build
update keycodemapdb for 6.0
debian/changelog | 6 +
...d-support-for
...we do those manually, and the build dir is not a git repo.
Signed-off-by: Stefan Reiter
---
debian/rules | 1 +
1 file changed, 1 insertion(+)
diff --git a/debian/rules b/debian/rules
index 349c610..1f623d8 100755
--- a/debian/rules
+++ b/debian/rules
@@ -41,6 +41,7 @@ ${BUILDDIR
rm boolean option 'nowait' deprecated
Please use wait=off instead
kvm: -vnc unix:/var/run/qemu-server/100.vnc,password: warning: short-form
boolean option 'password' deprecated
Please use password=on instead
The new syntax is backwards compatible to at least QEMU 4.0.
Signed-o
1 - 100 of 349 matches
Mail list logo
