Note: This bug is displayed in read-only format because the product is no longer active in Red Hat Bugzilla.
RHEL Engineering is moving the tracking of its product development work on RHEL 6 through RHEL 9 to Red Hat Jira (issues.redhat.com). If you're a Red Hat customer, please continue to file support cases via the Red Hat customer portal. If you're not, please head to the "RHEL project" in Red Hat Jira and file new tickets here. Individual Bugzilla bugs in the statuses "NEW", "ASSIGNED", and "POST" are being migrated throughout September 2023. Bugs of Red Hat partners with an assigned Engineering Partner Manager (EPM) are migrated in late September as per pre-agreed dates. Bugs against components "kernel", "kernel-rt", and "kpatch" are only migrated if still in "NEW" or "ASSIGNED". If you cannot log in to RH Jira, please consult article #7032570. That failing, please send an e-mail to the RH Jira admins at rh-issues@redhat.com to troubleshoot your issue as a user management inquiry. The email creates a ServiceNow ticket with Red Hat. Individual Bugzilla bugs that are migrated will be moved to status "CLOSED", resolution "MIGRATED", and set with "MigratedToJIRA" in "Keywords". The link to the successor Jira issue will be found under "Links", have a little "two-footprint" icon next to it, and direct you to the "RHEL project" in Red Hat Jira (issue links are of type "https://issues.redhat.com/browse/RHEL-XXXX", where "X" is a digit). This same link will be available in a blue banner at the top of the page informing you that that bug has been migrated.

Bug 2196178

Summary: libvirt: Changes to firmware selection
Product: Red Hat Enterprise Linux 9 Reporter: Andrea Bolognani <abologna>
Component: libvirtAssignee: Andrea Bolognani <abologna>
libvirt sub component: General QA Contact: Meina Li <meili>
Status: CLOSED ERRATA Docs Contact:
Severity: unspecified    
Priority: unspecified CC: dzheng, jdenemar, jsuchane, jwboyer, lmen, shdunne, virt-maint
Version: 9.1Keywords: AutomationTriaged, Triaged
Target Milestone: rcFlags: pm-rhel: mirror+
Target Release: ---   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: libvirt-9.5.0-6.el9 Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2023-11-07 08:31:17 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version: 9.7.0
Embargoed:
Attachments:
Description Flags
libvirtd.log none

Description Andrea Bolognani 2023-05-08 09:02:24 UTC 
As part of adding support for QCOW2 firmware images (Bug 2161965) I
have had to make some changes to the way libvirt performs firmware
selection.

Specifically:

  * firmware selection now happens only once, when the VM is defined,
    instead of each time it is started. This improves guest ABI
    stability;

  * firmware features (secure-boot, enrolled-keys) discovered via
    looking at JSON descriptors are now reflected in the XML.

While QCOW2 firmware images are only going to be used on aarch64, the
changes described above affect x86_64 as well.

All of this is included in libvirt 9.2.0.
Comment 2 Meina Li 2023-05-09 03:19:51 UTC 
Created attachment 1963449 [details]
libvirtd.log
Comment 3 Meina Li 2023-05-09 03:21:29 UTC 
Hi Andrea,

I've tested this bug and many scenarios tests passed. But there are still three test scenarios that I suspect are bugs. Or that's how it was designed. Could you help to review this scenarios again? Thank you very much.

Test Version:
libvirt-9.3.0-1.el9.x86_64
qemu-kvm-8.0.0-1.el9.x86_64

S1: Define a guest only with loader and file backed nvram template
Test Steps:
1. Prepare a guest xml with the following os xml:
<os>
  <type arch="x86_64" machine="pc-q35-rhel9.2.0">hvm</type>
  <loader readonly="yes" type="pflash">/usr/share/OVMF/OVMF_CODE.secboot.fd</loader>
  <nvram template="/usr/share/OVMF/OVMF_VARS.secboot.fd" type="file">
    <source file="/tmp/rhel_VARS.fd"/>
  </nvram>
  <boot dev="hd"/>
</os>
2. Define the guest.
# virsh define rhel.xml 
Domain 'rhel' defined from rhel.xml
3. Check the dumpxml.
# virsh dumpxml rhel --xpath //os
<os>
  <type arch="x86_64" machine="pc-q35-rhel9.2.0">hvm</type>
  <loader readonly="yes" secure="yes" type="pflash">/usr/share/OVMF/OVMF_CODE.secboot.fd</loader>
  <nvram template="/usr/share/OVMF/OVMF_VARS.fd" type="file">
    <source file="/tmp/rhel_VARS.fd"/>
  </nvram>
  <boot dev="hd"/>
</os>
------>> We can't get the firmware features automatically. And the nvram template has been changed from OVMF_CODE.secboot.fd to OVMF_VARS.fd.

S2: Define a guest with efi firmware and file backed nvram template. 
1. Prepare a guest xml with the following os xml:
<os firmware='efi'>
  <type arch="x86_64" machine="pc-q35-rhel9.2.0">hvm</type>
  <nvram template="/usr/share/OVMF/OVMF_VARS.secboot.fd" type="file">
  <source file="/tmp/rhel_VARS.fd"/>
  </nvram>
  <boot dev="hd"/>
</os>
2. Define the guest.
# virsh define rhel.xml 
error: Disconnected from qemu:///system due to end of file
error: Failed to define domain from rhel.xml
error: End of file while reading data: Input/output error
------>> Get unexpected error with no coredump and no error info in libvirtd.log(look at attachment.)

S3: Define a guest with readonly='no'.
1. Prepare a guest xml with readonly='no' in os loader.
<os>
  <type arch="x86_64" machine="pc-q35-rhel9.2.0">hvm</type>
  <loader readonly="no" secure="yes" type="pflash">/usr/share/edk2/ovmf/OVMF_CODE.secboot.fd</loader>
  <boot dev="hd"/>
</os>
2. Define the guest.
# virsh define rhel.xml 
Domain 'rhel' defined from rhel.xml
3. Check the dumpxml.
# virsh dumpxml rhel --xpath //os
<os firmware="efi">
  <type arch="x86_64" machine="pc-q35-rhel9.2.0">hvm</type>
  <firmware>
    <feature enabled="yes" name="enrolled-keys"/>
    <feature enabled="yes" name="secure-boot"/>
  </firmware>
  <loader readonly="yes" secure="yes" type="pflash">/usr/share/edk2/ovmf/OVMF_CODE.secboot.fd</loader>
  <nvram template="/usr/share/edk2/ovmf/OVMF_VARS.secboot.fd">/var/lib/libvirt/qemu/nvram/rhel_VARS.fd</nvram>
  <boot dev="hd"/>
</os>
------>> The value of readonly is changed to 'yes', is it expected?
Comment 4 Andrea Bolognani 2023-05-09 09:09:34 UTC 
(In reply to Meina Li from comment #3)
> Hi Andrea,
> 
> I've tested this bug and many scenarios tests passed. But there are still
> three test scenarios that I suspect are bugs. Or that's how it was designed.
> Could you help to review this scenarios again? Thank you very much.

Thanks for the report! These all seem unexpected. I'll look into
them and get back to you.
Comment 5 Andrea Bolognani 2023-05-11 13:32:26 UTC 
I have manage to reproduce all the scenarios that you've so very
nicely and accurately described, and yeah they're all undesired.

I'll start working on patches. In the meantime, moving the bug back
to ASSIGNED.
Comment 7 Meina Li 2023-05-24 02:26:12 UTC 
Another unreasonable test scenario.

Test Version:
libvirt-9.3.0-2.el9.x86_64
qemu-kvm-8.0.0-3.el9.x86_64

Test Steps:
1. Prepare a guest xml with /usr/share/edk2/ovmf/OVMF_CODE.secboot.fd loader and noexist template.
  <os>
    <type arch='x86_64' machine='pc-q35-rhel9.2.0'>hvm</type>
    <loader readonly='yes' secure='yes' type='pflash'>/usr/share/edk2/ovmf/OVMF_CODE.secboot.fd</loader>
    <nvram template='noexist'>/var/lib/libvirt/qemu/nvram/test_VARS.fd</nvram>
    <boot dev='hd'/>
  </os>
2. Define the guest.
# virsh define test.xml 
Domain 'test' defined from test.xml
# virsh dumpxml test --xpath //os
<os>
  <type arch="x86_64" machine="pc-q35-rhel9.2.0">hvm</type>
  <loader readonly="yes" secure="yes" type="pflash">/usr/share/edk2/ovmf/OVMF_CODE.secboot.fd</loader>
  <nvram template="noexist">/var/lib/libvirt/qemu/nvram/test_VARS.fd</nvram>
  <boot dev="hd"/>
</os>
3. Start the guest.
# virsh start test
error: Failed to start domain 'test'
error: Failed to open file 'noexist': No such file or directory
------> I think this is expected.

But if we use /usr/share/OVMF/OVMF_CODE.secboot.fd as the loader path. It will have different result, which is strange.
1. Prepare a guest xml with /usr/share/OVMF/OVMF_CODE.secboot.fd loader and noexist template.
  <os>
    <type arch='x86_64' machine='pc-q35-rhel9.2.0'>hvm</type>
    <loader readonly='yes' secure='yes' type='pflash'>/usr/share/OVMF/OVMF_CODE.secboot.fd</loader>
    <nvram template='noexist'>/var/lib/libvirt/qemu/nvram/test_VARS.fd</nvram>
    <boot dev='hd'/>
  </os>
2. Define the guest.
# virsh define test.xml 
Domain 'test' defined from test.xml
# virsh dumpxml test --xpath //os
<os>
  <type arch="x86_64" machine="pc-q35-rhel9.2.0">hvm</type>
  <loader readonly="yes" secure="yes" type="pflash">/usr/share/OVMF/OVMF_CODE.secboot.fd</loader>
  <nvram template="/usr/share/OVMF/OVMF_VARS.fd">/var/lib/libvirt/qemu/nvram/test_VARS.fd</nvram>
  <boot dev="hd"/>
</os>

------> After we defining the guest, the template will be changed and then certainly the guest start succssfully.

The relationship between /usr/share/OVMF/OVMF_CODE.secboot.fd and /usr/share/edk2/ovmf/OVMF_CODE.secboot.fd:
# ll /usr/share/OVMF/OVMF_CODE.secboot.fd
lrwxrwxrwx. 1 root root 33 May 22 04:38 /usr/share/OVMF/OVMF_CODE.secboot.fd -> ../edk2/ovmf/OVMF_CODE.secboot.fd
Comment 8 Andrea Bolognani 2023-05-25 13:55:47 UTC 
(In reply to Meina Li from comment #7)
> ------> I think this is expected.
> 
> But if we use /usr/share/OVMF/OVMF_CODE.secboot.fd as the loader path. It
> will have different result, which is strange.

Thanks for the additional testing! I'm still working on patches that
will hopefully improve the situation, and I will keep these scenarios
in mind too.
Comment 9 Andrea Bolognani 2023-08-10 16:53:33 UTC 
Patches posted upstream.

https://listman.redhat.com/archives/libvir-list/2023-August/241171.html
Comment 10 Andrea Bolognani 2023-08-10 17:15:01 UTC 
A few clarifications.

(In reply to Meina Li from comment #3)
> S1: Define a guest only with loader and file backed nvram template
> Test Steps:
> 1. Prepare a guest xml with the following os xml:
> <os>
>   <type arch="x86_64" machine="pc-q35-rhel9.2.0">hvm</type>
>   <loader readonly="yes" type="pflash">/usr/share/OVMF/OVMF_CODE.secboot.fd</loader>
>   <nvram template="/usr/share/OVMF/OVMF_VARS.secboot.fd" type="file">
>     <source file="/tmp/rhel_VARS.fd"/>
>   </nvram>
>   <boot dev="hd"/>
> </os>
> 2. Define the guest.
> # virsh define rhel.xml 
> Domain 'rhel' defined from rhel.xml
> 3. Check the dumpxml.
> # virsh dumpxml rhel --xpath //os
> <os>
>   <type arch="x86_64" machine="pc-q35-rhel9.2.0">hvm</type>
>   <loader readonly="yes" secure="yes" type="pflash">/usr/share/OVMF/OVMF_CODE.secboot.fd</loader>
>   <nvram template="/usr/share/OVMF/OVMF_VARS.fd" type="file">
>     <source file="/tmp/rhel_VARS.fd"/>
>   </nvram>
>   <boot dev="hd"/>
> </os>
> ------>> We can't get the firmware features automatically. And the nvram
> template has been changed from OVMF_CODE.secboot.fd to OVMF_VARS.fd.

The fact that the NVRAM template has been changed from
OVMF_CODE.secboot.fd to OVMF_VARS.fd is definitely undesired and
shouldn't happen.

The fact that firmware features are not automatically discovered,
however, is expected: that process relies on a corresponding JSON
firmware descriptor file existing on the system, and as of RHEL 9
those files only include references to modern paths such as
/usr/share/edk2/ovmf/OVMF_CODE.secboot.fd, not to legacy paths such
as /usr/share/OVMF/OVMF_CODE.secboot.fd. So using the legacy paths
will result in firmware features not being automatically discovered
and added to the XML.

> S3: Define a guest with readonly='no'.
> 1. Prepare a guest xml with readonly='no' in os loader.
> <os>
>   <type arch="x86_64" machine="pc-q35-rhel9.2.0">hvm</type>
>   <loader readonly="no" secure="yes" type="pflash">/usr/share/edk2/ovmf/OVMF_CODE.secboot.fd</loader>
>   <boot dev="hd"/>
> </os>
> 2. Define the guest.
> # virsh define rhel.xml
> Domain 'rhel' defined from rhel.xml
> 3. Check the dumpxml.
> # virsh dumpxml rhel --xpath //os
> <os firmware="efi">
>   <type arch="x86_64" machine="pc-q35-rhel9.2.0">hvm</type>
>   <firmware>
>     <feature enabled="yes" name="enrolled-keys"/>
>     <feature enabled="yes" name="secure-boot"/>
>   </firmware>
>   <loader readonly="yes" secure="yes" type="pflash">/usr/share/edk2/ovmf/OVMF_CODE.secboot.fd</loader>
>   <nvram> template="/usr/share/edk2/ovmf/OVMF_VARS.secboot.fd">/var/lib/libvirt/qemu/nvram/rhel_VARS.fd</nvram>
>   <boot dev="hd"/>
> </os>
> ------>> The value of readonly is changed to 'yes', is it expected?

loader.readonly=no should be used for combined firmware images, that
is, those where both the CODE and VARS parts are included in the same
file. That's something that's not really used anymore, if it ever has
been, because it prevents sharing the CODE part among different VMs.

Even if you wanted to use it for some reason, you'd have to bring
your own combined build - RHEL doesn't come with one. So the
configuration you're testing is fundamentally incorrect.

That said, the fact that libvirt will happily override some
user-provided value is still a problem.

(In reply to Meina Li from comment #7)
> But if we use /usr/share/OVMF/OVMF_CODE.secboot.fd as the loader path. It
> will have different result, which is strange.
>
[...]
>
> The relationship between /usr/share/OVMF/OVMF_CODE.secboot.fd and
> /usr/share/edk2/ovmf/OVMF_CODE.secboot.fd:
> # ll /usr/share/OVMF/OVMF_CODE.secboot.fd
> lrwxrwxrwx. 1 root root 33 May 22 04:38 /usr/share/OVMF/OVMF_CODE.secboot.fd > -> ../edk2/ovmf/OVMF_CODE.secboot.fd

See above. Even though the two paths refer to the same file, as far
as libvirt is concerned they are completely different. Specifically,
the modern path (/usr/share/edk2/ovmf/OVMF_CODE.secboot.fd) is
mentioned in a JSON firmware descriptor file, while the legacy one
(/usr/share/OVMF/OVMF_CODE.secboot.fd) is not and is instead
included in the DEFAULT_LOADER_NVRAM list, which acts as the default
value for the "nvram" key in qemu.conf if that hasn't been provided
by the admin.
Comment 13 Andrea Bolognani 2023-08-21 12:38:54 UTC 
Fixes pushed upstream.

  07b6189ef4 NEWS: Mention fixes to firmware selection
  7c328b6cf4 tests: Reintroduce firmware-auto-efi-format-mismatch
  48e5fe7af4 tests: Rename firmware-auto-efi-format-loader-qcow2-nvram-path
  10a8997cbb conf: Don't default to raw format for loader/NVRAM
  b845e376a4 qemu: Match NVRAM template extension for new domains
  e96e322725 qemu: Filter firmware based on loader.readonly
  ccbb987707 qemu: Generate NVRAM path in more cases
  4a49114ff4 qemu: Don't overwrite NVRAM template for legacy firmware
  1b3e9c67e3 tests: Include microvm in firmwaretest
  da6b98394b tests: Drop tags from BIOS firmware descriptor
  a97c56888c tests: Update firmware descriptor files
  e930f62a02 tests: Add more tests for firmware selection
  87d91e9e24 tests: Add some more DO_TEST*ABI_UPDATE* macros
  ac76386eda qemu: Fix lookup against stateless/combined pflash
  d917883b30 qemu: Fix return value for qemuFirmwareFillDomainLegacy()
  4ba04107d9 tests: Rename firmware-auto-efi-nvram-path
  8627ec167c tests: Turn abi-update.xml into a symlink
  1773526224 tests: Consistently use /path/to/guest_VARS.fd
  5c129c8e7a tests: Use virt-4.0 machine type for aarch64
  8c326914d8 tests: Switch to firmware autoselection for hvf
  751b0e6dbf tests: Use DO_TEST_CAPS_*_ABI_UPDATE() for ppc64

  v9.6.0-71-g07b6189ef4
Comment 14 Andrea Bolognani 2023-08-21 14:14:54 UTC 
CentOS Stream 9 backport prepared.

  https://gitlab.com/redhat/centos-stream/src/libvirt/-/merge_requests/7
Comment 18 Shelley Dunne 2023-08-24 20:54:05 UTC 
This was reviewed by the RHEL voting members and approved.
Comment 19 Meina Li 2023-08-30 07:03:47 UTC 
Pre-verified Version:
libvirt-9.5.0-6.el9.x86_64
qemu-kvm-8.0.0-13.el9.x86_64
edk2-ovmf-20230524-3.el9.noarch

Pre-verified Steps:
S1: Define and start a guest with readonly='no'.
1. Prepare a guest xml with readonly='no' in os loader.
<os>
  <type arch="x86_64" machine="pc-q35-rhel9.2.0">hvm</type>
  <loader readonly="no" secure="yes" type="pflash">/usr/share/edk2/ovmf/OVMF_CODE.secboot.fd</loader>
  <boot dev="hd"/>
</os>
2. Define the guest and check the dumpxml.
# virsh define rhel.xml 
Domain 'rhel' defined from rhel.xml
# virsh dumpxml rhel --xpath //os
<os>
  <type arch="x86_64" machine="pc-q35-rhel9.2.0">hvm</type>
  <loader readonly="no" type="pflash">/usr/share/edk2/ovmf/OVMF_CODE.secboot.fd</loader>
  <boot dev="hd"/>
</os>
3. Start the guest.
# virsh start rhel
error: Failed to start domain 'rhel'
error: internal error: QEMU unexpectedly closed the monitor (vm='rhel'): 2023-08-30T06:18:52.502400Z qemu-kvm: Could not open '/usr/share/edk2/ovmf/OVMF_CODE.secboot.fd': Permission denied
4. Repeat test with the legacy loader path, got the same expected error message: qemu-kvm: Could not open '/usr/share/edk2/ovmf/OVMF_CODE.secboot.fd': Permission denied.
5. Repeat test only with readonly='no' and efi firmware, got expected error:
error: operation failed: Unable to find any firmware to satisfy 'efi'.

S2: Define and start guest with nonexisted template.
1. Prepare a guest xml with legacy loader path and nonexisted template.
<os>
  <type arch="x86_64" machine="pc-q35-rhel9.2.0">hvm</type>
  <loader readonly="yes" secure="yes" type="pflash">/usr/share/OVMF/OVMF_CODE.secboot.fd</loader>
  <nvram template="noexist">/var/lib/libvirt/qemu/nvram/test_VARS.fd</nvram>
  <boot dev="hd"/>
</os>
2. Define the guest and check dumpxml.
# virsh define rhel.xml 
Domain 'rhel' defined from rhel.xml
# virsh dumpxml rhel --xpath //os
<os>
  <type arch="x86_64" machine="pc-q35-rhel9.2.0">hvm</type>
  <loader readonly="yes" secure="yes" type="pflash">/usr/share/OVMF/OVMF_CODE.secboot.fd</loader>
  <nvram template="noexist">/var/lib/libvirt/qemu/nvram/test_VARS.fd</nvram>
  <boot dev="hd"/>
</os>
3. Start the guest.
# virsh start rhel
error: Failed to start domain 'rhel'
error: Failed to open file 'noexist': No such file or directory
4. Repeat test with modern loader path and nonexisted template, can get same expected error.

S3: Start a guest only with loader and file backed nvram template.
1. Prepare a guest xml with file backend nvram template.
<os>
  <type arch="x86_64" machine="pc-q35-rhel9.2.0">hvm</type>
    <loader readonly='yes' secure='yes' type='pflash'>/usr/share/edk2/ovmf/OVMF_CODE.secboot.fd</loader>
    <nvram template='/usr/share/edk2/ovmf/OVMF_VARS.secboot.fd' type='file'>
      <source file='/tmp/rhel_VARS.fd'/>
    </nvram>
  <boot dev="hd"/>
</os>
2. Define and start the guest.
# virsh define rhel.xml 
Domain 'rhel' defined from rhel.xml
# virsh start rhel
Domain 'rhel' started
# virsh dumpxml rhel --xpath //os
<os firmware="efi">
  <type arch="x86_64" machine="pc-q35-rhel9.2.0">hvm</type>
  <firmware>
    <feature enabled="yes" name="enrolled-keys"/>
    <feature enabled="yes" name="secure-boot"/>
  </firmware>
  <loader readonly="yes" secure="yes" type="pflash">/usr/share/edk2/ovmf/OVMF_CODE.secboot.fd</loader>
  <nvram template="/usr/share/edk2/ovmf/OVMF_VARS.secboot.fd" type="file">
    <source file="/tmp/rhel_VARS.fd"/>
  </nvram>
  <boot dev="hd"/>
</os>

S4: Start a guest with efi firmware and file backed nvram template.
1. Prepare a guest xml with efi firmware and file backed nvram template.
  <os firmware='efi'>
    <type arch='x86_64' machine='pc-q35-rhel9.2.0'>hvm</type>
    <nvram template="/usr/share/edk2/ovmf/OVMF_VARS.secboot.fd" type="file">
    <source file="/tmp/rhel_VARS.fd"/>
    </nvram>
    <boot dev='hd'/>
  </os>
2. Define and start the guest.
# virsh define rhel.xml 
Domain 'rhel' defined from rhel.xml
# virsh start rhel
Domain 'rhel' started
# virsh dumpxml rhel --xpath //os
<os firmware="efi">
  <type arch="x86_64" machine="pc-q35-rhel9.2.0">hvm</type>
  <firmware>
    <feature enabled="yes" name="enrolled-keys"/>
    <feature enabled="yes" name="secure-boot"/>
  </firmware>
  <loader readonly="yes" secure="yes" type="pflash">/usr/share/edk2/ovmf/OVMF_CODE.secboot.fd</loader>
  <nvram template="/usr/share/edk2/ovmf/OVMF_VARS.secboot.fd" type="file">
    <source file="/tmp/rhel_VARS.fd"/>
  </nvram>
  <boot dev="hd"/>
</os>
Comment 20 Meina Li 2023-08-30 07:34:37 UTC 
Hi Andrea,

I found there are some patches in comment 13 related to the Fedora 38's edk2 package. For example,  the 4MB qcow2 format builds have been introduced on x86_64 in those new firmware descriptor files. Now we can't verify them. And it will block the verification of this bug. 

So could we open another bug to track those patches until the new edk2 package comes?

Thanks.
Comment 21 Andrea Bolognani 2023-08-30 08:57:41 UTC 
(In reply to Meina Li from comment #20)
> Hi Andrea,
> 
> I found there are some patches in comment 13 related to the Fedora 38's edk2
> package. For example,  the 4MB qcow2 format builds have been introduced on
> x86_64 in those new firmware descriptor files. Now we can't verify them. And
> it will block the verification of this bug. 
> 
> So could we open another bug to track those patches until the new edk2
> package comes?

There are no plans to switch x86_64 builds to qcow2 in RHEL.

We did the switch in Fedora because that gave us the chance to also
switch from 2M builds to 4M builds at the same time, which was long
overdue and we just didn't have a good mechanism for until now.

In RHEL x86_64 builds are already 4M, and unlike aarch64 there are no
benefits in terms of memory usage to using qcow2, so switching would
be pointless.

So, for any scenario in which upstream/Fedora uses qcow2 on x86_64,
RHEL should get raw instead. And for those in which qcow2 is
explicitly requested on x86_64, RHEL should report a failure.

Hope this helps. Please let me know if you need more information!
Comment 22 Meina Li 2023-08-30 09:14:19 UTC 
Sorry for the misunderstanding as I thought the qcow2 firmware will also be supported on rhel x86_64. So based on comment 19 and comment 21, move this bug to Verified: Tested.
Comment 23 Andrea Bolognani 2023-08-30 09:43:31 UTC 
(In reply to Meina Li from comment #22)
> Sorry for the misunderstanding as I thought the qcow2 firmware will
> also be supported on rhel x86_64.

No worries, it's a perfectly reasonable conclusion to reach based on
looking at the libvirt changes.

> So based on comment 19 and comment 21, move this bug to
> Verified: Tested.

Thanks :)
Comment 26 Meina Li 2023-09-06 08:11:21 UTC 
Verification test passed with the test scenarios in comment 19.
Comment 30 errata-xmlrpc 2023-11-07 08:31:17 UTC 
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory (Moderate: libvirt security, bug fix, and enhancement update), and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://access.redhat.com/errata/RHSA-2023:6409