Re: Question about error conditions for ACPI usage on aarch64
On Mon, Dec 15, 2025 at 07:08:28AM +0000, Kazuhiro Abe (Fujitsu) wrote:
On Tue, Dec 02, 2025 at 12:52:02AM +0000, Kazuhiro Abe (Fujitsu) wrote:
You've just identified one of those issues that I predicted I would run into and have to fix :)
I already have a patch addressing that, among other things, in my local branch. I expect to be able to post the whole series shortly, probably next week.
It would be great if you can share the updates of your patch.
Still working on it :) It's taking a bit longer than I had anticipated to get all the ducks in a row, but I'm hoping that I will be able to finally post the series this week or the next one.
When can we expect this patch to be ready? We would appreciate it if you could provide us with this information as soon as possible.
I plan have patches on the list by the end of the week. -- Andrea Bolognani / Red Hat / Virtualization
On Fri, Jan 09, 2026 at 04:44:23AM +0000, Kazuhiro Abe (Fujitsu) wrote:
Did you ensure that a suitable JSON firmware descriptor exists in the expected location? It should normally come in the same package as the firmware binary itself, though it wouldn't be surprising if that's not currently the case considering that the combination was not working until now. If it doesn't exist, that will have to be addressed at the distro level, but you can create the file yourself in the meantime for development purposes.
See the attached patch showing how the XML you provided is processed correctly by the libvirt test suite once the JSON firmware descriptor is added.
Thank you so much for your detailed explanation and the attached patch. Following your advice, I've placed a suitable JSON firmware descriptor in /usr/share/qemu/firmware and confirmed that the error is now resolved.
Excellent news :) The patches have now been pushed and will be included in the 12.0.0 release, coming out in a week or so. -- Andrea Bolognani / Red Hat / Virtualization
Hi Andrea,
On Mon, Dec 15, 2025 at 07:08:28AM +0000, Kazuhiro Abe (Fujitsu) wrote:
On Tue, Dec 02, 2025 at 12:52:02AM +0000, Kazuhiro Abe (Fujitsu) wrote:
You've just identified one of those issues that I predicted I would run into and have to fix :)
I already have a patch addressing that, among other things, in my local branch. I expect to be able to post the whole series shortly, probably next week.
It would be great if you can share the updates of your patch.
Still working on it :) It's taking a bit longer than I had anticipated to get all the ducks in a row, but I'm hoping that I will be able to finally post the series this week or the next one.
When can we expect this patch to be ready? We would appreciate it if you could provide us with this information as soon as possible.
I plan have patches on the list by the end of the week.
Thank you for your reply. I look forward to it.
-- Andrea Bolognani / Red Hat / Virtualization
On Tue, Dec 16, 2025 at 08:00:33AM +0000, Kazuhiro Abe (Fujitsu) wrote:
When can we expect this patch to be ready? We would appreciate it if you could provide us with this information as soon as possible.
I plan have patches on the list by the end of the week.
Thank you for your reply. I look forward to it.
That took longer than I anticipated, even though I had already exceeded my original estimate O:-) however patches are finally on the list. https://lists.libvirt.org/archives/list/devel@lists.libvirt.org/thread/N2ETT... I'd appreciate any testing you could give them, as well as any other kind of feedback you might have. Thanks once again for your patience. -- Andrea Bolognani / Red Hat / Virtualization
Hi Andrea,
On Tue, Dec 16, 2025 at 08:00:33AM +0000, Kazuhiro Abe (Fujitsu) wrote:
When can we expect this patch to be ready? We would appreciate it if you could provide us with this information as soon as possible.
I plan have patches on the list by the end of the week.
Thank you for your reply. I look forward to it.
That took longer than I anticipated, even though I had already exceeded my original estimate O:-) however patches are finally on the list.
https://lists.libvirt.org/archives/list/devel@lists.libvirt.org/thread/N2ETT... I5RWXGJG7DW5YYMZ7UGDYMHA/
I'd appreciate any testing you could give them, as well as any other kind of feedback you might have.
Thank you for posting the patches. I applied the patch and tested it on my side. However, it seems the issue my previous patch(linked below) aimed to resolve is still present. https://lists.libvirt.org/archives/list/devel@lists.libvirt.org/thread/AII3X... I expect the following XML not to result in an error. <os> <type arch='aarch64' machine='virt-4.0'>hvm</type> <loader type="rom">/usr/share/qemu-efi-aarch64/QEMU_EFI.fd</loader> </os> <features> <acpi/> </features> However, I am still seeing the same error as before. The error message observed when running "virsh create test.xml" is: error: Failed to create domain from test.xml error: unsupported configuration: ACPI requires UEFI on this architecture Could you please let me know whether this issue has already been resolved? Best regards, Kazuhiro Abe
Thanks once again for your patience.
-- Andrea Bolognani / Red Hat / Virtualization
On Wed, Jan 07, 2026 at 12:24:18AM +0000, Kazuhiro Abe (Fujitsu) wrote:
That took longer than I anticipated, even though I had already exceeded my original estimate O:-) however patches are finally on the list.
https://lists.libvirt.org/archives/list/devel@lists.libvirt.org/thread/N2ETT... I5RWXGJG7DW5YYMZ7UGDYMHA/
I'd appreciate any testing you could give them, as well as any other kind of feedback you might have.
Thank you for posting the patches.
I applied the patch and tested it on my side. However, it seems the issue my previous patch(linked below) aimed to resolve is still present. https://lists.libvirt.org/archives/list/devel@lists.libvirt.org/thread/AII3X...
I expect the following XML not to result in an error. <os> <type arch='aarch64' machine='virt-4.0'>hvm</type> <loader type="rom">/usr/share/qemu-efi-aarch64/QEMU_EFI.fd</loader> </os> <features> <acpi/> </features>
However, I am still seeing the same error as before. The error message observed when running "virsh create test.xml" is: error: Failed to create domain from test.xml error: unsupported configuration: ACPI requires UEFI on this architecture
Did you ensure that a suitable JSON firmware descriptor exists in the expected location? It should normally come in the same package as the firmware binary itself, though it wouldn't be surprising if that's not currently the case considering that the combination was not working until now. If it doesn't exist, that will have to be addressed at the distro level, but you can create the file yourself in the meantime for development purposes. See the attached patch showing how the XML you provided is processed correctly by the libvirt test suite once the JSON firmware descriptor is added. -- Andrea Bolognani / Red Hat / Virtualization
Hi Andrea,
That took longer than I anticipated, even though I had already exceeded my original estimate O:-) however patches are finally on the list.
https://lists.libvirt.org/archives/list/devel@lists.libvirt.org/thre ad/N2ETTZ3W I5RWXGJG7DW5YYMZ7UGDYMHA/
I'd appreciate any testing you could give them, as well as any other kind of feedback you might have.
Thank you for posting the patches.
I applied the patch and tested it on my side. However, it seems the issue my previous patch(linked below) aimed to resolve is still present. https://lists.libvirt.org/archives/list/devel@lists.libvirt.org/thread /AII3X37F7AK7AR234LYWNCS7BIR5AL2P/
I expect the following XML not to result in an error. <os> <type arch='aarch64' machine='virt-4.0'>hvm</type> <loader type="rom">/usr/share/qemu-efi-aarch64/QEMU_EFI.fd</loader> </os> <features> <acpi/> </features>
However, I am still seeing the same error as before. The error message observed when running "virsh create test.xml" is: error: Failed to create domain from test.xml error: unsupported configuration: ACPI requires UEFI on this architecture
Did you ensure that a suitable JSON firmware descriptor exists in the expected location? It should normally come in the same package as the firmware binary itself, though it wouldn't be surprising if that's not currently the case considering that the combination was not working until now. If it doesn't exist, that will have to be addressed at the distro level, but you can create the file yourself in the meantime for development purposes.
See the attached patch showing how the XML you provided is processed correctly by the libvirt test suite once the JSON firmware descriptor is added.
Thank you so much for your detailed explanation and the attached patch. Following your advice, I've placed a suitable JSON firmware descriptor in /usr/share/qemu/firmware and confirmed that the error is now resolved. Best Regards, Kazuhiro Abe
-- Andrea Bolognani / Red Hat / Virtualization
participants (2)
-
Andrea Bolognani -
Kazuhiro Abe (Fujitsu)