[pve-devel] [PATCH edk2-firmware v2] update to edk2-stable202308

[Fiona Ebner]

Am 31.08.23 um 14:47 schrieb Stefan Hanreich:
> On 8/31/23 14:39, Thomas Lamprecht wrote:
>> Am 30/08/2023 um 17:50 schrieb Stefan Hanreich:
>>> * Removed 0001-OvmfPkg-PlatformInitLib-limit-phys-bits-to-46.patch
>>>    since it has been included upstream since commit c1e85376 [1].
>>>
>>> * Updated the patch
>>>    Revert-ArmVirtPkg-make-EFI_LOADER_DATA-non-executabl so it applies
>>>    again.
>>>
>>> * had to increase FD_SIZE to 4M for the x64 build as otherwise the
>>>    package wouldn't build due to the size of the resulting image
>>
>> I mean new versions are great and what not, but is there a reason why you
>> send this update now? Maybe fix some issue our users face? If so, it'd be
>> great if you could record that here.
>>
> 
> I've been trying to reproduce an issue one of our users is facing and
> wanted to check whether newer firmware versions trigger the issue as
> well. Currently, this issue seems fixed with the newer firmware although
> it's hard to say since there is no definite reproducer.

To be a bit more precise, it's an issue were a RHEL VM with 4.x kernel
gets stuck after guest reboot, as far as we can tell right after OVMF
initialization. Mira and Stefan were able to reproduce the issue
semi-regularly with a bunch of VMs rebooting periodically. And Stefan
said it didn't yet trigger with the updated edk2 package for a few days
now. We can't be sure, as it is a race after all, but it does sound
promising.