Re: E820 memory allocation issue on Threadripper platforms

[Jan Beulich <jbeulich@xxxxxxxx>]

On 18.01.2024 07:23, Patrick Plenefisch wrote:
> On Wed, Jan 17, 2024 at 3:46 AM Jan Beulich <jbeulich@xxxxxxxx> wrote:
>> On 17.01.2024 07:12, Patrick Plenefisch wrote:
>>> I'm currently talking to the vendor's support team and testing a beta
>> BIOS
>>> for unrelated reasons, is there something specific I should forward to
>>> them, either as a question or as a request for a fix?
>>
>> Well, first it would need figuring whether the "interesting" regions
>> are being put in place by firmware of the boot loader. If it's firmware
>> (pretty likely at least for the region you're having trouble with), you
>> may want to ask them to re-do where they place that specific data.
> 
> This section changes boot-to-boot and grub vs EFI direct load, but my
> untrained eyes don't see an obvoius pattern. I've attached several logs.
> Name format:
> 
> xen-XENVERSION_LOADER_KERNELNAME_TYPE.log
> 
> where XENVERSION is 4.17 (packaged in debian 12) or 4.18 (I built from
> source) or 4.18p (I applied the patch you mention below and built from
> source)
> 
> where LOADER is grub for grub2 (from debian 12) or UEFI (direct boot via
> efibootmgr-configured UEFI entry)
> 
> where KERNELNAME is either empty (PVH failure), or linuxpatch (linux with
> the patch requested above), or linuxoffset (with PHYSICAL_START=2MiB), or
> linux6 (debian 12 kernel)
> 
> where TYPE is either pvh or pv
> 
> For the two logs that actually boooted (linuxoffset), I truncated them
> during pcie initialization, but they did go all the way to give me a login
> screen

The LOADER=UEFI logs confirm it's firmware (in the widest sense, as it could
also be a UEFI driver) which puts in place these unhelpful regions.

Jan