Re: Acer Aspire 5 A517-52G - No firmware ACIP and no boot

[Bruno <brunoce@xxxxxxxxxxxx>]

Jan, I am glad we see this now!

Should I report this on the to the linux  kernel dev group?

Thank you!

> On 3 Aug 2022, at 07:34, Jan Beulich <jbeulich@xxxxxxxx> wrote:
> 
> On 03.08.2022 08:25, Jan Beulich wrote:
>>> On 02.08.2022 18:07, Bruno wrote:
>>> On initramfs:
>>> 
>>> - xl will give me not found - I cannot do xl info or xl dmesg
>> 
>> As expected - you'd have to handcraft xl and ints libraries into the
>> image.
>> 
>>> - I cannot mount any disk, therefore, exporting logs from there has been 
>>> photographing and using OCR then reviewing and correcting - but I guess 
>>> the main difference is ACIP, as I cannot access the disks
>> 
>> Oh, I see. I was kind of expecting (hoping) you might be able to
>> export the logs via, say, a USB key.
>> 
>>> I have now booted on kernel 18 to get the kernel boot log for you (the 
>>> one on Xen-user is kernel 10).
>> 
>> At the first glance - similar ACPI errors, and a similar issue with
>> loading sound firmware. Using two different kernels won't really
>> lend itself for direct comparison, though. At the moment I don't see
>> a good way forward here as long as not even proper logs can be
>> obtained: It may be necessary to add debugging code, the output of
>> which would also need collecting. But maybe someone else sees a
>> viable route to help you ...
> 
> Oh, actually upon looking again I spotted something:
> 
> nvme nvme0: pci function 10000:e1:00.0
> 
> Note the wider-than-16-bits segment number. Which, afaict, is an effect of
> using the Volume Management Device:
> 
> vmd 0000:00:0e.0: PCI host bridge to bus 10000:e0
> 
> You may want to go through firmware setup options to see whether there's a
> way to disable it / its use. Neither Xen nor the Dom0 Linux kernel (i.e.
> when run under Xen) have been enabled so far to make this work (and, from
> inspection quite a while back, the kernel implementation, during its design,
> not having considered the possibility of running on a type-1 hypervisor will
> likely make adding support, well, interesting).
> 
> Jan