Re: Bug#927892: grub-ieee1275-bin: Please add bootinfo.txt for sparc and sparc64
On 28/04/2019 09:49, John Paul Adrian Glaubitz wrote:
> On 4/28/19 10:33 AM, Mark Cave-Ayland wrote:
>>>> https://cdimage.debian.org/cdimage/ports/grub-test/
>>
>> I grabbed the test image from the above URL to test under qemu-system-sparc64 and
>> initially got the error below:
>
> Please note: This image is not particularly tested with regards the installer itself,
> so please let's not start a discussion regarding other issues. The images are solely
> for testing the GRUB boot itself.
>
>> After some further poking it looks like the latest images bump the minimum memory
>> requirements from 256M to 512M so qemu-system-sparc64 needs to run with -m 512.
>
> I'm not sure this is actually required. debian-installer does actually have a low-mem
> installation mode with which I could the system even on 64 MiB.
Oh I didn't know that! From memory when you first started the ports images the
minimum memory required under qemu-system-sparc64 for the ISOs was 128M, then about
18 months ago it jumped to 256M and in the most recent it seems to be 512M.
Certainly let's not mix this with the grub discussion: sadly I don't have access to
real SPARC hardware but in a separate thread I'd be interested if you had a moment to
test the ISOs in a lower memory LDOM so I can determine whether qemu-system-sparc64
behaviour matches that of real hardware. If it isn't then my job is to fix that. And
of course don't let this stop you going ahead with merging the GRUB patches :)
>> One other slight annoyance is that when the grub menu loads it seems to hammer the
>> CPU for the first 60s or so which makes selection almost impossible. I'll see if I
>> can find some time to look at this later.
>
> I have not seen this on real hardware. Worked just fine.
My guess is that this is related to the way in which the keyboard/screen are being
polled trigging something in the emulation, so I see this as being a
qemu-system-sparc64 issue and not something to block the GRUB work.
>>>> Patches look good, although the second patch still uses /boot/grub/sparc64.elf
>>>> instead of /boot/grub/core.img for the grub-mkimage output parameter - is that correct?
>>> Good catch. Probably forgot a "git commit -a -amend" here.
>>>
>>> Btw, any suggestions for powerpc/ppc64? We are still using Yaboot here. I assume this
>>> case will be a bit more complicated due to the fact we also need to support PowerMacs.
>>
>> What's the current status? Sorry I had to drop out of that thread but I've recently
>> had a break because I got married :)
>
> Congrats!
>
>> I remember there being issues translating OF
>> pathnames for the OS, but was there progress using grub for HD installation and also
>> the installation media?
> The pathname translation issue was for grub-installer, i.e. the part where the installer
> installs GRUB to the target hard disk.
(goes and looks)
Wow there's a huge backlog of grub-related messages there :/ Do you have a link to
the latest summary update? I can start poking at the outstanding tasks as time allows.
ATB,
Mark.
Reply to: