[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: Loading different kernels based on platform identification (DMI)
From: |
Rajat Jain |
Subject: |
RE: Loading different kernels based on platform identification (DMI) |
Date: |
Thu, 29 Jan 2015 23:38:24 +0000 |
Hello Davig,
Thanks a bunch for the pointer!
I will try and test it out on my system next week and will let let know of the
results.
Thanks,
Rajat
> -----Original Message-----
> From: David Michael [mailto:address@hidden
> Sent: Thursday, January 29, 2015 2:05 PM
> To: Rajat Jain
> Cc: Andrei Borzenkov; address@hidden; Sanjay Jain; Raghuraman
> Thirumalairajan; Stu Grossman
> Subject: Re: Loading different kernels based on platform identification (DMI)
>
> Hi,
>
> On Thu, Jan 29, 2015 at 2:23 PM, Rajat Jain <address@hidden> wrote:
> > [+David]
> >
> > Hello,
> >
> >> DMI parsing was discussed a while back and proposed patch was posted
> >> on grub-devel. It was quite a while back, so patch probably needs
> rebasing.
> >> Somebody who cares enough should probably try to submit it for review
> >> again after testing with current git.
> >
> > I found a couple references, and the best bet I could find is this:
> >
> > http://lists.gnu.org/archive/html/grub-devel/2013-04/msg00153.html
> >
> > Can you please confirm if this is the patch you were referring to?
>
> You can download the latest version of the bare module file from the
> following link. It's been updated to support both EFI and traditional BIOS
> systems since that e-mail.
>
> https://github.com/dm0-/gnuxc/blob/master/patches/grub-2.02~beta2-
> smbios.c
>
> That module targets the last beta from alpha.gnu.org. Sorry for the lack of a
> patch file at the moment, but it can be built by moving the file to grub-
> core/commands/i386/smbios.c, appending "./grub-
> core/commands/i386/smbios.c" to po/POTFILES.in, appending the following
> block to grub-core/Makefile.core.def, then running autoreconf, configure,
> and make normally.
>
> module = {
> name = smbios;
> x86 = commands/i386/smbios.c;
> enable = x86;
> };
>
> If you want to test the module, it returns the requested data type at a given
> byte offset in an SMBIOS table entry. You can run "smbios --help" for the
> options list. For example, this will store the system manufacturer string in
> a
> variable:
>
> smbios --type 1 --get-string 4 --variable system_manufacturer echo
> $system_manufacturer
>
> > David: I'm running on Intel's latest 64 bit Haswell processor. Can you
> > please
> confirm if this patch should work on that (I ask because I found references
> about different ways to deal with 32 bit and 64 bit CPUs/BIOSes).
>
> I don't recall anything offhand that will cause a problem, but it shouldn't
> hurt
> to test. It seemed to work okay on every machine I've tried.
>
> > I'm interested and I think it would be highly useful to have this up
> streamed. Wanted to check if David or someone else is planning to submit it,
> or shall I try submitting it after rebasing to the latest and testing on my
> machine?
>
> I don't mind updating and trying to submit it to the development list for
> comments, maybe over the weekend. (I left that FSF copyright statement
> on it from when I originally planned to submit it long ago, but I haven't
> actually submitted copyright assignment papers. Oops.) Can someone tell
> me if this change would require copyright assignment?
>
> Thanks.
>
> David