From: Andrey Borzenkov <arvidjaar@gmail.com>
To: Holger Goetz <holgerg@hgsys.de>
Cc: grub-devel@gnu.org
Subject: Re: Obtaining the UUID of the system for a PXE boot
Date: Tue, 23 Jul 2013 20:45:03 +0400 [thread overview]
Message-ID: <20130723204503.16520724@opensuse.site> (raw)
In-Reply-To: <51EEAE57.8070700@hgsys.de>
В Tue, 23 Jul 2013 18:24:55 +0200
Holger Goetz <holgerg@hgsys.de> пишет:
> Hi,
>
> looks i could need some help here.
>
> I've moved the smbios module from commands/i386 to commands and added
> some casts to make the compiler happy. As I need a x86_64-efi grub.
> What i see now is that the "_SM_" signature is not found: "Failed to
> locate entry point structure" some added grub_dprintf's show all zeros
> in the memory area.
> Actually when i use the "dump" command to display the memory in the area
> that gets swiped (0xf0000..0x10000) it (also) only displays "00"s.
> Is there some memory mapping required in x86_64 mode to access the
> memory area properly? I did cross check w/ the Linux kernel efi code and
> syslinux all basically do the same searching.
>
You need to fetch SMBIOS table address from EFI system table. See as
example
https://git.ipxe.org/ipxe.git/blob/HEAD:/src/interface/efi/efi_smbios.c
You can use grub_machine_acpi_get_rsdpv2() as example implementation.
Just add SMBIOS GUID and any sanity checks as required.
> Thanks,
> Holger
>
>
>
> On 19.07.2013 18:31, David Michael wrote:
> > Hi,
> >
> > On Fri, Jul 19, 2013 at 10:24 AM, Holger Goetz <holgerg@hgsys.de> wrote:
> >> Yes it's towards the right direction. But it is 32bit only if i understand
> >> correctly, and it basically is a memory access to fixed/hardcoded MEMORY
> >> address (0x80000001). to pick the veondor id and machine info.
> > It's true that I've mostly been using the module on 32-bit virtual
> > machines, so it hasn't really been tested elsewhere. However, I'm not
> > sure I understand what you mean by the hard-coded memory address. The
> > first function grub_smbios_locate_eps searches for the SMBIOS entry
> > point structure as described in the spec. The table entries are then
> > read at the table address found in the EPS, not a hard-coded location.
> >
> >> I have only 64bit - UEFI here - therefore the approach w/ first searching
> >> the SMBIOS infoblock in memory is probably required. And then properly walk
> >> through the info-tables/blocks to get to the UUID entry. It doesn'T need to
> >> be a fixed info to be retrieved from the SMBIOS memory - maybe a generic
> >> function to query/search a specific entry and return that to be assigned to
> >> a variable would be more flexible.
> > The module's command-line interface does use a (dumb) query/search
> > method. You can specify the desired entry's type and/or handle and
> > the data to retrieve from it. For example the following command
> > prints the machine name (i.e. the string at offset 5 in an entry with
> > type 1).
> >
> > smbios -t 1 -s 5
> >
> > I think you may have found the first patch I sent in that old thread,
> > which was for different functionality. The SMBIOS module can be
> > downloaded from the list archive[1].
> >
> > Unfortunately it doesn't have any convenient functions to output a
> > usable UUID. It shouldn't take much to add one: the variable "entry"
> > in the "main" function is a pointer to the matched entry, so entry[8]
> > through entry[23] is the UUID in a call "smbios -t 1 ...". I've
> > verified these bytes correspond to dmidecode output on my physical
> > hardware with the following.
> >
> > for i in 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
> > do smbios -t 1 -b $i
> > done
> >
> > If the module isn't salvageable on UEFI, maybe I can send out an
> > updated version whenever I upgrade to such a system.
> >
> > Thanks.
> >
> > David
> >
> > [1] https://lists.gnu.org/archive/html/grub-devel/2013-04/binx8am8MvVSh.bin
> >
> > _______________________________________________
> > Grub-devel mailing list
> > Grub-devel@gnu.org
> > https://lists.gnu.org/mailman/listinfo/grub-devel
> >
> >
>
prev parent reply other threads:[~2013-07-23 16:45 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <51E67DC2.1090609@hgsys.de>
[not found] ` <CAA91j0XETaDEh16K0WJ3p5hNxxVXWk-PDBh2Gp-TYq6UjvRetw@mail.gmail.com>
[not found] ` <51E7C5F0.5000308@hgsys.de>
[not found] ` <20130718193501.629d9c69@opensuse.site>
2013-07-19 14:24 ` Obtaining the UUID of the system for a PXE boot Holger Goetz
2013-07-19 16:31 ` David Michael
2013-07-19 21:08 ` Vladimir 'φ-coder/phcoder' Serbinenko
2013-07-20 14:23 ` Holger Goetz
2013-07-23 16:24 ` Holger Goetz
2013-07-23 16:45 ` Andrey Borzenkov [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20130723204503.16520724@opensuse.site \
--to=arvidjaar@gmail.com \
--cc=grub-devel@gnu.org \
--cc=holgerg@hgsys.de \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).