grub-devel.gnu.org archive mirror
 help / color / mirror / Atom feed
From: Andrei Borzenkov <arvidjaar@gmail.com>
To: "Elliott, Robert (Persistent Memory)" <elliott@hpe.com>,
	The development of GNU GRUB <grub-devel@gnu.org>,
	"dan.j.williams@intel.com" <dan.j.williams@intel.com>,
	"linux-nvdimm@lists.01.org" <linux-nvdimm@lists.01.org>
Subject: Re: grub causing NVDIMMs to be treated as normal memory
Date: Thu, 26 Nov 2015 06:30:11 +0300	[thread overview]
Message-ID: <56567CC3.3040405@gmail.com> (raw)
In-Reply-To: <94D0CD8314A33A4D9D801C0FE68B40295BE5549A@G9W0745.americas.hpqcorp.net>

26.11.2015 03:12, Elliott, Robert (Persistent Memory) пишет:
> 
> 
>> -----Original Message-----
>> From: Linux-nvdimm [mailto:linux-nvdimm-bounces@lists.01.org] On
>> Behalf Of Andrei Borzenkov
>> Sent: Wednesday, November 25, 2015 12:37 PM
>> To: The development of GNU GRUB <grub-devel@gnu.org>;
>> dan.j.williams@intel.com; linux-nvdimm@lists.01.org
>> Subject: Re: grub causing NVDIMMs to be treated as normal memory
>>
>> 25.11.2015 02:52, Elliott, Robert (Persistent Memory) пишет:
>>> We've noticed that some combinations of grub and old linux kernels
>>>
>>> end up interpreting the UEFI memory map EfiPersistentMemory type 14
>>> (formerly a reserved value) as regular memory in the linux e820
>>> table, causing silent data corruption on the NVDIMMs.  That occurs
>>> even though grub prints this message suggesting everything is safe:
>>>
>>>     Unknown memory type 14, considering reserved
>>>
>>>
>>>
>>> In broken versions of grub, the code parsing the UEFI memory map
>>> has a "default" case that falls through to the
>>>
>>> GRUB_EFI_BOOT_SERVICES_DATA case, which marks the memory range
>>> as GRUB_MEMORY_AVAILABLE and ends up in e820 as regular memory.
>>
>> Could you test if attached patch works for you (compile tested)?
> 
> Thanks.
> 
> I think I finally got that to compile with
>     configure --with-platform=efi
>     make
> 
> but have no clue how to install it and try it out.  I'm using a
> fedora22 system, which has its own /sbin/grub2-install.  I 
> don't understand how that differs from the grub-install in the
> build directory or how to get either of them to work.
> 

From the build directory

pkgdatadir=$PWD ./grub-install --bootloader-id testgrub -d grub-core

This should install grub in \EFI\testgrub on ESP and add EFI menu for
it. You can add --no-nvram to skip EFI menu update and load it manually
then. It will install modules in /boot/grub (instead of /boot/grub2), so
you should probably copy /boot/grub2/grub.cfg there.

Of course you can simply add patch to Fedora package and rebuild this
package.

> Anyway, we should create another patch that does:
> * #define GRUB_EFI_PERSISTENT_MEMORY 14	per UEFI 2.5
> * #define GRUB_E820_PERSISTENT_MEMORY 7	per ACPI 6.0
> * add a GRUB_MEMORY_PMEM enum
> * map GRUB_EFI_PERSISTENT_MEMORY -> GRUM_PMEMORY_MEM 
>   -> GRUB_E820_PERSISTENT_MEMORY per ACPI 6.0
> 
> to explicitly handle the new types (in addition to handling
> unknown values correctly).
> 

That is much more involved than obvious bug fix. It can be done later.


  reply	other threads:[~2015-11-26  3:30 UTC|newest]

Thread overview: 23+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-11-24 23:52 grub causing NVDIMMs to be treated as normal memory Elliott, Robert (Persistent Memory)
2015-11-25 14:08 ` Andrei Borzenkov
2015-11-25 16:51   ` Dan Williams
2015-11-25 17:04   ` Elliott, Robert (Persistent Memory)
2015-11-25 17:07   ` Seth Goldberg
2015-11-25 18:36 ` Andrei Borzenkov
2015-11-26  0:12   ` Elliott, Robert (Persistent Memory)
2015-11-26  3:30     ` Andrei Borzenkov [this message]
2015-11-26  6:15       ` Elliott, Robert (Persistent Memory)
2015-11-26 16:55         ` Andrei Borzenkov
2015-11-26 23:24           ` Elliott, Robert (Persistent Memory)
2015-11-27  3:58             ` Andrei Borzenkov
2015-11-27  6:22               ` Elliott, Robert (Persistent Memory)
2015-11-27 11:08                 ` Vladimir 'φ-coder/phcoder' Serbinenko
2015-11-27 11:48                   ` Andrei Borzenkov
2015-11-27 13:55                     ` Vladimir 'φ-coder/phcoder' Serbinenko
2015-11-27 17:23                       ` Andrei Borzenkov
2015-11-28  6:41                   ` Elliott, Robert (Persistent Memory)
2015-12-01  0:25                     ` Elliott, Robert (Persistent Memory)
2015-12-03 17:50                     ` Elliott, Robert (Persistent Memory)
2015-12-08 17:15                       ` Andrei Borzenkov
2015-12-09  6:37                         ` Elliott, Robert (Persistent Memory)
2015-12-29 17:17                     ` Vladimir 'φ-coder/phcoder' Serbinenko

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=56567CC3.3040405@gmail.com \
    --to=arvidjaar@gmail.com \
    --cc=dan.j.williams@intel.com \
    --cc=elliott@hpe.com \
    --cc=grub-devel@gnu.org \
    --cc=linux-nvdimm@lists.01.org \
    /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).