Btrfs devel - Oct 2008 - crc32c implementation on x86 with SSE4.2... CONFIG_BTRFS_HW_SUM

Hi there,

I noticed that btrfs, in the git tree, has its own implementation of
crc32c for x86 with SSE4.2 that implement a crc32 instruction.. it
appears.
Its protected by a ifdef CONFIG_BTRFS_HW_SUM, and if we''re executing
on the right cpu, that will bypass the kernel implementation of
crc32c.

please see:
http://git.kernel.org/?p=linux/kernel/git/mason/btrfs-unstable.git;a=blob_plain;f=fs/btrfs/crc32c.h;hb=HEAD

Based on this understanding, should this be a implementation detail of
crc32c, and be correctly implemented (and therefore patched) in the
mainline kernel implementation of crc32c ?

I know that checksumming is cpu expensive, and btrfs wants the best
impl. possible. but that shouldn''t be a btrfs issue, that should be
crc32c issue, and its the crc32c implementation that should be fixed,
instead of having our own special crc32c in case we''re executing in
the right cpu.

Should I post a patch with this for the vanilla crc32c implementation?

kind regards,

-- 
Miguel Sousa Filipe
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs"
in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

On Thu, 2008-10-16 at 14:40 +0100, Miguel Sousa Filipe
wrote:
> Hi there,
> 
> I noticed that btrfs, in the git tree, has its own implementation of
> crc32c for x86 with SSE4.2 that implement a crc32 instruction.. it
> appears.
> Its protected by a ifdef CONFIG_BTRFS_HW_SUM, and if we''re
executing
> on the right cpu, that will bypass the kernel implementation of
> crc32c.
> 
The Btrfs support for intel''s faster crc32c came from a patch intel
posted to linux-kernel.  I just hacked it up a bit at the time because I
didn''t want to wait for mainline to include the work.

I don''t see intel''s patches in mainline yet, but I know there
was a plan
to get them there.

-chris


--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs"
in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Dnia 2008-10-16, czw o godzinie 09:49 -0400, Chris Mason
pisze:
> On Thu, 2008-10-16 at 14:40 +0100, Miguel Sousa Filipe wrote:
> > Hi there,
> > 
> > I noticed that btrfs, in the git tree, has its own implementation of
> > crc32c for x86 with SSE4.2 that implement a crc32 instruction.. it
> > appears.
> I don''t see intel''s patches in mainline yet, but I know
there was a
plan
> to get them there.
  It seems to be merged:
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=8cb51ba8e06570a5fff674b3744d12a1b089f2d0

  Using generic, replaceable by arch version is surely better. This way
you can avoid implementing crc32 for each architecture (like, for
example UltraSPARC T2, which computes crc at healthy 48 GB/s). 

-- 
Tomasz Torcz

--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs"
in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html