All of lore.kernel.org
 help / color / mirror / Atom feed
From: Bill Kendall <wkendall@sgi.com>
To: aelder@sgi.com
Cc: xfs@oss.sgi.com
Subject: Re: [PATCH v2 4/9] xfsrestore: mmap dirent names for faster lookups
Date: Mon, 15 Nov 2010 15:51:45 -0600	[thread overview]
Message-ID: <4CE1AB71.3000708@sgi.com> (raw)
In-Reply-To: <1289604322.2315.909.camel@doink>

On 11/12/2010 05:25 PM, Alex Elder wrote:
> On Fri, 2010-11-05 at 11:35 -0500, wkendall@sgi.com wrote:
>> plain text document attachment (namreg_map)
>> Pathname resolution in xfsrestore is about 4x faster if the file
>> containing dirent names ("namreg") is memory mapped.  If xfsrestore is
>> unable to map the file (e.g., due to virtual memory constraints)
>> fallback to the existing seek-and-read approach.
>>
>> The file is mapped after all directory entries have been written to
>> the "namreg" file. If the caller tries to add additional entries after
>> the file has been mapped, it will be unmapped and restore will resort
>> back to seek-and-read lookups.
>>
>> Signed-off-by: Bill Kendall<wkendall@sgi.com>
>
> I guess I might have created a simple namreg_flush_final()
> routine to encapsulate the namreg_flush() and namreg_map()
> calls, rather than adding a flag.  Then namreg_flush() could
> have been made to have static scope.  No big deal though.
>
> Also, you fall back to the non-mmapped method in case a
> flush or add happens after you've mapped the file.  That
> is in some sense comforting, but it would be nice to be
> assured it simply won't happen.  If you come to a point
> someday where you are sure of this it would be nice to
> see that code switched to assertions instead.
>
> You could define the new namreg_*map*() functions earlier
> in the file and skip the forward declarations.

But the comments told me to put the functions there! ;)
xfsdump/restore could benefit from some major refactoring.

>
> These aren't all that important though, I think it's a
> good change.  Let me know if you think you will
> re-work it but I think it's OK as-is.

I'll leave this patch as is, and make a note to look
at changing the mmap-checks into asserts.

Bill

>
> Reviewed-by: Alex Elder<aelder@sgi.com>
>
>

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

  reply	other threads:[~2010-11-15 21:50 UTC|newest]

Thread overview: 24+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-11-05 16:35 [PATCH v2 0/9] xfsrestore dirent limitations and scaling issues wkendall
2010-11-05 16:35 ` [PATCH v2 1/9] xfsrestore: turn off NODECHK wkendall
2010-11-12 23:23   ` Alex Elder
2010-11-05 16:35 ` [PATCH v2 2/9] xfsrestore: change nrh_t from 32 to 64 bits wkendall
2010-11-12 23:24   ` Alex Elder
2010-11-05 16:35 ` [PATCH v2 3/9] xfsrestore: cache path lookups wkendall
2010-11-12 23:25   ` Alex Elder
2010-11-05 16:35 ` [PATCH v2 4/9] xfsrestore: mmap dirent names for faster lookups wkendall
2010-11-12 23:25   ` Alex Elder
2010-11-15 21:51     ` Bill Kendall [this message]
2010-11-05 16:35 ` [PATCH v2 5/9] xfsrestore: cleanup node allocation wkendall
2010-11-15 20:38   ` Alex Elder
2010-11-15 21:36     ` Bill Kendall
2010-11-05 16:35 ` [PATCH v2 6/9] xfsrestore: fix node table setup wkendall
2010-11-15 20:38   ` Alex Elder
2010-11-15 21:30     ` Bill Kendall
2010-11-05 16:35 ` [PATCH v2 7/9] xfsrestore: make node lookup more efficient wkendall
2010-11-15 20:38   ` Alex Elder
2010-11-15 22:06     ` Bill Kendall
2010-11-05 16:35 ` [PATCH v2 8/9] xfsrestore: remove nix_t wkendall
2010-11-12 23:25   ` Alex Elder
2010-11-05 16:35 ` [PATCH v2 9/9] xfsrestore: check for compatible xfsrestore wkendall
2010-11-12 23:25   ` Alex Elder
2010-11-12 23:25 ` [PATCH v2 0/9] xfsrestore dirent limitations and scaling issues Alex Elder

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=4CE1AB71.3000708@sgi.com \
    --to=wkendall@sgi.com \
    --cc=aelder@sgi.com \
    --cc=xfs@oss.sgi.com \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.