[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: [PATCH] Improve single byte read stream performance

From: Johan Corveleyn <jcorvel_at_gmail.com>
Date: Mon, 15 Mar 2010 21:56:38 +0100

On Mon, Mar 15, 2010 at 6:18 PM, Ivan Zhakov <ivan_at_visualsvn.com> wrote:
> On Mon, Mar 15, 2010 at 17:01, C. Michael Pilato <cmpilato_at_collab.net> wrote:
>> Philip Martin wrote:
>>> Johan Corveleyn <jcorvel_at_gmail.com> writes:
>>>> On Tue, Mar 9, 2010 at 8:04 PM, Julian Foad <julian.foad_at_wandisco.com> wrote:
>>>>> On Sun, 2010-03-07, Stefan Fuhrman wrote:
>>>>>> Speed up input stream processing in config parser and
>>>>>> others that read single bytes from a stream.
>>>>>> * subversion/libsvn_subr/subst.c
>>>>>>   (translated_stream_read): Add an optimized code path
>>>>>>     for single byte read requests.
>>>>>> Patch by: Stefan Fuhrmann <stefanfuhrmann{_AT_}alice-dsl.de>
>>>>> Committed in r921057.
>>>> Any chance this could be backported to 1.6.x?
>>>> This seems like a nice performance improvement with low impact on the
>>>> rest of the code. Would be nice if this could be put out there as
>>>> quickly as possible (i.e. not having to wait for 1.7) for users to
>>>> enjoy ...
>>> Can you quantify the improvement?  I don't suppose a GUI like
>>> TortoiseSVN reads config files very often.  Any improvement is lost in
>>> the noise when I run the command line client on Linux.  It doesn't
>>> change the number of system calls reported by 'strace -c'.
>> Do our servers (mod_dav_svn and svnserve) parse authz files using this
>> codepath?
> Yes, they use the same code.
>> How often?  On a per-connection basis?
> mod_dav_svn caches authz file per connection.

I just tested with svnserve x fsfs. For every connection it uses this
code to read:
- fsfs.conf (from within find_repos -> svn_repos_open)
- svnserve.conf and authz (from within find_repos -> load_configs).
There's no caching whatsoever, these files are read everytime a client
requests something.

I have no idea of the real performance impact, but maybe I'll do some
measurements later tonight ...

Received on 2010-03-15 21:57:05 CET

This is an archived mail posted to the Subversion Dev mailing list.