Bug#238213: NPTL per char IO locking very slow

[GOTO Masanori <gotom@debian.or.jp>]

At Tue, 16 Mar 2004 07:56:10 +0800,
Michael Clark wrote:
> Package: libc6
> Version: 2.3.2.ds1-11
> 
> Did some tests with bonnie++ on 2.6.3 with latest glibc in sid and
> noticed per char IO is dreadfully slow. Remember a thread months
> back about this on LKML: http://kerneltrap.org/node/view/1574
> 
> Here are the relevant benchmarks run with glibc-2.3.2.ds1-11
> (using LD_ASSUME_KERNEL=2.4.1 to disable NPTL).
> 
> Version  1.03       ------Sequential Output------ --Sequential Input- --Random-
>                      -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
> Machine        Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP  /sec %CP
> 2.6.3 NPTL       1G  3313  95 16808   8  5619   3  3291  93 16130   2  74.1   0
> 2.6.3 LxThreads  1G 11562  92 17384   8  6155   2 11370  81 15846   2  65.1   0
> 2.6.3 NPTL-fixed 1G 10695  78 18139  10  5399   3 12534  86 16298   3  73.7   0
> 
> Fix for this was in glibc CVS and have included a patch (tested).

OK, I've put it in.  I believe this patch does not affect any changes except for
performance issue.

Regards,
-- gotom