Messages in this thread | | | Subject | Re: tbench regression in 2.6.25-rc1 | From | Valdis.Kletnieks@vt ... | Date | Mon, 18 Feb 2008 12:33:53 -0500 |
| |
On Mon, 18 Feb 2008 16:12:38 +0800, "Zhang, Yanmin" said:
> I also think __refcnt is the key. I did a new testing by adding 2 unsigned long > pading before lastuse, so the 3 members are moved to next cache line. The performance is > recovered. > > How about below patch? Almost all performance is recovered with the new patch. > > Signed-off-by: Zhang Yanmin <yanmin.zhang@intel.com>
Could you add a comment someplace that says "refcnt wants to be on a different cache line from input/output/ops or performance tanks badly", to warn some future kernel hacker who starts adding new fields to the structure? [unhandled content-type:application/pgp-signature] | |