On Wed, 28 Nov 2001, Anton Blanchard wrote:
> The dbench-o-matic results from removing the BKL from llseek can be
> found here:
>
> http://samba.org/~anton/linux/llseek/
>
> Same hardware as the earlier test, 12 way ppc64. The patch is pretty
> dumb, I just pushed the kernel lock down into the llseek methods and
> only replaced it with the inode semaphore in generic_llseek. Im sure
> Al will have a better fix for 2.5, I just wanted to highlight the
> problem :)
yeah i did that once (about a year ago or more) and it was rejected, cant
remember why :-) But it's indeed an important optimization and it shows up
in dbench overhead.
Ingo
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Fri Nov 30 2001 - 21:00:30 EST