* nfs shared mmap performance regression against 2.6.23
@ 2008-04-29 15:18 Arnd Bergmann
0 siblings, 0 replies; only message in thread
From: Arnd Bergmann @ 2008-04-29 15:18 UTC (permalink / raw)
To: linux-nfs
Cc: Trond.Myklebust, brad.benton, jroth, mkistler, adetsch, lxie,
mijo, gerhard.stenzel, uweigand, brokensh, hannsj_uhl, brianh,
tstaudt, linux-mm
Hi Trond,
We had an application writer report a performance regression between
2.6.23 and later kernels that we ultimately tracked down to your
patch 94387fb1aa, "NFS: Add the helper nfs_vm_page_mkwrite". It turns
out the application was accidentally using a writable shared mmap
between two processes on an NFS mount instead of tmpfs.
The setup is fixed now, but I'm still surprised by the 7x slowdown
caused by this patch for application runtime when the mmap is shared
between two processes writing to it. If there is only one process
that maps the file, there is no measureable slowdown.
Is that an expected side-effect of your patch, or something that
should not have happened?
Arnd <><
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2008-04-29 15:18 UTC | newest]
Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2008-04-29 15:18 nfs shared mmap performance regression against 2.6.23 Arnd Bergmann
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox