linux-ide.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH 0/1] sata_mv: Prevent PIO commands to be defered too long if traffic in progress.
@ 2009-10-12 22:43 Gwendal Grignou
  2009-10-12 22:44 ` [PATCH 1/1] " Gwendal Grignou
  0 siblings, 1 reply; 6+ messages in thread
From: Gwendal Grignou @ 2009-10-12 22:43 UTC (permalink / raw)
  To: mlord; +Cc: linux-ide

Hello,

I found a bug in the sata_mv driver, that defers PIO commands indefinitely
while there is a traffic on other ports of a sata port multiplier [PMP]:

without traffic:
time smartctl --all /dev/sdc > /dev/null
real    0m0.243s

with fio running and blasting read IOs on 5 drives behind the same port multiplier for 5 minutes:
real    4m59.353s

Obviously the PIO commands are defer while traffic is in progress.

With the patch following, smart_ctl return before fio complete. It still takes 4s, time for all NCQ commands to be flushed out.
real    0m4.074s

Gwendal.

^ permalink raw reply	[flat|nested] 6+ messages in thread

end of thread, other threads:[~2009-10-16 10:27 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2009-10-12 22:43 [PATCH 0/1] sata_mv: Prevent PIO commands to be defered too long if traffic in progress Gwendal Grignou
2009-10-12 22:44 ` [PATCH 1/1] " Gwendal Grignou
2009-10-13  1:44   ` Grant Grundler
2009-10-14 12:35   ` Mark Lord
2009-10-14 12:38   ` Jeff Garzik
2009-10-16 10:26   ` Jeff Garzik

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).