ahci: switch from 'threaded' to 'hardirq' interrupt handling
authorDan Williams <dan.j.williams@intel.com>
Thu, 12 Nov 2015 00:27:38 +0000 (16:27 -0800)
committerTejun Heo <tj@kernel.org>
Mon, 16 Nov 2015 16:29:10 +0000 (11:29 -0500)
commita6b7fb764ed2a6b7bb1ac96d93c06787aa589092
tree02bd0d5a1aa33e2fa3cb509ca4607b9ff11558bc
parentd684a90d38e24dcaf95fdb32c83efe05f80d152a
ahci: switch from 'threaded' to 'hardirq' interrupt handling

For high frequency I/O the overhead of threaded interrupts impacts
performance.  A quick out-of-the-box test (i.e. no affinity tuning)
shows ~10% random read performance at ~20% less cpu.  The cpu wins
appear to be from reduced lock contention.

Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
drivers/ata/libahci.c