ioat: fix tasklet tear down
authorDan Williams <dan.j.williams@intel.com>
Thu, 20 Feb 2014 00:19:35 +0000 (16:19 -0800)
committerDan Williams <dan.j.williams@intel.com>
Tue, 25 Feb 2014 17:44:20 +0000 (09:44 -0800)
commitda87ca4d4ca101f177fffd84f1f0a5e4c0343557
tree5b2be4209f11af7a3df5cf68997e9547612481ef
parent6d0abeca3242a88cab8232e4acd7e2bf088f3bc2
ioat: fix tasklet tear down

Since commit 77873803363c "net_dma: mark broken" we no longer pin dma
engines active for the network-receive-offload use case.  As a result
the ->free_chan_resources() that occurs after the driver self test no
longer has a NET_DMA induced ->alloc_chan_resources() to back it up.  A
late firing irq can lead to ksoftirqd spinning indefinitely due to the
tasklet_disable() performed by ->free_chan_resources().  Only
->alloc_chan_resources() can clear this condition in affected kernels.

This problem has been present since commit 3e037454bcfa "I/OAT: Add
support for MSI and MSI-X" in 2.6.24, but is now exposed. Given the
NET_DMA use case is deprecated we can revisit moving the driver to use
threaded irqs.  For now, just tear down the irq and tasklet properly by:

1/ Disable the irq from triggering the tasklet

2/ Disable the irq from re-arming

3/ Flush inflight interrupts

4/ Flush the timer

5/ Flush inflight tasklets

References:
https://lkml.org/lkml/2014/1/27/282
https://lkml.org/lkml/2014/2/19/672

Cc: Ingo Molnar <mingo@elte.hu>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: <stable@vger.kernel.org>
Reported-by: Mike Galbraith <bitbucket@online.de>
Reported-by: Stanislav Fomichev <stfomichev@yandex-team.ru>
Tested-by: Mike Galbraith <bitbucket@online.de>
Tested-by: Stanislav Fomichev <stfomichev@yandex-team.ru>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
drivers/dma/ioat/dma.c
drivers/dma/ioat/dma.h
drivers/dma/ioat/dma_v2.c
drivers/dma/ioat/dma_v3.c