OSDN Git Service

e1000e: Avoid missed interrupts following ICR read
authorBenjamin Poirier <bpoirier@suse.com>
Thu, 8 Feb 2018 06:47:14 +0000 (15:47 +0900)
committerJeff Kirsher <jeffrey.t.kirsher@intel.com>
Mon, 5 Mar 2018 18:08:20 +0000 (10:08 -0800)
commit116f4a640b3197401bc93b8adc6c35040308ceff
tree23900b859d705a00642a20bbc86a8ffffb138618
parent361a954e6a7215de11a6179ad9bdc07d7e394b04
e1000e: Avoid missed interrupts following ICR read

The 82574 specification update errata 12 states that interrupts may be
missed if ICR is read while INT_ASSERTED is not set. Avoid that problem by
setting all bits related to events that can trigger the Other interrupt in
IMS.

The Other interrupt is raised for such events regardless of whether or not
they are set in IMS. However, only when they are set is the INT_ASSERTED
bit also set in ICR.

By doing this, we ensure that INT_ASSERTED is always set when we read ICR
in e1000_msix_other() and steer clear of the errata. This also ensures that
ICR will automatically be cleared on read, therefore we no longer need to
clear bits explicitly.

Signed-off-by: Benjamin Poirier <bpoirier@suse.com>
Acked-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
drivers/net/ethernet/intel/e1000e/defines.h
drivers/net/ethernet/intel/e1000e/netdev.c