[RFC 0/2] Net softirq deferring to ksoftirqd

[Dmitry Safonov]

Based on events I've saw on out-of-tree drivers, but I believe that
can happen on the mainstream ones.
I managed to get some results in Qemu VMs, but this lacks testing
on a real mainstream hardware.

Also I'm not sure that's completely sane thing to-do, so sending early
to get some reviews.

Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
Cc: David Miller <davem@xxxxxxxxxxxxx>
Cc: Eric Dumazet <edumazet@xxxxxxxxxx>
Cc: Frederic Weisbecker <fweisbec@xxxxxxxxx>
Cc: Hannes Frederic Sowa <hannes@xxxxxxxxxxxxxxxxxxx>
Cc: Ingo Molnar <mingo@xxxxxxxxxx>
Cc: "Levin, Alexander (Sasha Levin)" <alexander.levin@xxxxxxxxxxx> 
Cc: Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx>
Cc: Paolo Abeni <pabeni@xxxxxxxxxx>
Cc: "Paul E. McKenney" <paulmck@xxxxxxxxxxxxxxxxxx> 
Cc: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
Cc: Radu Rendec <rrendec@xxxxxxxxxx>
Cc: Rik van Riel <riel@xxxxxxxxxx>
Cc: Stanislaw Gruszka <sgruszka@xxxxxxxxxx>
Cc: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
Cc: Wanpeng Li <wanpeng.li@xxxxxxxxxxx>

Dmitry Safonov (2):
  softirq: Defer net rx/tx processing to ksoftirqd context
  softirq: Introduce mask for __do_softirq()

 include/linux/interrupt.h |  8 ++++----
 kernel/softirq.c          | 48 ++++++++++++++++++++++++++++++++++-------------
 2 files changed, 39 insertions(+), 17 deletions(-)

-- 
2.13.6