On 2/14/2024 6:54 AM, Alexander Lobakin wrote:
From: Alexander Lobakin <aleksander.loba...@intel.com>
Date: Tue, 13 Feb 2024 14:16:47 +0100
From: Alan Brady <alan.br...@intel.com>
Date: Wed, 7 Feb 2024 16:42:43 -0800
From: Emil Tantilov <emil.s.tanti...@intel.com>
Fix softirq's not being handled during napi_schedule() call when
receiving marker packets for queue disable by disabling local bottom
half.
BTW, how exactly does this help?
__napi_schedule() already disables interrupts (local_irq_save()).
napi_schedule_prep() only has READ_ONCE() and other atomic read/write
helpers.
It's always been safe to call napi_schedule() with enabled BH, so I
don't really understand how this works.
It's been a while since I debugged this, I'll have to take a look again,
but its not so much about being safe as it is about making sure the
marker packets are received in those cases - like ifdown in the trace.
This also needs to be dropped from the fixes queue until investigated.
For now, it looks like a cheap hack (without the explanation how exactly
it does help), not a proper fix.
It does fix the issue at hand. Looking at the kernel code I see multiple
examples of napi_schedule() being wrapped with local_bh_disable/enable,
so this appears to be a common (not uncommon?) technique.
Thanks,
Emil
Thanks,
Olek
