[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20250206091714.684b592d@kernel.org>
Date: Thu, 6 Feb 2025 09:17:14 -0800
From: Jakub Kicinski <kuba@...nel.org>
To: Mina Almasry <almasrymina@...gle.com>
Cc: davem@...emloft.net, netdev@...r.kernel.org, edumazet@...gle.com,
pabeni@...hat.com, andrew+netdev@...n.ch, horms@...nel.org
Subject: Re: [PATCH net-next 3/3] netdevsim: allow normal queue reset while
down
On Thu, 6 Feb 2025 08:43:58 -0800 Mina Almasry wrote:
> > diff --git a/drivers/net/netdevsim/netdev.c b/drivers/net/netdevsim/netdev.c
> > index 42f247cbdcee..d26b2fb1cabc 100644
> > --- a/drivers/net/netdevsim/netdev.c
> > +++ b/drivers/net/netdevsim/netdev.c
> > @@ -644,6 +644,9 @@ nsim_queue_mem_alloc(struct net_device *dev, void *per_queue_mem, int idx)
> >
> > if (ns->rq_reset_mode > 3)
> > return -EINVAL;
> > + /* Only "mode 0" works when device is down */
>
> Just to make sure I understand: modes 2 & 3 should also work with the
> device down, and this is an artificial limitation that you're adding
> to netdevsim, right? I don't see how changing the call order to
> napi_del and napi_add_config breaks resetting while the device is
> down.
It should work, but I sprinkled a bunch of checks in page pool which
expect idle NAPIs to have the SCHED bit set. So we run into false
positive warnings if we create and destroy a page pool for NAPI which
was never initialized / listed. Let me take another look.
Powered by blists - more mailing lists