Re: [PATCH 1/2] net: Assert proper context while calling napi_schedule()

[Jakub Kicinski <kuba@kernel.org>]

On Thu, 13 Feb 2025 01:58:17 -0800 Breno Leitao wrote:
> > Looks like netcons is hitting this warning in netdevsim:
> > 
> > [   16.063196][  T219]  nsim_start_xmit+0x4e0/0x6f0 [netdevsim]
> > [   16.063219][  T219]  ? netif_skb_features+0x23e/0xa80
> > [   16.063237][  T219]  netpoll_start_xmit+0x3c3/0x670
> > [   16.063258][  T219]  __netpoll_send_skb+0x3e9/0x800
> > [   16.063287][  T219]  netpoll_send_skb+0x2a/0xa0
> > [   16.063298][  T219]  send_ext_msg_udp+0x286/0x350 [netconsole]
> > [   16.063325][  T219]  write_ext_msg+0x1c6/0x230 [netconsole]
> > [   16.063346][  T219]  console_emit_next_record+0x20d/0x430
> > 
> > https://netdev-3.bots.linux.dev/vmksft-net-drv-dbg/results/990261/7-netcons-basic-sh/stderr
> > 
> > We gotta fix that first.  
> 
> Thanks Jakub,
> 
> I understand that it will be fixed by this patchset, right?

The problem is a bit nasty, on a closer look. We don't know if netcons
is called in IRQ context or not. How about we add an hrtimer to netdevsim,
schedule it to fire 5usec in the future instead of scheduling NAPI
immediately? We can call napi_schedule() from a timer safely.

Unless there's another driver which schedules NAPI from xmit.
Then we'd need to try harder to fix this in netpoll.
veth does use NAPI on xmit but it sets IFF_DISABLE_NETPOLL already.