Re: [PATCH net] ipv4: igmp: hold wakelock to prevent delayed reports

[Tejaswi Tanikella <tejaswit@codeaurora.org>]

On Mon, Jun 04, 2018 at 11:06:40AM -0400, David Miller wrote:
> From: Tejaswi Tanikella <tejaswit@codeaurora.org>
> Date: Fri, 1 Jun 2018 19:35:41 +0530
> 
> > On receiving a IGMPv2/v3 query, based on max_delay set in the header a
> > timer is started to send out a response after a random time within
> > max_delay. If the system then moves into suspend state, Report is
> > delayed until system wakes up.
> > 
> > In one reported scenario, on arm64 devices, max_delay was set to 10s,
> > Reports were consistantly delayed if the timer is scheduled after 5 plus
> > seconds.
> > 
> > Hold wakelock while starting the timer to prevent moving into suspend
> > state.
> > 
> > Signed-off-by: Tejaswi Tanikella <tejaswit@codeaurora.org>
> 
> As Florian stated, this won't be the only networking facility to hit
> a problem like this.  So, if we go down this route, we probably want
> to generically solve this.
> 
> But I have a deeper concern.
> 
> Do we really want every timer based querying mechanism to prevent a
> system from being able to suspend?
> 
> We get to the point where external entities can generate traffic which
> can prevent a remote system from entering suspend state.

Like you suggested maybe aquiring a wakelock will unnecessarily stop the
system from suspending.
Would using alarmtimer be better? It seems similar to a hrtimer but can
wake the system up from suspend.

This should allow the system to suspend till the timer expires. But might
cause repeated wake-ups and suspends.