Jiri Kosina wrote:

On Wed, 24 Feb 2010, Rick L. Vinyard, Jr. wrote:

quoted

The key difference is the replacement of spin_lock() with spin_trylock()
such that if the non-interrupt code has already obtained the lock, the
interrupt will not deadlock but instead take the else path and schedule
a
framebuffer update at the next interval.

Why is _irqsave() and/or deferred work not enough? The aproach with
_trylock() seems to be overly complicated for no good reason (I personally
become very suspicious every time I see code that is using _trylock()).

I was concerned about _irqsave() because the lock is split across two
functions to protect the urb after it is handed off to the usb subsystem
with usb_submit_urb(). It's locked in g13_fb_send() and unlocked in the
urb completion callback.

As for deferred work, the g13_fb_send() is the I/O portion of the deferred
framebuffer callback. I was concerned that without a lock one deferred
update could hand the urb off to the usb subsystem and a second could try
to access it before it was handed back to the driver.

In this case the _trylock() would fail and in the else patch we would
defer yet again until the next update cycle.

I took this approach because usb_interrupt_msg() couldn't be used from an
interrupt context, such as a resume hook because eventually down the chain
it does a wait_for_completion_timeout().

It has the added benefit of reusing the urb instead of creating a new one
for each framebuffer sent out, but that wasn't a reason... just a side
effect.

The downside is that I had to manage the urb.

One thing I could do is forget about directly calling g13_fb_send() from
any context and instead use the deferred framebuffer workqueue.

That's probably a simpler approach anyway.

[ by the way, Rick, are you planning to resubmit the G13 driver with
incorporated feedback from the last review round? ]

Yes. I just wanted to get the details of suspend/resume worked out before
resubmitting.

-- 

Rick