[PATCH V3 39/63] GIC: Added dummy handlers for Power ManagementSuspend Resume

[Russell King - ARM Linux]

On Mon, Dec 20, 2010 at 05:50:37PM +0530, Santosh Shilimkar wrote:
> > -----Original Message-----
> > From: linux-arm-kernel-bounces at lists.infradead.org [mailto:linux-arm-
> > kernel-bounces at lists.infradead.org] On Behalf Of Russell King - ARM
> Linux
> > Sent: Monday, December 20, 2010 5:20 PM
> > To: viresh kumar
> > Cc: Rajeev KUMAR; Armando VISCONTI; Vipin KUMAR; Shiraz HASHIM; Amit
> > VIRDI; Vipul Kumar SAMAR; Deepak SIKRI; linux-arm-
> > kernel at lists.infradead.org
> > Subject: Re: [PATCH V3 39/63] GIC: Added dummy handlers for Power
> > ManagementSuspend Resume
> >
> > On Mon, Dec 20, 2010 at 05:02:17PM +0530, viresh kumar wrote:
> > > On 12/20/2010 04:40 PM, Russell King - ARM Linux wrote:
> > > > And still this patch gets reposted a few more times despite my
> > > > objections:
> > > >
> > > >
> >
> http://lists.arm.linux.org.uk/lurker/message/20100920.150749.c97eda0d.en.h
> > tml
> > >
> > > Russell,
> > >
> > > Actually, when we discussed all this, we didn't came to any
> conclusion,
> > > and so i asked you: should we go ahead with this patch or drop it?
> >
> > Yes, I didn't bother replying any further because it seemed that no one
> > was listening to me.
> >
> > I think over the four or five emails my position on the patch was pretty
> > clear: I do _not_ like it one bit, and I still do not like it.
> >
> > It is a hack, plain and simple.  You're adding code to misrepresent what
> > the hardware can do.  You're fooling the system into thinking that the
> > GIC can control wake-up sources, when in fact the GIC has zero wakeup
> > capabilities what so ever.
> >
> > As I pointed out in the message above, if you do this, then drivers have
> > NO WAY to detect whether the interrupt controller they're connected to
> > is wake-up capable or not.
> >
> >
> http://lists.arm.linux.org.uk/lurker/message/20100920.134808.634d6ea1.en.h
> > tml
> >
> > I still don't know what your driver code looks like, yet I've given you
> > a suggestion to solve your problem in a subsequent reply (see the URL
> > at the top of this message) which never really got a reply from you.
> >
> > It seems to me that as soon as I asked for driver code, ST lost all
> > interest in discussing the issue any further, as there was no further
> > technical discussion coming from _any_ ST people.
> >
> Just for information, we did found a serial driver BUG
> is similar aspect. Below is the thread.
> http://ns3.spinics.net/lists/linux-omap/msg41240.html

So, the serial_core layer isn't properly tracking whether enable_irq_wake()
succeeded, and is then calling disable_irq_wake().  That's a bug in
the serial_core layer which needs fixing.