Re: [RFT/PATCH] serial: omap: prevent resume if device is not suspended.

[Felipe Balbi <balbi@ti.com>]

[-- Attachment #1: Type: text/plain, Size: 2331 bytes --]

On Wed, Sep 19, 2012 at 02:52:18PM +0300, Grazvydas Ignotas wrote:
> On Tue, Sep 18, 2012 at 5:02 PM, Felipe Balbi <balbi@ti.com> wrote:
> > On Tue, Sep 18, 2012 at 06:10:50PM +0530, Sourav Poddar wrote:
> >> Greg's tty-next is not booting on 2420 based N800. The failure is
> >> observed at serial init itself. The reason might be that n800 tries to
> >> resume even though it is not suspended before.
> >>
> >> Reported-by: Paul Walmsley <paul@pwsan.com>
> >> Signed-off-by: Sourav Poddar <sourav.poddar@ti.com>
> >
> > FWIW:
> >
> > Reviewed-by: Felipe Balbi <balbi@ti.com>
> >
> > Paul does this fix the issue for you ? Note that it depends on a
> > previous patch Sourav sent [1]
> >
> > [1] http://marc.info/?l=linux-omap&m=134796819607889&w=2
> >
> > There's one thing that gets my attention, though, why only n800 would
> > fail here ?
> >
> > I wonder if we should be using:
> >
> > pm_runtime_set_active(dev);
> > pm_runtime_get_enable(dev);
> >
> > to prevent our runtime_resume() to be called from probe, but Sourav
> > tested and it doesn't work on BeagleBoard, but it works on PandaBoard.
> >
> > Does it ring any bell ??
> 
> Well I guess it's normal resume callback is called during probe as
> pm_runtime_get_sync() is called there, and according to documentation
> [1], devices are assumed to be suspended before probe is called.
> 
> According to [1], pm_runtime_set_active() should be called before
> pm_runtime_enable() to indicate to the core that device is active
> during probe if that's the case, I guess this means
> pm_runtime_get_sync() is not needed in that case, but I'm not sure
> what's the actual serial state during probe nowadays.
> 
> [1] chapter 5 of Documentation/power/runtime_pm.txt:
> The initial runtime PM status of all devices is 'suspended', but it
> need not reflect the actual physical state of the device. Thus, if the
> device is initially active (i.e. it is able to process I/O), its
> runtime PM status must be changed to 'active', with the help of
> pm_runtime_set_active(), before pm_runtime_enable() is called for the
> device.

this is what I mean, actually. If we remove pm_runtime_get_sync() in
exchange for pm_runtime_set_active() before pm_runtime_enable(), it
works on PandaBoard, but breaks BeagleBoard.

-- 
balbi

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 836 bytes --]