[Intel-gfx] [PATCH igt] igt/pm_rpm: Wait for PCI D3

[Chris Wilson]

On Mon, Mar 07, 2016 at 04:41:55PM +0100, Patrik Jakobsson wrote:
> On Mon, Mar 07, 2016 at 02:19:10PM +0000, Chris Wilson wrote:
> > The transition to PCI device state D3 is not instantaneous and only
> > started when runtime suspended. Allow the driver/hardware a little bit
> > of time to complete the transition before declaring a test failure.
> > 
> > References: https://bugs.freedesktop.org/show_bug.cgi?id=93123
> > Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
> > ---
> >  tests/pm_rpm.c | 12 ++++--------
> >  1 file changed, 4 insertions(+), 8 deletions(-)
> > 
> > diff --git a/tests/pm_rpm.c b/tests/pm_rpm.c
> > index e84a9d3..03ef2eb 100644
> > --- a/tests/pm_rpm.c
> > +++ b/tests/pm_rpm.c
> > @@ -1337,15 +1337,13 @@ static void reg_read_ioctl_subtest(void)
> >  
> >  static bool device_in_pci_d3(void)
> >  {
> > -	struct pci_device *pci_dev;
> > -	int rc;
> >  	uint16_t val;
> > +	int rc;
> >  
> > -	pci_dev = intel_get_pci_device();
> > -
> > -	rc = pci_device_cfg_read_u16(pci_dev, &val, 0xd4);
> > +	rc = pci_device_cfg_read_u16(intel_get_pci_device(), &val, 0xd4);
> >  	igt_assert_eq(rc, 0);
> >  
> > +	igt_debug("%s: PCI D3 state=%d\n", __func__, val & 0x3);
> >  	return (val & 0x3) == 0x3;
> >  }
> >  
> > @@ -1354,11 +1352,9 @@ static void pci_d3_state_subtest(void)
> >  	igt_require(has_runtime_pm);
> >  
> >  	disable_all_screens_and_wait(&ms_data);
> > -
> > -	igt_assert(device_in_pci_d3());
> > +	igt_assert(igt_wait(device_in_pci_d3(), 2000, 100));
> 
> Not sure we need to wait for 2 full secs here but not a big problem if we do.

2s was just a smallish number larger than what seemed reasonable for the
driver. It seems the longest wait may just be 120ms for certain Intel PCI
Express chipsets (drivers/pci/quirks.c) - so a safety factor of 10
before we declare the test a failure.
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre