| Message ID | 1413306342-23660-1-git-send-email-przanoni@gmail.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Tue, Oct 14, 2014 at 02:05:42PM -0300, Paulo Zanoni wrote: > From: Paulo Zanoni <paulo.r.zanoni@intel.com> > > When I look at BSpec, and at cursor_size_ok() (from the Kernel's > intel_display.c), I see that only 845g and i865g support non-square > displays, so SKIP the tests on HSW/BDW instead of failing them. > > Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=84268 > Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> > Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> http://lists.freedesktop.org/archives/intel-gfx/2014-September/052310.html Sadly I hit one small extra problem with the patches and didn't manage to send out revised versions before my new BSW arrived, at which point I had actual work to do again and couldn't play around with cursors :) Sometime soon I should hopefully have a few spare cycles to finish off that cursor stuff. > --- > tests/kms_cursor_crc.c | 5 +---- > 1 file changed, 1 insertion(+), 4 deletions(-) > > diff --git a/tests/kms_cursor_crc.c b/tests/kms_cursor_crc.c > index 92d1ce6..87b4f64 100644 > --- a/tests/kms_cursor_crc.c > +++ b/tests/kms_cursor_crc.c > @@ -371,10 +371,7 @@ static bool has_nonsquare_cursors(uint32_t devid) > * Test non-square cursors a bit on the platforms > * that support such things. > */ > - return devid == PCI_CHIP_845_G || > - devid == PCI_CHIP_I865_G || > - (IS_GEN7(devid) && !IS_VALLEYVIEW(devid)) || > - (IS_GEN8(devid) && !IS_CHERRYVIEW(devid)); > + return devid == PCI_CHIP_845_G || devid == PCI_CHIP_I865_G; > } > > static void test_cursor_size(data_t *data) > -- > 1.9.1
2014-10-14 16:32 GMT-03:00 Ville Syrjälä <ville.syrjala@linux.intel.com>: > On Tue, Oct 14, 2014 at 02:05:42PM -0300, Paulo Zanoni wrote: >> From: Paulo Zanoni <paulo.r.zanoni@intel.com> >> >> When I look at BSpec, and at cursor_size_ok() (from the Kernel's >> intel_display.c), I see that only 845g and i865g support non-square >> displays, so SKIP the tests on HSW/BDW instead of failing them. >> >> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=84268 >> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> >> Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> > > http://lists.freedesktop.org/archives/intel-gfx/2014-September/052310.html > > Sadly I hit one small extra problem with the patches and didn't manage > to send out revised versions before my new BSW arrived, at which point > I had actual work to do again and couldn't play around with cursors :) > > Sometime soon I should hopefully have a few spare cycles to finish off > that cursor stuff. So do we keep the bug open until there? Maybe it would be better to commit the IGT patch, close the bug, and then, when the variable-cursor-size support gets committed, update IGT. > >> --- >> tests/kms_cursor_crc.c | 5 +---- >> 1 file changed, 1 insertion(+), 4 deletions(-) >> >> diff --git a/tests/kms_cursor_crc.c b/tests/kms_cursor_crc.c >> index 92d1ce6..87b4f64 100644 >> --- a/tests/kms_cursor_crc.c >> +++ b/tests/kms_cursor_crc.c >> @@ -371,10 +371,7 @@ static bool has_nonsquare_cursors(uint32_t devid) >> * Test non-square cursors a bit on the platforms >> * that support such things. >> */ >> - return devid == PCI_CHIP_845_G || >> - devid == PCI_CHIP_I865_G || >> - (IS_GEN7(devid) && !IS_VALLEYVIEW(devid)) || >> - (IS_GEN8(devid) && !IS_CHERRYVIEW(devid)); >> + return devid == PCI_CHIP_845_G || devid == PCI_CHIP_I865_G; >> } >> >> static void test_cursor_size(data_t *data) >> -- >> 1.9.1 > > -- > Ville Syrjälä > Intel OTC
On Tue, Oct 14, 2014 at 02:05:42PM -0300, Paulo Zanoni wrote: > From: Paulo Zanoni <paulo.r.zanoni@intel.com> > > When I look at BSpec, and at cursor_size_ok() (from the Kernel's > intel_display.c), I see that only 845g and i865g support non-square > displays, so SKIP the tests on HSW/BDW instead of failing them. Why don't we just use the information provided by the kernel as to what it supports? -Chris
2014-10-14 16:50 GMT-03:00 Chris Wilson <chris@chris-wilson.co.uk>: > On Tue, Oct 14, 2014 at 02:05:42PM -0300, Paulo Zanoni wrote: >> From: Paulo Zanoni <paulo.r.zanoni@intel.com> >> >> When I look at BSpec, and at cursor_size_ok() (from the Kernel's >> intel_display.c), I see that only 845g and i865g support non-square >> displays, so SKIP the tests on HSW/BDW instead of failing them. > > Why don't we just use the information provided by the kernel as to what > it supports? (just documenting what we discussed on IRC yesterday) Apparently, there's no way to check if IVB+ has support for variable-size cursors with the current capability checks: the only way would be to check the return code from the cursor IOCTLs... > -Chris > > -- > Chris Wilson, Intel Open Source Technology Centre
2014-10-14 16:43 GMT-03:00 Paulo Zanoni <przanoni@gmail.com>: > 2014-10-14 16:32 GMT-03:00 Ville Syrjälä <ville.syrjala@linux.intel.com>: >> On Tue, Oct 14, 2014 at 02:05:42PM -0300, Paulo Zanoni wrote: >>> From: Paulo Zanoni <paulo.r.zanoni@intel.com> >>> >>> When I look at BSpec, and at cursor_size_ok() (from the Kernel's >>> intel_display.c), I see that only 845g and i865g support non-square >>> displays, so SKIP the tests on HSW/BDW instead of failing them. >>> >>> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=84268 >>> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> >>> Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> >> >> http://lists.freedesktop.org/archives/intel-gfx/2014-September/052310.html >> >> Sadly I hit one small extra problem with the patches and didn't manage >> to send out revised versions before my new BSW arrived, at which point >> I had actual work to do again and couldn't play around with cursors :) >> >> Sometime soon I should hopefully have a few spare cycles to finish off >> that cursor stuff. > > So do we keep the bug open until there? Maybe it would be better to > commit the IGT patch, close the bug, and then, when the > variable-cursor-size support gets committed, update IGT. Since no demonstrated to be opposed to the above plan, and I had to spend time debugging this, and we want to improve our bug statistics and IGT pass rate, I committed the patch and closed the bug. When we merge support for non-suqare cursor sizes, we can always update IGT accordingly. > >> >>> --- >>> tests/kms_cursor_crc.c | 5 +---- >>> 1 file changed, 1 insertion(+), 4 deletions(-) >>> >>> diff --git a/tests/kms_cursor_crc.c b/tests/kms_cursor_crc.c >>> index 92d1ce6..87b4f64 100644 >>> --- a/tests/kms_cursor_crc.c >>> +++ b/tests/kms_cursor_crc.c >>> @@ -371,10 +371,7 @@ static bool has_nonsquare_cursors(uint32_t devid) >>> * Test non-square cursors a bit on the platforms >>> * that support such things. >>> */ >>> - return devid == PCI_CHIP_845_G || >>> - devid == PCI_CHIP_I865_G || >>> - (IS_GEN7(devid) && !IS_VALLEYVIEW(devid)) || >>> - (IS_GEN8(devid) && !IS_CHERRYVIEW(devid)); >>> + return devid == PCI_CHIP_845_G || devid == PCI_CHIP_I865_G; >>> } >>> >>> static void test_cursor_size(data_t *data) >>> -- >>> 1.9.1 >> >> -- >> Ville Syrjälä >> Intel OTC > > > > -- > Paulo Zanoni
diff --git a/tests/kms_cursor_crc.c b/tests/kms_cursor_crc.c index 92d1ce6..87b4f64 100644 --- a/tests/kms_cursor_crc.c +++ b/tests/kms_cursor_crc.c @@ -371,10 +371,7 @@ static bool has_nonsquare_cursors(uint32_t devid) * Test non-square cursors a bit on the platforms * that support such things. */ - return devid == PCI_CHIP_845_G || - devid == PCI_CHIP_I865_G || - (IS_GEN7(devid) && !IS_VALLEYVIEW(devid)) || - (IS_GEN8(devid) && !IS_CHERRYVIEW(devid)); + return devid == PCI_CHIP_845_G || devid == PCI_CHIP_I865_G; } static void test_cursor_size(data_t *data)