... and I immediately regret that I've killed the return value
for igt_debugfs_init, since we have callers which need to work
without the forcewake stuff, e.g. the reg dumper needs to work
without i915 loaded.
Put this new helper to good use in the mmio code and the pm_pc8
testcase.
Cc: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Big cores seem to take forever to idle, at least my ivb here. Fairly
ridiculous, so maybe there's more room for our debooster to kick in.
Anyway, this gets the min-max-config-idle and reset subtests going
somewhat reliably on my ivb. They still occasionally fail with the
current frequency pegged to max (or close to it) for no apparent
reason at all. Rebooting tends to fix it.
Don't ask, don't tell.
Cc: Jeff McGee <jeff.mcgee@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Big core platforms need some seriuos omph to break a sweat.
This fixes min-max-config-loaded here on my ivb.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=75146
Cc: Jeff McGee <jeff.mcgee@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
We don't call cleanup handlers when exiting a subtest currently, only
when exiting the entire binary. Which means pm_rps falls over when it
fails more than one subtest.
Cc: Jeff McGee <jeff.mcgee@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Apparently the framework doesn't ignore disabled pipes correctly, so
help it out a bit.
Again an awesome track record for our QA and bug scrubbers :(
Oh and while I rant: Docs for this stuff, please ;-)
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Instead of a width/height combination. Since I've been lazy with the
math this now only accepts page-aligned copy operations, but that's
all we need really.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Doesn't affect the test since we only check the first few bytes. But I
want to switch the copy code to always copy entire pages around, so
use the right buffer size.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Just launching empty batches isn't enough load on my speedy g33
to force the kernel to enable interrupts an wait for them.
So use some real load. With that the test is now fully reliably on
my g33 amchine.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Doing (1 << 18) - 1 repeats isn't really. With the reduced time the
test is also useful in some shell scripted tests which need some real
load (but not too much), like drv_missed_irq_hang.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Oops, I didn't check to see if there was more than one call site where
we built and execbuffer.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=75237
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
The clone-single-crtc will test various cloning scenarios.
Currently it does limit the number of simultanously cloned
connectors to the number of pipes in the system. That restriction
can be lifted in the future if desired.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
I used to have a binary that would just disable all the screens - so
we can enter PC8/runtime PM - and then sleep forever. I used this
binary many times while debugging PC8 and runtime PM, and I also sent
the binary to many people so they would be able to test these things
without X running.
Since pm_pc8 already implements everything that the separate binary
needs, and it even has some additional code to try to configure the
environment to actually reach PC8, it's easier to just ask people to
run "sudo ./pm_pc8 --stay" instead of sending them a file, asking them
to compile it, setup the environment, and then run it.
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
We allocate more surfaces than used in a single pass in order to stress
the eviction code between batches. The intent here is not to exercise
swapping, and we fail to check that there is enough swap+memory to hold
all our surfaces. So limit the number of surfaces we allocate to fit
into RAM, and then require that the number of surfaces we need for
testing is less than the number of surfaces we can allocate.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Currently the test suite just looks at the files provided by the
runtime power management framework to check if the device is runtime
suspended. Add a test that reads the PCI config space to check if the
device is actually in PCI D3 state or not.
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
After I finally fixed the code that WARNs if we're runtime suspended
when reading registers I started getting the WARNs, so this test
should reproduce them on a Kernel with the problem.
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
This is a render correctness test, the intention is not to exercise the
swapper (but preferrably some eviction code).
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=75247
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
This is a render correctness test, the intention is not to exercise the
swapper (but preferrably some eviction code).
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=75247
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
dpms-vs-vblank-race and modeset-vs-vblank-race try to race wait for
vblank ioctls from one thread against modeset/dpms ioctls from another.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Context banning can prevent the page flip hang tests from actaully
testing anything, so make the relevant subtests fail in that case.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Just in case it's compile with M instead of Y. If the module is not
there, the other assertions will catch the problem.
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Just in case the module is compiled with M instead of Y. If the module
is not there, the other assertions will catch the problem.
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
This sleep was added because sometimes we didn't reach PC8+
residencies, but it was still not enough to prevent the problem every
time, and it is really not needed most of the times. I have
investigated more and it seems that we only have to wait until after
some minutes have past since the machine booted. So just remove the
sleep for now since when you run each subtest in a separate process,
you end up having to sleep at every subtest.
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
On a bad day, hanging the GPU may be terminal. Yet even if the GPU is
terminally wedged we expect modesetting (and pageflips) to continue.
That deserves to be a dedicated test, but in the meantime we should
strive to avoid falling over just because the code is not resilient.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Skipping this test until we have Cairo in the build.
Signed-off-by: Joao Santos <joao.santos@intel.com>
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
For those who wonder about the -j switch ... all the good ones have
been taken, so this is the Jesse-flag.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
