Quoting Joonas Lahtinen (2019-10-28 12:13:05)
> Quoting Chris Wilson (2019-10-27 17:03:13)
> > As we use hard coded offsets for a few locations within the context
> > image, include those in the selftests to assert that they are valid.
> >
> > Signed-off-by: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx>
> > Cc: Mika Kuoppala <mika.kuoppala@xxxxxxxxxxxxxxx>
>
> <SNIP>
>
> > +++ b/drivers/gpu/drm/i915/gt/selftest_lrc.c
> > @@ -3194,6 +3194,26 @@ static int live_lrc_fixed(void *arg)
> > u32 offset;
> > const char *name;
> > } tbl[] = {
> > + {
> > + i915_mmio_reg_offset(RING_START(engine->mmio_base)),
> > + CTX_RING_BUFFER_START - 1,
> > + "RING_START"
> > + },
> > + {
> > + i915_mmio_reg_offset(RING_CTL(engine->mmio_base)),
> > + CTX_RING_BUFFER_CONTROL - 1,
> > + "RING_CTL"
> > + },
> > + {
> > + i915_mmio_reg_offset(RING_HEAD(engine->mmio_base)),
> > + CTX_RING_HEAD - 1,
> > + "RING_HEAD"
> > + },
> > + {
> > + i915_mmio_reg_offset(RING_TAIL(engine->mmio_base)),
> > + CTX_RING_TAIL - 1,
> > + "RING_TAIL"
> > + },
>
> Quite repetitive. If you are not looking to add dissimilar patterns,
> could use a #define, but up to you.
The tables spares us even more repetition :)
I didn't look too hard at reducing the repetitiveness as I was expecting
a few more registers to break the mould. We shall see what the future
brings.
There's a few more registers we use hardcoded locations for, mostly in
perf. Hmm, I should probably make that a local test for selftest_perf.c
so that the logic and test are close together.
-Chris
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
