Quoting Tvrtko Ursulin (2018-12-06 13:12:35)
On 06/12/2018 08:44, Chris Wilson wrote:
Braswell is really picky about having our writes posted to memory before we execute or else the GPU may see stale values. A wmb() is insufficient as it only ensures the writes are visible to other cores, we need a full mb() to ensure the writes are in memory and visible to the GPU.
The most frequent failure in flushing before execution is that we see stale PTE values and execute the wrong pages.
References: 987abd5c62f9 ("drm/i915/execlists: Force write serialisation into context image vs execution") Signed-off-by: Chris Wilson chris@chris-wilson.co.uk Cc: Mika Kuoppala mika.kuoppala@linux.intel.com Cc: Tvrtko Ursulin tvrtko.ursulin@intel.com Cc: Joonas Lahtinen joonas.lahtinen@linux.intel.com Cc: stable@vger.kernel.org
drivers/gpu/drm/i915/intel_lrc.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index de1e9dc6aec0..e6a86fa4502d 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -379,8 +379,13 @@ static u64 execlists_update_context(struct i915_request *rq) * may not be visible to the HW prior to the completion of the UC * register write and that we may begin execution from the context * before its image is complete leading to invalid PD chasing.
** Furthermore, Braswell, at least, wants a full mb to be sure that* the writes are coherent in memory (visible to the GPU) prior to* execution, and not just visible to other CPUs (as is the result of* wmb). */
wmb();
mb(); return ce->lrc_desc;Too low level for me to really know what happens under the hood, but at least I know it can't break anything.
The alternative I'm considering is using a mmio read instead. However, the improvement in stability from switching to mb() here is already enough to proceed without necessarily finding the ideal solution. -Chris