When cancelling requests, also send the notification to any listeners
(gvt) that the request is no longer scheduler on hw. They may require to
keep the in/out exactly balanced, and so the reuse after the reset may
confuse the listener.
Fixes: 221ab9719bf3 ("drm/i915/execlists: Unwind incomplete requests on resets")
Signed-off-by: Chris Wilson <ch...@chris-wilson.co.uk>
Cc: "Zhenyu Wang" <zhen...@linux.intel.com>
Cc: "Wang, Zhi A" <zhi.a.w...@intel.com>
Cc: Michał Winiarski <michal.winiar...@intel.com>
Cc: Mika Kuoppala <mika.kuopp...@linux.intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursu...@intel.com>
---
drivers/gpu/drm/i915/intel_lrc.c | 6 +++++-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c
index 2c07f3c08bd3..61cac26a8b05 100644
--- a/drivers/gpu/drm/i915/intel_lrc.c
+++ b/drivers/gpu/drm/i915/intel_lrc.c
@@ -578,7 +578,11 @@ execlist_cancel_port_requests(struct
intel_engine_execlists *execlists)
unsigned int num_ports = ARRAY_SIZE(execlists->port);
while (num_ports-- && port_isset(port)) {
- i915_gem_request_put(port_request(port));
+ struct drm_i915_gem_request *rq = port_request(port);
+
+ execlists_context_status_change(rq, INTEL_CONTEXT_SCHEDULE_OUT);
+ i915_gem_request_put(rq);
+
memset(port, 0, sizeof(*port));
port++;
}
--
2.14.1
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
