gpu: nvgpu: More robust recovery

Make recovery a more straightforward process. When we detect a fault,
trigger MMU fault, and wait for it to trigger, and complete recovery.
Also reset engines before aborting channel to ensure no stray sync
point increments can happen.

Change-Id: Iac685db6534cb64fe62d9fb452391f43100f2999
Signed-off-by: Terje Bergstrom <tbergstrom@nvidia.com>
Reviewed-on: http://git-master/r/709060
(cherry picked from commit 95c62ffd9ac30a0d2eb88d033dcc6e6ff25efd6f)
Reviewed-on: http://git-master/r/707443

1 files changed, 1 insertions, 0 deletions

diff --git a/drivers/gpu/nvgpu/gk20a/fifo_gk20a.h b/drivers/gpu/nvgpu/gk20a/fifo_gk20a.h
index ecae970f..8fda38f5 100644
--- a/drivers/gpu/nvgpu/gk20a/fifo_gk20a.h
+++ b/drivers/gpu/nvgpu/gk20a/fifo_gk20a.h
@@ -162,6 +162,7 @@ void gk20a_fifo_recover(struct gk20a *g, u32 engine_ids, bool verbose);
 void gk20a_fifo_recover_ch(struct gk20a *g, u32 hw_chid, bool verbose);
 void gk20a_fifo_recover_tsg(struct gk20a *g, u32 tsgid, bool verbose);
 int gk20a_fifo_force_reset_ch(struct channel_gk20a *ch, bool verbose);
+void gk20a_fifo_reset_engine(struct gk20a *g, u32 engine_id);
 int gk20a_init_fifo_reset_enable_hw(struct gk20a *g);
 void gk20a_init_fifo(struct gpu_ops *gops);