gpu: nvgpu: Fix TLB invalidate race

TLB invalidate can have a race if several contexts use the same
address space. One thread starting an invalidate allows another
thread to submit before invalidate is completed.

Bug 1502332

Change-Id: I074ec493eac3b153c5f23d796a1dee1d8db24855
Signed-off-by: Terje Bergstrom <tbergstrom@nvidia.com>
Reviewed-on: http://git-master/r/407578
Reviewed-by: Riham Haidar <rhaidar@nvidia.com>
Tested-by: Riham Haidar <rhaidar@nvidia.com>

1 files changed, 0 insertions, 7 deletions

diff --git a/drivers/gpu/nvgpu/gk20a/channel_gk20a.c b/drivers/gpu/nvgpu/gk20a/channel_gk20a.c
index 1b7191db..7cfb3c24 100644
--- a/drivers/gpu/nvgpu/gk20a/channel_gk20a.c
+++ b/drivers/gpu/nvgpu/gk20a/channel_gk20a.c
@@ -1575,13 +1575,6 @@ static int gk20a_submit_channel_gpfifo(struct channel_gk20a *c,
                 incr_cmd->gp_put = c->gpfifo.put;
         }
 
-        /* Invalidate tlb if it's dirty...                                   */
-        /* TBD: this should be done in the cmd stream, not with PRIs.        */
-        /* We don't know what context is currently running...                */
-        /* Note also: there can be more than one context associated with the */
-        /* address space (vm).   */
-        gk20a_mm_tlb_invalidate(c->vm);
-
         trace_gk20a_channel_submitted_gpfifo(c->g->dev->name,
                                              c->hw_chid,
                                              num_entries,