gpu: nvgpu: preempt before adjusting fences

Current sequence in gk20a_disable_channel() is
- disable channel in gk20a_channel_abort()
- adjust pending fence in gk20a_channel_abort()
- preempt channel

But this leads to scenarios where syncpoint has
min > max value

Hence to fix this, make sequence in gk20a_disable_channel()
- disable channel in gk20a_channel_abort()
- preempt channel in gk20a_channel_abort()
- adjust pending fence in gk20a_channel_abort()

If gk20a_channel_abort() is called from other API where
preemption is not needed, then use channel_preempt
flag and do not preempt channel in those cases

Bug 1683059

Change-Id: I4d46d4294cf8597ae5f05f79dfe1b95c4187f2e3
Signed-off-by: Deepak Nibade <dnibade@nvidia.com>
Reviewed-on: http://git-master/r/921290
Reviewed-by: Terje Bergstrom <tbergstrom@nvidia.com>
Tested-by: Terje Bergstrom <tbergstrom@nvidia.com>

1 files changed, 6 insertions, 4 deletions

diff --git a/drivers/gpu/nvgpu/gk20a/channel_gk20a.c b/drivers/gpu/nvgpu/gk20a/channel_gk20a.c
index 7ec5ade4..b480c80a 100644
--- a/drivers/gpu/nvgpu/gk20a/channel_gk20a.c
+++ b/drivers/gpu/nvgpu/gk20a/channel_gk20a.c
@@ -392,7 +392,7 @@ void channel_gk20a_disable(struct channel_gk20a *ch)
                         ccsr_channel_enable_clr_true_f());
 }
 
-void gk20a_channel_abort(struct channel_gk20a *ch)
+void gk20a_channel_abort(struct channel_gk20a *ch, bool channel_preempt)
 {
         struct channel_gk20a_job *job, *n;
         bool released_job_semaphore = false;
@@ -404,6 +404,9 @@ void gk20a_channel_abort(struct channel_gk20a *ch)
 
         ch->g->ops.fifo.disable_channel(ch);
 
+        if (channel_preempt)
+                ch->g->ops.fifo.preempt_channel(ch->g, ch->hw_chid);
+
         /* ensure no fences are pending */
         mutex_lock(&ch->sync_lock);
         if (ch->sync)
@@ -455,8 +458,7 @@ int gk20a_wait_channel_idle(struct channel_gk20a *ch)
 
 void gk20a_disable_channel(struct channel_gk20a *ch)
 {
-        gk20a_channel_abort(ch);
-        ch->g->ops.fifo.preempt_channel(ch->g, ch->hw_chid);
+        gk20a_channel_abort(ch, true);
         channel_gk20a_update_runlist(ch, false);
 }
 
@@ -1621,7 +1623,7 @@ static void gk20a_channel_timeout_handler(struct work_struct *work)
                         gk20a_fifo_abort_tsg(g, ch->tsgid);
                 } else {
                         gk20a_fifo_set_ctx_mmu_error_ch(g, ch);
-                        gk20a_channel_abort(ch);
+                        gk20a_channel_abort(ch, false);
                 }
         } else {
                 /* If failing engine, trigger recovery */