gpu: nvgpu: fix bug in CHANNEL_FORCE_RESET

Channels belonging to a TSG did not have their error notifier
set correctly. This was due to using an incorrect TSG id.

Bug 1617046

Change-Id: Icb6911c7d79a9d02d7713bb47a7cbb24c3098dc1
Signed-off-by: Aingara Paramakuru <aparamakuru@nvidia.com>
Reviewed-on: http://git-master/r/1155293
Reviewed-by: Terje Bergstrom <tbergstrom@nvidia.com>
Tested-by: Terje Bergstrom <tbergstrom@nvidia.com>

1 files changed, 1 insertions, 1 deletions

diff --git a/drivers/gpu/nvgpu/gk20a/fifo_gk20a.c b/drivers/gpu/nvgpu/gk20a/fifo_gk20a.c
index 134a2480..2f7cb888 100644
--- a/drivers/gpu/nvgpu/gk20a/fifo_gk20a.c
+++ b/drivers/gpu/nvgpu/gk20a/fifo_gk20a.c
@@ -1419,7 +1419,7 @@ int gk20a_fifo_force_reset_ch(struct channel_gk20a *ch, bool verbose)
         struct gk20a *g = ch->g;
 
         if (gk20a_is_channel_marked_as_tsg(ch)) {
-                tsg = &g->fifo.tsg[ch->hw_chid];
+                tsg = &g->fifo.tsg[ch->tsgid];
 
                 mutex_lock(&tsg->ch_list_lock);