gpu: nvgpu: Add timeouts_disabled_refcount for enabling timeout

-timeouts will be enabled only when timeouts_disabled_refcount
 will reach 0
-timeouts_enabled debugfs will change from u32 type to file type
 to avoid race enabling/disabling timeout from debugfs and ioctl
-unify setting timeouts_enabled from debugfs and ioctl

Bug 1982434

Change-Id: I54bab778f1ae533872146dfb8d80deafd2a685c7
Signed-off-by: Seema Khowala <seemaj@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1588690
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>

3 files changed, 13 insertions, 7 deletions

diff --git a/drivers/gpu/nvgpu/gk20a/channel_gk20a.c b/drivers/gpu/nvgpu/gk20a/channel_gk20a.c
index a4637b8f..45ce0e77 100644
--- a/drivers/gpu/nvgpu/gk20a/channel_gk20a.c
+++ b/drivers/gpu/nvgpu/gk20a/channel_gk20a.c
@@ -1158,7 +1158,7 @@ int gk20a_channel_alloc_gpfifo(struct channel_gk20a *c,
                 }
         }
 
-        if (!c->g->timeouts_enabled || !c->timeout.enabled)
+        if (!nvgpu_is_timeouts_enabled(c->g) || !c->timeout.enabled)
                 acquire_timeout = 0;
         else
                 acquire_timeout = c->timeout.limit_ms;
@@ -1266,7 +1266,7 @@ bool gk20a_channel_update_and_check_timeout(struct channel_gk20a *ch,
 
         ch->timeout_gpfifo_get = gpfifo_get;
 
-        return ch->g->timeouts_enabled &&
+        return nvgpu_is_timeouts_enabled(ch->g) &&
                 ch->timeout_accumulated_ms > ch->timeout_ms_max;
 }
 
@@ -1303,7 +1303,7 @@ static void __gk20a_channel_timeout_start(struct channel_gk20a *ch)
  */
 static void gk20a_channel_timeout_start(struct channel_gk20a *ch)
 {
-        if (!ch->g->timeouts_enabled)
+        if (!nvgpu_is_timeouts_enabled(ch->g))
                 return;
 
         if (!ch->timeout.enabled)
diff --git a/drivers/gpu/nvgpu/gk20a/fifo_gk20a.c b/drivers/gpu/nvgpu/gk20a/fifo_gk20a.c
index 7f4a0948..964ccb03 100644
--- a/drivers/gpu/nvgpu/gk20a/fifo_gk20a.c
+++ b/drivers/gpu/nvgpu/gk20a/fifo_gk20a.c
@@ -2482,7 +2482,7 @@ unsigned int gk20a_fifo_handle_pbdma_intr_0(struct gk20a *g, u32 pbdma_id,
 
                 val &= ~pbdma_acquire_timeout_en_enable_f();
                 gk20a_writel(g, pbdma_acquire_r(pbdma_id), val);
-                if (g->timeouts_enabled) {
+                if (nvgpu_is_timeouts_enabled(g)) {
                         rc_type = RC_TYPE_PBDMA_FAULT;
                         nvgpu_err(g,
                                 "semaphore acquire timeout!");
diff --git a/drivers/gpu/nvgpu/gk20a/gk20a.h b/drivers/gpu/nvgpu/gk20a/gk20a.h
index 84d3f639..f6318257 100644
--- a/drivers/gpu/nvgpu/gk20a/gk20a.h
+++ b/drivers/gpu/nvgpu/gk20a/gk20a.h
@@ -1314,7 +1314,7 @@ struct gk20a {
         struct railgate_stats pstats;
 #endif
         u32 gr_idle_timeout_default;
-        bool timeouts_enabled;
+        bool timeouts_disabled_by_user;
         unsigned int ch_wdt_timeout_ms;
 
         struct nvgpu_mutex poweron_lock;
@@ -1376,7 +1376,8 @@ struct gk20a {
         /* also prevents debug sessions from attaching until released */
         struct nvgpu_mutex dbg_sessions_lock;
         int dbg_powergating_disabled_refcount; /*refcount for pg disable */
-        int dbg_timeout_disabled_refcount; /*refcount for timeout disable */
+        /*refcount for timeout disable */
+        nvgpu_atomic_t timeouts_disabled_refcount;
 
         /* must have dbg_sessions_lock before use */
         struct nvgpu_dbg_reg_op *dbg_regops_tmp_buf;
@@ -1508,9 +1509,14 @@ struct gk20a {
         struct nvgpu_list_node boardobjgrp_head;
 };
 
+static inline bool nvgpu_is_timeouts_enabled(struct gk20a *g)
+{
+        return nvgpu_atomic_read(&g->timeouts_disabled_refcount) == 0;
+}
+
 static inline unsigned long gk20a_get_gr_idle_timeout(struct gk20a *g)
 {
-        return g->timeouts_enabled ?
+        return nvgpu_is_timeouts_enabled(g) ?
                 g->gr_idle_timeout_default : ULONG_MAX;
 }