gpu: nvgpu: restore golden context without bind

Copy and restore golden context correctly with
context header. Removed parallel fecs bind method,
which can cause issues for context in execution.

Also added function pointer to freeing context
header during channel context free.

Bug 1834201

Change-Id: I7962d68338d5144f624375ab81436e86cb31051e
Signed-off-by: seshendra Gadagottu <sgadagottu@nvidia.com>
Reviewed-on: http://git-master/r/1275201
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>

2 files changed, 24 insertions, 20 deletions

diff --git a/drivers/gpu/nvgpu/gk20a/gk20a.h b/drivers/gpu/nvgpu/gk20a/gk20a.h
index 1dd42dc2..18e99d6f 100644
--- a/drivers/gpu/nvgpu/gk20a/gk20a.h
+++ b/drivers/gpu/nvgpu/gk20a/gk20a.h
@@ -332,6 +332,7 @@ struct gpu_ops {
                 int (*commit_global_timeslice)(struct gk20a *g,
                                         struct channel_gk20a *c, bool patch);
                 int (*commit_inst)(struct channel_gk20a *c, u64 gpu_va);
+                void (*restore_context_header)(struct gk20a *g, struct mem_desc *ctxheader);
         } gr;
         const char *name;
         struct {
@@ -420,6 +421,7 @@ struct gpu_ops {
                                         u32 *runlist);
                 u32 (*userd_gp_get)(struct gk20a *g, struct channel_gk20a *ch);
                 void (*userd_gp_put)(struct gk20a *g, struct channel_gk20a *ch);
+                void (*free_channel_ctx_header)(struct channel_gk20a *ch);
         } fifo;
         struct pmu_v {
                 /*used for change of enum zbc update cmd id from ver 0 to ver1*/
diff --git a/drivers/gpu/nvgpu/gk20a/gr_gk20a.c b/drivers/gpu/nvgpu/gk20a/gr_gk20a.c
index be8894d2..48b825a1 100644
--- a/drivers/gpu/nvgpu/gk20a/gr_gk20a.c
+++ b/drivers/gpu/nvgpu/gk20a/gr_gk20a.c
@@ -1571,18 +1571,6 @@ static int gr_gk20a_init_golden_ctx_image(struct gk20a *g,
         mutex_lock(&gr->ctx_mutex);
 
         if (gr->ctx_vars.golden_image_initialized) {
-                if (gk20a_mem_begin(g, ctxheader))
-                        return -ENOMEM;
-                if (ctxheader->gpu_va) {
-                        err = gr_gk20a_fecs_ctx_bind_channel(g, c);
-                        if (err)
-                                goto clean_up;
-
-                        err = gr_gk20a_wait_idle(g,
-                                                 gk20a_get_gr_idle_timeout(g),
-                                                 GR_IDLE_CHECK_DEFAULT);
-                }
-                gk20a_mem_end(g, ctxheader);
                 goto clean_up;
         }
         if (!platform->is_fmodel) {
@@ -1753,7 +1741,6 @@ restore_fe_go_idle:
                 data = gk20a_mem_rd32(g, gr_mem, i);
                 gk20a_mem_wr32(g, gold_mem, i, data);
         }
-
         gk20a_mem_wr(g, gold_mem, ctxsw_prog_main_image_zcull_o(),
                  ctxsw_prog_main_image_zcull_mode_no_ctxsw_v());
 
@@ -1763,6 +1750,9 @@ restore_fe_go_idle:
 
         gr_gk20a_fecs_ctx_image_save(c, gr_fecs_method_push_adr_wfi_golden_save_v());
 
+        if (gk20a_mem_begin(g, ctxheader))
+                goto clean_up;
+
         if (gr->ctx_vars.local_golden_image == NULL) {
 
                 gr->ctx_vars.local_golden_image =
@@ -1773,10 +1763,16 @@ restore_fe_go_idle:
                         goto clean_up;
                 }
 
-                gk20a_mem_rd_n(g, gold_mem, 0,
+                if (ctxheader->gpu_va)
+                        gk20a_mem_rd_n(g, ctxheader, 0,
+                                gr->ctx_vars.local_golden_image,
+                                gr->ctx_vars.golden_image_size);
+                else
+                        gk20a_mem_rd_n(g, gold_mem, 0,
                                 gr->ctx_vars.local_golden_image,
                                 gr->ctx_vars.golden_image_size);
         }
+        gk20a_mem_end(g, ctxheader);
 
         g->ops.gr.commit_inst(c, gr_mem->gpu_va);
 
@@ -2016,7 +2012,6 @@ int gr_gk20a_load_golden_ctx_image(struct gk20a *g,
         if (gr->ctx_vars.local_golden_image == NULL)
                 return -1;
 
-
         /* Channel gr_ctx buffer is gpu cacheable.
            Flush and invalidate before cpu update. */
         g->ops.mm.l2_flush(g, true);
@@ -2029,15 +2024,22 @@ int gr_gk20a_load_golden_ctx_image(struct gk20a *g,
                 goto clean_up_mem;
         }
 
-        gk20a_mem_wr_n(g, mem, 0,
+        if (ctxheader->gpu_va) {
+                if (g->ops.gr.restore_context_header)
+                        g->ops.gr.restore_context_header(g, ctxheader);
+        } else {
+                gk20a_mem_wr_n(g, mem, 0,
                         gr->ctx_vars.local_golden_image,
                         gr->ctx_vars.golden_image_size);
+                gk20a_mem_wr(g, mem,
+                        ctxsw_prog_main_image_num_save_ops_o(), 0);
+                gk20a_mem_wr(g, mem,
+                        ctxsw_prog_main_image_num_restore_ops_o(), 0);
+        }
 
         if (g->ops.gr.enable_cde_in_fecs && c->cde)
                 g->ops.gr.enable_cde_in_fecs(g, mem);
 
-        gk20a_mem_wr(g, mem, ctxsw_prog_main_image_num_save_ops_o(), 0);
-        gk20a_mem_wr(g, mem, ctxsw_prog_main_image_num_restore_ops_o(), 0);
         /* set priv access map */
         virt_addr_lo =
                  u64_lo32(ch_ctx->global_ctx_buffer_va[PRIV_ACCESS_MAP_VA]);
@@ -2145,9 +2147,7 @@ int gr_gk20a_load_golden_ctx_image(struct gk20a *g,
                         ctxsw_prog_main_image_pm_ptr_o(), virt_addr);
         }
 
-        gk20a_mem_end(g, mem);
         gk20a_mem_end(g, ctxheader);
-
 clean_up_mem:
         gk20a_mem_end(g, mem);
 
@@ -3065,6 +3065,8 @@ static void gr_gk20a_free_channel_pm_ctx(struct channel_gk20a *c)
 
 void gk20a_free_channel_ctx(struct channel_gk20a *c)
 {
+        if(c->g->ops.fifo.free_channel_ctx_header)
+                c->g->ops.fifo.free_channel_ctx_header(c);
         gr_gk20a_unmap_global_ctx_buffers(c);
         gr_gk20a_free_channel_patch_ctx(c);
         gr_gk20a_free_channel_pm_ctx(c);