gpu: nvgpu: move dump_ctxsw_stats_on_channel_close flag to gr_ctx_desc

Debug boolean flag dump_ctxsw_stats_on_channel_close is right now stored
in gr_gk20a.ctx_vars struct
This flag logically is property of gr.ctx units since it indicates
whether each context should dump ctxsw stats on channel/context close

Move this flag to struct nvgpu_gr_ctx_desc and remove it from
gr_gk20a.ctx_vars

Expose below API from gr.ctx unit to check if flag is set
nvgpu_gr_ctx_desc_dump_ctxsw_stats_on_channel_close()

Move debugfs creation code to create corresponding debugfs to
gr_gk20a_debugfs_init() and change debugfs type from "u32" to "file"

Struct gr.gr_ctx_desc is created only during first poweron.
Return error if this struct is not available.

Remove unnecessary initialization of this variable from platform
specific probe functions

Jira NVGPU-3112

Change-Id: Id675e047237f82e9b8198a42082e99c95824578f
Signed-off-by: Deepak Nibade <dnibade@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2099399
Reviewed-by: svc-mobile-coverity <svc-mobile-coverity@nvidia.com>
Reviewed-by: svc-mobile-misra <svc-mobile-misra@nvidia.com>
GVS: Gerrit_Virtual_Submit
Reviewed-by: Vinod Gopalakrishnakurup <vinodg@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>

drivers/gpu/nvgpu/common/gr/ctx.c

@@ -969,3 +969,9 @@ bool nvgpu_gr_ctx_desc_force_preemption_cilp(struct nvgpu_gr_ctx_desc *gr_ctx_de
 {
 	return gr_ctx_desc->force_preemption_cilp;
 }
+
+bool nvgpu_gr_ctx_desc_dump_ctxsw_stats_on_channel_close(
+		struct nvgpu_gr_ctx_desc *gr_ctx_desc)
+{
+	return gr_ctx_desc->dump_ctxsw_stats_on_channel_close;
+}

drivers/gpu/nvgpu/common/gr/ctx_priv.h

@@ -47,6 +47,8 @@ struct nvgpu_gr_ctx_desc {
 
 	bool force_preemption_gfxp;
 	bool force_preemption_cilp;
+
+	bool dump_ctxsw_stats_on_channel_close;
 };
 
 struct nvgpu_gr_ctx {

drivers/gpu/nvgpu/common/gr/gr_setup.c

@@ -181,7 +181,8 @@ void nvgpu_gr_setup_free_gr_ctx(struct gk20a *g,
 
 	if (gr_ctx != NULL) {
 		if ((g->ops.gr.ctxsw_prog.dump_ctxsw_stats != NULL) &&
-		     g->gr->ctx_vars.dump_ctxsw_stats_on_channel_close) {
+		     nvgpu_gr_ctx_desc_dump_ctxsw_stats_on_channel_close(
+					g->gr->gr_ctx_desc)) {
 			g->ops.gr.ctxsw_prog.dump_ctxsw_stats(g,
 				 nvgpu_gr_ctx_get_ctx_mem(gr_ctx));
 		}

drivers/gpu/nvgpu/gk20a/gr_gk20a.h

@@ -91,8 +91,6 @@ struct nvgpu_gr {
 		u32 preempt_image_size;
 
 		u32 zcull_image_size;
-
-		bool dump_ctxsw_stats_on_channel_close;
 	} ctx_vars;
 
 	struct nvgpu_mutex ctx_mutex; /* protect golden ctx init */

drivers/gpu/nvgpu/include/nvgpu/gr/ctx.h

@@ -236,4 +236,7 @@ bool nvgpu_gr_ctx_desc_force_preemption_gfxp(
 bool nvgpu_gr_ctx_desc_force_preemption_cilp(
 		struct nvgpu_gr_ctx_desc *gr_ctx_desc);
 
+bool nvgpu_gr_ctx_desc_dump_ctxsw_stats_on_channel_close(
+		struct nvgpu_gr_ctx_desc *gr_ctx_desc);
+
 #endif /* NVGPU_INCLUDE_GR_CTX_H */

drivers/gpu/nvgpu/os/linux/debug.c

@@ -411,11 +411,6 @@ void gk20a_debug_init(struct gk20a *g, const char *debugfs_symlink)
 					l->debugfs,
 					&g->runlist_interleave);
 
-	l->debugfs_dump_ctxsw_stats =
-		debugfs_create_bool("dump_ctxsw_stats_on_channel_close",
-			S_IRUGO|S_IWUSR, l->debugfs,
-			&g->gr->ctx_vars.dump_ctxsw_stats_on_channel_close);
-
 	gr_gk20a_debugfs_init(g);
 	gk20a_pmu_debugfs_init(g);
 	gk20a_railgating_debugfs_init(g);

drivers/gpu/nvgpu/os/linux/debug_gr.c

@@ -149,6 +149,58 @@ static struct file_operations force_preemption_cilp_fops = {
 	.write =	force_preemption_cilp_write,
 };
 
+static ssize_t dump_ctxsw_stats_on_channel_close_read(struct file *file,
+		char __user *user_buf, size_t count, loff_t *ppos)
+{
+	char buf[3];
+	struct gk20a *g = file->private_data;
+
+	if (g->gr->gr_ctx_desc == NULL) {
+		return -EFAULT;
+	}
+
+	if (g->gr->gr_ctx_desc->dump_ctxsw_stats_on_channel_close) {
+		buf[0] = 'Y';
+	} else {
+		buf[0] = 'N';
+	}
+
+	buf[1] = '\n';
+	buf[2] = 0x00;
+
+	return simple_read_from_buffer(user_buf, count, ppos, buf, 2);
+}
+
+static ssize_t dump_ctxsw_stats_on_channel_close_write(struct file *file,
+		const char __user *user_buf, size_t count, loff_t *ppos)
+{
+	char buf[32];
+	int buf_size;
+	bool val;
+	struct gk20a *g = file->private_data;
+
+	if (g->gr->gr_ctx_desc == NULL) {
+		return -EFAULT;
+	}
+
+	buf_size = min(count, (sizeof(buf)-1));
+	if (copy_from_user(buf, user_buf, buf_size)) {
+		return -EFAULT;
+	}
+
+	if (strtobool(buf, &val) == 0) {
+		g->gr->gr_ctx_desc->dump_ctxsw_stats_on_channel_close = val;
+	}
+
+	return count;
+}
+
+static struct file_operations dump_ctxsw_stats_on_channel_close_fops = {
+	.open =		simple_open,
+	.read =		dump_ctxsw_stats_on_channel_close_read,
+	.write =	dump_ctxsw_stats_on_channel_close_write,
+};
+
 int gr_gk20a_debugfs_init(struct gk20a *g)
 {
 	struct nvgpu_os_linux *l = nvgpu_os_linux_from_gk20a(g);
@@ -172,6 +224,13 @@ int gr_gk20a_debugfs_init(struct gk20a *g)
 	if (!d)
 		return -ENOMEM;
 
+	d = debugfs_create_file(
+		"dump_ctxsw_stats_on_channel_close", S_IRUGO|S_IWUSR,
+				l->debugfs, g,
+				&dump_ctxsw_stats_on_channel_close_fops);
+	if (!d)
+		return -ENOMEM;
+
 	return 0;
 }
 

drivers/gpu/nvgpu/os/linux/platform_gp10b_tegra.c

@@ -158,11 +158,6 @@ static int gp10b_tegra_probe(struct device *dev)
 
 	platform->disable_bigpage = !dev->archdata.iommu;
 
-	platform->g->gr->ctx_vars.dump_ctxsw_stats_on_channel_close
-		= false;
-	platform->g->gr->ctx_vars.dump_ctxsw_stats_on_channel_close
-		= false;
-
 	gp10b_tegra_get_clocks(dev);
 	nvgpu_linux_init_clk_support(platform->g);
 

drivers/gpu/nvgpu/os/linux/platform_gv11b_tegra.c

@@ -86,11 +86,6 @@ static int gv11b_tegra_probe(struct device *dev)
 
 	platform->disable_bigpage = !dev->archdata.iommu;
 
-	platform->g->gr->ctx_vars.dump_ctxsw_stats_on_channel_close
-		= false;
-	platform->g->gr->ctx_vars.dump_ctxsw_stats_on_channel_close
-		= false;
-
 	gp10b_tegra_get_clocks(dev);
 	nvgpu_linux_init_clk_support(platform->g);