gpu: nvgpu: add g->sw_ready flag

Fix a race condition where we'd still be booting up the gpu and/or
initializing the driver but elsewhere assume that all is done already.

Some userspace APIs to make sure that we're ready by testing
g->gr.sw_ready, but this flag is set in the middle of bootup; there are
other things after gr initialization. Add a new flag that is enabled
after bootup is fully complete at the end of finalize_poweron, and
change the checks in user API paths to test the new flag only.

These checks are only in the ioctl paths for ctrl, dbg and tsg, and in
the ctrl device's opening path.

The gr.sw_ready flag is still left there to signify whether just gr has
had its bookkeeping initialized.

Bug 200370011

Change-Id: I2995500e06de46430d9b835de1e9d60b3f01744e
Signed-off-by: Konsta Holtta <kholtta@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1640124
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>

drivers/gpu/nvgpu/common/linux/ioctl_ctrl.c

@@ -86,7 +86,7 @@ int gk20a_ctrl_dev_open(struct inode *inode, struct file *filp)
 	 */
 	priv->g = g;
 
-	if (!g->gr.sw_ready) {
+	if (!g->sw_ready) {
 		err = gk20a_busy(g);
 		if (err)
 			goto free_ref;
@@ -1556,7 +1556,7 @@ long gk20a_ctrl_dev_ioctl(struct file *filp, unsigned int cmd, unsigned long arg
 			return -EFAULT;
 	}
 
-	if (!g->gr.sw_ready) {
+	if (!g->sw_ready) {
 		err = gk20a_busy(g);
 		if (err)
 			return err;

drivers/gpu/nvgpu/common/linux/ioctl_dbg.c

@@ -1835,7 +1835,7 @@ long gk20a_dbg_gpu_dev_ioctl(struct file *filp, unsigned int cmd,
 			return -EFAULT;
 	}
 
-	if (!g->gr.sw_ready) {
+	if (!g->sw_ready) {
 		err = gk20a_busy(g);
 		if (err)
 			return err;

drivers/gpu/nvgpu/common/linux/ioctl_tsg.c

@@ -458,7 +458,7 @@ long nvgpu_ioctl_tsg_dev_ioctl(struct file *filp, unsigned int cmd,
 			return -EFAULT;
 	}
 
-	if (!g->gr.sw_ready) {
+	if (!g->sw_ready) {
 		err = gk20a_busy(g);
 		if (err)
 			return err;

drivers/gpu/nvgpu/common/linux/module.c

@@ -290,6 +290,8 @@ int gk20a_pm_finalize_poweron(struct device *dev)
 		return err;
 	}
 
+	g->sw_ready = true;
+
 done:
 	if (err)
 		g->power_on = false;
@@ -1041,6 +1043,7 @@ void gk20a_driver_start_unload(struct gk20a *g)
 	/* GR SW ready needs to be invalidated at this time with the busy lock
 	 * held to prevent a racing condition on the gr/mm code */
 	g->gr.sw_ready = false;
+	g->sw_ready = false;
 	up_write(&l->busy_lock);
 
 	if (g->is_virtual)

drivers/gpu/nvgpu/common/linux/sysfs.c

@@ -815,6 +815,7 @@ static ssize_t tpc_fs_mask_store(struct device *dev,
 		g->gr.ctx_vars.local_golden_image = NULL;
 		g->gr.ctx_vars.golden_image_initialized = false;
 		g->gr.ctx_vars.golden_image_size = 0;
+		/* Cause next poweron to reinit just gr */
 		g->gr.sw_ready = false;
 	}
 

drivers/gpu/nvgpu/common/linux/vgpu/vgpu.c

@@ -505,6 +505,8 @@ int vgpu_pm_finalize_poweron(struct device *dev)
 	gk20a_sched_ctrl_init(g);
 	gk20a_channel_resume(g);
 
+	g->sw_ready = true;
+
 done:
 	return err;
 }

drivers/gpu/nvgpu/gk20a/gk20a.h

@@ -1133,6 +1133,7 @@ struct gk20a {
 	bool gpu_reset_done;
 	bool power_on;
 	bool suspended;
+	bool sw_ready;
 
 	u32 log_mask;
 	u32 log_trace;