Merge v6.8-rc1 into drm-misc-fixes

Let's kickstart the 6.8 fix cycle.

Signed-off-by: Maxime Ripard <mripard@kernel.org>

1 files changed, 192 insertions, 7 deletions

diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c
index b4e31e59c799..d04660b60046 100644
--- a/drivers/gpu/drm/i915/i915_gpu_error.c
+++ b/drivers/gpu/drm/i915/i915_gpu_error.c
@@ -57,6 +57,7 @@
 #include "i915_memcpy.h"
 #include "i915_reg.h"
 #include "i915_scatterlist.h"
+#include "i915_sysfs.h"
 #include "i915_utils.h"
 
 #define ALLOW_FAIL (__GFP_KSWAPD_RECLAIM | __GFP_RETRY_MAYFAIL | __GFP_NOWARN)
@@ -520,7 +521,7 @@ __find_vma(struct i915_vma_coredump *vma, const char *name)
 	return NULL;
 }
 
-struct i915_vma_coredump *
+static struct i915_vma_coredump *
 intel_gpu_error_find_batch(const struct intel_engine_coredump *ee)
 {
 	return __find_vma(ee->vma, "batch");
@@ -609,9 +610,9 @@ void i915_error_printf(struct drm_i915_error_state_buf *e, const char *f, ...)
 	va_end(args);
 }
 
-void intel_gpu_error_print_vma(struct drm_i915_error_state_buf *m,
-			       const struct intel_engine_cs *engine,
-			       const struct i915_vma_coredump *vma)
+static void intel_gpu_error_print_vma(struct drm_i915_error_state_buf *m,
+				      const struct intel_engine_cs *engine,
+				      const struct i915_vma_coredump *vma)
 {
 	char out[ASCII85_BUFSZ];
 	struct page *page;
@@ -660,6 +661,7 @@ static void err_print_params(struct drm_i915_error_state_buf *m,
 	struct drm_printer p = i915_error_printer(m);
 
 	i915_params_dump(params, &p);
+	intel_display_params_dump(m->i915, &p);
 }
 
 static void err_print_pciid(struct drm_i915_error_state_buf *m,
@@ -1027,6 +1029,7 @@ static void i915_vma_coredump_free(struct i915_vma_coredump *vma)
 static void cleanup_params(struct i915_gpu_coredump *error)
 {
 	i915_params_free(&error->params);
+	intel_display_params_free(&error->display_params);
 }
 
 static void cleanup_uc(struct intel_uc_coredump *uc)
@@ -1988,6 +1991,7 @@ static void capture_gen(struct i915_gpu_coredump *error)
 	error->suspend_count = i915->suspend_count;
 
 	i915_params_copy(&error->params, &i915->params);
+	intel_display_params_copy(&error->display_params);
 	memcpy(&error->device_info,
 	       INTEL_INFO(i915),
 	       sizeof(error->device_info));
@@ -2137,7 +2141,7 @@ __i915_gpu_coredump(struct intel_gt *gt, intel_engine_mask_t engine_mask, u32 du
 	return error;
 }
 
-struct i915_gpu_coredump *
+static struct i915_gpu_coredump *
 i915_gpu_coredump(struct intel_gt *gt, intel_engine_mask_t engine_mask, u32 dump_flags)
 {
 	static DEFINE_MUTEX(capture_mutex);
@@ -2174,7 +2178,7 @@ void i915_error_state_store(struct i915_gpu_coredump *error)
 	    ktime_get_real_seconds() - DRIVER_TIMESTAMP < DAY_AS_SECONDS(180)) {
 		pr_info("GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace.\n");
 		pr_info("Please file a _new_ bug report at https://gitlab.freedesktop.org/drm/intel/issues/new.\n");
-		pr_info("Please see https://gitlab.freedesktop.org/drm/intel/-/wikis/How-to-file-i915-bugs for details.\n");
+		pr_info("Please see https://drm.pages.freedesktop.org/intel-docs/how-to-file-i915-bugs.html for details.\n");
 		pr_info("drm/i915 developers can then reassign to the right component if it's not a kernel issue.\n");
 		pr_info("The GPU crash dump is required to analyze GPU hangs, so please always attach it.\n");
 		pr_info("GPU crash dump saved to /sys/class/drm/card%d/error\n",
@@ -2208,7 +2212,7 @@ void i915_capture_error_state(struct intel_gt *gt,
 	i915_gpu_coredump_put(error);
 }
 
-struct i915_gpu_coredump *
+static struct i915_gpu_coredump *
 i915_first_error_state(struct drm_i915_private *i915)
 {
 	struct i915_gpu_coredump *error;
@@ -2375,3 +2379,184 @@ void intel_klog_error_capture(struct intel_gt *gt,
 	drm_info(&i915->drm, "[Capture/%d.%d] Dumped %zd bytes\n", l_count, line++, pos_err);
 }
 #endif
+
+static ssize_t gpu_state_read(struct file *file, char __user *ubuf,
+			      size_t count, loff_t *pos)
+{
+	struct i915_gpu_coredump *error;
+	ssize_t ret;
+	void *buf;
+
+	error = file->private_data;
+	if (!error)
+		return 0;
+
+	/* Bounce buffer required because of kernfs __user API convenience. */
+	buf = kmalloc(count, GFP_KERNEL);
+	if (!buf)
+		return -ENOMEM;
+
+	ret = i915_gpu_coredump_copy_to_buffer(error, buf, *pos, count);
+	if (ret <= 0)
+		goto out;
+
+	if (!copy_to_user(ubuf, buf, ret))
+		*pos += ret;
+	else
+		ret = -EFAULT;
+
+out:
+	kfree(buf);
+	return ret;
+}
+
+static int gpu_state_release(struct inode *inode, struct file *file)
+{
+	i915_gpu_coredump_put(file->private_data);
+	return 0;
+}
+
+static int i915_gpu_info_open(struct inode *inode, struct file *file)
+{
+	struct drm_i915_private *i915 = inode->i_private;
+	struct i915_gpu_coredump *gpu;
+	intel_wakeref_t wakeref;
+
+	gpu = NULL;
+	with_intel_runtime_pm(&i915->runtime_pm, wakeref)
+		gpu = i915_gpu_coredump(to_gt(i915), ALL_ENGINES, CORE_DUMP_FLAG_NONE);
+
+	if (IS_ERR(gpu))
+		return PTR_ERR(gpu);
+
+	file->private_data = gpu;
+	return 0;
+}
+
+static const struct file_operations i915_gpu_info_fops = {
+	.owner = THIS_MODULE,
+	.open = i915_gpu_info_open,
+	.read = gpu_state_read,
+	.llseek = default_llseek,
+	.release = gpu_state_release,
+};
+
+static ssize_t
+i915_error_state_write(struct file *filp,
+		       const char __user *ubuf,
+		       size_t cnt,
+		       loff_t *ppos)
+{
+	struct i915_gpu_coredump *error = filp->private_data;
+
+	if (!error)
+		return 0;
+
+	drm_dbg(&error->i915->drm, "Resetting error state\n");
+	i915_reset_error_state(error->i915);
+
+	return cnt;
+}
+
+static int i915_error_state_open(struct inode *inode, struct file *file)
+{
+	struct i915_gpu_coredump *error;
+
+	error = i915_first_error_state(inode->i_private);
+	if (IS_ERR(error))
+		return PTR_ERR(error);
+
+	file->private_data  = error;
+	return 0;
+}
+
+static const struct file_operations i915_error_state_fops = {
+	.owner = THIS_MODULE,
+	.open = i915_error_state_open,
+	.read = gpu_state_read,
+	.write = i915_error_state_write,
+	.llseek = default_llseek,
+	.release = gpu_state_release,
+};
+
+void i915_gpu_error_debugfs_register(struct drm_i915_private *i915)
+{
+	struct drm_minor *minor = i915->drm.primary;
+
+	debugfs_create_file("i915_error_state", 0644, minor->debugfs_root, i915,
+			    &i915_error_state_fops);
+	debugfs_create_file("i915_gpu_info", 0644, minor->debugfs_root, i915,
+			    &i915_gpu_info_fops);
+}
+
+static ssize_t error_state_read(struct file *filp, struct kobject *kobj,
+				struct bin_attribute *attr, char *buf,
+				loff_t off, size_t count)
+{
+
+	struct device *kdev = kobj_to_dev(kobj);
+	struct drm_i915_private *i915 = kdev_minor_to_i915(kdev);
+	struct i915_gpu_coredump *gpu;
+	ssize_t ret = 0;
+
+	/*
+	 * FIXME: Concurrent clients triggering resets and reading + clearing
+	 * dumps can cause inconsistent sysfs reads when a user calls in with a
+	 * non-zero offset to complete a prior partial read but the
+	 * gpu_coredump has been cleared or replaced.
+	 */
+
+	gpu = i915_first_error_state(i915);
+	if (IS_ERR(gpu)) {
+		ret = PTR_ERR(gpu);
+	} else if (gpu) {
+		ret = i915_gpu_coredump_copy_to_buffer(gpu, buf, off, count);
+		i915_gpu_coredump_put(gpu);
+	} else {
+		const char *str = "No error state collected\n";
+		size_t len = strlen(str);
+
+		if (off < len) {
+			ret = min_t(size_t, count, len - off);
+			memcpy(buf, str + off, ret);
+		}
+	}
+
+	return ret;
+}
+
+static ssize_t error_state_write(struct file *file, struct kobject *kobj,
+				 struct bin_attribute *attr, char *buf,
+				 loff_t off, size_t count)
+{
+	struct device *kdev = kobj_to_dev(kobj);
+	struct drm_i915_private *dev_priv = kdev_minor_to_i915(kdev);
+
+	drm_dbg(&dev_priv->drm, "Resetting error state\n");
+	i915_reset_error_state(dev_priv);
+
+	return count;
+}
+
+static const struct bin_attribute error_state_attr = {
+	.attr.name = "error",
+	.attr.mode = S_IRUSR | S_IWUSR,
+	.size = 0,
+	.read = error_state_read,
+	.write = error_state_write,
+};
+
+void i915_gpu_error_sysfs_setup(struct drm_i915_private *i915)
+{
+	struct device *kdev = i915->drm.primary->kdev;
+
+	if (sysfs_create_bin_file(&kdev->kobj, &error_state_attr))
+		drm_err(&i915->drm, "error_state sysfs setup failed\n");
+}
+
+void i915_gpu_error_sysfs_teardown(struct drm_i915_private *i915)
+{
+	struct device *kdev = i915->drm.primary->kdev;
+
+	sysfs_remove_bin_file(&kdev->kobj, &error_state_attr);
+}