Commit d5818410 authored by Jani Nikula's avatar Jani Nikula
Browse files

drm/i915: move gpu error sysfs to i915_gpu_error.c 



Hide gpu error specifics in i915_gpu_error.c. This is also cleaner wrt
conditional compilation, as i915_gpu_error.c is only built with
DRM_I915_CAPTURE_ERROR=y.

With this, we can also make i915_first_error_state() static.

Signed-off-by: default avatarJani Nikula <jani.nikula@intel.com>
Reviewed-by: default avatarNirmoy Das <nirmoy.das@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231031124502.1772160-3-jani.nikula@intel.com
parent 4fca5198

drivers/gpu/drm/i915/i915_gpu_error.c

+74 −1
Original line number Diff line number Diff line
@@ -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)
@@ -2208,7 +2209,7 @@ 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;
@@ -2484,3 +2485,75 @@ void i915_gpu_error_debugfs_register(struct drm_i915_private *i915) 
	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);

}

drivers/gpu/drm/i915/i915_gpu_error.h

+10 −7
Original line number Diff line number Diff line
@@ -323,11 +323,12 @@ static inline void i915_gpu_coredump_put(struct i915_gpu_coredump *gpu) 
		kref_put(&gpu->ref, __i915_gpu_coredump_free);

}



struct i915_gpu_coredump *i915_first_error_state(struct drm_i915_private *i915);

void i915_reset_error_state(struct drm_i915_private *i915);

void i915_disable_error_state(struct drm_i915_private *i915, int err);



void i915_gpu_error_debugfs_register(struct drm_i915_private *i915);

void i915_gpu_error_sysfs_setup(struct drm_i915_private *i915);

void i915_gpu_error_sysfs_teardown(struct drm_i915_private *i915);



#else
@@ -396,12 +397,6 @@ static inline void i915_gpu_coredump_put(struct i915_gpu_coredump *gpu) 
{

}



static inline struct i915_gpu_coredump *

i915_first_error_state(struct drm_i915_private *i915)

{

	return ERR_PTR(-ENODEV);

}



static inline void i915_reset_error_state(struct drm_i915_private *i915)

{

}
@@ -415,6 +410,14 @@ static inline void i915_gpu_error_debugfs_register(struct drm_i915_private *i915 
{

}



static inline void i915_gpu_error_sysfs_setup(struct drm_i915_private *i915)

{

}



static inline void i915_gpu_error_sysfs_teardown(struct drm_i915_private *i915)

{

}



#endif /* IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR) */



#endif /* _I915_GPU_ERROR_H_ */

drivers/gpu/drm/i915/i915_sysfs.c

+2 −77
Original line number Diff line number Diff line
@@ -155,81 +155,6 @@ static const struct bin_attribute dpf_attrs_1 = { 
	.private = (void *)1

};



#if IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR)



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,

};



static void i915_setup_error_capture(struct device *kdev)

{

	if (sysfs_create_bin_file(&kdev->kobj, &error_state_attr))

		drm_err(&kdev_minor_to_i915(kdev)->drm,

			"error_state sysfs setup failed\n");

}



static void i915_teardown_error_capture(struct device *kdev)

{

	sysfs_remove_bin_file(&kdev->kobj, &error_state_attr);

}

#else

static void i915_setup_error_capture(struct device *kdev) {}

static void i915_teardown_error_capture(struct device *kdev) {}

#endif



void i915_setup_sysfs(struct drm_i915_private *dev_priv)

{

	struct device *kdev = dev_priv->drm.primary->kdev;
@@ -255,7 +180,7 @@ void i915_setup_sysfs(struct drm_i915_private *dev_priv) 
		drm_warn(&dev_priv->drm,

			 "failed to register GT sysfs directory\n");



	i915_setup_error_capture(kdev);

	i915_gpu_error_sysfs_setup(dev_priv);



	intel_engines_add_sysfs(dev_priv);

}
@@ -264,7 +189,7 @@ void i915_teardown_sysfs(struct drm_i915_private *dev_priv) 
{

	struct device *kdev = dev_priv->drm.primary->kdev;



	i915_teardown_error_capture(kdev);

	i915_gpu_error_sysfs_teardown(dev_priv);



	device_remove_bin_file(kdev,  &dpf_attrs_1);

	device_remove_bin_file(kdev,  &dpf_attrs);