Skip to contentSkip to conversation

Introduce drmfs pseudo filesystem for drm subsystem

RFC In review [email protected] 5 participants 10 messages Dec 5, 2016 – Dec 14, 2016

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

cover Introduce drmfs pseudo filesystem for drm subsystem

From: Swati Dhingra <[email protected]>

Currently, we don't have a stable ABI which can be used for the purpose of providing output debug/loggging/crc and other such data from DRM. The ABI in current use (filesystems, ioctls, et al.) have their own constraints and are intended to output a particular type of data. Few cases in point:
sysfs - stable ABI, but constrained to one textual value per file
debugfs - unstable ABI, free-for-all
ioctls - not as suitable to many single purpose continuous data
dumping, we would very quickly run out ioctl space; requires more userspace support than "cat"
device nodes - a real possibilty, kernel instantiation is more tricky,
requires udev (+udev.rules) or userspace discovery of the dynamic major:minor (via sysfs) [mounting a registered filesystem is easy in comparison]
netlink - stream based, therefore involves numerous copies.

Debugfs is the lesser among the evils here, thereby we have grown used to the convenience and flexibility in presentation that debugfs gives us (including relayfs inodes) that we want some of that hierachy in stable user ABI form.

Due to these limitations, there is a need for a new pseudo filesytem, that would act as a stable 'free-for-all' ABI, with the heirarchial structure and thus convenience of debugfs. This will be managed by drm, thus named 'drmfs'. DRM would register this filesystem to manage a canonical mountpoint, but this wouldn't limit everyone to only using that pseudofs underneath.

This can serve to hold various kinds of output data from Linux DRM subsystems, for the files which can't truely fit anywhere else with existing ABI's but present so, for the lack of a better place.

In this patch series, we have introduced a pseudo filesystem named as 'drmfs' for now. The filesystem is introduced in the first patch, and the subsequent patches make use of the filesystem interfaces, in drm driver, and making them available for use by the drm subsystem components, one of which is i915. We've moved the location of i915 GuC logs from debugfs to drmfs in the last patch. Subsequently, more such files such as pipe_crc, error states, memory stats, etc. can be move to this filesystem, if the idea introduced here is acceptable per se. The filesystem introduced is being used to house the data generated by i915 driver in this patch series, but will hopefully be generic enough to provide scope for usage by any other drm subsystem component.

The patch series is being floated as RFC to gather feedback on the idea and infrastructure proposed here and it's suitability to address the specific problem statement/use case.

TODO: Create documentation. Will do so in next version.

v2: fix the bat failures caused due to missing config check

v3: Changes made:
- Move the location of drmfs from fs/ to drivers/gpu/drm/ (Chris)
- Moving config checks to header (Chris,Daniel)

Sourab Gupta (4):
drm: Introduce drmfs pseudo filesystem interfaces
drm: Register drmfs filesystem from drm init
drm: Create driver specific root directory inside drmfs drm/i915: Creating guc log file in drmfs instead of debugfs

 drivers/gpu/drm/Kconfig                    |   9 +
 drivers/gpu/drm/Makefile                   |   1 +
 drivers/gpu/drm/drm_drv.c                  |  12 +
 drivers/gpu/drm/drmfs.c                    | 555 +++++++++++++++++++++++++++++
 drivers/gpu/drm/i915/i915_guc_submission.c |  33 +-
 include/drm/drm_drv.h                      |   3 +
 include/drm/drmfs.h                        |  77 ++++
 include/uapi/linux/magic.h                 |   3 +
 8 files changed, 672 insertions(+), 21 deletions(-)
 create mode 100644 drivers/gpu/drm/drmfs.c
 create mode 100644 include/drm/drmfs.h
Signature · 6 lines
2.7.4

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#1 lore raw

RFC 1/4 drm: Introduce drmfs pseudo filesystem interfaces

From: Sourab Gupta <[email protected]>

The patch introduces a new pseudo filesystem type, named 'drmfs' which is intended to house the files for the data generated by drm subsystem that cannot be accommodated by any of the existing filesystems.
The filesystem is modelled on the lines of existing pseudo-filesystems such as debugfs/tracefs, and borrows ideas from their implementation. This filesystem will be appearing at sys/kernel/drm.

A new config 'CONFIG_DRMFS' is introduced to enable/disable the filesystem, which is dependent on CONFIG_DRM.
The filesystem will not be registered standalone during kernel init time, instead it is intended to be initialized/registered during drm initialization.

The intent for introduction of the filesystem is to act as a location to hold various kinds of data output from Linux DRM subsystems, which can't really fit anywhere else into the existing filesystems such as debugfs/sysfs etc. All these filesystems have their own constraints and are intended to output a particular type of data such as attributes and small debug parameter data. Due to these constraints, there is a need for a new pseudo filesytem, customizable to DRM specific requirements and catering to the needs to DRM subsystem components

---
 drivers/gpu/drm/Kconfig    |   9 +
 drivers/gpu/drm/Makefile   |   1 +
 drivers/gpu/drm/drmfs.c    | 555 +++++++++++++++++++++++++++++++++++++++++++++
 include/drm/drmfs.h        |  77 +++++++
 include/uapi/linux/magic.h |   3 +
 5 files changed, 645 insertions(+)
 create mode 100644 drivers/gpu/drm/drmfs.c
 create mode 100644 include/drm/drmfs.h
drivers/gpu/drm/Kconfig +9 -0
index 95fc041..5082840 100644
@@ -26,6 +26,15 @@ config DRM_MIPI_DSI
 	bool
 	depends on DRM
 
+config DRMFS
+	bool "Drmfs file system support"
+	depends on DRM
+	help
+	  Drmfs is a pseudo file system for drm subsystem output data.
+
+	  drmfs is a filesystem which serves as a stable ABI to hold
+	  miscellaneous output data from drm subsystems.
+
 config DRM_DP_AUX_CHARDEV
 	bool "DRM DP AUX Interface"
 	depends on DRM
drivers/gpu/drm/Makefile +1 -0
index 883f3e7..5021225 100644
@@ -42,6 +42,7 @@ CFLAGS_drm_trace_points.o := -I$(src)
 
 obj-$(CONFIG_DRM)	+= drm.o
 obj-$(CONFIG_DRM_MIPI_DSI) += drm_mipi_dsi.o
+obj-$(CONFIG_DRMFS)	+= drmfs.o
 obj-$(CONFIG_DRM_ARM)	+= arm/
 obj-$(CONFIG_DRM_TTM)	+= ttm/
 obj-$(CONFIG_DRM_TDFX)	+= tdfx/
drivers/gpu/drm/drmfs.c new file +555 -0

555 lines not on this page · raw message

include/drm/drmfs.h new file +77 -0

77 lines not on this page · raw message

include/uapi/linux/magic.h +3 -0
index 9bd5594..ef01eb7 100644
@@ -61,6 +61,9 @@
 #define STACK_END_MAGIC		0x57AC6E9D
 
 #define TRACEFS_MAGIC          0x74726163
+#define DRMFS_MAGIC	       0x84724143      /* some random number.
+						* Is there a better way?
+						*/
 
 #define V9FS_MAGIC		0x01021997
 
Signature · 6 lines
2.7.4

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#2 lore raw

RFC 2/4 drm: Register drmfs filesystem from drm init

In reply to [email protected]

From: Sourab Gupta <[email protected]>

The drmfs filesystem will not be registered standalone during kernel init time, instead it is intended to be initialized/registered during drm initialization.

---
 drivers/gpu/drm/drm_drv.c | 6 ++++++
 1 file changed, 6 insertions(+)
drivers/gpu/drm/drm_drv.c +6 -0
index f74b7d0..298013c 100644
@@ -34,6 +34,7 @@
 #include <linux/slab.h>
 
 #include <drm/drm_drv.h>
+#include <drm/drmfs.h>
 #include <drm/drmP.h>
 
 #include "drm_crtc_internal.h"
@@ -828,6 +829,7 @@ static void drm_core_exit(void)
 {
 	unregister_chrdev(DRM_MAJOR, "drm");
 	debugfs_remove(drm_debugfs_root);
+	drmfs_fini();
 	drm_sysfs_destroy();
 	idr_destroy(&drm_minors_idr);
 	drm_connector_ida_destroy();
@@ -848,6 +850,10 @@ static int __init drm_core_init(void)
 		goto error;
 	}
 
+	ret = drmfs_init();
+	if (ret < 0)
+		DRM_ERROR("Cannot create DRM FS: %d\n", ret);
+
 	drm_debugfs_root = debugfs_create_dir("dri", NULL);
 	if (!drm_debugfs_root) {
 		ret = -ENOMEM;
Signature · 6 lines
2.7.4

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#3 lore raw

RFC 3/4 drm: Create driver specific root directory inside drmfs

In reply to [email protected]

From: Sourab Gupta <[email protected]>

A driver specific directory is created inside drmfs root, and dentry of this directory is saved for subsequent use by the driver (e.g. i915). The driver can then create files/directories inside this root directory directly.

In case of i915 driver, the top directory is created at '/sys/kernel/drm/i915'.

---
 drivers/gpu/drm/drm_drv.c | 6 ++++++
 include/drm/drm_drv.h     | 3 +++
 2 files changed, 9 insertions(+)
drivers/gpu/drm/drm_drv.c +6 -0
index 298013c..4c32cf8 100644
@@ -671,6 +671,11 @@ EXPORT_SYMBOL(drm_dev_unref);
 int drm_dev_register(struct drm_device *dev, unsigned long flags)
 {
 	int ret;
+	struct dentry *drmfs_root;
+
+	drmfs_root = drmfs_create_dir(dev->driver->name, NULL);
+	if (!IS_ERR(drmfs_root))
+		dev->driver->drmfs_root = drmfs_root;
 
 	mutex_lock(&drm_global_mutex);
 
@@ -742,6 +747,7 @@ void drm_dev_unregister(struct drm_device *dev)
 	drm_minor_unregister(dev, DRM_MINOR_PRIMARY);
 	drm_minor_unregister(dev, DRM_MINOR_RENDER);
 	drm_minor_unregister(dev, DRM_MINOR_CONTROL);
+	drmfs_remove(dev->driver->drmfs_root);
 }
 EXPORT_SYMBOL(drm_dev_unregister);
 
include/drm/drm_drv.h +3 -0
index c4fc495..d646296 100644
@@ -405,6 +405,9 @@ struct drm_driver {
 
 	/* List of devices hanging off this driver with stealth attach. */
 	struct list_head legacy_dev_list;
+
+	/* drmfs parent directory dentry for this driver */
+	struct dentry *drmfs_root;
 };
 
 extern __printf(6, 7)
Signature · 6 lines
2.7.4

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#4 lore raw

RFC 4/4 drm/i915: Creating guc log file in drmfs instead of debugfs

In reply to [email protected]

From: Sourab Gupta <[email protected]>

In the current scenario, the relay API fit well only with debugfs, due to availability of parent dentry. Any other existing filesystem was not feasible for holding guc logs, due to incompatibility with relay. But this makes the guc_log file unavailable on the production kernels.

GuC log file can therefore be one of candidates for movement to the drmfs filesystem, which can satisfy all the requirements needed by relay API, and can house any relayfs based output file.

The patch moves the parent directory of guc 'log_dir' from debugfs_root to drmfs_root, while using the drmfs api's to create the requisite files.

---
 drivers/gpu/drm/i915/i915_guc_submission.c | 33 +++++++++++-------------------
 1 file changed, 12 insertions(+), 21 deletions(-)
drivers/gpu/drm/i915/i915_guc_submission.c +12 -21
index 5841380..3f5978a 100644
@@ -22,8 +22,8 @@
  *
  */
 #include <linux/circ_buf.h>
-#include <linux/debugfs.h>
 #include <linux/relay.h>
+#include <drm/drmfs.h>
 #include "i915_drv.h"
 #include "intel_uc.h"
 
@@ -812,7 +812,7 @@ static int subbuf_start_callback(struct rchan_buf *buf,
 }
 
 /*
- * file_create() callback. Creates relay file in debugfs.
+ * file_create() callback. Creates relay file.
  */
 static struct dentry *create_buf_file_callback(const char *filename,
 					       struct dentry *parent,
@@ -838,17 +838,19 @@ static struct dentry *create_buf_file_callback(const char *filename,
 	 * dentry of the file associated with the channel buffer and that file's
 	 * name need not be same as the filename passed as an argument.
 	 */
-	buf_file = debugfs_create_file("guc_log", mode,
+
+	buf_file = drmfs_create_file("guc_log", mode,
 				       parent, buf, &relay_file_operations);
+
 	return buf_file;
 }
 
 /*
- * file_remove() default callback. Removes relay file in debugfs.
+ * file_remove() default callback. Removes relay file.
  */
 static int remove_buf_file_callback(struct dentry *dentry)
 {
-	debugfs_remove(dentry);
+	drmfs_remove(dentry);
 	return 0;
 }
 
@@ -898,22 +900,11 @@ static int guc_log_create_relay_file(struct intel_guc *guc)
 	struct dentry *log_dir;
 	int ret;
 
-	/* For now create the log file in /sys/kernel/debug/dri/0 dir */
-	log_dir = dev_priv->drm.primary->debugfs_root;
-
-	/* If /sys/kernel/debug/dri/0 location do not exist, then debugfs is
-	 * not mounted and so can't create the relay file.
-	 * The relay API seems to fit well with debugfs only, for availing relay
-	 * there are 3 requirements which can be met for debugfs file only in a
-	 * straightforward/clean manner :-
-	 * i)   Need the associated dentry pointer of the file, while opening the
-	 *      relay channel.
-	 * ii)  Should be able to use 'relay_file_operations' fops for the file.
-	 * iii) Set the 'i_private' field of file's inode to the pointer of
-	 *	relay channel buffer.
-	 */
+	/* Create the log file in drmfs dir: /sys/kernel/drm/i915/ */
+	log_dir = dev_priv->drm.driver->drmfs_root;
+
 	if (!log_dir) {
-		DRM_ERROR("Debugfs dir not available yet for GuC log file\n");
+		DRM_ERROR("Root dir not available yet for GuC log file\n");
 		return -ENODEV;
 	}
 
@@ -1154,7 +1145,7 @@ static int guc_log_create_extras(struct intel_guc *guc)
 	if (!guc->log.relay_chan) {
 		/* Create a relay channel, so that we have buffers for storing
 		 * the GuC firmware logs, the channel will be linked with a file
-		 * later on when debugfs is registered.
+		 * later on when drmfs is registered.
 		 */
 		ret = guc_log_create_relay_channel(guc);
 		if (ret)
Signature · 6 lines
2.7.4

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#5 lore raw

✗ Fi.CI.BAT: warning for Introduce drmfs pseudo filesystem for drm subsystem (rev3)

In reply to [email protected]
== Series Details ==

== Series Details ==

Series: Introduce drmfs pseudo filesystem for drm subsystem (rev3)
URL : https://patchwork.freedesktop.org/series/16203/
State : warning

== Summary ==

Series 16203v3 Introduce drmfs pseudo filesystem for drm subsystem
https://patchwork.freedesktop.org/api/1.0/series/16203/revisions/3/mbox/

Test kms_pipe_crc_basic:

        Subgroup bad-nb-words-3:
                pass       -> DMESG-WARN (fi-ivb-3770)
fi-bdw-5557u     total:247  pass:232  dwarn:0   dfail:0   fail:0   skip:15 
fi-bsw-n3050     total:247  pass:207  dwarn:0   dfail:0   fail:0   skip:40 
fi-byt-j1900     total:247  pass:219  dwarn:0   dfail:0   fail:0   skip:28 
fi-byt-n2820     total:247  pass:215  dwarn:0   dfail:0   fail:0   skip:32 
fi-hsw-4770      total:247  pass:227  dwarn:0   dfail:0   fail:0   skip:20 
fi-hsw-4770r     total:247  pass:227  dwarn:0   dfail:0   fail:0   skip:20 
fi-ilk-650       total:247  pass:194  dwarn:0   dfail:0   fail:0   skip:53 
fi-ivb-3520m     total:247  pass:225  dwarn:0   dfail:0   fail:0   skip:22 
fi-ivb-3770      total:247  pass:224  dwarn:1   dfail:0   fail:0   skip:22 
fi-kbl-7500u     total:247  pass:225  dwarn:0   dfail:0   fail:0   skip:22 
fi-skl-6260u     total:247  pass:233  dwarn:0   dfail:0   fail:0   skip:14 
fi-skl-6700hq    total:247  pass:226  dwarn:0   dfail:0   fail:0   skip:21 
fi-skl-6700k     total:247  pass:223  dwarn:3   dfail:0   fail:0   skip:21 
fi-skl-6770hq    total:247  pass:233  dwarn:0   dfail:0   fail:0   skip:14 
fi-snb-2520m     total:247  pass:215  dwarn:0   dfail:0   fail:0   skip:32 
fi-snb-2600      total:247  pass:214  dwarn:0   dfail:0   fail:0   skip:33 

a17bd268c6f9e1076a9aea44573ff58b0caeb44d drm-tip: 2016y-12m-05d-11h-19m-41s UTC integration manifest 6987287 drm/i915: Creating guc log file in drmfs instead of debugfs
50a4a18 drm: Create driver specific root directory inside drmfs
df2fc99 drm: Register drmfs filesystem from drm init
d340379 drm: Introduce drmfs pseudo filesystem interfaces

== Logs ==

For more details see: https://intel-gfx-ci.01.org/CI/Patchwork_3186/
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

#6 lore raw

Re: ✗ Fi.CI.BAT: warning for Introduce drmfs pseudo filesystem for drm subsystem (rev3)

Quoted 6 lines

== Series Details ==

Series: Introduce drmfs pseudo filesystem for drm subsystem (rev3)
URL : https://patchwork.freedesktop.org/series/16203/
State : warning

Quoted 8 lines

== Summary ==

Series 16203v3 Introduce drmfs pseudo filesystem for drm subsystem
https://patchwork.freedesktop.org/api/1.0/series/16203/revisions/3/mbox/

Test kms_pipe_crc_basic:

        Subgroup bad-nb-words-3:
                pass       -> DMESG-WARN (fi-ivb-3770)

Stil, there: https://bugs.freedesktop.org/show_bug.cgi?id=98228

Quoted 9 lines

fi-ivb-3770 total:247 pass:224 dwarn:1 dfail:0 fail:0 skip:22

a17bd268c6f9e1076a9aea44573ff58b0caeb44d drm-tip: 2016y-12m-05d-11h- 19m-41s UTC integration manifest
6987287 drm/i915: Creating guc log file in drmfs instead of debugfs 50a4a18 drm: Create driver specific root directory inside drmfs df2fc99 drm: Register drmfs filesystem from drm init d340379 drm: Introduce drmfs pseudo filesystem interfaces

Quoted 3 lines

== Logs ==

For more details see: https://intel-gfx-ci.01.org/CI/Patchwork_3186/

Jani Saarinen
Intel Finland Oy - BIC 0357606-4 - Westendinkatu 7, 02160 Espoo

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

#7 lore raw
In reply to [email protected]
Quoted 75 lines · Swati

From: Swati Dhingra <[email protected]>

Currently, we don't have a stable ABI which can be used for the purpose of providing output debug/loggging/crc and other such data from DRM. The ABI in current use (filesystems, ioctls, et al.) have their own constraints and are intended to output a particular type of data. Few cases in point:
sysfs - stable ABI, but constrained to one textual value per file
debugfs - unstable ABI, free-for-all
ioctls - not as suitable to many single purpose continuous data
dumping, we would very quickly run out ioctl space; requires more userspace support than "cat"
device nodes - a real possibilty, kernel instantiation is more tricky,
requires udev (+udev.rules) or userspace discovery of the dynamic major:minor (via sysfs) [mounting a registered filesystem is easy in comparison]
netlink - stream based, therefore involves numerous copies.

Debugfs is the lesser among the evils here, thereby we have grown used to the convenience and flexibility in presentation that debugfs gives us (including relayfs inodes) that we want some of that hierachy in stable user ABI form.

Due to these limitations, there is a need for a new pseudo filesytem, that would act as a stable 'free-for-all' ABI, with the heirarchial structure and thus convenience of debugfs. This will be managed by drm, thus named 'drmfs'. DRM would register this filesystem to manage a canonical mountpoint, but this wouldn't limit everyone to only using that pseudofs underneath.

This can serve to hold various kinds of output data from Linux DRM subsystems, for the files which can't truely fit anywhere else with existing ABI's but present so, for the lack of a better place.

In this patch series, we have introduced a pseudo filesystem named as 'drmfs' for now. The filesystem is introduced in the first patch, and the subsequent patches make use of the filesystem interfaces, in drm driver, and making them available for use by the drm subsystem components, one of which is i915. We've moved the location of i915 GuC logs from debugfs to drmfs in the last patch. Subsequently, more such files such as pipe_crc, error states, memory stats, etc. can be move to this filesystem, if the idea introduced here is acceptable per se. The filesystem introduced is being used to house the data generated by i915 driver in this patch series, but will hopefully be generic enough to provide scope for usage by any other drm subsystem component.

The patch series is being floated as RFC to gather feedback on the idea and infrastructure proposed here and it's suitability to address the specific problem statement/use case.

TODO: Create documentation. Will do so in next version.

v2: fix the bat failures caused due to missing config check

v3: Changes made:
- Move the location of drmfs from fs/ to drivers/gpu/drm/ (Chris)
- Moving config checks to header (Chris,Daniel)

Sourab Gupta (4):
drm: Introduce drmfs pseudo filesystem interfaces
drm: Register drmfs filesystem from drm init
drm: Create driver specific root directory inside drmfs drm/i915: Creating guc log file in drmfs instead of debugfs

 drivers/gpu/drm/Kconfig                    |   9 +
 drivers/gpu/drm/Makefile                   |   1 +
 drivers/gpu/drm/drm_drv.c                  |  12 +
 drivers/gpu/drm/drmfs.c                    | 555 +++++++++++++++++++++++++++++
 drivers/gpu/drm/i915/i915_guc_submission.c |  33 +-
 include/drm/drm_drv.h                      |   3 +
 include/drm/drmfs.h                        |  77 ++++
 include/uapi/linux/magic.h                 |   3 +
 8 files changed, 672 insertions(+), 21 deletions(-)
 create mode 100644 drivers/gpu/drm/drmfs.c
 create mode 100644 include/drm/drmfs.h
Quoted 3 lines

--
2.7.4

Hi dri-devel folks,

Any feedback on the proposed drmfs infrastructure being proposed here?

Regards,
Sourab

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

#8 lore raw
Quoted 76 lines · sourab gupta
Quoted 75 lines · Swati

From: Swati Dhingra <[email protected]>

Currently, we don't have a stable ABI which can be used for the purpose of providing output debug/loggging/crc and other such data from DRM. The ABI in current use (filesystems, ioctls, et al.) have their own constraints and are intended to output a particular type of data. Few cases in point:
sysfs - stable ABI, but constrained to one textual value per file
debugfs - unstable ABI, free-for-all
ioctls - not as suitable to many single purpose continuous data
dumping, we would very quickly run out ioctl space; requires more userspace support than "cat"
device nodes - a real possibilty, kernel instantiation is more tricky,
requires udev (+udev.rules) or userspace discovery of the dynamic major:minor (via sysfs) [mounting a registered filesystem is easy in comparison]
netlink - stream based, therefore involves numerous copies.

Debugfs is the lesser among the evils here, thereby we have grown used to the convenience and flexibility in presentation that debugfs gives us (including relayfs inodes) that we want some of that hierachy in stable user ABI form.

Due to these limitations, there is a need for a new pseudo filesytem, that would act as a stable 'free-for-all' ABI, with the heirarchial structure and thus convenience of debugfs. This will be managed by drm, thus named 'drmfs'. DRM would register this filesystem to manage a canonical mountpoint, but this wouldn't limit everyone to only using that pseudofs underneath.

This can serve to hold various kinds of output data from Linux DRM subsystems, for the files which can't truely fit anywhere else with existing ABI's but present so, for the lack of a better place.

In this patch series, we have introduced a pseudo filesystem named as 'drmfs' for now. The filesystem is introduced in the first patch, and the subsequent patches make use of the filesystem interfaces, in drm driver, and making them available for use by the drm subsystem components, one of which is i915. We've moved the location of i915 GuC logs from debugfs to drmfs in the last patch. Subsequently, more such files such as pipe_crc, error states, memory stats, etc. can be move to this filesystem, if the idea introduced here is acceptable per se. The filesystem introduced is being used to house the data generated by i915 driver in this patch series, but will hopefully be generic enough to provide scope for usage by any other drm subsystem component.

The patch series is being floated as RFC to gather feedback on the idea and infrastructure proposed here and it's suitability to address the specific problem statement/use case.

TODO: Create documentation. Will do so in next version.

v2: fix the bat failures caused due to missing config check

v3: Changes made:
- Move the location of drmfs from fs/ to drivers/gpu/drm/ (Chris)
- Moving config checks to header (Chris,Daniel)

Sourab Gupta (4):
drm: Introduce drmfs pseudo filesystem interfaces
drm: Register drmfs filesystem from drm init
drm: Create driver specific root directory inside drmfs drm/i915: Creating guc log file in drmfs instead of debugfs

 drivers/gpu/drm/Kconfig                    |   9 +
 drivers/gpu/drm/Makefile                   |   1 +
 drivers/gpu/drm/drm_drv.c                  |  12 +
 drivers/gpu/drm/drmfs.c                    | 555 +++++++++++++++++++++++++++++
 drivers/gpu/drm/i915/i915_guc_submission.c |  33 +-
 include/drm/drm_drv.h                      |   3 +
 include/drm/drmfs.h                        |  77 ++++
 include/uapi/linux/magic.h                 |   3 +
 8 files changed, 672 insertions(+), 21 deletions(-)
 create mode 100644 drivers/gpu/drm/drmfs.c
 create mode 100644 include/drm/drmfs.h
Quoted 7 lines
Quoted 3 lines

--
2.7.4

Hi dri-devel folks,

Any feedback on the proposed drmfs infrastructure being proposed here?

I haven't looked too closely, but so far we haven't run into anything that we haven't been able to support via some combination of sysfs and debugfs.

Alex
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

#9 lore raw
Quoted 83 lines · Alex Deucher
Quoted 82 lines · sourab gupta
Quoted 75 lines · Swati

From: Swati Dhingra <[email protected]>

Currently, we don't have a stable ABI which can be used for the purpose of providing output debug/loggging/crc and other such data from DRM. The ABI in current use (filesystems, ioctls, et al.) have their own constraints and are intended to output a particular type of data. Few cases in point:
sysfs - stable ABI, but constrained to one textual value per file
debugfs - unstable ABI, free-for-all
ioctls - not as suitable to many single purpose continuous data
dumping, we would very quickly run out ioctl space; requires more userspace support than "cat"
device nodes - a real possibilty, kernel instantiation is more tricky,
requires udev (+udev.rules) or userspace discovery of the dynamic major:minor (via sysfs) [mounting a registered filesystem is easy in comparison]
netlink - stream based, therefore involves numerous copies.

Debugfs is the lesser among the evils here, thereby we have grown used to the convenience and flexibility in presentation that debugfs gives us (including relayfs inodes) that we want some of that hierachy in stable user ABI form.

Due to these limitations, there is a need for a new pseudo filesytem, that would act as a stable 'free-for-all' ABI, with the heirarchial structure and thus convenience of debugfs. This will be managed by drm, thus named 'drmfs'. DRM would register this filesystem to manage a canonical mountpoint, but this wouldn't limit everyone to only using that pseudofs underneath.

This can serve to hold various kinds of output data from Linux DRM subsystems, for the files which can't truely fit anywhere else with existing ABI's but present so, for the lack of a better place.

In this patch series, we have introduced a pseudo filesystem named as 'drmfs' for now. The filesystem is introduced in the first patch, and the subsequent patches make use of the filesystem interfaces, in drm driver, and making them available for use by the drm subsystem components, one of which is i915. We've moved the location of i915 GuC logs from debugfs to drmfs in the last patch. Subsequently, more such files such as pipe_crc, error states, memory stats, etc. can be move to this filesystem, if the idea introduced here is acceptable per se. The filesystem introduced is being used to house the data generated by i915 driver in this patch series, but will hopefully be generic enough to provide scope for usage by any other drm subsystem component.

The patch series is being floated as RFC to gather feedback on the idea and infrastructure proposed here and it's suitability to address the specific problem statement/use case.

TODO: Create documentation. Will do so in next version.

v2: fix the bat failures caused due to missing config check

v3: Changes made:
- Move the location of drmfs from fs/ to drivers/gpu/drm/ (Chris)
- Moving config checks to header (Chris,Daniel)

Sourab Gupta (4):
drm: Introduce drmfs pseudo filesystem interfaces
drm: Register drmfs filesystem from drm init
drm: Create driver specific root directory inside drmfs drm/i915: Creating guc log file in drmfs instead of debugfs

 drivers/gpu/drm/Kconfig                    |   9 +
 drivers/gpu/drm/Makefile                   |   1 +
 drivers/gpu/drm/drm_drv.c                  |  12 +
 drivers/gpu/drm/drmfs.c                    | 555 +++++++++++++++++++++++++++++
 drivers/gpu/drm/i915/i915_guc_submission.c |  33 +-
 include/drm/drm_drv.h                      |   3 +
 include/drm/drmfs.h                        |  77 ++++
 include/uapi/linux/magic.h                 |   3 +
 8 files changed, 672 insertions(+), 21 deletions(-)
 create mode 100644 drivers/gpu/drm/drmfs.c
 create mode 100644 include/drm/drmfs.h
Quoted 3 lines

--
2.7.4

Hi dri-devel folks,

Quoted 7 lines
Quoted 1 line

Any feedback on the proposed drmfs infrastructure being proposed here?

I haven't looked too closely, but so far we haven't run into anything that we haven't been able to support via some combination of sysfs and debugfs.

Alex

Hi Alex,

The intent here is to introduce a stable free-for-all ABI which has the hierarchical structure and convenience of debugfs. One of the possible candidate/usecase here is for holding the GuC (micro-controller) logs. Currently, we're using relayfs to provide these logs to userspace. Relayfs needs a parent dentry to be associated with log file, and also we need these logs to be available in production kernels. In absence of any other suitable candidate, debugfs is used currently, since it provides the heirarchial VFS structures available (dentry et al.). But still we felt the need for a stable ABI, which is modeled on lines of debugfs but which we can control fully.

Chris,
Can you add on if I'm not fully clear.

Regards,
Sourab

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx