Skip to contentSkip to conversation

debugfs:Fix terminology inconsistency of dir name to mount debugfs filesystem.

RFC PATCH Superseded GeunSik Lim 5 participants 29 messages May 26, 2009 – Jun 2, 2009

Subsystems (12)
lore
Share & tools
lore
#0 lore raw

RFC PATCH 1/1 debugfs:Fix terminology inconsistency of dir name to mount debugfs filesystem.

Hi all,

I want to hear your opinions about this proposal for consistency of mount directory to mount debugfs filesystem that need by ftrace infrastructure. Um... I appended "RFC" word in the subject because this mail is my proposal just.

commit 1766a83645706100b1829bb422f852ec757c6f4b

Author: GeunSik,Lim <[email protected]>
Date:   Tue May 26 13:49:23 2009 +0900

Fix terminology inconsistency of dir name to mount debugfs filesystem.

Many developers use "/debug/" directory name to mount debugfs filesystem for ftrace according to ./Documentation/tracers/ftrace.txt file.

But, two directory names(ex:/debug/, /debugfs/) is existed in kernel source like ftrace, DRM, Wireless, Network[sky2]files to mount debugfs filesystem. I think that we don't have to select either "/debug/" directory name or "/debugfs/" directory name for terminology consistency.

debugfs means debug filesystem for debugging easy to use by greg kroah hartman. "/debug/" name is suitable as directory name of debugfs for tracing using ftrace like "/sys/" of sysfs(system filesystem). -debugfs related reference: http://lwn.net/Articles/115405/

Fix inconsistency of directory name according to duplicated expression to mount debugfs.

---
 Documentation/DocBook/debugobjects.tmpl |    2 +-
 drivers/block/pktcdvd.c                 |    2 +-
 drivers/gpu/drm/drm_debugfs.c           |   12 ++++++------
 drivers/gpu/drm/drm_drv.c               |    2 +-
 drivers/gpu/drm/drm_stub.c              |    2 +-
 drivers/net/Kconfig                     |    4 ++--
 drivers/net/wimax/i2400m/i2400m.h       |    2 +-
 drivers/net/wireless/libertas/README    |   12 ++++++------
 include/linux/kernel.h                  |    2 +-
 kernel/trace/Kconfig                    |   10 +++++-----
 kernel/trace/trace.c                    |    2 +-
 11 files changed, 26 insertions(+), 26 deletions(-)
Documentation/DocBook/debugobjects.tmpl +1 -1
index 7f5f218..9beb77a 100644
@@ -106,7 +106,7 @@
       number of errors are printk'ed including a full stack trace.
     </para>
     <para>
-      The statistics are available via debugfs/debug_objects/stats.
+      The statistics are available via /debug/debug_objects/stats.
       They provide information about the number of warnings and the
       number of successful fixups along with information about the
       usage of the internal tracking objects and the state of the
drivers/block/pktcdvd.c +1 -1
index dc7a8c3..36989b0 100644
@@ -430,7 +430,7 @@ static void pkt_sysfs_cleanup(void)
 /********************************************************************
   entries in debugfs
 
-  /debugfs/pktcdvd[0-7]/
+  /debug/pktcdvd[0-7]/
 			info
 
  *******************************************************************/
drivers/gpu/drm/drm_debugfs.c +6 -6
index c77c6c6..04a2ccf 100644
@@ -105,7 +105,7 @@ int drm_debugfs_create_files(struct drm_info_list *files, int count,
 		ent = debugfs_create_file(files[i].name, S_IFREG | S_IRUGO,
 					  root, tmp, &drm_debugfs_fops);
 		if (!ent) {
-			DRM_ERROR("Cannot create /debugfs/dri/%s/%s\n",
+			DRM_ERROR("Cannot create /debug/dri/%s/%s\n",
 				  name, files[i].name);
 			drm_free(tmp, sizeof(struct drm_info_node),
 				 _DRM_DRIVER);
@@ -133,9 +133,9 @@ EXPORT_SYMBOL(drm_debugfs_create_files);
  * \param minor device minor number
  * \param root DRI debugfs dir entry.
  *
- * Create the DRI debugfs root entry "/debugfs/dri", the device debugfs root entry
- * "/debugfs/dri/%minor%/", and each entry in debugfs_list as
- * "/debugfs/dri/%minor%/%name%".
+ * Create the DRI debugfs root entry "/debug/dri", the device debugfs root entry
+ * "/debug/dri/%minor%/", and each entry in debugfs_list as
+ * "/debug/dri/%minor%/%name%".
  */
 int drm_debugfs_init(struct drm_minor *minor, int minor_id,
 		     struct dentry *root)
@@ -148,7 +148,7 @@ int drm_debugfs_init(struct drm_minor *minor, int minor_id,
 	sprintf(name, "%d", minor_id);
 	minor->debugfs_root = debugfs_create_dir(name, root);
 	if (!minor->debugfs_root) {
-		DRM_ERROR("Cannot create /debugfs/dri/%s\n", name);
+		DRM_ERROR("Cannot create /debug/dri/%s\n", name);
 		return -1;
 	}
 
@@ -165,7 +165,7 @@ int drm_debugfs_init(struct drm_minor *minor, int minor_id,
 		ret = dev->driver->debugfs_init(minor);
 		if (ret) {
 			DRM_ERROR("DRM: Driver failed to initialize "
-				  "/debugfs/dri.\n");
+				  "/debug/dri.\n");
 			return ret;
 		}
 	}
drivers/gpu/drm/drm_drv.c +1 -1
index 019b7c5..5472e67 100644
@@ -339,7 +339,7 @@ static int __init drm_core_init(void)
 
 	drm_debugfs_root = debugfs_create_dir("dri", NULL);
 	if (!drm_debugfs_root) {
-		DRM_ERROR("Cannot create /debugfs/dri\n");
+		DRM_ERROR("Cannot create /debug/dri\n");
 		ret = -1;
 		goto err_p3;
 	}
drivers/gpu/drm/drm_stub.c +1 -1
index b9631e3..9488769 100644
@@ -328,7 +328,7 @@ static int drm_get_minor(struct drm_device *dev, struct drm_minor **minor, int t
 #if defined(CONFIG_DEBUG_FS)
 	ret = drm_debugfs_init(new_minor, minor_id, drm_debugfs_root);
 	if (ret) {
-		DRM_ERROR("DRM: Failed to initialize /debugfs/dri.\n");
+		DRM_ERROR("DRM: Failed to initialize /debug/dri.\n");
 		goto err_g2;
 	}
 #endif
drivers/net/Kconfig +2 -2
index 214a92d..898c269 100644
@@ -2200,7 +2200,7 @@ config SKGE_DEBUG
        depends on SKGE && DEBUG_FS
        help
 	 This option adds the ability to dump driver state for debugging.
-	 The file debugfs/skge/ethX displays the state of the internal
+	 The file /debug/skge/ethX displays the state of the internal
 	 transmit and receive rings.
 
 	 If unsure, say N.
@@ -2226,7 +2226,7 @@ config SKY2_DEBUG
        depends on SKY2 && DEBUG_FS
        help
 	 This option adds the ability to dump driver state for debugging.
-	 The file debugfs/sky2/ethX displays the state of the internal
+	 The file /debug/sky2/ethX displays the state of the internal
 	 transmit and receive rings.
 
 	 If unsure, say N.
drivers/net/wimax/i2400m/i2400m.h +1 -1
index 3ae2df3..2366696 100644
@@ -389,7 +389,7 @@ struct i2400m {
 	unsigned ready:1;		/* all probing steps done */
 	unsigned rx_reorder:1;		/* RX reorder is enabled */
 	u8 trace_msg_from_user;		/* echo rx msgs to 'trace' pipe */
-					/* typed u8 so debugfs/u8 can tweak */
+					/* typed u8 so /debug/u8 can tweak */
 	enum i2400m_system_state state;
 	wait_queue_head_t state_wq;	/* Woken up when on state updates */
 
drivers/net/wireless/libertas/README +6 -6
index d860fc3..11616ba 100644
@@ -72,7 +72,7 @@ rdrf
 	location that is to be read.  This parameter must be specified in
 	hexadecimal (its possible to preceed preceding the number with a "0x").
 
-	Path: /debugfs/libertas_wireless/ethX/registers/
+	Path: /debug/libertas_wireless/ethX/registers/
 
 	Usage:
 		echo "0xa123" > rdmac ; cat rdmac
@@ -95,7 +95,7 @@ wrrf
 sleepparams
 	This command is used to set the sleepclock configurations
 
-	Path: /debugfs/libertas_wireless/ethX/
+	Path: /debug/libertas_wireless/ethX/
 
 	Usage:
 		cat sleepparams: reads the current sleepclock configuration
@@ -115,7 +115,7 @@ subscribed_events
 	The subscribed_events directory contains the interface for the
 	subscribed events API.
 
-	Path: /debugfs/libertas_wireless/ethX/subscribed_events/
+	Path: /debug/libertas_wireless/ethX/subscribed_events/
 
 	Each event is represented by a filename. Each filename consists of the
 	following three fields:
@@ -165,7 +165,7 @@ subscribed_events
 extscan
 	This command is used to do a specific scan.
 
-	Path: /debugfs/libertas_wireless/ethX/
+	Path: /debug/libertas_wireless/ethX/
 
 	Usage: echo "SSID" > extscan
 
@@ -179,7 +179,7 @@ getscantable
 	Display the current contents of the driver scan table (ie. get the
 	scan results).
 
-	Path: /debugfs/libertas_wireless/ethX/
+	Path: /debug/libertas_wireless/ethX/
 
 	Usage:
 		cat getscantable
@@ -188,7 +188,7 @@ setuserscan
 	Initiate a customized scan and retrieve the results
 
 
-	Path: /debugfs/libertas_wireless/ethX/
+	Path: /debug/libertas_wireless/ethX/
 
     Usage:
        echo "[ARGS]" > setuserscan
include/linux/kernel.h +1 -1
index 883cd44..7ffa753 100644
@@ -406,7 +406,7 @@ static inline char *pack_hex_byte(char *buf, u8 byte)
  *
  * Use tracing_on/tracing_off when you want to quickly turn on or off
  * tracing. It simply enables or disables the recording of the trace events.
- * This also corresponds to the user space debugfs/tracing/tracing_on
+ * This also corresponds to the user space /debug/tracing/tracing_on
  * file, which gives a means for the kernel and userspace to interact.
  * Place a tracing_off() in the kernel where you want tracing to end.
  * From user space, examine the trace, and then echo 1 > tracing_on
kernel/trace/Kconfig +5 -5
index 417d198..7fcc3d6 100644
@@ -120,7 +120,7 @@ config IRQSOFF_TRACER
 	  disabled by default and can be runtime (re-)started
 	  via:
 
-	      echo 0 > /debugfs/tracing/tracing_max_latency
+	      echo 0 > /debug/tracing/tracing_max_latency
 
 	  (Note that kernel size and overhead increases with this option
 	  enabled. This option and the preempt-off timing option can be
@@ -141,7 +141,7 @@ config PREEMPT_TRACER
 	  disabled by default and can be runtime (re-)started
 	  via:
 
-	      echo 0 > /debugfs/tracing/tracing_max_latency
+	      echo 0 > /debug/tracing/tracing_max_latency
 
 	  (Note that kernel size and overhead increases with this option
 	  enabled. This option and the irqs-off timing option can be
@@ -213,7 +213,7 @@ config TRACE_BRANCH_PROFILING
 	  This tracer profiles all the the likely and unlikely macros
 	  in the kernel. It will display the results in:
 
-	  /debugfs/tracing/profile_annotated_branch
+	  /debug/tracing/profile_annotated_branch
 
 	  Note: this will add a significant overhead, only turn this
 	  on if you need to profile the system's use of these macros.
@@ -228,7 +228,7 @@ config PROFILE_ALL_BRANCHES
 	  taken in the kernel is recorded whether it hit or miss.
 	  The results will be displayed in:
 
-	  /debugfs/tracing/profile_branch
+	  /debug/tracing/profile_branch
 
 	  This configuration, when enabled, will impose a great overhead
 	  on the system. This should only be enabled when the system
@@ -276,7 +276,7 @@ config STACK_TRACER
 	select KALLSYMS
 	help
 	  This special tracer records the maximum stack footprint of the
-	  kernel and displays it in debugfs/tracing/stack_trace.
+	  kernel and displays it in /debug/tracing/stack_trace.
 
 	  This tracer works by hooking into every function call that the
 	  kernel executes, and keeping a maximum stack depth value and
kernel/trace/trace.c +1 -1
index cda81ec..b474e8c 100644
@@ -335,7 +335,7 @@ static raw_spinlock_t ftrace_max_lock =
 /*
  * Copy the new maximum trace into the separate maximum-trace
  * structure. (this way the maximum trace is permanently saved,
- * for later retrieval via /debugfs/tracing/latency_trace)
+ * for later retrieval via /debug/tracing/latency_trace)
  */
 static void
 __update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu)

-----------------------------------------------
To unsubscribe from this list: send the line "unsubscribe linux-***" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

GeunSik Lim (ELS - OS Group - S/W Lab - SAIT - SAMSUNG) e-Mail :1) [email protected]
2) [email protected] , [email protected]
HomePage: http://intomail.dnip.net/invain/me/
-----------------------------------------------

Acked-by Inaky Perez-Gonzalez on v1 1/1
#1 lore raw
Quoted 27 lines · GeunSik Lim

Hi all,

I want to hear your opinions about this proposal for consistency of mount directory to mount debugfs filesystem that need by ftrace
infrastructure. Um... I appended "RFC" word in the subject because this mail is my proposal just.

commit 1766a83645706100b1829bb422f852ec757c6f4b

Author: GeunSik,Lim <[email protected]>
Date:   Tue May 26 13:49:23 2009 +0900

Fix terminology inconsistency of dir name to mount debugfs filesystem.

Many developers use "/debug/" directory name to mount debugfs
filesystem for ftrace according to ./Documentation/tracers/ftrace.txt file.

But, two directory names(ex:/debug/, /debugfs/) is existed in kernel
source like ftrace, DRM, Wireless, Network[sky2]files to mount debugfs filesystem. I think that we don't have to select either "/debug/" directory name or "/debugfs/" directory name for terminology consistency.

debugfs means debug filesystem for debugging easy to use by greg
kroah hartman. "/debug/" name is suitable as directory name of debugfs for tracing using ftrace like "/sys/" of sysfs(system filesystem). -debugfs related reference: http://lwn.net/Articles/115405/

Quoted 4 lines

Fix inconsistency of directory name according to duplicated
expression to mount debugfs.

For the WiMAX part, not really an issue:

maintainer
#2 lore raw
In reply to GeunSik Lim
Quoted 28 lines · GeunSik Lim

Hi all,

I want to hear your opinions about this proposal for consistency of mount directory to mount debugfs filesystem that need by ftrace infrastructure. Um... I appended "RFC" word in the subject because this mail is my proposal just.

commit 1766a83645706100b1829bb422f852ec757c6f4b

Author: GeunSik,Lim <[email protected]>
Date:   Tue May 26 13:49:23 2009 +0900

Fix terminology inconsistency of dir name to mount debugfs filesystem.

Many developers use "/debug/" directory name to mount debugfs filesystem for ftrace according to ./Documentation/tracers/ftrace.txt file.

But, two directory names(ex:/debug/, /debugfs/) is existed in kernel source like ftrace, DRM, Wireless, Network[sky2]files to mount debugfs filesystem. I think that we don't have to select either "/debug/" directory name or "/debugfs/" directory name for terminology consistency.

debugfs means debug filesystem for debugging easy to use by greg kroah hartman. "/debug/" name is suitable as directory name of debugfs for tracing using ftrace like "/sys/" of sysfs(system filesystem). -debugfs related reference: http://lwn.net/Articles/115405/

Quoted 4 lines

Fix inconsistency of directory name according to duplicated expression to mount debugfs.

I'm guilty of using both /debug and /debugfs in comments. In practice, I only use /debug.

But...

The filesystem Nazi's out there will argue that the proper place to mount the debugfs filesystem is /sys/kernel/debug. I personally find that location annoying.

-- Steve

#3 lore raw
Quoted 36 lines · Steven Rostedt
Quoted 28 lines · GeunSik Lim

Hi all,

I want to hear your opinions about this proposal for consistency of mount directory to mount debugfs filesystem that need by ftrace infrastructure. Um... I appended "RFC" word in the subject because this mail is my proposal just.

commit 1766a83645706100b1829bb422f852ec757c6f4b

Author: GeunSik,Lim <[email protected]>
Date:   Tue May 26 13:49:23 2009 +0900

Fix terminology inconsistency of dir name to mount debugfs filesystem.

Many developers use "/debug/" directory name to mount debugfs filesystem for ftrace according to ./Documentation/tracers/ftrace.txt file.

But, two directory names(ex:/debug/, /debugfs/) is existed in kernel source like ftrace, DRM, Wireless, Network[sky2]files to mount debugfs filesystem. I think that we don't have to select either "/debug/" directory name or "/debugfs/" directory name for terminology consistency.

debugfs means debug filesystem for debugging easy to use by greg kroah hartman. "/debug/" name is suitable as directory name of debugfs for tracing using ftrace like "/sys/" of sysfs(system filesystem). -debugfs related reference: http://lwn.net/Articles/115405/

Quoted 4 lines

Fix inconsistency of directory name according to duplicated expression to mount debugfs.

Quoted 8 lines

I'm guilty of using both /debug and /debugfs in comments. In practice, I only use /debug.

But...

The filesystem Nazi's out there will argue that the proper place to mount the debugfs filesystem is /sys/kernel/debug. I personally find that location annoying.

Why?

When I created debugfs, I was told that I had to pick a place to mount it so that everyone could write scripts that "knew" where it would be.

So we picked /sys/kernel/debug/ and the code creates the mount point to put it there. Now distros actually mount debugfs at that location.

So, now you don't like this? After it's become a standard? Ick.

Becides, /debug/ is pretty presumptious, it's as if debugging the kernel is a system-wide thing :)

So please, don't use /debug/ use the standard, defined, and ACTUALLY USED BY DISTROS TODAY mount point of /sys/kernel/debug/

If you want to create your own symlink from /debug/ to that location to make it easier to type, feel free, but do not hard code any scripts to use that location, or they are going to break.

thanks,

greg k-h

maintainer
#4 lore raw
Quoted 6 lines · Greg KH
Quoted 4 lines

The filesystem Nazi's out there will argue that the proper place to mount the debugfs filesystem is /sys/kernel/debug. I personally find that location annoying.

Why?

Because it is too much to type ;-)

Quoted 12 lines

When I created debugfs, I was told that I had to pick a place to mount it so that everyone could write scripts that "knew" where it would be.

So we picked /sys/kernel/debug/ and the code creates the mount point to put it there. Now distros actually mount debugfs at that location.

So, now you don't like this? After it's become a standard? Ick.

Becides, /debug/ is pretty presumptious, it's as if debugging the kernel is a system-wide thing :)

Quoted 6 lines

So please, don't use /debug/ use the standard, defined, and ACTUALLY USED BY DISTROS TODAY mount point of /sys/kernel/debug/

If you want to create your own symlink from /debug/ to that location to make it easier to type, feel free, but do not hard code any scripts to use that location, or they are going to break.

All my scripts search /proc/mounts to find where debugfs is mounted, then it uses that.

Hmm, Not sure what packages in the distros use the default. I don't see the debugfs in /etc/fstab in any of my machines.

-- Steve

#5 lore raw
Quoted 2 lines · Steven Rostedt

All my scripts search /proc/mounts to find where debugfs is mounted, then it uses that.

Good.

Quoted 2 lines

Hmm, Not sure what packages in the distros use the default. I don't see the debugfs in /etc/fstab in any of my machines.

It's in mine, I don't know what packages use it at the moment.

thanks,

greg k-h

#6 lore raw
In reply to Steven Rostedt
Quoted 10 lines · Steven Rostedt
Quoted 6 lines · Greg KH
Quoted 4 lines

The filesystem Nazi's out there will argue that the proper place to mount the debugfs filesystem is /sys/kernel/debug. I personally find that location annoying.

Why?

Because it is too much to type ;-)

I think so. But, when we consider many developers that use linux kernel source all over the world, I think that a regular rules or consistency is important
as greg k-h explained.

Recently, I read "ftrace-world.odp" presentation file of steven rostedt.

For example, about Mounting debugfs...
#> mount -t debugfs nodev /sys/kernel/debug
OR
#> mkdir /debug
#> mount -t debugfs nodev /debug

I want that we will reach a consensus about mount directory name of debugfs filesystem to avoid confusing among the kernel developers.

/debug/ ?
or /sys/kernel/debug/?
.....

Frankly speaking, I like /sys/kernel/debug/ rule is better than /debug/ rule for consistency.

Quoted 20 lines
Quoted 13 lines

When I created debugfs, I was told that I had to pick a place to mount it so that everyone could write scripts that "knew" where it would be.

So we picked /sys/kernel/debug/ and the code creates the mount point to put it there.  Now distros actually mount debugfs at that location.

So, now you don't like this?  After it's become a standard?  Ick.

Becides, /debug/ is pretty presumptious, it's as if debugging the kernel is a system-wide thing :)

Quoted 6 lines

So please, don't use /debug/ use the standard, defined, and ACTUALLY USED BY DISTROS TODAY mount point of /sys/kernel/debug/

If you want to create your own symlink from /debug/ to that location to make it easier to type, feel free, but do not hard code any scripts to use that location, or they are going to break.

Quoted 6 lines

All my scripts search /proc/mounts to find where debugfs is mounted, then it uses that.

Hmm, Not sure what packages in the distros use the default. I don't see the debugfs in /etc/fstab in any of my machines.

Yes.. In my case, I use linux distributions like Ubuntu 8.10, Fedora 9/10, CENTOS 5.3 currently. This distributions use proc , tmpfs, devpts, sysfs for official release
by default.

Quoted 3 lines

-- Steve

Signature · 7 lines
Regards,
GeunSik Lim
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/
maintainer
#7 lore raw
Quoted 27 lines · GeunSik Lim
Quoted 10 lines · Steven Rostedt
Quoted 6 lines · Greg KH
Quoted 4 lines

The filesystem Nazi's out there will argue that the proper place to mount the debugfs filesystem is /sys/kernel/debug. I personally find that location annoying.

Why?

Because it is too much to type ;-)

I think so. But, when we consider many developers that use linux kernel source all over the world, I think that a regular rules or consistency is important
as greg k-h explained.

Recently, I read "ftrace-world.odp" presentation file of steven rostedt.

For example, about Mounting debugfs...
#> mount -t debugfs nodev /sys/kernel/debug
OR
#> mkdir /debug
#> mount -t debugfs nodev /debug

I want that we will reach a consensus about mount directory name of debugfs filesystem to avoid confusing among the kernel developers.

Quoted 6 lines

/debug/ ?
or /sys/kernel/debug/?
.....

Frankly speaking, I like /sys/kernel/debug/ rule is better than /debug/ rule for consistency.

And this is why I always mention that directory.

Quoted 24 lines
Quoted 24 lines
Quoted 14 lines

When I created debugfs, I was told that I had to pick a place to mount it so that everyone could write scripts that "knew" where it would be.

So we picked /sys/kernel/debug/ and the code creates the mount point to put it there.  Now distros actually mount debugfs at that location.

So, now you don't like this?  After it's become a standard?  Ick.

Becides, /debug/ is pretty presumptious, it's as if debugging the kernel is a system-wide thing :)

Quoted 6 lines

So please, don't use /debug/ use the standard, defined, and ACTUALLY USED BY DISTROS TODAY mount point of /sys/kernel/debug/

If you want to create your own symlink from /debug/ to that location to make it easier to type, feel free, but do not hard code any scripts to use that location, or they are going to break.

All my scripts search /proc/mounts to find where debugfs is mounted, then it uses that.

Quoted 8 lines
Quoted 3 lines

Hmm, Not sure what packages in the distros use the default. I don't see the debugfs in /etc/fstab in any of my machines.

Yes.. In my case, I use linux distributions like Ubuntu 8.10, Fedora 9/10, CENTOS 5.3 currently. This distributions use proc , tmpfs, devpts, sysfs for official release
by default.

The key thing is not to have scripts depend on a path. When debugfs is enabled in a kernel, you automatically get /sys/kernel/debug, so that is a natural place to mount it.

I have also suggested to people (as Greg suggested earlier to me) that you can mount it there and then make a soft link to it. Which also works.

I personally just mount it at /debug.

To handle the scripts problem, all my scripts start with:

# grep debugfs /proc/mounts | cut -d ' ' -f 2

I also have a C function to do a search for debugfs in /proc/mounts.

If you look in Documentation/trace/ftrace.txt I have:

===
To mount the debugfs system:

  # mkdir /debug
  # mount -t debugfs nodev /debug

( Note: it is more common to mount at /sys/kernel/debug, but for
simplicity this document will use /debug)
===

Thus it is a personal preference of mine to put it at /debug. But I do not hide the fact that it is recommended to be in /sys/kernel/debug. Heck, I'm the one that brought to light that path in this thread. I just curse it everytime I have to say it ;-)

-- Steve

#8 lore raw
Quoted 97 lines · Steven Rostedt
Quoted 27 lines · GeunSik Lim
Quoted 10 lines · Steven Rostedt
Quoted 6 lines · Greg KH
Quoted 4 lines

The filesystem Nazi's out there will argue that the proper place to mount the debugfs filesystem is /sys/kernel/debug. I personally find that location annoying.

Why?

Because it is too much to type ;-)

I think so. But, when we consider many developers that use linux kernel source all over the world, I think that a regular rules or consistency is important
as greg k-h explained.

Recently, I read "ftrace-world.odp" presentation file  of steven rostedt.

For example, about Mounting debugfs...
#> mount -t debugfs nodev /sys/kernel/debug
OR
#> mkdir /debug
#> mount -t debugfs nodev /debug

I want that we will reach a consensus about mount directory name of debugfs filesystem to avoid confusing among the kernel developers.

Quoted 6 lines

/debug/ ?
or   /sys/kernel/debug/?
 .....

Frankly speaking, I like /sys/kernel/debug/ rule is better than /debug/ rule for consistency.

And this is why I always mention that directory.

Quoted 24 lines
Quoted 24 lines
Quoted 14 lines

When I created debugfs, I was told that I had to pick a place to mount it so that everyone could write scripts that "knew" where it would be.

So we picked /sys/kernel/debug/ and the code creates the mount point to put it there.  Now distros actually mount debugfs at that location.

So, now you don't like this?  After it's become a standard?  Ick.

Becides, /debug/ is pretty presumptious, it's as if debugging the kernel is a system-wide thing :)

Quoted 6 lines

So please, don't use /debug/ use the standard, defined, and ACTUALLY USED BY DISTROS TODAY mount point of /sys/kernel/debug/

If you want to create your own symlink from /debug/ to that location to make it easier to type, feel free, but do not hard code any scripts to use that location, or they are going to break.

All my scripts search /proc/mounts to find where debugfs is mounted, then it uses that.

Quoted 8 lines
Quoted 3 lines

Hmm, Not sure what packages in the distros use the default. I don't see the debugfs in /etc/fstab in any of my machines.

Yes.. In my case, I use linux distributions like Ubuntu 8.10, Fedora 9/10, CENTOS 5.3 currently. This distributions  use proc , tmpfs, devpts, sysfs for official release
by default.

The key thing is not to have scripts depend on a path. When debugfs is enabled in a kernel, you automatically get /sys/kernel/debug, so that is a natural place to mount it.

I have also suggested to people (as Greg suggested earlier to me) that you can mount it there and then make a soft link to it. Which also works.

I personally just mount it at /debug.

To handle the scripts problem, all my scripts start with:

# grep debugfs /proc/mounts | cut -d ' ' -f 2

I also have a C function to do a search for debugfs in /proc/mounts.

If you look in Documentation/trace/ftrace.txt I have:

===
To mount the debugfs system:

 # mkdir /debug
 # mount -t debugfs nodev /debug
Quoted 8 lines

( Note: it is more common to mount at /sys/kernel/debug, but for simplicity this document will use /debug)
===

Thus it is a personal preference of mine to put it at /debug. But I do not hide the fact that it is recommended to be in /sys/kernel/debug. Heck, I'm the one that brought to light that path in this thread. I just curse it everytime I have to say it ;-)

Thank you for explanation about history of "/debug/" and "/sys/kernel/debug/" in the ./Documentations/trace/ftrace.txt file.

Do I have to change contents of patch file from "/debug/***" to "debug/***" to consider both "/debug/"directory and
"/sys/kernel/debug/ (initial directory by debugfs)" directory in the my patch file?

Or,
Do I have to keep "/" character of "/debug" in my patch file that I proposed because many contributors already use "/debugfs/" or "/debug/" in their patch files?

Steven, Can you advise me?

Quoted 3 lines

-- Steve

Signature · 10 lines
Regards,
GeunSik Lim ( SAMSUNG ELECTRONICS)
Blog : http://blog.naver.com/invain/
e-Mail: [email protected]
           [email protected] , [email protected]
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/
#9 lore raw
In reply to Steven Rostedt

Hello Steve,

Quoted 5 lines · Steven Rostedt

If you look in Documentation/trace/ftrace.txt I have:

===
To mount the debugfs system:

Quoted 7 lines
  # mkdir /debug
  # mount -t debugfs nodev /debug

( Note: it is more common to mount at /sys/kernel/debug, but for
simplicity this document will use /debug)
===

Please standardize Documentation/trace/ftrace.txt and avoid confusion to users:

===
To mount the debugfs system:

  # mount -t debugfs nodev /sys/kernel/debug
===

Thanks,
--
JSR

maintainer
#10 lore raw
Quoted 17 lines · Jaswinder Singh Rajput

Hello Steve,

Quoted 5 lines · Steven Rostedt

If you look in Documentation/trace/ftrace.txt I have:

===
To mount the debugfs system:

Quoted 7 lines
  # mkdir /debug
  # mount -t debugfs nodev /debug

( Note: it is more common to mount at /sys/kernel/debug, but for
simplicity this document will use /debug)
===

Quoted 8 lines

Please standardize Documentation/trace/ftrace.txt and avoid confusion to users:

===
To mount the debugfs system:

  # mount -t debugfs nodev /sys/kernel/debug
===

I really don't want to do that. Then I would need to show

 # echo function_graph > /sys/kernel/debug/current_trace

Which I really do not like. I state that the common place to mount it is in /sys/kernel/debug, But my examples are much easier to understand if I just do: # echo function_graph > /debug/current_trace

-- Steve

#11 lore raw
Quoted 29 lines · Steven Rostedt
Quoted 17 lines · Jaswinder Singh Rajput

Hello Steve,

Quoted 5 lines · Steven Rostedt

If you look in Documentation/trace/ftrace.txt I have:

===
To mount the debugfs system:

Quoted 7 lines
  # mkdir /debug
  # mount -t debugfs nodev /debug

( Note: it is more common to mount at /sys/kernel/debug, but for
simplicity this document will use /debug)
===

Quoted 8 lines

Please standardize Documentation/trace/ftrace.txt and avoid confusion to users:

===
To mount the debugfs system:

  # mount -t debugfs nodev /sys/kernel/debug
===
Quoted 7 lines

I really don't want to do that. Then I would need to show

 # echo function_graph > /sys/kernel/debug/current_trace

Which I really do not like. I state that the common place to mount it is in /sys/kernel/debug, But my examples are much easier to understand if I just do: # echo function_graph > /debug/current_trace

It's not just the "common" place, it is the DOCUMENTED PLACE.

I'm sorry you have to type a long path, perhaps you wish to invest in a tab key for auto-completion :)

Again, distros already mount debugfs at this location, please don't confuse people just because you are annoyed at the length of the path.

thanks,

greg k-h

maintainer
#12 lore raw
Quoted 8 lines · Greg KH
Quoted 8 lines
Quoted 4 lines
Quoted 4 lines

( Note: it is more common to mount at /sys/kernel/debug, but for
simplicity this document will use /debug)
===

I really don't want to do that. Then I would need to show

Quoted 7 lines
Quoted 5 lines
 # echo function_graph > /sys/kernel/debug/current_trace

Which I really do not like. I state that the common place to mount it is in /sys/kernel/debug, But my examples are much easier to understand if I just do: # echo function_graph > /debug/current_trace

It's not just the "common" place, it is the DOCUMENTED PLACE.

I can change the wording to be "documented" or "distro mounted".

Quoted 6 lines

I'm sorry you have to type a long path, perhaps you wish to invest in a tab key for auto-completion :)

Again, distros already mount debugfs at this location, please don't confuse people just because you are annoyed at the length of the path.

Have you found which packages do that?

The point is that the document is more interested in showing the files that are being used, and if I add /sys/kernel/debug/ in front of every file, that just distracts from the eye. As stated:

"but for simplicity this document will use /debug"

I'll update ftrace.txt to explain it better.

Something like:

=====
Debugfs is mounted by some packages at /sys/kernel/debugfs. If it is not you can mount it by adding

debugfs /sys/kernel/debug debugfs defaults 0 0

to your /etc/fstab, or by manually mounting it:

 # mount -t debugfs nodev /sys/kernel/debugfs

For simplicity, this document will assume that you also added the following symbolic link:

 # ln -s /sys/kernel/debugfs /debug
=====

Then I can keep the document simplistic and still have the "DOCUMENTED" locating specified.

-- Steve

#13 lore raw
Quoted 7 lines · Steven Rostedt
Quoted 5 lines

I'm sorry you have to type a long path, perhaps you wish to invest in a tab key for auto-completion :)

Again, distros already mount debugfs at this location, please don't confuse people just because you are annoyed at the length of the path.

Have you found which packages do that?

The base system packages on all openSUSE and SLE11 releases. Possibly SLE10 also does it, but I don't have a running instance of that at the moment.

Quoted 5 lines

The point is that the document is more interested in showing the files that are being used, and if I add /sys/kernel/debug/ in front of every file, that just distracts from the eye. As stated:

"but for simplicity this document will use /debug"

Sorry for the "distraction", but that goes against documentation on your machine:

	$ man mount | grep -A 3 "Mount options for debugfs"
	Mount options for debugfs
	       The debugfs file system is a pseudo file system, traditionally  mounted
	       on /sys/kernel/debug.  There are no mount options.

It's not only a documented standard, it's a "tradition" :)

Please fix your documentation.

thanks,

greg k-h

#14 lore raw
In reply to Steven Rostedt
Quoted 29 lines · Steven Rostedt
Quoted 8 lines · Greg KH
Quoted 8 lines
Quoted 4 lines
Quoted 4 lines

( Note: it is more common to mount at /sys/kernel/debug, but for simplicity this document will use /debug)
===

I really don't want to do that. Then I would need to show

Quoted 7 lines
Quoted 5 lines
 # echo function_graph > /sys/kernel/debug/current_trace

Which I really do not like. I state that the common place to mount it is in /sys/kernel/debug, But my examples are much easier to understand if I just do: # echo function_graph > /debug/current_trace

It's not just the "common" place, it is the DOCUMENTED PLACE.

I can change the wording to be "documented" or "distro mounted".

Quoted 6 lines

I'm sorry you have to type a long path, perhaps you wish to invest in a tab key for auto-completion :)

Again, distros already mount debugfs at this location, please don't confuse people just because you are annoyed at the length of the path.

Have you found which packages do that?

Quoted 7 lines

The point is that the document is more interested in showing the files that are being used, and if I add /sys/kernel/debug/ in front of every file, that just distracts from the eye. As stated:

 "but for simplicity this document will use /debug"

I'll update ftrace.txt to explain it better.

Thank you for the decision to update debufs filesystem realted contents.

Quoted 11 lines

Something like:

=====
Debugfs is mounted by some packages at /sys/kernel/debugfs. If it is not you can mount it by adding

 debugfs   /sys/kernel/debug  debugfs   defaults   0   0

to your /etc/fstab, or by manually mounting it:

Quoted 7 lines
 # mount -t debugfs nodev /sys/kernel/debugfs

For simplicity, this document will assume that you also added the following symbolic link:

 # ln -s /sys/kernel/debugfs /debug
=====

About typo in the above statements.
/sys/kernel/debugfs --> /sys/kenrel/debug
Default directory of debugfs filesystem is /sys/kernel/debug/ as you know.

Quoted 7 lines

Then I can keep the document simplistic and still have the "DOCUMENTED" locating specified.

-- Steve

Signature · 10 lines
Regards,
GeunSik Lim ( SAMSUNG ELECTRONICS)
Blog : http://blog.naver.com/invain/
e-Mail: [email protected]
           [email protected] , [email protected]
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/
#15 lore raw

RFC PATCH v2 1/1 debugfs:Fix terminology inconsistency of dir name to mount debugfs filesystem.

In reply to GeunSik Lim

Dear Greg and Steen,

Thanks your advices and opinions about a confusion according to
too many mount directory names for debugfs. I made rfc v2 patch file again like belows. Can I hear your advices about below patch contents?.

* ref:
http://groups.google.com/group/linux.kernel/browse_thread/thread/c34e386587979340/dc52317f888e78b6?#dc52317f888e78b6

Regards,
GeunSik Lim.

Subject: [PATCH] Fix terminology inconsistency of dir name to mount debugfs filesystem.

Many developers use "/debug/" or "/debugfs/ directory name to mount debugfs filesystem for ftrace according to ./Documentation/tracers/ftrace.txt file.

And, three directory names(ex:/debug/, /debugfs/, /sys/kernel/debug/) is existed in kernel source like ftrace, DRM, Wireless, Network[sky2]files to mount debugfs filesystem.

debugfs means debug filesystem for debugging easy to use by greg kroah hartman. "/sys/kernel/debug/" name is suitable as directory name of debugfs filesystem.
-debugfs related reference: http://lwn.net/Articles/115405/

Fix inconsistency of directory name to mount debugfs filesystem.

---
 Documentation/DocBook/debugobjects.tmpl           |    2 +-
 Documentation/cdrom/packet-writing.txt            |    2 +-
 Documentation/fault-injection/fault-injection.txt |   70 +++++-----
 Documentation/kprobes.txt                         |    6 +-
 Documentation/trace/ftrace.txt                    |  166 ++++++++++----------
 Documentation/trace/mmiotrace.txt                 |   26 ++--
 drivers/block/pktcdvd.c                           |    2 +-
 drivers/gpu/drm/drm_debugfs.c                     |   12 +-
 drivers/gpu/drm/drm_drv.c                         |    2 +-
 drivers/gpu/drm/drm_stub.c                        |    2 +-
 drivers/net/Kconfig                               |    4 +-
 drivers/net/wimax/i2400m/i2400m.h                 |    2 +-
 drivers/net/wireless/ath5k/Kconfig                |    6 +-
 drivers/net/wireless/libertas/README              |   12 +-
 drivers/scsi/lpfc/lpfc_debugfs.c                  |    4 +-
 include/linux/kernel.h                            |    2 +-
 include/linux/tracepoint.h                        |    4 +-
 kernel/trace/Kconfig                              |   10 +-
 kernel/trace/trace.c                              |   24 ++--
 scripts/tracing/draw_functrace.py                 |    8 +-
 20 files changed, 183 insertions(+), 183 deletions(-)
Documentation/DocBook/debugobjects.tmpl +1 -1
index 7f5f218..08ff908 100644
@@ -106,7 +106,7 @@
       number of errors are printk'ed including a full stack trace.
     </para>
     <para>
-      The statistics are available via debugfs/debug_objects/stats.
+      The statistics are available via /sys/kernel/debug/debug_objects/stats.
       They provide information about the number of warnings and the
       number of successful fixups along with information about the
       usage of the internal tracking objects and the state of the
Documentation/cdrom/packet-writing.txt +1 -1
index cf1f812..1c40777 100644
@@ -117,7 +117,7 @@ Using the pktcdvd debugfs interface
 
 To read pktcdvd device infos in human readable form, do:
 
-	# cat /debug/pktcdvd/pktcdvd[0-7]/info
+	# cat /sys/kernel/debug/pktcdvd/pktcdvd[0-7]/info
 
 For a description of the debugfs interface look into the file:
 
Documentation/fault-injection/fault-injection.txt +35 -35

145 lines not on this page · raw message

Documentation/kprobes.txt +3 -3
index 1e7a769..053037a 100644
@@ -507,9 +507,9 @@ http://www.linuxsymposium.org/2006/linuxsymposium_procv2.pdf (pages 101-115)
 Appendix A: The kprobes debugfs interface
 
 With recent kernels (> 2.6.20) the list of registered kprobes is visible
-under the /debug/kprobes/ directory (assuming debugfs is mounted at /debug).
+under the /sys/kernel/debug/kprobes/ directory (assuming debugfs is mounted at //sys/kernel/debug).
 
-/debug/kprobes/list: Lists all registered probes on the system
+/sys/kernel/debug/kprobes/list: Lists all registered probes on the system
 
 c015d71a  k  vfs_read+0x0
 c011a316  j  do_fork+0x0
@@ -525,7 +525,7 @@ virtual addresses that correspond to modules that've been unloaded),
 such probes are marked with [GONE]. If the probe is temporarily disabled,
 such probes are marked with [DISABLED].
 
-/debug/kprobes/enabled: Turn kprobes ON/OFF forcibly.
+/sys/kernel/debug/kprobes/enabled: Turn kprobes ON/OFF forcibly.
 
 Provides a knob to globally and forcibly turn registered kprobes ON or OFF.
 By default, all kprobes are enabled. By echoing "0" to this file, all
Documentation/trace/ftrace.txt +83 -83

375 lines not on this page · raw message

Documentation/trace/mmiotrace.txt +13 -13
index 5731c67..162effb 100644
@@ -32,41 +32,41 @@ is no way to automatically detect if you are losing events due to CPUs racing.
 Usage Quick Reference
 ---------------------
 
-$ mount -t debugfs debugfs /debug
-$ echo mmiotrace > /debug/tracing/current_tracer
-$ cat /debug/tracing/trace_pipe > mydump.txt &
+$ mount -t debugfs debugfs /sys/kernel/debug
+$ echo mmiotrace > /sys/kernel/debug/tracing/current_tracer
+$ cat /sys/kernel/debug/tracing/trace_pipe > mydump.txt &
 Start X or whatever.
-$ echo "X is up" > /debug/tracing/trace_marker
-$ echo nop > /debug/tracing/current_tracer
+$ echo "X is up" > /sys/kernel/debug/tracing/trace_marker
+$ echo nop > /sys/kernel/debug/tracing/current_tracer
 Check for lost events.
 
 
 Usage
 -----
 
-Make sure debugfs is mounted to /debug. If not, (requires root privileges)
-$ mount -t debugfs debugfs /debug
+Make sure debugfs is mounted to /sys/kernel/debug. If not, (requires root privileges)
+$ mount -t debugfs debugfs /sys/kernel/debug
 
 Check that the driver you are about to trace is not loaded.
 
 Activate mmiotrace (requires root privileges):
-$ echo mmiotrace > /debug/tracing/current_tracer
+$ echo mmiotrace > /sys/kernel/debug/tracing/current_tracer
 
 Start storing the trace:
-$ cat /debug/tracing/trace_pipe > mydump.txt &
+$ cat /sys/kernel/debug/tracing/trace_pipe > mydump.txt &
 The 'cat' process should stay running (sleeping) in the background.
 
 Load the driver you want to trace and use it. Mmiotrace will only catch MMIO
 accesses to areas that are ioremapped while mmiotrace is active.
 
 During tracing you can place comments (markers) into the trace by
-$ echo "X is up" > /debug/tracing/trace_marker
+$ echo "X is up" > /sys/kernel/debug/tracing/trace_marker
 This makes it easier to see which part of the (huge) trace corresponds to
 which action. It is recommended to place descriptive markers about what you
 do.
 
 Shut down mmiotrace (requires root privileges):
-$ echo nop > /debug/tracing/current_tracer
+$ echo nop > /sys/kernel/debug/tracing/current_tracer
 The 'cat' process exits. If it does not, kill it by issuing 'fg' command and
 pressing ctrl+c.
 
@@ -78,10 +78,10 @@ to view your kernel log and look for "mmiotrace has lost events" warning. If
 events were lost, the trace is incomplete. You should enlarge the buffers and
 try again. Buffers are enlarged by first seeing how large the current buffers
 are:
-$ cat /debug/tracing/buffer_size_kb
+$ cat /sys/kernel/debug/tracing/buffer_size_kb
 gives you a number. Approximately double this number and write it back, for
 instance:
-$ echo 128000 > /debug/tracing/buffer_size_kb
+$ echo 128000 > /sys/kernel/debug/tracing/buffer_size_kb
 Then start again from the top.
 
 If you are doing a trace for a driver project, e.g. Nouveau, you should also
drivers/block/pktcdvd.c +1 -1
index dc7a8c3..9aa8932 100644
@@ -430,7 +430,7 @@ static void pkt_sysfs_cleanup(void)
 /********************************************************************
   entries in debugfs
 
-  /debugfs/pktcdvd[0-7]/
+  /sys/kernel/debug/pktcdvd[0-7]/
 			info
 
  *******************************************************************/
drivers/gpu/drm/drm_debugfs.c +6 -6
index c77c6c6..6ce0e26 100644
@@ -105,7 +105,7 @@ int drm_debugfs_create_files(struct drm_info_list *files, int count,
 		ent = debugfs_create_file(files[i].name, S_IFREG | S_IRUGO,
 					  root, tmp, &drm_debugfs_fops);
 		if (!ent) {
-			DRM_ERROR("Cannot create /debugfs/dri/%s/%s\n",
+			DRM_ERROR("Cannot create /sys/kernel/debug/dri/%s/%s\n",
 				  name, files[i].name);
 			drm_free(tmp, sizeof(struct drm_info_node),
 				 _DRM_DRIVER);
@@ -133,9 +133,9 @@ EXPORT_SYMBOL(drm_debugfs_create_files);
  * \param minor device minor number
  * \param root DRI debugfs dir entry.
  *
- * Create the DRI debugfs root entry "/debugfs/dri", the device debugfs root entry
- * "/debugfs/dri/%minor%/", and each entry in debugfs_list as
- * "/debugfs/dri/%minor%/%name%".
+ * Create the DRI debugfs root entry "/sys/kernel/debug/dri", the device debugfs root entry
+ * "/sys/kernel/debug/dri/%minor%/", and each entry in debugfs_list as
+ * "/sys/kernel/debug/dri/%minor%/%name%".
  */
 int drm_debugfs_init(struct drm_minor *minor, int minor_id,
 		     struct dentry *root)
@@ -148,7 +148,7 @@ int drm_debugfs_init(struct drm_minor *minor, int minor_id,
 	sprintf(name, "%d", minor_id);
 	minor->debugfs_root = debugfs_create_dir(name, root);
 	if (!minor->debugfs_root) {
-		DRM_ERROR("Cannot create /debugfs/dri/%s\n", name);
+		DRM_ERROR("Cannot create /sys/kernel/debug/dri/%s\n", name);
 		return -1;
 	}
 
@@ -165,7 +165,7 @@ int drm_debugfs_init(struct drm_minor *minor, int minor_id,
 		ret = dev->driver->debugfs_init(minor);
 		if (ret) {
 			DRM_ERROR("DRM: Driver failed to initialize "
-				  "/debugfs/dri.\n");
+				  "/sys/kernel/debug/dri.\n");
 			return ret;
 		}
 	}
drivers/gpu/drm/drm_drv.c +1 -1
index 019b7c5..1bf7efd 100644
@@ -339,7 +339,7 @@ static int __init drm_core_init(void)
 
 	drm_debugfs_root = debugfs_create_dir("dri", NULL);
 	if (!drm_debugfs_root) {
-		DRM_ERROR("Cannot create /debugfs/dri\n");
+		DRM_ERROR("Cannot create /sys/kernel/debug/dri\n");
 		ret = -1;
 		goto err_p3;
 	}
drivers/gpu/drm/drm_stub.c +1 -1
index b9631e3..b2a93c4 100644
@@ -328,7 +328,7 @@ static int drm_get_minor(struct drm_device *dev, struct drm_minor **minor, int t
 #if defined(CONFIG_DEBUG_FS)
 	ret = drm_debugfs_init(new_minor, minor_id, drm_debugfs_root);
 	if (ret) {
-		DRM_ERROR("DRM: Failed to initialize /debugfs/dri.\n");
+		DRM_ERROR("DRM: Failed to initialize /sys/kernel/debug/dri.\n");
 		goto err_g2;
 	}
 #endif
drivers/net/Kconfig +2 -2
index 214a92d..a170cea 100644
@@ -2200,7 +2200,7 @@ config SKGE_DEBUG
        depends on SKGE && DEBUG_FS
        help
 	 This option adds the ability to dump driver state for debugging.
-	 The file debugfs/skge/ethX displays the state of the internal
+	 The file /sys/kernel/debug/skge/ethX displays the state of the internal
 	 transmit and receive rings.
 
 	 If unsure, say N.
@@ -2226,7 +2226,7 @@ config SKY2_DEBUG
        depends on SKY2 && DEBUG_FS
        help
 	 This option adds the ability to dump driver state for debugging.
-	 The file debugfs/sky2/ethX displays the state of the internal
+	 The file /sys/kernel/debug/sky2/ethX displays the state of the internal
 	 transmit and receive rings.
 
 	 If unsure, say N.
drivers/net/wimax/i2400m/i2400m.h +1 -1
index 3ae2df3..da2e8f2 100644
@@ -389,7 +389,7 @@ struct i2400m {
 	unsigned ready:1;		/* all probing steps done */
 	unsigned rx_reorder:1;		/* RX reorder is enabled */
 	u8 trace_msg_from_user;		/* echo rx msgs to 'trace' pipe */
-					/* typed u8 so debugfs/u8 can tweak */
+					/* typed u8 so /sys/kernel/debug/u8 can tweak */
 	enum i2400m_system_state state;
 	wait_queue_head_t state_wq;	/* Woken up when on state updates */
 
drivers/net/wireless/ath5k/Kconfig +3 -3
index 75383a5..fdfa708 100644
@@ -27,11 +27,11 @@ config ATH5K_DEBUG
 	  Say Y, if and you will get debug options for ath5k.
 	  To use this, you need to mount debugfs:
 
-	  mkdir /debug/
-	  mount -t debugfs debug /debug/
+	  mkdir /sys/kernel/debug/
+	  mount -t debugfs debug /sys/kernel/debug/
 
 	  You will get access to files under:
-	  /debug/ath5k/phy0/
+	  /sys/kernel/debug/ath5k/phy0/
 
 	  To enable debug, pass the debug level to the debug module
 	  parameter. For example:
drivers/net/wireless/libertas/README +6 -6
index d860fc3..ab6a2d5 100644
@@ -72,7 +72,7 @@ rdrf
 	location that is to be read.  This parameter must be specified in
 	hexadecimal (its possible to preceed preceding the number with a "0x").
 
-	Path: /debugfs/libertas_wireless/ethX/registers/
+	Path: /sys/kernel/debug/libertas_wireless/ethX/registers/
 
 	Usage:
 		echo "0xa123" > rdmac ; cat rdmac
@@ -95,7 +95,7 @@ wrrf
 sleepparams
 	This command is used to set the sleepclock configurations
 
-	Path: /debugfs/libertas_wireless/ethX/
+	Path: /sys/kernel/debug/libertas_wireless/ethX/
 
 	Usage:
 		cat sleepparams: reads the current sleepclock configuration
@@ -115,7 +115,7 @@ subscribed_events
 	The subscribed_events directory contains the interface for the
 	subscribed events API.
 
-	Path: /debugfs/libertas_wireless/ethX/subscribed_events/
+	Path: /sys/kernel/debug/libertas_wireless/ethX/subscribed_events/
 
 	Each event is represented by a filename. Each filename consists of the
 	following three fields:
@@ -165,7 +165,7 @@ subscribed_events
 extscan
 	This command is used to do a specific scan.
 
-	Path: /debugfs/libertas_wireless/ethX/
+	Path: /sys/kernel/debug/libertas_wireless/ethX/
 
 	Usage: echo "SSID" > extscan
 
@@ -179,7 +179,7 @@ getscantable
 	Display the current contents of the driver scan table (ie. get the
 	scan results).
 
-	Path: /debugfs/libertas_wireless/ethX/
+	Path: /sys/kernel/debug/libertas_wireless/ethX/
 
 	Usage:
 		cat getscantable
@@ -188,7 +188,7 @@ setuserscan
 	Initiate a customized scan and retrieve the results
 
 
-	Path: /debugfs/libertas_wireless/ethX/
+	Path: /sys/kernel/debug/libertas_wireless/ethX/
 
     Usage:
        echo "[ARGS]" > setuserscan
drivers/scsi/lpfc/lpfc_debugfs.c +2 -2
index 52be564..9112380 100644
@@ -51,8 +51,8 @@
  * debugfs interface
  *
  * To access this interface the user should:
- * # mkdir /debug
- * # mount -t debugfs none /debug
+ * # mkdir /sys/kernel/debug
+ * # mount -t debugfs none /sys/kernel/debug
  *
  * The lpfc debugfs directory hierarchy is:
  * lpfc/lpfcX/vportY
include/linux/kernel.h +1 -1
index 883cd44..99b7aad 100644
@@ -406,7 +406,7 @@ static inline char *pack_hex_byte(char *buf, u8 byte)
  *
  * Use tracing_on/tracing_off when you want to quickly turn on or off
  * tracing. It simply enables or disables the recording of the trace events.
- * This also corresponds to the user space debugfs/tracing/tracing_on
+ * This also corresponds to the user space /sys/kernel/debug/tracing/tracing_on
  * file, which gives a means for the kernel and userspace to interact.
  * Place a tracing_off() in the kernel where you want tracing to end.
  * From user space, examine the trace, and then echo 1 > tracing_on
include/linux/tracepoint.h +2 -2
index d35a7ee..9c9ac7d 100644
@@ -197,7 +197,7 @@ static inline void tracepoint_synchronize_unregister(void)
  *	* This is how the trace record is structured and will
  *	* be saved into the ring buffer. These are the fields
  *	* that will be exposed to user-space in
- *	* /debug/tracing/events/<*>/format.
+ *	* /sys/kernel/debug/tracing/events/<*>/format.
  *	*
  *	* The declared 'local variable' is called '__entry'
  *	*
@@ -257,7 +257,7 @@ static inline void tracepoint_synchronize_unregister(void)
  * tracepoint callback (this is used by programmatic plugins and
  * can also by used by generic instrumentation like SystemTap), and
  * it is also used to expose a structured trace record in
- * /debug/tracing/events/.
+ * /sys/kernel/debug/tracing/events/.
  */
 
 #define TRACE_EVENT(name, proto, args, struct, assign, print)	\
kernel/trace/Kconfig +5 -5
index 417d198..e5ce28f 100644
@@ -120,7 +120,7 @@ config IRQSOFF_TRACER
 	  disabled by default and can be runtime (re-)started
 	  via:
 
-	      echo 0 > /debugfs/tracing/tracing_max_latency
+	      echo 0 > /sys/kernel/debug/tracing/tracing_max_latency
 
 	  (Note that kernel size and overhead increases with this option
 	  enabled. This option and the preempt-off timing option can be
@@ -141,7 +141,7 @@ config PREEMPT_TRACER
 	  disabled by default and can be runtime (re-)started
 	  via:
 
-	      echo 0 > /debugfs/tracing/tracing_max_latency
+	      echo 0 > /sys/kernel/debug/tracing/tracing_max_latency
 
 	  (Note that kernel size and overhead increases with this option
 	  enabled. This option and the irqs-off timing option can be
@@ -213,7 +213,7 @@ config TRACE_BRANCH_PROFILING
 	  This tracer profiles all the the likely and unlikely macros
 	  in the kernel. It will display the results in:
 
-	  /debugfs/tracing/profile_annotated_branch
+	  /sys/kernel/debug/tracing/profile_annotated_branch
 
 	  Note: this will add a significant overhead, only turn this
 	  on if you need to profile the system's use of these macros.
@@ -228,7 +228,7 @@ config PROFILE_ALL_BRANCHES
 	  taken in the kernel is recorded whether it hit or miss.
 	  The results will be displayed in:
 
-	  /debugfs/tracing/profile_branch
+	  /sys/kernel/debug/tracing/profile_branch
 
 	  This configuration, when enabled, will impose a great overhead
 	  on the system. This should only be enabled when the system
@@ -276,7 +276,7 @@ config STACK_TRACER
 	select KALLSYMS
 	help
 	  This special tracer records the maximum stack footprint of the
-	  kernel and displays it in debugfs/tracing/stack_trace.
+	  kernel and displays it in /sys/kernel/debug/tracing/stack_trace.
 
 	  This tracer works by hooking into every function call that the
 	  kernel executes, and keeping a maximum stack depth value and
kernel/trace/trace.c +12 -12
index cda81ec..9b94d57 100644
@@ -335,7 +335,7 @@ static raw_spinlock_t ftrace_max_lock =
 /*
  * Copy the new maximum trace into the separate maximum-trace
  * structure. (this way the maximum trace is permanently saved,
- * for later retrieval via /debugfs/tracing/latency_trace)
+ * for later retrieval via /sys/kernel/debug/tracing/latency_trace)
  */
 static void
 __update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu)
@@ -2366,21 +2366,21 @@ static const struct file_operations tracing_iter_fops = {
 
 static const char readme_msg[] =
 	"tracing mini-HOWTO:\n\n"
-	"# mkdir /debug\n"
-	"# mount -t debugfs nodev /debug\n\n"
-	"# cat /debug/tracing/available_tracers\n"
+	"# mkdir /sys/kernel/debug\n"
+	"# mount -t debugfs nodev /sys/kernel/debug\n\n"
+	"# cat /sys/kernel/debug/tracing/available_tracers\n"
 	"wakeup preemptirqsoff preemptoff irqsoff function sched_switch nop\n\n"
-	"# cat /debug/tracing/current_tracer\n"
+	"# cat /sys/kernel/debug/tracing/current_tracer\n"
 	"nop\n"
-	"# echo sched_switch > /debug/tracing/current_tracer\n"
-	"# cat /debug/tracing/current_tracer\n"
+	"# echo sched_switch > /sys/kernel/debug/tracing/current_tracer\n"
+	"# cat /sys/kernel/debug/tracing/current_tracer\n"
 	"sched_switch\n"
-	"# cat /debug/tracing/trace_options\n"
+	"# cat /sys/kernel/debug/tracing/trace_options\n"
 	"noprint-parent nosym-offset nosym-addr noverbose\n"
-	"# echo print-parent > /debug/tracing/trace_options\n"
-	"# echo 1 > /debug/tracing/tracing_enabled\n"
-	"# cat /debug/tracing/trace > /tmp/trace.txt\n"
-	"# echo 0 > /debug/tracing/tracing_enabled\n"
+	"# echo print-parent > /sys/kernel/debug/tracing/trace_options\n"
+	"# echo 1 > /sys/kernel/debug/tracing/tracing_enabled\n"
+	"# cat /sys/kernel/debug/tracing/trace > /tmp/trace.txt\n"
+	"# echo 0 > /sys/kernel/debug/tracing/tracing_enabled\n"
 ;
 
 static ssize_t
scripts/tracing/draw_functrace.py +4 -4
index 902f9a9..66f4d1b 100644
@@ -12,10 +12,10 @@ calls. Only the functions's names and the the call time are provided.
 
 Usage:
 	Be sure that you have CONFIG_FUNCTION_TRACER
-	# mkdir /debugfs
-	# mount -t debug debug /debug
-	# echo function > /debug/tracing/current_tracer
-	$ cat /debug/tracing/trace_pipe > ~/raw_trace_func
+	# mkdir /sys/kernel/debug
+	# mount -t debug debug /sys/kernel/debug
+	# echo function > /sys/kernel/debug/tracing/current_tracer
+	$ cat /sys/kernel/debug/tracing/trace_pipe > ~/raw_trace_func
 	Wait some times but not too much, the script is a bit slow.
 	Break the pipe (Ctrl + Z)
 	$ scripts/draw_functrace.py < raw_trace_func > draw_functrace
Signature · 14 lines
1.6.3.1

 
-----------------------------------------------
To unsubscribe from this list: send the line "unsubscribe linux-***" 
in the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

GeunSik Lim (  S A M S U N G     E L E C T R O N I C S  )
e-Mail  :1) [email protected]
         2) [email protected] , [email protected]
HomePage: http://blog.naver.com/invain/
-----------------------------------------------
maintainer
#16 lore raw

Documentation/trace/ftrace.txt well as the files to display output. · line 38

Quoted 11 lines · GeunSik Lim
Documentation/trace/ftrace.txt +1 -2
index fd9a3e6..cbc6cca 100644
@@ -35,11 +35,11 @@ well as the files to display output.
 
 To mount the debugfs system:
 
-  # mkdir /debug
-  # mount -t debugfs nodev /debug
+  # mkdir /sys/kernel/debug

If you configure debugfs into the kernel, the directory /sys/kernel/debug will automatically be created. Remember, /sys is a psuedo filesystem as well.

Documentation/trace/ftrace.txt well as the files to display output. · line 42

Quoted 5 lines
Documentation/trace/ftrace.txt +2 -1
+  # mount -t debugfs nodev /sys/kernel/debug
 
 ( Note: it is more common to mount at /sys/kernel/debug, but for
-  simplicity this document will use /debug)
+  simplicity this document will use /sys/kernel/debug)

LOL!

Quoted 99 lines
Documentation/trace/ftrace.txt +22 -22
 That's it! (assuming that you have ftrace configured into your kernel)
 
@@ -389,18 +389,18 @@ trace_options
 The trace_options file is used to control what gets printed in
 the trace output. To see what is available, simply cat the file:
 
-  cat /debug/tracing/trace_options
+  cat /sys/kernel/debug/tracing/trace_options
   print-parent nosym-offset nosym-addr noverbose noraw nohex nobin \
   noblock nostacktrace nosched-tree nouserstacktrace nosym-userobj
 
 To disable one of the options, echo in the option prepended with
 "no".
 
-  echo noprint-parent > /debug/tracing/trace_options
+  echo noprint-parent > /sys/kernel/debug/tracing/trace_options
 
 To enable an option, leave off the "no".
 
-  echo sym-offset > /debug/tracing/trace_options
+  echo sym-offset > /sys/kernel/debug/tracing/trace_options
 
 Here are the available options:
 
@@ -476,11 +476,11 @@ sched_switch
 This tracer simply records schedule switches. Here is an example
 of how to use it.
 
- # echo sched_switch > /debug/tracing/current_tracer
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo sched_switch > /sys/kernel/debug/tracing/current_tracer
+ # echo 1 > /sys/kernel/debug/tracing/tracing_enabled
  # sleep 1
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/trace
+ # echo 0 > /sys/kernel/debug/tracing/tracing_enabled
+ # cat /sys/kernel/debug/tracing/trace
 
 # tracer: sched_switch
 #
@@ -574,13 +574,13 @@ new trace is saved.
 To reset the maximum, echo 0 into tracing_max_latency. Here is
 an example:
 
- # echo irqsoff > /debug/tracing/current_tracer
- # echo 0 > /debug/tracing/tracing_max_latency
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo irqsoff > /sys/kernel/debug/tracing/current_tracer
+ # echo 0 > /sys/kernel/debug/tracing/tracing_max_latency
+ # echo 1 > /sys/kernel/debug/tracing/tracing_enabled
  # ls -ltr
[...]
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/latency_trace
+ # echo 0 > /sys/kernel/debug/tracing/tracing_enabled
+ # cat /sys/kernel/debug/tracing/latency_trace
 # tracer: irqsoff
 #
 irqsoff latency trace v1.1.5 on 2.6.26
@@ -681,13 +681,13 @@ Like the irqsoff tracer, it records the maximum latency for
 which preemption was disabled. The control of preemptoff tracer
 is much like the irqsoff tracer.
 
- # echo preemptoff > /debug/tracing/current_tracer
- # echo 0 > /debug/tracing/tracing_max_latency
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo preemptoff > /sys/kernel/debug/tracing/current_tracer
+ # echo 0 > /sys/kernel/debug/tracing/tracing_max_latency
+ # echo 1 > /sys/kernel/debug/tracing/tracing_enabled
  # ls -ltr
[...]
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/latency_trace
+ # echo 0 > /sys/kernel/debug/tracing/tracing_enabled
+ # cat /sys/kernel/debug/tracing/latency_trace
 # tracer: preemptoff
 #
 preemptoff latency trace v1.1.5 on 2.6.26-rc8
@@ -828,13 +828,13 @@ tracer.
 Again, using this trace is much like the irqsoff and preemptoff
 tracers.
 
- # echo preemptirqsoff > /debug/tracing/current_tracer
- # echo 0 > /debug/tracing/tracing_max_latency
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo preemptirqsoff > /sys/kernel/debug/tracing/current_tracer
+ # echo 0 > /sys/kernel/debug/tracing/tracing_max_latency
+ # echo 1 > /sys/kernel/debug/tracing/tracing_enabled
  # ls -ltr
[...]
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/latency_trace
+ # echo 0 > /sys/kernel/debug/tracing/tracing_enabled
+ # cat /sys/kernel/debug/tracing/latency_trace
 # tracer: preemptirqsoff
 #
 preemptirqsoff latency trace v1.1.5 on 2.6.26-rc8

Documentation/trace/ftrace.txt slightly differently than we did with the previous tracers. · line 995

Quoted 9 lines
Documentation/trace/ftrace.txt +3 -3
@@ -990,12 +990,12 @@ slightly differently than we did with the previous tracers.
 Instead of performing an 'ls', we will run 'sleep 1' under
 'chrt' which changes the priority of the task.
 
- # echo wakeup > /debug/tracing/current_tracer
- # echo 0 > /debug/tracing/tracing_max_latency
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo wakeup > /sys/kernel/debug/tracing/current_tracer
+ # echo 0 > /sys/kernel/debug/tracing/tracing_max_latency
+ # echo 1 > /sys/kernel/debug/tracing/tracing_enabled

Egad! This is exactly why I hate the path name. It distracts so much from the eyes to what you want to accomplish. It seems that because it is a tradition to use this /sys/kernel/debug path, that I'm destine to fail in my discuss for it. I will ask that this document suggest the following at the beginning.

===
When debugfs is configured into the kernel (which selecting any ftrace option will do) the directory /sys/kernel/debug will be created. To mount this directory, you can add to your /etc/fstab file:

debugfs /debug debugfs defaults 0 0

Or you can mount it at run time with:

mount -t debugfs nodev /sys/kernel/debug

For quicker access to that directory you may want to make a soft link to it:

ln -s /sys/kernel/debug /debug

Any selected ftrace option will also create a directory called tracing within the debugfs. The rest of the document will assume that you are in the ftrace directory (cd /sys/kernel/debugfs) and will only concentrate on the files within that directory and not distract from the content with the extended "/sys/kernel/debug/tracing" path name.
===

And then we can change all your other changes in this file as:

s,/sys/kernel/debug/tracing,,

-- Steve

Quoted 727 lines
Documentation/trace/ftrace.txt +55 -55
  # chrt -f 5 sleep 1
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/latency_trace
+ # echo 0 > /sys/kernel/debug/tracing/tracing_enabled
+ # cat /sys/kernel/debug/tracing/latency_trace
 # tracer: wakeup
 #
 wakeup latency trace v1.1.5 on 2.6.26-rc8
@@ -1105,11 +1105,11 @@ can be done from the debug file system. Make sure the
 ftrace_enabled is set; otherwise this tracer is a nop.
 
  # sysctl kernel.ftrace_enabled=1
- # echo function > /debug/tracing/current_tracer
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo function > /sys/kernel/debug/tracing/current_tracer
+ # echo 1 > /sys/kernel/debug/tracing/tracing_enabled
  # usleep 1
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/trace
+ # echo 0 > /sys/kernel/debug/tracing/tracing_enabled
+ # cat /sys/kernel/debug/tracing/trace
 # tracer: function
 #
 #           TASK-PID   CPU#    TIMESTAMP  FUNCTION
@@ -1146,7 +1146,7 @@ int trace_fd;
[...]
 int main(int argc, char *argv[]) {
[...]
-	trace_fd = open("/debug/tracing/tracing_enabled", O_WRONLY);
+	trace_fd = open("/sys/kernel/debug/tracing/tracing_enabled", O_WRONLY);
[...]
 	if (condition_hit()) {
 		write(trace_fd, "0", 1);
@@ -1155,7 +1155,7 @@ int main(int argc, char *argv[]) {
 }
 
 Note: Here we hard coded the path name. The debugfs mount is not
-guaranteed to be at /debug (and is more commonly at
+guaranteed to be at /sys/kernel/debug (and is more commonly at
 /sys/kernel/debug). For simple one time traces, the above is
 sufficent. For anything else, a search through /proc/mounts may
 be needed to find where the debugfs file-system is mounted.
@@ -1164,16 +1164,16 @@ be needed to find where the debugfs file-system is mounted.
 Single thread tracing
 ---------------------
 
-By writing into /debug/tracing/set_ftrace_pid you can trace a
+By writing into /sys/kernel/debug/tracing/set_ftrace_pid you can trace a
 single thread. For example:
 
-# cat /debug/tracing/set_ftrace_pid
+# cat /sys/kernel/debug/tracing/set_ftrace_pid
 no pid
-# echo 3111 > /debug/tracing/set_ftrace_pid
-# cat /debug/tracing/set_ftrace_pid
+# echo 3111 > /sys/kernel/debug/tracing/set_ftrace_pid
+# cat /sys/kernel/debug/tracing/set_ftrace_pid
 3111
-# echo function > /debug/tracing/current_tracer
-# cat /debug/tracing/trace | head
+# echo function > /sys/kernel/debug/tracing/current_tracer
+# cat /sys/kernel/debug/tracing/trace | head
  # tracer: function
  #
  #           TASK-PID    CPU#    TIMESTAMP  FUNCTION
@@ -1184,8 +1184,8 @@ no pid
      yum-updatesd-3111  [003]  1637.254683: lock_hrtimer_base <-hrtimer_try_to_cancel
      yum-updatesd-3111  [003]  1637.254685: fget_light <-do_sys_poll
      yum-updatesd-3111  [003]  1637.254686: pipe_poll <-do_sys_poll
-# echo -1 > /debug/tracing/set_ftrace_pid
-# cat /debug/tracing/trace |head
+# echo -1 > /sys/kernel/debug/tracing/set_ftrace_pid
+# cat /sys/kernel/debug/tracing/trace |head
  # tracer: function
  #
  #           TASK-PID    CPU#    TIMESTAMP  FUNCTION
@@ -1217,12 +1217,12 @@ int main (int argc, char **argv)
                 char line[64];
                 int s;
 
-                ffd = open("/debug/tracing/current_tracer", O_WRONLY);
+                ffd = open("/sys/kernel/debug/tracing/current_tracer", O_WRONLY);
                 if (ffd < 0)
                         exit(-1);
                 write(ffd, "nop", 3);
 
-                fd = open("/debug/tracing/set_ftrace_pid", O_WRONLY);
+                fd = open("/sys/kernel/debug/tracing/set_ftrace_pid", O_WRONLY);
                 s = sprintf(line, "%d\n", getpid());
                 write(fd, line, s);
 
@@ -1374,22 +1374,22 @@ want, depending on your needs.
   tracing_cpu_mask file) or you might sometimes see unordered
   function calls while cpu tracing switch.
 
-	hide: echo nofuncgraph-cpu > /debug/tracing/trace_options
-	show: echo funcgraph-cpu > /debug/tracing/trace_options
+	hide: echo nofuncgraph-cpu > /sys/kernel/debug/tracing/trace_options
+	show: echo funcgraph-cpu > /sys/kernel/debug/tracing/trace_options
 
 - The duration (function's time of execution) is displayed on
   the closing bracket line of a function or on the same line
   than the current function in case of a leaf one. It is default
   enabled.
 
-	hide: echo nofuncgraph-duration > /debug/tracing/trace_options
-	show: echo funcgraph-duration > /debug/tracing/trace_options
+	hide: echo nofuncgraph-duration > /sys/kernel/debug/tracing/trace_options
+	show: echo funcgraph-duration > /sys/kernel/debug/tracing/trace_options
 
 - The overhead field precedes the duration field in case of
   reached duration thresholds.
 
-	hide: echo nofuncgraph-overhead > /debug/tracing/trace_options
-	show: echo funcgraph-overhead > /debug/tracing/trace_options
+	hide: echo nofuncgraph-overhead > /sys/kernel/debug/tracing/trace_options
+	show: echo funcgraph-overhead > /sys/kernel/debug/tracing/trace_options
 	depends on: funcgraph-duration
 
   ie:
@@ -1418,8 +1418,8 @@ want, depending on your needs.
 - The task/pid field displays the thread cmdline and pid which
   executed the function. It is default disabled.
 
-	hide: echo nofuncgraph-proc > /debug/tracing/trace_options
-	show: echo funcgraph-proc > /debug/tracing/trace_options
+	hide: echo nofuncgraph-proc > /sys/kernel/debug/tracing/trace_options
+	show: echo funcgraph-proc > /sys/kernel/debug/tracing/trace_options
 
   ie:
 
@@ -1442,8 +1442,8 @@ want, depending on your needs.
   system clock since it started. A snapshot of this time is
   given on each entry/exit of functions
 
-	hide: echo nofuncgraph-abstime > /debug/tracing/trace_options
-	show: echo funcgraph-abstime > /debug/tracing/trace_options
+	hide: echo nofuncgraph-abstime > /sys/kernel/debug/tracing/trace_options
+	show: echo funcgraph-abstime > /sys/kernel/debug/tracing/trace_options
 
   ie:
 
@@ -1540,7 +1540,7 @@ listed in:
 
    available_filter_functions
 
- # cat /debug/tracing/available_filter_functions
+ # cat /sys/kernel/debug/tracing/available_filter_functions
 put_prev_task_idle
 kmem_cache_create
 pick_next_task_rt
@@ -1552,12 +1552,12 @@ mutex_lock
 If I am only interested in sys_nanosleep and hrtimer_interrupt:
 
  # echo sys_nanosleep hrtimer_interrupt \
-		> /debug/tracing/set_ftrace_filter
- # echo ftrace > /debug/tracing/current_tracer
- # echo 1 > /debug/tracing/tracing_enabled
+		> /sys/kernel/debug/tracing/set_ftrace_filter
+ # echo ftrace > /sys/kernel/debug/tracing/current_tracer
+ # echo 1 > /sys/kernel/debug/tracing/tracing_enabled
  # usleep 1
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/trace
+ # echo 0 > /sys/kernel/debug/tracing/tracing_enabled
+ # cat /sys/kernel/debug/tracing/trace
 # tracer: ftrace
 #
 #           TASK-PID   CPU#    TIMESTAMP  FUNCTION
@@ -1568,7 +1568,7 @@ If I am only interested in sys_nanosleep and hrtimer_interrupt:
 
 To see which functions are being traced, you can cat the file:
 
- # cat /debug/tracing/set_ftrace_filter
+ # cat /sys/kernel/debug/tracing/set_ftrace_filter
 hrtimer_interrupt
 sys_nanosleep
 
@@ -1588,7 +1588,7 @@ Note: It is better to use quotes to enclose the wild cards,
       otherwise the shell may expand the parameters into names
       of files in the local directory.
 
- # echo 'hrtimer_*' > /debug/tracing/set_ftrace_filter
+ # echo 'hrtimer_*' > /sys/kernel/debug/tracing/set_ftrace_filter
 
 Produces:
 
@@ -1609,7 +1609,7 @@ Produces:
 
 Notice that we lost the sys_nanosleep.
 
- # cat /debug/tracing/set_ftrace_filter
+ # cat /sys/kernel/debug/tracing/set_ftrace_filter
 hrtimer_run_queues
 hrtimer_run_pending
 hrtimer_init
@@ -1635,17 +1635,17 @@ To append to the filters, use '>>'
 To clear out a filter so that all functions will be recorded
 again:
 
- # echo > /debug/tracing/set_ftrace_filter
- # cat /debug/tracing/set_ftrace_filter
+ # echo > /sys/kernel/debug/tracing/set_ftrace_filter
+ # cat /sys/kernel/debug/tracing/set_ftrace_filter
  #
 
 Again, now we want to append.
 
- # echo sys_nanosleep > /debug/tracing/set_ftrace_filter
- # cat /debug/tracing/set_ftrace_filter
+ # echo sys_nanosleep > /sys/kernel/debug/tracing/set_ftrace_filter
+ # cat /sys/kernel/debug/tracing/set_ftrace_filter
 sys_nanosleep
- # echo 'hrtimer_*' >> /debug/tracing/set_ftrace_filter
- # cat /debug/tracing/set_ftrace_filter
+ # echo 'hrtimer_*' >> /sys/kernel/debug/tracing/set_ftrace_filter
+ # cat /sys/kernel/debug/tracing/set_ftrace_filter
 hrtimer_run_queues
 hrtimer_run_pending
 hrtimer_init
@@ -1668,7 +1668,7 @@ hrtimer_init_sleeper
 The set_ftrace_notrace prevents those functions from being
 traced.
 
- # echo '*preempt*' '*lock*' > /debug/tracing/set_ftrace_notrace
+ # echo '*preempt*' '*lock*' > /sys/kernel/debug/tracing/set_ftrace_notrace
 
 Produces:
 
@@ -1758,13 +1758,13 @@ the effect on the tracing is different. Every read from
 trace_pipe is consumed. This means that subsequent reads will be
 different. The trace is live.
 
- # echo function > /debug/tracing/current_tracer
- # cat /debug/tracing/trace_pipe > /tmp/trace.out &
+ # echo function > /sys/kernel/debug/tracing/current_tracer
+ # cat /sys/kernel/debug/tracing/trace_pipe > /tmp/trace.out &
 [1] 4153
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo 1 > /sys/kernel/debug/tracing/tracing_enabled
  # usleep 1
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/trace
+ # echo 0 > /sys/kernel/debug/tracing/tracing_enabled
+ # cat /sys/kernel/debug/tracing/trace
 # tracer: function
 #
 #           TASK-PID   CPU#    TIMESTAMP  FUNCTION
@@ -1800,7 +1800,7 @@ number listed is the number of entries that can be recorded per
 CPU. To know the full size, multiply the number of possible CPUS
 with the number of entries.
 
- # cat /debug/tracing/buffer_size_kb
+ # cat /sys/kernel/debug/tracing/buffer_size_kb
 1408 (units kilobytes)
 
 Note, to modify this, you must have tracing completely disabled.
@@ -1808,18 +1808,18 @@ To do that, echo "nop" into the current_tracer. If the
 current_tracer is not set to "nop", an EINVAL error will be
 returned.
 
- # echo nop > /debug/tracing/current_tracer
- # echo 10000 > /debug/tracing/buffer_size_kb
- # cat /debug/tracing/buffer_size_kb
+ # echo nop > /sys/kernel/debug/tracing/current_tracer
+ # echo 10000 > /sys/kernel/debug/tracing/buffer_size_kb
+ # cat /sys/kernel/debug/tracing/buffer_size_kb
 10000 (units kilobytes)
 
 The number of pages which will be allocated is limited to a
 percentage of available memory. Allocating too much will produce
 an error.
 
- # echo 1000000000000 > /debug/tracing/buffer_size_kb
+ # echo 1000000000000 > /sys/kernel/debug/tracing/buffer_size_kb
 -bash: echo: write error: Cannot allocate memory
- # cat /debug/tracing/buffer_size_kb
+ # cat /sys/kernel/debug/tracing/buffer_size_kb
 85
 
 -----------
Documentation/trace/mmiotrace.txt +13 -13
index 5731c67..162effb 100644
@@ -32,41 +32,41 @@ is no way to automatically detect if you are losing events due to CPUs racing.
 Usage Quick Reference
 ---------------------
 
-$ mount -t debugfs debugfs /debug
-$ echo mmiotrace > /debug/tracing/current_tracer
-$ cat /debug/tracing/trace_pipe > mydump.txt &
+$ mount -t debugfs debugfs /sys/kernel/debug
+$ echo mmiotrace > /sys/kernel/debug/tracing/current_tracer
+$ cat /sys/kernel/debug/tracing/trace_pipe > mydump.txt &
 Start X or whatever.
-$ echo "X is up" > /debug/tracing/trace_marker
-$ echo nop > /debug/tracing/current_tracer
+$ echo "X is up" > /sys/kernel/debug/tracing/trace_marker
+$ echo nop > /sys/kernel/debug/tracing/current_tracer
 Check for lost events.
 
 
 Usage
 -----
 
-Make sure debugfs is mounted to /debug. If not, (requires root privileges)
-$ mount -t debugfs debugfs /debug
+Make sure debugfs is mounted to /sys/kernel/debug. If not, (requires root privileges)
+$ mount -t debugfs debugfs /sys/kernel/debug
 
 Check that the driver you are about to trace is not loaded.
 
 Activate mmiotrace (requires root privileges):
-$ echo mmiotrace > /debug/tracing/current_tracer
+$ echo mmiotrace > /sys/kernel/debug/tracing/current_tracer
 
 Start storing the trace:
-$ cat /debug/tracing/trace_pipe > mydump.txt &
+$ cat /sys/kernel/debug/tracing/trace_pipe > mydump.txt &
 The 'cat' process should stay running (sleeping) in the background.
 
 Load the driver you want to trace and use it. Mmiotrace will only catch MMIO
 accesses to areas that are ioremapped while mmiotrace is active.
 
 During tracing you can place comments (markers) into the trace by
-$ echo "X is up" > /debug/tracing/trace_marker
+$ echo "X is up" > /sys/kernel/debug/tracing/trace_marker
 This makes it easier to see which part of the (huge) trace corresponds to
 which action. It is recommended to place descriptive markers about what you
 do.
 
 Shut down mmiotrace (requires root privileges):
-$ echo nop > /debug/tracing/current_tracer
+$ echo nop > /sys/kernel/debug/tracing/current_tracer
 The 'cat' process exits. If it does not, kill it by issuing 'fg' command and
 pressing ctrl+c.
 
@@ -78,10 +78,10 @@ to view your kernel log and look for "mmiotrace has lost events" warning. If
 events were lost, the trace is incomplete. You should enlarge the buffers and
 try again. Buffers are enlarged by first seeing how large the current buffers
 are:
-$ cat /debug/tracing/buffer_size_kb
+$ cat /sys/kernel/debug/tracing/buffer_size_kb
 gives you a number. Approximately double this number and write it back, for
 instance:
-$ echo 128000 > /debug/tracing/buffer_size_kb
+$ echo 128000 > /sys/kernel/debug/tracing/buffer_size_kb
 Then start again from the top.
 
 If you are doing a trace for a driver project, e.g. Nouveau, you should also
drivers/block/pktcdvd.c +1 -1
index dc7a8c3..9aa8932 100644
@@ -430,7 +430,7 @@ static void pkt_sysfs_cleanup(void)
 /********************************************************************
   entries in debugfs
 
-  /debugfs/pktcdvd[0-7]/
+  /sys/kernel/debug/pktcdvd[0-7]/
 			info
 
  *******************************************************************/
drivers/gpu/drm/drm_debugfs.c +6 -6
index c77c6c6..6ce0e26 100644
@@ -105,7 +105,7 @@ int drm_debugfs_create_files(struct drm_info_list *files, int count,
 		ent = debugfs_create_file(files[i].name, S_IFREG | S_IRUGO,
 					  root, tmp, &drm_debugfs_fops);
 		if (!ent) {
-			DRM_ERROR("Cannot create /debugfs/dri/%s/%s\n",
+			DRM_ERROR("Cannot create /sys/kernel/debug/dri/%s/%s\n",
 				  name, files[i].name);
 			drm_free(tmp, sizeof(struct drm_info_node),
 				 _DRM_DRIVER);
@@ -133,9 +133,9 @@ EXPORT_SYMBOL(drm_debugfs_create_files);
  * \param minor device minor number
  * \param root DRI debugfs dir entry.
  *
- * Create the DRI debugfs root entry "/debugfs/dri", the device debugfs root entry
- * "/debugfs/dri/%minor%/", and each entry in debugfs_list as
- * "/debugfs/dri/%minor%/%name%".
+ * Create the DRI debugfs root entry "/sys/kernel/debug/dri", the device debugfs root entry
+ * "/sys/kernel/debug/dri/%minor%/", and each entry in debugfs_list as
+ * "/sys/kernel/debug/dri/%minor%/%name%".
  */
 int drm_debugfs_init(struct drm_minor *minor, int minor_id,
 		     struct dentry *root)
@@ -148,7 +148,7 @@ int drm_debugfs_init(struct drm_minor *minor, int minor_id,
 	sprintf(name, "%d", minor_id);
 	minor->debugfs_root = debugfs_create_dir(name, root);
 	if (!minor->debugfs_root) {
-		DRM_ERROR("Cannot create /debugfs/dri/%s\n", name);
+		DRM_ERROR("Cannot create /sys/kernel/debug/dri/%s\n", name);
 		return -1;
 	}
 
@@ -165,7 +165,7 @@ int drm_debugfs_init(struct drm_minor *minor, int minor_id,
 		ret = dev->driver->debugfs_init(minor);
 		if (ret) {
 			DRM_ERROR("DRM: Driver failed to initialize "
-				  "/debugfs/dri.\n");
+				  "/sys/kernel/debug/dri.\n");
 			return ret;
 		}
 	}
drivers/gpu/drm/drm_drv.c +1 -1
index 019b7c5..1bf7efd 100644
@@ -339,7 +339,7 @@ static int __init drm_core_init(void)
 
 	drm_debugfs_root = debugfs_create_dir("dri", NULL);
 	if (!drm_debugfs_root) {
-		DRM_ERROR("Cannot create /debugfs/dri\n");
+		DRM_ERROR("Cannot create /sys/kernel/debug/dri\n");
 		ret = -1;
 		goto err_p3;
 	}
drivers/gpu/drm/drm_stub.c +1 -1
index b9631e3..b2a93c4 100644
@@ -328,7 +328,7 @@ static int drm_get_minor(struct drm_device *dev, struct drm_minor **minor, int t
 #if defined(CONFIG_DEBUG_FS)
 	ret = drm_debugfs_init(new_minor, minor_id, drm_debugfs_root);
 	if (ret) {
-		DRM_ERROR("DRM: Failed to initialize /debugfs/dri.\n");
+		DRM_ERROR("DRM: Failed to initialize /sys/kernel/debug/dri.\n");
 		goto err_g2;
 	}
 #endif
drivers/net/Kconfig +2 -2
index 214a92d..a170cea 100644
@@ -2200,7 +2200,7 @@ config SKGE_DEBUG
        depends on SKGE && DEBUG_FS
        help
 	 This option adds the ability to dump driver state for debugging.
-	 The file debugfs/skge/ethX displays the state of the internal
+	 The file /sys/kernel/debug/skge/ethX displays the state of the internal
 	 transmit and receive rings.
 
 	 If unsure, say N.
@@ -2226,7 +2226,7 @@ config SKY2_DEBUG
        depends on SKY2 && DEBUG_FS
        help
 	 This option adds the ability to dump driver state for debugging.
-	 The file debugfs/sky2/ethX displays the state of the internal
+	 The file /sys/kernel/debug/sky2/ethX displays the state of the internal
 	 transmit and receive rings.
 
 	 If unsure, say N.
drivers/net/wimax/i2400m/i2400m.h +1 -1
index 3ae2df3..da2e8f2 100644
@@ -389,7 +389,7 @@ struct i2400m {
 	unsigned ready:1;		/* all probing steps done */
 	unsigned rx_reorder:1;		/* RX reorder is enabled */
 	u8 trace_msg_from_user;		/* echo rx msgs to 'trace' pipe */
-					/* typed u8 so debugfs/u8 can tweak */
+					/* typed u8 so /sys/kernel/debug/u8 can tweak */
 	enum i2400m_system_state state;
 	wait_queue_head_t state_wq;	/* Woken up when on state updates */
 
drivers/net/wireless/ath5k/Kconfig +3 -3
index 75383a5..fdfa708 100644
@@ -27,11 +27,11 @@ config ATH5K_DEBUG
 	  Say Y, if and you will get debug options for ath5k.
 	  To use this, you need to mount debugfs:
 
-	  mkdir /debug/
-	  mount -t debugfs debug /debug/
+	  mkdir /sys/kernel/debug/
+	  mount -t debugfs debug /sys/kernel/debug/
 
 	  You will get access to files under:
-	  /debug/ath5k/phy0/
+	  /sys/kernel/debug/ath5k/phy0/
 
 	  To enable debug, pass the debug level to the debug module
 	  parameter. For example:
drivers/net/wireless/libertas/README +6 -6
index d860fc3..ab6a2d5 100644
@@ -72,7 +72,7 @@ rdrf
 	location that is to be read.  This parameter must be specified in
 	hexadecimal (its possible to preceed preceding the number with a "0x").
 
-	Path: /debugfs/libertas_wireless/ethX/registers/
+	Path: /sys/kernel/debug/libertas_wireless/ethX/registers/
 
 	Usage:
 		echo "0xa123" > rdmac ; cat rdmac
@@ -95,7 +95,7 @@ wrrf
 sleepparams
 	This command is used to set the sleepclock configurations
 
-	Path: /debugfs/libertas_wireless/ethX/
+	Path: /sys/kernel/debug/libertas_wireless/ethX/
 
 	Usage:
 		cat sleepparams: reads the current sleepclock configuration
@@ -115,7 +115,7 @@ subscribed_events
 	The subscribed_events directory contains the interface for the
 	subscribed events API.
 
-	Path: /debugfs/libertas_wireless/ethX/subscribed_events/
+	Path: /sys/kernel/debug/libertas_wireless/ethX/subscribed_events/
 
 	Each event is represented by a filename. Each filename consists of the
 	following three fields:
@@ -165,7 +165,7 @@ subscribed_events
 extscan
 	This command is used to do a specific scan.
 
-	Path: /debugfs/libertas_wireless/ethX/
+	Path: /sys/kernel/debug/libertas_wireless/ethX/
 
 	Usage: echo "SSID" > extscan
 
@@ -179,7 +179,7 @@ getscantable
 	Display the current contents of the driver scan table (ie. get the
 	scan results).
 
-	Path: /debugfs/libertas_wireless/ethX/
+	Path: /sys/kernel/debug/libertas_wireless/ethX/
 
 	Usage:
 		cat getscantable
@@ -188,7 +188,7 @@ setuserscan
 	Initiate a customized scan and retrieve the results
 
 
-	Path: /debugfs/libertas_wireless/ethX/
+	Path: /sys/kernel/debug/libertas_wireless/ethX/
 
     Usage:
        echo "[ARGS]" > setuserscan
drivers/scsi/lpfc/lpfc_debugfs.c +2 -2
index 52be564..9112380 100644
@@ -51,8 +51,8 @@
  * debugfs interface
  *
  * To access this interface the user should:
- * # mkdir /debug
- * # mount -t debugfs none /debug
+ * # mkdir /sys/kernel/debug
+ * # mount -t debugfs none /sys/kernel/debug
  *
  * The lpfc debugfs directory hierarchy is:
  * lpfc/lpfcX/vportY
include/linux/kernel.h +1 -1
index 883cd44..99b7aad 100644
@@ -406,7 +406,7 @@ static inline char *pack_hex_byte(char *buf, u8 byte)
  *
  * Use tracing_on/tracing_off when you want to quickly turn on or off
  * tracing. It simply enables or disables the recording of the trace events.
- * This also corresponds to the user space debugfs/tracing/tracing_on
+ * This also corresponds to the user space /sys/kernel/debug/tracing/tracing_on
  * file, which gives a means for the kernel and userspace to interact.
  * Place a tracing_off() in the kernel where you want tracing to end.
  * From user space, examine the trace, and then echo 1 > tracing_on
include/linux/tracepoint.h +2 -2
index d35a7ee..9c9ac7d 100644
@@ -197,7 +197,7 @@ static inline void tracepoint_synchronize_unregister(void)
  *	* This is how the trace record is structured and will
  *	* be saved into the ring buffer. These are the fields
  *	* that will be exposed to user-space in
- *	* /debug/tracing/events/<*>/format.
+ *	* /sys/kernel/debug/tracing/events/<*>/format.
  *	*
  *	* The declared 'local variable' is called '__entry'
  *	*
@@ -257,7 +257,7 @@ static inline void tracepoint_synchronize_unregister(void)
  * tracepoint callback (this is used by programmatic plugins and
  * can also by used by generic instrumentation like SystemTap), and
  * it is also used to expose a structured trace record in
- * /debug/tracing/events/.
+ * /sys/kernel/debug/tracing/events/.
  */
 
 #define TRACE_EVENT(name, proto, args, struct, assign, print)	\
kernel/trace/Kconfig +5 -5
index 417d198..e5ce28f 100644
@@ -120,7 +120,7 @@ config IRQSOFF_TRACER
 	  disabled by default and can be runtime (re-)started
 	  via:
 
-	      echo 0 > /debugfs/tracing/tracing_max_latency
+	      echo 0 > /sys/kernel/debug/tracing/tracing_max_latency
 
 	  (Note that kernel size and overhead increases with this option
 	  enabled. This option and the preempt-off timing option can be
@@ -141,7 +141,7 @@ config PREEMPT_TRACER
 	  disabled by default and can be runtime (re-)started
 	  via:
 
-	      echo 0 > /debugfs/tracing/tracing_max_latency
+	      echo 0 > /sys/kernel/debug/tracing/tracing_max_latency
 
 	  (Note that kernel size and overhead increases with this option
 	  enabled. This option and the irqs-off timing option can be
@@ -213,7 +213,7 @@ config TRACE_BRANCH_PROFILING
 	  This tracer profiles all the the likely and unlikely macros
 	  in the kernel. It will display the results in:
 
-	  /debugfs/tracing/profile_annotated_branch
+	  /sys/kernel/debug/tracing/profile_annotated_branch
 
 	  Note: this will add a significant overhead, only turn this
 	  on if you need to profile the system's use of these macros.
@@ -228,7 +228,7 @@ config PROFILE_ALL_BRANCHES
 	  taken in the kernel is recorded whether it hit or miss.
 	  The results will be displayed in:
 
-	  /debugfs/tracing/profile_branch
+	  /sys/kernel/debug/tracing/profile_branch
 
 	  This configuration, when enabled, will impose a great overhead
 	  on the system. This should only be enabled when the system
@@ -276,7 +276,7 @@ config STACK_TRACER
 	select KALLSYMS
 	help
 	  This special tracer records the maximum stack footprint of the
-	  kernel and displays it in debugfs/tracing/stack_trace.
+	  kernel and displays it in /sys/kernel/debug/tracing/stack_trace.
 
 	  This tracer works by hooking into every function call that the
 	  kernel executes, and keeping a maximum stack depth value and
kernel/trace/trace.c +12 -12
index cda81ec..9b94d57 100644
@@ -335,7 +335,7 @@ static raw_spinlock_t ftrace_max_lock =
 /*
  * Copy the new maximum trace into the separate maximum-trace
  * structure. (this way the maximum trace is permanently saved,
- * for later retrieval via /debugfs/tracing/latency_trace)
+ * for later retrieval via /sys/kernel/debug/tracing/latency_trace)
  */
 static void
 __update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu)
@@ -2366,21 +2366,21 @@ static const struct file_operations tracing_iter_fops = {
 
 static const char readme_msg[] =
 	"tracing mini-HOWTO:\n\n"
-	"# mkdir /debug\n"
-	"# mount -t debugfs nodev /debug\n\n"
-	"# cat /debug/tracing/available_tracers\n"
+	"# mkdir /sys/kernel/debug\n"
+	"# mount -t debugfs nodev /sys/kernel/debug\n\n"
+	"# cat /sys/kernel/debug/tracing/available_tracers\n"
 	"wakeup preemptirqsoff preemptoff irqsoff function sched_switch nop\n\n"
-	"# cat /debug/tracing/current_tracer\n"
+	"# cat /sys/kernel/debug/tracing/current_tracer\n"
 	"nop\n"
-	"# echo sched_switch > /debug/tracing/current_tracer\n"
-	"# cat /debug/tracing/current_tracer\n"
+	"# echo sched_switch > /sys/kernel/debug/tracing/current_tracer\n"
+	"# cat /sys/kernel/debug/tracing/current_tracer\n"
 	"sched_switch\n"
-	"# cat /debug/tracing/trace_options\n"
+	"# cat /sys/kernel/debug/tracing/trace_options\n"
 	"noprint-parent nosym-offset nosym-addr noverbose\n"
-	"# echo print-parent > /debug/tracing/trace_options\n"
-	"# echo 1 > /debug/tracing/tracing_enabled\n"
-	"# cat /debug/tracing/trace > /tmp/trace.txt\n"
-	"# echo 0 > /debug/tracing/tracing_enabled\n"
+	"# echo print-parent > /sys/kernel/debug/tracing/trace_options\n"
+	"# echo 1 > /sys/kernel/debug/tracing/tracing_enabled\n"
+	"# cat /sys/kernel/debug/tracing/trace > /tmp/trace.txt\n"
+	"# echo 0 > /sys/kernel/debug/tracing/tracing_enabled\n"
 ;
 
 static ssize_t
scripts/tracing/draw_functrace.py +4 -4
index 902f9a9..66f4d1b 100644
@@ -12,10 +12,10 @@ calls. Only the functions's names and the the call time are provided.
 
 Usage:
 	Be sure that you have CONFIG_FUNCTION_TRACER
-	# mkdir /debugfs
-	# mount -t debug debug /debug
-	# echo function > /debug/tracing/current_tracer
-	$ cat /debug/tracing/trace_pipe > ~/raw_trace_func
+	# mkdir /sys/kernel/debug
+	# mount -t debug debug /sys/kernel/debug
+	# echo function > /sys/kernel/debug/tracing/current_tracer
+	$ cat /sys/kernel/debug/tracing/trace_pipe > ~/raw_trace_func
 	Wait some times but not too much, the script is a bit slow.
 	Break the pipe (Ctrl + Z)
 	$ scripts/draw_functrace.py < raw_trace_func > draw_functrace
Signature · 14 lines
1.6.3.1

 
-----------------------------------------------
To unsubscribe from this list: send the line "unsubscribe linux-***" 
in the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

GeunSik Lim (  S A M S U N G     E L E C T R O N I C S  )
e-Mail  :1) [email protected]
         2) [email protected] , [email protected]
HomePage: http://blog.naver.com/invain/
-----------------------------------------------
maintainer
#17 lore raw
Quoted 6 lines · Steven Rostedt

===
When debugfs is configured into the kernel (which selecting any ftrace option will do) the directory /sys/kernel/debug will be created. To mount this directory, you can add to your /etc/fstab file:

debugfs /debug debugfs defaults 0 0

s,/debug,/sys/kernel/debug,

Quoted 8 lines

Or you can mount it at run time with:

mount -t debugfs nodev /sys/kernel/debug

For quicker access to that directory you may want to make a soft link to it:

Quoted 8 lines

ln -s /sys/kernel/debug /debug

Any selected ftrace option will also create a directory called tracing within the debugfs. The rest of the document will assume that you are in the ftrace directory (cd /sys/kernel/debugfs) and will only concentrate on the files within that directory and not distract from the content with the extended "/sys/kernel/debug/tracing" path name.
===

-- Steve

maintainer
#18 lore raw
In reply to Steven Rostedt
Quoted 11 lines · Steven Rostedt

===
When debugfs is configured into the kernel (which selecting any ftrace option will do) the directory /sys/kernel/debug will be created. To mount this directory, you can add to your /etc/fstab file:

debugfs /debug debugfs defaults 0 0

Or you can mount it at run time with:

mount -t debugfs nodev /sys/kernel/debug

Quoted 8 lines

For quicker access to that directory you may want to make a soft link to it:

ln -s /sys/kernel/debug /debug

Any selected ftrace option will also create a directory called tracing within the debugfs. The rest of the document will assume that you are in the ftrace directory (cd /sys/kernel/debugfs) and will only concentrate

s,/sys/kernel/debugfs,/sys/kernel/debug/tracing,

Quoted 3 lines

on the files within that directory and not distract from the content with the extended "/sys/kernel/debug/tracing" path name.
===

-- Steve

maintainer
#19 lore raw
In reply to GeunSik Lim
Quoted 21 lines · GeunSik Lim

Dear Greg and Steen,

Thanks your advices and opinions about a confusion according to
too many mount directory names for debugfs. I made rfc v2 patch file again like belows. Can I hear your advices about below patch contents?.

* ref:
http://groups.google.com/group/linux.kernel/browse_thread/thread/c34e386587979340/dc52317f888e78b6?#dc52317f888e78b6

Regards,
GeunSik Lim.

Subject: [PATCH] Fix terminology inconsistency of dir name to mount debugfs filesystem.

Many developers use "/debug/" or "/debugfs/ directory name to mount debugfs filesystem for ftrace according to ./Documentation/tracers/ftrace.txt file.

Quoted 8 lines

And, three directory names(ex:/debug/, /debugfs/, /sys/kernel/debug/) is existed in kernel source like ftrace, DRM, Wireless, Network[sky2]files to mount debugfs filesystem.

debugfs means debug filesystem for debugging easy to use by greg kroah hartman. "/sys/kernel/debug/" name is suitable as directory name of debugfs filesystem.
-debugfs related reference: http://lwn.net/Articles/115405/

Note, LWN has an update this week:

http://lwn.net/Articles/334546/

But it currently is subscriber only. Others would need to either buy a subscription (plug) or wait a week for it to become open.

Still for some aspects, just having "debugfs" without the leading "/" might just be self explanatory. We can assume the user already knows where debugfs is mounted.

-- Steve

Quoted 8 lines

Fix inconsistency of directory name to mount debugfs filesystem.

#20 lore raw

Thanks your advice. I agree with your reply about this patch. I updated contents that you advised like belows.
If I mistake, Please correct me.

And, I also need your review about below 2 contents in the ftrace.txt file.

Thanks,
GeunSik Lim.

1)
@@ -1146,7 +1158,7 @@ int trace_fd;
[...]
int main(int argc, char *argv[]) {
[...]
- trace_fd = open("/debug/tracing/tracing_enabled", O_WRONLY);
+ trace_fd = open("/sys/kernel/debug/tracing/tracing_enabled", O_WRONLY);
[...]
if (condition_hit()) {
write(trace_fd, "0", 1);
@@ -1154,26 +1166,23 @@ int main(int argc, char *argv[]) {
[...]
}

@@ -1217,12 +1226,12 @@ int main (int argc, char **argv)

                 char line[64];
                 int s;

- ffd = open("/debug/tracing/current_tracer", O_WRONLY);
+ ffd =

open("/sys/kernel/debug/tracing/current_tracer", O_WRONLY);
                 if (ffd < 0)
                         exit(-1);
                 write(ffd, "nop", 3);

- fd = open("/debug/tracing/set_ftrace_pid", O_WRONLY);
+ fd = open("/sys/kernel/debug/tracing/set_ftrace_pid",

O_WRONLY);
                 s = sprintf(line, "%d\n", getpid());
                 write(fd, line, s);

-Note: Here we hard coded the path name. The debugfs mount is not -guaranteed to be at /debug (and is more commonly at -/sys/kernel/debug). For simple one time traces, the above is -sufficent. For anything else, a search through /proc/mounts may +Note: Here we hard coded the path name. A search through /proc/mounts may
be needed to find where the debugfs file-system is mounted.

2)
More details can be found in the source code, in the -kernel/tracing/*.c files.
+kernel/trace/*.c files

* about ftrace.txt file
---

Documentation/trace/ftrace.txt +30 -18
index fd9a3e6..790199b 100644
@@ -7,7 +7,6 @@ Copyright 2008 Red Hat Inc.
                (dual licensed under the GPL v2)
 Reviewers:   Elias Oltmanns, Randy Dunlap, Andrew Morton,
 	     John Kacur, and David Teigland.
-
 Written for: 2.6.28-rc2
 
 Introduction
@@ -33,13 +32,26 @@ The File System
 Ftrace uses the debugfs file system to hold the control files as
 well as the files to display output.
 
-To mount the debugfs system:
+When debugfs is configured into the kernel (which selecting any ftrace
+option will do) the directory /sys/kernel/debug will be created. To mount
+this directory, you can add to your /etc/fstab file:
+
+ debugfs       /sys/kernel/debug          debugfs defaults        0       0
+
+Or you can mount it at run time with:
+
+ mount -t debugfs nodev /sys/kernel/debug
+
+For quicker access to that directory you may want to make a soft link to
+it:
 
-  # mkdir /debug
-  # mount -t debugfs nodev /debug
+ ln -s /sys/kernel/debug /debug
 
-( Note: it is more common to mount at /sys/kernel/debug, but for
-  simplicity this document will use /debug)
+Any selected ftrace option will also create a directory called tracing
+within the debugfs. The rest of the document will assume that you are in
+the ftrace directory (cd /sys/kernel/debug/tracing) and will only concentrate
+on the files within that directory and not distract from the content with
+the extended "/sys/kernel/debug/tracing" path name.
 
 That's it! (assuming that you have ftrace configured into your kernel)
 
@@ -389,18 +401,18 @@ trace_options
 The trace_options file is used to control what gets printed in
 the trace output. To see what is available, simply cat the file:
 
-  cat /debug/tracing/trace_options
+  cat ./trace_options
   print-parent nosym-offset nosym-addr noverbose noraw nohex nobin \
   noblock nostacktrace nosched-tree nouserstacktrace nosym-userobj
 
 To disable one of the options, echo in the option prepended with
 "no".
 
-  echo noprint-parent > /debug/tracing/trace_options
+  echo noprint-parent > ./trace_options
 
 To enable an option, leave off the "no".
 
-  echo sym-offset > /debug/tracing/trace_options
+  echo sym-offset > ./trace_options
 
 Here are the available options:
 
@@ -476,11 +488,11 @@ sched_switch
 This tracer simply records schedule switches. Here is an example
 of how to use it.
 
- # echo sched_switch > /debug/tracing/current_tracer
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo sched_switch > ./current_tracer
+ # echo 1 > ./tracing_enabled
  # sleep 1
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/trace
+ # echo 0 > ./tracing_enabled
+ # cat ./trace
 
 # tracer: sched_switch
 #
@@ -574,13 +586,13 @@ new trace is saved.
 To reset the maximum, echo 0 into tracing_max_latency. Here is
 an example:
 
- # echo irqsoff > /debug/tracing/current_tracer
- # echo 0 > /debug/tracing/tracing_max_latency
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo irqsoff > ./current_tracer
+ # echo 0 > ./tracing_max_latency
+ # echo 1 > ./tracing_enabled
  # ls -ltr
  [...]
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/latency_trace
+ # echo 0 > ./tracing_enabled
+ # cat ./latency_trace
 # tracer: irqsoff
 #
 irqsoff latency trace v1.1.5 on 2.6.26
@@ -681,13 +693,13 @@ Like the irqsoff tracer, it records the maximum

latency for
which preemption was disabled. The control of preemptoff tracer is much like the irqsoff tracer.

- # echo preemptoff > /debug/tracing/current_tracer
- # echo 0 > /debug/tracing/tracing_max_latency
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo preemptoff > ./current_tracer
+ # echo 0 > ./tracing_max_latency
+ # echo 1 > ./tracing_enabled

  # ls -ltr
  [...]
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/latency_trace
+ # echo 0 > ./tracing_enabled
+ # cat ./latency_trace
 # tracer: preemptoff
 #
 preemptoff latency trace v1.1.5 on 2.6.26-rc8
@@ -828,13 +840,13 @@ tracer.
 Again, using this trace is much like the irqsoff and preemptoff
 tracers.

- # echo preemptirqsoff > /debug/tracing/current_tracer
- # echo 0 > /debug/tracing/tracing_max_latency
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo preemptirqsoff > ./current_tracer
+ # echo 0 > ./tracing_max_latency
+ # echo 1 > ./tracing_enabled

  # ls -ltr
  [...]
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/latency_trace
+ # echo 0 > ./tracing_enabled
+ # cat ./latency_trace
 # tracer: preemptirqsoff
 #
 preemptirqsoff latency trace v1.1.5 on 2.6.26-rc8
@@ -990,12 +1002,12 @@ slightly differently than we did with the
previous tracers.
 Instead of performing an 'ls', we will run 'sleep 1' under
 'chrt' which changes the priority of the task.

- # echo wakeup > /debug/tracing/current_tracer
- # echo 0 > /debug/tracing/tracing_max_latency
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo wakeup > ./current_tracer
+ # echo 0 > ./tracing_max_latency
+ # echo 1 > ./tracing_enabled

  # chrt -f 5 sleep 1
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/latency_trace
+ # echo 0 > ./tracing_enabled
+ # cat ./latency_trace
 # tracer: wakeup
 #
 wakeup latency trace v1.1.5 on 2.6.26-rc8
@@ -1105,11 +1117,11 @@ can be done from the debug file system. Make sure the
 ftrace_enabled is set; otherwise this tracer is a nop.
  # sysctl kernel.ftrace_enabled=1
- # echo function > /debug/tracing/current_tracer
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo function > ./current_tracer
+ # echo 1 > ./tracing_enabled
  # usleep 1
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/trace
+ # echo 0 > ./tracing_enabled
+ # cat ./trace
 # tracer: function
 #
 #           TASK-PID   CPU#    TIMESTAMP  FUNCTION
@@ -1146,7 +1158,7 @@ int trace_fd;
 [...]
 int main(int argc, char *argv[]) {
 	[...]
-	trace_fd = open("/debug/tracing/tracing_enabled", O_WRONLY);
+	trace_fd = open("/sys/kernel/debug/tracing/tracing_enabled", O_WRONLY);
 	[...]
 	if (condition_hit()) {
 		write(trace_fd, "0", 1);
@@ -1154,26 +1166,23 @@ int main(int argc, char *argv[]) {
 	[...]
 }

-Note: Here we hard coded the path name. The debugfs mount is not -guaranteed to be at /debug (and is more commonly at -/sys/kernel/debug). For simple one time traces, the above is -sufficent. For anything else, a search through /proc/mounts may +Note: Here we hard coded the path name. A search through /proc/mounts may
be needed to find where the debugfs file-system is mounted.

Single thread tracing
---------------------

-By writing into /debug/tracing/set_ftrace_pid you can trace a +By writing into ./set_ftrace_pid you can trace a
single thread. For example:

-# cat /debug/tracing/set_ftrace_pid
+# cat ./set_ftrace_pid
no pid
-# echo 3111 > /debug/tracing/set_ftrace_pid
-# cat /debug/tracing/set_ftrace_pid
+# echo 3111 > ./set_ftrace_pid
+# cat ./set_ftrace_pid
3111
-# echo function > /debug/tracing/current_tracer
-# cat /debug/tracing/trace | head
+# echo function > ./current_tracer
+# cat ./trace | head
# tracer: function
#
# TASK-PID CPU# TIMESTAMP FUNCTION
@@ -1184,8 +1193,8 @@ no pid

      yum-updatesd-3111  [003]  1637.254683: lock_hrtimer_base
<-hrtimer_try_to_cancel
      yum-updatesd-3111  [003]  1637.254685: fget_light <-do_sys_poll
      yum-updatesd-3111  [003]  1637.254686: pipe_poll <-do_sys_poll

-# echo -1 > /debug/tracing/set_ftrace_pid
-# cat /debug/tracing/trace |head
+# echo -1 > ./set_ftrace_pid
+# cat ./trace |head
# tracer: function
#
# TASK-PID CPU# TIMESTAMP FUNCTION
@@ -1217,12 +1226,12 @@ int main (int argc, char **argv)

                 char line[64];
                 int s;

- ffd = open("/debug/tracing/current_tracer", O_WRONLY);
+ ffd =

open("/sys/kernel/debug/tracing/current_tracer", O_WRONLY);
                 if (ffd < 0)
                         exit(-1);
                 write(ffd, "nop", 3);

- fd = open("/debug/tracing/set_ftrace_pid", O_WRONLY);
+ fd = open("/sys/kernel/debug/tracing/set_ftrace_pid",

O_WRONLY);
                 s = sprintf(line, "%d\n", getpid());
                 write(fd, line, s);

@@ -1374,22 +1383,22 @@ want, depending on your needs.
tracing_cpu_mask file) or you might sometimes see unordered function calls while cpu tracing switch.

- hide: echo nofuncgraph-cpu > /debug/tracing/trace_options
- show: echo funcgraph-cpu > /debug/tracing/trace_options
+ hide: echo nofuncgraph-cpu > ./trace_options
+ show: echo funcgraph-cpu > ./trace_options

- The duration (function's time of execution) is displayed on the closing bracket line of a function or on the same line than the current function in case of a leaf one. It is default enabled.

- hide: echo nofuncgraph-duration > /debug/tracing/trace_options
- show: echo funcgraph-duration > /debug/tracing/trace_options
+ hide: echo nofuncgraph-duration > ./trace_options
+ show: echo funcgraph-duration > ./trace_options

- The overhead field precedes the duration field in case of reached duration thresholds.

- hide: echo nofuncgraph-overhead > /debug/tracing/trace_options
- show: echo funcgraph-overhead > /debug/tracing/trace_options
+ hide: echo nofuncgraph-overhead > ./trace_options
+ show: echo funcgraph-overhead > ./trace_options
depends on: funcgraph-duration

ie:
@@ -1418,8 +1427,8 @@ want, depending on your needs.
- The task/pid field displays the thread cmdline and pid which executed the function. It is default disabled.

- hide: echo nofuncgraph-proc > /debug/tracing/trace_options
- show: echo funcgraph-proc > /debug/tracing/trace_options
+ hide: echo nofuncgraph-proc > ./trace_options
+ show: echo funcgraph-proc > ./trace_options

ie:

@@ -1442,8 +1451,8 @@ want, depending on your needs.
system clock since it started. A snapshot of this time is given on each entry/exit of functions

- hide: echo nofuncgraph-abstime > /debug/tracing/trace_options
- show: echo funcgraph-abstime > /debug/tracing/trace_options
+ hide: echo nofuncgraph-abstime > ./trace_options
+ show: echo funcgraph-abstime > ./trace_options

ie:

@@ -1540,7 +1549,7 @@ listed in:

available_filter_functions

- # cat /debug/tracing/available_filter_functions
+ # cat ./available_filter_functions put_prev_task_idle
kmem_cache_create
pick_next_task_rt
@@ -1552,12 +1561,12 @@ mutex_lock
If I am only interested in sys_nanosleep and hrtimer_interrupt:

  # echo sys_nanosleep hrtimer_interrupt \
-		> /debug/tracing/set_ftrace_filter
- # echo ftrace > /debug/tracing/current_tracer
- # echo 1 > /debug/tracing/tracing_enabled
+		> ./set_ftrace_filter
+ # echo ftrace > ./current_tracer
+ # echo 1 > ./tracing_enabled
  # usleep 1
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/trace
+ # echo 0 > ./tracing_enabled
+ # cat ./trace
 # tracer: ftrace
 #
 #           TASK-PID   CPU#    TIMESTAMP  FUNCTION
@@ -1568,7 +1577,7 @@ If I am only interested in sys_nanosleep and
hrtimer_interrupt:

To see which functions are being traced, you can cat the file:

- # cat /debug/tracing/set_ftrace_filter
+ # cat ./set_ftrace_filter
hrtimer_interrupt
sys_nanosleep

@@ -1588,7 +1597,7 @@ Note: It is better to use quotes to enclose the wild cards,
otherwise the shell may expand the parameters into names of files in the local directory.

- # echo 'hrtimer_*' > /debug/tracing/set_ftrace_filter
+ # echo 'hrtimer_*' > ./set_ftrace_filter

Produces:

@@ -1609,7 +1618,7 @@ Produces:

Notice that we lost the sys_nanosleep.

- # cat /debug/tracing/set_ftrace_filter
+ # cat ./set_ftrace_filter
hrtimer_run_queues
hrtimer_run_pending
hrtimer_init
@@ -1635,17 +1644,17 @@ To append to the filters, use '>>'
To clear out a filter so that all functions will be recorded again:

- # echo > /debug/tracing/set_ftrace_filter
- # cat /debug/tracing/set_ftrace_filter
+ # echo > ./set_ftrace_filter
+ # cat ./set_ftrace_filter
#

Again, now we want to append.

- # echo sys_nanosleep > /debug/tracing/set_ftrace_filter
- # cat /debug/tracing/set_ftrace_filter
+ # echo sys_nanosleep > ./set_ftrace_filter
+ # cat ./set_ftrace_filter
sys_nanosleep
- # echo 'hrtimer_*' >> /debug/tracing/set_ftrace_filter
- # cat /debug/tracing/set_ftrace_filter
+ # echo 'hrtimer_*' >> ./set_ftrace_filter
+ # cat ./set_ftrace_filter
hrtimer_run_queues
hrtimer_run_pending
hrtimer_init
@@ -1668,7 +1677,7 @@ hrtimer_init_sleeper
The set_ftrace_notrace prevents those functions from being traced.

- # echo '*preempt*' '*lock*' > /debug/tracing/set_ftrace_notrace
+ # echo '*preempt*' '*lock*' > ./set_ftrace_notrace

Produces:

@@ -1758,13 +1767,13 @@ the effect on the tracing is different. Every read from
trace_pipe is consumed. This means that subsequent reads will be different. The trace is live.

- # echo function > /debug/tracing/current_tracer
- # cat /debug/tracing/trace_pipe > /tmp/trace.out &
+ # echo function > ./current_tracer
+ # cat ./trace_pipe > /tmp/trace.out &
[1] 4153
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo 1 > ./tracing_enabled

  # usleep 1
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/trace
+ # echo 0 > ./tracing_enabled
+ # cat ./trace
 # tracer: function
 #
 #           TASK-PID   CPU#    TIMESTAMP  FUNCTION
@@ -1800,7 +1809,7 @@ number listed is the number of entries that can
be recorded per
 CPU. To know the full size, multiply the number of possible CPUS
 with the number of entries.

- # cat /debug/tracing/buffer_size_kb
+ # cat ./buffer_size_kb
1408 (units kilobytes)

Note, to modify this, you must have tracing completely disabled.
@@ -1808,21 +1817,21 @@ To do that, echo "nop" into the current_tracer. If the
current_tracer is not set to "nop", an EINVAL error will be returned.

- # echo nop > /debug/tracing/current_tracer
- # echo 10000 > /debug/tracing/buffer_size_kb
- # cat /debug/tracing/buffer_size_kb
+ # echo nop > ./current_tracer
+ # echo 10000 > ./buffer_size_kb
+ # cat ./buffer_size_kb
10000 (units kilobytes)

The number of pages which will be allocated is limited to a percentage of available memory. Allocating too much will produce an error.

- # echo 1000000000000 > /debug/tracing/buffer_size_kb
+ # echo 1000000000000 > ./buffer_size_kb -bash: echo: write error: Cannot allocate memory
- # cat /debug/tracing/buffer_size_kb
+ # cat ./buffer_size_kb
85

---

Quoted 58 lines · Steven Rostedt
Quoted 21 lines · GeunSik Lim

Dear Greg and Steen,

Thanks your advices and opinions about a confusion according to
too many mount directory names for debugfs. I made rfc v2 patch file again like belows. Can I hear your advices about below patch contents?.

* ref:
http://groups.google.com/group/linux.kernel/browse_thread/thread/c34e386587979340/dc52317f888e78b6?#dc52317f888e78b6

Regards,
GeunSik Lim.

Subject: [PATCH]   Fix terminology inconsistency of dir name to mount debugfs filesystem.

      Many developers use "/debug/" or "/debugfs/ directory name to mount debugfs filesystem for ftrace according to ./Documentation/tracers/ftrace.txt file.

Quoted 8 lines

      And, three directory names(ex:/debug/, /debugfs/, /sys/kernel/debug/) is existed in kernel source like ftrace, DRM, Wireless, Network[sky2]files to mount debugfs filesystem.

      debugfs means debug filesystem for debugging easy to use by greg kroah hartman. "/sys/kernel/debug/" name is suitable as directory name of debugfs filesystem.
      -debugfs related reference: http://lwn.net/Articles/115405/

Note, LWN has an update this week:

 http://lwn.net/Articles/334546/

But it currently is subscriber only. Others would need to either buy a subscription (plug) or wait a week for it to become open.

Still for some aspects, just having "debugfs" without the leading "/" might just be self explanatory. We can assume the user already knows where debugfs is mounted.

-- Steve

Quoted 8 lines

      Fix inconsistency of directory name to mount debugfs filesystem.

Signature · 10 lines
Regards,
GeunSik Lim ( SAMSUNG ELECTRONICS)
Blog : http://blog.naver.com/invain/
e-Mail: [email protected]
           [email protected] , [email protected]
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/
maintainer
#21 lore raw
Quoted 41 lines · GeunSik Lim

Thanks your advice. I agree with your reply about this patch. I updated contents that you advised like belows.
If I mistake, Please correct me.

And, I also need your review about below 2 contents in the ftrace.txt file.

Thanks,
GeunSik Lim.

1)

(quoted diff) +3 -3
@@ -1146,7 +1158,7 @@ int trace_fd;
[...]
 int main(int argc, char *argv[]) {
[...]
-	trace_fd = open("/debug/tracing/tracing_enabled", O_WRONLY);
+	trace_fd = open("/sys/kernel/debug/tracing/tracing_enabled", O_WRONLY);
[...]
 	if (condition_hit()) {
 		write(trace_fd, "0", 1);
@@ -1154,26 +1166,23 @@ int main(int argc, char *argv[]) {
[...]
 }
@@ -1217,12 +1226,12 @@ int main (int argc, char **argv)
                 char line[64];
                 int s;
 
-                ffd = open("/debug/tracing/current_tracer", O_WRONLY);
+                ffd =
 open("/sys/kernel/debug/tracing/current_tracer", O_WRONLY);
                 if (ffd < 0)
                         exit(-1);
                 write(ffd, "nop", 3);
 
-                fd = open("/debug/tracing/set_ftrace_pid", O_WRONLY);
+                fd = open("/sys/kernel/debug/tracing/set_ftrace_pid",
 O_WRONLY);
                 s = sprintf(line, "%d\n", getpid());
                 write(fd, line, s);

Review

Quoted 6 lines
(quoted diff) +1 -4
-Note: Here we hard coded the path name. The debugfs mount is not
-guaranteed to be at /debug (and is more commonly at
-/sys/kernel/debug). For simple one time traces, the above is
-sufficent. For anything else, a search through /proc/mounts may
+Note: Here we hard coded the path name. A search through /proc/mounts may
 be needed to find where the debugfs file-system is mounted.

Here's the function I use:

#define _STR(x) x
#define STR(x) _STR(x)
#define MAX_PATH 256

const char *find_debugfs(void)

{
	static char debugfs[MAX_PATH+1];
	static int debugfs_found;
	char type[100];
	FILE *fp;

	if (debugfs_found)
		return debugfs;


	if ((fp = fopen("/proc/mounts","r")) == NULL) {
		perror("/proc/mounts");
		return NULL;
	}

	while (fscanf(fp, "%*s %"
		      STR(MAX_PATH)
		      "s %99s %*s %*d %*d\n",
		      debugfs, type) == 2) {
		if (strcmp(type, "debugfs") == 0)
			break;
	}
	fclose(fp);

	if (strcmp(type, "debugfs") != 0) {
		fprintf(stderr, "debugfs not mounted");
		return NULL;
	}

	debugfs_found = 1;

	return debugfs;
}

This might be a good idea to show it. This is the most robust since it does not hard code the debugfs directory.

Quoted 55 lines

2)
More details can be found in the source code, in the

(quoted diff) +1 -1
-kernel/tracing/*.c files.
+kernel/trace/*.c files

* about ftrace.txt file
---

Documentation/trace/ftrace.txt +18 -6
index fd9a3e6..790199b 100644
@@ -7,7 +7,6 @@ Copyright 2008 Red Hat Inc.
                (dual licensed under the GPL v2)
 Reviewers:   Elias Oltmanns, Randy Dunlap, Andrew Morton,
 	     John Kacur, and David Teigland.
-
 Written for: 2.6.28-rc2
 
 Introduction
@@ -33,13 +32,26 @@ The File System
 Ftrace uses the debugfs file system to hold the control files as
 well as the files to display output.
 
-To mount the debugfs system:
+When debugfs is configured into the kernel (which selecting any ftrace
+option will do) the directory /sys/kernel/debug will be created. To mount
+this directory, you can add to your /etc/fstab file:
+
+ debugfs       /sys/kernel/debug          debugfs defaults        0       0
+
+Or you can mount it at run time with:
+
+ mount -t debugfs nodev /sys/kernel/debug
+
+For quicker access to that directory you may want to make a soft link to
+it:
 
-  # mkdir /debug
-  # mount -t debugfs nodev /debug
+ ln -s /sys/kernel/debug /debug
 
-( Note: it is more common to mount at /sys/kernel/debug, but for
-  simplicity this document will use /debug)
+Any selected ftrace option will also create a directory called tracing
+within the debugfs. The rest of the document will assume that you are in
+the ftrace directory (cd /sys/kernel/debug/tracing) and will only concentrate
+on the files within that directory and not distract from the content with
+the extended "/sys/kernel/debug/tracing" path name.
 
 That's it! (assuming that you have ftrace configured into your kernel)
 

Documentation/trace/ftrace.txt trace_options · line 404

Quoted 5 lines
Documentation/trace/ftrace.txt +1 -1
@@ -389,18 +401,18 @@ trace_options
 The trace_options file is used to control what gets printed in
 the trace output. To see what is available, simply cat the file:
 
-  cat /debug/tracing/trace_options
+  cat ./trace_options

Please remove the "./", it is also distracting. We state that we "cd" to the location. "cat trace_options" is good enough.

Same for the all the others below.

Also, since I'm the most picky about this, you may want to remove the ftrace files from the patch and move it to another patch. That way the owners of the other files can ack them and get those changes in. When we have something for ftrace, I'll give my ack. But lets not hold up the other changes because of me ;-)

-- Steve

Quoted 465 lines
Documentation/trace/ftrace.txt +11 -11
   print-parent nosym-offset nosym-addr noverbose noraw nohex nobin \
   noblock nostacktrace nosched-tree nouserstacktrace nosym-userobj
 
 To disable one of the options, echo in the option prepended with
 "no".
 
-  echo noprint-parent > /debug/tracing/trace_options
+  echo noprint-parent > ./trace_options
 
 To enable an option, leave off the "no".
 
-  echo sym-offset > /debug/tracing/trace_options
+  echo sym-offset > ./trace_options
 
 Here are the available options:
 
@@ -476,11 +488,11 @@ sched_switch
 This tracer simply records schedule switches. Here is an example
 of how to use it.
 
- # echo sched_switch > /debug/tracing/current_tracer
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo sched_switch > ./current_tracer
+ # echo 1 > ./tracing_enabled
  # sleep 1
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/trace
+ # echo 0 > ./tracing_enabled
+ # cat ./trace
 
 # tracer: sched_switch
 #
@@ -574,13 +586,13 @@ new trace is saved.
 To reset the maximum, echo 0 into tracing_max_latency. Here is
 an example:
 
- # echo irqsoff > /debug/tracing/current_tracer
- # echo 0 > /debug/tracing/tracing_max_latency
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo irqsoff > ./current_tracer
+ # echo 0 > ./tracing_max_latency
+ # echo 1 > ./tracing_enabled
  # ls -ltr
[...]
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/latency_trace
+ # echo 0 > ./tracing_enabled
+ # cat ./latency_trace
 # tracer: irqsoff
 #
 irqsoff latency trace v1.1.5 on 2.6.26
@@ -681,13 +693,13 @@ Like the irqsoff tracer, it records the maximum

latency for
which preemption was disabled. The control of preemptoff tracer is much like the irqsoff tracer.

- # echo preemptoff > /debug/tracing/current_tracer
- # echo 0 > /debug/tracing/tracing_max_latency
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo preemptoff > ./current_tracer
+ # echo 0 > ./tracing_max_latency
+ # echo 1 > ./tracing_enabled

  # ls -ltr
  [...]
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/latency_trace
+ # echo 0 > ./tracing_enabled
+ # cat ./latency_trace
 # tracer: preemptoff
 #
 preemptoff latency trace v1.1.5 on 2.6.26-rc8
Documentation/trace/ftrace.txt +5 -5
@@ -828,13 +840,13 @@ tracer.
 Again, using this trace is much like the irqsoff and preemptoff
 tracers.
 
- # echo preemptirqsoff > /debug/tracing/current_tracer
- # echo 0 > /debug/tracing/tracing_max_latency
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo preemptirqsoff > ./current_tracer
+ # echo 0 > ./tracing_max_latency
+ # echo 1 > ./tracing_enabled
  # ls -ltr
[...]
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/latency_trace
+ # echo 0 > ./tracing_enabled
+ # cat ./latency_trace
 # tracer: preemptirqsoff
 #
 preemptirqsoff latency trace v1.1.5 on 2.6.26-rc8
@@ -990,12 +1002,12 @@ slightly differently than we did with the

previous tracers.
Instead of performing an 'ls', we will run 'sleep 1' under 'chrt' which changes the priority of the task.

- # echo wakeup > /debug/tracing/current_tracer
- # echo 0 > /debug/tracing/tracing_max_latency
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo wakeup > ./current_tracer
+ # echo 0 > ./tracing_max_latency
+ # echo 1 > ./tracing_enabled

  # chrt -f 5 sleep 1
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/latency_trace
+ # echo 0 > ./tracing_enabled
+ # cat ./latency_trace
 # tracer: wakeup
 #
 wakeup latency trace v1.1.5 on 2.6.26-rc8
Documentation/trace/ftrace.txt +12 -15
@@ -1105,11 +1117,11 @@ can be done from the debug file system. Make sure the
 ftrace_enabled is set; otherwise this tracer is a nop.
 
  # sysctl kernel.ftrace_enabled=1
- # echo function > /debug/tracing/current_tracer
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo function > ./current_tracer
+ # echo 1 > ./tracing_enabled
  # usleep 1
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/trace
+ # echo 0 > ./tracing_enabled
+ # cat ./trace
 # tracer: function
 #
 #           TASK-PID   CPU#    TIMESTAMP  FUNCTION
@@ -1146,7 +1158,7 @@ int trace_fd;
[...]
 int main(int argc, char *argv[]) {
[...]
-	trace_fd = open("/debug/tracing/tracing_enabled", O_WRONLY);
+	trace_fd = open("/sys/kernel/debug/tracing/tracing_enabled", O_WRONLY);
[...]
 	if (condition_hit()) {
 		write(trace_fd, "0", 1);
@@ -1154,26 +1166,23 @@ int main(int argc, char *argv[]) {
[...]
 }
 
-Note: Here we hard coded the path name. The debugfs mount is not
-guaranteed to be at /debug (and is more commonly at
-/sys/kernel/debug). For simple one time traces, the above is
-sufficent. For anything else, a search through /proc/mounts may
+Note: Here we hard coded the path name. A search through /proc/mounts may
 be needed to find where the debugfs file-system is mounted.
 
 
 Single thread tracing
 ---------------------
 
-By writing into /debug/tracing/set_ftrace_pid you can trace a
+By writing into ./set_ftrace_pid you can trace a
 single thread. For example:
 
-# cat /debug/tracing/set_ftrace_pid
+# cat ./set_ftrace_pid
 no pid
-# echo 3111 > /debug/tracing/set_ftrace_pid
-# cat /debug/tracing/set_ftrace_pid
+# echo 3111 > ./set_ftrace_pid
+# cat ./set_ftrace_pid
 3111
-# echo function > /debug/tracing/current_tracer
-# cat /debug/tracing/trace | head
+# echo function > ./current_tracer
+# cat ./trace | head
  # tracer: function
  #
  #           TASK-PID    CPU#    TIMESTAMP  FUNCTION
@@ -1184,8 +1193,8 @@ no pid
      yum-updatesd-3111  [003]  1637.254683: lock_hrtimer_base

<-hrtimer_try_to_cancel

      yum-updatesd-3111  [003]  1637.254685: fget_light <-do_sys_poll
      yum-updatesd-3111  [003]  1637.254686: pipe_poll <-do_sys_poll
Documentation/trace/ftrace.txt +4 -4
-# echo -1 > /debug/tracing/set_ftrace_pid
-# cat /debug/tracing/trace |head
+# echo -1 > ./set_ftrace_pid
+# cat ./trace |head
  # tracer: function
  #
  #           TASK-PID    CPU#    TIMESTAMP  FUNCTION
@@ -1217,12 +1226,12 @@ int main (int argc, char **argv)
                 char line[64];
                 int s;
 
-                ffd = open("/debug/tracing/current_tracer", O_WRONLY);
+                ffd =
 open("/sys/kernel/debug/tracing/current_tracer", O_WRONLY);
                 if (ffd < 0)
                         exit(-1);
                 write(ffd, "nop", 3);
 
-                fd = open("/debug/tracing/set_ftrace_pid", O_WRONLY);
+                fd = open("/sys/kernel/debug/tracing/set_ftrace_pid",
 O_WRONLY);
                 s = sprintf(line, "%d\n", getpid());
                 write(fd, line, s);
Documentation/trace/ftrace.txt +16 -16
@@ -1374,22 +1383,22 @@ want, depending on your needs.
   tracing_cpu_mask file) or you might sometimes see unordered
   function calls while cpu tracing switch.
 
-	hide: echo nofuncgraph-cpu > /debug/tracing/trace_options
-	show: echo funcgraph-cpu > /debug/tracing/trace_options
+	hide: echo nofuncgraph-cpu > ./trace_options
+	show: echo funcgraph-cpu > ./trace_options
 
 - The duration (function's time of execution) is displayed on
   the closing bracket line of a function or on the same line
   than the current function in case of a leaf one. It is default
   enabled.
 
-	hide: echo nofuncgraph-duration > /debug/tracing/trace_options
-	show: echo funcgraph-duration > /debug/tracing/trace_options
+	hide: echo nofuncgraph-duration > ./trace_options
+	show: echo funcgraph-duration > ./trace_options
 
 - The overhead field precedes the duration field in case of
   reached duration thresholds.
 
-	hide: echo nofuncgraph-overhead > /debug/tracing/trace_options
-	show: echo funcgraph-overhead > /debug/tracing/trace_options
+	hide: echo nofuncgraph-overhead > ./trace_options
+	show: echo funcgraph-overhead > ./trace_options
 	depends on: funcgraph-duration
 
   ie:
@@ -1418,8 +1427,8 @@ want, depending on your needs.
 - The task/pid field displays the thread cmdline and pid which
   executed the function. It is default disabled.
 
-	hide: echo nofuncgraph-proc > /debug/tracing/trace_options
-	show: echo funcgraph-proc > /debug/tracing/trace_options
+	hide: echo nofuncgraph-proc > ./trace_options
+	show: echo funcgraph-proc > ./trace_options
 
   ie:
 
@@ -1442,8 +1451,8 @@ want, depending on your needs.
   system clock since it started. A snapshot of this time is
   given on each entry/exit of functions
 
-	hide: echo nofuncgraph-abstime > /debug/tracing/trace_options
-	show: echo funcgraph-abstime > /debug/tracing/trace_options
+	hide: echo nofuncgraph-abstime > ./trace_options
+	show: echo funcgraph-abstime > ./trace_options
 
   ie:
 
@@ -1540,7 +1549,7 @@ listed in:
 
    available_filter_functions
 
- # cat /debug/tracing/available_filter_functions
+ # cat ./available_filter_functions
 put_prev_task_idle
 kmem_cache_create
 pick_next_task_rt
@@ -1552,12 +1561,12 @@ mutex_lock
 If I am only interested in sys_nanosleep and hrtimer_interrupt:
 
  # echo sys_nanosleep hrtimer_interrupt \
-		> /debug/tracing/set_ftrace_filter
- # echo ftrace > /debug/tracing/current_tracer
- # echo 1 > /debug/tracing/tracing_enabled
+		> ./set_ftrace_filter
+ # echo ftrace > ./current_tracer
+ # echo 1 > ./tracing_enabled
  # usleep 1
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/trace
+ # echo 0 > ./tracing_enabled
+ # cat ./trace
 # tracer: ftrace
 #
 #           TASK-PID   CPU#    TIMESTAMP  FUNCTION
@@ -1568,7 +1577,7 @@ If I am only interested in sys_nanosleep and

hrtimer_interrupt:

To see which functions are being traced, you can cat the file:

- # cat /debug/tracing/set_ftrace_filter
+ # cat ./set_ftrace_filter
hrtimer_interrupt
sys_nanosleep

Documentation/trace/ftrace.txt +14 -14
@@ -1588,7 +1597,7 @@ Note: It is better to use quotes to enclose the
 wild cards,
       otherwise the shell may expand the parameters into names
       of files in the local directory.
 
- # echo 'hrtimer_*' > /debug/tracing/set_ftrace_filter
+ # echo 'hrtimer_*' > ./set_ftrace_filter
 
 Produces:
@@ -1609,7 +1618,7 @@ Produces:
 
 Notice that we lost the sys_nanosleep.
 
- # cat /debug/tracing/set_ftrace_filter
+ # cat ./set_ftrace_filter
 hrtimer_run_queues
 hrtimer_run_pending
 hrtimer_init
@@ -1635,17 +1644,17 @@ To append to the filters, use '>>'
 To clear out a filter so that all functions will be recorded
 again:
 
- # echo > /debug/tracing/set_ftrace_filter
- # cat /debug/tracing/set_ftrace_filter
+ # echo > ./set_ftrace_filter
+ # cat ./set_ftrace_filter
  #
 
 Again, now we want to append.
 
- # echo sys_nanosleep > /debug/tracing/set_ftrace_filter
- # cat /debug/tracing/set_ftrace_filter
+ # echo sys_nanosleep > ./set_ftrace_filter
+ # cat ./set_ftrace_filter
 sys_nanosleep
- # echo 'hrtimer_*' >> /debug/tracing/set_ftrace_filter
- # cat /debug/tracing/set_ftrace_filter
+ # echo 'hrtimer_*' >> ./set_ftrace_filter
+ # cat ./set_ftrace_filter
 hrtimer_run_queues
 hrtimer_run_pending
 hrtimer_init
@@ -1668,7 +1677,7 @@ hrtimer_init_sleeper
 The set_ftrace_notrace prevents those functions from being
 traced.
 
- # echo '*preempt*' '*lock*' > /debug/tracing/set_ftrace_notrace
+ # echo '*preempt*' '*lock*' > ./set_ftrace_notrace
 
 Produces:
 
@@ -1758,13 +1767,13 @@ the effect on the tracing is different. Every read from
 trace_pipe is consumed. This means that subsequent reads will be
 different. The trace is live.
 
- # echo function > /debug/tracing/current_tracer
- # cat /debug/tracing/trace_pipe > /tmp/trace.out &
+ # echo function > ./current_tracer
+ # cat ./trace_pipe > /tmp/trace.out &
 [1] 4153
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo 1 > ./tracing_enabled
  # usleep 1
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/trace
+ # echo 0 > ./tracing_enabled
+ # cat ./trace
 # tracer: function
 #
 #           TASK-PID   CPU#    TIMESTAMP  FUNCTION
@@ -1800,7 +1809,7 @@ number listed is the number of entries that can

be recorded per
CPU. To know the full size, multiply the number of possible CPUS with the number of entries.

- # cat /debug/tracing/buffer_size_kb
+ # cat ./buffer_size_kb
1408 (units kilobytes)

Note, to modify this, you must have tracing completely disabled.

Documentation/trace/ftrace.txt +5 -5
@@ -1808,21 +1817,21 @@ To do that, echo "nop" into the current_tracer. If the
 current_tracer is not set to "nop", an EINVAL error will be
 returned.
 
- # echo nop > /debug/tracing/current_tracer
- # echo 10000 > /debug/tracing/buffer_size_kb
- # cat /debug/tracing/buffer_size_kb
+ # echo nop > ./current_tracer
+ # echo 10000 > ./buffer_size_kb
+ # cat ./buffer_size_kb
 10000 (units kilobytes)
 
 The number of pages which will be allocated is limited to a
 percentage of available memory. Allocating too much will produce
 an error.
 
- # echo 1000000000000 > /debug/tracing/buffer_size_kb
+ # echo 1000000000000 > ./buffer_size_kb
 -bash: echo: write error: Cannot allocate memory
- # cat /debug/tracing/buffer_size_kb
+ # cat ./buffer_size_kb
 85
 
 ---
 
Quoted 58 lines · Steven Rostedt
Quoted 21 lines · GeunSik Lim

Dear Greg and Steen,

Thanks your advices and opinions about a confusion according to
too many mount directory names for debugfs. I made rfc v2 patch file again like belows. Can I hear your advices about below patch contents?.

* ref:
http://groups.google.com/group/linux.kernel/browse_thread/thread/c34e386587979340/dc52317f888e78b6?#dc52317f888e78b6

Regards,
GeunSik Lim.

Subject: [PATCH]   Fix terminology inconsistency of dir name to mount debugfs filesystem.

      Many developers use "/debug/" or "/debugfs/ directory name to mount debugfs filesystem for ftrace according to ./Documentation/tracers/ftrace.txt file.

Quoted 8 lines

      And, three directory names(ex:/debug/, /debugfs/, /sys/kernel/debug/) is existed in kernel source like ftrace, DRM, Wireless, Network[sky2]files to mount debugfs filesystem.

      debugfs means debug filesystem for debugging easy to use by greg kroah hartman. "/sys/kernel/debug/" name is suitable as directory name of debugfs filesystem.
      -debugfs related reference: http://lwn.net/Articles/115405/

Note, LWN has an update this week:

 http://lwn.net/Articles/334546/

But it currently is subscriber only. Others would need to either buy a subscription (plug) or wait a week for it to become open.

Still for some aspects, just having "debugfs" without the leading "/" might just be self explanatory. We can assume the user already knows where debugfs is mounted.

-- Steve

Quoted 8 lines

      Fix inconsistency of directory name to mount debugfs filesystem.

Signature · 10 lines
Regards,
GeunSik Lim ( SAMSUNG ELECTRONICS)
Blog : http://blog.naver.com/invain/
e-Mail: [email protected]
           [email protected] , [email protected]
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/
#22 lore raw
Quoted 87 lines · Steven Rostedt
Quoted 46 lines · GeunSik Lim

Thanks your advice. I agree with your reply about this patch. I updated contents that you advised like belows.
If I mistake, Please correct me.

And, I also need your review  about below 2 contents in the ftrace.txt file.

Thanks,
GeunSik Lim.

1)

(quoted diff) +0 -0
@@ -1146,7 +1158,7 @@ int trace_fd;
[...]

 int main(int argc, char *argv[]) {
      [...]
-     trace_fd = open("/debug/tracing/tracing_enabled", O_WRONLY);
+     trace_fd = open("/sys/kernel/debug/tracing/tracing_enabled", O_WRONLY); [...]
      if (condition_hit()) {
              write(trace_fd, "0", 1);

(quoted diff) +0 -0
@@ -1154,26 +1166,23 @@ int main(int argc, char *argv[]) {
[...]

 }

(quoted diff) +2 -2
@@ -1217,12 +1226,12 @@ int main (int argc, char **argv)
                  char line[64];
                  int s;
 
-                ffd = open("/debug/tracing/current_tracer", O_WRONLY);
+                ffd =
 open("/sys/kernel/debug/tracing/current_tracer", O_WRONLY);
                  if (ffd < 0)
                          exit(-1);
                  write(ffd, "nop", 3);
 
-                fd = open("/debug/tracing/set_ftrace_pid", O_WRONLY);
+                fd = open("/sys/kernel/debug/tracing/set_ftrace_pid",
 O_WRONLY);
                  s = sprintf(line, "%d\n", getpid());

                 write(fd, line, s);

(quoted diff) +1 -4
-Note: Here we hard coded the path name. The debugfs mount is not
-guaranteed to be at /debug (and is more commonly at
-/sys/kernel/debug). For simple one time traces, the above is
-sufficent. For anything else, a search through /proc/mounts may
+Note: Here we hard coded the path name. A search through /proc/mounts may
Quoted 1 line

 be needed to find where the debugfs file-system is mounted.

Here's the function I use:

#define _STR(x) x
#define STR(x) _STR(x)
#define MAX_PATH 256

const char *find_debugfs(void)
{
       static char debugfs[MAX_PATH+1];
       static int debugfs_found;
       char type[100];
       FILE *fp;

       if (debugfs_found)
               return debugfs;

       if ((fp = fopen("/proc/mounts","r")) == NULL) { perror("/proc/mounts");
               return NULL;
       }

       while (fscanf(fp, "%*s %"
                     STR(MAX_PATH)
                     "s %99s %*s %*d %*d\n",
                     debugfs, type) == 2) {
               if (strcmp(type, "debugfs") == 0)
                       break;
       }
       fclose(fp);

       if (strcmp(type, "debugfs") != 0) { fprintf(stderr, "debugfs not mounted"); return NULL;
       }

Quoted 7 lines

       debugfs_found = 1;

       return debugfs;
}

This might be a good idea to show it. This is the most robust since it does not hard code the debugfs directory.

Yes, I think so. I have one opinion in the above find_debugfs() function. Default directory is "/sys/kernel/debug/tracing" in ftrace.txt file as you mentioned.
I will append below code for "***/tracing/" directory name.

         int size;
         [...]
         size = MAX_PATH - strlen(debugfs);
         strncat(debugfs, "/tracing/", size);
         return debugfs;
Quoted 59 lines
Quoted 59 lines

2)
More details can be found in the source code, in the

(quoted diff) +1 -1
-kernel/tracing/*.c files.
+kernel/trace/*.c files

* about ftrace.txt file
---

Documentation/trace/ftrace.txt +0 -0
index fd9a3e6..790199b 100644
@@ -7,7 +7,6 @@ Copyright 2008 Red Hat Inc.
                 (dual licensed under the GPL v2)
  Reviewers:   Elias Oltmanns, Randy Dunlap, Andrew Morton,

           John Kacur, and David Teigland.
-
 Written for: 2.6.28-rc2

 Introduction

Documentation/trace/ftrace.txt +0 -0
@@ -33,13 +32,26 @@ The File System

 Ftrace uses the debugfs file system to hold the control files as well as the files to display output.

Documentation/trace/ftrace.txt +18 -5
-To mount the debugfs system:
+When debugfs is configured into the kernel (which selecting any ftrace
+option will do) the directory /sys/kernel/debug will be created. To mount
+this directory, you can add to your /etc/fstab file:
+
+ debugfs       /sys/kernel/debug          debugfs defaults        0       0
+
+Or you can mount it at run time with:
+
+ mount -t debugfs nodev /sys/kernel/debug
+
+For quicker access to that directory you may want to make a soft link to
+it:
 
-  # mkdir /debug
-  # mount -t debugfs nodev /debug
+ ln -s /sys/kernel/debug /debug
 
-( Note: it is more common to mount at /sys/kernel/debug, but for
-  simplicity this document will use /debug)
+Any selected ftrace option will also create a directory called tracing
+within the debugfs. The rest of the document will assume that you are in
+the ftrace directory (cd /sys/kernel/debug/tracing) and will only concentrate
+on the files within that directory and not distract from the content with
+the extended "/sys/kernel/debug/tracing" path name.

 That's it! (assuming that you have ftrace configured into your kernel)

Documentation/trace/ftrace.txt +0 -0
@@ -389,18 +401,18 @@ trace_options

 The trace_options file is used to control what gets printed in the trace output. To see what is available, simply cat the file:

Quoted 5 lines
Quoted 2 lines

-  cat /debug/tracing/trace_options
+  cat ./trace_options

Please remove the "./", it is also distracting. We state that we "cd" to the location. "cat trace_options" is good enough.

It's right. This is just my private style. I will remove "./".

Quoted 5 lines

Same for the all the others below.

Also, since I'm the most picky about this, you may want to remove the ftrace files from the patch and move it to another patch. That way the

I am not think that you are the most picky about this. I like your advice and review about my proposal for consensus.

Quoted 4 lines

owners of the other files can ack them and get those changes in. When we have something for ftrace, I'll give my ack. But lets not hold up the other changes because of me ;-)

Don't care. Hope it helps. ;-)
I will modify ftrace files from the patch according to your advice.

Quoted 467 lines

-- Steve

Quoted 465 lines

   print-parent nosym-offset nosym-addr noverbose noraw nohex nobin \
   noblock nostacktrace nosched-tree nouserstacktrace nosym-userobj

 To disable one of the options, echo in the option prepended with "no".

-  echo noprint-parent > /debug/tracing/trace_options
+  echo noprint-parent > ./trace_options

 To enable an option, leave off the "no".

-  echo sym-offset > /debug/tracing/trace_options
+  echo sym-offset > ./trace_options

 Here are the available options:

Documentation/trace/ftrace.txt +0 -0
@@ -476,11 +488,11 @@ sched_switch

 This tracer simply records schedule switches. Here is an example of how to use it.

- # echo sched_switch > /debug/tracing/current_tracer
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo sched_switch > ./current_tracer
+ # echo 1 > ./tracing_enabled

  # sleep 1
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/trace
+ # echo 0 > ./tracing_enabled
+ # cat ./trace

 # tracer: sched_switch
 #

Documentation/trace/ftrace.txt +0 -0
@@ -574,13 +586,13 @@ new trace is saved.

 To reset the maximum, echo 0 into tracing_max_latency. Here is an example:

- # echo irqsoff > /debug/tracing/current_tracer
- # echo 0 > /debug/tracing/tracing_max_latency
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo irqsoff > ./current_tracer
+ # echo 0 > ./tracing_max_latency
+ # echo 1 > ./tracing_enabled

  # ls -ltr
  [...]
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/latency_trace
+ # echo 0 > ./tracing_enabled
+ # cat ./latency_trace
 # tracer: irqsoff
 #
 irqsoff latency trace v1.1.5 on 2.6.26
Documentation/trace/ftrace.txt +0 -0
@@ -681,13 +693,13 @@ Like the irqsoff tracer, it records the maximum

latency for
 which preemption was disabled. The control of preemptoff tracer is much like the irqsoff tracer.

- # echo preemptoff > /debug/tracing/current_tracer
- # echo 0 > /debug/tracing/tracing_max_latency
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo preemptoff > ./current_tracer
+ # echo 0 > ./tracing_max_latency
+ # echo 1 > ./tracing_enabled

  # ls -ltr
  [...]
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/latency_trace
+ # echo 0 > ./tracing_enabled
+ # cat ./latency_trace
 # tracer: preemptoff
 #
 preemptoff latency trace v1.1.5 on 2.6.26-rc8
Documentation/trace/ftrace.txt +0 -0
@@ -828,13 +840,13 @@ tracer.

 Again, using this trace is much like the irqsoff and preemptoff tracers.

- # echo preemptirqsoff > /debug/tracing/current_tracer
- # echo 0 > /debug/tracing/tracing_max_latency
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo preemptirqsoff > ./current_tracer
+ # echo 0 > ./tracing_max_latency
+ # echo 1 > ./tracing_enabled

  # ls -ltr
  [...]
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/latency_trace
+ # echo 0 > ./tracing_enabled
+ # cat ./latency_trace
 # tracer: preemptirqsoff
 #
 preemptirqsoff latency trace v1.1.5 on 2.6.26-rc8
Documentation/trace/ftrace.txt +0 -0
@@ -990,12 +1002,12 @@ slightly differently than we did with the

previous tracers.
 Instead of performing an 'ls', we will run 'sleep 1' under 'chrt' which changes the priority of the task.

- # echo wakeup > /debug/tracing/current_tracer
- # echo 0 > /debug/tracing/tracing_max_latency
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo wakeup > ./current_tracer
+ # echo 0 > ./tracing_max_latency
+ # echo 1 > ./tracing_enabled

  # chrt -f 5 sleep 1
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/latency_trace
+ # echo 0 > ./tracing_enabled
+ # cat ./latency_trace
 # tracer: wakeup
 #
 wakeup latency trace v1.1.5 on 2.6.26-rc8
Documentation/trace/ftrace.txt +0 -0
@@ -1105,11 +1117,11 @@ can be done from the debug file system. Make sure the

 ftrace_enabled is set; otherwise this tracer is a nop.

  # sysctl kernel.ftrace_enabled=1
- # echo function > /debug/tracing/current_tracer
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo function > ./current_tracer
+ # echo 1 > ./tracing_enabled
  # usleep 1
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/trace
+ # echo 0 > ./tracing_enabled
+ # cat ./trace
 # tracer: function
 #
 #           TASK-PID   CPU#    TIMESTAMP  FUNCTION
Documentation/trace/ftrace.txt +0 -0
@@ -1146,7 +1158,7 @@ int trace_fd;
[...]

 int main(int argc, char *argv[]) {
      [...]
-     trace_fd = open("/debug/tracing/tracing_enabled", O_WRONLY);
+     trace_fd = open("/sys/kernel/debug/tracing/tracing_enabled", O_WRONLY); [...]
      if (condition_hit()) {
              write(trace_fd, "0", 1);

Documentation/trace/ftrace.txt +0 -0
@@ -1154,26 +1166,23 @@ int main(int argc, char *argv[]) {
[...]

 }

Documentation/trace/ftrace.txt +1 -4
-Note: Here we hard coded the path name. The debugfs mount is not
-guaranteed to be at /debug (and is more commonly at
-/sys/kernel/debug). For simple one time traces, the above is
-sufficent. For anything else, a search through /proc/mounts may
+Note: Here we hard coded the path name. A search through /proc/mounts may

 be needed to find where the debugfs file-system is mounted.

 Single thread tracing
 ---------------------

Documentation/trace/ftrace.txt +1 -1
-By writing into /debug/tracing/set_ftrace_pid you can trace a
+By writing into ./set_ftrace_pid you can trace a

 single thread. For example:

Documentation/trace/ftrace.txt +1 -1
-# cat /debug/tracing/set_ftrace_pid
+# cat ./set_ftrace_pid

 no pid

Documentation/trace/ftrace.txt +2 -2
-# echo 3111 > /debug/tracing/set_ftrace_pid
-# cat /debug/tracing/set_ftrace_pid
+# echo 3111 > ./set_ftrace_pid
+# cat ./set_ftrace_pid

 3111

Documentation/trace/ftrace.txt +2 -2
-# echo function > /debug/tracing/current_tracer
-# cat /debug/tracing/trace | head
+# echo function > ./current_tracer
+# cat ./trace | head

  # tracer: function
  #
  #           TASK-PID    CPU#    TIMESTAMP  FUNCTION

Documentation/trace/ftrace.txt +0 -0
@@ -1184,8 +1193,8 @@ no pid

      yum-updatesd-3111  [003]  1637.254683: lock_hrtimer_base <-hrtimer_try_to_cancel
      yum-updatesd-3111  [003]  1637.254685: fget_light <-do_sys_poll yum-updatesd-3111  [003]  1637.254686: pipe_poll <-do_sys_poll

Documentation/trace/ftrace.txt +2 -2
-# echo -1 > /debug/tracing/set_ftrace_pid
-# cat /debug/tracing/trace |head
+# echo -1 > ./set_ftrace_pid
+# cat ./trace |head

  # tracer: function
  #
  #           TASK-PID    CPU#    TIMESTAMP  FUNCTION

Documentation/trace/ftrace.txt +2 -2
@@ -1217,12 +1226,12 @@ int main (int argc, char **argv)
                  char line[64];
                  int s;
 
-                ffd = open("/debug/tracing/current_tracer", O_WRONLY);
+                ffd =
 open("/sys/kernel/debug/tracing/current_tracer", O_WRONLY);
                  if (ffd < 0)
                          exit(-1);
                  write(ffd, "nop", 3);
 
-                fd = open("/debug/tracing/set_ftrace_pid", O_WRONLY);
+                fd = open("/sys/kernel/debug/tracing/set_ftrace_pid",
 O_WRONLY);
                  s = sprintf(line, "%d\n", getpid());

                 write(fd, line, s);

Documentation/trace/ftrace.txt +0 -0
@@ -1374,22 +1383,22 @@ want, depending on your needs.

   tracing_cpu_mask file) or you might sometimes see unordered function calls while cpu tracing switch.

-     hide: echo nofuncgraph-cpu > /debug/tracing/trace_options
-     show: echo funcgraph-cpu > /debug/tracing/trace_options
+     hide: echo nofuncgraph-cpu > ./trace_options
+     show: echo funcgraph-cpu > ./trace_options

 - The duration (function's time of execution) is displayed on the closing bracket line of a function or on the same line than the current function in case of a leaf one. It is default enabled.

-     hide: echo nofuncgraph-duration > /debug/tracing/trace_options
-     show: echo funcgraph-duration > /debug/tracing/trace_options
+     hide: echo nofuncgraph-duration > ./trace_options
+     show: echo funcgraph-duration > ./trace_options

 - The overhead field precedes the duration field in case of reached duration thresholds.

-     hide: echo nofuncgraph-overhead > /debug/tracing/trace_options
-     show: echo funcgraph-overhead > /debug/tracing/trace_options
+     hide: echo nofuncgraph-overhead > ./trace_options
+     show: echo funcgraph-overhead > ./trace_options depends on: funcgraph-duration

   ie:

Documentation/trace/ftrace.txt +0 -0
@@ -1418,8 +1427,8 @@ want, depending on your needs.

 - The task/pid field displays the thread cmdline and pid which executed the function. It is default disabled.

-     hide: echo nofuncgraph-proc > /debug/tracing/trace_options
-     show: echo funcgraph-proc > /debug/tracing/trace_options
+     hide: echo nofuncgraph-proc > ./trace_options
+     show: echo funcgraph-proc > ./trace_options

   ie:

Documentation/trace/ftrace.txt +0 -0
@@ -1442,8 +1451,8 @@ want, depending on your needs.

   system clock since it started. A snapshot of this time is given on each entry/exit of functions

-     hide: echo nofuncgraph-abstime > /debug/tracing/trace_options
-     show: echo funcgraph-abstime > /debug/tracing/trace_options
+     hide: echo nofuncgraph-abstime > ./trace_options
+     show: echo funcgraph-abstime > ./trace_options

   ie:

Documentation/trace/ftrace.txt +0 -0
@@ -1540,7 +1549,7 @@ listed in:
 

    available_filter_functions

- # cat /debug/tracing/available_filter_functions
+ # cat ./available_filter_functions
 put_prev_task_idle
 kmem_cache_create
 pick_next_task_rt

Documentation/trace/ftrace.txt +0 -0
@@ -1552,12 +1561,12 @@ mutex_lock

 If I am only interested in sys_nanosleep and hrtimer_interrupt:

  # echo sys_nanosleep hrtimer_interrupt \
-             > /debug/tracing/set_ftrace_filter
- # echo ftrace > /debug/tracing/current_tracer
- # echo 1 > /debug/tracing/tracing_enabled
+             > ./set_ftrace_filter
+ # echo ftrace > ./current_tracer
+ # echo 1 > ./tracing_enabled
  # usleep 1
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/trace
+ # echo 0 > ./tracing_enabled
+ # cat ./trace
 # tracer: ftrace
 #
 #           TASK-PID   CPU#    TIMESTAMP  FUNCTION
Documentation/trace/ftrace.txt +0 -0
@@ -1568,7 +1577,7 @@ If I am only interested in sys_nanosleep and

hrtimer_interrupt:

 To see which functions are being traced, you can cat the file:

- # cat /debug/tracing/set_ftrace_filter
+ # cat ./set_ftrace_filter
 hrtimer_interrupt
 sys_nanosleep

Documentation/trace/ftrace.txt +0 -0
@@ -1588,7 +1597,7 @@ Note: It is better to use quotes to enclose the
 wild cards,

       otherwise the shell may expand the parameters into names of files in the local directory.

- # echo 'hrtimer_*' > /debug/tracing/set_ftrace_filter
+ # echo 'hrtimer_*' > ./set_ftrace_filter

 Produces:

Documentation/trace/ftrace.txt +0 -0
@@ -1609,7 +1618,7 @@ Produces:
 

 Notice that we lost the sys_nanosleep.

- # cat /debug/tracing/set_ftrace_filter
+ # cat ./set_ftrace_filter
 hrtimer_run_queues
 hrtimer_run_pending
 hrtimer_init

Documentation/trace/ftrace.txt +0 -0
@@ -1635,17 +1644,17 @@ To append to the filters, use '>>'

 To clear out a filter so that all functions will be recorded again:

- # echo > /debug/tracing/set_ftrace_filter
- # cat /debug/tracing/set_ftrace_filter
+ # echo > ./set_ftrace_filter
+ # cat ./set_ftrace_filter
  #

 Again, now we want to append.

- # echo sys_nanosleep > /debug/tracing/set_ftrace_filter
- # cat /debug/tracing/set_ftrace_filter
+ # echo sys_nanosleep > ./set_ftrace_filter
+ # cat ./set_ftrace_filter
 sys_nanosleep
- # echo 'hrtimer_*' >> /debug/tracing/set_ftrace_filter
- # cat /debug/tracing/set_ftrace_filter
+ # echo 'hrtimer_*' >> ./set_ftrace_filter
+ # cat ./set_ftrace_filter
 hrtimer_run_queues
 hrtimer_run_pending
 hrtimer_init

Documentation/trace/ftrace.txt +0 -0
@@ -1668,7 +1677,7 @@ hrtimer_init_sleeper

 The set_ftrace_notrace prevents those functions from being traced.

- # echo '*preempt*' '*lock*' > /debug/tracing/set_ftrace_notrace
+ # echo '*preempt*' '*lock*' > ./set_ftrace_notrace

 Produces:

Documentation/trace/ftrace.txt +0 -0
@@ -1758,13 +1767,13 @@ the effect on the tracing is different. Every read from

 trace_pipe is consumed. This means that subsequent reads will be different. The trace is live.

- # echo function > /debug/tracing/current_tracer
- # cat /debug/tracing/trace_pipe > /tmp/trace.out &
+ # echo function > ./current_tracer
+ # cat ./trace_pipe > /tmp/trace.out &
 [1] 4153
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo 1 > ./tracing_enabled

  # usleep 1
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/trace
+ # echo 0 > ./tracing_enabled
+ # cat ./trace
 # tracer: function
 #
 #           TASK-PID   CPU#    TIMESTAMP  FUNCTION
Documentation/trace/ftrace.txt +0 -0
@@ -1800,7 +1809,7 @@ number listed is the number of entries that can

be recorded per
 CPU. To know the full size, multiply the number of possible CPUS with the number of entries.

- # cat /debug/tracing/buffer_size_kb
+ # cat ./buffer_size_kb
 1408 (units kilobytes)

 Note, to modify this, you must have tracing completely disabled.

Documentation/trace/ftrace.txt +0 -0
@@ -1808,21 +1817,21 @@ To do that, echo "nop" into the current_tracer. If the

 current_tracer is not set to "nop", an EINVAL error will be returned.

- # echo nop > /debug/tracing/current_tracer
- # echo 10000 > /debug/tracing/buffer_size_kb
- # cat /debug/tracing/buffer_size_kb
+ # echo nop > ./current_tracer
+ # echo 10000 > ./buffer_size_kb
+ # cat ./buffer_size_kb
 10000 (units kilobytes)

 The number of pages which will be allocated is limited to a percentage of available memory. Allocating too much will produce an error.

- # echo 1000000000000 > /debug/tracing/buffer_size_kb
+ # echo 1000000000000 > ./buffer_size_kb -bash: echo: write error: Cannot allocate memory
- # cat /debug/tracing/buffer_size_kb
+ # cat ./buffer_size_kb
 85

 ---

Quoted 39 lines · Steven Rostedt
Quoted 21 lines · GeunSik Lim

Dear Greg and Steen,

Thanks your advices and opinions about a confusion according to
too many mount directory names for debugfs. I made rfc v2 patch file again like belows. Can I hear your advices about below patch contents?.

* ref:
http://groups.google.com/group/linux.kernel/browse_thread/thread/c34e386587979340/dc52317f888e78b6?#dc52317f888e78b6

Regards,
GeunSik Lim.

Subject: [PATCH]   Fix terminology inconsistency of dir name to mount debugfs filesystem.

      Many developers use "/debug/" or "/debugfs/ directory name to mount debugfs filesystem for ftrace according to ./Documentation/tracers/ftrace.txt file.

Quoted 8 lines

      And, three directory names(ex:/debug/, /debugfs/, /sys/kernel/debug/) is existed in kernel source like ftrace, DRM, Wireless, Network[sky2]files to mount debugfs filesystem.

      debugfs means debug filesystem for debugging easy to use by greg kroah hartman. "/sys/kernel/debug/" name is suitable as directory name of debugfs filesystem.
      -debugfs related reference: http://lwn.net/Articles/115405/

Note, LWN has an update this week:

 http://lwn.net/Articles/334546/

Quoted 19 lines

But it currently is subscriber only. Others would need to either buy a subscription (plug) or wait a week for it to become open.

Still for some aspects, just having "debugfs" without the leading "/" might just be self explanatory. We can assume the user already knows where debugfs is mounted.

-- Steve

Quoted 8 lines

      Fix inconsistency of directory name to mount debugfs filesystem.

--
Regards,
GeunSik Lim ( SAMSUNG ELECTRONICS)
Blog : http://blog.naver.com/invain/
e-Mail: [email protected]
           [email protected] , [email protected]
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Signature · 10 lines
Regards,
GeunSik Lim ( SAMSUNG ELECTRONICS)
Blog : http://blog.naver.com/invain/
e-Mail: [email protected]
           [email protected] , [email protected]
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/
#23 lore raw
Quoted 10 lines · GeunSik Lim
Quoted 2 lines

Also, since I'm the most picky about this, you may want to remove the ftrace files from the patch and move it to another patch. That way the

I am not think that you are the most picky about this. I like your advice and review about my proposal for consensus.

Quoted 4 lines

owners of the other files can ack them and get those changes in. When we have something for ftrace, I'll give my ack. But lets not hold up the other changes because of me ;-)

Don't care. Hope it helps. ;-)
I will modify  ftrace files from the patch according to your advice.

I updated ftrace related contents according to Steven's help like belows. If I misunderstood wrongly, Please correct me.
Thanks.

* About ftrace.txt file/
---

Documentation/trace/ftrace.txt +30 -18
index fd9a3e6..caaf5d9 100644
@@ -7,7 +7,6 @@ Copyright 2008 Red Hat Inc.
                (dual licensed under the GPL v2)
 Reviewers:   Elias Oltmanns, Randy Dunlap, Andrew Morton,
 	     John Kacur, and David Teigland.
-
 Written for: 2.6.28-rc2
 
 Introduction
@@ -33,13 +32,26 @@ The File System
 Ftrace uses the debugfs file system to hold the control files as
 well as the files to display output.
 
-To mount the debugfs system:
+When debugfs is configured into the kernel (which selecting any ftrace
+option will do) the directory /sys/kernel/debug will be created. To mount
+this directory, you can add to your /etc/fstab file:
+
+ debugfs       /sys/kernel/debug          debugfs defaults        0       0
+
+Or you can mount it at run time with:
+
+ mount -t debugfs nodev /sys/kernel/debug
+
+For quicker access to that directory you may want to make a soft link to
+it:
 
-  # mkdir /debug
-  # mount -t debugfs nodev /debug
+ ln -s /sys/kernel/debug /debug
 
-( Note: it is more common to mount at /sys/kernel/debug, but for
-  simplicity this document will use /debug)
+Any selected ftrace option will also create a directory called tracing
+within the debugfs. The rest of the document will assume that you are in
+the ftrace directory (cd /sys/kernel/debug/tracing) and will only concentrate
+on the files within that directory and not distract from the content with
+the extended "/sys/kernel/debug/tracing" path name.
 
 That's it! (assuming that you have ftrace configured into your kernel)
 
@@ -389,18 +401,18 @@ trace_options
 The trace_options file is used to control what gets printed in
 the trace output. To see what is available, simply cat the file:
 
-  cat /debug/tracing/trace_options
+  cat trace_options
   print-parent nosym-offset nosym-addr noverbose noraw nohex nobin \
   noblock nostacktrace nosched-tree nouserstacktrace nosym-userobj
 
 To disable one of the options, echo in the option prepended with
 "no".
 
-  echo noprint-parent > /debug/tracing/trace_options
+  echo noprint-parent > trace_options
 
 To enable an option, leave off the "no".
 
-  echo sym-offset > /debug/tracing/trace_options
+  echo sym-offset > trace_options
 
 Here are the available options:
 
@@ -476,11 +488,11 @@ sched_switch
 This tracer simply records schedule switches. Here is an example
 of how to use it.
 
- # echo sched_switch > /debug/tracing/current_tracer
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo sched_switch > current_tracer
+ # echo 1 > tracing_enabled
  # sleep 1
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/trace
+ # echo 0 > tracing_enabled
+ # cat trace
 
 # tracer: sched_switch
 #
@@ -574,13 +586,13 @@ new trace is saved.
 To reset the maximum, echo 0 into tracing_max_latency. Here is
 an example:
 
- # echo irqsoff > /debug/tracing/current_tracer
- # echo 0 > /debug/tracing/tracing_max_latency
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo irqsoff > current_tracer
+ # echo 0 > tracing_max_latency
+ # echo 1 > tracing_enabled
  # ls -ltr
  [...]
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/latency_trace
+ # echo 0 > tracing_enabled
+ # cat latency_trace
 # tracer: irqsoff
 #
 irqsoff latency trace v1.1.5 on 2.6.26
@@ -681,13 +693,13 @@ Like the irqsoff tracer, it records the maximum

latency for
which preemption was disabled. The control of preemptoff tracer is much like the irqsoff tracer.

- # echo preemptoff > /debug/tracing/current_tracer
- # echo 0 > /debug/tracing/tracing_max_latency
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo preemptoff > current_tracer
+ # echo 0 > tracing_max_latency
+ # echo 1 > tracing_enabled

  # ls -ltr
  [...]
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/latency_trace
+ # echo 0 > tracing_enabled
+ # cat latency_trace
 # tracer: preemptoff
 #
 preemptoff latency trace v1.1.5 on 2.6.26-rc8
@@ -828,13 +840,13 @@ tracer.
 Again, using this trace is much like the irqsoff and preemptoff
 tracers.

- # echo preemptirqsoff > /debug/tracing/current_tracer
- # echo 0 > /debug/tracing/tracing_max_latency
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo preemptirqsoff > current_tracer
+ # echo 0 > tracing_max_latency
+ # echo 1 > tracing_enabled

  # ls -ltr
  [...]
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/latency_trace
+ # echo 0 > tracing_enabled
+ # cat latency_trace
 # tracer: preemptirqsoff
 #
 preemptirqsoff latency trace v1.1.5 on 2.6.26-rc8
@@ -990,12 +1002,12 @@ slightly differently than we did with the
previous tracers.
 Instead of performing an 'ls', we will run 'sleep 1' under
 'chrt' which changes the priority of the task.

- # echo wakeup > /debug/tracing/current_tracer
- # echo 0 > /debug/tracing/tracing_max_latency
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo wakeup > current_tracer
+ # echo 0 > tracing_max_latency
+ # echo 1 > tracing_enabled

  # chrt -f 5 sleep 1
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/latency_trace
+ # echo 0 > tracing_enabled
+ # cat latency_trace
 # tracer: wakeup
 #
 wakeup latency trace v1.1.5 on 2.6.26-rc8
@@ -1105,11 +1117,11 @@ can be done from the debug file system. Make sure the
 ftrace_enabled is set; otherwise this tracer is a nop.
  # sysctl kernel.ftrace_enabled=1
- # echo function > /debug/tracing/current_tracer
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo function > current_tracer
+ # echo 1 > tracing_enabled
  # usleep 1
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/trace
+ # echo 0 > tracing_enabled
+ # cat trace
 # tracer: function
 #
 #           TASK-PID   CPU#    TIMESTAMP  FUNCTION
@@ -1146,7 +1158,7 @@ int trace_fd;
 [...]
 int main(int argc, char *argv[]) {
 	[...]
-	trace_fd = open("/debug/tracing/tracing_enabled", O_WRONLY);
+	trace_fd = open(find_debugfs("tracing_enabled"), O_WRONLY);
 	[...]
 	if (condition_hit()) {
 		write(trace_fd, "0", 1);
@@ -1154,26 +1166,20 @@ int main(int argc, char *argv[]) {
 	[...]
 }

-Note: Here we hard coded the path name. The debugfs mount is not -guaranteed to be at /debug (and is more commonly at -/sys/kernel/debug). For simple one time traces, the above is -sufficent. For anything else, a search through /proc/mounts may -be needed to find where the debugfs file-system is mounted. -

Single thread tracing
---------------------

-By writing into /debug/tracing/set_ftrace_pid you can trace a +By writing into set_ftrace_pid you can trace a
single thread. For example:

-# cat /debug/tracing/set_ftrace_pid
+# cat set_ftrace_pid
no pid
-# echo 3111 > /debug/tracing/set_ftrace_pid
-# cat /debug/tracing/set_ftrace_pid
+# echo 3111 > set_ftrace_pid
+# cat set_ftrace_pid
3111
-# echo function > /debug/tracing/current_tracer
-# cat /debug/tracing/trace | head
+# echo function > current_tracer
+# cat trace | head
# tracer: function
#
# TASK-PID CPU# TIMESTAMP FUNCTION
@@ -1184,8 +1190,8 @@ no pid

      yum-updatesd-3111  [003]  1637.254683: lock_hrtimer_base
<-hrtimer_try_to_cancel
      yum-updatesd-3111  [003]  1637.254685: fget_light <-do_sys_poll
      yum-updatesd-3111  [003]  1637.254686: pipe_poll <-do_sys_poll

-# echo -1 > /debug/tracing/set_ftrace_pid
-# cat /debug/tracing/trace |head
+# echo -1 > set_ftrace_pid
+# cat trace |head
# tracer: function
#
# TASK-PID CPU# TIMESTAMP FUNCTION
@@ -1202,11 +1208,57 @@ something like this simple program:

#include <stdio.h>
#include <stdlib.h>
+#include <string.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <unistd.h>

+#define _STR(x) #x
+#define STR(x) _STR(x)
+#define MAX_PATH 256
+
+const char *find_debugfs(char *file_name)
+{

+       static char debugfs[MAX_PATH+1];
+       static int debugfs_found;
+       char type[100];
+       FILE *fp;
+       int size;

+
+ if (debugfs_found)
+ return debugfs;
+
+ if ((fp = fopen("/proc/mounts","r")) == NULL) {
+ perror("/proc/mounts");
+ return NULL;
+ }
+
+ while (fscanf(fp, "%*s %"

+                     STR(MAX_PATH)
+                     "s %99s %*s %*d %*d\n",
+                     debugfs, type) == 2) {

+ if (strcmp(type, "debugfs") == 0)
+ break;
+ }
+ fclose(fp);
+
+ if (strcmp(type, "debugfs") != 0) {
+ fprintf(stderr, "debugfs not mounted");
+ return NULL;
+ }
+
+ debugfs_found = 1;
+
+ size = MAX_PATH - strlen(debugfs);
+ strncat(debugfs, "/tracing/", size);
+
+ size = MAX_PATH - strlen(debugfs);
+ strncat(debugfs, file_name, size);
+
+ return debugfs;
+}
+
int main (int argc, char **argv)

 {
         if (argc < 1)
@@ -1217,12 +1269,12 @@ int main (int argc, char **argv)
                 char line[64];
                 int s;

- ffd = open("/debug/tracing/current_tracer", O_WRONLY);

+                ffd = open(find_debugfs("current_tracer"), O_WRONLY);
                 if (ffd < 0)
                         exit(-1);
                 write(ffd, "nop", 3);

- fd = open("/debug/tracing/set_ftrace_pid", O_WRONLY);

+                fd = open(find_debugfs("set_ftrace_pid"), O_WRONLY);
                 s = sprintf(line, "%d\n", getpid());
                 write(fd, line, s);

@@ -1374,22 +1426,22 @@ want, depending on your needs.
tracing_cpu_mask file) or you might sometimes see unordered function calls while cpu tracing switch.

- hide: echo nofuncgraph-cpu > /debug/tracing/trace_options
- show: echo funcgraph-cpu > /debug/tracing/trace_options
+ hide: echo nofuncgraph-cpu > trace_options
+ show: echo funcgraph-cpu > trace_options

- The duration (function's time of execution) is displayed on the closing bracket line of a function or on the same line than the current function in case of a leaf one. It is default enabled.

- hide: echo nofuncgraph-duration > /debug/tracing/trace_options
- show: echo funcgraph-duration > /debug/tracing/trace_options
+ hide: echo nofuncgraph-duration > trace_options
+ show: echo funcgraph-duration > trace_options

- The overhead field precedes the duration field in case of reached duration thresholds.

- hide: echo nofuncgraph-overhead > /debug/tracing/trace_options
- show: echo funcgraph-overhead > /debug/tracing/trace_options
+ hide: echo nofuncgraph-overhead > trace_options
+ show: echo funcgraph-overhead > trace_options
depends on: funcgraph-duration

ie:
@@ -1418,8 +1470,8 @@ want, depending on your needs.
- The task/pid field displays the thread cmdline and pid which executed the function. It is default disabled.

- hide: echo nofuncgraph-proc > /debug/tracing/trace_options
- show: echo funcgraph-proc > /debug/tracing/trace_options
+ hide: echo nofuncgraph-proc > trace_options
+ show: echo funcgraph-proc > trace_options

ie:

@@ -1442,8 +1494,8 @@ want, depending on your needs.
system clock since it started. A snapshot of this time is given on each entry/exit of functions

- hide: echo nofuncgraph-abstime > /debug/tracing/trace_options
- show: echo funcgraph-abstime > /debug/tracing/trace_options
+ hide: echo nofuncgraph-abstime > trace_options
+ show: echo funcgraph-abstime > trace_options

ie:

@@ -1540,7 +1592,7 @@ listed in:

available_filter_functions

- # cat /debug/tracing/available_filter_functions
+ # cat available_filter_functions
put_prev_task_idle
kmem_cache_create
pick_next_task_rt
@@ -1552,12 +1604,12 @@ mutex_lock
If I am only interested in sys_nanosleep and hrtimer_interrupt:

  # echo sys_nanosleep hrtimer_interrupt \
-		> /debug/tracing/set_ftrace_filter
- # echo ftrace > /debug/tracing/current_tracer
- # echo 1 > /debug/tracing/tracing_enabled
+		> set_ftrace_filter
+ # echo ftrace > current_tracer
+ # echo 1 > tracing_enabled
  # usleep 1
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/trace
+ # echo 0 > tracing_enabled
+ # cat trace
 # tracer: ftrace
 #
 #           TASK-PID   CPU#    TIMESTAMP  FUNCTION
@@ -1568,7 +1620,7 @@ If I am only interested in sys_nanosleep and
hrtimer_interrupt:

To see which functions are being traced, you can cat the file:

- # cat /debug/tracing/set_ftrace_filter
+ # cat set_ftrace_filter
hrtimer_interrupt
sys_nanosleep

@@ -1588,7 +1640,7 @@ Note: It is better to use quotes to enclose the wild cards,
otherwise the shell may expand the parameters into names of files in the local directory.

- # echo 'hrtimer_*' > /debug/tracing/set_ftrace_filter
+ # echo 'hrtimer_*' > set_ftrace_filter

Produces:

@@ -1609,7 +1661,7 @@ Produces:

Notice that we lost the sys_nanosleep.

- # cat /debug/tracing/set_ftrace_filter
+ # cat set_ftrace_filter
hrtimer_run_queues
hrtimer_run_pending
hrtimer_init
@@ -1635,17 +1687,17 @@ To append to the filters, use '>>'
To clear out a filter so that all functions will be recorded again:

- # echo > /debug/tracing/set_ftrace_filter
- # cat /debug/tracing/set_ftrace_filter
+ # echo > set_ftrace_filter
+ # cat set_ftrace_filter
#

Again, now we want to append.

- # echo sys_nanosleep > /debug/tracing/set_ftrace_filter
- # cat /debug/tracing/set_ftrace_filter
+ # echo sys_nanosleep > set_ftrace_filter
+ # cat set_ftrace_filter
sys_nanosleep
- # echo 'hrtimer_*' >> /debug/tracing/set_ftrace_filter
- # cat /debug/tracing/set_ftrace_filter
+ # echo 'hrtimer_*' >> set_ftrace_filter
+ # cat set_ftrace_filter
hrtimer_run_queues
hrtimer_run_pending
hrtimer_init
@@ -1668,7 +1720,7 @@ hrtimer_init_sleeper
The set_ftrace_notrace prevents those functions from being traced.

- # echo '*preempt*' '*lock*' > /debug/tracing/set_ftrace_notrace
+ # echo '*preempt*' '*lock*' > set_ftrace_notrace

Produces:

@@ -1758,13 +1810,13 @@ the effect on the tracing is different. Every read from
trace_pipe is consumed. This means that subsequent reads will be different. The trace is live.

- # echo function > /debug/tracing/current_tracer
- # cat /debug/tracing/trace_pipe > /tmp/trace.out &
+ # echo function > current_tracer
+ # cat trace_pipe > /tmp/trace.out &
[1] 4153
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo 1 > tracing_enabled

  # usleep 1
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/trace
+ # echo 0 > tracing_enabled
+ # cat trace
 # tracer: function
 #
 #           TASK-PID   CPU#    TIMESTAMP  FUNCTION
@@ -1800,7 +1852,7 @@ number listed is the number of entries that can
be recorded per
 CPU. To know the full size, multiply the number of possible CPUS
 with the number of entries.

- # cat /debug/tracing/buffer_size_kb
+ # cat buffer_size_kb
1408 (units kilobytes)

Note, to modify this, you must have tracing completely disabled.
@@ -1808,18 +1860,18 @@ To do that, echo "nop" into the current_tracer. If the
current_tracer is not set to "nop", an EINVAL error will be returned.

- # echo nop > /debug/tracing/current_tracer
- # echo 10000 > /debug/tracing/buffer_size_kb
- # cat /debug/tracing/buffer_size_kb
+ # echo nop > current_tracer
+ # echo 10000 > buffer_size_kb
+ # cat buffer_size_kb
10000 (units kilobytes)

The number of pages which will be allocated is limited to a percentage of available memory. Allocating too much will produce an error.

- # echo 1000000000000 > /debug/tracing/buffer_size_kb
+ # echo 1000000000000 > buffer_size_kb -bash: echo: write error: Cannot allocate memory
- # cat /debug/tracing/buffer_size_kb
+ # cat buffer_size_kb
85

---

Regards,
GeunSik Lim ( SAMSUNG ELECTRONICS)
Blog : http://blog.naver.com/invain/
e-Mail: [email protected]
[email protected] , [email protected]

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

maintainer
#24 lore raw
Quoted 41 lines · GeunSik Lim
(quoted diff) +32 -0
@@ -1202,11 +1208,57 @@ something like this simple program:
 
 #include <stdio.h>
 #include <stdlib.h>
+#include <string.h>
 #include <sys/types.h>
 #include <sys/stat.h>
 #include <fcntl.h>
 #include <unistd.h>
 
+#define _STR(x) #x
+#define STR(x) _STR(x)
+#define MAX_PATH 256
+
+const char *find_debugfs(char *file_name)
+{
+       static char debugfs[MAX_PATH+1];
+       static int debugfs_found;
+       char type[100];
+       FILE *fp;
+       int size;
+
+       if (debugfs_found)
+               return debugfs;
+
+       if ((fp = fopen("/proc/mounts","r")) == NULL) {
+               perror("/proc/mounts");
+               return NULL;
+       }
+
+       while (fscanf(fp, "%*s %"
+                     STR(MAX_PATH)
+                     "s %99s %*s %*d %*d\n",
+                     debugfs, type) == 2) {
+               if (strcmp(type, "debugfs") == 0)
+                       break;
+       }
+       fclose(fp);
+
+       if (strcmp(type, "debugfs") != 0) {
+               fprintf(stderr, "debugfs not mounted");

Review something like this simple program: · line 1257

Quoted 10 lines
(quoted diff) +10 -0
+               return NULL;
+       }
+
+       debugfs_found = 1;
+
+       size = MAX_PATH - strlen(debugfs);
+       strncat(debugfs, "/tracing/", size);
+
+       size = MAX_PATH - strlen(debugfs);
+       strncat(debugfs, file_name, size);

You need to make two functions. One to find the debugfs and one to add the file. The "debugfs_found" will make the function return just the first file_name every time. Something like this: (remove the file_name from find_debugfs).

const char *tracing_file(const char *file_name)

{
	static char trace_file[MAX_PATH+1];

	snprintf(trace_file, MAX_PATH, "%s/%s", find_debugfs(), file_name);

	return trace_file;
}

The above is not thread safe, but is OK for non threaded programs.

-- Steve

Quoted 22 lines
(quoted diff) +6 -2
+
+       return debugfs;
+}
+
 int main (int argc, char **argv)
 {
         if (argc < 1)
@@ -1217,12 +1269,12 @@ int main (int argc, char **argv)
                 char line[64];
                 int s;
 
-                ffd = open("/debug/tracing/current_tracer", O_WRONLY);
+                ffd = open(find_debugfs("current_tracer"), O_WRONLY);
                 if (ffd < 0)
                         exit(-1);
                 write(ffd, "nop", 3);
 
-                fd = open("/debug/tracing/set_ftrace_pid", O_WRONLY);
+                fd = open(find_debugfs("set_ftrace_pid"), O_WRONLY);
                 s = sprintf(line, "%d\n", getpid());
                 write(fd, line, s);
 
#25 lore raw
Quoted 9 lines · Steven Rostedt

You need to make two functions. One to find the debugfs and one to add the file. The "debugfs_found" will make the function return just the first file_name every time. Something like this: (remove the file_name from find_debugfs).

const char *tracing_file(const char *file_name)
{
       static char trace_file[MAX_PATH+1];

Quoted 6 lines

       snprintf(trace_file, MAX_PATH, "%s/%s", find_debugfs(), file_name);

       return trace_file;
}

The above is not thread safe, but is OK for non threaded programs.

Thanks.
I modified pseudo code using tracing_file function that you sent.

* ftrace.txt info after modification.
---

Documentation/trace/ftrace.txt +30 -18
index fd9a3e6..05f8378 100644
@@ -7,7 +7,6 @@ Copyright 2008 Red Hat Inc.
                (dual licensed under the GPL v2)
 Reviewers:   Elias Oltmanns, Randy Dunlap, Andrew Morton,
 	     John Kacur, and David Teigland.
-
 Written for: 2.6.28-rc2
 
 Introduction
@@ -33,13 +32,26 @@ The File System
 Ftrace uses the debugfs file system to hold the control files as
 well as the files to display output.
 
-To mount the debugfs system:
+When debugfs is configured into the kernel (which selecting any ftrace
+option will do) the directory /sys/kernel/debug will be created. To mount
+this directory, you can add to your /etc/fstab file:
+
+ debugfs       /sys/kernel/debug          debugfs defaults        0       0
+
+Or you can mount it at run time with:
+
+ mount -t debugfs nodev /sys/kernel/debug
 
-  # mkdir /debug
-  # mount -t debugfs nodev /debug
+For quicker access to that directory you may want to make a soft link to
+it:
 
-( Note: it is more common to mount at /sys/kernel/debug, but for
-  simplicity this document will use /debug)
+ ln -s /sys/kernel/debug /debug
+
+Any selected ftrace option will also create a directory called tracing
+within the debugfs. The rest of the document will assume that you are in
+the ftrace directory (cd /sys/kernel/debug/tracing) and will only concentrate
+on the files within that directory and not distract from the content with
+the extended "/sys/kernel/debug/tracing" path name.
 
 That's it! (assuming that you have ftrace configured into your kernel)
 
@@ -389,18 +401,18 @@ trace_options
 The trace_options file is used to control what gets printed in
 the trace output. To see what is available, simply cat the file:
 
-  cat /debug/tracing/trace_options
+  cat trace_options
   print-parent nosym-offset nosym-addr noverbose noraw nohex nobin \
   noblock nostacktrace nosched-tree nouserstacktrace nosym-userobj
 
 To disable one of the options, echo in the option prepended with
 "no".
 
-  echo noprint-parent > /debug/tracing/trace_options
+  echo noprint-parent > trace_options
 
 To enable an option, leave off the "no".
 
-  echo sym-offset > /debug/tracing/trace_options
+  echo sym-offset > trace_options
 
 Here are the available options:
 
@@ -476,11 +488,11 @@ sched_switch
 This tracer simply records schedule switches. Here is an example
 of how to use it.
 
- # echo sched_switch > /debug/tracing/current_tracer
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo sched_switch > current_tracer
+ # echo 1 > tracing_enabled
  # sleep 1
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/trace
+ # echo 0 > tracing_enabled
+ # cat trace
 
 # tracer: sched_switch
 #
@@ -574,13 +586,13 @@ new trace is saved.
 To reset the maximum, echo 0 into tracing_max_latency. Here is
 an example:
 
- # echo irqsoff > /debug/tracing/current_tracer
- # echo 0 > /debug/tracing/tracing_max_latency
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo irqsoff > current_tracer
+ # echo 0 > tracing_max_latency
+ # echo 1 > tracing_enabled
  # ls -ltr
  [...]
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/latency_trace
+ # echo 0 > tracing_enabled
+ # cat latency_trace
 # tracer: irqsoff
 #
 irqsoff latency trace v1.1.5 on 2.6.26
@@ -681,13 +693,13 @@ Like the irqsoff tracer, it records the maximum

latency for
which preemption was disabled. The control of preemptoff tracer is much like the irqsoff tracer.

- # echo preemptoff > /debug/tracing/current_tracer
- # echo 0 > /debug/tracing/tracing_max_latency
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo preemptoff > current_tracer
+ # echo 0 > tracing_max_latency
+ # echo 1 > tracing_enabled

  # ls -ltr
  [...]
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/latency_trace
+ # echo 0 > tracing_enabled
+ # cat latency_trace
 # tracer: preemptoff
 #
 preemptoff latency trace v1.1.5 on 2.6.26-rc8
@@ -828,13 +840,13 @@ tracer.
 Again, using this trace is much like the irqsoff and preemptoff
 tracers.

- # echo preemptirqsoff > /debug/tracing/current_tracer
- # echo 0 > /debug/tracing/tracing_max_latency
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo preemptirqsoff > current_tracer
+ # echo 0 > tracing_max_latency
+ # echo 1 > tracing_enabled

  # ls -ltr
  [...]
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/latency_trace
+ # echo 0 > tracing_enabled
+ # cat latency_trace
 # tracer: preemptirqsoff
 #
 preemptirqsoff latency trace v1.1.5 on 2.6.26-rc8
@@ -990,12 +1002,12 @@ slightly differently than we did with the
previous tracers.
 Instead of performing an 'ls', we will run 'sleep 1' under
 'chrt' which changes the priority of the task.

- # echo wakeup > /debug/tracing/current_tracer
- # echo 0 > /debug/tracing/tracing_max_latency
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo wakeup > current_tracer
+ # echo 0 > tracing_max_latency
+ # echo 1 > tracing_enabled

  # chrt -f 5 sleep 1
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/latency_trace
+ # echo 0 > tracing_enabled
+ # cat latency_trace
 # tracer: wakeup
 #
 wakeup latency trace v1.1.5 on 2.6.26-rc8
@@ -1105,11 +1117,11 @@ can be done from the debug file system. Make sure the
 ftrace_enabled is set; otherwise this tracer is a nop.
  # sysctl kernel.ftrace_enabled=1
- # echo function > /debug/tracing/current_tracer
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo function > current_tracer
+ # echo 1 > tracing_enabled
  # usleep 1
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/trace
+ # echo 0 > tracing_enabled
+ # cat trace
 # tracer: function
 #
 #           TASK-PID   CPU#    TIMESTAMP  FUNCTION
@@ -1146,7 +1158,7 @@ int trace_fd;
 [...]
 int main(int argc, char *argv[]) {
 	[...]
-	trace_fd = open("/debug/tracing/tracing_enabled", O_WRONLY);
+	trace_fd = open(tracing_file("tracing_enabled"), O_WRONLY);
 	[...]
 	if (condition_hit()) {
 		write(trace_fd, "0", 1);
@@ -1154,26 +1166,20 @@ int main(int argc, char *argv[]) {
 	[...]
 }

-Note: Here we hard coded the path name. The debugfs mount is not -guaranteed to be at /debug (and is more commonly at -/sys/kernel/debug). For simple one time traces, the above is -sufficent. For anything else, a search through /proc/mounts may -be needed to find where the debugfs file-system is mounted. -

Single thread tracing
---------------------

-By writing into /debug/tracing/set_ftrace_pid you can trace a +By writing into set_ftrace_pid you can trace a
single thread. For example:

-# cat /debug/tracing/set_ftrace_pid
+# cat set_ftrace_pid
no pid
-# echo 3111 > /debug/tracing/set_ftrace_pid
-# cat /debug/tracing/set_ftrace_pid
+# echo 3111 > set_ftrace_pid
+# cat set_ftrace_pid
3111
-# echo function > /debug/tracing/current_tracer
-# cat /debug/tracing/trace | head
+# echo function > current_tracer
+# cat trace | head
# tracer: function
#
# TASK-PID CPU# TIMESTAMP FUNCTION
@@ -1184,8 +1190,8 @@ no pid

      yum-updatesd-3111  [003]  1637.254683: lock_hrtimer_base
<-hrtimer_try_to_cancel
      yum-updatesd-3111  [003]  1637.254685: fget_light <-do_sys_poll
      yum-updatesd-3111  [003]  1637.254686: pipe_poll <-do_sys_poll

-# echo -1 > /debug/tracing/set_ftrace_pid
-# cat /debug/tracing/trace |head
+# echo -1 > set_ftrace_pid
+# cat trace |head
# tracer: function
#
# TASK-PID CPU# TIMESTAMP FUNCTION
@@ -1207,6 +1213,51 @@ something like this simple program:
#include <fcntl.h>
#include <unistd.h>

+#define _STR(x) #x
+#define STR(x) _STR(x)
+#define MAX_PATH 256
+
+const char *find_debugfs(void)
+{

+       static char debugfs[MAX_PATH+1];
+       static int debugfs_found;
+       char type[100];
+       FILE *fp;

+
+ if (debugfs_found)
+ return debugfs;
+
+ if ((fp = fopen("/proc/mounts","r")) == NULL) {
+ perror("/proc/mounts");
+ return NULL;
+ }
+
+ while (fscanf(fp, "%*s %"

+                     STR(MAX_PATH)
+                     "s %99s %*s %*d %*d\n",
+                     debugfs, type) == 2) {

+ if (strcmp(type, "debugfs") == 0)
+ break;
+ }
+ fclose(fp);
+
+ if (strcmp(type, "debugfs") != 0) {
+ fprintf(stderr, "debugfs not mounted");
+ return NULL;
+ }
+
+ debugfs_found = 1;
+
+ return debugfs;
+}
+
+const char *tracing_file(const char *file_name)
+{

+       static char trace_file[MAX_PATH+1];
+       snprintf(trace_file, MAX_PATH, "%s/%s", find_debugfs(), file_name);
+       return trace_file;
+}
+
 int main (int argc, char **argv)
 {
         if (argc < 1)
@@ -1217,12 +1268,12 @@ int main (int argc, char **argv)
                 char line[64];
                 int s;

- ffd = open("/debug/tracing/current_tracer", O_WRONLY);

+                ffd = open(tracing_file("current_tracer"), O_WRONLY);
                 if (ffd < 0)
                         exit(-1);
                 write(ffd, "nop", 3);

- fd = open("/debug/tracing/set_ftrace_pid", O_WRONLY);

+                fd = open(tracing_file("set_ftrace_pid"), O_WRONLY);
                 s = sprintf(line, "%d\n", getpid());
                 write(fd, line, s);

@@ -1374,22 +1425,22 @@ want, depending on your needs.
tracing_cpu_mask file) or you might sometimes see unordered function calls while cpu tracing switch.

- hide: echo nofuncgraph-cpu > /debug/tracing/trace_options
- show: echo funcgraph-cpu > /debug/tracing/trace_options
+ hide: echo nofuncgraph-cpu > trace_options
+ show: echo funcgraph-cpu > trace_options

- The duration (function's time of execution) is displayed on the closing bracket line of a function or on the same line than the current function in case of a leaf one. It is default enabled.

- hide: echo nofuncgraph-duration > /debug/tracing/trace_options
- show: echo funcgraph-duration > /debug/tracing/trace_options
+ hide: echo nofuncgraph-duration > trace_options
+ show: echo funcgraph-duration > trace_options

- The overhead field precedes the duration field in case of reached duration thresholds.

- hide: echo nofuncgraph-overhead > /debug/tracing/trace_options
- show: echo funcgraph-overhead > /debug/tracing/trace_options
+ hide: echo nofuncgraph-overhead > trace_options
+ show: echo funcgraph-overhead > trace_options
depends on: funcgraph-duration

ie:
@@ -1418,8 +1469,8 @@ want, depending on your needs.
- The task/pid field displays the thread cmdline and pid which executed the function. It is default disabled.

- hide: echo nofuncgraph-proc > /debug/tracing/trace_options
- show: echo funcgraph-proc > /debug/tracing/trace_options
+ hide: echo nofuncgraph-proc > trace_options
+ show: echo funcgraph-proc > trace_options

ie:

@@ -1442,8 +1493,8 @@ want, depending on your needs.
system clock since it started. A snapshot of this time is given on each entry/exit of functions

- hide: echo nofuncgraph-abstime > /debug/tracing/trace_options
- show: echo funcgraph-abstime > /debug/tracing/trace_options
+ hide: echo nofuncgraph-abstime > trace_options
+ show: echo funcgraph-abstime > trace_options

ie:

@@ -1540,7 +1591,7 @@ listed in:

available_filter_functions

- # cat /debug/tracing/available_filter_functions
+ # cat available_filter_functions
put_prev_task_idle
kmem_cache_create
pick_next_task_rt
@@ -1552,12 +1603,12 @@ mutex_lock
If I am only interested in sys_nanosleep and hrtimer_interrupt:

  # echo sys_nanosleep hrtimer_interrupt \
-		> /debug/tracing/set_ftrace_filter
- # echo ftrace > /debug/tracing/current_tracer
- # echo 1 > /debug/tracing/tracing_enabled
+		> set_ftrace_filter
+ # echo ftrace > current_tracer
+ # echo 1 > tracing_enabled
  # usleep 1
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/trace
+ # echo 0 > tracing_enabled
+ # cat trace
 # tracer: ftrace
 #
 #           TASK-PID   CPU#    TIMESTAMP  FUNCTION
@@ -1568,7 +1619,7 @@ If I am only interested in sys_nanosleep and
hrtimer_interrupt:

To see which functions are being traced, you can cat the file:

- # cat /debug/tracing/set_ftrace_filter
+ # cat set_ftrace_filter
hrtimer_interrupt
sys_nanosleep

@@ -1588,7 +1639,7 @@ Note: It is better to use quotes to enclose the wild cards,
otherwise the shell may expand the parameters into names of files in the local directory.

- # echo 'hrtimer_*' > /debug/tracing/set_ftrace_filter
+ # echo 'hrtimer_*' > set_ftrace_filter

Produces:

@@ -1609,7 +1660,7 @@ Produces:

Notice that we lost the sys_nanosleep.

- # cat /debug/tracing/set_ftrace_filter
+ # cat set_ftrace_filter
hrtimer_run_queues
hrtimer_run_pending
hrtimer_init
@@ -1635,17 +1686,17 @@ To append to the filters, use '>>'
To clear out a filter so that all functions will be recorded again:

- # echo > /debug/tracing/set_ftrace_filter
- # cat /debug/tracing/set_ftrace_filter
+ # echo > set_ftrace_filter
+ # cat set_ftrace_filter
#

Again, now we want to append.

- # echo sys_nanosleep > /debug/tracing/set_ftrace_filter
- # cat /debug/tracing/set_ftrace_filter
+ # echo sys_nanosleep > set_ftrace_filter
+ # cat set_ftrace_filter
sys_nanosleep
- # echo 'hrtimer_*' >> /debug/tracing/set_ftrace_filter
- # cat /debug/tracing/set_ftrace_filter
+ # echo 'hrtimer_*' >> set_ftrace_filter
+ # cat set_ftrace_filter
hrtimer_run_queues
hrtimer_run_pending
hrtimer_init
@@ -1668,7 +1719,7 @@ hrtimer_init_sleeper
The set_ftrace_notrace prevents those functions from being traced.

- # echo '*preempt*' '*lock*' > /debug/tracing/set_ftrace_notrace
+ # echo '*preempt*' '*lock*' > set_ftrace_notrace

Produces:

@@ -1758,13 +1809,13 @@ the effect on the tracing is different. Every read from
trace_pipe is consumed. This means that subsequent reads will be different. The trace is live.

- # echo function > /debug/tracing/current_tracer
- # cat /debug/tracing/trace_pipe > /tmp/trace.out &
+ # echo function > current_tracer
+ # cat trace_pipe > /tmp/trace.out &
[1] 4153
- # echo 1 > /debug/tracing/tracing_enabled
+ # echo 1 > tracing_enabled

  # usleep 1
- # echo 0 > /debug/tracing/tracing_enabled
- # cat /debug/tracing/trace
+ # echo 0 > tracing_enabled
+ # cat trace
 # tracer: function
 #
 #           TASK-PID   CPU#    TIMESTAMP  FUNCTION
@@ -1800,7 +1851,7 @@ number listed is the number of entries that can
be recorded per
 CPU. To know the full size, multiply the number of possible CPUS
 with the number of entries.

- # cat /debug/tracing/buffer_size_kb
+ # cat buffer_size_kb
1408 (units kilobytes)

Note, to modify this, you must have tracing completely disabled.
@@ -1808,18 +1859,18 @@ To do that, echo "nop" into the current_tracer. If the
current_tracer is not set to "nop", an EINVAL error will be returned.

- # echo nop > /debug/tracing/current_tracer
- # echo 10000 > /debug/tracing/buffer_size_kb
- # cat /debug/tracing/buffer_size_kb
+ # echo nop > current_tracer
+ # echo 10000 > buffer_size_kb
+ # cat buffer_size_kb
10000 (units kilobytes)

The number of pages which will be allocated is limited to a percentage of available memory. Allocating too much will produce an error.

- # echo 1000000000000 > /debug/tracing/buffer_size_kb
+ # echo 1000000000000 > buffer_size_kb -bash: echo: write error: Cannot allocate memory
- # cat /debug/tracing/buffer_size_kb
+ # cat buffer_size_kb
85

---

Signature · 10 lines
Regards,
GeunSik Lim ( SAMSUNG ELECTRONICS)
Blog : http://blog.naver.com/invain/
e-Mail: [email protected]
           [email protected] , [email protected]
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/
maintainer
#26 lore raw
In reply to GeunSik Lim
Quoted 48 lines · GeunSik Lim
Quoted 36 lines

Here's the function I use:

#define _STR(x) x
#define STR(x) _STR(x)
#define MAX_PATH 256

const char *find_debugfs(void)
{
       static char debugfs[MAX_PATH+1];
       static int debugfs_found;
       char type[100];
       FILE *fp;

       if (debugfs_found)
               return debugfs;

       if ((fp = fopen("/proc/mounts","r")) == NULL) { perror("/proc/mounts");
               return NULL;
       }

       while (fscanf(fp, "%*s %"
                     STR(MAX_PATH)
                     "s %99s %*s %*d %*d\n",
                     debugfs, type) == 2) {
               if (strcmp(type, "debugfs") == 0)
                       break;
       }
       fclose(fp);

       if (strcmp(type, "debugfs") != 0) { fprintf(stderr, "debugfs not mounted"); return NULL;
       }

Quoted 7 lines

       debugfs_found = 1;

       return debugfs;
}

This might be a good idea to show it. This is the most robust since it does not hard code the debugfs directory.

Yes, I think so. I have one opinion in the above find_debugfs() function. Default directory is "/sys/kernel/debug/tracing" in ftrace.txt file as you mentioned.
I will append below code for "***/tracing/" directory name.

Quoted 5 lines
         int size;
         [...]
         size = MAX_PATH - strlen(debugfs);
         strncat(debugfs, "/tracing/", size);
         return debugfs;

Heh, I usually have another function to add the /tracing part. But sure, we could do that too.

-- Steve

#27 lore raw

RFC PATCH v3 1/1 debugfs:Fix terminology inconsistency of dir name to mount debugfs filesystem.

In reply to GeunSik Lim

Dear Greg and Steven,

Thanks your advices and opinions about a confusion according to too many mount directory names for debugfs filesystem. I made rfc v3 like belows thanks to steven and greg. Can I hear your advices and comments about below patch contents?

* LKML about rfc v2:
http://lkml.org/lkml/2009/5/29/219

* LKML about rfc v1:
http://lkml.org/lkml/2009/5/26/17

Regards,
GeunSik Lim.

Subject: [PATCH] Fix terminology inconsistency of dir name to mount debugfs filesystem.

Many developers use "/debug/" or "/debugfs/ directory name to mount debugfs filesystem for ftrace according to ./Documentation/tracers/ftrace.txt file.

And, three directory names(ex:/debug/, /debugfs/, /sys/kernel/debug/) is existed in kernel source like ftrace, DRM, Wireless, Documentation, Network[sky2]files to mount debugfs filesystem.

debugfs means debug filesystem for debugging easy to use by greg kroah hartman. "/sys/kernel/debug/" name is suitable as directory name of debugfs filesystem.
-debugfs related reference: http://lwn.net/Articles/334546/

Fix inconsistency of directory name to mount debugfs filesystem.

CC Jiri Kosina <[email protected]>
CC David Airlie <[email protected]>
CC Peter Osterlund <[email protected]>
CC James Smart <[email protected]>
CC Ananth N Mavinakayanahalli <[email protected]>
CC Anil S Keshavamurthy <[email protected]>
CC Masami Hiramatsu <[email protected]>
CC linux-kernel <[email protected]>
---
 Documentation/DocBook/debugobjects.tmpl           |    2 +-
 Documentation/cdrom/packet-writing.txt            |    2 +-
 Documentation/fault-injection/fault-injection.txt |   70 +++---
 Documentation/kprobes.txt                         |    6 +-
 Documentation/trace/ftrace.txt                    |  233 +++++++++++++--------
 Documentation/trace/mmiotrace.txt                 |   26 ++--
 drivers/block/pktcdvd.c                           |    2 +-
 drivers/gpu/drm/drm_debugfs.c                     |   12 +-
 drivers/gpu/drm/drm_drv.c                         |    2 +-
 drivers/gpu/drm/drm_stub.c                        |    2 +-
 drivers/net/Kconfig                               |    4 +-
 drivers/net/wimax/i2400m/i2400m.h                 |    2 +-
 drivers/net/wireless/ath5k/Kconfig                |    6 +-
 drivers/net/wireless/libertas/README              |   12 +-
 drivers/scsi/lpfc/lpfc_debugfs.c                  |    4 +-
 include/linux/kernel.h                            |    2 +-
 include/linux/tracepoint.h                        |    4 +-
 kernel/trace/Kconfig                              |   10 +-
 kernel/trace/trace.c                              |   24 +-
 scripts/tracing/draw_functrace.py                 |    8 +-
 20 files changed, 242 insertions(+), 191 deletions(-)
Documentation/DocBook/debugobjects.tmpl +1 -1
index 7f5f218..08ff908 100644
@@ -106,7 +106,7 @@
       number of errors are printk'ed including a full stack trace.
     </para>
     <para>
-      The statistics are available via debugfs/debug_objects/stats.
+      The statistics are available via /sys/kernel/debug/debug_objects/stats.
       They provide information about the number of warnings and the
       number of successful fixups along with information about the
       usage of the internal tracking objects and the state of the
Documentation/cdrom/packet-writing.txt +1 -1
index cf1f812..1c40777 100644
@@ -117,7 +117,7 @@ Using the pktcdvd debugfs interface
 
 To read pktcdvd device infos in human readable form, do:
 
-	# cat /debug/pktcdvd/pktcdvd[0-7]/info
+	# cat /sys/kernel/debug/pktcdvd/pktcdvd[0-7]/info
 
 For a description of the debugfs interface look into the file:
 
Documentation/fault-injection/fault-injection.txt +35 -35

145 lines not on this page · raw message

Documentation/kprobes.txt +3 -3
index 1e7a769..053037a 100644
@@ -507,9 +507,9 @@ http://www.linuxsymposium.org/2006/linuxsymposium_procv2.pdf (pages 101-115)
 Appendix A: The kprobes debugfs interface
 
 With recent kernels (> 2.6.20) the list of registered kprobes is visible
-under the /debug/kprobes/ directory (assuming debugfs is mounted at /debug).
+under the /sys/kernel/debug/kprobes/ directory (assuming debugfs is mounted at //sys/kernel/debug).
 
-/debug/kprobes/list: Lists all registered probes on the system
+/sys/kernel/debug/kprobes/list: Lists all registered probes on the system
 
 c015d71a  k  vfs_read+0x0
 c011a316  j  do_fork+0x0
@@ -525,7 +525,7 @@ virtual addresses that correspond to modules that've been unloaded),
 such probes are marked with [GONE]. If the probe is temporarily disabled,
 such probes are marked with [DISABLED].
 
-/debug/kprobes/enabled: Turn kprobes ON/OFF forcibly.
+/sys/kernel/debug/kprobes/enabled: Turn kprobes ON/OFF forcibly.
 
 Provides a knob to globally and forcibly turn registered kprobes ON or OFF.
 By default, all kprobes are enabled. By echoing "0" to this file, all
Documentation/trace/ftrace.txt +142 -91

452 lines not on this page · raw message

Documentation/trace/mmiotrace.txt +13 -13
index 5731c67..162effb 100644
@@ -32,41 +32,41 @@ is no way to automatically detect if you are losing events due to CPUs racing.
 Usage Quick Reference
 ---------------------
 
-$ mount -t debugfs debugfs /debug
-$ echo mmiotrace > /debug/tracing/current_tracer
-$ cat /debug/tracing/trace_pipe > mydump.txt &
+$ mount -t debugfs debugfs /sys/kernel/debug
+$ echo mmiotrace > /sys/kernel/debug/tracing/current_tracer
+$ cat /sys/kernel/debug/tracing/trace_pipe > mydump.txt &
 Start X or whatever.
-$ echo "X is up" > /debug/tracing/trace_marker
-$ echo nop > /debug/tracing/current_tracer
+$ echo "X is up" > /sys/kernel/debug/tracing/trace_marker
+$ echo nop > /sys/kernel/debug/tracing/current_tracer
 Check for lost events.
 
 
 Usage
 -----
 
-Make sure debugfs is mounted to /debug. If not, (requires root privileges)
-$ mount -t debugfs debugfs /debug
+Make sure debugfs is mounted to /sys/kernel/debug. If not, (requires root privileges)
+$ mount -t debugfs debugfs /sys/kernel/debug
 
 Check that the driver you are about to trace is not loaded.
 
 Activate mmiotrace (requires root privileges):
-$ echo mmiotrace > /debug/tracing/current_tracer
+$ echo mmiotrace > /sys/kernel/debug/tracing/current_tracer
 
 Start storing the trace:
-$ cat /debug/tracing/trace_pipe > mydump.txt &
+$ cat /sys/kernel/debug/tracing/trace_pipe > mydump.txt &
 The 'cat' process should stay running (sleeping) in the background.
 
 Load the driver you want to trace and use it. Mmiotrace will only catch MMIO
 accesses to areas that are ioremapped while mmiotrace is active.
 
 During tracing you can place comments (markers) into the trace by
-$ echo "X is up" > /debug/tracing/trace_marker
+$ echo "X is up" > /sys/kernel/debug/tracing/trace_marker
 This makes it easier to see which part of the (huge) trace corresponds to
 which action. It is recommended to place descriptive markers about what you
 do.
 
 Shut down mmiotrace (requires root privileges):
-$ echo nop > /debug/tracing/current_tracer
+$ echo nop > /sys/kernel/debug/tracing/current_tracer
 The 'cat' process exits. If it does not, kill it by issuing 'fg' command and
 pressing ctrl+c.
 
@@ -78,10 +78,10 @@ to view your kernel log and look for "mmiotrace has lost events" warning. If
 events were lost, the trace is incomplete. You should enlarge the buffers and
 try again. Buffers are enlarged by first seeing how large the current buffers
 are:
-$ cat /debug/tracing/buffer_size_kb
+$ cat /sys/kernel/debug/tracing/buffer_size_kb
 gives you a number. Approximately double this number and write it back, for
 instance:
-$ echo 128000 > /debug/tracing/buffer_size_kb
+$ echo 128000 > /sys/kernel/debug/tracing/buffer_size_kb
 Then start again from the top.
 
 If you are doing a trace for a driver project, e.g. Nouveau, you should also
drivers/block/pktcdvd.c +1 -1
index dc7a8c3..9aa8932 100644
@@ -430,7 +430,7 @@ static void pkt_sysfs_cleanup(void)
 /********************************************************************
   entries in debugfs
 
-  /debugfs/pktcdvd[0-7]/
+  /sys/kernel/debug/pktcdvd[0-7]/
 			info
 
  *******************************************************************/
drivers/gpu/drm/drm_debugfs.c +6 -6
index c77c6c6..6ce0e26 100644
@@ -105,7 +105,7 @@ int drm_debugfs_create_files(struct drm_info_list *files, int count,
 		ent = debugfs_create_file(files[i].name, S_IFREG | S_IRUGO,
 					  root, tmp, &drm_debugfs_fops);
 		if (!ent) {
-			DRM_ERROR("Cannot create /debugfs/dri/%s/%s\n",
+			DRM_ERROR("Cannot create /sys/kernel/debug/dri/%s/%s\n",
 				  name, files[i].name);
 			drm_free(tmp, sizeof(struct drm_info_node),
 				 _DRM_DRIVER);
@@ -133,9 +133,9 @@ EXPORT_SYMBOL(drm_debugfs_create_files);
  * \param minor device minor number
  * \param root DRI debugfs dir entry.
  *
- * Create the DRI debugfs root entry "/debugfs/dri", the device debugfs root entry
- * "/debugfs/dri/%minor%/", and each entry in debugfs_list as
- * "/debugfs/dri/%minor%/%name%".
+ * Create the DRI debugfs root entry "/sys/kernel/debug/dri", the device debugfs root entry
+ * "/sys/kernel/debug/dri/%minor%/", and each entry in debugfs_list as
+ * "/sys/kernel/debug/dri/%minor%/%name%".
  */
 int drm_debugfs_init(struct drm_minor *minor, int minor_id,
 		     struct dentry *root)
@@ -148,7 +148,7 @@ int drm_debugfs_init(struct drm_minor *minor, int minor_id,
 	sprintf(name, "%d", minor_id);
 	minor->debugfs_root = debugfs_create_dir(name, root);
 	if (!minor->debugfs_root) {
-		DRM_ERROR("Cannot create /debugfs/dri/%s\n", name);
+		DRM_ERROR("Cannot create /sys/kernel/debug/dri/%s\n", name);
 		return -1;
 	}
 
@@ -165,7 +165,7 @@ int drm_debugfs_init(struct drm_minor *minor, int minor_id,
 		ret = dev->driver->debugfs_init(minor);
 		if (ret) {
 			DRM_ERROR("DRM: Driver failed to initialize "
-				  "/debugfs/dri.\n");
+				  "/sys/kernel/debug/dri.\n");
 			return ret;
 		}
 	}
drivers/gpu/drm/drm_drv.c +1 -1
index 019b7c5..1bf7efd 100644
@@ -339,7 +339,7 @@ static int __init drm_core_init(void)
 
 	drm_debugfs_root = debugfs_create_dir("dri", NULL);
 	if (!drm_debugfs_root) {
-		DRM_ERROR("Cannot create /debugfs/dri\n");
+		DRM_ERROR("Cannot create /sys/kernel/debug/dri\n");
 		ret = -1;
 		goto err_p3;
 	}
drivers/gpu/drm/drm_stub.c +1 -1
index b9631e3..b2a93c4 100644
@@ -328,7 +328,7 @@ static int drm_get_minor(struct drm_device *dev, struct drm_minor **minor, int t
 #if defined(CONFIG_DEBUG_FS)
 	ret = drm_debugfs_init(new_minor, minor_id, drm_debugfs_root);
 	if (ret) {
-		DRM_ERROR("DRM: Failed to initialize /debugfs/dri.\n");
+		DRM_ERROR("DRM: Failed to initialize /sys/kernel/debug/dri.\n");
 		goto err_g2;
 	}
 #endif
drivers/net/Kconfig +2 -2
index 214a92d..a170cea 100644
@@ -2200,7 +2200,7 @@ config SKGE_DEBUG
        depends on SKGE && DEBUG_FS
        help
 	 This option adds the ability to dump driver state for debugging.
-	 The file debugfs/skge/ethX displays the state of the internal
+	 The file /sys/kernel/debug/skge/ethX displays the state of the internal
 	 transmit and receive rings.
 
 	 If unsure, say N.
@@ -2226,7 +2226,7 @@ config SKY2_DEBUG
        depends on SKY2 && DEBUG_FS
        help
 	 This option adds the ability to dump driver state for debugging.
-	 The file debugfs/sky2/ethX displays the state of the internal
+	 The file /sys/kernel/debug/sky2/ethX displays the state of the internal
 	 transmit and receive rings.
 
 	 If unsure, say N.
drivers/net/wimax/i2400m/i2400m.h +1 -1
index 3ae2df3..da2e8f2 100644
@@ -389,7 +389,7 @@ struct i2400m {
 	unsigned ready:1;		/* all probing steps done */
 	unsigned rx_reorder:1;		/* RX reorder is enabled */
 	u8 trace_msg_from_user;		/* echo rx msgs to 'trace' pipe */
-					/* typed u8 so debugfs/u8 can tweak */
+					/* typed u8 so /sys/kernel/debug/u8 can tweak */
 	enum i2400m_system_state state;
 	wait_queue_head_t state_wq;	/* Woken up when on state updates */
 
drivers/net/wireless/ath5k/Kconfig +3 -3
index 75383a5..fdfa708 100644
@@ -27,11 +27,11 @@ config ATH5K_DEBUG
 	  Say Y, if and you will get debug options for ath5k.
 	  To use this, you need to mount debugfs:
 
-	  mkdir /debug/
-	  mount -t debugfs debug /debug/
+	  mkdir /sys/kernel/debug/
+	  mount -t debugfs debug /sys/kernel/debug/
 
 	  You will get access to files under:
-	  /debug/ath5k/phy0/
+	  /sys/kernel/debug/ath5k/phy0/
 
 	  To enable debug, pass the debug level to the debug module
 	  parameter. For example:
drivers/net/wireless/libertas/README +6 -6
index d860fc3..ab6a2d5 100644
@@ -72,7 +72,7 @@ rdrf
 	location that is to be read.  This parameter must be specified in
 	hexadecimal (its possible to preceed preceding the number with a "0x").
 
-	Path: /debugfs/libertas_wireless/ethX/registers/
+	Path: /sys/kernel/debug/libertas_wireless/ethX/registers/
 
 	Usage:
 		echo "0xa123" > rdmac ; cat rdmac
@@ -95,7 +95,7 @@ wrrf
 sleepparams
 	This command is used to set the sleepclock configurations
 
-	Path: /debugfs/libertas_wireless/ethX/
+	Path: /sys/kernel/debug/libertas_wireless/ethX/
 
 	Usage:
 		cat sleepparams: reads the current sleepclock configuration
@@ -115,7 +115,7 @@ subscribed_events
 	The subscribed_events directory contains the interface for the
 	subscribed events API.
 
-	Path: /debugfs/libertas_wireless/ethX/subscribed_events/
+	Path: /sys/kernel/debug/libertas_wireless/ethX/subscribed_events/
 
 	Each event is represented by a filename. Each filename consists of the
 	following three fields:
@@ -165,7 +165,7 @@ subscribed_events
 extscan
 	This command is used to do a specific scan.
 
-	Path: /debugfs/libertas_wireless/ethX/
+	Path: /sys/kernel/debug/libertas_wireless/ethX/
 
 	Usage: echo "SSID" > extscan
 
@@ -179,7 +179,7 @@ getscantable
 	Display the current contents of the driver scan table (ie. get the
 	scan results).
 
-	Path: /debugfs/libertas_wireless/ethX/
+	Path: /sys/kernel/debug/libertas_wireless/ethX/
 
 	Usage:
 		cat getscantable
@@ -188,7 +188,7 @@ setuserscan
 	Initiate a customized scan and retrieve the results
 
 
-	Path: /debugfs/libertas_wireless/ethX/
+	Path: /sys/kernel/debug/libertas_wireless/ethX/
 
     Usage:
        echo "[ARGS]" > setuserscan
drivers/scsi/lpfc/lpfc_debugfs.c +2 -2
index 52be564..9112380 100644
@@ -51,8 +51,8 @@
  * debugfs interface
  *
  * To access this interface the user should:
- * # mkdir /debug
- * # mount -t debugfs none /debug
+ * # mkdir /sys/kernel/debug
+ * # mount -t debugfs none /sys/kernel/debug
  *
  * The lpfc debugfs directory hierarchy is:
  * lpfc/lpfcX/vportY
include/linux/kernel.h +1 -1
index 883cd44..99b7aad 100644
@@ -406,7 +406,7 @@ static inline char *pack_hex_byte(char *buf, u8 byte)
  *
  * Use tracing_on/tracing_off when you want to quickly turn on or off
  * tracing. It simply enables or disables the recording of the trace events.
- * This also corresponds to the user space debugfs/tracing/tracing_on
+ * This also corresponds to the user space /sys/kernel/debug/tracing/tracing_on
  * file, which gives a means for the kernel and userspace to interact.
  * Place a tracing_off() in the kernel where you want tracing to end.
  * From user space, examine the trace, and then echo 1 > tracing_on
include/linux/tracepoint.h +2 -2
index d35a7ee..9c9ac7d 100644
@@ -197,7 +197,7 @@ static inline void tracepoint_synchronize_unregister(void)
  *	* This is how the trace record is structured and will
  *	* be saved into the ring buffer. These are the fields
  *	* that will be exposed to user-space in
- *	* /debug/tracing/events/<*>/format.
+ *	* /sys/kernel/debug/tracing/events/<*>/format.
  *	*
  *	* The declared 'local variable' is called '__entry'
  *	*
@@ -257,7 +257,7 @@ static inline void tracepoint_synchronize_unregister(void)
  * tracepoint callback (this is used by programmatic plugins and
  * can also by used by generic instrumentation like SystemTap), and
  * it is also used to expose a structured trace record in
- * /debug/tracing/events/.
+ * /sys/kernel/debug/tracing/events/.
  */
 
 #define TRACE_EVENT(name, proto, args, struct, assign, print)	\
kernel/trace/Kconfig +5 -5
index 417d198..e5ce28f 100644
@@ -120,7 +120,7 @@ config IRQSOFF_TRACER
 	  disabled by default and can be runtime (re-)started
 	  via:
 
-	      echo 0 > /debugfs/tracing/tracing_max_latency
+	      echo 0 > /sys/kernel/debug/tracing/tracing_max_latency
 
 	  (Note that kernel size and overhead increases with this option
 	  enabled. This option and the preempt-off timing option can be
@@ -141,7 +141,7 @@ config PREEMPT_TRACER
 	  disabled by default and can be runtime (re-)started
 	  via:
 
-	      echo 0 > /debugfs/tracing/tracing_max_latency
+	      echo 0 > /sys/kernel/debug/tracing/tracing_max_latency
 
 	  (Note that kernel size and overhead increases with this option
 	  enabled. This option and the irqs-off timing option can be
@@ -213,7 +213,7 @@ config TRACE_BRANCH_PROFILING
 	  This tracer profiles all the the likely and unlikely macros
 	  in the kernel. It will display the results in:
 
-	  /debugfs/tracing/profile_annotated_branch
+	  /sys/kernel/debug/tracing/profile_annotated_branch
 
 	  Note: this will add a significant overhead, only turn this
 	  on if you need to profile the system's use of these macros.
@@ -228,7 +228,7 @@ config PROFILE_ALL_BRANCHES
 	  taken in the kernel is recorded whether it hit or miss.
 	  The results will be displayed in:
 
-	  /debugfs/tracing/profile_branch
+	  /sys/kernel/debug/tracing/profile_branch
 
 	  This configuration, when enabled, will impose a great overhead
 	  on the system. This should only be enabled when the system
@@ -276,7 +276,7 @@ config STACK_TRACER
 	select KALLSYMS
 	help
 	  This special tracer records the maximum stack footprint of the
-	  kernel and displays it in debugfs/tracing/stack_trace.
+	  kernel and displays it in /sys/kernel/debug/tracing/stack_trace.
 
 	  This tracer works by hooking into every function call that the
 	  kernel executes, and keeping a maximum stack depth value and
kernel/trace/trace.c +12 -12
index cda81ec..9b94d57 100644
@@ -335,7 +335,7 @@ static raw_spinlock_t ftrace_max_lock =
 /*
  * Copy the new maximum trace into the separate maximum-trace
  * structure. (this way the maximum trace is permanently saved,
- * for later retrieval via /debugfs/tracing/latency_trace)
+ * for later retrieval via /sys/kernel/debug/tracing/latency_trace)
  */
 static void
 __update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu)
@@ -2366,21 +2366,21 @@ static const struct file_operations tracing_iter_fops = {
 
 static const char readme_msg[] =
 	"tracing mini-HOWTO:\n\n"
-	"# mkdir /debug\n"
-	"# mount -t debugfs nodev /debug\n\n"
-	"# cat /debug/tracing/available_tracers\n"
+	"# mkdir /sys/kernel/debug\n"
+	"# mount -t debugfs nodev /sys/kernel/debug\n\n"
+	"# cat /sys/kernel/debug/tracing/available_tracers\n"
 	"wakeup preemptirqsoff preemptoff irqsoff function sched_switch nop\n\n"
-	"# cat /debug/tracing/current_tracer\n"
+	"# cat /sys/kernel/debug/tracing/current_tracer\n"
 	"nop\n"
-	"# echo sched_switch > /debug/tracing/current_tracer\n"
-	"# cat /debug/tracing/current_tracer\n"
+	"# echo sched_switch > /sys/kernel/debug/tracing/current_tracer\n"
+	"# cat /sys/kernel/debug/tracing/current_tracer\n"
 	"sched_switch\n"
-	"# cat /debug/tracing/trace_options\n"
+	"# cat /sys/kernel/debug/tracing/trace_options\n"
 	"noprint-parent nosym-offset nosym-addr noverbose\n"
-	"# echo print-parent > /debug/tracing/trace_options\n"
-	"# echo 1 > /debug/tracing/tracing_enabled\n"
-	"# cat /debug/tracing/trace > /tmp/trace.txt\n"
-	"# echo 0 > /debug/tracing/tracing_enabled\n"
+	"# echo print-parent > /sys/kernel/debug/tracing/trace_options\n"
+	"# echo 1 > /sys/kernel/debug/tracing/tracing_enabled\n"
+	"# cat /sys/kernel/debug/tracing/trace > /tmp/trace.txt\n"
+	"# echo 0 > /sys/kernel/debug/tracing/tracing_enabled\n"
 ;
 
 static ssize_t
scripts/tracing/draw_functrace.py +4 -4
index 902f9a9..66f4d1b 100644
@@ -12,10 +12,10 @@ calls. Only the functions's names and the the call time are provided.
 
 Usage:
 	Be sure that you have CONFIG_FUNCTION_TRACER
-	# mkdir /debugfs
-	# mount -t debug debug /debug
-	# echo function > /debug/tracing/current_tracer
-	$ cat /debug/tracing/trace_pipe > ~/raw_trace_func
+	# mkdir /sys/kernel/debug
+	# mount -t debug debug /sys/kernel/debug
+	# echo function > /sys/kernel/debug/tracing/current_tracer
+	$ cat /sys/kernel/debug/tracing/trace_pipe > ~/raw_trace_func
 	Wait some times but not too much, the script is a bit slow.
 	Break the pipe (Ctrl + Z)
 	$ scripts/draw_functrace.py < raw_trace_func > draw_functrace
Signature · 16 lines
1.6.3.1




-----------------------------------------------
To unsubscribe from this list: send the line "unsubscribe linux-***" 
in the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

GeunSik Lim (  S A M S U N G     E L E C T R O N I C S  )
e-Mail  :1) [email protected]
         2) [email protected] , [email protected]
HomePage: http://blog.naver.com/invain/
-----------------------------------------------
#28 lore raw

PATCH 1/1 debugfs:Fix terminology inconsistency of dir name to mount debugfs filesystem.

Dear Greg and Steven,

This is debugfs's mount directory related patch files. Please, merge my patch. Any Comment?

I read ./linux-2.6.git/MAINTAINERS file to submit debugfs related this patch file. But, I don't know a maintainer about this patch exactly.
Um... If I mistake, Please, correct me.

Thanks,

Subject: [PATCH V1] Fix terminology inconsistency of dir name to mount debugfs filesystem.

Many developers use "/debug/" or "/debugfs/" or "/sys/kernel/debug/" directory name to mount debugfs filesystem for ftrace according to ./Documentation/tracers/ftrace.txt file.

And, three directory names(ex:/debug/, /debugfs/, /sys/kernel/debug/) is existed in kernel source like ftrace, DRM, Wireless, Documentation, Network[sky2]files to mount debugfs filesystem.

debugfs means debug filesystem for debugging easy to use by greg kroah hartman. "/sys/kernel/debug/" name is suitable as directory name of debugfs filesystem.
- debugfs related reference: http://lwn.net/Articles/334546/

Fix inconsistency of directory name to mount debugfs filesystem.

* From Steven Rostedt
- find_debugfs() and tracing_files() in this patch.

CC Jiri Kosina <[email protected]>
CC David Airlie <[email protected]>
CC Peter Osterlund <[email protected]>
CC Ananth N Mavinakayanahalli <[email protected]>
CC Anil S Keshavamurthy <[email protected]>
CC Masami Hiramatsu <[email protected]>
CC linux-kernel <[email protected]>
---
 Documentation/DocBook/debugobjects.tmpl           |    2 +-
 Documentation/cdrom/packet-writing.txt            |    2 +-
 Documentation/fault-injection/fault-injection.txt |   70 +++---
 Documentation/kprobes.txt                         |    6 +-
 Documentation/trace/ftrace.txt                    |  233 +++++++++++++--------
 Documentation/trace/mmiotrace.txt                 |   26 ++--
 drivers/block/pktcdvd.c                           |    2 +-
 drivers/gpu/drm/drm_debugfs.c                     |   12 +-
 drivers/gpu/drm/drm_drv.c                         |    2 +-
 drivers/gpu/drm/drm_stub.c                        |    2 +-
 drivers/net/Kconfig                               |    4 +-
 drivers/net/wimax/i2400m/i2400m.h                 |    2 +-
 drivers/net/wireless/ath5k/Kconfig                |    6 +-
 drivers/net/wireless/libertas/README              |   12 +-
 drivers/scsi/lpfc/lpfc_debugfs.c                  |    4 +-
 include/linux/kernel.h                            |    2 +-
 include/linux/tracepoint.h                        |    4 +-
 kernel/trace/Kconfig                              |   10 +-
 kernel/trace/trace.c                              |   24 +-
 scripts/tracing/draw_functrace.py                 |    8 +-
 20 files changed, 242 insertions(+), 191 deletions(-)
Documentation/DocBook/debugobjects.tmpl +1 -1
index 7f5f218..08ff908 100644
@@ -106,7 +106,7 @@
       number of errors are printk'ed including a full stack trace.
     </para>
     <para>
-      The statistics are available via debugfs/debug_objects/stats.
+      The statistics are available via /sys/kernel/debug/debug_objects/stats.
       They provide information about the number of warnings and the
       number of successful fixups along with information about the
       usage of the internal tracking objects and the state of the
Documentation/cdrom/packet-writing.txt +1 -1
index cf1f812..1c40777 100644
@@ -117,7 +117,7 @@ Using the pktcdvd debugfs interface
 
 To read pktcdvd device infos in human readable form, do:
 
-	# cat /debug/pktcdvd/pktcdvd[0-7]/info
+	# cat /sys/kernel/debug/pktcdvd/pktcdvd[0-7]/info
 
 For a description of the debugfs interface look into the file:
 
Documentation/fault-injection/fault-injection.txt +35 -35

145 lines not on this page · raw message

Documentation/kprobes.txt +3 -3
index 1e7a769..053037a 100644
@@ -507,9 +507,9 @@ http://www.linuxsymposium.org/2006/linuxsymposium_procv2.pdf (pages 101-115)
 Appendix A: The kprobes debugfs interface
 
 With recent kernels (> 2.6.20) the list of registered kprobes is visible
-under the /debug/kprobes/ directory (assuming debugfs is mounted at /debug).
+under the /sys/kernel/debug/kprobes/ directory (assuming debugfs is mounted at //sys/kernel/debug).
 
-/debug/kprobes/list: Lists all registered probes on the system
+/sys/kernel/debug/kprobes/list: Lists all registered probes on the system
 
 c015d71a  k  vfs_read+0x0
 c011a316  j  do_fork+0x0
@@ -525,7 +525,7 @@ virtual addresses that correspond to modules that've been unloaded),
 such probes are marked with [GONE]. If the probe is temporarily disabled,
 such probes are marked with [DISABLED].
 
-/debug/kprobes/enabled: Turn kprobes ON/OFF forcibly.
+/sys/kernel/debug/kprobes/enabled: Turn kprobes ON/OFF forcibly.
 
 Provides a knob to globally and forcibly turn registered kprobes ON or OFF.
 By default, all kprobes are enabled. By echoing "0" to this file, all
Documentation/trace/ftrace.txt +142 -91

452 lines not on this page · raw message

Documentation/trace/mmiotrace.txt +13 -13
index 5731c67..162effb 100644
@@ -32,41 +32,41 @@ is no way to automatically detect if you are losing events due to CPUs racing.
 Usage Quick Reference
 ---------------------
 
-$ mount -t debugfs debugfs /debug
-$ echo mmiotrace > /debug/tracing/current_tracer
-$ cat /debug/tracing/trace_pipe > mydump.txt &
+$ mount -t debugfs debugfs /sys/kernel/debug
+$ echo mmiotrace > /sys/kernel/debug/tracing/current_tracer
+$ cat /sys/kernel/debug/tracing/trace_pipe > mydump.txt &
 Start X or whatever.
-$ echo "X is up" > /debug/tracing/trace_marker
-$ echo nop > /debug/tracing/current_tracer
+$ echo "X is up" > /sys/kernel/debug/tracing/trace_marker
+$ echo nop > /sys/kernel/debug/tracing/current_tracer
 Check for lost events.
 
 
 Usage
 -----
 
-Make sure debugfs is mounted to /debug. If not, (requires root privileges)
-$ mount -t debugfs debugfs /debug
+Make sure debugfs is mounted to /sys/kernel/debug. If not, (requires root privileges)
+$ mount -t debugfs debugfs /sys/kernel/debug
 
 Check that the driver you are about to trace is not loaded.
 
 Activate mmiotrace (requires root privileges):
-$ echo mmiotrace > /debug/tracing/current_tracer
+$ echo mmiotrace > /sys/kernel/debug/tracing/current_tracer
 
 Start storing the trace:
-$ cat /debug/tracing/trace_pipe > mydump.txt &
+$ cat /sys/kernel/debug/tracing/trace_pipe > mydump.txt &
 The 'cat' process should stay running (sleeping) in the background.
 
 Load the driver you want to trace and use it. Mmiotrace will only catch MMIO
 accesses to areas that are ioremapped while mmiotrace is active.
 
 During tracing you can place comments (markers) into the trace by
-$ echo "X is up" > /debug/tracing/trace_marker
+$ echo "X is up" > /sys/kernel/debug/tracing/trace_marker
 This makes it easier to see which part of the (huge) trace corresponds to
 which action. It is recommended to place descriptive markers about what you
 do.
 
 Shut down mmiotrace (requires root privileges):
-$ echo nop > /debug/tracing/current_tracer
+$ echo nop > /sys/kernel/debug/tracing/current_tracer
 The 'cat' process exits. If it does not, kill it by issuing 'fg' command and
 pressing ctrl+c.
 
@@ -78,10 +78,10 @@ to view your kernel log and look for "mmiotrace has lost events" warning. If
 events were lost, the trace is incomplete. You should enlarge the buffers and
 try again. Buffers are enlarged by first seeing how large the current buffers
 are:
-$ cat /debug/tracing/buffer_size_kb
+$ cat /sys/kernel/debug/tracing/buffer_size_kb
 gives you a number. Approximately double this number and write it back, for
 instance:
-$ echo 128000 > /debug/tracing/buffer_size_kb
+$ echo 128000 > /sys/kernel/debug/tracing/buffer_size_kb
 Then start again from the top.
 
 If you are doing a trace for a driver project, e.g. Nouveau, you should also
drivers/block/pktcdvd.c +1 -1
index dc7a8c3..9aa8932 100644
@@ -430,7 +430,7 @@ static void pkt_sysfs_cleanup(void)
 /********************************************************************
   entries in debugfs
 
-  /debugfs/pktcdvd[0-7]/
+  /sys/kernel/debug/pktcdvd[0-7]/
 			info
 
  *******************************************************************/
drivers/gpu/drm/drm_debugfs.c +6 -6
index c77c6c6..6ce0e26 100644
@@ -105,7 +105,7 @@ int drm_debugfs_create_files(struct drm_info_list *files, int count,
 		ent = debugfs_create_file(files[i].name, S_IFREG | S_IRUGO,
 					  root, tmp, &drm_debugfs_fops);
 		if (!ent) {
-			DRM_ERROR("Cannot create /debugfs/dri/%s/%s\n",
+			DRM_ERROR("Cannot create /sys/kernel/debug/dri/%s/%s\n",
 				  name, files[i].name);
 			drm_free(tmp, sizeof(struct drm_info_node),
 				 _DRM_DRIVER);
@@ -133,9 +133,9 @@ EXPORT_SYMBOL(drm_debugfs_create_files);
  * \param minor device minor number
  * \param root DRI debugfs dir entry.
  *
- * Create the DRI debugfs root entry "/debugfs/dri", the device debugfs root entry
- * "/debugfs/dri/%minor%/", and each entry in debugfs_list as
- * "/debugfs/dri/%minor%/%name%".
+ * Create the DRI debugfs root entry "/sys/kernel/debug/dri", the device debugfs root entry
+ * "/sys/kernel/debug/dri/%minor%/", and each entry in debugfs_list as
+ * "/sys/kernel/debug/dri/%minor%/%name%".
  */
 int drm_debugfs_init(struct drm_minor *minor, int minor_id,
 		     struct dentry *root)
@@ -148,7 +148,7 @@ int drm_debugfs_init(struct drm_minor *minor, int minor_id,
 	sprintf(name, "%d", minor_id);
 	minor->debugfs_root = debugfs_create_dir(name, root);
 	if (!minor->debugfs_root) {
-		DRM_ERROR("Cannot create /debugfs/dri/%s\n", name);
+		DRM_ERROR("Cannot create /sys/kernel/debug/dri/%s\n", name);
 		return -1;
 	}
 
@@ -165,7 +165,7 @@ int drm_debugfs_init(struct drm_minor *minor, int minor_id,
 		ret = dev->driver->debugfs_init(minor);
 		if (ret) {
 			DRM_ERROR("DRM: Driver failed to initialize "
-				  "/debugfs/dri.\n");
+				  "/sys/kernel/debug/dri.\n");
 			return ret;
 		}
 	}
drivers/gpu/drm/drm_drv.c +1 -1
index 019b7c5..1bf7efd 100644
@@ -339,7 +339,7 @@ static int __init drm_core_init(void)
 
 	drm_debugfs_root = debugfs_create_dir("dri", NULL);
 	if (!drm_debugfs_root) {
-		DRM_ERROR("Cannot create /debugfs/dri\n");
+		DRM_ERROR("Cannot create /sys/kernel/debug/dri\n");
 		ret = -1;
 		goto err_p3;
 	}
drivers/gpu/drm/drm_stub.c +1 -1
index b9631e3..b2a93c4 100644
@@ -328,7 +328,7 @@ static int drm_get_minor(struct drm_device *dev, struct drm_minor **minor, int t
 #if defined(CONFIG_DEBUG_FS)
 	ret = drm_debugfs_init(new_minor, minor_id, drm_debugfs_root);
 	if (ret) {
-		DRM_ERROR("DRM: Failed to initialize /debugfs/dri.\n");
+		DRM_ERROR("DRM: Failed to initialize /sys/kernel/debug/dri.\n");
 		goto err_g2;
 	}
 #endif
drivers/net/Kconfig +2 -2
index 214a92d..a170cea 100644
@@ -2200,7 +2200,7 @@ config SKGE_DEBUG
        depends on SKGE && DEBUG_FS
        help
 	 This option adds the ability to dump driver state for debugging.
-	 The file debugfs/skge/ethX displays the state of the internal
+	 The file /sys/kernel/debug/skge/ethX displays the state of the internal
 	 transmit and receive rings.
 
 	 If unsure, say N.
@@ -2226,7 +2226,7 @@ config SKY2_DEBUG
        depends on SKY2 && DEBUG_FS
        help
 	 This option adds the ability to dump driver state for debugging.
-	 The file debugfs/sky2/ethX displays the state of the internal
+	 The file /sys/kernel/debug/sky2/ethX displays the state of the internal
 	 transmit and receive rings.
 
 	 If unsure, say N.
drivers/net/wimax/i2400m/i2400m.h +1 -1
index 3ae2df3..da2e8f2 100644
@@ -389,7 +389,7 @@ struct i2400m {
 	unsigned ready:1;		/* all probing steps done */
 	unsigned rx_reorder:1;		/* RX reorder is enabled */
 	u8 trace_msg_from_user;		/* echo rx msgs to 'trace' pipe */
-					/* typed u8 so debugfs/u8 can tweak */
+					/* typed u8 so /sys/kernel/debug/u8 can tweak */
 	enum i2400m_system_state state;
 	wait_queue_head_t state_wq;	/* Woken up when on state updates */
 
drivers/net/wireless/ath5k/Kconfig +3 -3
index 75383a5..fdfa708 100644
@@ -27,11 +27,11 @@ config ATH5K_DEBUG
 	  Say Y, if and you will get debug options for ath5k.
 	  To use this, you need to mount debugfs:
 
-	  mkdir /debug/
-	  mount -t debugfs debug /debug/
+	  mkdir /sys/kernel/debug/
+	  mount -t debugfs debug /sys/kernel/debug/
 
 	  You will get access to files under:
-	  /debug/ath5k/phy0/
+	  /sys/kernel/debug/ath5k/phy0/
 
 	  To enable debug, pass the debug level to the debug module
 	  parameter. For example:
drivers/net/wireless/libertas/README +6 -6
index d860fc3..ab6a2d5 100644
@@ -72,7 +72,7 @@ rdrf
 	location that is to be read.  This parameter must be specified in
 	hexadecimal (its possible to preceed preceding the number with a "0x").
 
-	Path: /debugfs/libertas_wireless/ethX/registers/
+	Path: /sys/kernel/debug/libertas_wireless/ethX/registers/
 
 	Usage:
 		echo "0xa123" > rdmac ; cat rdmac
@@ -95,7 +95,7 @@ wrrf
 sleepparams
 	This command is used to set the sleepclock configurations
 
-	Path: /debugfs/libertas_wireless/ethX/
+	Path: /sys/kernel/debug/libertas_wireless/ethX/
 
 	Usage:
 		cat sleepparams: reads the current sleepclock configuration
@@ -115,7 +115,7 @@ subscribed_events
 	The subscribed_events directory contains the interface for the
 	subscribed events API.
 
-	Path: /debugfs/libertas_wireless/ethX/subscribed_events/
+	Path: /sys/kernel/debug/libertas_wireless/ethX/subscribed_events/
 
 	Each event is represented by a filename. Each filename consists of the
 	following three fields:
@@ -165,7 +165,7 @@ subscribed_events
 extscan
 	This command is used to do a specific scan.
 
-	Path: /debugfs/libertas_wireless/ethX/
+	Path: /sys/kernel/debug/libertas_wireless/ethX/
 
 	Usage: echo "SSID" > extscan
 
@@ -179,7 +179,7 @@ getscantable
 	Display the current contents of the driver scan table (ie. get the
 	scan results).
 
-	Path: /debugfs/libertas_wireless/ethX/
+	Path: /sys/kernel/debug/libertas_wireless/ethX/
 
 	Usage:
 		cat getscantable
@@ -188,7 +188,7 @@ setuserscan
 	Initiate a customized scan and retrieve the results
 
 
-	Path: /debugfs/libertas_wireless/ethX/
+	Path: /sys/kernel/debug/libertas_wireless/ethX/
 
     Usage:
        echo "[ARGS]" > setuserscan
drivers/scsi/lpfc/lpfc_debugfs.c +2 -2
index 52be564..9112380 100644
@@ -51,8 +51,8 @@
  * debugfs interface
  *
  * To access this interface the user should:
- * # mkdir /debug
- * # mount -t debugfs none /debug
+ * # mkdir /sys/kernel/debug
+ * # mount -t debugfs none /sys/kernel/debug
  *
  * The lpfc debugfs directory hierarchy is:
  * lpfc/lpfcX/vportY
include/linux/kernel.h +1 -1
index 883cd44..99b7aad 100644
@@ -406,7 +406,7 @@ static inline char *pack_hex_byte(char *buf, u8 byte)
  *
  * Use tracing_on/tracing_off when you want to quickly turn on or off
  * tracing. It simply enables or disables the recording of the trace events.
- * This also corresponds to the user space debugfs/tracing/tracing_on
+ * This also corresponds to the user space /sys/kernel/debug/tracing/tracing_on
  * file, which gives a means for the kernel and userspace to interact.
  * Place a tracing_off() in the kernel where you want tracing to end.
  * From user space, examine the trace, and then echo 1 > tracing_on
include/linux/tracepoint.h +2 -2
index d35a7ee..9c9ac7d 100644
@@ -197,7 +197,7 @@ static inline void tracepoint_synchronize_unregister(void)
  *	* This is how the trace record is structured and will
  *	* be saved into the ring buffer. These are the fields
  *	* that will be exposed to user-space in
- *	* /debug/tracing/events/<*>/format.
+ *	* /sys/kernel/debug/tracing/events/<*>/format.
  *	*
  *	* The declared 'local variable' is called '__entry'
  *	*
@@ -257,7 +257,7 @@ static inline void tracepoint_synchronize_unregister(void)
  * tracepoint callback (this is used by programmatic plugins and
  * can also by used by generic instrumentation like SystemTap), and
  * it is also used to expose a structured trace record in
- * /debug/tracing/events/.
+ * /sys/kernel/debug/tracing/events/.
  */
 
 #define TRACE_EVENT(name, proto, args, struct, assign, print)	\
kernel/trace/Kconfig +5 -5
index 417d198..e5ce28f 100644
@@ -120,7 +120,7 @@ config IRQSOFF_TRACER
 	  disabled by default and can be runtime (re-)started
 	  via:
 
-	      echo 0 > /debugfs/tracing/tracing_max_latency
+	      echo 0 > /sys/kernel/debug/tracing/tracing_max_latency
 
 	  (Note that kernel size and overhead increases with this option
 	  enabled. This option and the preempt-off timing option can be
@@ -141,7 +141,7 @@ config PREEMPT_TRACER
 	  disabled by default and can be runtime (re-)started
 	  via:
 
-	      echo 0 > /debugfs/tracing/tracing_max_latency
+	      echo 0 > /sys/kernel/debug/tracing/tracing_max_latency
 
 	  (Note that kernel size and overhead increases with this option
 	  enabled. This option and the irqs-off timing option can be
@@ -213,7 +213,7 @@ config TRACE_BRANCH_PROFILING
 	  This tracer profiles all the the likely and unlikely macros
 	  in the kernel. It will display the results in:
 
-	  /debugfs/tracing/profile_annotated_branch
+	  /sys/kernel/debug/tracing/profile_annotated_branch
 
 	  Note: this will add a significant overhead, only turn this
 	  on if you need to profile the system's use of these macros.
@@ -228,7 +228,7 @@ config PROFILE_ALL_BRANCHES
 	  taken in the kernel is recorded whether it hit or miss.
 	  The results will be displayed in:
 
-	  /debugfs/tracing/profile_branch
+	  /sys/kernel/debug/tracing/profile_branch
 
 	  This configuration, when enabled, will impose a great overhead
 	  on the system. This should only be enabled when the system
@@ -276,7 +276,7 @@ config STACK_TRACER
 	select KALLSYMS
 	help
 	  This special tracer records the maximum stack footprint of the
-	  kernel and displays it in debugfs/tracing/stack_trace.
+	  kernel and displays it in /sys/kernel/debug/tracing/stack_trace.
 
 	  This tracer works by hooking into every function call that the
 	  kernel executes, and keeping a maximum stack depth value and
kernel/trace/trace.c +12 -12
index cda81ec..9b94d57 100644
@@ -335,7 +335,7 @@ static raw_spinlock_t ftrace_max_lock =
 /*
  * Copy the new maximum trace into the separate maximum-trace
  * structure. (this way the maximum trace is permanently saved,
- * for later retrieval via /debugfs/tracing/latency_trace)
+ * for later retrieval via /sys/kernel/debug/tracing/latency_trace)
  */
 static void
 __update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu)
@@ -2366,21 +2366,21 @@ static const struct file_operations tracing_iter_fops = {
 
 static const char readme_msg[] =
 	"tracing mini-HOWTO:\n\n"
-	"# mkdir /debug\n"
-	"# mount -t debugfs nodev /debug\n\n"
-	"# cat /debug/tracing/available_tracers\n"
+	"# mkdir /sys/kernel/debug\n"
+	"# mount -t debugfs nodev /sys/kernel/debug\n\n"
+	"# cat /sys/kernel/debug/tracing/available_tracers\n"
 	"wakeup preemptirqsoff preemptoff irqsoff function sched_switch nop\n\n"
-	"# cat /debug/tracing/current_tracer\n"
+	"# cat /sys/kernel/debug/tracing/current_tracer\n"
 	"nop\n"
-	"# echo sched_switch > /debug/tracing/current_tracer\n"
-	"# cat /debug/tracing/current_tracer\n"
+	"# echo sched_switch > /sys/kernel/debug/tracing/current_tracer\n"
+	"# cat /sys/kernel/debug/tracing/current_tracer\n"
 	"sched_switch\n"
-	"# cat /debug/tracing/trace_options\n"
+	"# cat /sys/kernel/debug/tracing/trace_options\n"
 	"noprint-parent nosym-offset nosym-addr noverbose\n"
-	"# echo print-parent > /debug/tracing/trace_options\n"
-	"# echo 1 > /debug/tracing/tracing_enabled\n"
-	"# cat /debug/tracing/trace > /tmp/trace.txt\n"
-	"# echo 0 > /debug/tracing/tracing_enabled\n"
+	"# echo print-parent > /sys/kernel/debug/tracing/trace_options\n"
+	"# echo 1 > /sys/kernel/debug/tracing/tracing_enabled\n"
+	"# cat /sys/kernel/debug/tracing/trace > /tmp/trace.txt\n"
+	"# echo 0 > /sys/kernel/debug/tracing/tracing_enabled\n"
 ;
 
 static ssize_t
scripts/tracing/draw_functrace.py +4 -4
index 902f9a9..66f4d1b 100644
@@ -12,10 +12,10 @@ calls. Only the functions's names and the the call time are provided.
 
 Usage:
 	Be sure that you have CONFIG_FUNCTION_TRACER
-	# mkdir /debugfs
-	# mount -t debug debug /debug
-	# echo function > /debug/tracing/current_tracer
-	$ cat /debug/tracing/trace_pipe > ~/raw_trace_func
+	# mkdir /sys/kernel/debug
+	# mount -t debug debug /sys/kernel/debug
+	# echo function > /sys/kernel/debug/tracing/current_tracer
+	$ cat /sys/kernel/debug/tracing/trace_pipe > ~/raw_trace_func
 	Wait some times but not too much, the script is a bit slow.
 	Break the pipe (Ctrl + Z)
 	$ scripts/draw_functrace.py < raw_trace_func > draw_functrace
Signature · 16 lines
1.6.3.1



 
-----------------------------------------------
To unsubscribe from this list: send the line "unsubscribe linux-***" 
in the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

GeunSik Lim ( Samsung Electronics )
e-Mail  :1) [email protected]
         2) [email protected] , [email protected]
HomePage: http://blog.naver.com/invain/
-----------------------------------------------