debugfs for 2.6.10-rc3
Subsystems (4)
PATCH 1/1 debugfs for 2.6.10-rc3
I've added debugfs to my bk driver tree (located at bk://kernel.bkbits.net/gregkh/linux/driver-2.6) so it will show up in the next -mm release. For those who want to see the patch version, or want to put it in any other kernel tree (Fedora perhaps?) I've included it below. I haven't added the kobject interface yet, I'll try to get to that next week.
thanks,
greg k-h
----------------------------
debugfs: add debugfs
debugfs is a filesystem that is just for debug data. Start moving stuff out of proc and sysfs now :)
Documentation/DocBook/kernel-api.tmpl +9 -0
@@ -105,6 +105,15 @@ </sect1> </chapter> + <chapter id="debugfs"> + <title>The debugfs filesystem</title> + + <sect1><title>debugfs interface</title> +!Efs/debugfs/inode.c +!Efs/debugfs/file.c + </sect1> + </chapter> + <chapter id="vfs"> <title>The Linux VFS</title> <sect1><title>The Directory Cache</title>
fs/Makefile +1 -0
@@ -94,3 +94,4 @@ obj-$(CONFIG_BEFS_FS) += befs/ obj-$(CONFIG_HOSTFS) += hostfs/ obj-$(CONFIG_HPPFS) += hppfs/ +obj-$(CONFIG_DEBUG_FS) += debugfs/
fs/debugfs/Makefile new file +4 -0
@@ -0,0 +1,4 @@ +debugfs-objs := inode.o file.o + +obj-$(CONFIG_DEBUG_FS) += debugfs.o +
fs/debugfs/file.c new file +262 -0
262 lines not on this page · raw message
fs/debugfs/inode.c new file +315 -0
315 lines not on this page · raw message
include/linux/debugfs.h new file +90 -0
90 lines not on this page · raw message
lib/Kconfig.debug +10 -0
@@ -107,6 +107,16 @@ If you're truly short on disk space or don't expect to report any bugs back to the UML developers, say N, otherwise say Y. +config DEBUG_FS + bool "Debug Filesystem" + depends on DEBUG_KERNEL + help + debugfs is a virtual file system that kernel developers use to put + debugging files into. Enable this option to be able to read and + write to these files. + + If unsure, say N. + if !X86_64 config FRAME_POINTER bool "Compile the kernel with frame pointers"
Quoted 9 lines · Greg KH
I've added debugfs to my bk driver tree (located at bk://kernel.bkbits.net/gregkh/linux/driver-2.6) so it will show up in the next -mm release. For those who want to see the patch version, or want to put it in any other kernel tree (Fedora perhaps?) I've included it below. I haven't added the kobject interface yet, I'll try to get to that next week.
thanks,
Quoted 8 lines
greg k-h
----------------------------
debugfs: add debugfs
debugfs is a filesystem that is just for debug data. Start moving stuff out of proc and sysfs now :)
Hi Greg
Is there any reason why this shouldn't be buildable as module? I saw no apparent reason and added "|| defined(MODULE)" to debugfs.h and made the Kconfig option tristate. This allowed it to be built as a working module (yay, i didn't have to restart to try it out!).
Also I saw there was no debugfs_create_u64() for us on 64bit machines ;)
Thanks
Alexander
Quoted 19 lines · Alexander Nyberg
Quoted 9 lines · Greg KH
I've added debugfs to my bk driver tree (located at bk://kernel.bkbits.net/gregkh/linux/driver-2.6) so it will show up in the next -mm release. For those who want to see the patch version, or want to put it in any other kernel tree (Fedora perhaps?) I've included it below. I haven't added the kobject interface yet, I'll try to get to that next week.
thanks,
Quoted 8 lines
greg k-h
----------------------------
debugfs: add debugfs
debugfs is a filesystem that is just for debug data. Start moving stuff out of proc and sysfs now :)
Quoted 6 lines
Hi Greg
Is there any reason why this shouldn't be buildable as module? I saw no apparent reason and added "|| defined(MODULE)" to
debugfs.hand made theKconfigoption tristate. This allowed it to be built as a working module (yay, i didn't have to restart to try it out!).
If debugfs calls are used by any code that is built into the kernel, then the build would fail if debugfs is a module. As that would soon get _real_ messy from a Kbuild standpoint, I took the easy way out and just made it a Y/N type option :)
Quoted 1 line
Also I saw there was no
debugfs_create_u64()for us on 64bit machines ;)
If you have a need for it, I'll be glad to add it :)
thanks,
greg k-h
Quoted 3 lines · Greg KH
I've added debugfs to my bk driver tree (located at bk://kernel.bkbits.net/gregkh/linux/driver-2.6) so it will show up in the next -mm release.
Debugfs is a very natural name, but it will undoubtedly cause confusion since we already have debugfs(8) in e2fsprogs. One is a filesystem, and the other is a user-mode command, so it's not a total name collision, but it could cause some communication mixups.
On the other hand, I couldn't think of a better name, so perhaps we should just live with it. I did want to point out the potential problem now while there's still a chance to change it, though....
- Ted
Quoted 5 lines · Theodore Ts'o
Quoted 3 lines · Greg KH
I've added debugfs to my bk driver tree (located at bk://kernel.bkbits.net/gregkh/linux/driver-2.6) so it will show up in the next -mm release.
Quoted 8 lines
Debugfs is a very natural name, but it will undoubtedly cause confusion since we already have debugfs(8) in e2fsprogs. One is a filesystem, and the other is a user-mode command, so it's not a total name collision, but it could cause some communication mixups.
On the other hand, I couldn't think of a better name, so perhaps we should just live with it. I did want to point out the potential problem now while there's still a chance to change it, though....
I agree about the name clash (I pointed it out in my original debugfs post) and like you, I couldn't think of a better name either.
Think of it like a trademark, they both have the same name, yet they live in different locations, hence they are both allowed to exist :)
thanks,
greg k-h