- 12 Oct, 2007 40 commits
-
-
Eric W. Biederman authored
Since sysfs no longer stores fs directory information in the dcache on a permanent basis kill_litter_super it is inappropriate and actively wrong. It will decrement the count on all dentries left in the dcache before trying to free them. At the moment this is not biting us only because we never unmount sysfs. Signed-off-by:
Eric W. Biederman <ebiederm@xmission.com> Signed-off-by:
Tejun Heo <htejun@gmail.com> Cc: Cornelia Huck <cornelia.huck@de.ibm.com> Signed-off-by:
Greg Kroah-Hartman <gregkh@suse.de>
-
Eric W. Biederman authored
Now that sysfs_get_inode is dropping the inode lock we no longer have a need from sysfs_instantiate. Signed-off-by:
-
Eric W. Biederman authored
Signed-off-by:
-
Tejun Heo authored
lookup_one_len_kern() should be called with the parent's i_mutex locked. Fix it. Spotted by Eric W. Biederman. Signed-off-by:
-
Jean Delvare authored
While platform_device.id is a u32, platform_device_add() handles "-1" as a special id value. This has potential for confusion and bugs. Making it an int instead should prevent problems from happening in the future. Signed-off-by:
Jean Delvare <khali@linux-fr.org> Signed-off-by:
-
Akinobu Mita authored
No one uses sysdev_drivers. Because no one calls sysdev_driver_register with NULL class. And it is difficult to imagine that someone want to implement a global sysdev driver which is called with all sys_device on any kind of sysdev_class. So this patch removes global sysdev_drivers list and update comments for this change. Signed-off-by:
Akinobu Mita <akinobu.mita@gmail.com> Cc: Tejun Heo <htejun@gmail.com> Acked-by:
Cornelia Huck <cornelia.huck@de.ibm.com> Signed-off-by:
-
Rolf Eike Beer authored
Signed-off-by:
Rolf Eike Beer <eike-kernel@sf-tec.de> Signed-off-by:
Andrew Morton <akpm@linux-foundation.org> Signed-off-by:
-
Adrian Bunk authored
suspend_device() and resume_device() can now become static. Signed-off-by:
Adrian Bunk <bunk@stusta.de> Signed-off-by:
-
Markus Rechberger authored
following patch fixes the i2c name collision with i2c-dev. http://mcentral.de/wiki/index.php/Bugtracker#i2c_core_problem This issue has been experienced with em28xx and saa7133 based devices. I discussed that problem with Jean Delvare a while ago and he proposed to add a prefix to the class name. Signed-off-by:
Markus Rechberger <markus.rechberger@amd.com> Acked-by:
Marcel Holtmann <marcel@holtmann.org> Signed-off-by:
-
Kay Sievers authored
Signed-off-by:
Kay Sievers <kay.sievers@vrfy.org> Signed-off-by:
-
Kay Sievers authored
This has been in the SuSE kernels for some time now. Signed-off-by:
-
Kay Sievers authored
Move uevent specific logic from the core into kobject_uevent.c, which does no longer require to link the unused string array if hotplug is not compiled in. Signed-off-by:
-
tonyj@suse.de authored
Convert from class_device to device in drivers/char. Signed-off-by:
Tony Jones <tonyj@suse.de> Signed-off-by:
-
tonyj@suse.de authored
Convert from class_device to device for drivers/video. Signed-off-by:
-
Jean Delvare authored
The DEFINE_DMI_ATTR macro has a single user left so we can expand it for slightly shorter/simpler code. Signed-off-by:
Lennart Poettering <mzxreary@0pointer.de> Signed-off-by:
-
Jean Delvare authored
We can use sysfs attributes with an extra parameter for dmi id attributes. This makes it possible to use the same callback function for all attributes, reducing the binary size significantly (-18% on x86_64.) Signed-off-by:
-
Tejun Heo authored
With the previous sysfs_add_one() update, there is only one user of the return value of sysfs_addrm_finish() and the user can switch to testing @sd easily. Make sysfs_addrm_finish() return void for cleaner semantics as suggested by Satyam Sharma. This patch doesn't introduce any noticeable behavior change. Signed-off-by:
-
Tejun Heo authored
Make sysfs_add_one() check for duplicate entry and return -EEXIST if such entry exists. This simplifies node addition code a bit. This patch doesn't introduce any noticeable behavior change. Signed-off-by:
-
Tejun Heo authored
When adding or removing a sysfs_dirent, the user used to be required to call link/unlink separately. It was for two reasons - code looked like that before sysfs_addrm_cxt conversion and to avoid looping through parent_sd->children list twice during removal. Performance optimization during removal just isn't worth it. Make sysfs_add/remove_one() call sysfs_link/unlink_sibing() implicitly. This makes code simpler albeit slightly less efficient. This change doesn't introduce any noticeable behavior change. Signed-off-by:
-
Tejun Heo authored
With the shadow directories gone, sysfs_rename_dir() can be simplified. * parent doesn't need to be grabbed separately. Just access old_dentry->d_parent. * parent sd can never change. Remove code to move under the new parent. * Massage comments a bit. Signed-off-by:
-
Tejun Heo authored
* remove space between * and symbol name in variable declaration. * kill unnecessary new line. * kill 'found' and test 'sd' instead. Signed-off-by:
-
Eric W. Biederman authored
While shadow directories appear to be a good idea, the current scheme of controlling their creation and destruction outside of sysfs appears to be a locking and maintenance nightmare in the face of sysfs directories dynamically coming and going. Which can now occur for directories containing network devices when CONFIG_SYSFS_DEPRECATED is not set. This patch removes everything from the initial shadow directory support that allowed the shadow directory creation to be controlled at a higher level. So except for a few bits of sysfs_rename_dir everything from commit b592fcfe is now gone. Signed-off-by:
-
Dave Young authored
Cleanup semaphore.h Signed-off-by:
Dave Young <hidave.darkstar@gmail.com> Signed-off-by:
-
Dave Young authored
Use mutex instead of semaphore in sysfs/file.c : sys_buffer. Signed-off-by:
-
Robin Getz authored
Allows debugfs helper functions to have a hex output, rather than just decimal Signed-off-by:
Robin Getz <rgetz@blackfin.uclinux.org> Signed-off-by:
-
Cornelia Huck authored
Signed-off-by:
-
Greg Kroah-Hartman authored
Due to historical reasons, struct kobject contained a static array for the name, and a dynamic pointer in case the name got bigger than the array. That's just dumb, as people didn't always know which variable to reference, even with the accessor for the kobject name. This patch removes the static array, potentially saving a lot of memory as the majority of kobjects do not have a very long name. Thanks to Kay for the idea to do this. Cc: Kay Sievers <kay.sievers@vrfy.org> Signed-off-by: -
Greg Kroah-Hartman authored
A kset should not have its name set directly, so dynamically set the name at runtime. This is needed to remove the static array in the kobject structure which will be changed in a future patch. Signed-off-by: -
Greg Kroah-Hartman authored
struct cdev does not need the kobject name to be set, as it is never used. This patch fixes up the few places it is set. Signed-off-by:
Mauro Carvalho Chehab <mchehab@infradead.org>
-
Greg Kroah-Hartman authored
A number of different drivers incorrect access the kobject name field directly. This is not correct as the name might not be in the array. Use the proper accessor function instead.
-
Greg Kroah-Hartman authored
get_bus() should not be globally visable as it is not used by anything other than drivers/base/bus.c. This patch removes the visability of it, and renames it to match all of the other *_get() functions in the kernel. Signed-off-by: -
Greg Kroah-Hartman authored
put_bus() should not be globally visable as it is not used by anything other than drivers/base/bus.c. This patch removes the visability of it, and renames it to match all of the other *_put() functions in the kernel. Signed-off-by: -
Greg Kroah-Hartman authored
There are no more subsystems, it's a kset now so remove the function and the only two users, which are in the driver core. Signed-off-by: -
Greg Kroah-Hartman authored
There are no more subsystems, it's a kset now so remove the function and the only two users, which are in the driver core. Signed-off-by: -
Greg Kroah-Hartman authored
This macro is only used by the driver core and is held over from when we had subsystems. It is not needed anymore. Signed-off-by: -
Greg Kroah-Hartman authored
This macro is only used by the driver core and is held over from when we had subsystems. It is not needed anymore. Signed-off-by: -
Kay Sievers authored
The kernel creates a process for every event that is send, even when there is no binary it could execute. We are needlessly creating around 200-300 failing processes during early bootup, until we have the chance to disable it from userspace. This change allows us to disable /sbin/hotplug entirely, if you want to, by setting UEVENT_HELPER_PATH="" in the kernel config. Signed-off-by:
-
Kay Sievers authored
This changes the uevent buffer functions to use a struct instead of a long list of parameters. It does no longer require the caller to do the proper buffer termination and size accounting, which is currently wrong in some places. It fixes a known bug where parts of the uevent environment are overwritten because of wrong index calculations. Many thanks to Mathieu Desnoyers for finding bugs and improving the error handling. Signed-off-by:
-
Kay Sievers authored
Attributes do not have an owner(module) anymore, so there is no need to carry the attributes in every single bus instance. Signed-off-by:
-
Tsugikazu Shibata authored
Here is another sync patch of Documentation/ja_JP/HOWTO Japanese developer sent me some cosmetic changes and also follow changes of HOWTO Cross reference URL (sosdg.org/qiyong/lxr) known_regression explanations on kernel dev. process Signed-off-by:Tsugikazu Shibata <tshibata@ab.jp.nec.com> Cc: stable <stable@kernel.org> Signed-off-by:
-
