summaryrefslogtreecommitdiffstats
path: root/net
diff options
context:
space:
mode:
authorBenjamin Tissoires <benjamin.tissoires@redhat.com>2013-07-24 19:38:04 +0200
committerJiri Kosina <jkosina@suse.cz>2013-07-31 10:12:28 +0200
commitabf832bfc349b54fd500f1e3b612f7f3cd9dfcc6 (patch)
tree4ece0e75cf3fd463be58376e3f900c1cfd60dd05 /net
parent3366dd9fa887ebbda4872e9554f853eaeda764be (diff)
downloadlinux-stable-abf832bfc349b54fd500f1e3b612f7f3cd9dfcc6.tar.gz
linux-stable-abf832bfc349b54fd500f1e3b612f7f3cd9dfcc6.tar.bz2
linux-stable-abf832bfc349b54fd500f1e3b612f7f3cd9dfcc6.zip
HID: trivial devm conversion for special hid drivers
It is safe to use devres allocation within the hid subsystem: - the devres release is called _after_ the call to .remove(), meaning that no freed pointers will exists while removing the device - if a .probe() fails, devres releases all the allocated ressources before going to the next driver: there will not be ghost ressources attached to a hid device if several drivers are probed. Given that, we can clean up a little some of the HID drivers. These ones are trivial: - there is only one kzalloc in the driver - the .remove() callback contains only one kfree on top of hid_hw_stop() - the error path in the probe is easy enough to be manually checked Signed-off-by: Benjamin Tissoires <benjamin.tissoires@redhat.com> Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com> Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Diffstat (limited to 'net')
0 files changed, 0 insertions, 0 deletions