summaryrefslogtreecommitdiff
path: root/tools/testing/selftests/sysctl/Makefile
blob: b3c33e071f10069ad01d0029c5d7c0350e413a18 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
# Makefile for sysctl selftests.
# Expects kernel.sysctl_writes_strict=1.

# No binaries, but make sure arg-less "make" doesn't trigger "run_tests".
all:

TEST_PROGS := run_numerictests run_stringtests
TEST_FILES := common_tests

include ../lib.mk

# Nothing to clean up.
clean:
hang <jszhang@marvell.com>2016-11-10 17:21:29 +0800 committerMark Brown <broonie@kernel.org>2016-11-11 15:38:08 +0000 commit09f2ba0b0b7c44ecea49cf69a708203b76ba5535 (patch) tree042df33ac99f77d8b86ac427431e267c33561c35 /sound/firewire parent1001354ca34179f3db924eb66672442a173147dc (diff)
regulator: gpio: properly check return value of of_get_named_gpio
The function of_get_named_gpio() could return -ENOENT, -EPROBE_DEFER -EINVAL and so on. Currently, for the optional property "enable-gpio", we only check -EPROBE_DEFER, this is not enough since there may be misconfigured "enable-gpio" in the DTB, of_get_named_gpio() will return -EINVAL in this case, we should return immediately here. And for the optional property "gpios", we didn't check the return value, the driver will continue to the point where gpio_request_array() is called, it doesn't make sense to continue if we got -EPROBE_DEFER or -EINVAL here. This patch tries to address these two issues by properly checking the return value of of_get_named_gpio. Signed-off-by: Jisheng Zhang <jszhang@marvell.com> Signed-off-by: Mark Brown <broonie@kernel.org>
Diffstat (limited to 'sound/firewire')