2016-11-22 14:34:20

by Arnd Bergmann

Subject: [PATCH] clocksource: nps: avoid maybe-uninitialized warning

We get a harmless false-positive warning with the newly added nps
clocksource driver:

drivers/clocksource/timer-nps.c: In function 'nps_setup_clocksource':
drivers/clocksource/timer-nps.c:102:6: error: 'nps_timer1_freq' may be used uninitialized in this function [-Werror=maybe-uninitialized]

Gcc here fails to identify that IS_ERR() is only true if PTR_ERR()
has a nonzero value. Using PTR_ERR_OR_ZERO() to convert the result
first makes this obvious and shuts up the warning.

Fixes: 0ee4d9922df5 ("clocksource: Add clockevent support to NPS400 driver")
Signed-off-by: Arnd Bergmann <[email protected]>
---
drivers/clocksource/timer-nps.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/drivers/clocksource/timer-nps.c b/drivers/clocksource/timer-nps.c
index b4c8a023a2d4..8da5e93b6810 100644
--- a/drivers/clocksource/timer-nps.c
+++ b/drivers/clocksource/timer-nps.c
@@ -53,9 +53,10 @@ static int __init nps_get_timer_clk(struct device_node *node,
int ret;

*clk = of_clk_get(node, 0);
- if (IS_ERR(*clk)) {
+ ret = PTR_ERR_OR_ZERO(*clk);
+ if (ret) {
pr_err("timer missing clk");
- return PTR_ERR(*clk);
+ return ret;
}

ret = clk_prepare_enable(*clk);
--
2.9.0

2016-11-22 18:25:50

by Vineet Gupta

Subject: Re: [PATCH] clocksource: nps: avoid maybe-uninitialized warning

On 11/22/2016 06:34 AM, Arnd Bergmann wrote:
> We get a harmless false-positive warning with the newly added nps
> clocksource driver:
>
> drivers/clocksource/timer-nps.c: In function 'nps_setup_clocksource':
> drivers/clocksource/timer-nps.c:102:6: error: 'nps_timer1_freq' may be used uninitialized in this function [-Werror=maybe-uninitialized]
>
> Gcc here fails to identify that IS_ERR() is only true if PTR_ERR()
> has a nonzero value. Using PTR_ERR_OR_ZERO() to convert the result
> first makes this obvious and shuts up the warning.
>
> Fixes: 0ee4d9922df5 ("clocksource: Add clockevent support to NPS400 driver")
> Signed-off-by: Arnd Bergmann <[email protected]>

Thx for this fix Arnd. Since the driver is going via my tree, I added it to the
patchset.

-Vineet

> ---
> drivers/clocksource/timer-nps.c | 5 +++--
> 1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/clocksource/timer-nps.c b/drivers/clocksource/timer-nps.c
> index b4c8a023a2d4..8da5e93b6810 100644
> --- a/drivers/clocksource/timer-nps.c
> +++ b/drivers/clocksource/timer-nps.c
> @@ -53,9 +53,10 @@ static int __init nps_get_timer_clk(struct device_node *node,
> int ret;
>
> *clk = of_clk_get(node, 0);
> - if (IS_ERR(*clk)) {
> + ret = PTR_ERR_OR_ZERO(*clk);
> + if (ret) {
> pr_err("timer missing clk");
> - return PTR_ERR(*clk);
> + return ret;
> }
>
> ret = clk_prepare_enable(*clk);