| Message ID | 6fbb3819-a348-4cc3-a1d0-951ca1c380d6@moroto.mountain (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
| Series | [v2,led] leds: trigger: netdev: uninitialized variable in netdev_trig_activate() | expand |
| Context | Check | Description |
|---|---|---|
| netdev/tree_selection | success | Not a local patch |
On Tue, Jun 13, 2023 at 09:46:53AM +0300, Dan Carpenter wrote: > The qca8k_cled_hw_control_get() function which implements ->hw_control_get > sets the appropriate bits but does not clear them. This leads to an > uninitialized variable bug. Fix this by setting mode to zero at the > start. > > Fixes: e0256648c831 ("net: dsa: qca8k: implement hw_control ops") > Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org> > --- > v2: In the original patch I fixed qca8k_cled_hw_control_get() instead > so that patch went to netdev instead of to the led subsystem. I think his actually needs to be merged via netdev. e0256648c831 ("net: dsa: qca8k: implement hw_control ops") is in net-next/main. I don't see it in leds/master, leds/for-leds-next. Also, git blame shows mode was added by 0316cc5629d1 ("leds: trigger: netdev: init mode if hw control already active") which also appears only to be in net-next/main. A lot of these LED patches were merged via netdev because they are cross subsystem. Reviewed-by: Andrew Lunn <andrew@lunn.ch> Andrew
On Wed, Jun 14, 2023 at 02:51:59AM +0200, Andrew Lunn wrote: > I think his actually needs to be merged via netdev. e0256648c831 > ("net: dsa: qca8k: implement hw_control ops") is in net-next/main. I > don't see it in leds/master, leds/for-leds-next. Also, git blame shows > mode was added by 0316cc5629d1 ("leds: trigger: netdev: init mode if > hw control already active") which also appears only to be in > net-next/main. > > A lot of these LED patches were merged via netdev because they are > cross subsystem. Ok. Thanks, Andrew. In that case, let me resend. I just used get_maintainer.pl and it didn't add the necessary CC list. Plus the netdev scripts won't run if you don't have the correct subject... regards, dan carpenter
diff --git a/drivers/leds/trigger/ledtrig-netdev.c b/drivers/leds/trigger/ledtrig-netdev.c index b0a6f2749552..2311dae7f070 100644 --- a/drivers/leds/trigger/ledtrig-netdev.c +++ b/drivers/leds/trigger/ledtrig-netdev.c @@ -445,7 +445,7 @@ static void netdev_trig_work(struct work_struct *work) static int netdev_trig_activate(struct led_classdev *led_cdev) { struct led_netdev_data *trigger_data; - unsigned long mode; + unsigned long mode = 0; struct device *dev; int rc;
The qca8k_cled_hw_control_get() function which implements ->hw_control_get sets the appropriate bits but does not clear them. This leads to an uninitialized variable bug. Fix this by setting mode to zero at the start. Fixes: e0256648c831 ("net: dsa: qca8k: implement hw_control ops") Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org> --- v2: In the original patch I fixed qca8k_cled_hw_control_get() instead so that patch went to netdev instead of to the led subsystem. https://lore.kernel.org/all/5dff3719-f827-45b6-a0d3-a00efed1099b@moroto.mountain/ Fixing it here is a more reliable way to do it. drivers/leds/trigger/ledtrig-netdev.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)