Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1673491 > unrolled thread

[PATCH] gpiolib: fix filtering out unwanted events

Back to article view | Back to linux.kernel

Contents

  [PATCH] gpiolib: fix filtering out unwanted events Bartosz Golaszewski <brgl@bgdev.pl> - 2017-06-23 13:50 +0200
    [PATCH] gpiolib: fix filtering out unwanted events Bartosz Golaszewski <brgl@bgdev.pl> - 2017-06-23 13:50 +0200
      Re: [PATCH] gpiolib: fix filtering out unwanted events Linus Walleij <linus.walleij@linaro.org> - 2017-06-29 11:40 +0200
        Re: [PATCH] gpiolib: fix filtering out unwanted events Bartosz Golaszewski <brgl@bgdev.pl> - 2017-06-29 13:10 +0200
          Re: [PATCH] gpiolib: fix filtering out unwanted events Linus Walleij <linus.walleij@linaro.org> - 2017-06-29 15:20 +0200
    Re: [PATCH] gpiolib: fix filtering out unwanted events Bartosz Golaszewski <brgl@bgdev.pl> - 2017-06-23 16:10 +0200

#1673491 — [PATCH] gpiolib: fix filtering out unwanted events

Hi Linux,

I've just fixed an issue in libgpiod which accidentally also led to
filtering out unwanted events in user space instead of gpiolib kernel
code. I then noticed that programs don't behave correctly and tracked
it down to an issue in gpiolib.c.

Please take a look at the fix. The bug has been there for some time and
it's quite late in the release cycle, but I believe it should still go
in v4.12 since currently the character device interface doesn't do what
it promises to.

Bartosz Golaszewski (1):
  gpiolib: fix filtering out unwanted events

 drivers/gpio/gpiolib.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

-- 
2.9.3

[toc] | [next] | [standalone]

#1673492

GPIOEVENT_REQUEST_BOTH_EDGES is not a single flag, but a binary OR of
GPIOEVENT_REQUEST_RISING_EDGE and GPIOEVENT_REQUEST_FALLING_EDGE.

The expression 'le->eflags & GPIOEVENT_REQUEST_BOTH_EDGES' we'll get
evaluated to true even if only one event type was requested.

Fix it by checking both RISING & FALLING flags explicitly.

Signed-off-by: Bartosz Golaszewski <brgl@bgdev.pl>
---
 drivers/gpio/gpiolib.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/gpio/gpiolib.c b/drivers/gpio/gpiolib.c
index 5db4413..a42a1ee 100644
--- a/drivers/gpio/gpiolib.c
+++ b/drivers/gpio/gpiolib.c
@@ -708,7 +708,8 @@ static irqreturn_t lineevent_irq_thread(int irq, void *p)
 
 	ge.timestamp = ktime_get_real_ns();
 
-	if (le->eflags & GPIOEVENT_REQUEST_BOTH_EDGES) {
+	if (le->eflags & GPIOEVENT_REQUEST_RISING_EDGE
+	    && le->eflags & GPIOEVENT_REQUEST_FALLING_EDGE) {
 		int level = gpiod_get_value_cansleep(le->desc);
 
 		if (level)
-- 
2.9.3

[toc] | [prev] | [next] | [standalone]

#1677559

On Fri, Jun 23, 2017 at 1:45 PM, Bartosz Golaszewski <brgl@bgdev.pl> wrote:

> GPIOEVENT_REQUEST_BOTH_EDGES is not a single flag, but a binary OR of
> GPIOEVENT_REQUEST_RISING_EDGE and GPIOEVENT_REQUEST_FALLING_EDGE.
>
> The expression 'le->eflags & GPIOEVENT_REQUEST_BOTH_EDGES' we'll get
> evaluated to true even if only one event type was requested.
>
> Fix it by checking both RISING & FALLING flags explicitly.
>
> Signed-off-by: Bartosz Golaszewski <brgl@bgdev.pl>

Patch applied for fixes.

Yours,
Linus Walleij

[toc] | [prev] | [next] | [standalone]

#1677613

2017-06-29 11:33 GMT+02:00 Linus Walleij <linus.walleij@linaro.org>:
> On Fri, Jun 23, 2017 at 1:45 PM, Bartosz Golaszewski <brgl@bgdev.pl> wrote:
>
>> GPIOEVENT_REQUEST_BOTH_EDGES is not a single flag, but a binary OR of
>> GPIOEVENT_REQUEST_RISING_EDGE and GPIOEVENT_REQUEST_FALLING_EDGE.
>>
>> The expression 'le->eflags & GPIOEVENT_REQUEST_BOTH_EDGES' we'll get
>> evaluated to true even if only one event type was requested.
>>
>> Fix it by checking both RISING & FALLING flags explicitly.
>>
>> Signed-off-by: Bartosz Golaszewski <brgl@bgdev.pl>
>
> Patch applied for fixes.
>
> Yours,
> Linus Walleij

Thanks. I think this should also go to linux-stable (e.g. 4.9 LTS is
affected). Will you forward it or should I do it myself?

Thanks,
Bartosz

[toc] | [prev] | [next] | [standalone]

#1677746

On Thu, Jun 29, 2017 at 1:03 PM, Bartosz Golaszewski <brgl@bgdev.pl> wrote:
> 2017-06-29 11:33 GMT+02:00 Linus Walleij <linus.walleij@linaro.org>:
>> On Fri, Jun 23, 2017 at 1:45 PM, Bartosz Golaszewski <brgl@bgdev.pl> wrote:
>>
>>> GPIOEVENT_REQUEST_BOTH_EDGES is not a single flag, but a binary OR of
>>> GPIOEVENT_REQUEST_RISING_EDGE and GPIOEVENT_REQUEST_FALLING_EDGE.
>>>
>>> The expression 'le->eflags & GPIOEVENT_REQUEST_BOTH_EDGES' we'll get
>>> evaluated to true even if only one event type was requested.
>>>
>>> Fix it by checking both RISING & FALLING flags explicitly.
>>>
>>> Signed-off-by: Bartosz Golaszewski <brgl@bgdev.pl>
>>
>> Patch applied for fixes.
>>
>> Yours,
>> Linus Walleij
>
> Thanks. I think this should also go to linux-stable (e.g. 4.9 LTS is
> affected). Will you forward it or should I do it myself?

I have added the stable tag so it should JustWork(TM).

Yours,
Linus Walleij

[toc] | [prev] | [next] | [standalone]

#1673578

2017-06-23 13:45 GMT+02:00 Bartosz Golaszewski <brgl@bgdev.pl>:
> Hi Linux,
>

s/Linux/Linus :)

Bartosz

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web