2004-06-16 11:34:43

by Sau Dan Lee

Subject: 'uinput' Oops upon select() or poll() on 2.6.7

Re-reporting this bug found 3 weeks ago, because it's still in 2.6.7.
(It was found and reported for 2.6.7-rc1.)

For details, please check the bugzilla entry

http://bugzilla.kernel.org/show_bug.cgi?id=2786


Dmitry Torokhov has already offered a patch on 2004-05-28 12:52.

Vojtech Pavlik said on 2004-05-28 13:59 that he has fixed it in his
input tree.


But the bug is STILL in 2.6.7. I don't understand. What are the -rc*
supposed for, if bugs do not get fixed?



from Dmitry Torokhov 2004-05-28 12:52:

--- 2.6.6/drivers/input/misc/uinput.c 2004-05-11 14:57:25.793252800 -0500
+++ linux-2.6.6/drivers/input/misc/uinput.c 2004-05-28 14:42:36.922273600 -0500
@@ -279,10 +279,11 @@
{
struct uinput_device *udev = file->private_data;

- poll_wait(file, &udev->waitq, wait);
-
- if (udev->head != udev->tail)
- return POLLIN | POLLRDNORM;
+ if (test_bit(UIST_CREATED, &udev->state)) {
+ poll_wait(file, &udev->waitq, wait);
+ if (udev->head != udev->tail)
+ return POLLIN | POLLRDNORM;
+ }

return 0;
}


--
Sau Dan LEE ???u??(Big5) ~{@nJX6X~}(HZ)

E-mail: [email protected]
Home page: http://www.informatik.uni-freiburg.de/~danlee

2004-06-16 11:56:25

by Dmitry Torokhov

Subject: Re: 'uinput' Oops upon select() or poll() on 2.6.7

On Wednesday 16 June 2004 06:34 am, Sau Dan Lee wrote:
>
> But the bug is STILL in 2.6.7. I don't understand. What are the -rc*
> supposed for, if bugs do not get fixed?
>

Hi,

Given the fact that it affects exactly one person and you can avoid it by
fixing your program which is broken in this regard anyway (why would you
poll device before you create it?) I don't think it is that urgent...

I believe it will be merged when Linus does next pull.

--
Dmitry