On Wednesday 16 May 2012 14:23:32 Khem Raj wrote:
> On Wed, May 16, 2012 at 11:00 AM, Eugene Rudoy wrote:
> > After taking a look at what glibc does, I would suggest the following
> > (not yet tested) fix (s. attached patch)
>
> Looks ok. send with sign-offs and preferably a testcase now that you
On Wed, May 16, 2012 at 11:00 AM, Eugene Rudoy
wrote:
>
> After taking a look at what glibc does, I would suggest the following (not
> yet tested) fix (s. attached patch)
>
Looks ok. send with sign-offs and preferably a testcase now that you have one.
> Best regards,
> Gene
_
Hi Natanael,
On 16 May 2012 15:55, Natanael Copa wrote:
>
> I think you are absolutely right. We bumped into this issue when
> upgrading to glib-2.32.
>
believe it or not, I also ran into the issue after updating glib to 2.32 ;-)
> I suggest this patch but it will require kernel2.6.27+ kernel:
On Wed, May 16, 2012 at 3:55 PM, Natanael Copa wrote:
>> I believe the implementation of eventfd recently added to uClibc (s. [1],
>> [2]) is incorrect.
..
> I think you are absolutely right. We bumped into this issue when
> upgrading to glib-2.32.
I should probably mention that all glib applica
On Wed, May 16, 2012 at 2:50 PM, Eugene Rudoy wrote:
> Hi,
>
> I believe the implementation of eventfd recently added to uClibc (s. [1],
> [2]) is incorrect. It incorrectly assumes eventfd takes two arguments
> whereas in reality it expects just one. It's eventfd2 which expects two
> arguments. Fu
Hi,
I believe the implementation of eventfd recently added to uClibc (s. [1],
[2]) is incorrect. It incorrectly assumes eventfd takes two arguments
whereas in reality it expects just one. It's eventfd2 which expects two
arguments. Furthermore it doesn't properly support kernel versions which do
pr