Re: [Valgrind-developers] Issue with glibc and futex system call w/ FUTEX_WAIT_BITSET

[Petar J. <mip...@gm...>]

> This worries me .. why would the kernel cause the same syscall with
> same arguments, to fail with different errors on different platforms?
> It does not make sense from a software engineering point of view.

It should not worry you. The trick is that the arguments are not the same.
The 6th argument is a "random" value (and not initialized for MIPS).

So, in MIPS case, this is a value on stack, and the 6th argument happens to
be zero value. In that case, kernel will complain about bitset value being
all zeros.

Check here:
http://lxr.free-electrons.com/source/kernel/futex.c?v=2.6.29#L1179

In x86 case, this happens to be previously set non-zero value,

Check here:
http://lxr.free-electrons.com/source/kernel/futex.c?v=2.6.29#L1225

Petar


On Thu, Jul 19, 2012 at 5:45 PM, Julian Seward <js...@ac...> wrote:

>
> Petar, thank you for analysing this, and for the proposed patch.
>
> > Kernel finds different errors with this call on different platforms.
>
> This worries me .. why would the kernel cause the same syscall with
> same arguments, to fail with different errors on different platforms?
> It does not make sense from a software engineering point of view.
>
> Can you maybe send a few more details, showing how/where the kernel
> causes the same call to fail differently?
>
> J
>