[PATCH] fix length bug in affinity mask displays (try 3)

[Dmitry V. Levin]

Hi,

On Fri, Jan 08, 2016 at 11:28:33AM -0500, Joe Korty wrote:
[...]
> Hi Dmitry,
> Thanks for replying.  I have had a hard time posting to
> the list, so it is nice to know that the email actually
> got through to someone on the third try ...

I suppose it has something to do with anti-spam filters
on the sourceforge.net side.

> PS: not sure that the 'set of integers' idea is correct in
> a big-endian, 64-bit environment.

Since it's implemented using the same algorithm that's used for printing
fd_set arguments of select/pselect6 syscalls, it should be correct.
Anyway, I've added sched_xetaffinity.test along with this change.
This tests passes on all platforms I checked (including s390x),
so I suppose it is correct.

> I know the version I
> posted is correct in any mix of 32 and 64 bit environments
> and big and little endian environments.

Yes, it was correct, with one exception: tracing of 32-bit executables
by 64-bit strace fails to print cpuset properly if sizeof(long) > cpusetsize.


-- 
ldv
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 819 bytes
Desc: not available
URL: <http://lists.strace.io/pipermail/strace-devel/attachments/20160109/0a880f0d/attachment.bin>