Hello Folks, This is my first post to the list. I have been investigating porting software to the Linaro system (on OMAP 4) and I came across this issue. Basically, the kernel appears to report that it does not provide Robust Mutex support? This would seem to be some kind of kernel build/configuration issue?
The reason I'm asking is that our software system checks for (and uses) robust mutex features and I had originally built everything on the Ubuntu 10.10 release for Pandaboard only to discover this problem. I tried an 11.04 alpha 3 install on my Pandaboard and robust mutex appeared to work, so I transferred all our work to that... But now again (after a few updates on that system) robust mutex seems to be again turned off somewhere.
Any suggestions on how to resolve? Should one expect robust mutex support?
-Sincerely, Todd Pack
On Friday 25 March 2011 20:22:06 Pack, Robert wrote:
Any suggestions on how to resolve? Should one expect robust mutex
support?
There was a patch to support for robust futexes on SMP kernel last year, from Anton Vorontsov:
http://lists.infradead.org/pipermail/linux-arm-kernel/2010-June/017374.html
There was some discussion on this patch, but I think the work was never completed to address the concerns.
Maybe you were running with this patch at an earlier time?
Arnd
On Fri, Mar 25, 2011 at 08:57:50PM +0100, Arnd Bergmann wrote:
On Friday 25 March 2011 20:22:06 Pack, Robert wrote:
Any suggestions on how to resolve? Should one expect robust mutex
support?
There was a patch to support for robust futexes on SMP kernel last year, from Anton Vorontsov:
http://lists.infradead.org/pipermail/linux-arm-kernel/2010-June/017374.html
There was some discussion on this patch, but I think the work was never completed to address the concerns.
Maybe you were running with this patch at an earlier time?
The patch was waiting for Catalin Marinas' "Remove the domain switching on ARMv6k/v7 CPUs" patch to hit Linus' tree. Now that the patch is in, the robust mutex support should be OK with the patch above (but a minor change[1] would be needed to get it into mainline).
Thanks,
[1] http://lists.infradead.org/pipermail/linux-arm-kernel/2010-June/017395.html
On Fri, Mar 25, 2011 at 08:57:50PM +0100, Arnd Bergmann wrote:
On Friday 25 March 2011 20:22:06 Pack, Robert wrote:
Any suggestions on how to resolve? Should one expect robust mutex
support?
There was a patch to support for robust futexes on SMP kernel last year, from Anton Vorontsov:
http://lists.infradead.org/pipermail/linux-arm-kernel/2010-June/017374.html
There was some discussion on this patch, but I think the work was never completed to address the concerns.
Maybe you were running with this patch at an earlier time?
The patch was waiting for Catalin Marinas' "Remove the domain switching on ARMv6k/v7 CPUs" patch to hit Linus' tree. Now that the patch is in, the robust mutex support should be OK with the patch above (but a minor change[1] would be needed to get it into mainline).
Thanks,
[1] http://lists.infradead.org/pipermail/linux-arm-kernel/2010-June/017395.html
--------------------------------------------------------------------
Hello Folks, Thanks for the help it was very informative.
It sounds as if we're almost there and you folks have already done most of the hard work. How can someone help to get this patch into the pipeline? It is basically a blocker for my work on the ARM at this point. I'd love to see this make it in before Linaro 11.05 comes out if at all possible. What is the right way to pursue it?
-Sincerely, Todd Pack
On Sun, 27 Mar 2011, Pack, Robert Todd wrote:
The patch was waiting for Catalin Marinas' "Remove the domain switching on ARMv6k/v7 CPUs" patch to hit Linus' tree. Now that the patch is in, the robust mutex support should be OK with the patch above (but a minor change[1] would be needed to get it into mainline).
Thanks,
[1] http://lists.infradead.org/pipermail/linux-arm-kernel/2010-June/017395.html
Hello Folks, Thanks for the help it was very informative.
It sounds as if we're almost there and you folks have already done most of the hard work. How can someone help to get this patch into the pipeline? It is basically a blocker for my work on the ARM at this point. I'd love to see this make it in before Linaro 11.05 comes out if at all possible. What is the right way to pursue it?
I'll take care of including it into the linaro 11.05 kernel. Someone (Anton?) should also push it into mainline via Russell.
Nicolas
linaro-kernel@lists.linaro.org