[PATCH 1/5] Fix spurious segmentation faults

[Nicolas Pitre]

On Fri, 2 Jul 2010, Sebastien Jan wrote:

> This commit (f22c4b02) appears in the kernel-ubuntu tree (in the for-ubuntu-2.6.34 branch for example) (and comes from our kernel-omap4 tree, both on dev.omapzoom.org).
> See this link: http://dev.omapzoom.org/?p=integration/kernel-ubuntu.git;a=commit;h=f22c4b0221b283f05a648d35f19facf80d8a0b23
> 
> >> But there have been some updates to this patch series, which the current patch aims to cover, as it fixes some very ugly side effects.
> >>
> > 
> > Yeah, I think this patch will be helpful, but we need some testing. Do we have
> > launchpad bug tracker of this bug? Ogra and Lee, are you guys aware of this bug?
> > I am not playing with the Maverick rootfs.
> > 
> >> Regarding patch author: I specified Rob as he did the investigations on the segfault we had, and find out the missing parts. We may have to credit Catalin instead/additionaly.
> > 
> > Rob, thanks for fixing this. But we need know more information about the
> > original patches and the status of this patch? Will this patch in mainline or in
> > the OMAP4 upstream queue? That will be much easier for us to review.
> 
> The 1st version of this patch series has been integrated into our omap4 tree (see above commit information).
> Since then, new versions of this series are being discussed on the ARM ML [1]. From what I can read, it seems to be a proposal for the 2.6.36 merge window.
> 
> Rob's patch contains a part of the new patch series, which fixes a memory related issue (which used to cause the segfaults when using git v1.7.0 for example).
> Without this patch, I also used to see segfaults when using dpkg or gcc, making my maverick FS impossible to use for intensive build tasks.
> With this patch, I was able to build the latest kernel image on my omap4 board.

How different is this patch from Catalin's latest?

Did you test with the latest from Catalin?  If not, then why not?

It would really be a good idea to test the latest patch instead of 
trying to fix an older incarnation.  Even more so if the latest still 
exhibits problems.  And in that case it would be really important to 
disclose those segmentation fault problems on the ARM mailing list in 
order to have the original fixed.


Nicolas