Re: [PATCH] remove hugetlb_instantiation_mutex

[Adam Litke <agl@us.ibm.com>]

On Mon, 2007-07-30 at 15:15 +0800, Zhang, Yanmin wrote:
> On Fri, 2007-07-27 at 11:37 -0500, Adam Litke wrote:
> > Hey... I am amazed at how quickly you came back with a patch for this :)
> > Thanks for looking at it.  Unfortunately there is one show-stopper and I
> > have some reservations (pun definitely intended) with your approach:
> Thanks for your great comments.

Sorry for such a long delay in responding.  I have been pretty busy
lately.

> > First, your patch does not pass the libhugetlbfs test
> > 'alloc-instantiate-race' which was written to tickle the the race which
> > the mutex was introduced to solve.  Your patch works for shared
> > mappings, but not for the private case.
> My testing about private might not be thorough. Function hugetlb_cow has a race
> for multi-thread to fault on the same private page index. But after I fixed it,
> alloc-instantiate-race still failed.
> 
> I tried to google the source code tarball of libhugetlbfs test suite, but couldn't
> find it. Would you like to send me a copy of the test source codes?

http://libhugetlbfs.ozlabs.org/releases/libhugetlbfs-1.2-pre1.tar.gz

The tarball will contain a test called alloc-instantiate-race.  Make
sure to run it in private and shared mode.  Let me know what you find
out.

-- 
Adam Litke - (agl at us.ibm.com)
IBM Linux Technology Center