linux-kernel - [BUG] new copy_hugetlb_page

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [day] [month] [year] [list]

Date:	Thu, 17 Jul 2014 18:23:45 +0200
From:	Guillaume Morin <guillaume@...infr.org>
To:	Naoya Horiguchi <n-horiguchi@...jp.nec.com>,
	Hugh Dickins <hughd@...gle.com>
Cc:	linux-kernel@...r.kernel.org, stable@...r.kernel.org
Subject: [BUG] new copy_hugetlb_page_range() causing crashes

Naoya, Hugh,

I am seeing lots of crashes with the new copy_hugetlb_page_range() code
added by 4a705fef986231a3e7a6b1a6d3c37025f021f49f for some set of
programs.

Specifically, I am running some test programs which use huge pages for
malloc (through libhugetlbfs with HUGETLB_MORECORE=yes and
HUGETLB_NO_PREFAULT= set) that also do fork() a lot.  The crashes are
very diverse: assertion failures in malloc() or the python GC code, some
SIGSEGV as well.

I started observing these crashes with 3.4.98 and 3.14.12 which just got
a backport of the patch above (as
2bcdd4933ff4dc46445dcae93cb37c648283b782 in the stable branch).  The 3.4
and 3.14 patches are identical to the upstream commit so that's not a patch
backport issue.

If I revert only 2bcdd4933ff4dc46445dcae93cb37c648283b782 in my 3.4
tree, the crashes disappear right away and everything is stable. 

-- 
Guillaume Morin <guillaume@...infr.org>
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/