On Thu, Apr 26, 2018 at 09:54:56AM +0800, Eryu Guan wrote: > On Wed, Apr 25, 2018 at 07:53:41AM -0400, Brian Foster wrote: > > On Wed, Apr 25, 2018 at 11:22:21AM +0800, Eryu Guan wrote: > > > On Fri, Apr 06, 2018 at 10:18:15AM -0400, Brian Foster wrote: > > > > The XFS filestreams allocator caches dir inode -> agno mappings in > > > > an MRU mechanism that holds elements in memory for an amount of time > > > > and then cleans up expired elements in the background. The elements > > > > typically held inode pointers without holding a reference to the > > > > associated inode. This means that if the inode is reclaimed before > > > > an expired entry is cleaned up, the MRU reaper can access freed > > > > memory and cause a panic. > > > > > > > > Test for this problem by performing continuous filestreams > > > > allocations under short-lived parent directory inodes. This will > > > > produce KASAN use-after-free splats if enabled during the test. > > > > > > > > Signed-off-by: Brian Foster <bfoster@xxxxxxxxxx> > > > > --- > > > > > > > > This test reproduces the problem described in this[1] thread. It's > > > > XFS-specific because of the filestream option and specific geometry used > > > > to format the scratch device. > > > > > > > > Brian > > > > > > > > [1] https://marc.info/?l=linux-xfs&m=152293031029161&w=2 > > > > > > From above thread, it seems that we don't need the kernel change > > > anymore, and the test itself would trigger dmesg check failure when > > > testing on kernel with kasan enabled, right? > > > > > > > Yep... > > > > > But I've looped the test for 200 times and it all passed without > > > triggering any KASAN warnings, kernel is v4.17-rc2. > > > > > > > The kernel fix ended up being a patch from Christoph. It looks like it > > made it into v4.17-rc1 as commit 7fcd3efa1e ("xfs: remove filestream > > item xfs_inode reference"). Could you perhaps try an older kernel or one > > with that patch reverted? > > Sure, I'll try reverting that patch. Yeah, I hit KASAN warning quite quickly after reverting that patch. Thanks! Eryu -- To unsubscribe from this list: send the line "unsubscribe fstests" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
