Re: [RFC][PATCH] ipc: Use private shmem or hugetlbfs inodes for shm segments.

[Stephen Smalley]

On 07/23/2015 08:11 PM, Dave Chinner wrote:
> On Thu, Jul 23, 2015 at 12:28:33PM -0400, Stephen Smalley wrote:
>> The shm implementation internally uses shmem or hugetlbfs inodes
>> for shm segments.  As these inodes are never directly exposed to
>> userspace and only accessed through the shm operations which are
>> already hooked by security modules, mark the inodes with the
>> S_PRIVATE flag so that inode security initialization and permission
>> checking is skipped.
>>
>> This was motivated by the following lockdep warning:
>> ===================================================
>> [ INFO: possible circular locking dependency detected ]
>> 4.2.0-0.rc3.git0.1.fc24.x86_64+debug #1 Tainted: G        W
>> -------------------------------------------------------
>> httpd/1597 is trying to acquire lock:
>> (&ids->rwsem){+++++.}, at: [<ffffffff81385354>] shm_close+0x34/0x130
>> (&mm->mmap_sem){++++++}, at: [<ffffffff81386bbb>] SyS_shmdt+0x4b/0x180
>>       [<ffffffff81109a07>] lock_acquire+0xc7/0x270
>>       [<ffffffff81217baa>] __might_fault+0x7a/0xa0
>>       [<ffffffff81284a1e>] filldir+0x9e/0x130
>>       [<ffffffffa019bb08>] xfs_dir2_block_getdents.isra.12+0x198/0x1c0 [xfs]
>>       [<ffffffffa019c5b4>] xfs_readdir+0x1b4/0x330 [xfs]
>>       [<ffffffffa019f38b>] xfs_file_readdir+0x2b/0x30 [xfs]
>>       [<ffffffff812847e7>] iterate_dir+0x97/0x130
>>       [<ffffffff81284d21>] SyS_getdents+0x91/0x120
>>       [<ffffffff81871d2e>] entry_SYSCALL_64_fastpath+0x12/0x76
>>       [<ffffffff81109a07>] lock_acquire+0xc7/0x270
>>       [<ffffffff81101e97>] down_read_nested+0x57/0xa0
>>       [<ffffffffa01b0e57>] xfs_ilock+0x167/0x350 [xfs]
>>       [<ffffffffa01b10b8>] xfs_ilock_attr_map_shared+0x38/0x50 [xfs]
>>       [<ffffffffa014799d>] xfs_attr_get+0xbd/0x190 [xfs]
>>       [<ffffffffa01c17ad>] xfs_xattr_get+0x3d/0x70 [xfs]
>>       [<ffffffff8129962f>] generic_getxattr+0x4f/0x70
>>       [<ffffffff8139ba52>] inode_doinit_with_dentry+0x162/0x670
>>       [<ffffffff8139cf69>] sb_finish_set_opts+0xd9/0x230
>>       [<ffffffff8139d66c>] selinux_set_mnt_opts+0x35c/0x660
>>       [<ffffffff8139ff97>] superblock_doinit+0x77/0xf0
>>       [<ffffffff813a0020>] delayed_superblock_init+0x10/0x20
>>       [<ffffffff81272d23>] iterate_supers+0xb3/0x110
>>       [<ffffffff813a4e5f>] selinux_complete_init+0x2f/0x40
>>       [<ffffffff813b47a3>] security_load_policy+0x103/0x600
>>       [<ffffffff813a6901>] sel_write_load+0xc1/0x750
>>       [<ffffffff8126e817>] __vfs_write+0x37/0x100
>>       [<ffffffff8126f229>] vfs_write+0xa9/0x1a0
>>       [<ffffffff8126ff48>] SyS_write+0x58/0xd0
>>       [<ffffffff81871d2e>] entry_SYSCALL_64_fastpath+0x12/0x76
>>       [<ffffffff81109a07>] lock_acquire+0xc7/0x270
>>       [<ffffffff8186de8f>] mutex_lock_nested+0x7f/0x3e0
>>       [<ffffffff8139b9a9>] inode_doinit_with_dentry+0xb9/0x670
>>       [<ffffffff8139bf7c>] selinux_d_instantiate+0x1c/0x20
>>       [<ffffffff813955f6>] security_d_instantiate+0x36/0x60
>>       [<ffffffff81287c34>] d_instantiate+0x54/0x70
>>       [<ffffffff8120111c>] __shmem_file_setup+0xdc/0x240
>>       [<ffffffff81201290>] shmem_file_setup+0x10/0x20
>>       [<ffffffff813856e0>] newseg+0x290/0x3a0
>>       [<ffffffff8137e278>] ipcget+0x208/0x2d0
>>       [<ffffffff81386074>] SyS_shmget+0x54/0x70
>>       [<ffffffff81871d2e>] entry_SYSCALL_64_fastpath+0x12/0x76
>>       [<ffffffff81108df8>] __lock_acquire+0x1a78/0x1d00
>>       [<ffffffff81109a07>] lock_acquire+0xc7/0x270
>>       [<ffffffff8186efba>] down_write+0x5a/0xc0
>>       [<ffffffff81385354>] shm_close+0x34/0x130
>>       [<ffffffff812203a5>] remove_vma+0x45/0x80
>>       [<ffffffff81222a30>] do_munmap+0x2b0/0x460
>>       [<ffffffff81386c25>] SyS_shmdt+0xb5/0x180
>>       [<ffffffff81871d2e>] entry_SYSCALL_64_fastpath+0x12/0x76
> 
> That's a completely screwed up stack trace. There are *4* syscall
> entry points with 4 separate, unrelated syscall chains on that
> stack trace, all starting at the same address. How is this a valid
> stack trace and not a lockdep bug of some kind?

Sorry, I mangled it when I tried to reformat it from Morten Steven's
original report.  Fixed in v2.



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