memory leak in new_inode_pseudo
Subsystems (4)
syzbot
syzbotHello,
syzbot found the following crash on:
HEAD commit: f49aa1de Merge tag 'for-5.2-rc1-tag' of git://git.kernel.o.. git tree: upstream console output: https://syzkaller.appspot.com/x/log.txt?x=12dccd9ca00000 kernel config: https://syzkaller.appspot.com/x/.config?x=61dd9e15a761691d dashboard link: https://syzkaller.appspot.com/bug?extid=111cb28d9f583693aefa compiler: gcc (GCC) 9.0.0 20181231 (experimental) syz repro: https://syzkaller.appspot.com/x/repro.syz?x=1587839ca00000 C reproducer: https://syzkaller.appspot.com/x/repro.c?x=16fdb38ca00000
IMPORTANT: if you fix the bug, please add the following tag to the commit:
e list of known hosts.
executing program
BUG: memory leak
unreferenced object 0xffff888121bdfd00 (size 632):
comm "syz-executor073", pid 7035, jiffies 4294943279 (age 8.130s)
hex dump (first 32 bytes):
01 00 00 00 02 00 00 00 00 00 00 00 00 00 00 00 ................
80 1e b6 1f 81 88 ff ff 00 00 00 00 00 00 00 00 ................
backtrace:
[<0000000088b15f0d>] kmemleak_alloc_recursive
include/linux/kmemleak.h:55 [inline]
[<0000000088b15f0d>] slab_post_alloc_hook mm/slab.h:439 [inline]
[<0000000088b15f0d>] slab_alloc mm/slab.c:3326 [inline]
[<0000000088b15f0d>] kmem_cache_alloc+0x134/0x270 mm/slab.c:3488
[<000000001a541de3>] sock_alloc_inode+0x1d/0xe0 net/socket.c:252
[<00000000226aca0e>] alloc_inode+0x2c/0xe0 fs/inode.c:226
[<0000000054963a4b>] new_inode_pseudo+0x18/0x70 fs/inode.c:915
[<00000000a577ddf0>] sock_alloc+0x1c/0x90 net/socket.c:575
[<000000000e50448f>] __sock_create+0x8f/0x250 net/socket.c:1394
[<00000000248d5219>] sock_create_kern+0x3b/0x50 net/socket.c:1499
[<0000000081cd440d>] io_uring_get_fd fs/io_uring.c:2997 [inline]
[<0000000081cd440d>] io_uring_create fs/io_uring.c:3080 [inline]
[<0000000081cd440d>] io_uring_setup+0x4ea/0x990 fs/io_uring.c:3128
[<00000000b95ec5c9>] __do_sys_io_uring_setup fs/io_uring.c:3141 [inline]
[<00000000b95ec5c9>] __se_sys_io_uring_setup fs/io_uring.c:3138 [inline]
[<00000000b95ec5c9>] __x64_sys_io_uring_setup+0x1a/0x20
fs/io_uring.c:3138
[<0000000072749d9e>] do_syscall_64+0x76/0x1a0
arch/x86/entry/common.c:301
[<0000000053106e40>] entry_SYSCALL_64_after_hwframe+0x44/0xa9
BUG: memory leak
unreferenced object 0xffff88811fb61e80 (size 64):
comm "syz-executor073", pid 7035, jiffies 4294943279 (age 8.130s)
hex dump (first 32 bytes):
00 00 00 00 81 88 ff ff 88 1e b6 1f 81 88 ff ff ................
88 1e b6 1f 81 88 ff ff 00 00 00 00 00 00 00 00 ................
backtrace:
[<000000008c8eddf3>] kmemleak_alloc_recursive
include/linux/kmemleak.h:55 [inline]
[<000000008c8eddf3>] slab_post_alloc_hook mm/slab.h:439 [inline]
[<000000008c8eddf3>] slab_alloc mm/slab.c:3326 [inline]
[<000000008c8eddf3>] kmem_cache_alloc_trace+0x13d/0x280 mm/slab.c:3553
[<000000005afcf70e>] kmalloc include/linux/slab.h:547 [inline]
[<000000005afcf70e>] sock_alloc_inode+0x44/0xe0 net/socket.c:255
[<00000000226aca0e>] alloc_inode+0x2c/0xe0 fs/inode.c:226
[<0000000054963a4b>] new_inode_pseudo+0x18/0x70 fs/inode.c:915
[<00000000a577ddf0>] sock_alloc+0x1c/0x90 net/socket.c:575
[<000000000e50448f>] __sock_create+0x8f/0x250 net/socket.c:1394
[<00000000248d5219>] sock_create_kern+0x3b/0x50 net/socket.c:1499
[<0000000081cd440d>] io_uring_get_fd fs/io_uring.c:2997 [inline]
[<0000000081cd440d>] io_uring_create fs/io_uring.c:3080 [inline]
[<0000000081cd440d>] io_uring_setup+0x4ea/0x990 fs/io_uring.c:3128
[<00000000b95ec5c9>] __do_sys_io_uring_setup fs/io_uring.c:3141 [inline]
[<00000000b95ec5c9>] __se_sys_io_uring_setup fs/io_uring.c:3138 [inline]
[<00000000b95ec5c9>] __x64_sys_io_uring_setup+0x1a/0x20
fs/io_uring.c:3138
[<0000000072749d9e>] do_syscall_64+0x76/0x1a0
arch/x86/entry/common.c:301
[<0000000053106e40>] entry_SYSCALL_64_after_hwframe+0x44/0xa9
BUG: memory leak
unreferenced object 0xffff88811dd21c08 (size 56):
comm "syz-executor073", pid 7035, jiffies 4294943279 (age 8.130s)
hex dump (first 32 bytes):
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
30 fd bd 21 81 88 ff ff 20 1c d2 1d 81 88 ff ff 0..!.... .......
backtrace:
[<0000000088b15f0d>] kmemleak_alloc_recursive
include/linux/kmemleak.h:55 [inline]
[<0000000088b15f0d>] slab_post_alloc_hook mm/slab.h:439 [inline]
[<0000000088b15f0d>] slab_alloc mm/slab.c:3326 [inline]
[<0000000088b15f0d>] kmem_cache_alloc+0x134/0x270 mm/slab.c:3488
[<00000000dec41212>] kmem_cache_zalloc include/linux/slab.h:732 [inline]
[<00000000dec41212>] lsm_inode_alloc security/security.c:523 [inline]
[<00000000dec41212>] security_inode_alloc+0x33/0xb0
security/security.c:876
[<0000000093ac97c3>] inode_init_always+0x108/0x200 fs/inode.c:168
[<000000006cedb4e8>] alloc_inode+0x49/0xe0 fs/inode.c:233
[<0000000054963a4b>] new_inode_pseudo+0x18/0x70 fs/inode.c:915
[<00000000a577ddf0>] sock_alloc+0x1c/0x90 net/socket.c:575
[<000000000e50448f>] __sock_create+0x8f/0x250 net/socket.c:1394
[<00000000248d5219>] sock_create_kern+0x3b/0x50 net/socket.c:1499
[<0000000081cd440d>] io_uring_get_fd fs/io_uring.c:2997 [inline]
[<0000000081cd440d>] io_uring_create fs/io_uring.c:3080 [inline]
[<0000000081cd440d>] io_uring_setup+0x4ea/0x990 fs/io_uring.c:3128
[<00000000b95ec5c9>] __do_sys_io_uring_setup fs/io_uring.c:3141 [inline]
[<00000000b95ec5c9>] __se_sys_io_uring_setup fs/io_uring.c:3138 [inline]
[<00000000b95ec5c9>] __x64_sys_io_uring_setup+0x1a/0x20
fs/io_uring.c:3138
[<0000000072749d9e>] do_syscall_64+0x76/0x1a0
arch/x86/entry/common.c:301
[<0000000053106e40>] entry_SYSCALL_64_after_hwframe+0x44/0xa9---
This bug is generated by a bot. It may contain errors. See https://goo.gl/tpsmEJ for more information about syzbot. syzbot engineers can be reached at [email protected].
syzbot will keep track of this bug report. See:
https://goo.gl/tpsmEJ#status for how to communicate with syzbot. syzbot can test patches for this bug, for details see:
https://goo.gl/tpsmEJ#testing-patches
PATCH 1/1 io_uring: fix memory leak of UNIX domain socket inode
From: Eric Biggers <[email protected]>
Opening and closing an io_uring instance leaks a UNIX domain socket inode. This is because the ->file of the io_uring instance's internal UNIX domain socket is set to point to the io_uring file, but then sock_release() sees the non-NULL ->file and assumes the inode reference is held by the file so doesn't call iput(). That's not the case here, since the reference is still meant to be held by the socket; the actual inode of the io_uring file is different.
Fix this leak by NULL-ing out ->file before releasing the socket.
--- fs/io_uring.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
fs/io_uring.c +3 -1
@@ -2777,8 +2777,10 @@ static void io_ring_ctx_free(struct io_ring_ctx *ctx)io_eventfd_unregister (ctx);#if defined(CONFIG_UNIX ) -if (ctx->ring_sock) +if (ctx->ring_sock) { + ctx->ring_sock->file =NULL ;/* so that iput() is called */ sock_release (ctx->ring_sock); + }#endif io_mem_free (ctx->sq_ring);
Signature · 1 line
2.22.0.rc2.383.gf4fbbf30c2-goog
Re: [PATCH] io_uring: fix memory leak of UNIX domain socket inode
Quoted 11 lines · Eric Biggers
From: Eric Biggers <[email protected]>
Opening and closing an
io_uringinstance leaks a UNIX domain socket inode. This is because the ->file of theio_uringinstance's internal UNIX domain socket is set to point to theio_uringfile, but thensock_release()sees the non-NULL ->file and assumes the inode reference is held by the file so doesn't calliput(). That's not the case here, since the reference is still meant to be held by the socket; the actual inode of theio_uringfile is different.Fix this leak by NULL-ing out ->file before releasing the socket.
Thanks, applied.
Signature · 1 line
Jens Axboe