Re: KASAN: use-after-free Read in io_uring_setup

[Jens Axboe]

On 10/28/19 1:22 AM, syzbot wrote:
> Hello,
> 
> syzbot found the following crash on:
> 
> HEAD commit:    5a1e843c Merge tag 'mips_fixes_5.4_3' of git://git.kernel...
> git tree:       upstream
> console output: https://syzkaller.appspot.com/x/log.txt?x=10e2001f600000
> kernel config:  https://syzkaller.appspot.com/x/.config?x=420126a10fdda0f1
> dashboard link: https://syzkaller.appspot.com/bug?extid=6f03d895a6cd0d06187f
> compiler:       gcc (GCC) 9.0.0 20181231 (experimental)
> syz repro:      https://syzkaller.appspot.com/x/repro.syz?x=11d4fa97600000
> 
> IMPORTANT: if you fix the bug, please add the following tag to the commit:
> Reported-by: syzbot+6f03d895a6cd0d06187f@xxxxxxxxxxxxxxxxxxxxxxxxx
> 
> ==================================================================
> BUG: KASAN: use-after-free in io_uring_create fs/io_uring.c:3842 [inline]
> BUG: KASAN: use-after-free in io_uring_setup+0x1877/0x18c0
> fs/io_uring.c:3881
> Read of size 8 at addr ffff888082284048 by task syz-executor.5/11342
> 
> CPU: 1 PID: 11342 Comm: syz-executor.5 Not tainted 5.4.0-rc4+ #0
> Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS
> Google 01/01/2011
> Call Trace:
>    __dump_stack lib/dump_stack.c:77 [inline]
>    dump_stack+0x172/0x1f0 lib/dump_stack.c:113
>    print_address_description.constprop.0.cold+0xd4/0x30b mm/kasan/report.c:374
>    __kasan_report.cold+0x1b/0x41 mm/kasan/report.c:506
>    kasan_report+0x12/0x20 mm/kasan/common.c:634
>    __asan_report_load8_noabort+0x14/0x20 mm/kasan/generic_report.c:132
>    io_uring_create fs/io_uring.c:3842 [inline]
>    io_uring_setup+0x1877/0x18c0 fs/io_uring.c:3881
>    __do_sys_io_uring_setup fs/io_uring.c:3894 [inline]
>    __se_sys_io_uring_setup fs/io_uring.c:3891 [inline]
>    __x64_sys_io_uring_setup+0x54/0x80 fs/io_uring.c:3891
>    do_syscall_64+0xfa/0x760 arch/x86/entry/common.c:290
>    entry_SYSCALL_64_after_hwframe+0x49/0xbe
> RIP: 0033:0x459f39
> Code: ad b6 fb ff c3 66 2e 0f 1f 84 00 00 00 00 00 66 90 48 89 f8 48 89 f7
> 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff
> ff 0f 83 7b b6 fb ff c3 66 2e 0f 1f 84 00 00 00 00
> RSP: 002b:00007f313e126c78 EFLAGS: 00000246 ORIG_RAX: 00000000000001a9
> RAX: ffffffffffffffda RBX: 0000000000000002 RCX: 0000000000459f39
> RDX: 0000000000000000 RSI: 00000000200005c0 RDI: 000000040000000e
> RBP: 000000000075bf20 R08: 0000000000000000 R09: 0000000000000000
> R10: 0000000000000000 R11: 0000000000000246 R12: 00007f313e1276d4
> R13: 00000000004c1512 R14: 00000000004d4da8 R15: 00000000ffffffff
> 
> Allocated by task 11342:
>    save_stack+0x23/0x90 mm/kasan/common.c:69
>    set_track mm/kasan/common.c:77 [inline]
>    __kasan_kmalloc mm/kasan/common.c:510 [inline]
>    __kasan_kmalloc.constprop.0+0xcf/0xe0 mm/kasan/common.c:483
>    kasan_kmalloc+0x9/0x10 mm/kasan/common.c:524
>    kmem_cache_alloc_trace+0x158/0x790 mm/slab.c:3550
>    kmalloc include/linux/slab.h:556 [inline]
>    kzalloc include/linux/slab.h:690 [inline]
>    io_ring_ctx_alloc fs/io_uring.c:393 [inline]
>    io_uring_create fs/io_uring.c:3811 [inline]
>    io_uring_setup+0xec6/0x18c0 fs/io_uring.c:3881
>    __do_sys_io_uring_setup fs/io_uring.c:3894 [inline]
>    __se_sys_io_uring_setup fs/io_uring.c:3891 [inline]
>    __x64_sys_io_uring_setup+0x54/0x80 fs/io_uring.c:3891
>    do_syscall_64+0xfa/0x760 arch/x86/entry/common.c:290
>    entry_SYSCALL_64_after_hwframe+0x49/0xbe
> 
> Freed by task 11335:
>    save_stack+0x23/0x90 mm/kasan/common.c:69
>    set_track mm/kasan/common.c:77 [inline]
>    kasan_set_free_info mm/kasan/common.c:332 [inline]
>    __kasan_slab_free+0x102/0x150 mm/kasan/common.c:471
>    kasan_slab_free+0xe/0x10 mm/kasan/common.c:480
>    __cache_free mm/slab.c:3425 [inline]
>    kfree+0x10a/0x2c0 mm/slab.c:3756
>    io_ring_ctx_free fs/io_uring.c:3552 [inline]
>    io_ring_ctx_wait_and_kill+0x4d7/0x6c0 fs/io_uring.c:3592
>    io_uring_release+0x42/0x50 fs/io_uring.c:3600
>    __fput+0x2ff/0x890 fs/file_table.c:280
>    ____fput+0x16/0x20 fs/file_table.c:313
>    task_work_run+0x145/0x1c0 kernel/task_work.c:113
>    tracehook_notify_resume include/linux/tracehook.h:188 [inline]
>    exit_to_usermode_loop+0x316/0x380 arch/x86/entry/common.c:163
>    prepare_exit_to_usermode arch/x86/entry/common.c:194 [inline]
>    syscall_return_slowpath arch/x86/entry/common.c:274 [inline]
>    do_syscall_64+0x65f/0x760 arch/x86/entry/common.c:300
>    entry_SYSCALL_64_after_hwframe+0x49/0xbe
> 
> The buggy address belongs to the object at ffff888082284000
>    which belongs to the cache kmalloc-2k of size 2048
> The buggy address is located 72 bytes inside of
>    2048-byte region [ffff888082284000, ffff888082284800)
> The buggy address belongs to the page:
> page:ffffea000208a100 refcount:1 mapcount:0 mapping:ffff8880aa400e00
> index:0x0
> flags: 0x1fffc0000000200(slab)
> raw: 01fffc0000000200 ffffea0002a1bc88 ffffea00023fa248 ffff8880aa400e00
> raw: 0000000000000000 ffff888082284000 0000000100000001 0000000000000000
> page dumped because: kasan: bad access detected
> 
> Memory state around the buggy address:
>    ffff888082283f00: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
>    ffff888082283f80: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
>> ffff888082284000: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
>                                                 ^
>    ffff888082284080: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
>    ffff888082284100: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
> ==================================================================

Interesting, looks like a malicious case that attempts to close the
fd as soon as it's installed. As a result of that, the rest of the
setup will be done on a ring that's already torn down. The below should
fix that.

Totally untested, haven't tried the reproducer yet.


diff --git a/fs/io_uring.c b/fs/io_uring.c
index ba1431046c98..749637ca9cf7 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -3821,6 +3821,12 @@ static int io_uring_create(unsigned entries, struct io_uring_params *p)
 	ctx->account_mem = account_mem;
 	ctx->user = user;
 
+	/*
+	 * Grab an initial reference to the ctx, so we ensure we have one
+	 * by the time the fd is installed.
+	 */
+	percpu_ref_get(&ctx->refs);
+
 	ret = io_allocate_scq_urings(ctx, p);
 	if (ret)
 		goto err;
@@ -3851,8 +3857,10 @@ static int io_uring_create(unsigned entries, struct io_uring_params *p)
 	p->cq_off.cqes = offsetof(struct io_rings, cqes);
 
 	p->features = IORING_FEAT_SINGLE_MMAP;
+	percpu_ref_put(&ctx->refs);
 	return ret;
 err:
+	percpu_ref_put(&ctx->refs);
 	io_ring_ctx_wait_and_kill(ctx);
 	return ret;
 }

-- 
Jens Axboe