Re: KASAN: use-after-free Write in xfrm_hash_rebuild

From: Dmitry Vyukov
Date: Tue Jul 02 2019 - 02:44:07 EST


On Tue, Jul 2, 2019 at 8:38 AM Hillf Danton <hdanton@xxxxxxxx> wrote:
>
>
> On Wed, 26 Jun 2019 20:59:05 -0700 (PDT)
> > syzbot has found a reproducer for the following crash on:
> >
> > HEAD commit: 249155c2 Merge branch 'parisc-5.2-4' of git://git.kernel.o..
> > git tree: upstream
> > console output: https://syzkaller.appspot.com/x/log.txt?x=10f017c3a00000
> > kernel config: https://syzkaller.appspot.com/x/.config?x=9a31528e58cc12e2
> > dashboard link: https://syzkaller.appspot.com/bug?extid=0165480d4ef07360eeda
> > compiler: clang version 9.0.0 (/home/glider/llvm/clang 80fee25776c2fb61e74c1ecb1a523375c2500b69)
> > syz repro: https://syzkaller.appspot.com/x/repro.syz?x=16cf37c3a00000
> >
> > IMPORTANT: if you fix the bug, please add the following tag to the commit:
> > Reported-by: syzbot+0165480d4ef07360eeda@xxxxxxxxxxxxxxxxxxxxxxxxx
> >
> > ==================================================================
> > BUG: KASAN: use-after-free in __write_once_size include/linux/compiler.h:221 [inline]
> > BUG: KASAN: use-after-free in __hlist_del include/linux/list.h:748 [inline]
> > BUG: KASAN: use-after-free in hlist_del_rcu include/linux/rculist.h:455 [inline]
> > BUG: KASAN: use-after-free in xfrm_hash_rebuild+0xa0d/0x1000 net/xfrm/xfrm_policy.c:1318
> > Write of size 8 at addr ffff888095e79c00 by task kworker/1:3/8066
> >
> > CPU: 1 PID: 8066 Comm: kworker/1:3 Not tainted 5.2.0-rc6+ #7
> > Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS
> > Google 01/01/2011
> > Workqueue: events xfrm_hash_rebuild
> > Call Trace:
> > __dump_stack lib/dump_stack.c:77 [inline]
> > dump_stack+0x1d8/0x2f8 lib/dump_stack.c:113
> > print_address_description+0x6d/0x310 mm/kasan/report.c:188
> > __kasan_report+0x14b/0x1c0 mm/kasan/report.c:317
> > kasan_report+0x26/0x50 mm/kasan/common.c:614
> > __asan_report_store8_noabort+0x17/0x20 mm/kasan/generic_report.c:137
> > __write_once_size include/linux/compiler.h:221 [inline]
> > __hlist_del include/linux/list.h:748 [inline]
> > hlist_del_rcu include/linux/rculist.h:455 [inline]
> > xfrm_hash_rebuild+0xa0d/0x1000 net/xfrm/xfrm_policy.c:1318
> > process_one_work+0x814/0x1130 kernel/workqueue.c:2269
> > worker_thread+0xc01/0x1640 kernel/workqueue.c:2415
> > kthread+0x325/0x350 kernel/kthread.c:255
> > ret_from_fork+0x24/0x30 arch/x86/entry/entry_64.S:352
> >
> > Allocated by task 8064:
> > save_stack mm/kasan/common.c:71 [inline]
> > set_track mm/kasan/common.c:79 [inline]
> > __kasan_kmalloc+0x11c/0x1b0 mm/kasan/common.c:489
> > kasan_kmalloc+0x9/0x10 mm/kasan/common.c:503
> > __do_kmalloc mm/slab.c:3660 [inline]
> > __kmalloc+0x23c/0x310 mm/slab.c:3669
> > kmalloc include/linux/slab.h:552 [inline]
> > kzalloc include/linux/slab.h:742 [inline]
> > xfrm_hash_alloc+0x38/0xe0 net/xfrm/xfrm_hash.c:21
> > xfrm_policy_init net/xfrm/xfrm_policy.c:4036 [inline]
> > xfrm_net_init+0x269/0xd60 net/xfrm/xfrm_policy.c:4120
> > ops_init+0x336/0x420 net/core/net_namespace.c:130
> > setup_net+0x212/0x690 net/core/net_namespace.c:316
> > copy_net_ns+0x224/0x380 net/core/net_namespace.c:439
> > create_new_namespaces+0x4ec/0x700 kernel/nsproxy.c:103
> > unshare_nsproxy_namespaces+0x12a/0x190 kernel/nsproxy.c:202
> > ksys_unshare+0x540/0xac0 kernel/fork.c:2692
> > __do_sys_unshare kernel/fork.c:2760 [inline]
> > __se_sys_unshare kernel/fork.c:2758 [inline]
> > __x64_sys_unshare+0x38/0x40 kernel/fork.c:2758
> > do_syscall_64+0xfe/0x140 arch/x86/entry/common.c:301
> > entry_SYSCALL_64_after_hwframe+0x49/0xbe
> >
> > Freed by task 17:
> > save_stack mm/kasan/common.c:71 [inline]
> > set_track mm/kasan/common.c:79 [inline]
> > __kasan_slab_free+0x12a/0x1e0 mm/kasan/common.c:451
> > kasan_slab_free+0xe/0x10 mm/kasan/common.c:459
> > __cache_free mm/slab.c:3432 [inline]
> > kfree+0xae/0x120 mm/slab.c:3755
> > xfrm_hash_free+0x38/0xd0 net/xfrm/xfrm_hash.c:35
> > xfrm_bydst_resize net/xfrm/xfrm_policy.c:602 [inline]
> > xfrm_hash_resize+0x13f1/0x1840 net/xfrm/xfrm_policy.c:680
> > process_one_work+0x814/0x1130 kernel/workqueue.c:2269
> > worker_thread+0xc01/0x1640 kernel/workqueue.c:2415
> > kthread+0x325/0x350 kernel/kthread.c:255
> > ret_from_fork+0x24/0x30 arch/x86/entry/entry_64.S:352
> >
> > The buggy address belongs to the object at ffff888095e79c00
> > which belongs to the cache kmalloc-64 of size 64
> > The buggy address is located 0 bytes inside of
> > 64-byte region [ffff888095e79c00, ffff888095e79c40)
> > The buggy address belongs to the page:
> > page:ffffea0002579e40 refcount:1 mapcount:0 mapping:ffff8880aa400340
> > index:0x0
> > flags: 0x1fffc0000000200(slab)
> > raw: 01fffc0000000200 ffffea0002540888 ffffea0002907548 ffff8880aa400340
> > raw: 0000000000000000 ffff888095e79000 0000000100000020 0000000000000000
> > page dumped because: kasan: bad access detected
> >
> > Memory state around the buggy address:
> > ffff888095e79b00: 00 00 00 00 00 00 00 00 fc fc fc fc fc fc fc fc
> > ffff888095e79b80: 00 00 00 00 00 00 00 00 fc fc fc fc fc fc fc fc
> > > ffff888095e79c00: fb fb fb fb fb fb fb fb fc fc fc fc fc fc fc fc
> > ^
> > ffff888095e79c80: fb fb fb fb fb fb fb fb fc fc fc fc fc fc fc fc
> > ffff888095e79d00: 00 00 00 00 00 00 00 00 fc fc fc fc fc fc fc fc
> > ==================================================================
> >
>
> --- a/net/xfrm/xfrm_policy.c
> +++ b/net/xfrm/xfrm_policy.c
> @@ -1203,6 +1203,11 @@ xfrm_policy_inexact_insert(struct xfrm_policy *policy, u8 dir, int excl)
> return delpol;
> }
>
> +static inline bool xfrm_policy_node_hashed(struct hlist_node *node)
> +{
> + return node->pprev && node->pprev != LIST_POISON2;

Is it right to open code LIST_POISON2 use here? As far as I see all
current uses of LIST_POISON2 are encapsulated in list functions.

> +}
> +
> static void xfrm_hash_rebuild(struct work_struct *work)
> {
> struct net *net = container_of(work, struct net,
> @@ -1315,7 +1320,9 @@ static void xfrm_hash_rebuild(struct work_struct *work)
> chain = policy_hash_bysel(net, &policy->selector,
> policy->family, dir);
>
> - hlist_del_rcu(&policy->bydst);
> + /* check bydst still hashed in case that policy survived bydst resize */
> + if (xfrm_policy_node_hashed(&policy->bydst))
> + hlist_del_rcu(&policy->bydst);
>
> if (!chain) {
> void *p = xfrm_policy_inexact_insert(policy, dir, 0);
> --