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

From: Steffen Klassert
Date: Thu Sep 27 2018 - 05:12:03 EST


On Thu, Sep 27, 2018 at 10:35:48AM +0200, Ard Biesheuvel wrote:
> (+ Stefan)
>
> On Wed, 26 Sep 2018 at 22:24, syzbot
> <syzbot+6d3612ba5e254e387153@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote:
> >
> > Hello,
> >
> > syzbot found the following crash on:
> >
> > HEAD commit: 739d0def85ca Merge branch 'hv_netvsc-Support-LRO-RSC-in-th..
> > git tree: net-next
> > console output: https://syzkaller.appspot.com/x/log.txt?x=1146ffae400000
> > kernel config: https://syzkaller.appspot.com/x/.config?x=e79b9299baeb2298
> > dashboard link: https://syzkaller.appspot.com/bug?extid=6d3612ba5e254e387153
> > compiler: gcc (GCC) 8.0.1 20180413 (experimental)
> > syz repro: https://syzkaller.appspot.com/x/repro.syz?x=16200b9e400000
> > C reproducer: https://syzkaller.appspot.com/x/repro.c?x=102311c6400000
> >
> > IMPORTANT: if you fix the bug, please add the following tag to the commit:
> > Reported-by: syzbot+6d3612ba5e254e387153@xxxxxxxxxxxxxxxxxxxxxxxxx
> >
> > random: sshd: uninitialized urandom read (32 bytes read)
> > ==================================================================
> > BUG: KASAN: use-after-free in memcpy include/linux/string.h:345 [inline]
> > BUG: KASAN: use-after-free in generic_gcmaes_encrypt+0xc6/0x186
> > arch/x86/crypto/aesni-intel_glue.c:1291
> > Read of size 12 at addr ffff8801d7ad0900 by task kworker/0:1/14
> >
> > CPU: 0 PID: 14 Comm: kworker/0:1 Not tainted 4.19.0-rc4+ #228
> > Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS
> > Google 01/01/2011
> > Workqueue: pencrypt padata_parallel_worker
> > Call Trace:
> > __dump_stack lib/dump_stack.c:77 [inline]
> > dump_stack+0x1c4/0x2b4 lib/dump_stack.c:113
> > print_address_description.cold.8+0x9/0x1ff mm/kasan/report.c:256
> > kasan_report_error mm/kasan/report.c:354 [inline]
> > kasan_report.cold.9+0x242/0x309 mm/kasan/report.c:412
> > check_memory_region_inline mm/kasan/kasan.c:260 [inline]
> > check_memory_region+0x13e/0x1b0 mm/kasan/kasan.c:267
> > memcpy+0x23/0x50 mm/kasan/kasan.c:302
> > memcpy include/linux/string.h:345 [inline]
> > generic_gcmaes_encrypt+0xc6/0x186 arch/x86/crypto/aesni-intel_glue.c:1291
> > crypto_aead_encrypt include/crypto/aead.h:335 [inline]
> > gcmaes_wrapper_encrypt+0x162/0x200 arch/x86/crypto/aesni-intel_glue.c:1127
> > crypto_aead_encrypt include/crypto/aead.h:335 [inline]
> > pcrypt_aead_enc+0xcb/0x190 crypto/pcrypt.c:143
> > padata_parallel_worker+0x49d/0x760 kernel/padata.c:86
> > process_one_work+0xc90/0x1b90 kernel/workqueue.c:2153
> > worker_thread+0x17f/0x1390 kernel/workqueue.c:2296
> > kthread+0x35a/0x420 kernel/kthread.c:246
> > ret_from_fork+0x3a/0x50 arch/x86/entry/entry_64.S:413
> >
> > Allocated by task 5474:
> > save_stack+0x43/0xd0 mm/kasan/kasan.c:448
> > set_track mm/kasan/kasan.c:460 [inline]
> > kasan_kmalloc+0xc7/0xe0 mm/kasan/kasan.c:553
> > kmem_cache_alloc_trace+0x152/0x750 mm/slab.c:3620
> > kmalloc include/linux/slab.h:513 [inline]
> > tls_set_sw_offload+0xcc4/0x14b0 net/tls/tls_sw.c:1741
> > do_tls_setsockopt_conf net/tls/tls_main.c:467 [inline]
> > do_tls_setsockopt net/tls/tls_main.c:514 [inline]
> > tls_setsockopt+0x689/0x770 net/tls/tls_main.c:533
> > sock_common_setsockopt+0x9a/0xe0 net/core/sock.c:3038
> > __sys_setsockopt+0x1ba/0x3c0 net/socket.c:1902
> > __do_sys_setsockopt net/socket.c:1913 [inline]
> > __se_sys_setsockopt net/socket.c:1910 [inline]
> > __x64_sys_setsockopt+0xbe/0x150 net/socket.c:1910
> > do_syscall_64+0x1b9/0x820 arch/x86/entry/common.c:290
> > entry_SYSCALL_64_after_hwframe+0x49/0xbe
> >
> > Freed by task 5472:
> > save_stack+0x43/0xd0 mm/kasan/kasan.c:448
> > set_track mm/kasan/kasan.c:460 [inline]
> > __kasan_slab_free+0x102/0x150 mm/kasan/kasan.c:521
> > kasan_slab_free+0xe/0x10 mm/kasan/kasan.c:528
> > __cache_free mm/slab.c:3498 [inline]
> > kfree+0xcf/0x230 mm/slab.c:3813
> > tls_sk_proto_close+0x5fd/0x750 net/tls/tls_main.c:277

I'd say tls_sk_proto_close does not handle asynchronous crypto properly.
The pcrypt template is used here, so all crypto requests are processed
asynchronously. tls_sk_proto_close frees the needed recources before
the remaining crypto requests are finalized.