use-after-free in sctp_do_sm

Hello,

The following program triggers use-after-free in sctp_do_sm:

// autogenerated by syzkaller (http://github.com/google/syzkaller)
#include <syscall.h>
#include <string.h>
#include <stdint.h>

int main()

{
        long r0 = syscall(SYS_socket, 0xaul, 0x80805ul, 0x0ul, 0, 0, 0);
        long r1 = syscall(SYS_mmap, 0x20000000ul, 0x10000ul, 0x3ul,
0x32ul, 0xfffffffffffffffful, 0x0ul);
        memcpy((void*)0x20002fe4,
"\x0a\x00\x33\xe7\xeb\x9d\xcf\x61\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x01\xc5\xc8\x88\x64",
28);
        long r3 = syscall(SYS_bind, r0, 0x20002fe4ul, 0x1cul, 0, 0, 0);
        memcpy((void*)0x20000faa,
"\x9b\x01\x7d\xcd\xb8\x6a\xc7\x3d\x09\x3a\x07\x00\xa7\xc4\xe9\xee\x0a\xd6\xec\xde\x26\x75\x5f\x22\xae\x4e\x33\x00\xb0\x76\x10\x70\xd6\xca\x19\xbc\x15\x83\xcf\x2e\xbc\x99\x0c\x5e\x83\x89\xc1\x44\x9c\x6e\x74\xd8\x5d\x5d\xd0\xf0\xdf\x47\xc0\x00\x71\x0b\x55\x4c\xab\xf0\xd8\x90\xd5\x92\x8c\x6e\x33\x22\x15\x5b\x19\xfb\xed\xdd\xa6\xac\xcb\x60\xcf\xe2\xde\xed\xdb\x95\x5c\xaa\x20\xa3",
94);
        memcpy((void*)0x2000033a,
"\x02\x00\x33\xe2\x7f\x00\x00\x01\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00",
128);
        long r6 = syscall(SYS_sendto, r0, 0x20000faaul, 0x5eul,
0x81ul, 0x2000033aul, 0x80ul);
        return 0;
}

=================================

BUG: KASAN: use-after-free in sctp_do_sm+0x42f6/0x4f60 at addr ffff880036fa80a8
Read of size 4 by task a.out/5664

======================================> BUG kmalloc-4096 (Tainted: G B ): kasan: bad access detected
-----------------------------------------------------------------------------

INFO: Allocated in sctp_association_new+0x6f/0x1ea0 age=8 cpu=1 pidV64

[<      none      >] kmem_cache_alloc_trace+0x1cf/0x220 ./mm/slab.c:3707
[<      none      >] sctp_association_new+0x6f/0x1ea0
[<      none      >] sctp_sendmsg+0x1954/0x28e0
[<      none      >] inet_sendmsg+0x316/0x4f0 ./net/ipv4/af_inet.c:802
[<     inline     >] __sock_sendmsg_nosec ./net/socket.c:641
[<     inline     >] __sock_sendmsg ./net/socket.c:651
[<      none      >] sock_sendmsg+0xca/0x110 ./net/socket.c:662
[<      none      >] SYSC_sendto+0x208/0x350 ./net/socket.c:1841
[<      none      >] SyS_sendto+0x40/0x50 ./net/socket.c:1862
[<      none      >] entry_SYSCALL_64_fastpath+0x16/0x7a

INFO: Freed in sctp_association_put+0x150/0x250 age cpu=1 pidV64

[<      none      >] kfree+0x199/0x1b0 ./mm/slab.c:1211
[<      none      >] sctp_association_put+0x150/0x250
[<      none      >] sctp_association_free+0x498/0x630
[<      none      >] sctp_do_sm+0xd8b/0x4f60
[<      none      >] sctp_primitive_SHUTDOWN+0xa9/0xd0
[<      none      >] sctp_close+0x616/0x790
[<      none      >] inet_release+0xed/0x1c0 ./net/ipv4/af_inet.c:471
[<      none      >] inet6_release+0x50/0x70 ./net/ipv6/af_inet6.c:416
[<     inline     >] constant_test_bit ././arch/x86/include/asm/bitops.h:321
[<      none      >] sock_release+0x8d/0x200 ./net/socket.c:601
[<      none      >] sock_close+0x16/0x20 ./net/socket.c:1188
[<      none      >] __fput+0x21d/0x6e0 ./fs/file_table.c:265
[<      none      >] ____fput+0x15/0x20 ./fs/file_table.c:84
[<      none      >] task_work_run+0x163/0x1f0 ./include/trace/events/rcu.h:20
[<     inline     >] __list_add ./include/linux/list.h:42
[<     inline     >] list_add_tail ./include/linux/list.h:76
[<     inline     >] list_move_tail ./include/linux/list.h:168
[<     inline     >] reparent_leader ./kernel/exit.c:618
[<     inline     >] forget_original_parent ./kernel/exit.c:669
[<     inline     >] exit_notify ./kernel/exit.c:697
[<      none      >] do_exit+0x809/0x2b90 ./kernel/exit.c:878
[<      none      >] do_group_exit+0x108/0x320 ./kernel/exit.c:985

INFO: Slab 0xffffea0000dbea00 objects=7 used=1 fp=0xffff880036fa8000 flags=0x100000000004080
INFO: Object 0xffff880036fa8000 @offset=0 fp=0xffff880036fad668

CPU: 1 PID: 5664 Comm: a.out Tainted: G    B           4.4.0-rc1+ #81
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
 00000000ffffffff ffff880061d6f700 ffffffff825d3336 ffff88003e806d00
 ffff880036fa8000 ffff880036fa8000 ffff880061d6f730 ffffffff81618784
 ffff88003e806d00 ffffea0000dbea00 ffff880036fa8000 0000000000000000

Call Trace:
 [<ffffffff8162131e>] __asan_report_load4_noabort+0x3e/0x40
 [<ffffffff8475ac76>] sctp_do_sm+0x42f6/0x4f60
 [<ffffffff847b50e9>] sctp_primitive_SHUTDOWN+0xa9/0xd0
 [<ffffffff847a1426>] sctp_close+0x616/0x790
 [<ffffffff8409bb0d>] inet_release+0xed/0x1c0 ./net/ipv4/af_inet.c:471
 [<ffffffff84192cc0>] inet6_release+0x50/0x70 ./net/ipv6/af_inet6.c:416
 [<     inline     >] constant_test_bit ././arch/x86/include/asm/bitops.h:321
 [<ffffffff83dc78cd>] sock_release+0x8d/0x200 ./net/socket.c:601
 [<ffffffff83dc7a56>] sock_close+0x16/0x20 ./net/socket.c:1188
 [<ffffffff81662f5d>] __fput+0x21d/0x6e0 ./fs/file_table.c:265
 [<ffffffff816634a5>] ____fput+0x15/0x20 ./fs/file_table.c:84
 [<ffffffff812a33d3>] task_work_run+0x163/0x1f0 ./include/trace/events/rcu.h:20
 [<     inline     >] __list_add ./include/linux/list.h:42
 [<     inline     >] list_add_tail ./include/linux/list.h:76
 [<     inline     >] list_move_tail ./include/linux/list.h:168
 [<     inline     >] reparent_leader ./kernel/exit.c:618
 [<     inline     >] forget_original_parent ./kernel/exit.c:669
 [<     inline     >] exit_notify ./kernel/exit.c:697
 [<ffffffff812505d9>] do_exit+0x809/0x2b90 ./kernel/exit.c:878
 [<ffffffff81252ad8>] do_group_exit+0x108/0x320 ./kernel/exit.c:985
 [<ffffffff81252d0d>] SyS_exit_group+0x1d/0x20 ./kernel/exit.c:1002
 [<ffffffff84bf0c36>] entry_SYSCALL_64_fastpath+0x16/0x7a
=================================

I am on commit 90b55590c43258a157a2a143748455dcc50fbb53 of net-next (Nov 22).

Thanks

Quoted 107 lines · Dmitry Vyukov

Hello,

The following program triggers use-after-free in sctp_do_sm:

// autogenerated by syzkaller (http://github.com/google/syzkaller)
#include <syscall.h>
#include <string.h>
#include <stdint.h>

int main()

{
        long r0 = syscall(SYS_socket, 0xaul, 0x80805ul, 0x0ul, 0, 0, 0);
        long r1 = syscall(SYS_mmap, 0x20000000ul, 0x10000ul, 0x3ul,
0x32ul, 0xfffffffffffffffful, 0x0ul);
        memcpy((void*)0x20002fe4,
"\x0a\x00\x33\xe7\xeb\x9d\xcf\x61\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x01\xc5\xc8\x88\x64",
28);
        long r3 = syscall(SYS_bind, r0, 0x20002fe4ul, 0x1cul, 0, 0, 0);
        memcpy((void*)0x20000faa,
"\x9b\x01\x7d\xcd\xb8\x6a\xc7\x3d\x09\x3a\x07\x00\xa7\xc4\xe9\xee\x0a\xd6\xec\xde\x26\x75\x5f\x22\xae\x4e\x33\x00\xb0\x76\x10\x70\xd6\xca\x19\xbc\x15\x83\xcf\x2e\xbc\x99\x0c\x5e\x83\x89\xc1\x44\x9c\x6e\x74\xd8\x5d\x5d\xd0\xf0\xdf\x47\xc0\x00\x71\x0b\x55\x4c\xab\xf0\xd8\x90\xd5\x92\x8c\x6e\x33\x22\x15\x5b\x19\xfb\xed\xdd\xa6\xac\xcb\x60\xcf\xe2\xde\xed\xdb\x95\x5c\xaa\x20\xa3",
94);
        memcpy((void*)0x2000033a,
"\x02\x00\x33\xe2\x7f\x00\x00\x01\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00",
128);
        long r6 = syscall(SYS_sendto, r0, 0x20000faaul, 0x5eul,
0x81ul, 0x2000033aul, 0x80ul);
        return 0;
}

=================================

BUG: KASAN: use-after-free in sctp_do_sm+0x42f6/0x4f60 at addr ffff880036fa80a8
Read of size 4 by task a.out/5664

======================================>> BUG kmalloc-4096 (Tainted: G B ): kasan: bad access detected
-----------------------------------------------------------------------------

INFO: Allocated in sctp_association_new+0x6f/0x1ea0 age=8 cpu=1 pidV64

[<      none      >] kmem_cache_alloc_trace+0x1cf/0x220 ./mm/slab.c:3707
[<      none      >] sctp_association_new+0x6f/0x1ea0
[<      none      >] sctp_sendmsg+0x1954/0x28e0
[<      none      >] inet_sendmsg+0x316/0x4f0 ./net/ipv4/af_inet.c:802
[<     inline     >] __sock_sendmsg_nosec ./net/socket.c:641
[<     inline     >] __sock_sendmsg ./net/socket.c:651
[<      none      >] sock_sendmsg+0xca/0x110 ./net/socket.c:662
[<      none      >] SYSC_sendto+0x208/0x350 ./net/socket.c:1841
[<      none      >] SyS_sendto+0x40/0x50 ./net/socket.c:1862
[<      none      >] entry_SYSCALL_64_fastpath+0x16/0x7a

INFO: Freed in sctp_association_put+0x150/0x250 age cpu=1 pidV64

[<      none      >] kfree+0x199/0x1b0 ./mm/slab.c:1211
[<      none      >] sctp_association_put+0x150/0x250
[<      none      >] sctp_association_free+0x498/0x630
[<      none      >] sctp_do_sm+0xd8b/0x4f60
[<      none      >] sctp_primitive_SHUTDOWN+0xa9/0xd0
[<      none      >] sctp_close+0x616/0x790
[<      none      >] inet_release+0xed/0x1c0 ./net/ipv4/af_inet.c:471
[<      none      >] inet6_release+0x50/0x70 ./net/ipv6/af_inet6.c:416
[<     inline     >] constant_test_bit ././arch/x86/include/asm/bitops.h:321
[<      none      >] sock_release+0x8d/0x200 ./net/socket.c:601
[<      none      >] sock_close+0x16/0x20 ./net/socket.c:1188
[<      none      >] __fput+0x21d/0x6e0 ./fs/file_table.c:265
[<      none      >] ____fput+0x15/0x20 ./fs/file_table.c:84
[<      none      >] task_work_run+0x163/0x1f0 ./include/trace/events/rcu.h:20
[<     inline     >] __list_add ./include/linux/list.h:42
[<     inline     >] list_add_tail ./include/linux/list.h:76
[<     inline     >] list_move_tail ./include/linux/list.h:168
[<     inline     >] reparent_leader ./kernel/exit.c:618
[<     inline     >] forget_original_parent ./kernel/exit.c:669
[<     inline     >] exit_notify ./kernel/exit.c:697
[<      none      >] do_exit+0x809/0x2b90 ./kernel/exit.c:878
[<      none      >] do_group_exit+0x108/0x320 ./kernel/exit.c:985

INFO: Slab 0xffffea0000dbea00 objects=7 used=1 fp=0xffff880036fa8000 flags=0x100000000004080
INFO: Object 0xffff880036fa8000 @offset=0 fp=0xffff880036fad668

CPU: 1 PID: 5664 Comm: a.out Tainted: G    B           4.4.0-rc1+ #81
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
 00000000ffffffff ffff880061d6f700 ffffffff825d3336 ffff88003e806d00
 ffff880036fa8000 ffff880036fa8000 ffff880061d6f730 ffffffff81618784
 ffff88003e806d00 ffffea0000dbea00 ffff880036fa8000 0000000000000000

Call Trace:
 [<ffffffff8162131e>] __asan_report_load4_noabort+0x3e/0x40
 [<ffffffff8475ac76>] sctp_do_sm+0x42f6/0x4f60
 [<ffffffff847b50e9>] sctp_primitive_SHUTDOWN+0xa9/0xd0
 [<ffffffff847a1426>] sctp_close+0x616/0x790
 [<ffffffff8409bb0d>] inet_release+0xed/0x1c0 ./net/ipv4/af_inet.c:471
 [<ffffffff84192cc0>] inet6_release+0x50/0x70 ./net/ipv6/af_inet6.c:416
 [<     inline     >] constant_test_bit ././arch/x86/include/asm/bitops.h:321
 [<ffffffff83dc78cd>] sock_release+0x8d/0x200 ./net/socket.c:601
 [<ffffffff83dc7a56>] sock_close+0x16/0x20 ./net/socket.c:1188
 [<ffffffff81662f5d>] __fput+0x21d/0x6e0 ./fs/file_table.c:265
 [<ffffffff816634a5>] ____fput+0x15/0x20 ./fs/file_table.c:84
 [<ffffffff812a33d3>] task_work_run+0x163/0x1f0 ./include/trace/events/rcu.h:20
 [<     inline     >] __list_add ./include/linux/list.h:42
 [<     inline     >] list_add_tail ./include/linux/list.h:76
 [<     inline     >] list_move_tail ./include/linux/list.h:168
 [<     inline     >] reparent_leader ./kernel/exit.c:618
 [<     inline     >] forget_original_parent ./kernel/exit.c:669
 [<     inline     >] exit_notify ./kernel/exit.c:697
 [<ffffffff812505d9>] do_exit+0x809/0x2b90 ./kernel/exit.c:878
 [<ffffffff81252ad8>] do_group_exit+0x108/0x320 ./kernel/exit.c:985
 [<ffffffff81252d0d>] SyS_exit_group+0x1d/0x20 ./kernel/exit.c:1002
 [<ffffffff84bf0c36>] entry_SYSCALL_64_fastpath+0x16/0x7a
=================================

Quoted 1 line

I am on commit 90b55590c43258a157a2a143748455dcc50fbb53 of net-next (Nov 22).

Quoted 108 lines · Dmitry Vyukov

Quoted 107 lines · Dmitry Vyukov

Hello,

The following program triggers use-after-free in sctp_do_sm:

// autogenerated by syzkaller (http://github.com/google/syzkaller)
#include <syscall.h>
#include <string.h>
#include <stdint.h>

int main()

{
        long r0 = syscall(SYS_socket, 0xaul, 0x80805ul, 0x0ul, 0, 0, 0);
        long r1 = syscall(SYS_mmap, 0x20000000ul, 0x10000ul, 0x3ul,
0x32ul, 0xfffffffffffffffful, 0x0ul);
        memcpy((void*)0x20002fe4,
"\x0a\x00\x33\xe7\xeb\x9d\xcf\x61\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x01\xc5\xc8\x88\x64",
28);
        long r3 = syscall(SYS_bind, r0, 0x20002fe4ul, 0x1cul, 0, 0, 0);
        memcpy((void*)0x20000faa,
"\x9b\x01\x7d\xcd\xb8\x6a\xc7\x3d\x09\x3a\x07\x00\xa7\xc4\xe9\xee\x0a\xd6\xec\xde\x26\x75\x5f\x22\xae\x4e\x33\x00\xb0\x76\x10\x70\xd6\xca\x19\xbc\x15\x83\xcf\x2e\xbc\x99\x0c\x5e\x83\x89\xc1\x44\x9c\x6e\x74\xd8\x5d\x5d\xd0\xf0\xdf\x47\xc0\x00\x71\x0b\x55\x4c\xab\xf0\xd8\x90\xd5\x92\x8c\x6e\x33\x22\x15\x5b\x19\xfb\xed\xdd\xa6\xac\xcb\x60\xcf\xe2\xde\xed\xdb\x95\x5c\xaa\x20\xa3",
94);
        memcpy((void*)0x2000033a,
"\x02\x00\x33\xe2\x7f\x00\x00\x01\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00",
128);
        long r6 = syscall(SYS_sendto, r0, 0x20000faaul, 0x5eul,
0x81ul, 0x2000033aul, 0x80ul);
        return 0;
}

=================================

BUG: KASAN: use-after-free in sctp_do_sm+0x42f6/0x4f60 at addr ffff880036fa80a8
Read of size 4 by task a.out/5664

======================================> >> BUG kmalloc-4096 (Tainted: G B ): kasan: bad access detected
-----------------------------------------------------------------------------

INFO: Allocated in sctp_association_new+0x6f/0x1ea0 age=8 cpu=1 pidV64

[<      none      >] kmem_cache_alloc_trace+0x1cf/0x220 ./mm/slab.c:3707
[<      none      >] sctp_association_new+0x6f/0x1ea0
[<      none      >] sctp_sendmsg+0x1954/0x28e0
[<      none      >] inet_sendmsg+0x316/0x4f0 ./net/ipv4/af_inet.c:802
[<     inline     >] __sock_sendmsg_nosec ./net/socket.c:641
[<     inline     >] __sock_sendmsg ./net/socket.c:651
[<      none      >] sock_sendmsg+0xca/0x110 ./net/socket.c:662
[<      none      >] SYSC_sendto+0x208/0x350 ./net/socket.c:1841
[<      none      >] SyS_sendto+0x40/0x50 ./net/socket.c:1862
[<      none      >] entry_SYSCALL_64_fastpath+0x16/0x7a

INFO: Freed in sctp_association_put+0x150/0x250 age cpu=1 pidV64

[<      none      >] kfree+0x199/0x1b0 ./mm/slab.c:1211
[<      none      >] sctp_association_put+0x150/0x250
[<      none      >] sctp_association_free+0x498/0x630
[<      none      >] sctp_do_sm+0xd8b/0x4f60
[<      none      >] sctp_primitive_SHUTDOWN+0xa9/0xd0
[<      none      >] sctp_close+0x616/0x790
[<      none      >] inet_release+0xed/0x1c0 ./net/ipv4/af_inet.c:471
[<      none      >] inet6_release+0x50/0x70 ./net/ipv6/af_inet6.c:416
[<     inline     >] constant_test_bit ././arch/x86/include/asm/bitops.h:321
[<      none      >] sock_release+0x8d/0x200 ./net/socket.c:601
[<      none      >] sock_close+0x16/0x20 ./net/socket.c:1188
[<      none      >] __fput+0x21d/0x6e0 ./fs/file_table.c:265
[<      none      >] ____fput+0x15/0x20 ./fs/file_table.c:84
[<      none      >] task_work_run+0x163/0x1f0 ./include/trace/events/rcu.h:20
[<     inline     >] __list_add ./include/linux/list.h:42
[<     inline     >] list_add_tail ./include/linux/list.h:76
[<     inline     >] list_move_tail ./include/linux/list.h:168
[<     inline     >] reparent_leader ./kernel/exit.c:618
[<     inline     >] forget_original_parent ./kernel/exit.c:669
[<     inline     >] exit_notify ./kernel/exit.c:697
[<      none      >] do_exit+0x809/0x2b90 ./kernel/exit.c:878
[<      none      >] do_group_exit+0x108/0x320 ./kernel/exit.c:985

INFO: Slab 0xffffea0000dbea00 objects=7 used=1 fp=0xffff880036fa8000 flags=0x100000000004080
INFO: Object 0xffff880036fa8000 @offset=0 fp=0xffff880036fad668

CPU: 1 PID: 5664 Comm: a.out Tainted: G    B           4.4.0-rc1+ #81
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
 00000000ffffffff ffff880061d6f700 ffffffff825d3336 ffff88003e806d00
 ffff880036fa8000 ffff880036fa8000 ffff880061d6f730 ffffffff81618784
 ffff88003e806d00 ffffea0000dbea00 ffff880036fa8000 0000000000000000

Call Trace:
 [<ffffffff8162131e>] __asan_report_load4_noabort+0x3e/0x40
 [<ffffffff8475ac76>] sctp_do_sm+0x42f6/0x4f60
 [<ffffffff847b50e9>] sctp_primitive_SHUTDOWN+0xa9/0xd0
 [<ffffffff847a1426>] sctp_close+0x616/0x790
 [<ffffffff8409bb0d>] inet_release+0xed/0x1c0 ./net/ipv4/af_inet.c:471
 [<ffffffff84192cc0>] inet6_release+0x50/0x70 ./net/ipv6/af_inet6.c:416
 [<     inline     >] constant_test_bit ././arch/x86/include/asm/bitops.h:321
 [<ffffffff83dc78cd>] sock_release+0x8d/0x200 ./net/socket.c:601
 [<ffffffff83dc7a56>] sock_close+0x16/0x20 ./net/socket.c:1188
 [<ffffffff81662f5d>] __fput+0x21d/0x6e0 ./fs/file_table.c:265
 [<ffffffff816634a5>] ____fput+0x15/0x20 ./fs/file_table.c:84
 [<ffffffff812a33d3>] task_work_run+0x163/0x1f0 ./include/trace/events/rcu.h:20
 [<     inline     >] __list_add ./include/linux/list.h:42
 [<     inline     >] list_add_tail ./include/linux/list.h:76
 [<     inline     >] list_move_tail ./include/linux/list.h:168
 [<     inline     >] reparent_leader ./kernel/exit.c:618
 [<     inline     >] forget_original_parent ./kernel/exit.c:669
 [<     inline     >] exit_notify ./kernel/exit.c:697
 [<ffffffff812505d9>] do_exit+0x809/0x2b90 ./kernel/exit.c:878
 [<ffffffff81252ad8>] do_group_exit+0x108/0x320 ./kernel/exit.c:985
 [<ffffffff81252d0d>] SyS_exit_group+0x1d/0x20 ./kernel/exit.c:1002
 [<ffffffff84bf0c36>] entry_SYSCALL_64_fastpath+0x16/0x7a
=================================

Quoted 1 line

Quoted 1 line

I am on commit 90b55590c43258a157a2a143748455dcc50fbb53 of net-next (Nov 22).

The right commit is:

commit 7d267278a9ece963d77eefec61630223fce08c6c

Author: Rainer Weikusat
Date:   Fri Nov 20 22:07:23 2015 +0000

unix: avoid use-after-free in ep_remove_wait_queue

--
To unsubscribe from this list: send the line "unsubscribe linux-sctp" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

Quoted 111 lines · Dmitry Vyukov

Quoted 109 lines · Dmitry Vyukov

Quoted 108 lines · Dmitry Vyukov

Hello,

The following program triggers use-after-free in sctp_do_sm:

// autogenerated by syzkaller (http://github.com/google/syzkaller)
#include <syscall.h>
#include <string.h>
#include <stdint.h>

int main()

{
        long r0 = syscall(SYS_socket, 0xaul, 0x80805ul, 0x0ul, 0, 0, 0);
        long r1 = syscall(SYS_mmap, 0x20000000ul, 0x10000ul, 0x3ul,
0x32ul, 0xfffffffffffffffful, 0x0ul);
        memcpy((void*)0x20002fe4,
"\x0a\x00\x33\xe7\xeb\x9d\xcf\x61\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x01\xc5\xc8\x88\x64",
28);
        long r3 = syscall(SYS_bind, r0, 0x20002fe4ul, 0x1cul, 0, 0, 0);
        memcpy((void*)0x20000faa,
"\x9b\x01\x7d\xcd\xb8\x6a\xc7\x3d\x09\x3a\x07\x00\xa7\xc4\xe9\xee\x0a\xd6\xec\xde\x26\x75\x5f\x22\xae\x4e\x33\x00\xb0\x76\x10\x70\xd6\xca\x19\xbc\x15\x83\xcf\x2e\xbc\x99\x0c\x5e\x83\x89\xc1\x44\x9c\x6e\x74\xd8\x5d\x5d\xd0\xf0\xdf\x47\xc0\x00\x71\x0b\x55\x4c\xab\xf0\xd8\x90\xd5\x92\x8c\x6e\x33\x22\x15\x5b\x19\xfb\xed\xdd\xa6\xac\xcb\x60\xcf\xe2\xde\xed\xdb\x95\x5c\xaa\x20\xa3",
94);
        memcpy((void*)0x2000033a,
"\x02\x00\x33\xe2\x7f\x00\x00\x01\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00",
128);
        long r6 = syscall(SYS_sendto, r0, 0x20000faaul, 0x5eul,
0x81ul, 0x2000033aul, 0x80ul);
        return 0;
}

=================================

BUG: KASAN: use-after-free in sctp_do_sm+0x42f6/0x4f60 at addr ffff880036fa80a8
Read of size 4 by task a.out/5664

======================================> > >> BUG kmalloc-4096 (Tainted: G B ): kasan: bad access detected
-----------------------------------------------------------------------------

INFO: Allocated in sctp_association_new+0x6f/0x1ea0 age=8 cpu=1 pidV64

[<      none      >] kmem_cache_alloc_trace+0x1cf/0x220 ./mm/slab.c:3707
[<      none      >] sctp_association_new+0x6f/0x1ea0
[<      none      >] sctp_sendmsg+0x1954/0x28e0
[<      none      >] inet_sendmsg+0x316/0x4f0 ./net/ipv4/af_inet.c:802
[<     inline     >] __sock_sendmsg_nosec ./net/socket.c:641
[<     inline     >] __sock_sendmsg ./net/socket.c:651
[<      none      >] sock_sendmsg+0xca/0x110 ./net/socket.c:662
[<      none      >] SYSC_sendto+0x208/0x350 ./net/socket.c:1841
[<      none      >] SyS_sendto+0x40/0x50 ./net/socket.c:1862
[<      none      >] entry_SYSCALL_64_fastpath+0x16/0x7a

INFO: Freed in sctp_association_put+0x150/0x250 age cpu=1 pidV64

[<      none      >] kfree+0x199/0x1b0 ./mm/slab.c:1211
[<      none      >] sctp_association_put+0x150/0x250
[<      none      >] sctp_association_free+0x498/0x630
[<      none      >] sctp_do_sm+0xd8b/0x4f60
[<      none      >] sctp_primitive_SHUTDOWN+0xa9/0xd0
[<      none      >] sctp_close+0x616/0x790
[<      none      >] inet_release+0xed/0x1c0 ./net/ipv4/af_inet.c:471
[<      none      >] inet6_release+0x50/0x70 ./net/ipv6/af_inet6.c:416
[<     inline     >] constant_test_bit ././arch/x86/include/asm/bitops.h:321
[<      none      >] sock_release+0x8d/0x200 ./net/socket.c:601
[<      none      >] sock_close+0x16/0x20 ./net/socket.c:1188
[<      none      >] __fput+0x21d/0x6e0 ./fs/file_table.c:265
[<      none      >] ____fput+0x15/0x20 ./fs/file_table.c:84
[<      none      >] task_work_run+0x163/0x1f0 ./include/trace/events/rcu.h:20
[<     inline     >] __list_add ./include/linux/list.h:42
[<     inline     >] list_add_tail ./include/linux/list.h:76
[<     inline     >] list_move_tail ./include/linux/list.h:168
[<     inline     >] reparent_leader ./kernel/exit.c:618
[<     inline     >] forget_original_parent ./kernel/exit.c:669
[<     inline     >] exit_notify ./kernel/exit.c:697
[<      none      >] do_exit+0x809/0x2b90 ./kernel/exit.c:878
[<      none      >] do_group_exit+0x108/0x320 ./kernel/exit.c:985

INFO: Slab 0xffffea0000dbea00 objects=7 used=1 fp=0xffff880036fa8000 flags=0x100000000004080
INFO: Object 0xffff880036fa8000 @offset=0 fp=0xffff880036fad668

CPU: 1 PID: 5664 Comm: a.out Tainted: G    B           4.4.0-rc1+ #81
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
 00000000ffffffff ffff880061d6f700 ffffffff825d3336 ffff88003e806d00
 ffff880036fa8000 ffff880036fa8000 ffff880061d6f730 ffffffff81618784
 ffff88003e806d00 ffffea0000dbea00 ffff880036fa8000 0000000000000000

Call Trace:
 [<ffffffff8162131e>] __asan_report_load4_noabort+0x3e/0x40
 [<ffffffff8475ac76>] sctp_do_sm+0x42f6/0x4f60
 [<ffffffff847b50e9>] sctp_primitive_SHUTDOWN+0xa9/0xd0
 [<ffffffff847a1426>] sctp_close+0x616/0x790
 [<ffffffff8409bb0d>] inet_release+0xed/0x1c0 ./net/ipv4/af_inet.c:471
 [<ffffffff84192cc0>] inet6_release+0x50/0x70 ./net/ipv6/af_inet6.c:416
 [<     inline     >] constant_test_bit ././arch/x86/include/asm/bitops.h:321
 [<ffffffff83dc78cd>] sock_release+0x8d/0x200 ./net/socket.c:601
 [<ffffffff83dc7a56>] sock_close+0x16/0x20 ./net/socket.c:1188
 [<ffffffff81662f5d>] __fput+0x21d/0x6e0 ./fs/file_table.c:265
 [<ffffffff816634a5>] ____fput+0x15/0x20 ./fs/file_table.c:84
 [<ffffffff812a33d3>] task_work_run+0x163/0x1f0 ./include/trace/events/rcu.h:20
 [<     inline     >] __list_add ./include/linux/list.h:42
 [<     inline     >] list_add_tail ./include/linux/list.h:76
 [<     inline     >] list_move_tail ./include/linux/list.h:168
 [<     inline     >] reparent_leader ./kernel/exit.c:618
 [<     inline     >] forget_original_parent ./kernel/exit.c:669
 [<     inline     >] exit_notify ./kernel/exit.c:697
 [<ffffffff812505d9>] do_exit+0x809/0x2b90 ./kernel/exit.c:878
 [<ffffffff81252ad8>] do_group_exit+0x108/0x320 ./kernel/exit.c:985
 [<ffffffff81252d0d>] SyS_exit_group+0x1d/0x20 ./kernel/exit.c:1002
 [<ffffffff84bf0c36>] entry_SYSCALL_64_fastpath+0x16/0x7a
=================================

I am on commit 90b55590c43258a157a2a143748455dcc50fbb53 of net-next (Nov 22).

Quoted 6 lines

The right commit is:

commit 7d267278a9ece963d77eefec61630223fce08c6c

Author: Rainer Weikusat
Date:   Fri Nov 20 22:07:23 2015 +0000

unix: avoid use-after-free in ep_remove_wait_queue

This commit doesn't seem to exist

Quoted 6 lines

The right commit is:

commit 7d267278a9ece963d77eefec61630223fce08c6c

Author: Rainer Weikusat
Date:   Fri Nov 20 22:07:23 2015 +0000

unix: avoid use-after-free in ep_remove_wait_queue

This commit doesn't seem to exist

Quoted 111 lines · Dmitry Vyukov

Quoted 109 lines · Dmitry Vyukov

Quoted 108 lines · Dmitry Vyukov

Hello,

The following program triggers use-after-free in sctp_do_sm:

// autogenerated by syzkaller (http://github.com/google/syzkaller)
#include <syscall.h>
#include <string.h>
#include <stdint.h>

int main()

{
        long r0 = syscall(SYS_socket, 0xaul, 0x80805ul, 0x0ul, 0, 0, 0);
        long r1 = syscall(SYS_mmap, 0x20000000ul, 0x10000ul, 0x3ul,
0x32ul, 0xfffffffffffffffful, 0x0ul);
        memcpy((void*)0x20002fe4,
"\x0a\x00\x33\xe7\xeb\x9d\xcf\x61\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x01\xc5\xc8\x88\x64",
28);
        long r3 = syscall(SYS_bind, r0, 0x20002fe4ul, 0x1cul, 0, 0, 0);
        memcpy((void*)0x20000faa,
"\x9b\x01\x7d\xcd\xb8\x6a\xc7\x3d\x09\x3a\x07\x00\xa7\xc4\xe9\xee\x0a\xd6\xec\xde\x26\x75\x5f\x22\xae\x4e\x33\x00\xb0\x76\x10\x70\xd6\xca\x19\xbc\x15\x83\xcf\x2e\xbc\x99\x0c\x5e\x83\x89\xc1\x44\x9c\x6e\x74\xd8\x5d\x5d\xd0\xf0\xdf\x47\xc0\x00\x71\x0b\x55\x4c\xab\xf0\xd8\x90\xd5\x92\x8c\x6e\x33\x22\x15\x5b\x19\xfb\xed\xdd\xa6\xac\xcb\x60\xcf\xe2\xde\xed\xdb\x95\x5c\xaa\x20\xa3",
94);
        memcpy((void*)0x2000033a,
"\x02\x00\x33\xe2\x7f\x00\x00\x01\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00",
128);
        long r6 = syscall(SYS_sendto, r0, 0x20000faaul, 0x5eul,
0x81ul, 0x2000033aul, 0x80ul);
        return 0;
}

=================================

BUG: KASAN: use-after-free in sctp_do_sm+0x42f6/0x4f60 at addr ffff880036fa80a8
Read of size 4 by task a.out/5664

======================================>>>> BUG kmalloc-4096 (Tainted: G B ): kasan: bad access detected
-----------------------------------------------------------------------------

INFO: Allocated in sctp_association_new+0x6f/0x1ea0 age=8 cpu=1 pidV64

[<      none      >] kmem_cache_alloc_trace+0x1cf/0x220 ./mm/slab.c:3707
[<      none      >] sctp_association_new+0x6f/0x1ea0
[<      none      >] sctp_sendmsg+0x1954/0x28e0
[<      none      >] inet_sendmsg+0x316/0x4f0 ./net/ipv4/af_inet.c:802
[<     inline     >] __sock_sendmsg_nosec ./net/socket.c:641
[<     inline     >] __sock_sendmsg ./net/socket.c:651
[<      none      >] sock_sendmsg+0xca/0x110 ./net/socket.c:662
[<      none      >] SYSC_sendto+0x208/0x350 ./net/socket.c:1841
[<      none      >] SyS_sendto+0x40/0x50 ./net/socket.c:1862
[<      none      >] entry_SYSCALL_64_fastpath+0x16/0x7a

INFO: Freed in sctp_association_put+0x150/0x250 age cpu=1 pidV64

[<      none      >] kfree+0x199/0x1b0 ./mm/slab.c:1211
[<      none      >] sctp_association_put+0x150/0x250
[<      none      >] sctp_association_free+0x498/0x630
[<      none      >] sctp_do_sm+0xd8b/0x4f60
[<      none      >] sctp_primitive_SHUTDOWN+0xa9/0xd0
[<      none      >] sctp_close+0x616/0x790
[<      none      >] inet_release+0xed/0x1c0 ./net/ipv4/af_inet.c:471
[<      none      >] inet6_release+0x50/0x70 ./net/ipv6/af_inet6.c:416
[<     inline     >] constant_test_bit ././arch/x86/include/asm/bitops.h:321
[<      none      >] sock_release+0x8d/0x200 ./net/socket.c:601
[<      none      >] sock_close+0x16/0x20 ./net/socket.c:1188
[<      none      >] __fput+0x21d/0x6e0 ./fs/file_table.c:265
[<      none      >] ____fput+0x15/0x20 ./fs/file_table.c:84
[<      none      >] task_work_run+0x163/0x1f0 ./include/trace/events/rcu.h:20
[<     inline     >] __list_add ./include/linux/list.h:42
[<     inline     >] list_add_tail ./include/linux/list.h:76
[<     inline     >] list_move_tail ./include/linux/list.h:168
[<     inline     >] reparent_leader ./kernel/exit.c:618
[<     inline     >] forget_original_parent ./kernel/exit.c:669
[<     inline     >] exit_notify ./kernel/exit.c:697
[<      none      >] do_exit+0x809/0x2b90 ./kernel/exit.c:878
[<      none      >] do_group_exit+0x108/0x320 ./kernel/exit.c:985

INFO: Slab 0xffffea0000dbea00 objects=7 used=1 fp=0xffff880036fa8000 flags=0x100000000004080
INFO: Object 0xffff880036fa8000 @offset=0 fp=0xffff880036fad668

CPU: 1 PID: 5664 Comm: a.out Tainted: G    B           4.4.0-rc1+ #81
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
 00000000ffffffff ffff880061d6f700 ffffffff825d3336 ffff88003e806d00
 ffff880036fa8000 ffff880036fa8000 ffff880061d6f730 ffffffff81618784
 ffff88003e806d00 ffffea0000dbea00 ffff880036fa8000 0000000000000000

Call Trace:
 [<ffffffff8162131e>] __asan_report_load4_noabort+0x3e/0x40
 [<ffffffff8475ac76>] sctp_do_sm+0x42f6/0x4f60
 [<ffffffff847b50e9>] sctp_primitive_SHUTDOWN+0xa9/0xd0
 [<ffffffff847a1426>] sctp_close+0x616/0x790
 [<ffffffff8409bb0d>] inet_release+0xed/0x1c0 ./net/ipv4/af_inet.c:471
 [<ffffffff84192cc0>] inet6_release+0x50/0x70 ./net/ipv6/af_inet6.c:416
 [<     inline     >] constant_test_bit ././arch/x86/include/asm/bitops.h:321
 [<ffffffff83dc78cd>] sock_release+0x8d/0x200 ./net/socket.c:601
 [<ffffffff83dc7a56>] sock_close+0x16/0x20 ./net/socket.c:1188
 [<ffffffff81662f5d>] __fput+0x21d/0x6e0 ./fs/file_table.c:265
 [<ffffffff816634a5>] ____fput+0x15/0x20 ./fs/file_table.c:84
 [<ffffffff812a33d3>] task_work_run+0x163/0x1f0 ./include/trace/events/rcu.h:20
 [<     inline     >] __list_add ./include/linux/list.h:42
 [<     inline     >] list_add_tail ./include/linux/list.h:76
 [<     inline     >] list_move_tail ./include/linux/list.h:168
 [<     inline     >] reparent_leader ./kernel/exit.c:618
 [<     inline     >] forget_original_parent ./kernel/exit.c:669
 [<     inline     >] exit_notify ./kernel/exit.c:697
 [<ffffffff812505d9>] do_exit+0x809/0x2b90 ./kernel/exit.c:878
 [<ffffffff81252ad8>] do_group_exit+0x108/0x320 ./kernel/exit.c:985
 [<ffffffff81252d0d>] SyS_exit_group+0x1d/0x20 ./kernel/exit.c:1002
 [<ffffffff84bf0c36>] entry_SYSCALL_64_fastpath+0x16/0x7a
=================================

I am on commit 90b55590c43258a157a2a143748455dcc50fbb53 of net-next (Nov 22).

Quoted 6 lines

The right commit is:

commit 7d267278a9ece963d77eefec61630223fce08c6c

Author: Rainer Weikusat
Date:   Fri Nov 20 22:07:23 2015 +0000

unix: avoid use-after-free in ep_remove_wait_queue

This commit doesn't seem to exist

Neil

Quoted 5 lines

--
To unsubscribe from this list: send the line "unsubscribe linux-sctp" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

Quoted 118 lines · Neil Horman

Quoted 117 lines · Dmitry Vyukov

Quoted 108 lines · Dmitry Vyukov

Quoted 107 lines · Dmitry Vyukov

Hello,

The following program triggers use-after-free in sctp_do_sm:

// autogenerated by syzkaller (http://github.com/google/syzkaller)
#include <syscall.h>
#include <string.h>
#include <stdint.h>

int main()

{
        long r0 = syscall(SYS_socket, 0xaul, 0x80805ul, 0x0ul, 0, 0, 0);
        long r1 = syscall(SYS_mmap, 0x20000000ul, 0x10000ul, 0x3ul,
0x32ul, 0xfffffffffffffffful, 0x0ul);
        memcpy((void*)0x20002fe4,
"\x0a\x00\x33\xe7\xeb\x9d\xcf\x61\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x01\xc5\xc8\x88\x64",
28);
        long r3 = syscall(SYS_bind, r0, 0x20002fe4ul, 0x1cul, 0, 0, 0);
        memcpy((void*)0x20000faa,
"\x9b\x01\x7d\xcd\xb8\x6a\xc7\x3d\x09\x3a\x07\x00\xa7\xc4\xe9\xee\x0a\xd6\xec\xde\x26\x75\x5f\x22\xae\x4e\x33\x00\xb0\x76\x10\x70\xd6\xca\x19\xbc\x15\x83\xcf\x2e\xbc\x99\x0c\x5e\x83\x89\xc1\x44\x9c\x6e\x74\xd8\x5d\x5d\xd0\xf0\xdf\x47\xc0\x00\x71\x0b\x55\x4c\xab\xf0\xd8\x90\xd5\x92\x8c\x6e\x33\x22\x15\x5b\x19\xfb\xed\xdd\xa6\xac\xcb\x60\xcf\xe2\xde\xed\xdb\x95\x5c\xaa\x20\xa3",
94);
        memcpy((void*)0x2000033a,
"\x02\x00\x33\xe2\x7f\x00\x00\x01\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00",
128);
        long r6 = syscall(SYS_sendto, r0, 0x20000faaul, 0x5eul,
0x81ul, 0x2000033aul, 0x80ul);
        return 0;
}

=================================

BUG: KASAN: use-after-free in sctp_do_sm+0x42f6/0x4f60 at addr ffff880036fa80a8
Read of size 4 by task a.out/5664

======================================>>>>> BUG kmalloc-4096 (Tainted: G B ): kasan: bad access detected
-----------------------------------------------------------------------------

INFO: Allocated in sctp_association_new+0x6f/0x1ea0 age=8 cpu=1 pidV64

[<      none      >] kmem_cache_alloc_trace+0x1cf/0x220 ./mm/slab.c:3707
[<      none      >] sctp_association_new+0x6f/0x1ea0
[<      none      >] sctp_sendmsg+0x1954/0x28e0
[<      none      >] inet_sendmsg+0x316/0x4f0 ./net/ipv4/af_inet.c:802
[<     inline     >] __sock_sendmsg_nosec ./net/socket.c:641
[<     inline     >] __sock_sendmsg ./net/socket.c:651
[<      none      >] sock_sendmsg+0xca/0x110 ./net/socket.c:662
[<      none      >] SYSC_sendto+0x208/0x350 ./net/socket.c:1841
[<      none      >] SyS_sendto+0x40/0x50 ./net/socket.c:1862
[<      none      >] entry_SYSCALL_64_fastpath+0x16/0x7a

INFO: Freed in sctp_association_put+0x150/0x250 age cpu=1 pidV64

[<      none      >] kfree+0x199/0x1b0 ./mm/slab.c:1211
[<      none      >] sctp_association_put+0x150/0x250
[<      none      >] sctp_association_free+0x498/0x630
[<      none      >] sctp_do_sm+0xd8b/0x4f60
[<      none      >] sctp_primitive_SHUTDOWN+0xa9/0xd0
[<      none      >] sctp_close+0x616/0x790
[<      none      >] inet_release+0xed/0x1c0 ./net/ipv4/af_inet.c:471
[<      none      >] inet6_release+0x50/0x70 ./net/ipv6/af_inet6.c:416
[<     inline     >] constant_test_bit ././arch/x86/include/asm/bitops.h:321
[<      none      >] sock_release+0x8d/0x200 ./net/socket.c:601
[<      none      >] sock_close+0x16/0x20 ./net/socket.c:1188
[<      none      >] __fput+0x21d/0x6e0 ./fs/file_table.c:265
[<      none      >] ____fput+0x15/0x20 ./fs/file_table.c:84
[<      none      >] task_work_run+0x163/0x1f0 ./include/trace/events/rcu.h:20
[<     inline     >] __list_add ./include/linux/list.h:42
[<     inline     >] list_add_tail ./include/linux/list.h:76
[<     inline     >] list_move_tail ./include/linux/list.h:168
[<     inline     >] reparent_leader ./kernel/exit.c:618
[<     inline     >] forget_original_parent ./kernel/exit.c:669
[<     inline     >] exit_notify ./kernel/exit.c:697
[<      none      >] do_exit+0x809/0x2b90 ./kernel/exit.c:878
[<      none      >] do_group_exit+0x108/0x320 ./kernel/exit.c:985

INFO: Slab 0xffffea0000dbea00 objects=7 used=1 fp=0xffff880036fa8000 flags=0x100000000004080
INFO: Object 0xffff880036fa8000 @offset=0 fp=0xffff880036fad668

CPU: 1 PID: 5664 Comm: a.out Tainted: G    B           4.4.0-rc1+ #81
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
 00000000ffffffff ffff880061d6f700 ffffffff825d3336 ffff88003e806d00
 ffff880036fa8000 ffff880036fa8000 ffff880061d6f730 ffffffff81618784
 ffff88003e806d00 ffffea0000dbea00 ffff880036fa8000 0000000000000000

Call Trace:
 [<ffffffff8162131e>] __asan_report_load4_noabort+0x3e/0x40
 [<ffffffff8475ac76>] sctp_do_sm+0x42f6/0x4f60
 [<ffffffff847b50e9>] sctp_primitive_SHUTDOWN+0xa9/0xd0
 [<ffffffff847a1426>] sctp_close+0x616/0x790
 [<ffffffff8409bb0d>] inet_release+0xed/0x1c0 ./net/ipv4/af_inet.c:471
 [<ffffffff84192cc0>] inet6_release+0x50/0x70 ./net/ipv6/af_inet6.c:416
 [<     inline     >] constant_test_bit ././arch/x86/include/asm/bitops.h:321
 [<ffffffff83dc78cd>] sock_release+0x8d/0x200 ./net/socket.c:601
 [<ffffffff83dc7a56>] sock_close+0x16/0x20 ./net/socket.c:1188
 [<ffffffff81662f5d>] __fput+0x21d/0x6e0 ./fs/file_table.c:265
 [<ffffffff816634a5>] ____fput+0x15/0x20 ./fs/file_table.c:84
 [<ffffffff812a33d3>] task_work_run+0x163/0x1f0 ./include/trace/events/rcu.h:20
 [<     inline     >] __list_add ./include/linux/list.h:42
 [<     inline     >] list_add_tail ./include/linux/list.h:76
 [<     inline     >] list_move_tail ./include/linux/list.h:168
 [<     inline     >] reparent_leader ./kernel/exit.c:618
 [<     inline     >] forget_original_parent ./kernel/exit.c:669
 [<     inline     >] exit_notify ./kernel/exit.c:697
 [<ffffffff812505d9>] do_exit+0x809/0x2b90 ./kernel/exit.c:878
 [<ffffffff81252ad8>] do_group_exit+0x108/0x320 ./kernel/exit.c:985
 [<ffffffff81252d0d>] SyS_exit_group+0x1d/0x20 ./kernel/exit.c:1002
 [<ffffffff84bf0c36>] entry_SYSCALL_64_fastpath+0x16/0x7a
=================================

Quoted 1 line

Quoted 1 line

I am on commit 90b55590c43258a157a2a143748455dcc50fbb53 of net-next (Nov 22).

The right commit is:

commit 7d267278a9ece963d77eefec61630223fce08c6c

Author: Rainer Weikusat
Date:   Fri Nov 20 22:07:23 2015 +0000

unix: avoid use-after-free in ep_remove_wait_queue

Quoted 1 line

This commit doesn't seem to exist

I don't think this matters... I think what's happening is that a close is happening on a socket still in connection initialization phase and we've never handled that particularly well...

Net-next kernel with mem debugging hangs on boot for me with a ton of printks suppressed. Will try the net kernel to see if that's better

This also seems to lead the the following WARNINGS:

------------[ cut here ]------------
WARNING: CPU: 3 PID: 21734 at kernel/jump_label.c:77
__static_key_slow_dec+0xfb/0x120()
jump label: negative count!
Modules linked in:
CPU: 3 PID: 21734 Comm: executor Tainted: G    B   W       4.4.0-rc2+ #3
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
 00000000ffffffff ffff88006083f660 ffffffff82719fc6 ffff88006083f6d0
 ffff88003bbf8000 ffffffff85a612e0 ffff88006083f6a0 ffffffff81244ec9
 ffffffff8152c54b ffffed000c107ed6 ffffffff85a612e0 000000000000004d
Call Trace:
 [<     inline     >] __dump_stack lib/dump_stack.c:15
 [<ffffffff82719fc6>] dump_stack+0x68/0x92 lib/dump_stack.c:50
 [<ffffffff81244ec9>] warn_slowpath_common+0xd9/0x140 kernel/panic.c:460
 [<ffffffff81244fd9>] warn_slowpath_fmt+0xa9/0xd0 kernel/panic.c:472
 [<ffffffff8152c54b>] __static_key_slow_dec+0xfb/0x120 kernel/jump_label.c:76
 [<ffffffff8152c5c1>] static_key_slow_dec+0x51/0x90 kernel/jump_label.c:100
 [<ffffffff84962d9b>] net_disable_timestamp+0x3b/0x50 net/core/dev.c:1709
 [<ffffffff84914d43>] sock_disable_timestamp+0x93/0xb0 net/core/sock.c:444
 [<ffffffff8491f82c>] sk_destruct+0xec/0x440 net/core/sock.c:1457
 [<ffffffff8491fbd7>] __sk_free+0x57/0x200 net/core/sock.c:1476
 [<ffffffff8491fdb0>] sk_free+0x30/0x40 net/core/sock.c:1487
 [<     inline     >] sock_put include/net/sock.h:1623
 [<ffffffff854c8a18>] sctp_close+0x628/0x790 net/sctp/socket.c:1546
 [<ffffffff84d4b3ed>] inet_release+0xed/0x1c0 net/ipv4/af_inet.c:413
 [<ffffffff84e70240>] inet6_release+0x50/0x70 net/ipv6/af_inet6.c:406
 [<ffffffff84909bbd>] sock_release+0x8d/0x1d0 net/socket.c:571
 [<ffffffff84909d16>] sock_close+0x16/0x20 net/socket.c:1022
 [<ffffffff81663a00>] __fput+0x220/0x770 fs/file_table.c:208
 [<ffffffff81663fd5>] ____fput+0x15/0x20 fs/file_table.c:244
 [<ffffffff8129f673>] task_work_run+0x163/0x1f0 kernel/task_work.c:115
 [<     inline     >] exit_task_work include/linux/task_work.h:21
 [<ffffffff8124d9e9>] do_exit+0x809/0x2ae0 kernel/exit.c:750
 [<ffffffff8124fe38>] do_group_exit+0x108/0x320 kernel/exit.c:880
 [<ffffffff81271df7>] get_signal+0x597/0x1630 kernel/signal.c:2307
 [<ffffffff8114c77f>] do_signal+0x7f/0x18e0 arch/x86/kernel/signal.c:709
 [<ffffffff81003901>] exit_to_usermode_loop+0xf1/0x1a0
arch/x86/entry/common.c:247
 [<     inline     >] prepare_exit_to_usermode arch/x86/entry/common.c:282
 [<ffffffff8100616f>] syscall_return_slowpath+0x19f/0x210
arch/x86/entry/common.c:344
 [<ffffffff85955362>] int_ret_from_sys_call+0x25/0x9f
arch/x86/entry/entry_64.S:281
---[ end trace 3e42717665ff2020 ]---

These WARNINGS always go with the original use-after-free reports. And I was not able to reproduce this WARNING with commented out sctp_association_destroy.

For the reference here is syzkaller program that triggers the WARNING.

r0 = socket(0xa, 0x1, 0x84)
mmap(&(0x7f0000000000)=nil, (0x10000), 0x3, 0x32, 0xffffffffffffffff, 0x0)
bind(r0, &(0x7f0000000000)="0a0033e049d02e70000000000000000000000000000000014c37ffc4",
0x1c)
connect(r0, &(0x7f0000001000)="020033d97f000001000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000", 0x80)
setsockopt$sock_int(r0, 0x1, 0x1d, &(0x7f0000001000+0x336)=0x1, 0x4)
listen(r0, 0xbb3)
r1 = accept(r0, &(0x7f0000003000+0xfd6)=nil, &(0x7f0000004000-0x2)=nil)

Quoted 54 lines · Dmitry Vyukov

This also seems to lead the the following WARNINGS:

------------[ cut here ]------------
WARNING: CPU: 3 PID: 21734 at kernel/jump_label.c:77
__static_key_slow_dec+0xfb/0x120()
jump label: negative count!
Modules linked in:
CPU: 3 PID: 21734 Comm: executor Tainted: G    B   W       4.4.0-rc2+ #3
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
 00000000ffffffff ffff88006083f660 ffffffff82719fc6 ffff88006083f6d0
 ffff88003bbf8000 ffffffff85a612e0 ffff88006083f6a0 ffffffff81244ec9
 ffffffff8152c54b ffffed000c107ed6 ffffffff85a612e0 000000000000004d
Call Trace:
 [<     inline     >] __dump_stack lib/dump_stack.c:15
 [<ffffffff82719fc6>] dump_stack+0x68/0x92 lib/dump_stack.c:50
 [<ffffffff81244ec9>] warn_slowpath_common+0xd9/0x140 kernel/panic.c:460
 [<ffffffff81244fd9>] warn_slowpath_fmt+0xa9/0xd0 kernel/panic.c:472
 [<ffffffff8152c54b>] __static_key_slow_dec+0xfb/0x120 kernel/jump_label.c:76
 [<ffffffff8152c5c1>] static_key_slow_dec+0x51/0x90 kernel/jump_label.c:100
 [<ffffffff84962d9b>] net_disable_timestamp+0x3b/0x50 net/core/dev.c:1709
 [<ffffffff84914d43>] sock_disable_timestamp+0x93/0xb0 net/core/sock.c:444
 [<ffffffff8491f82c>] sk_destruct+0xec/0x440 net/core/sock.c:1457
 [<ffffffff8491fbd7>] __sk_free+0x57/0x200 net/core/sock.c:1476
 [<ffffffff8491fdb0>] sk_free+0x30/0x40 net/core/sock.c:1487
 [<     inline     >] sock_put include/net/sock.h:1623
 [<ffffffff854c8a18>] sctp_close+0x628/0x790 net/sctp/socket.c:1546
 [<ffffffff84d4b3ed>] inet_release+0xed/0x1c0 net/ipv4/af_inet.c:413
 [<ffffffff84e70240>] inet6_release+0x50/0x70 net/ipv6/af_inet6.c:406
 [<ffffffff84909bbd>] sock_release+0x8d/0x1d0 net/socket.c:571
 [<ffffffff84909d16>] sock_close+0x16/0x20 net/socket.c:1022
 [<ffffffff81663a00>] __fput+0x220/0x770 fs/file_table.c:208
 [<ffffffff81663fd5>] ____fput+0x15/0x20 fs/file_table.c:244
 [<ffffffff8129f673>] task_work_run+0x163/0x1f0 kernel/task_work.c:115
 [<     inline     >] exit_task_work include/linux/task_work.h:21
 [<ffffffff8124d9e9>] do_exit+0x809/0x2ae0 kernel/exit.c:750
 [<ffffffff8124fe38>] do_group_exit+0x108/0x320 kernel/exit.c:880
 [<ffffffff81271df7>] get_signal+0x597/0x1630 kernel/signal.c:2307
 [<ffffffff8114c77f>] do_signal+0x7f/0x18e0 arch/x86/kernel/signal.c:709
 [<ffffffff81003901>] exit_to_usermode_loop+0xf1/0x1a0
arch/x86/entry/common.c:247
 [<     inline     >] prepare_exit_to_usermode arch/x86/entry/common.c:282
 [<ffffffff8100616f>] syscall_return_slowpath+0x19f/0x210
arch/x86/entry/common.c:344
 [<ffffffff85955362>] int_ret_from_sys_call+0x25/0x9f
arch/x86/entry/entry_64.S:281
---[ end trace 3e42717665ff2020 ]---

These WARNINGS always go with the original use-after-free reports. And I was not able to reproduce this WARNING with commented out sctp_association_destroy.

For the reference here is syzkaller program that triggers the WARNING.

Quoted 9 lines

r0 = socket(0xa, 0x1, 0x84)
mmap(&(0x7f0000000000)=nil, (0x10000), 0x3, 0x32, 0xffffffffffffffff, 0x0)
bind(r0, &(0x7f0000000000)="0a0033e049d02e70000000000000000000000000000000014c37ffc4",
0x1c)
connect(r0, &(0x7f0000001000)="020033d97f000001000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000", 0x80)
setsockopt$sock_int(r0, 0x1, 0x1d, &(0x7f0000001000+0x336)=0x1, 0x4)
listen(r0, 0xbb3)
r1 = accept(r0, &(0x7f0000003000+0xfd6)=nil, &(0x7f0000004000-0x2)=nil)

These two are unrelated, actually.

Do you know if this accept() returned something? Seems so. Seems to be originated on sctp_v6_create_accept_sk() -> sctp_copy_sock():

void sctp_copy_sock(struct sock *newsk, struct sock *sk,
struct sctp_association *asoc)

{
        struct inet_sock *inet = inet_sk(sk);
        struct inet_sock *newinet;

        newsk->sk_type = sk->sk_type;
        newsk->sk_bound_dev_if = sk->sk_bound_dev_if;
        newsk->sk_flags = sk->sk_flags;               <---

As it enabled SO_TIMESTAMP on listening socket, this flag will be copied and will trigger the second net_disable_timestamp() by the time the second socket is destroyed, because it never had its enable counterpart called.

This also happens via sctp peeloff operation.

Vlad, others,

It's been a long time but this was introduced by commit 914e1c8b6980 ("sctp: Inherit all socket options from parent correctly."). This is not very consistent with how other protocols work and it will be hard to keep tracking a negative mask of flags that we can't copy.

I reviewed the list of options and I'm thinking that only SO_BINDTODEVICE is worth copying, leaving the others for the application to re-set, as it is for other protocols. So I'm thinking on simply:

(quoted diff) +1 -1

-       newsk->sk_flags = sk->sk_flags;
+       newsk->sk_flags = sk->sk_flags & SO_BINDTODEVICE;

in the above.

What do you think?

Quoted 5 lines

Vlad, others,

It's been a long time but this was introduced by commit 914e1c8b6980 ("sctp: Inherit all socket options from parent correctly."). This is

not

Quoted 5 lines

very consistent with how other protocols work and it will be hard to keep tracking a negative mask of flags that we can't copy.

I reviewed the list of options and I'm thinking that only SO_BINDTODEVICE is worth copying, leaving the others for the

application

Quoted 5 lines

to re-set, as it is for other protocols. So I'm thinking on simply:

(quoted diff) +1 -1

-       newsk->sk_flags = sk->sk_flags;
+       newsk->sk_flags = sk->sk_flags & SO_BINDTODEVICE;

Quoted 3 lines

in the above.

What do you think?

I think SO_BINDTODEVICE is not a flag ;)

#define SO_BINDTODEVICE 25

Em 3 de dezembro de 2015 15:59:10 BRST, Eric Dumazet <[email protected]> escreveu:

Quoted 22 lines · Marcelo Ricardo Leitner

Quoted 5 lines

Vlad, others,

It's been a long time but this was introduced by commit 914e1c8b6980 ("sctp: Inherit all socket options from parent correctly."). This is

not

Quoted 5 lines

very consistent with how other protocols work and it will be hard to keep tracking a negative mask of flags that we can't copy.

I reviewed the list of options and I'm thinking that only SO_BINDTODEVICE is worth copying, leaving the others for the

application

Quoted 8 lines

to re-set, as it is for other protocols. So I'm thinking on simply:

(quoted diff) +1 -1

-       newsk->sk_flags = sk->sk_flags;
+       newsk->sk_flags = sk->sk_flags & SO_BINDTODEVICE;

in the above.

What do you think?

Quoted 3 lines

I think SO_BINDTODEVICE is not a flag ;)

#define SO_BINDTODEVICE 25

Oops, indeed!
Idea persists.
Thx!

Quoted 29 lines · Marcelo

Em 3 de dezembro de 2015 15:59:10 BRST, Eric Dumazet <[email protected]> escreveu:

Quoted 25 lines · Marcelo Ricardo Leitner

Quoted 5 lines

Vlad, others,

It's been a long time but this was introduced by commit 914e1c8b6980 ("sctp: Inherit all socket options from parent correctly."). This is

not

Quoted 5 lines

very consistent with how other protocols work and it will be hard to keep tracking a negative mask of flags that we can't copy.

I reviewed the list of options and I'm thinking that only SO_BINDTODEVICE is worth copying, leaving the others for the

application

Quoted 8 lines

to re-set, as it is for other protocols. So I'm thinking on simply:

(quoted diff) +1 -1

-       newsk->sk_flags = sk->sk_flags;
+       newsk->sk_flags = sk->sk_flags & SO_BINDTODEVICE;

in the above.

What do you think?

I think SO_BINDTODEVICE is not a flag ;)

#define SO_BINDTODEVICE 25

Quoted 4 lines

Oops, indeed!
Idea persists.
Thx!

Hmm... sk_clone_lock() appears to copy the flags as well, so it would appear the tcp accept() sockets would also have timestamping set.

I can see how we probably shouldn't being copying sk_flags as there isn't much there that need to be set.

SCTP echoes a cookie o INIT ACK chunks that contains a timestamp, for detecting stale cookies. This cookie is echoed back to the server by the client and then that timestamp is checked.

Thing is, if the listening socket is using packet timestamping, the cookie is encoded with ktime_get() value and checked against ktime_get_real(), as done by __net_timestamp().

The fix is to sctp also use ktime_get_real(), so we can compare bananas with bananas later no matter if packet timestamping was enabled or not.

Fixes 52db882f3fc2 ("net: sctp: migrate cookie life from timeval to ktime")

Signed-off-by Marcelo Ricardo Leitner <[email protected]>

---
 net/sctp/sm_make_chunk.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

net/sctp/sm_make_chunk.c +2 -2

index 763e06a55155b2a9e0a9d918ecc1fe2dd6d9e0c0..5d6a03fad3789a12290f5f14c5a7efa69c98f41a 100644
@@ -1652,7 +1652,7 @@ static sctp_cookie_param_t *sctp_pack_cookie(const struct sctp_endpoint *ep,
 
 	/* Set an expiration time for the cookie.  */
 	cookie->c.expiration = ktime_add(asoc->cookie_life,
-					 ktime_get());
+					 ktime_get_real());
 
 	/* Copy the peer's init packet.  */
 	memcpy(&cookie->c.peer_init[0], init_chunk->chunk_hdr,

@@ -1780,7 +1780,7 @@ no_hmac:
 	if (sock_flag(ep->base.sk, SOCK_TIMESTAMP))
 		kt = skb_get_ktime(skb);
 	else
-		kt = ktime_get();
+		kt = ktime_get_real();
 
 	if (!asoc && ktime_before(bear_cookie->expiration, kt)) {
 		/*

Dmitry Vyukov reported that SCTP was triggering a WARN on socket destroy related to disabling sock timestamp.

When SCTP accepts an association or peel one off, it copies sock flags but forgot to call net_enable_timestamp() if a packet timestamping flag was copied, leading to extra calls to net_disable_timestamp() whenever such clones were closed.

The fix is to call net_enable_timestamp() whenever we copy a sock with that flag on, like tcp does.

Reported-by Dmitry Vyukov <[email protected]>

Signed-off-by Marcelo Ricardo Leitner <[email protected]>

---
 include/net/sock.h | 2 ++
 net/core/sock.c    | 2 --
 net/sctp/socket.c  | 3 +++
 3 files changed, 5 insertions(+), 2 deletions(-)

include/net/sock.h +2 -0

index 52d27ee924f47867026d8f65c65551a9137219d3..b1d475b5db6825e13df3e3e147fed8654e1cf086 100644
@@ -740,6 +740,8 @@ enum sock_flags {
 	SOCK_SELECT_ERR_QUEUE, /* Wake select on error queue */
 };
 
+#define SK_FLAGS_TIMESTAMP ((1UL << SOCK_TIMESTAMP) | (1UL << SOCK_TIMESTAMPING_RX_SOFTWARE))
+
 static inline void sock_copy_flags(struct sock *nsk, struct sock *osk)
 {
 	nsk->sk_flags = osk->sk_flags;

net/core/sock.c +0 -2

index e31dfcee1729aa23bdd2ed692fda1b90bd75afb8..d01c8f42dbb2f040fd48009b2767bd4e80aea8ab 100644
@@ -433,8 +433,6 @@ static bool sock_needs_netstamp(const struct sock *sk)
 	}
 }
 
-#define SK_FLAGS_TIMESTAMP ((1UL << SOCK_TIMESTAMP) | (1UL << SOCK_TIMESTAMPING_RX_SOFTWARE))
-
 static void sock_disable_timestamp(struct sock *sk, unsigned long flags)
 {
 	if (sk->sk_flags & flags) {

net/sctp/socket.c +3 -0

index 03c8256063ec6355fcce034366aa5d005d75b5f7..4c9282bdd06790a0cca7f7c33986e7eb6c541398 100644
@@ -7199,6 +7199,9 @@ void sctp_copy_sock(struct sock *newsk, struct sock *sk,
 	newinet->mc_ttl = 1;
 	newinet->mc_index = 0;
 	newinet->mc_list = NULL;
+
+	if (newsk->sk_flags & SK_FLAGS_TIMESTAMP)
+		net_enable_timestamp();
 }
 
 static inline void sctp_copy_descendant(struct sock *sk_to,

As we are keeping timestamps on when copying the socket, we also have to copy sk_tsflags.

This is needed since b9f40e21ef42 ("net-timestamp: move timestamp flags out of sk_flags").

Signed-off-by Marcelo Ricardo Leitner <[email protected]>

---
 net/sctp/socket.c | 1 +
 1 file changed, 1 insertion(+)

net/sctp/socket.c +1 -0

index 4c9282bdd06790a0cca7f7c33986e7eb6c541398..1a32ecdb8bae98de2e76591f0f5ffee1441ff04d 100644
@@ -7167,6 +7167,7 @@ void sctp_copy_sock(struct sock *newsk, struct sock *sk,
 	newsk->sk_type = sk->sk_type;
 	newsk->sk_bound_dev_if = sk->sk_bound_dev_if;
 	newsk->sk_flags = sk->sk_flags;
+	newsk->sk_tsflags = sk->sk_tsflags;
 	newsk->sk_no_check_tx = sk->sk_no_check_tx;
 	newsk->sk_no_check_rx = sk->sk_no_check_rx;
 	newsk->sk_reuse = sk->sk_reuse;

These a couple of fixes regarding sctp/packet timestamps.

Dmitry Vyukov <[email protected]> reported the counter leak on missing net_enable_timestamp() (2nd patch) and further testing here revealed the other two issues.

Please consider these to -stable.

Call Trace:
 [<ffffffff8162131e>] __asan_report_load4_noabort+0x3e/0x40
 [<ffffffff8475ac76>] sctp_do_sm+0x42f6/0x4f60
 [<ffffffff847b50e9>] sctp_primitive_SHUTDOWN+0xa9/0xd0
 [<ffffffff847a1426>] sctp_close+0x616/0x790
 [<ffffffff8409bb0d>] inet_release+0xed/0x1c0 ./net/ipv4/af_inet.c:471
 [<ffffffff84192cc0>] inet6_release+0x50/0x70 ./net/ipv6/af_inet6.c:416
 [<     inline     >] constant_test_bit ././arch/x86/include/asm/bitops.h:321
 [<ffffffff83dc78cd>] sock_release+0x8d/0x200 ./net/socket.c:601
 [<ffffffff83dc7a56>] sock_close+0x16/0x20 ./net/socket.c:1188
 [<ffffffff81662f5d>] __fput+0x21d/0x6e0 ./fs/file_table.c:265
 [<ffffffff816634a5>] ____fput+0x15/0x20 ./fs/file_table.c:84
 [<ffffffff812a33d3>] task_work_run+0x163/0x1f0 ./include/trace/events/rcu.h:20
 [<     inline     >] __list_add ./include/linux/list.h:42

Hi,

Quoted 15 lines · Dmitry Vyukov

Call Trace:
 [<ffffffff8162131e>] __asan_report_load4_noabort+0x3e/0x40
 [<ffffffff8475ac76>] sctp_do_sm+0x42f6/0x4f60
 [<ffffffff847b50e9>] sctp_primitive_SHUTDOWN+0xa9/0xd0
 [<ffffffff847a1426>] sctp_close+0x616/0x790
 [<ffffffff8409bb0d>] inet_release+0xed/0x1c0 ./net/ipv4/af_inet.c:471
 [<ffffffff84192cc0>] inet6_release+0x50/0x70 ./net/ipv6/af_inet6.c:416
 [<     inline     >] constant_test_bit ././arch/x86/include/asm/bitops.h:321
 [<ffffffff83dc78cd>] sock_release+0x8d/0x200 ./net/socket.c:601
 [<ffffffff83dc7a56>] sock_close+0x16/0x20 ./net/socket.c:1188
 [<ffffffff81662f5d>] __fput+0x21d/0x6e0 ./fs/file_table.c:265
 [<ffffffff816634a5>] ____fput+0x15/0x20 ./fs/file_table.c:84
 [<ffffffff812a33d3>] task_work_run+0x163/0x1f0 ./include/trace/events/rcu.h:20
 [<     inline     >] __list_add ./include/linux/list.h:42

By any chance, did you have the pr_debug()s enabled? Because that would trigger a use-after-free on debug_post_sfx() macro expansion when the asoc is freed:

#define debug_post_sfx() \
        pr_debug("%s[post-sfx]: error:%d, asoc:%p[%s]\n", __func__, error, \
                 asoc, sctp_state_tbl[(asoc && sctp_id2assoc(ep->base.sk, \
                 sctp_assoc2id(asoc))) ? asoc->state : SCTP_STATE_CLOSED])

No, I don't. But pr_debug always computes its arguments. See no_printk in printk.h. So this use-after-free happens for all users.

Quoted 3 lines

No, I don't. But pr_debug always computes its arguments. See no_printk in printk.h. So this use-after-free happens for all users.

Hmm.

pr_debug() should be a nop unless either DEBUG or CONFIG_DYNAMIC_DEBUG are set

On our production kernels, pr_debug() is a nop.

Can you double check ? Thanks !

Quoted 10 lines · Eric Dumazet

Quoted 3 lines

No, I don't. But pr_debug always computes its arguments. See no_printk in printk.h. So this use-after-free happens for all users.

Hmm.

pr_debug() should be a nop unless either DEBUG or CONFIG_DYNAMIC_DEBUG are set

On our production kernels, pr_debug() is a nop.

Quoted 1 line

Can you double check ? Thanks !

Why should it be nop? no_printk thing in printk.h pretty much explicitly makes it not a nop...

Double-checked: debug_post_sfx leads to some generated code:

Quoted 11 lines · Eric Dumazet

Quoted 3 lines

No, I don't. But pr_debug always computes its arguments. See no_printk in printk.h. So this use-after-free happens for all users.

Hmm.

pr_debug() should be a nop unless either DEBUG or CONFIG_DYNAMIC_DEBUG are set

On our production kernels, pr_debug() is a nop.

Can you double check ? Thanks !

Why should it be nop? no_printk thing in printk.h pretty much explicitly makes it not a nop...

Double-checked: debug_post_sfx leads to some generated code:

        debug_post_sfx();
ffffffff8229f256:       48 8b 85 58 fe ff ff    mov    -0x1a8(%rbp),%rax
ffffffff8229f25d:       48 85 c0                test   %rax,%rax
ffffffff8229f260:       74 24                   je
ffffffff8229f286 <sctp_do_sm+0x176>
ffffffff8229f262:       8b b0 a8 00 00 00       mov    0xa8(%rax),%esi
ffffffff8229f268:       48 8b 85 60 fe ff ff    mov    -0x1a0(%rbp),%rax
ffffffff8229f26f:       44 89 85 74 fe ff ff    mov    %r8d,-0x18c(%rbp)
ffffffff8229f276:       48 8b 78 20             mov    0x20(%rax),%rdi
ffffffff8229f27a:       e8 71 28 01 00          callq

ffffffff822b1af0 <sctp_id2assoc>
ffffffff8229f27f: 44 8b 85 74 fe ff ff mov -0x18c(%rbp),%r8d

        return error;
}
ffffffff8229f286:       48 81 c4 a0 01 00 00    add    $0x1a0,%rsp
ffffffff8229f28d:       44 89 c0                mov    %r8d,%eax
ffffffff8229f290:       5b                      pop    %rbx
ffffffff8229f291:       41 5c                   pop    %r12
ffffffff8229f293:       41 5d                   pop    %r13
ffffffff8229f295:       41 5e                   pop    %r14
ffffffff8229f297:       41 5f                   pop    %r15
ffffffff8229f299:       5d                      pop    %rbp
ffffffff8229f29a:       c3                      retq

Quoted 32 lines · Dmitry Vyukov

Quoted 11 lines · Eric Dumazet

Quoted 3 lines

No, I don't. But pr_debug always computes its arguments. See no_printk in printk.h. So this use-after-free happens for all users.

Hmm.

pr_debug() should be a nop unless either DEBUG or CONFIG_DYNAMIC_DEBUG are set

On our production kernels, pr_debug() is a nop.

Can you double check ? Thanks !

Why should it be nop? no_printk thing in printk.h pretty much explicitly makes it not a nop...

Double-checked: debug_post_sfx leads to some generated code:

        debug_post_sfx();
ffffffff8229f256:       48 8b 85 58 fe ff ff    mov    -0x1a8(%rbp),%rax
ffffffff8229f25d:       48 85 c0                test   %rax,%rax
ffffffff8229f260:       74 24                   je
ffffffff8229f286 <sctp_do_sm+0x176>
ffffffff8229f262:       8b b0 a8 00 00 00       mov    0xa8(%rax),%esi
ffffffff8229f268:       48 8b 85 60 fe ff ff    mov    -0x1a0(%rbp),%rax
ffffffff8229f26f:       44 89 85 74 fe ff ff    mov    %r8d,-0x18c(%rbp)
ffffffff8229f276:       48 8b 78 20             mov    0x20(%rax),%rdi
ffffffff8229f27a:       e8 71 28 01 00          callq

ffffffff822b1af0 <sctp_id2assoc>
ffffffff8229f27f: 44 8b 85 74 fe ff ff mov -0x18c(%rbp),%r8d

Quoted 11 lines

        return error;
}
ffffffff8229f286:       48 81 c4 a0 01 00 00    add    $0x1a0,%rsp
ffffffff8229f28d:       44 89 c0                mov    %r8d,%eax
ffffffff8229f290:       5b                      pop    %rbx
ffffffff8229f291:       41 5c                   pop    %r12
ffffffff8229f293:       41 5d                   pop    %r13
ffffffff8229f295:       41 5e                   pop    %r14
ffffffff8229f297:       41 5f                   pop    %r15
ffffffff8229f299:       5d                      pop    %rbp
ffffffff8229f29a:       c3                      retq

This is a serious concern, because we let in the past lot of patches converting traditional

#ifdef DEBUG

# define some_hand_coded_ugly_debug()  printk( ...._
#else
# define some_hand_coded_ugly_debug()
#endif

On the premise pr_debug() would be a nop.

It seems it is not always the case. This is a very serious problem.

We probably have hundred of potential bugs, because few people actually make sure all debugging stuff is correct,
like comments can be wrong because they are not updated properly as time flies.

It is definitely a nop for many cases.

(quoted diff) +6 -1

+void eric_test_pr_debug(struct sock *sk)
+{
+       if (atomic_read(&sk->sk_omem_alloc))
+               pr_debug("%s: optmem leakage for sock %p\n",
+                        __func__, sk);
+}
 
->

0000000000004740 <eric_test_pr_debug>:

    4740: e8 00 00 00 00       callq  4745 <eric_test_pr_debug+0x5>

4741: R_X86_64_PC32 __fentry__-0x4
4745: 55 push %rbp
4746: 8b 87 24 01 00 00 mov 0x124(%rdi),%eax //
atomic_read() but nothing follows

    474c: 48 89 e5             mov    %rsp,%rbp
    474f: 5d                   pop    %rbp
    4750: c3                   retq

Quoted 11 lines · Eric Dumazet

Quoted 10 lines · Dmitry Vyukov

Quoted 9 lines · Eric Dumazet

Quoted 3 lines

No, I don't. But pr_debug always computes its arguments. See no_printk in printk.h. So this use-after-free happens for all users.

Hmm.

pr_debug() should be a nop unless either DEBUG or
CONFIG_DYNAMIC_DEBUG are set

Quoted 7 lines

Quoted 7 lines

Quoted 3 lines

On our production kernels, pr_debug() is a nop.

Can you double check ? Thanks !

Why should it be nop? no_printk thing in printk.h pretty much explicitly makes it not a nop...

Quoted 28 lines

Quoted 27 lines

Double-checked: debug_post_sfx leads to some generated code:

        debug_post_sfx();
ffffffff8229f256:       48 8b 85 58 fe ff ff    mov    -0x1a8(%rbp),%rax
ffffffff8229f25d:       48 85 c0                test   %rax,%rax
ffffffff8229f260:       74 24                   je
ffffffff8229f286 <sctp_do_sm+0x176>
ffffffff8229f262:       8b b0 a8 00 00 00       mov    0xa8(%rax),%esi
ffffffff8229f268:       48 8b 85 60 fe ff ff    mov    -0x1a0(%rbp),%rax
ffffffff8229f26f:       44 89 85 74 fe ff ff    mov    %r8d,-0x18c(%rbp)
ffffffff8229f276:       48 8b 78 20             mov    0x20(%rax),%rdi
ffffffff8229f27a:       e8 71 28 01 00          callq

ffffffff822b1af0 <sctp_id2assoc>
ffffffff8229f27f: 44 8b 85 74 fe ff ff mov -0x18c(%rbp),%r8d

        return error;
}
ffffffff8229f286:       48 81 c4 a0 01 00 00    add    $0x1a0,%rsp
ffffffff8229f28d:       44 89 c0                mov    %r8d,%eax
ffffffff8229f290:       5b                      pop    %rbx
ffffffff8229f291:       41 5c                   pop    %r12
ffffffff8229f293:       41 5d                   pop    %r13
ffffffff8229f295:       41 5e                   pop    %r14
ffffffff8229f297:       41 5f                   pop    %r15
ffffffff8229f299:       5d                      pop    %rbp
ffffffff8229f29a:       c3                      retq

Quoted 2 lines

This is a serious concern, because we let in the past lot of patches converting traditional

Quoted 9 lines

#ifdef DEBUG

# define some_hand_coded_ugly_debug()  printk( ...._
#else
# define some_hand_coded_ugly_debug()
#endif

On the premise pr_debug() would be a nop.

It seems it is not always the case. This is a very serious problem.

Quoted 17 lines

We probably have hundred of potential bugs, because few people actually make sure all debugging stuff is correct,
like comments can be wrong because they are not updated properly as time flies.

It is definitely a nop for many cases.

(quoted diff) +6 -1

+void eric_test_pr_debug(struct sock *sk)
+{
+       if (atomic_read(&sk->sk_omem_alloc))
+               pr_debug("%s: optmem leakage for sock %p\n",
+                        __func__, sk);
+}
 
->

0000000000004740 <eric_test_pr_debug>:

Quoted 8 lines

    4740: e8 00 00 00 00       callq  4745 <eric_test_pr_debug+0x5>

4741: R_X86_64_PC32 __fentry__-0x4
4745: 55 push %rbp
4746: 8b 87 24 01 00 00 mov 0x124(%rdi),%eax //
atomic_read() but nothing follows

    474c: 48 89 e5             mov    %rsp,%rbp
    474f: 5d                   pop    %rbp
    4750: c3                   retq

I would expect that it is nop when argument evaluation does not have

side-effects. For example, for a load of a variable compiler will most likely elide it (though, it does not have to elide it, because the load is spelled in the code, so it can also legally emit the load and doesn't use the result).
But if argument computation has side-effect (or compiler can't prove otherwise), it must emit code. It must emit code for function calls when the function is defined in a different translation unit, and for volatile accesses (most likely including atomic accesses), etc

Quoted 12 lines · Dmitry Vyukov

Quoted 11 lines · Eric Dumazet

Quoted 10 lines · Dmitry Vyukov

Quoted 9 lines · Eric Dumazet

Quoted 3 lines

No, I don't. But pr_debug always computes its arguments. See no_printk in printk.h. So this use-after-free happens for all users.

Hmm.

pr_debug() should be a nop unless either DEBUG or
CONFIG_DYNAMIC_DEBUG are set

Quoted 7 lines

Quoted 7 lines

Quoted 7 lines

Quoted 3 lines

On our production kernels, pr_debug() is a nop.

Can you double check ? Thanks !

Why should it be nop? no_printk thing in printk.h pretty much explicitly makes it not a nop...

Because it was until commit 5264f2f75d8. It also violates my reading of the following from printk.h:

 * All of these will print unconditionally, although note that pr_debug()
 * and other debug macros are compiled out unless either DEBUG is defined
 * or CONFIG_DYNAMIC_DEBUG is set.

Quoted 28 lines

Quoted 28 lines

Quoted 27 lines

Double-checked: debug_post_sfx leads to some generated code:

        debug_post_sfx();
ffffffff8229f256:       48 8b 85 58 fe ff ff    mov    -0x1a8(%rbp),%rax ffffffff8229f25d:       48 85 c0                test   %rax,%rax ffffffff8229f260:       74 24                   je ffffffff8229f286
ffffffff8229f262:       8b b0 a8 00 00 00       mov    0xa8(%rax),%esi ffffffff8229f268:       48 8b 85 60 fe ff ff    mov    -0x1a0(%rbp),%rax ffffffff8229f26f:       44 89 85 74 fe ff ff    mov    %r8d,-0x18c(%rbp) ffffffff8229f276:       48 8b 78 20             mov    0x20(%rax),%rdi ffffffff8229f27a:       e8 71 28 01 00          callq ffffffff822b1af0
ffffffff8229f27f:       44 8b 85 74 fe ff ff    mov    -0x18c(%rbp),%r8d

        return error;
}
ffffffff8229f286:       48 81 c4 a0 01 00 00    add    $0x1a0,%rsp ffffffff8229f28d:       44 89 c0                mov    %r8d,%eax ffffffff8229f290:       5b                      pop    %rbx ffffffff8229f291:       41 5c                   pop    %r12 ffffffff8229f293:       41 5d                   pop    %r13 ffffffff8229f295:       41 5e                   pop    %r14 ffffffff8229f297:       41 5f                   pop    %r15 ffffffff8229f299:       5d                      pop    %rbp ffffffff8229f29a:       c3                      retq

Quoted 2 lines

Quoted 2 lines

This is a serious concern, because we let in the past lot of patches converting traditional

+1

Quoted 9 lines

Quoted 9 lines

#ifdef DEBUG

# define some_hand_coded_ugly_debug()  printk( ...._
#else
# define some_hand_coded_ugly_debug()
#endif

On the premise pr_debug() would be a nop.

It seems it is not always the case. This is a very serious problem.

+1

Quoted 29 lines

Quoted 16 lines

We probably have hundred of potential bugs, because few people actually make sure all debugging stuff is correct,
like comments can be wrong because they are not updated properly as time flies.

It is definitely a nop for many cases.

(quoted diff) +6 -1

+void eric_test_pr_debug(struct sock *sk)
+{
+       if (atomic_read(&sk->sk_omem_alloc))
+               pr_debug("%s: optmem leakage for sock %p\n",
+                        __func__, sk);
+}
 
->

Quoted 9 lines

0000000000004740 :
    4740: e8 00 00 00 00       callq  4745
4741: R_X86_64_PC32 __fentry__-0x4
    4745: 55                   push   %rbp
    4746: 8b 87 24 01 00 00     mov    0x124(%rdi),%eax     // atomic_read()  but nothing follows
    474c: 48 89 e5             mov    %rsp,%rbp 474f: 5d                   pop    %rbp
    4750: c3                   retq

I would expect that it is nop when argument evaluation does not have

Quoted 8 lines

side-effects. For example, for a load of a variable compiler will most likely elide it (though, it does not have to elide it, because the load is spelled in the code, so it can also legally emit the load and doesn't use the result).
But if argument computation has side-effect (or compiler can't prove otherwise), it must emit code. It must emit code for function calls when the function is defined in a different translation unit, and for volatile accesses (most likely including atomic accesses), etc

This isn't 100% true. As you state, in order to reach the return 0, all side effects must be evaluated. Load generally does not have side effects, so it can be safely elided, but function() must be emitted.

However, that is _not_ required to get the desired warning emission on a printf argument function, see http://pastebin.com/UHuaydkj for an example.

I think that as a minimum, the following patch should be evaluted, but am unsure to whom I should submit it (after I test):

Quoted 12 lines · Dmitry Vyukov

Quoted 11 lines · Eric Dumazet

Quoted 10 lines · Dmitry Vyukov

Quoted 9 lines · Eric Dumazet

Quoted 3 lines

No, I don't. But pr_debug always computes its arguments. See no_printk in printk.h. So this use-after-free happens for all users.

Hmm.

pr_debug() should be a nop unless either DEBUG or
CONFIG_DYNAMIC_DEBUG are set

Quoted 7 lines

Quoted 7 lines

Quoted 7 lines

Quoted 3 lines

On our production kernels, pr_debug() is a nop.

Can you double check ? Thanks !

Why should it be nop? no_printk thing in printk.h pretty much explicitly makes it not a nop...

Because it was until commit 5264f2f75d8. It also violates my reading of the following from printk.h:

* All of these will print unconditionally, although note that pr_debug()
* and other debug macros are compiled out unless either DEBUG is defined
* or CONFIG_DYNAMIC_DEBUG is set.

Quoted 28 lines

Quoted 28 lines

Quoted 27 lines

Double-checked: debug_post_sfx leads to some generated code:

        debug_post_sfx();
ffffffff8229f256:       48 8b 85 58 fe ff ff    mov    -0x1a8(%rbp),%rax
ffffffff8229f25d:       48 85 c0                test   %rax,%rax
ffffffff8229f260:       74 24                   je
ffffffff8229f286 <sctp_do_sm+0x176>
ffffffff8229f262:       8b b0 a8 00 00 00       mov    0xa8(%rax),%esi
ffffffff8229f268:       48 8b 85 60 fe ff ff    mov    -0x1a0(%rbp),%rax
ffffffff8229f26f:       44 89 85 74 fe ff ff    mov    %r8d,-0x18c(%rbp)
ffffffff8229f276:       48 8b 78 20             mov    0x20(%rax),%rdi
ffffffff8229f27a:       e8 71 28 01 00          callq

ffffffff822b1af0 <sctp_id2assoc>
ffffffff8229f27f: 44 8b 85 74 fe ff ff mov -0x18c(%rbp),%r8d

        return error;
}
ffffffff8229f286:       48 81 c4 a0 01 00 00    add    $0x1a0,%rsp
ffffffff8229f28d:       44 89 c0                mov    %r8d,%eax
ffffffff8229f290:       5b                      pop    %rbx
ffffffff8229f291:       41 5c                   pop    %r12
ffffffff8229f293:       41 5d                   pop    %r13
ffffffff8229f295:       41 5e                   pop    %r14
ffffffff8229f297:       41 5f                   pop    %r15
ffffffff8229f299:       5d                      pop    %rbp
ffffffff8229f29a:       c3                      retq

Quoted 2 lines

Quoted 2 lines

This is a serious concern, because we let in the past lot of patches converting traditional

+1

Quoted 9 lines

Quoted 9 lines

#ifdef DEBUG

# define some_hand_coded_ugly_debug()  printk( ...._
#else
# define some_hand_coded_ugly_debug()
#endif

On the premise pr_debug() would be a nop.

It seems it is not always the case. This is a very serious problem.

+1

Quoted 29 lines

Quoted 17 lines

We probably have hundred of potential bugs, because few people actually make sure all debugging stuff is correct,
like comments can be wrong because they are not updated properly as time flies.

It is definitely a nop for many cases.

(quoted diff) +6 -1

+void eric_test_pr_debug(struct sock *sk)
+{
+       if (atomic_read(&sk->sk_omem_alloc))
+               pr_debug("%s: optmem leakage for sock %p\n",
+                        __func__, sk);
+}
 
->

0000000000004740 <eric_test_pr_debug>:

Quoted 8 lines

    4740: e8 00 00 00 00       callq  4745 <eric_test_pr_debug+0x5>

4741: R_X86_64_PC32 __fentry__-0x4
4745: 55 push %rbp
4746: 8b 87 24 01 00 00 mov 0x124(%rdi),%eax //
atomic_read() but nothing follows

    474c: 48 89 e5             mov    %rsp,%rbp
    474f: 5d                   pop    %rbp
    4750: c3                   retq

I would expect that it is nop when argument evaluation does not have

Quoted 8 lines

side-effects. For example, for a load of a variable compiler will most likely elide it (though, it does not have to elide it, because the load is spelled in the code, so it can also legally emit the load and doesn't use the result).
But if argument computation has side-effect (or compiler can't prove otherwise), it must emit code. It must emit code for function calls when the function is defined in a different translation unit, and for volatile accesses (most likely including atomic accesses), etc

This isn't 100% true. As you state, in order to reach the return 0, all side effects must be evaluated. Load generally does not have side effects, so it can be safely elided, but function() must be emitted.

However, that is _not_ required to get the desired warning emission on a printf argument function, see http://pastebin.com/UHuaydkj for an example.

I think that as a minimum, the following patch should be evaluted, but am unsure to whom I should submit it (after I test):

include/linux/printk.h +1 -2

index 9729565..cd24d2d 100644
@@ -286,7 +286,7 @@ extern asmlinkage void dump_stack(void) __cold;
        printk(KERN_DEBUG pr_fmt(fmt), ##__VA_ARGS__)
 #else
 #define pr_debug(fmt, ...) \
-       no_printk(KERN_DEBUG pr_fmt(fmt), ##__VA_ARGS__)
+       ({ if(0) printk(KERN_DEBUG pr_fmt(fmt), ##__VA_ARGS__); 0;})
 #endif
 
 /*
--

To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

Quoted 2 lines · Aaron Conole

I think that as a minimum, the following patch should be evaluted, but am unsure to whom I should submit it (after I test):

Agreed - the intention here is certainly to have no side effects. It looks like 'no_printk()' is used in quite a few other places that would benefit from this change. So we probably want a generic 'really_no_printk()' macro.

Quoted 3 lines · Jason Baron

Quoted 2 lines · Aaron Conole

I think that as a minimum, the following patch should be evaluted, but am unsure to whom I should submit it (after I test):

[]

Quoted 4 lines

Agreed - the intention here is certainly to have no side effects. It looks like 'no_printk()' is used in quite a few other places that would benefit from this change. So we probably want a generic 'really_no_printk()' macro.

https://lkml.org/lkml/2012/6/17/231

Quoted 5 lines · Joe Perches

Quoted 3 lines · Jason Baron

Quoted 2 lines · Aaron Conole

I think that as a minimum, the following patch should be evaluted, but am unsure to whom I should submit it (after I test):

[]

Quoted 6 lines

Quoted 4 lines

Agreed - the intention here is certainly to have no side effects. It looks like 'no_printk()' is used in quite a few other places that would benefit from this change. So we probably want a generic 'really_no_printk()' macro.

https://lkml.org/lkml/2012/6/17/231

I don't see this in the tree.

Also maybe we should just convert no_printk() to do what your 'eliminated_printk()'.

So we can convert all users with this change?

Quoted 6 lines · Jason Baron

Quoted 5 lines · Joe Perches

Quoted 3 lines · Jason Baron

Quoted 2 lines · Aaron Conole

I think that as a minimum, the following patch should be evaluted, but am unsure to whom I should submit it (after I test):

[]

Quoted 8 lines

Quoted 6 lines

Quoted 4 lines

Agreed - the intention here is certainly to have no side effects. It looks like 'no_printk()' is used in quite a few other places that would benefit from this change. So we probably want a generic 'really_no_printk()' macro.

https://lkml.org/lkml/2012/6/17/231

I don't see this in the tree.

It never got applied.

Quoted 2 lines

Also maybe we should just convert no_printk() to do what your 'eliminated_printk()'.

Some of them at least.

Quoted 1 line

So we can convert all users with this change?

I don't think so, I think there are some
function evaluation/side effects that are
required. I believe some do hardware I/O.

It'd be good to at least isolate them.

I'm not sure how to find them via some
automated tool/mechanism though.

I asked Julia Lawall about it once in this
thread: https://lkml.org/lkml/2014/12/3/696

Quoted 34 lines · Joe Perches

Quoted 6 lines · Jason Baron

Quoted 5 lines · Joe Perches

Quoted 3 lines · Jason Baron

Quoted 2 lines · Aaron Conole

I think that as a minimum, the following patch should be evaluted, but am unsure to whom I should submit it (after I test):

[]

Quoted 8 lines

Quoted 6 lines

Quoted 4 lines

Agreed - the intention here is certainly to have no side effects. It looks like 'no_printk()' is used in quite a few other places that would benefit from this change. So we probably want a generic 'really_no_printk()' macro.

https://lkml.org/lkml/2012/6/17/231

I don't see this in the tree.

It never got applied.

Quoted 2 lines

Also maybe we should just convert no_printk() to do what your 'eliminated_printk()'.

Some of them at least.

Quoted 1 line

So we can convert all users with this change?

I don't think so, I think there are some
function evaluation/side effects that are
required.  I believe some do hardware I/O.

It'd be good to at least isolate them.

I'm not sure how to find them via some
automated tool/mechanism though.

Quoted 3 lines

I asked Julia Lawall about it once in this
thread:  https://lkml.org/lkml/2014/12/3/696

Seems rather fragile to have side effects that we rely upon hidden in a printk().

Just convert them and see what breaks :)

(adding lkml as this is likely better discussed there)

Quoted 35 lines · Jason Baron

Quoted 34 lines · Joe Perches

Quoted 6 lines · Jason Baron

Quoted 5 lines · Joe Perches

Quoted 3 lines · Jason Baron

Quoted 2 lines · Aaron Conole

I think that as a minimum, the following patch should be evaluted, but am unsure to whom I should submit it (after I test):

[]

Quoted 8 lines

Quoted 6 lines

Quoted 4 lines

Agreed - the intention here is certainly to have no side effects. It looks like 'no_printk()' is used in quite a few other places that would benefit from this change. So we probably want a generic 'really_no_printk()' macro.

https://lkml.org/lkml/2012/6/17/231

I don't see this in the tree.

It never got applied.

Quoted 2 lines

Also maybe we should just convert no_printk() to do what your 'eliminated_printk()'.

Some of them at least.

Quoted 1 line

So we can convert all users with this change?

I don't think so, I think there are some
function evaluation/side effects that are
required. I believe some do hardware I/O.

It'd be good to at least isolate them.

I'm not sure how to find them via some
automated tool/mechanism though.

Quoted 6 lines

Quoted 3 lines

I asked Julia Lawall about it once in this
thread: https://lkml.org/lkml/2014/12/3/696

Seems rather fragile to have side effects that we rely upon hidden in a printk().

Yup.

Quoted 1 line

Just convert them and see what breaks :)

I appreciate your optimism. It's very 1995.
Try it and see what happens.

Quoted 49 lines · Joe Perches

(adding lkml as this is likely better discussed there)

Quoted 35 lines · Jason Baron

Quoted 34 lines · Joe Perches

Quoted 6 lines · Jason Baron

Quoted 5 lines · Joe Perches

Quoted 3 lines · Jason Baron

Quoted 2 lines · Aaron Conole

I think that as a minimum, the following patch should be evaluted, but am unsure to whom I should submit it (after I test):

[]

Quoted 8 lines

Quoted 6 lines

Quoted 4 lines

Agreed - the intention here is certainly to have no side effects. It looks like 'no_printk()' is used in quite a few other places that would benefit from this change. So we probably want a generic 'really_no_printk()' macro.

https://lkml.org/lkml/2012/6/17/231

I don't see this in the tree.

It never got applied.

Quoted 2 lines

Also maybe we should just convert no_printk() to do what your 'eliminated_printk()'.

Some of them at least.

Quoted 1 line

So we can convert all users with this change?

I don't think so, I think there are some
function evaluation/side effects that are
required. I believe some do hardware I/O.

It'd be good to at least isolate them.

I'm not sure how to find them via some
automated tool/mechanism though.

Quoted 6 lines

Quoted 3 lines

I asked Julia Lawall about it once in this
thread: https://lkml.org/lkml/2014/12/3/696

Seems rather fragile to have side effects that we rely upon hidden in a printk().

Yup.

Quoted 1 line

Just convert them and see what breaks :)

Quoted 2 lines

I appreciate your optimism. It's very 1995.
Try it and see what happens.

Whatever is the resolution for pr_debug, we still need to fix this particular use-after-free. It affects stability of debug builds, gives invalid debug output, prevents us from finding more bugs in SCTP. And maybe somebody uses CONFIG_DYNAMIC_DEBUG in production.

Quoted 50 lines · Dmitry Vyukov

Quoted 49 lines · Joe Perches

(adding lkml as this is likely better discussed there)

Quoted 35 lines · Jason Baron

Quoted 34 lines · Joe Perches

Quoted 6 lines · Jason Baron

Quoted 5 lines · Joe Perches

Quoted 3 lines · Jason Baron

Quoted 2 lines · Aaron Conole

I think that as a minimum, the following patch should be evaluted, but am unsure to whom I should submit it (after I test):

[]

Quoted 8 lines

Quoted 6 lines

Quoted 4 lines

Agreed - the intention here is certainly to have no side effects. It looks like 'no_printk()' is used in quite a few other places that would benefit from this change. So we probably want a generic 'really_no_printk()' macro.

https://lkml.org/lkml/2012/6/17/231

I don't see this in the tree.

It never got applied.

Quoted 2 lines

Also maybe we should just convert no_printk() to do what your 'eliminated_printk()'.

Some of them at least.

Quoted 1 line

So we can convert all users with this change?

I don't think so, I think there are some
function evaluation/side effects that are
required. I believe some do hardware I/O.

It'd be good to at least isolate them.

I'm not sure how to find them via some
automated tool/mechanism though.

Quoted 6 lines

Quoted 3 lines

I asked Julia Lawall about it once in this
thread: https://lkml.org/lkml/2014/12/3/696

Seems rather fragile to have side effects that we rely upon hidden in a printk().

Yup.

Quoted 1 line

Just convert them and see what breaks :)

Quoted 8 lines

Quoted 2 lines

I appreciate your optimism. It's very 1995.
Try it and see what happens.

Whatever is the resolution for pr_debug, we still need to fix this particular use-after-free. It affects stability of debug builds, gives invalid debug output, prevents us from finding more bugs in SCTP. And maybe somebody uses CONFIG_DYNAMIC_DEBUG in production.

Agreed. I'm already working on a fix for this particular use-after-free.

Another interesting thing about this is that sctp_do_sm() is called for nearly every movement that happens on a sctp socket. Said that, that always-running IDR search hidden on that debug statement do have some nasty performance impact, specially because it's serialized on a spinlock.

This wouldn't be happening if it was fully ellided and would be ok if that pr_debug() was really being printed, but not as it is. Kudos to this report that I could notice this. I'm trying to fix this on SCTP-side as well.

Marcelo

Quoted 60 lines · Marcelo Ricardo Leitner

Quoted 50 lines · Dmitry Vyukov

Quoted 49 lines · Joe Perches

(adding lkml as this is likely better discussed there)

Quoted 35 lines · Jason Baron

Quoted 34 lines · Joe Perches

Quoted 6 lines · Jason Baron

Quoted 5 lines · Joe Perches

Quoted 3 lines · Jason Baron

Quoted 2 lines · Aaron Conole

I think that as a minimum, the following patch should be evaluted, but am unsure to whom I should submit it (after I test):

[]

Quoted 8 lines

Quoted 6 lines

Quoted 4 lines

Agreed - the intention here is certainly to have no side effects. It looks like 'no_printk()' is used in quite a few other places that would benefit from this change. So we probably want a generic 'really_no_printk()' macro.

https://lkml.org/lkml/2012/6/17/231

I don't see this in the tree.

It never got applied.

Quoted 2 lines

Also maybe we should just convert no_printk() to do what your 'eliminated_printk()'.

Some of them at least.

Quoted 1 line

So we can convert all users with this change?

I don't think so, I think there are some
function evaluation/side effects that are
required. I believe some do hardware I/O.

It'd be good to at least isolate them.

I'm not sure how to find them via some
automated tool/mechanism though.

Quoted 6 lines

Quoted 3 lines

I asked Julia Lawall about it once in this
thread: https://lkml.org/lkml/2014/12/3/696

Seems rather fragile to have side effects that we rely upon hidden in a printk().

Yup.

Quoted 1 line

Just convert them and see what breaks :)

Quoted 8 lines

Quoted 2 lines

I appreciate your optimism. It's very 1995.
Try it and see what happens.

Whatever is the resolution for pr_debug, we still need to fix this particular use-after-free. It affects stability of debug builds, gives invalid debug output, prevents us from finding more bugs in SCTP. And maybe somebody uses CONFIG_DYNAMIC_DEBUG in production.

Quoted 7 lines

Agreed. I'm already working on a fix for this particular use-after-free.

Another interesting thing about this is that sctp_do_sm() is called for nearly every movement that happens on a sctp socket. Said that, that always-running IDR search hidden on that debug statement do have some nasty performance impact, specially because it's serialized on a spinlock.

YUCK! I didn't really pay much attention to those debug macros before, but debug_post_sfx() is truly awful.

This wasn't such a bad thing where these macros depended on CONFIG_SCTP_DEBUG, but now that they are always built, we need fix them.

-vlad

Quoted 7 lines

This wouldn't be happening if it was fully ellided and would be ok if that pr_debug() was really being printed, but not as it is. Kudos to this report that I could notice this. I'm trying to fix this on SCTP-side as well.

Marcelo

(adding lkml as this is likely better discussed there)

Quoted 35 lines · Jason Baron

Quoted 34 lines · Joe Perches

Quoted 6 lines · Jason Baron

Quoted 5 lines · Joe Perches

Quoted 3 lines · Jason Baron

Quoted 2 lines · Aaron Conole

I think that as a minimum, the following patch should be evaluted, but am unsure to whom I should submit it (after I test):

[]

Quoted 8 lines

Quoted 6 lines

Quoted 4 lines

Agreed - the intention here is certainly to have no side effects. It looks like 'no_printk()' is used in quite a few other places that would benefit from this change. So we probably want a generic 'really_no_printk()' macro.

https://lkml.org/lkml/2012/6/17/231

I don't see this in the tree.

It never got applied.

Quoted 2 lines

Also maybe we should just convert no_printk() to do what your 'eliminated_printk()'.

Some of them at least.

Quoted 1 line

So we can convert all users with this change?

I don't think so, I think there are some
function evaluation/side effects that are
required. I believe some do hardware I/O.

It'd be good to at least isolate them.

I'm not sure how to find them via some
automated tool/mechanism though.

Quoted 6 lines

Quoted 3 lines

I asked Julia Lawall about it once in this
thread: https://lkml.org/lkml/2014/12/3/696

Seems rather fragile to have side effects that we rely upon hidden in a printk().

Yup.

Quoted 1 line

Just convert them and see what breaks :)

I appreciate your optimism. It's very 1995.
Try it and see what happens.

Quoted 49 lines · Joe Perches

(adding lkml as this is likely better discussed there)

Quoted 35 lines · Jason Baron

Quoted 34 lines · Joe Perches

Quoted 6 lines · Jason Baron

Quoted 5 lines · Joe Perches

Quoted 3 lines · Jason Baron

Quoted 2 lines · Aaron Conole

I think that as a minimum, the following patch should be evaluted, but am unsure to whom I should submit it (after I test):

[]

Quoted 8 lines

Quoted 6 lines

Quoted 4 lines

Agreed - the intention here is certainly to have no side effects. It looks like 'no_printk()' is used in quite a few other places that would benefit from this change. So we probably want a generic 'really_no_printk()' macro.

https://lkml.org/lkml/2012/6/17/231

I don't see this in the tree.

It never got applied.

Quoted 2 lines

Also maybe we should just convert no_printk() to do what your 'eliminated_printk()'.

Some of them at least.

Quoted 1 line

So we can convert all users with this change?

I don't think so, I think there are some
function evaluation/side effects that are
required. I believe some do hardware I/O.

It'd be good to at least isolate them.

I'm not sure how to find them via some
automated tool/mechanism though.

Quoted 6 lines

Quoted 3 lines

I asked Julia Lawall about it once in this
thread: https://lkml.org/lkml/2014/12/3/696

Seems rather fragile to have side effects that we rely upon hidden in a printk().

Yup.

Quoted 1 line

Just convert them and see what breaks :)

Quoted 2 lines

I appreciate your optimism. It's very 1995.
Try it and see what happens.

But Aaron says that DYNAMIC_DEBUG is enabled in most major distributions, and all these side-effects don't happen with DYNAMIC_DEBUG.

This suggests that we can make these side-effects not happen without DYNAMIC_DEBUG as well.
Or I am missing something here?

When DYNAMIC_DEBUG is enabled we have this wrapper from include/linux/dynamic_debug.h:

if (unlikely(descriptor.flags & _DPRINTK_FLAGS_PRINT))
<do debug stuff>

So the compiler is not emitting the side-effects in this case.

Quoted 8 lines · Jason Baron

When DYNAMIC_DEBUG is enabled we have this wrapper from include/linux/dynamic_debug.h:

if (unlikely(descriptor.flags & _DPRINTK_FLAGS_PRINT))
<do debug stuff>

So the compiler is not emitting the side-effects in this case.

Huh? Do I misunderstand what you are writing?

You are testing a variable that is not generally set so the call is not being performed in the general case, but the compiler can not elide the code.

If the variable was enabled via the control file, the __dynamic_pr_debug would be performed with the use-after-free.

Quoted 32 lines · Dmitry Vyukov

Quoted 11 lines · Eric Dumazet

Quoted 3 lines

No, I don't. But pr_debug always computes its arguments. See no_printk in printk.h. So this use-after-free happens for all users.

Hmm.

pr_debug() should be a nop unless either DEBUG or CONFIG_DYNAMIC_DEBUG are set

On our production kernels, pr_debug() is a nop.

Can you double check ? Thanks !

Why should it be nop? no_printk thing in printk.h pretty much explicitly makes it not a nop...

Double-checked: debug_post_sfx leads to some generated code:

        debug_post_sfx();
ffffffff8229f256:       48 8b 85 58 fe ff ff    mov    -0x1a8(%rbp),%rax
ffffffff8229f25d:       48 85 c0                test   %rax,%rax
ffffffff8229f260:       74 24                   je
ffffffff8229f286 <sctp_do_sm+0x176>
ffffffff8229f262:       8b b0 a8 00 00 00       mov    0xa8(%rax),%esi
ffffffff8229f268:       48 8b 85 60 fe ff ff    mov    -0x1a0(%rbp),%rax
ffffffff8229f26f:       44 89 85 74 fe ff ff    mov    %r8d,-0x18c(%rbp)
ffffffff8229f276:       48 8b 78 20             mov    0x20(%rax),%rdi
ffffffff8229f27a:       e8 71 28 01 00          callq

ffffffff822b1af0 <sctp_id2assoc>
ffffffff8229f27f: 44 8b 85 74 fe ff ff mov -0x18c(%rbp),%r8d

Quoted 11 lines

        return error;
}
ffffffff8229f286:       48 81 c4 a0 01 00 00    add    $0x1a0,%rsp
ffffffff8229f28d:       44 89 c0                mov    %r8d,%eax
ffffffff8229f290:       5b                      pop    %rbx
ffffffff8229f291:       41 5c                   pop    %r12
ffffffff8229f293:       41 5d                   pop    %r13
ffffffff8229f295:       41 5e                   pop    %r14
ffffffff8229f297:       41 5f                   pop    %r15
ffffffff8229f299:       5d                      pop    %rbp
ffffffff8229f29a:       c3                      retq

This is a serious concern, because we let in the past lot of patches converting traditional

#ifdef DEBUG

# define some_hand_coded_ugly_debug()  printk( ...._
#else
# define some_hand_coded_ugly_debug()
#endif

On the premise pr_debug() would be a nop.

It seems it is not always the case. This is a very serious problem.

We probably have hundred of potential bugs, because few people actually make sure all debugging stuff is correct,
like comments can be wrong because they are not updated properly as time flies.

It is definitely a nop for many cases.

(quoted diff) +6 -1

+void eric_test_pr_debug(struct sock *sk)
+{
+       if (atomic_read(&sk->sk_omem_alloc))
+               pr_debug("%s: optmem leakage for sock %p\n",
+                        __func__, sk);
+}
 
->

0000000000004740 <eric_test_pr_debug>:

    4740: e8 00 00 00 00       callq  4745 <eric_test_pr_debug+0x5>

4741: R_X86_64_PC32 __fentry__-0x4
4745: 55 push %rbp
4746: 8b 87 24 01 00 00 mov 0x124(%rdi),%eax //
atomic_read() but nothing follows

    474c: 48 89 e5             mov    %rsp,%rbp
    474f: 5d                   pop    %rbp
    4750: c3                   retq

Hi Dmitry,

Can you please test this patch?
I'll re-post with proper subject if it works.

---8<---

Dmitry Vyukov reported a use-after-free in the code expanded by the macro debug_post_sfx, which is caused by the use of the asoc pointer after it was freed within sctp_side_effect() scope.

This patch fixes it by allowing sctp_side_effect to clear that asoc pointer when the TCB is freed.

The macro is already prepared to handle such NULL pointer.

Reported-by Dmitry Vyukov <[email protected]>

Signed-off-by Marcelo Ricardo Leitner <[email protected]>

---
 net/sctp/sm_sideeffect.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)

net/sctp/sm_sideeffect.c +5 -5

index 6098d4c42fa91287d3cde36ac05d860f76d4fe32..05594dcd93e0d649cace5215d225bef2713f9310 100644
@@ -63,7 +63,7 @@ static int sctp_cmd_interpreter(sctp_event_t event_type,
 static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
                             sctp_state_t state,
                             struct sctp_endpoint *ep,
-                            struct sctp_association *asoc,
+                            struct sctp_association **asoc,
                             void *event_arg,
                             sctp_disposition_t status,
                             sctp_cmd_seq_t *commands,

@@ -1123,7 +1123,7 @@ int sctp_do_sm(struct net *net, sctp_event_t event_type, sctp_subtype_t subtype,
        debug_post_sfn();
 
        error = sctp_side_effects(event_type, subtype, state,
-                                 ep, asoc, event_arg, status,
+                                 ep, &asoc, event_arg, status,
                                  &commands, gfp);
        debug_post_sfx();
 

@@ -1136,7 +1136,7 @@ int sctp_do_sm(struct net *net, sctp_event_t event_type, sctp_subtype_t subtype,
 static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
                             sctp_state_t state,
                             struct sctp_endpoint *ep,
-                            struct sctp_association *asoc,
+                            struct sctp_association **asoc,
                             void *event_arg,
                             sctp_disposition_t status,
                             sctp_cmd_seq_t *commands,

@@ -1151,7 +1151,7 @@ static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
         * disposition SCTP_DISPOSITION_CONSUME.
         */
        if (0 != (error = sctp_cmd_interpreter(event_type, subtype, state,
-                                              ep, asoc,
+                                              ep, *asoc,
                                               event_arg, status,
                                               commands, gfp)))
                goto bail;

@@ -1175,6 +1175,7 @@ static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
 
        case SCTP_DISPOSITION_DELETE_TCB:
                /* This should now be a command. */
+               *asoc = NULL;
                break;
 
        case SCTP_DISPOSITION_CONSUME:
--

2.5.0

Quoted 4 lines · Marcelo Ricardo Leitner

Hi Dmitry,

Can you please test this patch?
I'll re-post with proper subject if it works.

Still happening with the same stacks.

Quoted 8 lines · Dmitry Vyukov

Quoted 4 lines · Marcelo Ricardo Leitner

Hi Dmitry,

Can you please test this patch?
I'll re-post with proper subject if it works.

Still happening with the same stacks.

Then there may be another one, I'm afraid.

I'm using the testapp you shared in the first email, with that debug line enabled and added a new one:

+       pr_debug("%p %d\n", asoc, asoc ? asoc->state : 0);
        debug_post_sfx();

(should have used %x, but ok)

Also enabled slub_debug=PUZ, and I get:

without the patch:

[   87.873640] sctp: ffff8800b71533d8 1
[   87.873647] sctp: sctp_do_sm[post-sfx]: error:0,
asoc:ffff8800b71533d8[STATE_CLOSED]
[   87.873739] sctp: ffff8800b71533d8 1
[   87.873742] sctp: sctp_do_sm[post-sfx]: error:0,
asoc:ffff8800b71533d8[STATE_CLOSED]
[   87.875149] sctp: ffff8800b71533d8 1802201963
[   87.875238] sctp: sctp_do_sm[post-sfx]: error:0,

asoc:ffff8800b71533d8[STATE_CLOSED]

1802201963 = 0x6b6b6b6b, poison

with the patch:

[   81.071265] sctp: ffff880137571148 1
[   81.071273] sctp: sctp_do_sm[post-sfx]: error:0,
asoc:ffff880137571148[STATE_CLOSED]
[   81.071372] sctp: ffff880137571148 1
[   81.071375] sctp: sctp_do_sm[post-sfx]: error:0,
asoc:ffff880137571148[STATE_CLOSED]
[   81.072423] sctp:           (null) 0
[   81.072427] sctp: sctp_do_sm[post-sfx]: error:0, asoc:

(null)[STATE_CLOSED]

This one, at least, is gone with this patch.

Quoted 8 lines · Dmitry Vyukov

Quoted 4 lines · Marcelo Ricardo Leitner

Hi Dmitry,

Can you please test this patch?
I'll re-post with proper subject if it works.

Still happening with the same stacks.

Then there may be another one, I'm afraid.

I'm using the testapp you shared in the first email, with that debug line enabled and added a new one:

+       pr_debug("%p %d\n", asoc, asoc ? asoc->state : 0);
        debug_post_sfx();

(should have used %x, but ok)

Also enabled slub_debug=PUZ, and I get:

without the patch:

[   87.873640] sctp: ffff8800b71533d8 1
[   87.873647] sctp: sctp_do_sm[post-sfx]: error:0,
asoc:ffff8800b71533d8[STATE_CLOSED]
[   87.873739] sctp: ffff8800b71533d8 1
[   87.873742] sctp: sctp_do_sm[post-sfx]: error:0,
asoc:ffff8800b71533d8[STATE_CLOSED]
[   87.875149] sctp: ffff8800b71533d8 1802201963
[   87.875238] sctp: sctp_do_sm[post-sfx]: error:0,

asoc:ffff8800b71533d8[STATE_CLOSED]

1802201963 = 0x6b6b6b6b, poison

with the patch:

[   81.071265] sctp: ffff880137571148 1
[   81.071273] sctp: sctp_do_sm[post-sfx]: error:0,
asoc:ffff880137571148[STATE_CLOSED]
[   81.071372] sctp: ffff880137571148 1
[   81.071375] sctp: sctp_do_sm[post-sfx]: error:0,
asoc:ffff880137571148[STATE_CLOSED]
[   81.072423] sctp:           (null) 0
[   81.072427] sctp: sctp_do_sm[post-sfx]: error:0, asoc:

(null)[STATE_CLOSED]

This one, at least, is gone with this patch.

Marcelo

Quoted 42 lines · Marcelo Ricardo Leitner

Quoted 8 lines · Dmitry Vyukov

Quoted 4 lines · Marcelo Ricardo Leitner

Hi Dmitry,

Can you please test this patch?
I'll re-post with proper subject if it works.

Still happening with the same stacks.

Then there may be another one, I'm afraid.

I'm using the testapp you shared in the first email, with that debug line enabled and added a new one:

+       pr_debug("%p %d\n", asoc, asoc ? asoc->state : 0);
        debug_post_sfx();

(should have used %x, but ok)

Also enabled slub_debug=PUZ, and I get:

without the patch:

[   87.873640] sctp: ffff8800b71533d8 1
[   87.873647] sctp: sctp_do_sm[post-sfx]: error:0,
asoc:ffff8800b71533d8[STATE_CLOSED]
[   87.873739] sctp: ffff8800b71533d8 1
[   87.873742] sctp: sctp_do_sm[post-sfx]: error:0,
asoc:ffff8800b71533d8[STATE_CLOSED]
[   87.875149] sctp: ffff8800b71533d8 1802201963
[   87.875238] sctp: sctp_do_sm[post-sfx]: error:0,

asoc:ffff8800b71533d8[STATE_CLOSED]

1802201963 = 0x6b6b6b6b, poison

with the patch:

[   81.071265] sctp: ffff880137571148 1
[   81.071273] sctp: sctp_do_sm[post-sfx]: error:0,
asoc:ffff880137571148[STATE_CLOSED]
[   81.071372] sctp: ffff880137571148 1
[   81.071375] sctp: sctp_do_sm[post-sfx]: error:0,
asoc:ffff880137571148[STATE_CLOSED]
[   81.072423] sctp:           (null) 0
[   81.072427] sctp: sctp_do_sm[post-sfx]: error:0, asoc:

Quoted 6 lines

(null)[STATE_CLOSED]

This one, at least, is gone with this patch.

Marcelo

Hi Marcelo

I think you also need to catch the SCTP_DISPOSITION_ABORT and update the pointer. There are some issues there though as some functions report that code without actually destroying the association. This happens when the ABORT chunk may be dropped.

I think this might be why we still see the issue.

Quoted 57 lines · Vlad Yasevich

Quoted 42 lines · Marcelo Ricardo Leitner

Quoted 8 lines · Dmitry Vyukov

Quoted 4 lines · Marcelo Ricardo Leitner

Hi Dmitry,

Can you please test this patch?
I'll re-post with proper subject if it works.

Still happening with the same stacks.

Then there may be another one, I'm afraid.

I'm using the testapp you shared in the first email, with that debug line enabled and added a new one:

+       pr_debug("%p %d\n", asoc, asoc ? asoc->state : 0);
        debug_post_sfx();

(should have used %x, but ok)

Also enabled slub_debug=PUZ, and I get:

without the patch:

[   87.873640] sctp: ffff8800b71533d8 1
[   87.873647] sctp: sctp_do_sm[post-sfx]: error:0,
asoc:ffff8800b71533d8[STATE_CLOSED]
[   87.873739] sctp: ffff8800b71533d8 1
[   87.873742] sctp: sctp_do_sm[post-sfx]: error:0,
asoc:ffff8800b71533d8[STATE_CLOSED]
[   87.875149] sctp: ffff8800b71533d8 1802201963
[   87.875238] sctp: sctp_do_sm[post-sfx]: error:0,

asoc:ffff8800b71533d8[STATE_CLOSED]

1802201963 = 0x6b6b6b6b, poison

with the patch:

[   81.071265] sctp: ffff880137571148 1
[   81.071273] sctp: sctp_do_sm[post-sfx]: error:0,
asoc:ffff880137571148[STATE_CLOSED]
[   81.071372] sctp: ffff880137571148 1
[   81.071375] sctp: sctp_do_sm[post-sfx]: error:0,
asoc:ffff880137571148[STATE_CLOSED]
[   81.072423] sctp:           (null) 0
[   81.072427] sctp: sctp_do_sm[post-sfx]: error:0, asoc:

Quoted 6 lines

(null)[STATE_CLOSED]

This one, at least, is gone with this patch.

Marcelo

Hi Marcelo

I think you also need to catch the SCTP_DISPOSITION_ABORT and update the pointer. There are some issues there though as some functions report that code without actually destroying the association. This happens when the ABORT chunk may be dropped.

Quoted 1 line

I think this might be why we still see the issue.

Marcelo,

Is this info enough for you to cook another fix?

Quoted 63 lines · Dmitry Vyukov

Quoted 50 lines · Vlad Yasevich

Quoted 42 lines · Marcelo Ricardo Leitner

Quoted 8 lines · Dmitry Vyukov

Quoted 4 lines · Marcelo Ricardo Leitner

Hi Dmitry,

Can you please test this patch?
I'll re-post with proper subject if it works.

Still happening with the same stacks.

Then there may be another one, I'm afraid.

I'm using the testapp you shared in the first email, with that debug line enabled and added a new one:

+       pr_debug("%p %d\n", asoc, asoc ? asoc->state : 0);
        debug_post_sfx();

(should have used %x, but ok)

Also enabled slub_debug=PUZ, and I get:

without the patch:

[   87.873640] sctp: ffff8800b71533d8 1
[   87.873647] sctp: sctp_do_sm[post-sfx]: error:0,
asoc:ffff8800b71533d8[STATE_CLOSED]
[   87.873739] sctp: ffff8800b71533d8 1
[   87.873742] sctp: sctp_do_sm[post-sfx]: error:0,
asoc:ffff8800b71533d8[STATE_CLOSED]
[   87.875149] sctp: ffff8800b71533d8 1802201963
[   87.875238] sctp: sctp_do_sm[post-sfx]: error:0,

asoc:ffff8800b71533d8[STATE_CLOSED]

1802201963 = 0x6b6b6b6b, poison

with the patch:

[   81.071265] sctp: ffff880137571148 1
[   81.071273] sctp: sctp_do_sm[post-sfx]: error:0,
asoc:ffff880137571148[STATE_CLOSED]
[   81.071372] sctp: ffff880137571148 1
[   81.071375] sctp: sctp_do_sm[post-sfx]: error:0,
asoc:ffff880137571148[STATE_CLOSED]
[   81.072423] sctp:           (null) 0
[   81.072427] sctp: sctp_do_sm[post-sfx]: error:0, asoc:

Quoted 6 lines

(null)[STATE_CLOSED]

This one, at least, is gone with this patch.

Marcelo

Quoted 8 lines

Hi Marcelo

I think you also need to catch the SCTP_DISPOSITION_ABORT and update the pointer. There are some issues there though as some functions report that code without actually destroying the association. This happens when the ABORT chunk may be dropped.

I think this might be why we still see the issue.

Marcelo,

Quoted 1 line

Is this info enough for you to cook another fix?

Hi, I think so. I was really wondering how you could trigger that issue without the timestamp fix and Vlad's comment does shed some light on it.

I'll do more tests later today, but what did you have connecting to the listening socket? Somehow you made that accept() call to return..

Quoted 2 lines · Marcelo Ricardo Leitner

Quoted 1 line · Dmitry Vyukov

On Sat, Dec 5, 2015 at 5:39 PM, Vlad Yasevich <[email protected]> wrote:

Quoted 14 lines

Quoted 13 lines

Quoted 8 lines

Hi Marcelo

I think you also need to catch the SCTP_DISPOSITION_ABORT and update the pointer. There are some issues there though as some functions report that code without actually destroying the association. This happens when the ABORT chunk may be dropped.

I think this might be why we still see the issue.

Marcelo,

Is this info enough for you to cook another fix?

Quoted 5 lines

Hi, I think so. I was really wondering how you could trigger that issue without the timestamp fix and Vlad's comment does shed some light on it.

I'll do more tests later today, but what did you have connecting to the listening socket? Somehow you made that accept() call to return..

Local connect in another thread I guess.

Quoted 4 lines · Dmitry Vyukov

Quoted 2 lines · Marcelo Ricardo Leitner

Quoted 1 line · Dmitry Vyukov

On Sat, Dec 5, 2015 at 5:39 PM, Vlad Yasevich <[email protected]> wrote:

...

Quoted 20 lines

Quoted 19 lines

Quoted 7 lines

Quoted 7 lines

Hi Marcelo

I think you also need to catch the SCTP_DISPOSITION_ABORT and update the pointer. There are some issues there though as some functions report that code without actually destroying the association. This happens when the ABORT chunk may be dropped.

Quoted 6 lines

Quoted 1 line

I think this might be why we still see the issue.

Marcelo,

Is this info enough for you to cook another fix?

Hi, I think so. I was really wondering how you could trigger that issue without the timestamp fix and Vlad's comment does shed some light on it.

I'll do more tests later today, but what did you have connecting to the listening socket? Somehow you made that accept() call to return..

Quoted 1 line

Local connect in another thread I guess.

Vlad, I reviewed the places on which it returns SCTP_DISPOSITION_ABORT, and if I didn't miss something in there all of them either issue SCTP_CMD_ASSOC_FAILED or SCTP_CMD_INIT_FAILED before returning it, thus delaying DELETE_TCB and with that the asoc free.

There is one place,
though, that may not do it that way, it's sctp_sf_abort_violation(), but then that code only runs if asoc is already NULL by then.

Dmitry, still no luck here, cannot reproduce another hit. I'm using sctp_test and a custom test of mine, both on localhost so I would catch it in server or client side, nothing..

I need more info. Please enable the pr_debug() on debug_post_sfn() macro and see which status is being reported when you trigger the issue. And/or share a traffic capture so we can see what's going on with the association.

Marcelo

Quoted 26 lines · Marcelo Ricardo Leitner

Quoted 4 lines · Dmitry Vyukov

Quoted 2 lines · Marcelo Ricardo Leitner

Quoted 1 line · Dmitry Vyukov

On Sat, Dec 5, 2015 at 5:39 PM, Vlad Yasevich <[email protected]> wrote:

...

Quoted 20 lines

Quoted 19 lines

Quoted 7 lines

Quoted 7 lines

Hi Marcelo

I think you also need to catch the SCTP_DISPOSITION_ABORT and update the pointer. There are some issues there though as some functions report that code without actually destroying the association. This happens when the ABORT chunk may be dropped.

Quoted 6 lines

Quoted 1 line

I think this might be why we still see the issue.

Marcelo,

Is this info enough for you to cook another fix?

Hi, I think so. I was really wondering how you could trigger that issue without the timestamp fix and Vlad's comment does shed some light on it.

I'll do more tests later today, but what did you have connecting to the listening socket? Somehow you made that accept() call to return..

Quoted 6 lines

Quoted 1 line

Local connect in another thread I guess.

Vlad, I reviewed the places on which it returns SCTP_DISPOSITION_ABORT, and if I didn't miss something in there all of them either issue SCTP_CMD_ASSOC_FAILED or SCTP_CMD_INIT_FAILED before returning it, thus delaying DELETE_TCB and with that the asoc free.

They delay it from the perspective of the command interpreter since the command to delete the TCB happens a little later, but status code is checked after all commands are processed and command processing doesn't change it. So the 'status' code would still be SCTP_DISPOSITION_ABORT after DELETE_TCB command was processed. So, I think we may still have an use-after-free issue here.

Quoted 3 lines

There is one place,
though, that may not do it that way, it's sctp_sf_abort_violation(), but then that code only runs if asoc is already NULL by then.

I don't believe so. The violation state function can run with a non-NULL association if we are encountering protocol violations after the association is established.

Quoted 34 lines · Vlad Yasevich

Quoted 26 lines · Marcelo Ricardo Leitner

Quoted 4 lines · Dmitry Vyukov

Quoted 2 lines · Marcelo Ricardo Leitner

Quoted 1 line · Dmitry Vyukov

On Sat, Dec 5, 2015 at 5:39 PM, Vlad Yasevich <[email protected]> wrote:

...

Quoted 20 lines

Quoted 19 lines

Quoted 7 lines

Quoted 7 lines

Hi Marcelo

I think you also need to catch the SCTP_DISPOSITION_ABORT and update the pointer. There are some issues there though as some functions report that code without actually destroying the association. This happens when the ABORT chunk may be dropped.

Quoted 6 lines

Quoted 1 line

I think this might be why we still see the issue.

Marcelo,

Is this info enough for you to cook another fix?

Hi, I think so. I was really wondering how you could trigger that issue without the timestamp fix and Vlad's comment does shed some light on it.

I'll do more tests later today, but what did you have connecting to the listening socket? Somehow you made that accept() call to return..

Quoted 6 lines

Quoted 1 line

Local connect in another thread I guess.

Vlad, I reviewed the places on which it returns SCTP_DISPOSITION_ABORT, and if I didn't miss something in there all of them either issue SCTP_CMD_ASSOC_FAILED or SCTP_CMD_INIT_FAILED before returning it, thus delaying DELETE_TCB and with that the asoc free.

Quoted 5 lines

They delay it from the perspective of the command interpreter since the command to delete the TCB happens a little later, but status code is checked after all commands are processed and command processing doesn't change it. So the 'status' code would still be SCTP_DISPOSITION_ABORT after DELETE_TCB command was processed. So, I think we may still have an use-after-free issue here.

Gotcha! That's pretty much it then. From that point of view now, there shouldn't be a case that it returns _ABORT without freeing the asoc in the same loop. (more below)

Quoted 4 lines

Quoted 3 lines

There is one place,
though, that may not do it that way, it's sctp_sf_abort_violation(), but then that code only runs if asoc is already NULL by then.

Quoted 2 lines

I don't believe so. The violation state function can run with a non-NULL association if we are encountering protocol violations after the association is established.

Yup, that's correct. I just tried to reference one case on which it would return _ABORT without issuing any of those _FAILEDs before doing so (meaning the association could still be valid) but that in that case, the asoc was already NULL.

Dmitry, please give this one a run, as I still cannot reproduce your use case..

---8<---

commit b63ad8dc45257dd6c536ac0227fcc623efd9328b

Author: Marcelo Ricardo Leitner <[email protected]>
Date:   Fri Dec 4 15:30:23 2015 -0200

sctp: fix use-after-free in pr_debug statement

Dmitry Vyukov reported a use-after-free in the code expanded by the macro debug_post_sfx, which is caused by the use of the asoc pointer after it was freed within sctp_side_effect() scope.

This patch fixes it by allowing sctp_side_effect to clear that asoc pointer when the TCB is freed.

As Vlad explained, we also have to cover the SCTP_DISPOSITION_ABORT case because it will trigger DELETE_TCB too on that same loop.

The macro is already prepared to handle such NULL pointer.

Reported-by Dmitry Vyukov <[email protected]>

net/sctp/sm_sideeffect.c +6 -5

index 6098d4c42fa9..be23d5c2074f 100644
@@ -63,7 +63,7 @@ static int sctp_cmd_interpreter(sctp_event_t event_type,
 static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
 			     sctp_state_t state,
 			     struct sctp_endpoint *ep,
-			     struct sctp_association *asoc,
+			     struct sctp_association **asoc,
 			     void *event_arg,
 			     sctp_disposition_t status,
 			     sctp_cmd_seq_t *commands,

@@ -1123,7 +1123,7 @@ int sctp_do_sm(struct net *net, sctp_event_t event_type, sctp_subtype_t subtype,
 	debug_post_sfn();
 
 	error = sctp_side_effects(event_type, subtype, state,
-				  ep, asoc, event_arg, status,
+				  ep, &asoc, event_arg, status,
 				  &commands, gfp);
 	debug_post_sfx();
 

@@ -1136,7 +1136,7 @@ int sctp_do_sm(struct net *net, sctp_event_t event_type, sctp_subtype_t subtype,
 static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
 			     sctp_state_t state,
 			     struct sctp_endpoint *ep,
-			     struct sctp_association *asoc,
+			     struct sctp_association **asoc,
 			     void *event_arg,
 			     sctp_disposition_t status,
 			     sctp_cmd_seq_t *commands,

@@ -1151,7 +1151,7 @@ static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
 	 * disposition SCTP_DISPOSITION_CONSUME.
 	 */
 	if (0 != (error = sctp_cmd_interpreter(event_type, subtype, state,
-					       ep, asoc,
+					       ep, *asoc,
 					       event_arg, status,
 					       commands, gfp)))
 		goto bail;

@@ -1174,11 +1174,12 @@ static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
 		break;
 
 	case SCTP_DISPOSITION_DELETE_TCB:
+	case SCTP_DISPOSITION_ABORT:
 		/* This should now be a command. */
+		*asoc = NULL;
 		break;
 
 	case SCTP_DISPOSITION_CONSUME:
-	case SCTP_DISPOSITION_ABORT:
 		/*
 		 * We should no longer have much work to do here as the
 		 * real work has been done as explicit commands above.

Quoted 21 lines · Vlad Yasevich

Quoted 11 lines · Marcelo Ricardo Leitner

Quoted 4 lines

Vlad, I reviewed the places on which it returns SCTP_DISPOSITION_ABORT, and if I didn't miss something in there all of them either issue SCTP_CMD_ASSOC_FAILED or SCTP_CMD_INIT_FAILED before returning it, thus delaying DELETE_TCB and with that the asoc free.

They delay it from the perspective of the command interpreter since the command to delete the TCB happens a little later, but status code is checked after all commands are processed and command processing doesn't change it. So the 'status' code would still be SCTP_DISPOSITION_ABORT after DELETE_TCB command was processed. So, I think we may still have an use-after-free issue here.

Gotcha! That's pretty much it then. From that point of view now, there shouldn't be a case that it returns _ABORT without freeing the asoc in the same loop. (more below)

Quoted 4 lines

Quoted 3 lines

There is one place,
though, that may not do it that way, it's sctp_sf_abort_violation(), but then that code only runs if asoc is already NULL by then.

Quoted 7 lines

Quoted 2 lines

I don't believe so. The violation state function can run with a non-NULL association if we are encountering protocol violations after the association is established.

Yup, that's correct. I just tried to reference one case on which it would return _ABORT without issuing any of those _FAILEDs before doing so (meaning the association could still be valid) but that in that case, the asoc was already NULL.

I think it is possible to hit the 'discard:' tag in that function while still having a valid association. That happens when ABORT chunk is required to be authenticated. This that case, instead of generating an ABORT and terminating the current association, we just drop the packet, but still report an _ABORT disposition code.

This probably need to change if we are going to catch the _ABORT disposition and clear the asoc pointer.

Em 07-12-2015 18:37, Vlad Yasevich escreveu:

Quoted 55 lines · Marcelo Ricardo Leitner

Quoted 38 lines · Vlad Yasevich

Quoted 9 lines · Marcelo Ricardo Leitner

Quoted 5 lines

Vlad, I reviewed the places on which it returns SCTP_DISPOSITION_ABORT, and if I didn't miss something in there all of them either issue SCTP_CMD_ASSOC_FAILED or SCTP_CMD_INIT_FAILED before returning it, thus delaying DELETE_TCB and with that the asoc free.

Quoted 9 lines

They delay it from the perspective of the command interpreter since the command
to delete the TCB happens a little later, but status code is checked after all
commands are processed and command processing doesn't change it. So the 'status'
code would still be SCTP_DISPOSITION_ABORT after DELETE_TCB command was processed.
So, I think we may still have an use-after-free issue here.

Gotcha! That's pretty much it then. From that point of view now, there shouldn't be a case that it returns _ABORT without freeing the asoc in the same loop. (more below)

Quoted 10 lines

Quoted 4 lines

There is one place,
though, that may not do it that way, it's sctp_sf_abort_violation(), but
then that code only runs if asoc is already NULL by then.

I don't believe so. The violation state function can run with a non-NULL association
if we are encountering protocol violations after the association is established.

Quoted 4 lines

Yup, that's correct. I just tried to reference one case on which it would return _ABORT without issuing any of those _FAILEDs before doing so (meaning the association could still be valid) but that in that case, the asoc was already NULL.

I think it is possible to hit the 'discard:' tag in that function while still
having a valid association. That happens when ABORT chunk is required to be
authenticated. This that case, instead of generating an ABORT and terminating the
current association, we just drop the packet, but still report an _ABORT disposition code.

Quoted 3 lines

This probably need to change if we are going to catch the _ABORT disposition and
clear the asoc pointer.

Oups. Nice one. I'll switch it to SCTP_DISPOSITION_DISCARD if it hits that if() then. Thanks Vlad.

Quoted 61 lines · Marcelo Ricardo Leitner

Em 07-12-2015 18:37, Vlad Yasevich escreveu:

Quoted 58 lines · Marcelo Ricardo Leitner

Quoted 38 lines · Vlad Yasevich

Quoted 9 lines · Marcelo Ricardo Leitner

Quoted 5 lines

Vlad, I reviewed the places on which it returns SCTP_DISPOSITION_ABORT, and if I didn't miss something in there all of them either issue SCTP_CMD_ASSOC_FAILED or SCTP_CMD_INIT_FAILED before returning it, thus delaying DELETE_TCB and with that the asoc free.

Quoted 9 lines

They delay it from the perspective of the command interpreter since the command
to delete the TCB happens a little later, but status code is checked after all
commands are processed and command processing doesn't change it. So the 'status'
code would still be SCTP_DISPOSITION_ABORT after DELETE_TCB command was processed.
So, I think we may still have an use-after-free issue here.

Gotcha! That's pretty much it then. From that point of view now, there shouldn't be a case that it returns _ABORT without freeing the asoc in the same loop. (more below)

Quoted 10 lines

Quoted 4 lines

There is one place,
though, that may not do it that way, it's sctp_sf_abort_violation(), but
then that code only runs if asoc is already NULL by then.

I don't believe so. The violation state function can run with a non-NULL association
if we are encountering protocol violations after the association is established.

Quoted 4 lines

Yup, that's correct. I just tried to reference one case on which it would return _ABORT without issuing any of those _FAILEDs before doing so (meaning the association could still be valid) but that in that case, the asoc was already NULL.

I think it is possible to hit the 'discard:' tag in that function while still
having a valid association. That happens when ABORT chunk is required to be
authenticated. This that case, instead of generating an ABORT and terminating the
current association, we just drop the packet, but still report an _ABORT disposition code.

This probably need to change if we are going to catch the _ABORT disposition and
clear the asoc pointer.

Quoted 2 lines

Oups. Nice one. I'll switch it to SCTP_DISPOSITION_DISCARD if it hits that if() then. Thanks Vlad.

So I am waiting for a new patch, right?
Can you please combine all changes into a single patch (as far as I understand the previous one must be applied on top of the first one)?

Quoted 63 lines · Dmitry Vyukov

Quoted 61 lines · Marcelo Ricardo Leitner

Em 07-12-2015 18:37, Vlad Yasevich escreveu:

Quoted 58 lines · Marcelo Ricardo Leitner

Quoted 38 lines · Vlad Yasevich

Quoted 9 lines · Marcelo Ricardo Leitner

Quoted 5 lines

Vlad, I reviewed the places on which it returns SCTP_DISPOSITION_ABORT, and if I didn't miss something in there all of them either issue SCTP_CMD_ASSOC_FAILED or SCTP_CMD_INIT_FAILED before returning it, thus delaying DELETE_TCB and with that the asoc free.

Quoted 9 lines

They delay it from the perspective of the command interpreter since the command
to delete the TCB happens a little later, but status code is checked after all
commands are processed and command processing doesn't change it. So the 'status'
code would still be SCTP_DISPOSITION_ABORT after DELETE_TCB command was processed.
So, I think we may still have an use-after-free issue here.

Gotcha! That's pretty much it then. From that point of view now, there shouldn't be a case that it returns _ABORT without freeing the asoc in the same loop. (more below)

Quoted 10 lines

Quoted 4 lines

There is one place,
though, that may not do it that way, it's sctp_sf_abort_violation(), but
then that code only runs if asoc is already NULL by then.

I don't believe so. The violation state function can run with a non-NULL association
if we are encountering protocol violations after the association is established.

Quoted 4 lines

Yup, that's correct. I just tried to reference one case on which it would return _ABORT without issuing any of those _FAILEDs before doing so (meaning the association could still be valid) but that in that case, the asoc was already NULL.

I think it is possible to hit the 'discard:' tag in that function while still
having a valid association. That happens when ABORT chunk is required to be
authenticated. This that case, instead of generating an ABORT and terminating the
current association, we just drop the packet, but still report an _ABORT disposition code.

This probably need to change if we are going to catch the _ABORT disposition and
clear the asoc pointer.

Quoted 7 lines

Quoted 2 lines

Oups. Nice one. I'll switch it to SCTP_DISPOSITION_DISCARD if it hits that if() then. Thanks Vlad.

So I am waiting for a new patch, right?
Can you please combine all changes into a single patch (as far as I understand the previous one must be applied on top of the first one)?

The patches were combined already, but this last pick by Vlad is just not yet patched. It's not necessary for your testing and I didn't want to interrupt it in case you were already testing it.

You can use my last patch here, from 2 emails ago, the one which contains this line:
- case SCTP_DISPOSITION_ABORT:

Quoted 76 lines · Marcelo Ricardo Leitner

Quoted 63 lines · Dmitry Vyukov

Quoted 61 lines · Marcelo Ricardo Leitner

Em 07-12-2015 18:37, Vlad Yasevich escreveu:

Quoted 58 lines · Marcelo Ricardo Leitner

Quoted 38 lines · Vlad Yasevich

Quoted 9 lines · Marcelo Ricardo Leitner

Quoted 5 lines

Vlad, I reviewed the places on which it returns SCTP_DISPOSITION_ABORT, and if I didn't miss something in there all of them either issue SCTP_CMD_ASSOC_FAILED or SCTP_CMD_INIT_FAILED before returning it, thus delaying DELETE_TCB and with that the asoc free.

Quoted 9 lines

They delay it from the perspective of the command interpreter since the command
to delete the TCB happens a little later, but status code is checked after all
commands are processed and command processing doesn't change it. So the 'status'
code would still be SCTP_DISPOSITION_ABORT after DELETE_TCB command was processed.
So, I think we may still have an use-after-free issue here.

Gotcha! That's pretty much it then. From that point of view now, there shouldn't be a case that it returns _ABORT without freeing the asoc in the same loop. (more below)

Quoted 10 lines

Quoted 4 lines

There is one place,
though, that may not do it that way, it's sctp_sf_abort_violation(), but
then that code only runs if asoc is already NULL by then.

I don't believe so. The violation state function can run with a non-NULL association
if we are encountering protocol violations after the association is established.

Quoted 4 lines

Yup, that's correct. I just tried to reference one case on which it would return _ABORT without issuing any of those _FAILEDs before doing so (meaning the association could still be valid) but that in that case, the asoc was already NULL.

I think it is possible to hit the 'discard:' tag in that function while still
having a valid association. That happens when ABORT chunk is required to be
authenticated. This that case, instead of generating an ABORT and terminating the
current association, we just drop the packet, but still report an _ABORT disposition code.

This probably need to change if we are going to catch the _ABORT disposition and
clear the asoc pointer.

Quoted 7 lines

Quoted 2 lines

Oups. Nice one. I'll switch it to SCTP_DISPOSITION_DISCARD if it hits that if() then. Thanks Vlad.

So I am waiting for a new patch, right?
Can you please combine all changes into a single patch (as far as I understand the previous one must be applied on top of the first one)?

The patches were combined already, but this last pick by Vlad is just not yet patched. It's not necessary for your testing and I didn't want to interrupt it in case you were already testing it.

Quoted 3 lines

You can use my last patch here, from 2 emails ago, the one which contains this line:
- case SCTP_DISPOSITION_ABORT:

You are right. I missed that they are combined. Testing with it now.

Quoted 2 lines · Dmitry Vyukov

On Tue, Dec 8, 2015 at 6:40 PM, Marcelo Ricardo Leitner <[email protected]> wrote:

Quoted 9 lines

Quoted 7 lines

The patches were combined already, but this last pick by Vlad is just not yet patched. It's not necessary for your testing and I didn't want to interrupt it in case you were already testing it.

You can use my last patch here, from 2 emails ago, the one which contains this line:
- case SCTP_DISPOSITION_ABORT:

Quoted 1 line

You are right. I missed that they are combined. Testing with it now.

Use-after-free still happens.
I am on commit aa53685549a2cfb5f175b0c4a20bc9aa1e5a1b85 (Dec 8) plus the following sctp-related changes:

Quoted 3 lines · Dmitry Vyukov

Quoted 2 lines · Dmitry Vyukov

On Tue, Dec 8, 2015 at 6:40 PM, Marcelo Ricardo Leitner <[email protected]> wrote:

...

Quoted 14 lines

Quoted 10 lines

Quoted 7 lines

The patches were combined already, but this last pick by Vlad is just not yet patched. It's not necessary for your testing and I didn't want to interrupt it in case you were already testing it.

You can use my last patch here, from 2 emails ago, the one which contains this line:
- case SCTP_DISPOSITION_ABORT:

You are right. I missed that they are combined. Testing with it now.

Quoted 3 lines

Use-after-free still happens.
I am on commit aa53685549a2cfb5f175b0c4a20bc9aa1e5a1b85 (Dec 8) plus the following sctp-related changes:

Changes are fine. Ugh. Ok, I'll try your new reproducer here.

Quoted 19 lines · Marcelo Ricardo Leitner

Quoted 3 lines · Dmitry Vyukov

Quoted 2 lines · Dmitry Vyukov

On Tue, Dec 8, 2015 at 6:40 PM, Marcelo Ricardo Leitner <[email protected]> wrote:

...

Quoted 14 lines

Quoted 10 lines

Quoted 7 lines

The patches were combined already, but this last pick by Vlad is just not yet patched. It's not necessary for your testing and I didn't want to interrupt it in case you were already testing it.

You can use my last patch here, from 2 emails ago, the one which contains this line:
- case SCTP_DISPOSITION_ABORT:

You are right. I missed that they are combined. Testing with it now.

Quoted 5 lines

Quoted 3 lines

Use-after-free still happens.
I am on commit aa53685549a2cfb5f175b0c4a20bc9aa1e5a1b85 (Dec 8) plus the following sctp-related changes:

Changes are fine. Ugh. Ok, I'll try your new reproducer here.

Heh I wasn't going to reproduce this by myself anytime soon, I think. It's using the same socket to connect to itself, and only happens if the connect() gets there before the listen() call. Figured this out because I could only reproduce it under strace at first.

Please give this other patch a try. A state command (sctp_sf_cookie_wait_prm_abort) was issuing SCTP_CMD_INIT_FAILED, which leads to SCTP_CMD_DELETE_TCB, but returning SCTP_DISPOSITION_CONSUME, which fooled the patch.

---8<---
commit 9f84d50e36cee0ce66e4ce9b3b1665e0a1dbcdd3

Author: Marcelo Ricardo Leitner <[email protected]>
Date:   Fri Dec 4 15:30:23 2015 -0200

sctp: fix use-after-free in pr_debug statement

Dmitry Vyukov reported a use-after-free in the code expanded by the macro debug_post_sfx, which is caused by the use of the asoc pointer after it was freed within sctp_side_effect() scope.

This patch fixes it by allowing sctp_side_effect to clear that asoc pointer when the TCB is freed.

As Vlad explained, we also have to cover the SCTP_DISPOSITION_ABORT case because it will trigger DELETE_TCB too on that same loop.

Also, there was a place issuing SCTP_CMD_INIT_FAILED but returning SCTP_DISPOSITION_CONSUME, which would fool the scheme above. Fix it by returning SCTP_DISPOSITION_ABORT instead.

The macro is already prepared to handle such NULL pointer.

Reported-by Dmitry Vyukov <[email protected]>

net/sctp/sm_sideeffect.c +6 -5

index 6098d4c42fa9..be23d5c2074f 100644
@@ -63,7 +63,7 @@ static int sctp_cmd_interpreter(sctp_event_t event_type,
 static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
                             sctp_state_t state,
                             struct sctp_endpoint *ep,
-                            struct sctp_association *asoc,
+                            struct sctp_association **asoc,
                             void *event_arg,
                             sctp_disposition_t status,
                             sctp_cmd_seq_t *commands,

@@ -1123,7 +1123,7 @@ int sctp_do_sm(struct net *net, sctp_event_t event_type, sctp_subtype_t subtype,
        debug_post_sfn();
 
        error = sctp_side_effects(event_type, subtype, state,
-                                 ep, asoc, event_arg, status,
+                                 ep, &asoc, event_arg, status,
                                  &commands, gfp);
        debug_post_sfx();
 

@@ -1136,7 +1136,7 @@ int sctp_do_sm(struct net *net, sctp_event_t event_type, sctp_subtype_t subtype,
 static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
                             sctp_state_t state,
                             struct sctp_endpoint *ep,
-                            struct sctp_association *asoc,
+                            struct sctp_association **asoc,
                             void *event_arg,
                             sctp_disposition_t status,
                             sctp_cmd_seq_t *commands,

@@ -1151,7 +1151,7 @@ static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
         * disposition SCTP_DISPOSITION_CONSUME.
         */
        if (0 != (error = sctp_cmd_interpreter(event_type, subtype, state,
-                                              ep, asoc,
+                                              ep, *asoc,
                                               event_arg, status,
                                               commands, gfp)))
                goto bail;

@@ -1174,11 +1174,12 @@ static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
                break;
 
        case SCTP_DISPOSITION_DELETE_TCB:
+       case SCTP_DISPOSITION_ABORT:
                /* This should now be a command. */
+               *asoc = NULL;
                break;
 
        case SCTP_DISPOSITION_CONSUME:
-       case SCTP_DISPOSITION_ABORT:
                /*
                 * We should no longer have much work to do here as the
                 * real work has been done as explicit commands above.

net/sctp/sm_statefuns.c +0 -2

index 6f46aa16cb76..d801e151498a 100644
@@ -4959,12 +4959,10 @@ sctp_disposition_t sctp_sf_cookie_wait_prm_abort(
        sctp_cmd_seq_t *commands)
 {
        struct sctp_chunk *abort = arg;
-       sctp_disposition_t retval;
 
        /* Stop T1-init timer */
        sctp_add_cmd_sf(commands, SCTP_CMD_TIMER_STOP,
                        SCTP_TO(SCTP_EVENT_TIMEOUT_T1_INIT));
-       retval = SCTP_DISPOSITION_CONSUME;
 
        sctp_add_cmd_sf(commands, SCTP_CMD_REPLY, SCTP_CHUNK(abort));
 

Quoted 140 lines · Marcelo Ricardo Leitner

Quoted 19 lines · Marcelo Ricardo Leitner

Quoted 3 lines · Dmitry Vyukov

Quoted 2 lines · Dmitry Vyukov

On Tue, Dec 8, 2015 at 6:40 PM, Marcelo Ricardo Leitner <[email protected]> wrote:

...

Quoted 14 lines

Quoted 10 lines

Quoted 7 lines

The patches were combined already, but this last pick by Vlad is just not yet patched. It's not necessary for your testing and I didn't want to interrupt it in case you were already testing it.

You can use my last patch here, from 2 emails ago, the one which contains this line:
- case SCTP_DISPOSITION_ABORT:

You are right. I missed that they are combined. Testing with it now.

Quoted 5 lines

Quoted 3 lines

Use-after-free still happens.
I am on commit aa53685549a2cfb5f175b0c4a20bc9aa1e5a1b85 (Dec 8) plus the following sctp-related changes:

Changes are fine. Ugh. Ok, I'll try your new reproducer here.

Heh I wasn't going to reproduce this by myself anytime soon, I think. It's using the same socket to connect to itself, and only happens if the connect() gets there before the listen() call. Figured this out because I could only reproduce it under strace at first.

Please give this other patch a try. A state command (sctp_sf_cookie_wait_prm_abort) was issuing SCTP_CMD_INIT_FAILED, which leads to SCTP_CMD_DELETE_TCB, but returning SCTP_DISPOSITION_CONSUME, which fooled the patch.

---8<---
commit 9f84d50e36cee0ce66e4ce9b3b1665e0a1dbcdd3

Author: Marcelo Ricardo Leitner <[email protected]>
Date:   Fri Dec 4 15:30:23 2015 -0200

sctp: fix use-after-free in pr_debug statement

Dmitry Vyukov reported a use-after-free in the code expanded by the macro debug_post_sfx, which is caused by the use of the asoc pointer after it was freed within sctp_side_effect() scope.

This patch fixes it by allowing sctp_side_effect to clear that asoc pointer when the TCB is freed.

As Vlad explained, we also have to cover the SCTP_DISPOSITION_ABORT case because it will trigger DELETE_TCB too on that same loop.

Also, there was a place issuing SCTP_CMD_INIT_FAILED but returning SCTP_DISPOSITION_CONSUME, which would fool the scheme above. Fix it by returning SCTP_DISPOSITION_ABORT instead.

The macro is already prepared to handle such NULL pointer.

Reported-by Dmitry Vyukov <[email protected]>

net/sctp/sm_sideeffect.c +6 -5

index 6098d4c42fa9..be23d5c2074f 100644
@@ -63,7 +63,7 @@ static int sctp_cmd_interpreter(sctp_event_t event_type,
  static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
                              sctp_state_t state,
                              struct sctp_endpoint *ep,
-                            struct sctp_association *asoc,
+                            struct sctp_association **asoc,
                              void *event_arg,
                              sctp_disposition_t status,
                              sctp_cmd_seq_t *commands,

@@ -1123,7 +1123,7 @@ int sctp_do_sm(struct net *net, sctp_event_t event_type, sctp_subtype_t subtype,
         debug_post_sfn();
 
         error = sctp_side_effects(event_type, subtype, state,
-                                 ep, asoc, event_arg, status,
+                                 ep, &asoc, event_arg, status,
                                   &commands, gfp);
         debug_post_sfx();
 

@@ -1136,7 +1136,7 @@ int sctp_do_sm(struct net *net, sctp_event_t event_type, sctp_subtype_t subtype,
  static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
                              sctp_state_t state,
                              struct sctp_endpoint *ep,
-                            struct sctp_association *asoc,
+                            struct sctp_association **asoc,
                              void *event_arg,
                              sctp_disposition_t status,
                              sctp_cmd_seq_t *commands,

@@ -1151,7 +1151,7 @@ static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
          * disposition SCTP_DISPOSITION_CONSUME.
          */
         if (0 != (error = sctp_cmd_interpreter(event_type, subtype, state,
-                                              ep, asoc,
+                                              ep, *asoc,
                                                event_arg, status,
                                                commands, gfp)))
                 goto bail;

@@ -1174,11 +1174,12 @@ static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
                 break;
 
         case SCTP_DISPOSITION_DELETE_TCB:
+       case SCTP_DISPOSITION_ABORT:
                 /* This should now be a command. */
+               *asoc = NULL;
                 break;
 
         case SCTP_DISPOSITION_CONSUME:
-       case SCTP_DISPOSITION_ABORT:
                 /*
                  * We should no longer have much work to do here as the
                  * real work has been done as explicit commands above.

net/sctp/sm_statefuns.c +1 -3

index 6f46aa16cb76..d801e151498a 100644
@@ -4959,12 +4959,10 @@ sctp_disposition_t sctp_sf_cookie_wait_prm_abort(
         sctp_cmd_seq_t *commands)
  {
         struct sctp_chunk *abort = arg;
-       sctp_disposition_t retval;
 
         /* Stop T1-init timer */
         sctp_add_cmd_sf(commands, SCTP_CMD_TIMER_STOP,
                         SCTP_TO(SCTP_EVENT_TIMEOUT_T1_INIT));
-       retval = SCTP_DISPOSITION_CONSUME;
 
         sctp_add_cmd_sf(commands, SCTP_CMD_REPLY, SCTP_CHUNK(abort));
 

@@ -4983,7 +4981,7 @@ sctp_disposition_t sctp_sf_cookie_wait_prm_abort(
         sctp_add_cmd_sf(commands, SCTP_CMD_INIT_FAILED,
                         SCTP_PERR(SCTP_ERROR_USER_ABORT));
 
-       return retval;
+       return SCTP_DISPOSITION_ABORT;
  }
 

net/sctp/sm_statefuns.c sctp_disposition_t sctp_sf_cookie_wait_prm_abort( · line 4987

Quoted 1 line

net/sctp/sm_statefuns.c +0 -0

  /*

Still happens...
I am on commit aa53685549a2cfb5f175b0c4a20bc9aa1e5a1b85 with your latest patch applied.
Can you figure out what happens now from the report below? If not I can create a repro, it's just somewhat time consuming.

Em 11-12-2015 11:35, Dmitry Vyukov escreveu:

Quoted 142 lines

Quoted 140 lines · Marcelo Ricardo Leitner

Quoted 19 lines · Marcelo Ricardo Leitner

Quoted 3 lines · Dmitry Vyukov

Quoted 2 lines · Dmitry Vyukov

On Tue, Dec 8, 2015 at 6:40 PM, Marcelo Ricardo Leitner <[email protected]> wrote:

...

Quoted 14 lines

Quoted 10 lines

Quoted 7 lines

The patches were combined already, but this last pick by Vlad is just not yet patched. It's not necessary for your testing and I didn't want to interrupt it in case you were already testing it.

You can use my last patch here, from 2 emails ago, the one which contains this line:
- case SCTP_DISPOSITION_ABORT:

You are right. I missed that they are combined. Testing with it now.

Quoted 5 lines

Quoted 3 lines

Use-after-free still happens.
I am on commit aa53685549a2cfb5f175b0c4a20bc9aa1e5a1b85 (Dec 8) plus the following sctp-related changes:

Changes are fine. Ugh. Ok, I'll try your new reproducer here.

Heh I wasn't going to reproduce this by myself anytime soon, I think. It's using the same socket to connect to itself, and only happens if the connect() gets there before the listen() call. Figured this out because I could only reproduce it under strace at first.

Please give this other patch a try. A state command (sctp_sf_cookie_wait_prm_abort) was issuing SCTP_CMD_INIT_FAILED, which leads to SCTP_CMD_DELETE_TCB, but returning SCTP_DISPOSITION_CONSUME, which fooled the patch.

---8<---
commit 9f84d50e36cee0ce66e4ce9b3b1665e0a1dbcdd3

Author: Marcelo Ricardo Leitner <[email protected]>
Date:   Fri Dec 4 15:30:23 2015 -0200

sctp: fix use-after-free in pr_debug statement

Dmitry Vyukov reported a use-after-free in the code expanded by the macro debug_post_sfx, which is caused by the use of the asoc pointer after it was freed within sctp_side_effect() scope.

This patch fixes it by allowing sctp_side_effect to clear that asoc pointer when the TCB is freed.

As Vlad explained, we also have to cover the SCTP_DISPOSITION_ABORT case because it will trigger DELETE_TCB too on that same loop.

Also, there was a place issuing SCTP_CMD_INIT_FAILED but returning SCTP_DISPOSITION_CONSUME, which would fool the scheme above. Fix it by returning SCTP_DISPOSITION_ABORT instead.

The macro is already prepared to handle such NULL pointer.

Reported-by Dmitry Vyukov <[email protected]>

net/sctp/sm_sideeffect.c +6 -5

index 6098d4c42fa9..be23d5c2074f 100644
@@ -63,7 +63,7 @@ static int sctp_cmd_interpreter(sctp_event_t event_type,
 static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
                             sctp_state_t state,
                             struct sctp_endpoint *ep,
-                            struct sctp_association *asoc,
+                            struct sctp_association **asoc,
                             void *event_arg,
                             sctp_disposition_t status,
                             sctp_cmd_seq_t *commands,

@@ -1123,7 +1123,7 @@ int sctp_do_sm(struct net *net, sctp_event_t event_type, sctp_subtype_t subtype,
        debug_post_sfn();
 
        error = sctp_side_effects(event_type, subtype, state,
-                                 ep, asoc, event_arg, status,
+                                 ep, &asoc, event_arg, status,
                                  &commands, gfp);
        debug_post_sfx();
 

@@ -1136,7 +1136,7 @@ int sctp_do_sm(struct net *net, sctp_event_t event_type, sctp_subtype_t subtype,
 static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
                             sctp_state_t state,
                             struct sctp_endpoint *ep,
-                            struct sctp_association *asoc,
+                            struct sctp_association **asoc,
                             void *event_arg,
                             sctp_disposition_t status,
                             sctp_cmd_seq_t *commands,

@@ -1151,7 +1151,7 @@ static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
         * disposition SCTP_DISPOSITION_CONSUME.
         */
        if (0 != (error = sctp_cmd_interpreter(event_type, subtype, state,
-                                              ep, asoc,
+                                              ep, *asoc,
                                               event_arg, status,
                                               commands, gfp)))
                goto bail;

@@ -1174,11 +1174,12 @@ static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
                break;
 
        case SCTP_DISPOSITION_DELETE_TCB:
+       case SCTP_DISPOSITION_ABORT:
                /* This should now be a command. */
+               *asoc = NULL;
                break;
 
        case SCTP_DISPOSITION_CONSUME:
-       case SCTP_DISPOSITION_ABORT:
                /*
                 * We should no longer have much work to do here as the
                 * real work has been done as explicit commands above.

net/sctp/sm_statefuns.c +1 -3

index 6f46aa16cb76..d801e151498a 100644
@@ -4959,12 +4959,10 @@ sctp_disposition_t sctp_sf_cookie_wait_prm_abort(
        sctp_cmd_seq_t *commands)
 {
        struct sctp_chunk *abort = arg;
-       sctp_disposition_t retval;
 
        /* Stop T1-init timer */
        sctp_add_cmd_sf(commands, SCTP_CMD_TIMER_STOP,
                        SCTP_TO(SCTP_EVENT_TIMEOUT_T1_INIT));
-       retval = SCTP_DISPOSITION_CONSUME;
 
        sctp_add_cmd_sf(commands, SCTP_CMD_REPLY, SCTP_CHUNK(abort));
 

@@ -4983,7 +4981,7 @@ sctp_disposition_t sctp_sf_cookie_wait_prm_abort(
        sctp_add_cmd_sf(commands, SCTP_CMD_INIT_FAILED,
                        SCTP_PERR(SCTP_ERROR_USER_ABORT));
 
-       return retval;
+       return SCTP_DISPOSITION_ABORT;
 }
 

Quoted 8 lines

net/sctp/sm_statefuns.c sctp_disposition_t sctp_sf_cookie_wait_prm_abort( · line 4987

Quoted 1 line

net/sctp/sm_statefuns.c +0 -0

 /*

Still happens...
I am on commit aa53685549a2cfb5f175b0c4a20bc9aa1e5a1b85 with your latest patch applied.
Can you figure out what happens now from the report below? If not I can create a repro, it's just somewhat time consuming.

I can imagine. I don't know how this fuzzer works, but it would be nice if this reproducer extractor could be executed easier. So far, we have identified 3 different issues already leading to this bug:
- 1st, the handling on DELETE_TCB
- 2nd, the handling on DISPOSITION_ABORT
- 3rd, the bad combination on internal state-machine command to a return value

I can and will review it again, but it's doing nasty stuff like using the same socket to connect to itself. It's hard to imagine all those combinations in mind that might lead to that use-after-free.

Keep you posted.. thanks.

Quoted 152 lines · Marcelo Ricardo Leitner

Em 11-12-2015 11:35, Dmitry Vyukov escreveu:

Quoted 150 lines

Quoted 133 lines · Marcelo Ricardo Leitner

Quoted 19 lines · Marcelo Ricardo Leitner

Quoted 3 lines · Dmitry Vyukov

Quoted 2 lines · Dmitry Vyukov

On Tue, Dec 8, 2015 at 6:40 PM, Marcelo Ricardo Leitner <[email protected]> wrote:

...

Quoted 14 lines

Quoted 10 lines

Quoted 7 lines

The patches were combined already, but this last pick by Vlad is just not yet patched. It's not necessary for your testing and I didn't want to interrupt it in case you were already testing it.

You can use my last patch here, from 2 emails ago, the one which contains this line:
- case SCTP_DISPOSITION_ABORT:

You are right. I missed that they are combined. Testing with it now.

Quoted 5 lines

Quoted 3 lines

Use-after-free still happens.
I am on commit aa53685549a2cfb5f175b0c4a20bc9aa1e5a1b85 (Dec 8) plus the following sctp-related changes:

Changes are fine. Ugh. Ok, I'll try your new reproducer here.

Heh I wasn't going to reproduce this by myself anytime soon, I think. It's using the same socket to connect to itself, and only happens if the connect() gets there before the listen() call. Figured this out because I could only reproduce it under strace at first.

Please give this other patch a try. A state command (sctp_sf_cookie_wait_prm_abort) was issuing SCTP_CMD_INIT_FAILED, which leads to SCTP_CMD_DELETE_TCB, but returning SCTP_DISPOSITION_CONSUME, which fooled the patch.

---8<---
commit 9f84d50e36cee0ce66e4ce9b3b1665e0a1dbcdd3

Author: Marcelo Ricardo Leitner <[email protected]>
Date:   Fri Dec 4 15:30:23 2015 -0200

sctp: fix use-after-free in pr_debug statement

Dmitry Vyukov reported a use-after-free in the code expanded by the macro debug_post_sfx, which is caused by the use of the asoc pointer after it was freed within sctp_side_effect() scope.

This patch fixes it by allowing sctp_side_effect to clear that asoc pointer when the TCB is freed.

As Vlad explained, we also have to cover the SCTP_DISPOSITION_ABORT case because it will trigger DELETE_TCB too on that same loop.

Also, there was a place issuing SCTP_CMD_INIT_FAILED but returning SCTP_DISPOSITION_CONSUME, which would fool the scheme above. Fix it by returning SCTP_DISPOSITION_ABORT instead.

The macro is already prepared to handle such NULL pointer.

Reported-by Dmitry Vyukov <[email protected]>

net/sctp/sm_sideeffect.c +6 -5

index 6098d4c42fa9..be23d5c2074f 100644
@@ -63,7 +63,7 @@ static int sctp_cmd_interpreter(sctp_event_t event_type,
 static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
                             sctp_state_t state,
                             struct sctp_endpoint *ep,
-                            struct sctp_association *asoc,
+                            struct sctp_association **asoc,
                             void *event_arg,
                             sctp_disposition_t status,
                             sctp_cmd_seq_t *commands,

@@ -1123,7 +1123,7 @@ int sctp_do_sm(struct net *net, sctp_event_t event_type, sctp_subtype_t subtype,
        debug_post_sfn();
 
        error = sctp_side_effects(event_type, subtype, state,
-                                 ep, asoc, event_arg, status,
+                                 ep, &asoc, event_arg, status,
                                  &commands, gfp);
        debug_post_sfx();
 

@@ -1136,7 +1136,7 @@ int sctp_do_sm(struct net *net, sctp_event_t event_type, sctp_subtype_t subtype,
 static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
                             sctp_state_t state,
                             struct sctp_endpoint *ep,
-                            struct sctp_association *asoc,
+                            struct sctp_association **asoc,
                             void *event_arg,
                             sctp_disposition_t status,
                             sctp_cmd_seq_t *commands,

@@ -1151,7 +1151,7 @@ static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
         * disposition SCTP_DISPOSITION_CONSUME.
         */
        if (0 != (error = sctp_cmd_interpreter(event_type, subtype, state,
-                                              ep, asoc,
+                                              ep, *asoc,
                                               event_arg, status,
                                               commands, gfp)))
                goto bail;

@@ -1174,11 +1174,12 @@ static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
                break;
 
        case SCTP_DISPOSITION_DELETE_TCB:
+       case SCTP_DISPOSITION_ABORT:
                /* This should now be a command. */
+               *asoc = NULL;
                break;
 
        case SCTP_DISPOSITION_CONSUME:
-       case SCTP_DISPOSITION_ABORT:
                /*
                 * We should no longer have much work to do here as the
                 * real work has been done as explicit commands above.

net/sctp/sm_statefuns.c +0 -2

index 6f46aa16cb76..d801e151498a 100644
@@ -4959,12 +4959,10 @@ sctp_disposition_t sctp_sf_cookie_wait_prm_abort(
        sctp_cmd_seq_t *commands)
 {
        struct sctp_chunk *abort = arg;
-       sctp_disposition_t retval;
 
        /* Stop T1-init timer */
        sctp_add_cmd_sf(commands, SCTP_CMD_TIMER_STOP,
                        SCTP_TO(SCTP_EVENT_TIMEOUT_T1_INIT));
-       retval = SCTP_DISPOSITION_CONSUME;
 
        sctp_add_cmd_sf(commands, SCTP_CMD_REPLY, SCTP_CHUNK(abort));
 

net/sctp/sm_statefuns.c sctp_disposition_t sctp_sf_cookie_wait_prm_abort( · line 4987

Quoted 8 lines

net/sctp/sm_statefuns.c +1 -1

@@ -4983,7 +4981,7 @@ sctp_disposition_t sctp_sf_cookie_wait_prm_abort(
        sctp_add_cmd_sf(commands, SCTP_CMD_INIT_FAILED,
                        SCTP_PERR(SCTP_ERROR_USER_ABORT));
 
-       return retval;
+       return SCTP_DISPOSITION_ABORT;
 }
 
 /*

Still happens...
I am on commit aa53685549a2cfb5f175b0c4a20bc9aa1e5a1b85 with your latest patch applied.
Can you figure out what happens now from the report below? If not I can create a repro, it's just somewhat time consuming.

Quoted 2 lines

I can imagine. I don't know how this fuzzer works, but it would be nice if this reproducer extractor could be executed easier. So far, we have

Quoted 10 lines

identified 3 different issues already leading to this bug:
- 1st, the handling on DELETE_TCB
- 2nd, the handling on DISPOSITION_ABORT
- 3rd, the bad combination on internal state-machine command to a return value

I can and will review it again, but it's doing nasty stuff like using the same socket to connect to itself. It's hard to imagine all those combinations in mind that might lead to that use-after-free.

Quoted 1 line

Keep you posted.. thanks.

Found a similar place in abort primitive handling like in this last patch update, it's probably the issue you're still triggering.

Also found another place that may lead to this use after free, in case we receive a packet with a chunk that has no data.

Oh my.. :)

Marcelo

Quoted 153 lines · Marcelo Ricardo Leitner

Quoted 152 lines · Marcelo Ricardo Leitner

Em 11-12-2015 11:35, Dmitry Vyukov escreveu:

Quoted 150 lines

Quoted 133 lines · Marcelo Ricardo Leitner

Quoted 19 lines · Marcelo Ricardo Leitner

Quoted 3 lines · Dmitry Vyukov

Quoted 2 lines · Dmitry Vyukov

On Tue, Dec 8, 2015 at 6:40 PM, Marcelo Ricardo Leitner <[email protected]> wrote:

...

Quoted 14 lines

Quoted 10 lines

Quoted 7 lines

The patches were combined already, but this last pick by Vlad is just not yet patched. It's not necessary for your testing and I didn't want to interrupt it in case you were already testing it.

You can use my last patch here, from 2 emails ago, the one which contains this line:
- case SCTP_DISPOSITION_ABORT:

You are right. I missed that they are combined. Testing with it now.

Quoted 5 lines

Quoted 3 lines

Use-after-free still happens.
I am on commit aa53685549a2cfb5f175b0c4a20bc9aa1e5a1b85 (Dec 8) plus the following sctp-related changes:

Changes are fine. Ugh. Ok, I'll try your new reproducer here.

Heh I wasn't going to reproduce this by myself anytime soon, I think. It's using the same socket to connect to itself, and only happens if the connect() gets there before the listen() call. Figured this out because I could only reproduce it under strace at first.

Please give this other patch a try. A state command (sctp_sf_cookie_wait_prm_abort) was issuing SCTP_CMD_INIT_FAILED, which leads to SCTP_CMD_DELETE_TCB, but returning SCTP_DISPOSITION_CONSUME, which fooled the patch.

---8<---
commit 9f84d50e36cee0ce66e4ce9b3b1665e0a1dbcdd3

Author: Marcelo Ricardo Leitner <[email protected]>
Date:   Fri Dec 4 15:30:23 2015 -0200

sctp: fix use-after-free in pr_debug statement

Dmitry Vyukov reported a use-after-free in the code expanded by the macro debug_post_sfx, which is caused by the use of the asoc pointer after it was freed within sctp_side_effect() scope.

This patch fixes it by allowing sctp_side_effect to clear that asoc pointer when the TCB is freed.

As Vlad explained, we also have to cover the SCTP_DISPOSITION_ABORT case because it will trigger DELETE_TCB too on that same loop.

Also, there was a place issuing SCTP_CMD_INIT_FAILED but returning SCTP_DISPOSITION_CONSUME, which would fool the scheme above. Fix it by returning SCTP_DISPOSITION_ABORT instead.

The macro is already prepared to handle such NULL pointer.

Reported-by Dmitry Vyukov <[email protected]>

net/sctp/sm_sideeffect.c +6 -5

index 6098d4c42fa9..be23d5c2074f 100644
@@ -63,7 +63,7 @@ static int sctp_cmd_interpreter(sctp_event_t event_type,
  static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
                              sctp_state_t state,
                              struct sctp_endpoint *ep,
-                            struct sctp_association *asoc,
+                            struct sctp_association **asoc,
                              void *event_arg,
                              sctp_disposition_t status,
                              sctp_cmd_seq_t *commands,

@@ -1123,7 +1123,7 @@ int sctp_do_sm(struct net *net, sctp_event_t event_type, sctp_subtype_t subtype,
         debug_post_sfn();
 
         error = sctp_side_effects(event_type, subtype, state,
-                                 ep, asoc, event_arg, status,
+                                 ep, &asoc, event_arg, status,
                                   &commands, gfp);
         debug_post_sfx();
 

@@ -1136,7 +1136,7 @@ int sctp_do_sm(struct net *net, sctp_event_t event_type, sctp_subtype_t subtype,
  static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
                              sctp_state_t state,
                              struct sctp_endpoint *ep,
-                            struct sctp_association *asoc,
+                            struct sctp_association **asoc,
                              void *event_arg,
                              sctp_disposition_t status,
                              sctp_cmd_seq_t *commands,

@@ -1151,7 +1151,7 @@ static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
          * disposition SCTP_DISPOSITION_CONSUME.
          */
         if (0 != (error = sctp_cmd_interpreter(event_type, subtype, state,
-                                              ep, asoc,
+                                              ep, *asoc,
                                                event_arg, status,
                                                commands, gfp)))
                 goto bail;

@@ -1174,11 +1174,12 @@ static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
                 break;
 
         case SCTP_DISPOSITION_DELETE_TCB:
+       case SCTP_DISPOSITION_ABORT:
                 /* This should now be a command. */
+               *asoc = NULL;
                 break;
 
         case SCTP_DISPOSITION_CONSUME:
-       case SCTP_DISPOSITION_ABORT:
                 /*
                  * We should no longer have much work to do here as the
                  * real work has been done as explicit commands above.

net/sctp/sm_statefuns.c +0 -2

index 6f46aa16cb76..d801e151498a 100644
@@ -4959,12 +4959,10 @@ sctp_disposition_t sctp_sf_cookie_wait_prm_abort(
         sctp_cmd_seq_t *commands)
  {
         struct sctp_chunk *abort = arg;
-       sctp_disposition_t retval;
 
         /* Stop T1-init timer */
         sctp_add_cmd_sf(commands, SCTP_CMD_TIMER_STOP,
                         SCTP_TO(SCTP_EVENT_TIMEOUT_T1_INIT));
-       retval = SCTP_DISPOSITION_CONSUME;
 
         sctp_add_cmd_sf(commands, SCTP_CMD_REPLY, SCTP_CHUNK(abort));
 

net/sctp/sm_statefuns.c sctp_disposition_t sctp_sf_cookie_wait_prm_abort( · line 4987

Quoted 8 lines

net/sctp/sm_statefuns.c +1 -1

@@ -4983,7 +4981,7 @@ sctp_disposition_t sctp_sf_cookie_wait_prm_abort(
         sctp_add_cmd_sf(commands, SCTP_CMD_INIT_FAILED,
                         SCTP_PERR(SCTP_ERROR_USER_ABORT));
 
-       return retval;
+       return SCTP_DISPOSITION_ABORT;
  }
 
  /*

Still happens...
I am on commit aa53685549a2cfb5f175b0c4a20bc9aa1e5a1b85 with your latest patch applied.
Can you figure out what happens now from the report below? If not I can create a repro, it's just somewhat time consuming.

Quoted 2 lines

Quoted 2 lines

I can imagine. I don't know how this fuzzer works, but it would be nice if this reproducer extractor could be executed easier. So far, we have

It would be very nice, but it is not always trivial.

Fuzzer pretty much tried to trigger everything that is triggerable from user-space. Sometimes what it does can make no sense. But it is still super-important for contexts like Android, where programs can be as malicious as you can imagine and the system heavily relies on kernel for protection.

Quoted 10 lines

Quoted 10 lines

identified 3 different issues already leading to this bug:
- 1st, the handling on DELETE_TCB
- 2nd, the handling on DISPOSITION_ABORT
- 3rd, the bad combination on internal state-machine command to a return value

I can and will review it again, but it's doing nasty stuff like using the same socket to connect to itself. It's hard to imagine all those combinations in mind that might lead to that use-after-free.

Quoted 7 lines

Quoted 1 line

Keep you posted.. thanks.

Found a similar place in abort primitive handling like in this last patch update, it's probably the issue you're still triggering.

Also found another place that may lead to this use after free, in case we receive a packet with a chunk that has no data.

I see that sctp_cmd_interpreter does:

     sctp_cmd_delete_tcb(commands, asoc);
     asoc = NULL;

Won't it be simpler to pass sctp_association ** to this function and let it clear it whenever it decides to free the objects, rather than try to duplicate its logic on higher level. Just a blind thought.

Couldn't get syzkaller working over here, so I still need your help on testing this. I expect this will be the last cycle, though.

If it does generate another trace, I'll need the reproducer too because I can't find anything else just with code review.

Thanks

--8<--

Dmitry Vyukov reported a use-after-free in the code expanded by the macro debug_post_sfx, which is caused by the use of the asoc pointer after it was freed within sctp_side_effect() scope.

This patch fixes it by allowing sctp_side_effect to clear that asoc pointer when the TCB is freed.

As Vlad explained, we also have to cover the SCTP_DISPOSITION_ABORT case because it will trigger DELETE_TCB too on that same loop.

Also, there were places issuing SCTP_CMD_INIT_FAILED and ASSOC_FAILED but returning SCTP_DISPOSITION_CONSUME, which would fool the scheme above. Fix it by returning SCTP_DISPOSITION_ABORT instead.

The macro is already prepared to handle such NULL pointer.

Reported-by Dmitry Vyukov <[email protected]>

Signed-off-by Marcelo Ricardo Leitner <[email protected]>

---
 net/sctp/sm_sideeffect.c | 11 ++++++-----
 net/sctp/sm_statefuns.c  | 17 ++++-------------
 2 files changed, 10 insertions(+), 18 deletions(-)

net/sctp/sm_sideeffect.c +6 -5

index 4f170ad38ff4f7d345d8e3a3fee7d691df64d9cb..2e21384697c2a6a5fd045142bcd9c39992d3867f 100644
@@ -63,7 +63,7 @@ static int sctp_cmd_interpreter(sctp_event_t event_type,
 static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
 			     sctp_state_t state,
 			     struct sctp_endpoint *ep,
-			     struct sctp_association *asoc,
+			     struct sctp_association **asoc,
 			     void *event_arg,
 			     sctp_disposition_t status,
 			     sctp_cmd_seq_t *commands,

@@ -1125,7 +1125,7 @@ int sctp_do_sm(struct net *net, sctp_event_t event_type, sctp_subtype_t subtype,
 	debug_post_sfn();
 
 	error = sctp_side_effects(event_type, subtype, state,
-				  ep, asoc, event_arg, status,
+				  ep, &asoc, event_arg, status,
 				  &commands, gfp);
 	debug_post_sfx();
 

@@ -1138,7 +1138,7 @@ int sctp_do_sm(struct net *net, sctp_event_t event_type, sctp_subtype_t subtype,
 static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
 			     sctp_state_t state,
 			     struct sctp_endpoint *ep,
-			     struct sctp_association *asoc,
+			     struct sctp_association **asoc,
 			     void *event_arg,
 			     sctp_disposition_t status,
 			     sctp_cmd_seq_t *commands,

@@ -1153,7 +1153,7 @@ static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
 	 * disposition SCTP_DISPOSITION_CONSUME.
 	 */
 	if (0 != (error = sctp_cmd_interpreter(event_type, subtype, state,
-					       ep, asoc,
+					       ep, *asoc,
 					       event_arg, status,
 					       commands, gfp)))
 		goto bail;

@@ -1176,11 +1176,12 @@ static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
 		break;
 
 	case SCTP_DISPOSITION_DELETE_TCB:
+	case SCTP_DISPOSITION_ABORT:
 		/* This should now be a command. */
+		*asoc = NULL;
 		break;
 
 	case SCTP_DISPOSITION_CONSUME:
-	case SCTP_DISPOSITION_ABORT:
 		/*
 		 * We should no longer have much work to do here as the
 		 * real work has been done as explicit commands above.

net/sctp/sm_statefuns.c +4 -13

index 22c2bf367d7e8c7025065f33eabfd7e93a7f4021..f1f08c8f277bd8719299d1ed21eb23e36d55f7e2 100644
@@ -2976,7 +2976,7 @@ sctp_disposition_t sctp_sf_eat_data_6_2(struct net *net,
 		SCTP_INC_STATS(net, SCTP_MIB_IN_DATA_CHUNK_DISCARDS);
 		goto discard_force;
 	case SCTP_IERROR_NO_DATA:
-		goto consume;
+		return SCTP_DISPOSITION_ABORT;
 	case SCTP_IERROR_PROTO_VIOLATION:
 		return sctp_sf_abort_violation(net, ep, asoc, chunk, commands,
 			(u8 *)chunk->subh.data_hdr, sizeof(sctp_datahdr_t));

@@ -3043,9 +3043,6 @@ discard_noforce:
 		sctp_add_cmd_sf(commands, SCTP_CMD_GEN_SACK, force);
 
 	return SCTP_DISPOSITION_DISCARD;
-consume:
-	return SCTP_DISPOSITION_CONSUME;
-
 }
 
 /*

@@ -3093,7 +3090,7 @@ sctp_disposition_t sctp_sf_eat_data_fast_4_4(struct net *net,
 	case SCTP_IERROR_BAD_STREAM:
 		break;
 	case SCTP_IERROR_NO_DATA:
-		goto consume;
+		return SCTP_DISPOSITION_ABORT;
 	case SCTP_IERROR_PROTO_VIOLATION:
 		return sctp_sf_abort_violation(net, ep, asoc, chunk, commands,
 			(u8 *)chunk->subh.data_hdr, sizeof(sctp_datahdr_t));

@@ -3119,7 +3116,6 @@ sctp_disposition_t sctp_sf_eat_data_fast_4_4(struct net *net,
 				SCTP_TO(SCTP_EVENT_TIMEOUT_T2_SHUTDOWN));
 	}
 
-consume:
 	return SCTP_DISPOSITION_CONSUME;
 }
 

@@ -4825,9 +4821,6 @@ sctp_disposition_t sctp_sf_do_9_1_prm_abort(
 	 * if necessary to fill gaps.
 	 */
 	struct sctp_chunk *abort = arg;
-	sctp_disposition_t retval;
-
-	retval = SCTP_DISPOSITION_CONSUME;
 
 	if (abort)
 		sctp_add_cmd_sf(commands, SCTP_CMD_REPLY, SCTP_CHUNK(abort));

@@ -4845,7 +4838,7 @@ sctp_disposition_t sctp_sf_do_9_1_prm_abort(
 	SCTP_INC_STATS(net, SCTP_MIB_ABORTEDS);
 	SCTP_DEC_STATS(net, SCTP_MIB_CURRESTAB);
 
-	return retval;
+	return SCTP_DISPOSITION_ABORT;
 }
 
 /* We tried an illegal operation on an association which is closed.  */

@@ -4960,12 +4953,10 @@ sctp_disposition_t sctp_sf_cookie_wait_prm_abort(
 	sctp_cmd_seq_t *commands)
 {
 	struct sctp_chunk *abort = arg;
-	sctp_disposition_t retval;
 
 	/* Stop T1-init timer */
 	sctp_add_cmd_sf(commands, SCTP_CMD_TIMER_STOP,
 			SCTP_TO(SCTP_EVENT_TIMEOUT_T1_INIT));
-	retval = SCTP_DISPOSITION_CONSUME;
 
 	if (abort)
 		sctp_add_cmd_sf(commands, SCTP_CMD_REPLY, SCTP_CHUNK(abort));

@@ -4985,7 +4976,7 @@ sctp_disposition_t sctp_sf_cookie_wait_prm_abort(
 	sctp_add_cmd_sf(commands, SCTP_CMD_INIT_FAILED,
 			SCTP_PERR(SCTP_ERROR_USER_ABORT));
 
-	return retval;
+	return SCTP_DISPOSITION_ABORT;
 }
 
 /*

Dmitry Vyukov reported a use-after-free in the code expanded by the macro debug_post_sfx, which is caused by the use of the asoc pointer after it was freed within sctp_side_effect() scope.

This patch fixes it by allowing sctp_side_effect to clear that asoc pointer when the TCB is freed.

As Vlad explained, we also have to cover the SCTP_DISPOSITION_ABORT case because it will trigger DELETE_TCB too on that same loop.

Also, there were places issuing SCTP_CMD_INIT_FAILED and ASSOC_FAILED but returning SCTP_DISPOSITION_CONSUME, which would fool the scheme above. Fix it by returning SCTP_DISPOSITION_ABORT instead.

The macro is already prepared to handle such NULL pointer.

Reported-by Dmitry Vyukov <[email protected]>

Signed-off-by Marcelo Ricardo Leitner <[email protected]>

From: Marcelo Ricardo Leitner <[email protected]>
Date: Fri, 8 Jan 2016 11:00:54 -0200

Quoted 14 lines

Dmitry Vyukov reported a use-after-free in the code expanded by the macro debug_post_sfx, which is caused by the use of the asoc pointer after it was freed within sctp_side_effect() scope.

This patch fixes it by allowing sctp_side_effect to clear that asoc pointer when the TCB is freed.

As Vlad explained, we also have to cover the SCTP_DISPOSITION_ABORT case because it will trigger DELETE_TCB too on that same loop.

Also, there were places issuing SCTP_CMD_INIT_FAILED and ASSOC_FAILED but returning SCTP_DISPOSITION_CONSUME, which would fool the scheme above. Fix it by returning SCTP_DISPOSITION_ABORT instead.

Quoted 4 lines

The macro is already prepared to handle such NULL pointer.

Reported-by Dmitry Vyukov <[email protected]>

Signed-off-by Marcelo Ricardo Leitner <[email protected]>

Applied, thank you.

Quoted 160 lines · Marcelo Ricardo Leitner

Em 11-12-2015 11:35, Dmitry Vyukov escreveu:

Quoted 142 lines

Quoted 140 lines · Marcelo Ricardo Leitner

Quoted 19 lines · Marcelo Ricardo Leitner

Quoted 3 lines · Dmitry Vyukov

Quoted 2 lines · Dmitry Vyukov

On Tue, Dec 8, 2015 at 6:40 PM, Marcelo Ricardo Leitner <[email protected]> wrote:

...

Quoted 14 lines

Quoted 10 lines

Quoted 7 lines

The patches were combined already, but this last pick by Vlad is just not yet patched. It's not necessary for your testing and I didn't want to interrupt it in case you were already testing it.

You can use my last patch here, from 2 emails ago, the one which contains this line:
- case SCTP_DISPOSITION_ABORT:

You are right. I missed that they are combined. Testing with it now.

Quoted 5 lines

Quoted 3 lines

Use-after-free still happens.
I am on commit aa53685549a2cfb5f175b0c4a20bc9aa1e5a1b85 (Dec 8) plus the following sctp-related changes:

Changes are fine. Ugh. Ok, I'll try your new reproducer here.

Heh I wasn't going to reproduce this by myself anytime soon, I think. It's using the same socket to connect to itself, and only happens if the connect() gets there before the listen() call. Figured this out because I could only reproduce it under strace at first.

Please give this other patch a try. A state command (sctp_sf_cookie_wait_prm_abort) was issuing SCTP_CMD_INIT_FAILED, which leads to SCTP_CMD_DELETE_TCB, but returning SCTP_DISPOSITION_CONSUME, which fooled the patch.

---8<---
commit 9f84d50e36cee0ce66e4ce9b3b1665e0a1dbcdd3

Author: Marcelo Ricardo Leitner <[email protected]>
Date:   Fri Dec 4 15:30:23 2015 -0200

sctp: fix use-after-free in pr_debug statement

Dmitry Vyukov reported a use-after-free in the code expanded by the macro debug_post_sfx, which is caused by the use of the asoc pointer after it was freed within sctp_side_effect() scope.

This patch fixes it by allowing sctp_side_effect to clear that asoc pointer when the TCB is freed.

As Vlad explained, we also have to cover the SCTP_DISPOSITION_ABORT case because it will trigger DELETE_TCB too on that same loop.

Also, there was a place issuing SCTP_CMD_INIT_FAILED but returning SCTP_DISPOSITION_CONSUME, which would fool the scheme above. Fix it by returning SCTP_DISPOSITION_ABORT instead.

The macro is already prepared to handle such NULL pointer.

Reported-by Dmitry Vyukov <[email protected]>

net/sctp/sm_sideeffect.c +6 -5

index 6098d4c42fa9..be23d5c2074f 100644
@@ -63,7 +63,7 @@ static int sctp_cmd_interpreter(sctp_event_t event_type,
 static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
                             sctp_state_t state,
                             struct sctp_endpoint *ep,
-                            struct sctp_association *asoc,
+                            struct sctp_association **asoc,
                             void *event_arg,
                             sctp_disposition_t status,
                             sctp_cmd_seq_t *commands,

@@ -1123,7 +1123,7 @@ int sctp_do_sm(struct net *net, sctp_event_t event_type, sctp_subtype_t subtype,
        debug_post_sfn();
 
        error = sctp_side_effects(event_type, subtype, state,
-                                 ep, asoc, event_arg, status,
+                                 ep, &asoc, event_arg, status,
                                  &commands, gfp);
        debug_post_sfx();
 

@@ -1136,7 +1136,7 @@ int sctp_do_sm(struct net *net, sctp_event_t event_type, sctp_subtype_t subtype,
 static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
                             sctp_state_t state,
                             struct sctp_endpoint *ep,
-                            struct sctp_association *asoc,
+                            struct sctp_association **asoc,
                             void *event_arg,
                             sctp_disposition_t status,
                             sctp_cmd_seq_t *commands,

@@ -1151,7 +1151,7 @@ static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
         * disposition SCTP_DISPOSITION_CONSUME.
         */
        if (0 != (error = sctp_cmd_interpreter(event_type, subtype, state,
-                                              ep, asoc,
+                                              ep, *asoc,
                                               event_arg, status,
                                               commands, gfp)))
                goto bail;

@@ -1174,11 +1174,12 @@ static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
                break;
 
        case SCTP_DISPOSITION_DELETE_TCB:
+       case SCTP_DISPOSITION_ABORT:
                /* This should now be a command. */
+               *asoc = NULL;
                break;
 
        case SCTP_DISPOSITION_CONSUME:
-       case SCTP_DISPOSITION_ABORT:
                /*
                 * We should no longer have much work to do here as the
                 * real work has been done as explicit commands above.

net/sctp/sm_statefuns.c +1 -3

index 6f46aa16cb76..d801e151498a 100644
@@ -4959,12 +4959,10 @@ sctp_disposition_t sctp_sf_cookie_wait_prm_abort(
        sctp_cmd_seq_t *commands)
 {
        struct sctp_chunk *abort = arg;
-       sctp_disposition_t retval;
 
        /* Stop T1-init timer */
        sctp_add_cmd_sf(commands, SCTP_CMD_TIMER_STOP,
                        SCTP_TO(SCTP_EVENT_TIMEOUT_T1_INIT));
-       retval = SCTP_DISPOSITION_CONSUME;
 
        sctp_add_cmd_sf(commands, SCTP_CMD_REPLY, SCTP_CHUNK(abort));
 

@@ -4983,7 +4981,7 @@ sctp_disposition_t sctp_sf_cookie_wait_prm_abort(
        sctp_add_cmd_sf(commands, SCTP_CMD_INIT_FAILED,
                        SCTP_PERR(SCTP_ERROR_USER_ABORT));
 
-       return retval;
+       return SCTP_DISPOSITION_ABORT;
 }
 

Quoted 8 lines

net/sctp/sm_statefuns.c sctp_disposition_t sctp_sf_cookie_wait_prm_abort( · line 4987

Quoted 1 line

net/sctp/sm_statefuns.c +0 -0

 /*

Still happens...
I am on commit aa53685549a2cfb5f175b0c4a20bc9aa1e5a1b85 with your latest patch applied.
Can you figure out what happens now from the report below? If not I can create a repro, it's just somewhat time consuming.

I can imagine. I don't know how this fuzzer works, but it would be nice if this reproducer extractor could be executed easier. So far, we have identified 3 different issues already leading to this bug:
- 1st, the handling on DELETE_TCB
- 2nd, the handling on DISPOSITION_ABORT
- 3rd, the bad combination on internal state-machine command to a return value

Quoted 5 lines

I can and will review it again, but it's doing nasty stuff like using the same socket to connect to itself. It's hard to imagine all those combinations in mind that might lead to that use-after-free.

Keep you posted.. thanks.