Linux kernel mirror (for testing) git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
kernel os linux

net: add validation for the socket syscall protocol argument

郭永刚 reported that one could simply crash the kernel as root by
using a simple program:

int socket_fd;
struct sockaddr_in addr;
addr.sin_port = 0;
addr.sin_addr.s_addr = INADDR_ANY;
addr.sin_family = 10;

socket_fd = socket(10,3,0x40000000);
connect(socket_fd , &addr,16);

AF_INET, AF_INET6 sockets actually only support 8-bit protocol
identifiers. inet_sock's skc_protocol field thus is sized accordingly,
thus larger protocol identifiers simply cut off the higher bits and
store a zero in the protocol fields.

This could lead to e.g. NULL function pointer because as a result of
the cut off inet_num is zero and we call down to inet_autobind, which
is NULL for raw sockets.

kernel: Call Trace:
kernel: [<ffffffff816db90e>] ? inet_autobind+0x2e/0x70
kernel: [<ffffffff816db9a4>] inet_dgram_connect+0x54/0x80
kernel: [<ffffffff81645069>] SYSC_connect+0xd9/0x110
kernel: [<ffffffff810ac51b>] ? ptrace_notify+0x5b/0x80
kernel: [<ffffffff810236d8>] ? syscall_trace_enter_phase2+0x108/0x200
kernel: [<ffffffff81645e0e>] SyS_connect+0xe/0x10
kernel: [<ffffffff81779515>] tracesys_phase2+0x84/0x89

I found no particular commit which introduced this problem.

CVE: CVE-2015-8543
Cc: Cong Wang <cwang@twopensource.com>
Reported-by: 郭永刚 <guoyonggang@360.cn>
Signed-off-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

authored by

Hannes Frederic Sowa and committed by
David S. Miller
79462ad0 20b08e1a

+16
+1
include/net/sock.h
··· 403 403 sk_no_check_rx : 1, 404 404 sk_userlocks : 4, 405 405 sk_protocol : 8, 406 + #define SK_PROTOCOL_MAX U8_MAX 406 407 sk_type : 16; 407 408 kmemcheck_bitfield_end(flags); 408 409 int sk_wmem_queued;
+3
net/ax25/af_ax25.c
··· 805 805 struct sock *sk; 806 806 ax25_cb *ax25; 807 807 808 + if (protocol < 0 || protocol > SK_PROTOCOL_MAX) 809 + return -EINVAL; 810 + 808 811 if (!net_eq(net, &init_net)) 809 812 return -EAFNOSUPPORT; 810 813
+3
net/decnet/af_decnet.c
··· 678 678 { 679 679 struct sock *sk; 680 680 681 + if (protocol < 0 || protocol > SK_PROTOCOL_MAX) 682 + return -EINVAL; 683 + 681 684 if (!net_eq(net, &init_net)) 682 685 return -EAFNOSUPPORT; 683 686
+3
net/ipv4/af_inet.c
··· 257 257 int try_loading_module = 0; 258 258 int err; 259 259 260 + if (protocol < 0 || protocol >= IPPROTO_MAX) 261 + return -EINVAL; 262 + 260 263 sock->state = SS_UNCONNECTED; 261 264 262 265 /* Look for the requested type/protocol pair. */
+3
net/ipv6/af_inet6.c
··· 109 109 int try_loading_module = 0; 110 110 int err; 111 111 112 + if (protocol < 0 || protocol >= IPPROTO_MAX) 113 + return -EINVAL; 114 + 112 115 /* Look for the requested type/protocol pair. */ 113 116 lookup_protocol: 114 117 err = -ESOCKTNOSUPPORT;
+3
net/irda/af_irda.c
··· 1086 1086 struct sock *sk; 1087 1087 struct irda_sock *self; 1088 1088 1089 + if (protocol < 0 || protocol > SK_PROTOCOL_MAX) 1090 + return -EINVAL; 1091 + 1089 1092 if (net != &init_net) 1090 1093 return -EAFNOSUPPORT; 1091 1094