[LTP] [PATCH] Test for CVE-2016-7042 in /proc/keys show function

[Guangwen Feng]

Hi!

Thanks for your review.

在 07/25/2017 04:07 PM, Richard Palethorpe 写道:
> Hello,
> 
> Guangwen Feng writes:
> 
>> Signed-off-by: Guangwen Feng <fenggw-fnst@cn.fujitsu.com>
>> ---
>>  runtest/cve                   |  1 +
>>  testcases/cve/.gitignore      |  1 +
>>  testcases/cve/cve-2016-7042.c | 98 +++++++++++++++++++++++++++++++++++++++++++
>>  3 files changed, 100 insertions(+)
>>  create mode 100644 testcases/cve/cve-2016-7042.c
>>
>> diff --git a/runtest/cve b/runtest/cve
>> index ee0614a..149d4a2 100644
>> --- a/runtest/cve
>> +++ b/runtest/cve
>> @@ -3,6 +3,7 @@ cve-2012-0957 cve-2012-0957
>>  cve-2014-0196 cve-2014-0196
>>  cve-2016-4997 cve-2016-4997
>>  cve-2016-5195 dirtyc0w
>> +cve-2016-7042 cve-2016-7042
>>  cve-2016-7117 cve-2016-7117
>>  cve-2017-5669 cve-2017-5669
>>  cve-2017-6951 cve-2017-6951
>> diff --git a/testcases/cve/.gitignore b/testcases/cve/.gitignore
>> index bdb73f3..4922bef 100644
>> --- a/testcases/cve/.gitignore
>> +++ b/testcases/cve/.gitignore
>> @@ -1,6 +1,7 @@
>>  cve-2012-0957
>>  cve-2014-0196
>>  cve-2016-4997
>> +cve-2016-7042
>>  cve-2016-7117
>>  cve-2017-6951
>>  cve-2017-5669
>> diff --git a/testcases/cve/cve-2016-7042.c b/testcases/cve/cve-2016-7042.c
>> new file mode 100644
>> index 0000000..89c59c9
>> --- /dev/null
>> +++ b/testcases/cve/cve-2016-7042.c
>> @@ -0,0 +1,98 @@
>> +/*
>> + * Copyright (c) 2017 Fujitsu Ltd.
>> + * Author: Guangwen Feng <fenggw-fnst@cn.fujitsu.com>
>> + *
>> + * This program is free software: you can redistribute it and/or modify
>> + * it under the terms of the GNU General Public License as published by
>> + * the Free Software Foundation, either version 2 of the License, or
>> + * (at your option) any later version.
>> + *
>> + * This program is distributed in the hope that it will be useful,
>> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
>> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
>> + * GNU General Public License for more details.
>> + *
>> + * You should have received a copy of the GNU General Public License
>> + * along with this program, if not, see <http://www.gnu.org/licenses/>.
>> + */
>> +
>> +/*
>> + * Test for CVE-2016-7042, this regression test can crash the buggy kernel
>> + * when the stack-protector is enabled, and the bug was fixed in:
>> + *
>> + *  commit 03dab869b7b239c4e013ec82aea22e181e441cfc
>> + *  Author: David Howells <dhowells@redhat.com>
>> + *  Date:   Wed Oct 26 15:01:54 2016 +0100
>> + *
>> + *  KEYS: Fix short sprintf buffer in /proc/keys show function
>> + */
>> +
>> +#include "config.h"
>> +#include <errno.h>
>> +#include <stdio.h>
>> +#include <sys/types.h>
>> +#ifdef HAVE_KEYUTILS_H
>> +# include <keyutils.h>
>> +#endif
> 
> This file is only being included for a typedef of int and some
> constants, so you could provide fallback definitions if it is not
> present. e.g. typedef int32_t key_serial_t. Otherwise it won't compile
> on a lot of systems by defualt where the CVE is easily exploitable.

This is very reasonable, thanks!

> 
> We should probably have something in include/lapi for keyutils, but you
> could just include the definitions here as well.

OK, I see.

> 
>> +#include "tst_test.h"
>> +#include "linux_syscall_numbers.h"
>> +
>> +#ifdef HAVE_KEYUTILS_H
>> +
>> +#define PATH_KEYS	"/proc/keys"
>> +
>> +static key_serial_t key;
>> +static int fd;
>> +
>> +static void do_test(void)
>> +{
>> +	char buf[BUFSIZ];
>> +
>> +	key = tst_syscall(__NR_add_key,
>> +		"user", "ltptestkey", "a", 1, KEY_SPEC_SESSION_KEYRING);
>> +	if (key == -1)
>> +		tst_brk(TBROK, "Failed to add key");
>> +
>> +	if (tst_syscall(__NR_keyctl, KEYCTL_UPDATE, key, "b", 1))
>> +		tst_brk(TBROK, "Failed to update key");
>> +
>> +	fd = SAFE_OPEN(PATH_KEYS, O_RDONLY);
>> +
>> +	// Will cause a panic due to stack corruption if bug occurs
>> +	SAFE_READ(0, fd, buf, BUFSIZ);
> 
> Please replace the comment with something like tst_res(TINFO,
> "Attempting to crash system...").

It sounds better, got it.
I will send v2 as your suggested, thanks!


Best Regards,
Guangwen Feng

> 
>> +
>> +	tst_res(TPASS, "Bug not reproduced");
>> +
>> +	SAFE_CLOSE(fd);
>> +
>> +	if (tst_syscall(__NR_keyctl, KEYCTL_UNLINK, key,
>> +		KEY_SPEC_SESSION_KEYRING))
>> +		tst_brk(TBROK, "Failed to unlink key");
>> +	key = 0;
>> +}
>> +
>> +static void setup(void)
>> +{
>> +	if (access(PATH_KEYS, F_OK))
>> +		tst_brk(TCONF, "%s does not exist", PATH_KEYS);
>> +}
>> +
>> +static void cleanup(void)
>> +{
>> +	if (key > 0 && tst_syscall(__NR_keyctl, KEYCTL_UNLINK, key,
>> +		KEY_SPEC_SESSION_KEYRING))
>> +		tst_res(TWARN, "Failed to unlink key");
>> +
>> +	if (fd > 0)
>> +		SAFE_CLOSE(fd);
>> +}
>> +
>> +static struct tst_test test = {
>> +	.setup = setup,
>> +	.cleanup = cleanup,
>> +	.test_all = do_test,
>> +};
>> +
>> +#else
>> +	TST_TEST_TCONF("keyutils.h does not exist");
>> +#endif /* HAVE_KEYUTILS_H */
>> -- 
>> 2.9.4
> 
> Great!
>