[LTP] [PATCH] Add regression test for CVE-2017-17052

[Cyril Hrubis]

Hi!
> ---
>  testcases/cve/Makefile         |  2 ++
>  testcases/cve/cve-2017-17052.c | 68 ++++++++++++++++++++++++++++++++++++++++++

You are missing entry in cve runtest file runtest/cve and also
.gitignore entry.


>  2 files changed, 70 insertions(+)
>  create mode 100644 testcases/cve/cve-2017-17052.c
> 
> diff --git a/testcases/cve/Makefile b/testcases/cve/Makefile
> index 0905fd95c..22dca3b3f 100644
> --- a/testcases/cve/Makefile
> +++ b/testcases/cve/Makefile
> @@ -30,4 +30,6 @@ cve-2014-0196:  LDLIBS += -lrt -lutil
>  cve-2017-2671:	CFLAGS += -pthread
>  cve-2017-2671:	LDLIBS += -lrt
>  
> +cve-2017-17052:	CFLAGS += -pthread
> +
>  include $(top_srcdir)/include/mk/generic_leaf_target.mk
> diff --git a/testcases/cve/cve-2017-17052.c b/testcases/cve/cve-2017-17052.c
> new file mode 100644
> index 000000000..7103685f2
> --- /dev/null
> +++ b/testcases/cve/cve-2017-17052.c
> @@ -0,0 +1,68 @@
> +/*
> + * Copyright (c) 2018 Michael Moese <mmoese@suse.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-2017-17052, original reproducer can be found here:
> + * https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=2b7e8665b4ff51c034c55df3cff76518d1a9ee3a
> + */
> +
> +#include <unistd.h>
> +#include <pthread.h>
> +#include <stdlib.h>
> +#include <sys/mman.h>
> +#include <sys/wait.h>
> +#include <sys/syscall.h>
> +
> +#include "tst_test.h"
> +#include "lapi/syscalls.h"
> +
> +static void *mmap_thread(void *_arg)
> +{
> +	for (;;) {
> +		mmap(NULL, 0x1000000, PROT_READ,
> +			MAP_POPULATE|MAP_ANONYMOUS|MAP_PRIVATE, -1, 0);
> +	}
> +}
> +
> +static void *fork_thread(void *_arg)
> +{
> +	usleep(rand() % 10000);
> +	fork();
> +}
> +
> +static void run(void)
> +{
> +	fork();
> +	fork();
> +	fork();
> +	for(;;) {
> +		if (fork() == 0) {
> +			pthread_t t;
> +
> +			pthread_create(&t, NULL, mmap_thread, NULL);
> +			pthread_create(&t, NULL, fork_thread, NULL);
> +			usleep(rand() % 10000);
> +			syscall(__NR_exit_group, 0);
> +		}
> +		wait(NULL);
> +	}

You are supposed to use SAFE_ macros for most of these calls, i.e.
SAFE_FORK(), SAFE_PTHREAD_CREATE(), SAVE_WAIT() etc.

Also this test will timeout on kernel without this bug. We should choose
minimal number of iterations for the for() loop here that reasonably
reliably reproduces the problem so that the test has chance to finish in
reasonable time. And also print tst_res(TPASS, "kernel survived"); or
something here at the end of the run() function, otherwise the test will
fail because it haven't reported any results.

Moreover what happens when the NULL dereference is triggered? I suppose
that the offending process is killed with SIGKILL? Then we should check
the return value from the wait() inside of the for() loop and report
that the failure has been reproduced with something as
tst_res(TFAIL, "bug reproduced!");

And also if the number of iterations needed for the bug to be reproduced
takes more than a second or so, you should consider using the fuzzy sync
library see include/tst_fuzzy_sync.h.

-- 
Cyril Hrubis
chrubis@suse.cz