[Intel-gfx] [PATCH i-g-t v3] lib/tests/igt_segfault.c Add unit test to test segfault handling

[Morton, Derek J]

>-----Original Message-----
>From: Daniel Vetter [mailto:daniel.vetter at ffwll.ch] On Behalf Of Daniel Vetter
>Sent: Thursday, May 28, 2015 10:05 AM
>To: Morton, Derek J
>Cc: intel-gfx at lists.freedesktop.org; Wood, Thomas
>Subject: Re: [Intel-gfx] [PATCH i-g-t v3] lib/tests/igt_segfault.c Add unit test to test segfault handling
>
>On Wed, May 27, 2015 at 04:54:24PM +0100, Derek Morton wrote:
>> Unit test to check a segfaulting subtest is handled correctly.
>> 
>> v2: Added script to check subtest results
>> v3: Removed script. Updated test to use fork to monitor return status.
>> 
>> Signed-off-by: Derek Morton <derek.j.morton at intel.com>
>> ---
>>  lib/tests/Makefile.sources |   1 +
>>  lib/tests/igt_segfault.c   | 139 +++++++++++++++++++++++++++++++++++++++++++++
>>  2 files changed, 140 insertions(+)
>>  create mode 100644 lib/tests/igt_segfault.c
>> 
>> diff --git a/lib/tests/Makefile.sources b/lib/tests/Makefile.sources 
>> index 10e0617..5fa0b31 100644
>> --- a/lib/tests/Makefile.sources
>> +++ b/lib/tests/Makefile.sources
>> @@ -8,6 +8,7 @@ check_PROGRAMS = \
>>  	igt_simple_test_subtests \
>>  	igt_timeout \
>>  	igt_invalid_subtest_name \
>> +	igt_segfault \
>>  	$(NULL)
>>  
>>  check_SCRIPTS = \
>> diff --git a/lib/tests/igt_segfault.c b/lib/tests/igt_segfault.c new 
>> file mode 100644 index 0000000..b420b1a
>> --- /dev/null
>> +++ b/lib/tests/igt_segfault.c
>> @@ -0,0 +1,139 @@
>> +/*
>> + * Copyright (c) 2015 Intel Corporation
>> + *
>> + * Permission is hereby granted, free of charge, to any person 
>> +obtaining a
>> + * copy of this software and associated documentation files (the 
>> +"Software"),
>> + * to deal in the Software without restriction, including without 
>> +limitation
>> + * the rights to use, copy, modify, merge, publish, distribute, 
>> +sublicense,
>> + * and/or sell copies of the Software, and to permit persons to whom 
>> +the
>> + * Software is furnished to do so, subject to the following conditions:
>> + *
>> + * The above copyright notice and this permission notice (including 
>> +the next
>> + * paragraph) shall be included in all copies or substantial portions 
>> +of the
>> + * Software.
>> + *
>> + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, 
>> +EXPRESS OR
>> + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF 
>> +MERCHANTABILITY,
>> + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT 
>> +SHALL
>> + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES 
>> +OR OTHER
>> + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, 
>> +ARISING
>> + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR 
>> +OTHER DEALINGS
>> + * IN THE SOFTWARE.
>> + *
>> + * Authors:
>> + *    Derek Morton <derek.j.morton at intel.com>
>> + *
>> + */
>> +
>> +/*
>> + * Testcase: Test the framework catches a segfault and returns an error.
>> + *
>> + * 1. Test a crashing simple test is reported.
>> + * 2. Test a crashing subtest is reported.
>> + * 3. Test a crashing subtest following a passing subtest is reported.
>> + * 4. Test a crashing subtest preceeding a passing subtest is reported.
>> + */
>> +
>> +#include <stdlib.h>
>> +#include <sys/wait.h>
>> +#include <sys/types.h>
>> +#include <assert.h>
>> +#include <errno.h>
>> +
>> +#include "drmtest.h"
>> +#include "igt_core.h"
>> +
>> +/*
>> + * We need to hide assert from the cocci igt test refactor spatch.
>> + *
>> + * IMPORTANT: Test infrastructure tests are the only valid places 
>> +where using
>> + * assert is allowed.
>> + */
>> +#define internal_assert assert
>> +
>> +bool simple;
>> +bool runa;
>> +bool runc;
>> +char test[] = "test";
>> +char *argv_run[] = { test };
>> +
>> +static int do_fork(void)
>> +{
>> +	int pid, status;
>> +	int argc;
>> +	void (*crashme)(void) = NULL;
>> +
>> +	switch (pid = fork()) {
>> +	case -1:
>> +		internal_assert(0);
>> +	case 0:
>> +		if (simple) {
>> +			argc = 1;
>> +			igt_simple_init(argc, argv_run);
>> +			crashme();
>> +
>> +			igt_exit();
>> +		} else {
>> +
>> +			argc = 1;
>> +			igt_subtest_init(argc, argv_run);
>> +
>> +			if(runa)
>> +				igt_subtest("A")
>> +					;
>> +
>> +			igt_subtest("B")
>> +				crashme();
>> +
>> +			if(runc)
>> +				igt_subtest("C")
>> +					;
>> +
>> +			igt_exit();
>> +		}
>> +	default:
>> +		while (waitpid(pid, &status, 0) == -1 &&
>> +		       errno == EINTR)
>> +			;
>> +
>> +		if(WIFSIGNALED(status))
>> +			return WTERMSIG(status) + 128;
>> +
>> +		return WEXITSTATUS(status);
>> +	}
>> +}
>> +
>> +int main(int argc, char **argv)
>> +{
>> +	/* Test Crash in simple test is reported */
>> +	simple = true;
>> +	runa=false;
>> +	runc=false;
>> +	igt_info("Simple test.\n");
>> +	fflush(stdout);
>> +	internal_assert(do_fork() == SIGSEGV + 128);
>> +
>> +	/* Test crash in a single subtest is reported */
>> +	simple = false;
>> +	igt_info("Single subtest.\n");
>> +	fflush(stdout);
>> +	internal_assert(do_fork() == SIGSEGV + 128);
>
>This one here fails for me when running make check. I didn't look into details. Also something seemed off with the content-encoding of your mail, at least git apply-mbox complained. I didn't debug but applied it manually.
>
>Also please add igt_segfault to the relevant .gitignore.
>-Daniel
>

I submitted a separate patch "[PATCH i-g-t] lib/igt_core.c: Flag the test as failing after a segfault" as segfaults were not being flagged as a failure. If that patch has not yet been merged I would expect this test to fail at this point. (I was asked to write this UT when I submitted that patch.)
Git send-email asked about encoding when I submitted the patch. I assumed it was the (c) character in the copyright header. I just selected whatever it recommended. (UTF8 I think). Previous patches have not been a problem. 
I will update the .gitignore and resubmit this patch.

//Derek

>> +
>> +	/* Test crash in a subtest following a pass is reported */
>> +	simple = false;
>> +	runa=true;
>> +	igt_info("Passing then crashing subtest.\n");
>> +	fflush(stdout);
>> +	internal_assert(do_fork() == SIGSEGV + 128);
>> +
>> +	/* Test crash in a subtest preceeding a pass is reported */
>> +	simple = false;
>> +	runa=false;
>> +	runc=true;
>> +	igt_info("Crashing then passing subtest.\n");
>> +	fflush(stdout);
>> +	internal_assert(do_fork() == SIGSEGV + 128);
>> +
>> +	return 0;
>> +}
>> +
>> --
>> 1.9.1
>> 
>
>
>-- 
>Daniel Vetter
>Software Engineer, Intel Corporation
>http://blog.ffwll.ch
>