[LTP] [PATCH v3] Hugetlb: Migrating libhugetlbfs test truncate_reserve_wraparound.c

[Cyril Hrubis]

Hi!
> Test to verify the Hugepages_Rsvd before and after mmap, truncate and sigbus operations.
> At one stage, improper handling of tests against i_size could mess
> up accounting of reserved hugepages on certain truncate operations.
> 
> Tested and verified that test passes for iterations: i2, i5 and i10
> 
> Signed-off-by: Geetika <geetika@linux.ibm.com>
> ---
> Changes in v3:
> - Fixed the indentation
> - Corrected all the tst_res() format string
> ---
> Changes in v2:
> - Modified the Description section of test
> - Removed definition of unsued RANDOM_CONSTANT
> - Fixed warning: unused parameter ‘signum’ [-Wunused-parameter]
> - Fixed warning: unused parameter ‘si’ [-Wunused-parameter]
> - Fixed warning: unused parameter ‘uc’ [-Wunused-parameter]
> ---
>  runtest/hugetlb                               |   1 +
>  testcases/kernel/mem/.gitignore               |   1 +
>  .../kernel/mem/hugetlb/hugemmap/hugemmap38.c  | 122 ++++++++++++++++++
>  3 files changed, 124 insertions(+)
>  create mode 100644 testcases/kernel/mem/hugetlb/hugemmap/hugemmap38.c
> 
> diff --git a/runtest/hugetlb b/runtest/hugetlb
> index 299c07ac9..2ffd35aec 100644
> --- a/runtest/hugetlb
> +++ b/runtest/hugetlb
> @@ -35,6 +35,7 @@ hugemmap29 hugemmap29
>  hugemmap30 hugemmap30
>  hugemmap31 hugemmap31
>  hugemmap32 hugemmap32
> +hugemmap38 hugemmap38
>  hugemmap05_1 hugemmap05 -m
>  hugemmap05_2 hugemmap05 -s
>  hugemmap05_3 hugemmap05 -s -m
> diff --git a/testcases/kernel/mem/.gitignore b/testcases/kernel/mem/.gitignore
> index c96fe8bfc..dd0c59e98 100644
> --- a/testcases/kernel/mem/.gitignore
> +++ b/testcases/kernel/mem/.gitignore
> @@ -34,6 +34,7 @@
>  /hugetlb/hugemmap/hugemmap30
>  /hugetlb/hugemmap/hugemmap31
>  /hugetlb/hugemmap/hugemmap32
> +/hugetlb/hugemmap/hugemmap38
>  /hugetlb/hugeshmat/hugeshmat01
>  /hugetlb/hugeshmat/hugeshmat02
>  /hugetlb/hugeshmat/hugeshmat03
> diff --git a/testcases/kernel/mem/hugetlb/hugemmap/hugemmap38.c b/testcases/kernel/mem/hugetlb/hugemmap/hugemmap38.c
> new file mode 100644
> index 000000000..20845b946
> --- /dev/null
> +++ b/testcases/kernel/mem/hugetlb/hugemmap/hugemmap38.c
> @@ -0,0 +1,122 @@
> +// SPDX-License-Identifier: LGPL-2.1-or-later
> +/*
> + * Copyright (C) 2005-2006 David Gibson & Adam Litke, IBM Corporation.
> + */
> +
> +/*\
> + *[Descripiton]
> + *
> + * Origin: https://github.com/libhugetlbfs/libhugetlbfs/blob/master/tests/truncate_reserve_wraparound.c
> + *
> + * At one stage, improper handling of tests against i_size could mess
> + * up accounting of reserved hugepages on certain truncate
> + * operations.
> + *
> + */
> +
> +#include <signal.h>
> +#include <setjmp.h>
> +#include "hugetlb.h"
> +
> +#define MNTPOINT "hugetlbfs/"
> +
> +static long hpage_size;
> +static int fd = -1;

> +static int sigbus_count;
> +static unsigned long initial_rsvd, after_map_rsvd, after_touch_rsvd;
> +static unsigned long after_trunc_rsvd, after_unmap_rsvd, after_sigbus_rsvd;
> +static volatile unsigned int *q;

All these can be defined in the run_test() function instead, no need to
keep them global.

> +static sigjmp_buf sig_escape;
> +
> +static void sigbus_handler(int signum LTP_ATTRIBUTE_UNUSED, siginfo_t *si LTP_ATTRIBUTE_UNUSED, void *uc LTP_ATTRIBUTE_UNUSED)
> +{
> +	siglongjmp(sig_escape, 17);
> +}
> +
> +static void run_test(void)
> +{
> +	void *p;
> +	struct sigaction sa = {
> +		.sa_sigaction = sigbus_handler,
> +		.sa_flags = SA_SIGINFO,
> +	};

If I remmeber correctly I asked this to be moved into the test setup()
along with the corresponding SAFE_SIGACTION() call.

Tere is no point in setting up the signal handler in each iteration of
the test.

> +	sigbus_count = 0;
> +
> +	initial_rsvd = SAFE_READ_MEMINFO(MEMINFO_HPAGE_RSVD);
> +	tst_res(TINFO, "Reserve count before map: %lu", initial_rsvd);
> +
> +	p = SAFE_MMAP(NULL, hpage_size, PROT_READ|PROT_WRITE, MAP_SHARED,
> +			fd, 0);
> +	q = p;
> +
> +	after_map_rsvd = SAFE_READ_MEMINFO(MEMINFO_HPAGE_RSVD);
> +	tst_res(TINFO, "Reserve count after map: %lu", after_map_rsvd);
> +
> +	*q = 0;
> +	after_touch_rsvd = SAFE_READ_MEMINFO(MEMINFO_HPAGE_RSVD);
> +	tst_res(TINFO, "Reserve count after touch: %lu", after_touch_rsvd);

So if I understand this correctly the after_touch_rsvd should be back to
initial_rsvd because since we faulted the hugepage it's no longer
reserved. Shouldn't we check that after_touch_rsvd == initial_rsvd as
well?

> +	SAFE_FTRUNCATE(fd, 0);
> +	after_trunc_rsvd = SAFE_READ_MEMINFO(MEMINFO_HPAGE_RSVD);
> +	tst_res(TINFO, "Reserve count after truncate: %lu", after_trunc_rsvd);
> +
> +	if (after_trunc_rsvd != initial_rsvd) {
> +		tst_res(TFAIL, "Reserved count is not restored after truncate: %lu instead of %lu", after_trunc_rsvd, initial_rsvd);
                                             ^
					     I would omit the "count is not restored"

So that the message would look like "Reserved after truncate %lu instead of %lu"

That would make the message shorter and more to the point while keeping
most of the information in there.

> +		goto windup;
> +	}
> +
> +	SAFE_SIGACTION(SIGBUS, &sa, NULL);
> +	if (sigsetjmp(sig_escape, 1) == 0)
> +		*q; /* Fault, triggering a SIGBUS */
> +	else
> +		sigbus_count++;
> +
> +	if (sigbus_count != 1) {
> +		tst_res(TFAIL, "Didn't SIGBUS after truncate");
> +		goto windup;
> +	}
> +
> +	after_sigbus_rsvd = SAFE_READ_MEMINFO(MEMINFO_HPAGE_RSVD);
> +	tst_res(TINFO, "Reserve count after sigbus: %lu", after_sigbus_rsvd);
> +
> +	if (after_trunc_rsvd != initial_rsvd) {
             ^
	     Shouldn't this be after_sigbus_rsvd?

> +		tst_res(TFAIL, "Reserved count is altered by SIGBUS fault: %lu instead of %lu", after_sigbus_rsvd, initial_rsvd);
                                                 ^
						 And changed this to:

						 "Reserved after SIGBUS %lu instead of %lu"


All in all these are really minor and the code looks mostly fine now.

-- 
Cyril Hrubis
chrubis@suse.cz