[LTP] [PATCH] getrusage03: mlock the memory area in consume_mb

[LTP] [PATCH] getrusage03: mlock the memory area in consume_mb

[Li Wang]

Seems it trying to test without swap happen to guarantee less
disturbing for the ’->ru_maxrss‘ counting.

Therefore add mlock() to prevent that memory allocated by consume_mb
from being paged to the swap area.

Signed-off-by: Li Wang <liwang@redhat.com>
---
 testcases/kernel/syscalls/getrusage/getrusage03.h | 1 +
 1 file changed, 1 insertion(+)

diff --git a/testcases/kernel/syscalls/getrusage/getrusage03.h b/testcases/kernel/syscalls/getrusage/getrusage03.h
index f1bbe9be5..d46fdff85 100644
--- a/testcases/kernel/syscalls/getrusage/getrusage03.h
+++ b/testcases/kernel/syscalls/getrusage/getrusage03.h
@@ -18,6 +18,7 @@ static void consume_mb(int consume_nr)
 
 	size = consume_nr * 1024 * 1024;
 	ptr = SAFE_MALLOC(size);
+	mlock(ptr, size);
 	memset(ptr, 0, size);
 
 	SAFE_FILE_LINES_SCANF("/proc/self/status", "VmSwap: %lu", &vmswap_size);
-- 
2.31.1


-- 
Mailing list info: https://lists.linux.it/listinfo/ltp

Re: [LTP] [PATCH] getrusage03: mlock the memory area in consume_mb

[Li Wang]

[-- Attachment #1.1: Type: text/plain, Size: 1357 bytes --]

On Tue, Feb 8, 2022 at 9:24 PM Li Wang <liwang@redhat.com> wrote:

> Seems it trying to test without swap happen to guarantee less
> disturbing for the ’->ru_maxrss‘ counting.
>
> Therefore add mlock() to prevent that memory allocated by consume_mb
> from being paged to the swap area.
>
> Signed-off-by: Li Wang <liwang@redhat.com>
> ---
>  testcases/kernel/syscalls/getrusage/getrusage03.h | 1 +
>  1 file changed, 1 insertion(+)
>
> diff --git a/testcases/kernel/syscalls/getrusage/getrusage03.h
> b/testcases/kernel/syscalls/getrusage/getrusage03.h
> index f1bbe9be5..d46fdff85 100644
> --- a/testcases/kernel/syscalls/getrusage/getrusage03.h
> +++ b/testcases/kernel/syscalls/getrusage/getrusage03.h
> @@ -18,6 +18,7 @@ static void consume_mb(int consume_nr)
>
>         size = consume_nr * 1024 * 1024;
>         ptr = SAFE_MALLOC(size);
> +       mlock(ptr, size);
>

Hmm, seems using mlockall() will be better than mlock().

Because that locks all pages mapped into the address space of the calling
process.
(includes code, data, stack segment, shared libraries, etc).



>         memset(ptr, 0, size);
>
>         SAFE_FILE_LINES_SCANF("/proc/self/status", "VmSwap: %lu",
> &vmswap_size);
> --
> 2.31.1
>
>
> --
> Mailing list info: https://lists.linux.it/listinfo/ltp
>


-- 
Regards,
Li Wang

[-- Attachment #1.2: Type: text/html, Size: 2480 bytes --]

[-- Attachment #2: Type: text/plain, Size: 60 bytes --]


-- 
Mailing list info: https://lists.linux.it/listinfo/ltp

[LTP] [PATCH v2] getrusage03: mlock all address space for process calling consume_mb

[Li Wang]

It is trying to test without swap happen to guarantee less disturbing
for the ’->ru_maxrss‘ counting.

Therefore add mlockall() to prevent that memory address space
of the calling process from being paged to the swap area.

Due to that lock needs CAP_IPC_LOCK capability and child can not
inherit memory locks from parent, here add the definition and put
the mlockall() in consume_mb.

Signed-off-by: Li Wang <liwang@redhat.com>
---
 include/lapi/capability.h                         | 4 ++++
 testcases/kernel/syscalls/getrusage/getrusage03.c | 4 ++++
 testcases/kernel/syscalls/getrusage/getrusage03.h | 2 ++
 3 files changed, 10 insertions(+)

diff --git a/include/lapi/capability.h b/include/lapi/capability.h
index 8cabd0f28..17ec107b1 100644
--- a/include/lapi/capability.h
+++ b/include/lapi/capability.h
@@ -24,6 +24,10 @@
 # define CAP_NET_RAW          13
 #endif
 
+#ifndef CAP_IPC_LOCK
+# define CAP_IPC_LOCK         14
+#endif
+
 #ifndef CAP_SYS_CHROOT
 # define CAP_SYS_CHROOT       18
 #endif
diff --git a/testcases/kernel/syscalls/getrusage/getrusage03.c b/testcases/kernel/syscalls/getrusage/getrusage03.c
index 201d258fa..43fab0605 100644
--- a/testcases/kernel/syscalls/getrusage/getrusage03.c
+++ b/testcases/kernel/syscalls/getrusage/getrusage03.c
@@ -193,4 +193,8 @@ static struct tst_test test = {
 	.setup = setup,
 	.test = run,
 	.tcnt = ARRAY_SIZE(testfunc_list),
+	.caps = (struct tst_cap []) {
+		TST_CAP(TST_CAP_REQ, CAP_IPC_LOCK),
+		{}
+	},
 };
diff --git a/testcases/kernel/syscalls/getrusage/getrusage03.h b/testcases/kernel/syscalls/getrusage/getrusage03.h
index f1bbe9be5..b28b9d4c3 100644
--- a/testcases/kernel/syscalls/getrusage/getrusage03.h
+++ b/testcases/kernel/syscalls/getrusage/getrusage03.h
@@ -16,6 +16,8 @@ static void consume_mb(int consume_nr)
 	size_t size;
 	unsigned long vmswap_size;
 
+	mlockall(MCL_CURRENT|MCL_FUTURE);
+
 	size = consume_nr * 1024 * 1024;
 	ptr = SAFE_MALLOC(size);
 	memset(ptr, 0, size);
-- 
2.31.1


-- 
Mailing list info: https://lists.linux.it/listinfo/ltp

Re: [LTP] [PATCH v2] getrusage03: mlock all address space for process calling consume_mb

[Cyril Hrubis]

Hi!
> --- a/testcases/kernel/syscalls/getrusage/getrusage03.c
> +++ b/testcases/kernel/syscalls/getrusage/getrusage03.c
> @@ -193,4 +193,8 @@ static struct tst_test test = {
>  	.setup = setup,
>  	.test = run,
>  	.tcnt = ARRAY_SIZE(testfunc_list),
> +	.caps = (struct tst_cap []) {
> +		TST_CAP(TST_CAP_REQ, CAP_IPC_LOCK),
> +		{}
> +	},
>  };

I wonder if we should start to track the capabilities like this instead
of the .needs_root approach. I guess that it may be useful to keep
things granular in the future.

-- 
Cyril Hrubis
chrubis@suse.cz

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp

Re: [LTP] [PATCH v2] getrusage03: mlock all address space for process calling consume_mb

[Li Wang]

[-- Attachment #1.1: Type: text/plain, Size: 1007 bytes --]

On Wed, Feb 16, 2022 at 5:54 PM Cyril Hrubis <chrubis@suse.cz> wrote:

> Hi!
> > --- a/testcases/kernel/syscalls/getrusage/getrusage03.c
> > +++ b/testcases/kernel/syscalls/getrusage/getrusage03.c
> > @@ -193,4 +193,8 @@ static struct tst_test test = {
> >       .setup = setup,
> >       .test = run,
> >       .tcnt = ARRAY_SIZE(testfunc_list),
> > +     .caps = (struct tst_cap []) {
> > +             TST_CAP(TST_CAP_REQ, CAP_IPC_LOCK),
> > +             {}
> > +     },
> >  };
>
> I wonder if we should start to track the capabilities like this instead
> of the .needs_root approach. I guess that it may be useful to keep
> things granular in the future.
>

I think yes, Linux Capabilities makes it possible for specific privileges to
the test process, and avoid that .needs_root limited many tests scenario
to skip directly with non-root users. Especially beneficial for running LTP
in the docker with only part of root privileges for reasons like strict
security
verification.

-- 
Regards,
Li Wang

[-- Attachment #1.2: Type: text/html, Size: 1948 bytes --]

[-- Attachment #2: Type: text/plain, Size: 60 bytes --]


-- 
Mailing list info: https://lists.linux.it/listinfo/ltp

Re: [LTP] [PATCH v2] getrusage03: mlock all address space for process calling consume_mb

[Cyril Hrubis]

Hi!
> I think yes, Linux Capabilities makes it possible for specific privileges to
> the test process, and avoid that .needs_root limited many tests scenario
> to skip directly with non-root users. Especially beneficial for running LTP
> in the docker with only part of root privileges for reasons like strict
> security
> verification.

Looks good then:

Reviewed-by: Cyril Hrubis <chrubis@suse.cz>

-- 
Cyril Hrubis
chrubis@suse.cz

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp

Re: [LTP] [PATCH v2] getrusage03: mlock all address space for process calling consume_mb

[Li Wang]

[-- Attachment #1.1: Type: text/plain, Size: 38 bytes --]

Pushed, thanks.

-- 
Regards,
Li Wang

[-- Attachment #1.2: Type: text/html, Size: 309 bytes --]

[-- Attachment #2: Type: text/plain, Size: 60 bytes --]


-- 
Mailing list info: https://lists.linux.it/listinfo/ltp