All of lore.kernel.org
 help / color / mirror / Atom feed
* [LTP] [PATCH v2] lib: Do not fail a test if oom score cannot be adjusted.
@ 2021-12-22 13:52 Cyril Hrubis
  2021-12-22 14:03 ` Petr Vorel
  0 siblings, 1 reply; 8+ messages in thread
From: Cyril Hrubis @ 2021-12-22 13:52 UTC (permalink / raw)
  To: ltp

From: Petr Vorel <pvorel@suse.cz>

Setting value < 0 in /proc/*/oom_score_adj requires CAP_SYS_RESOURCE or
CAP_SYS_ADMIN. However setting the library process score is a best
effort operation, so let's skip it silently when the user is not
privileged to do so.

Fixes: 8a0827766d ("lib: add functions to adjust oom score")
Signed-off-by: Petr Vorel <pvorel@suse.cz>
Signed-off-by: Cyril Hrubis <chrubis@suse.cz>
---
 include/tst_memutils.h |  6 ++++-
 lib/tst_memutils.c     | 55 ++++++++++++++++++++++++++++++++++++------
 2 files changed, 52 insertions(+), 9 deletions(-)

diff --git a/include/tst_memutils.h b/include/tst_memutils.h
index 68a6e3771..855c6f289 100644
--- a/include/tst_memutils.h
+++ b/include/tst_memutils.h
@@ -28,13 +28,17 @@ long long tst_available_mem(void);
 /*
  * Enable OOM protection to prevent process($PID) being killed by OOM Killer.
  *   echo -1000 >/proc/$PID/oom_score_adj
+ *
  * If the pid is 0 which means it will set on current(self) process.
  *
+ *  Unless the process has CAP_SYS_RESOURCE or CAP_SYS_ADMIN this call will be
+ *  no-op because setting adj value < 0 requires it.
+ *
  * Note:
  *  This exported tst_enable_oom_protection function can be used at anywhere
  *  you want to protect, but please remember that if you do enable protection
  *  on a process($PID) that all the children will inherit its score and be
- *  ignored by OOM Killer as well. So that's why tst_disable_oom_protection
+ *  ignored by OOM Killer as well. So that's why tst_disable_oom_protection()
  *  to be used in combination.
  */
 void tst_enable_oom_protection(pid_t pid);
diff --git a/lib/tst_memutils.c b/lib/tst_memutils.c
index 4dea30330..4a47bbb33 100644
--- a/lib/tst_memutils.c
+++ b/lib/tst_memutils.c
@@ -11,6 +11,8 @@
 
 #define TST_NO_DEFAULT_MAIN
 #include "tst_test.h"
+#include "tst_capability.h"
+#include "lapi/syscalls.h"
 
 #define BLOCKSIZE (16 * 1024 * 1024)
 
@@ -93,6 +95,42 @@ long long tst_available_mem(void)
 	return mem_available;
 }
 
+static int has_caps(void)
+{
+	struct tst_cap_user_header hdr = {
+		.version = 0x20080522,
+		.pid = tst_syscall(__NR_gettid),
+	};
+
+	struct tst_cap_user_data caps[2];
+
+	if (tst_capget(&hdr, caps))
+		tst_brk(TBROK | TERRNO, "tst_capget()");
+
+	if ((caps[0].effective & (1U << CAP_SYS_ADMIN)) ||
+	    (caps[0].effective & (1U << CAP_SYS_RESOURCE)))
+		return 1;
+
+	return 0;
+}
+
+static int write_score(const char *path, int score)
+{
+	FILE *f;
+
+	f = fopen(path, "w");
+	if (!f)
+		return 1;
+
+	if (fprintf(f, "%d", score) <= 0)
+		return 1;
+
+	if (fclose(f))
+		return 1;
+
+	return 0;
+}
+
 static void set_oom_score_adj(pid_t pid, int value)
 {
 	int val;
@@ -111,17 +149,18 @@ static void set_oom_score_adj(pid_t pid, int value)
 			tst_brk(TBROK, "%s does not exist, please check if PID is valid", score_path);
 	}
 
-	FILE_PRINTF(score_path, "%d", value);
+	if (write_score(score_path, value)) {
+		if (!has_caps())
+			return;
+
+		tst_res(TWARN, "Can't adjust score, even with capabilities!?");
+		return;
+	}
+
 	FILE_SCANF(score_path, "%d", &val);
 
-	if (val != value) {
-		if (value < 0) {
-			tst_res(TWARN, "'%s' cannot be set to %i, are you root?",
-				score_path, value);
-			return;
-		}
+	if (val != value)
 		tst_brk(TBROK, "oom_score_adj = %d, but expect %d.", val, value);
-	}
 }
 
 void tst_enable_oom_protection(pid_t pid)
-- 
2.32.0


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

^ permalink raw reply related	[flat|nested] 8+ messages in thread

* Re: [LTP] [PATCH v2] lib: Do not fail a test if oom score cannot be adjusted.
  2021-12-22 13:52 [LTP] [PATCH v2] lib: Do not fail a test if oom score cannot be adjusted Cyril Hrubis
@ 2021-12-22 14:03 ` Petr Vorel
  2021-12-23  1:17   ` xuyang2018.jy
  0 siblings, 1 reply; 8+ messages in thread
From: Petr Vorel @ 2021-12-22 14:03 UTC (permalink / raw)
  To: Cyril Hrubis; +Cc: ltp

Hi Cyril, all,

> From: Petr Vorel <pvorel@suse.cz>

> Setting value < 0 in /proc/*/oom_score_adj requires CAP_SYS_RESOURCE or
> CAP_SYS_ADMIN. However setting the library process score is a best
> effort operation, so let's skip it silently when the user is not
> privileged to do so.
+1

LGTM, thanks for this version, I'm for merging it.

Also tested locally and on CI, working as expected.
https://github.com/pevik/ltp/runs/4607602484?check_suite_focus=true

Kind regards,
Petr

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

^ permalink raw reply	[flat|nested] 8+ messages in thread

* Re: [LTP] [PATCH v2] lib: Do not fail a test if oom score cannot be adjusted.
  2021-12-22 14:03 ` Petr Vorel
@ 2021-12-23  1:17   ` xuyang2018.jy
  2021-12-23  2:13     ` Li Wang
  0 siblings, 1 reply; 8+ messages in thread
From: xuyang2018.jy @ 2021-12-23  1:17 UTC (permalink / raw)
  To: Petr Vorel; +Cc: ltp

Hi Petr, All
> Hi Cyril, all,
>
>> From: Petr Vorel<pvorel@suse.cz>
>
>> Setting value<  0 in /proc/*/oom_score_adj requires CAP_SYS_RESOURCE or
>> CAP_SYS_ADMIN.
Here will mislead user.
Since the default oom_score_adj value is 0, so we can not set a value < 0.

The value of /proc/<pid>/oom_score_adj may be reduced no lower than the 
last value set by a CAP_SYS_RESOURCE process. To reduce the value any 
lower requires CAP_SYS_RESOURCE.

Also looks man 7 capabilities, CAP_SYS_ADMIN doesn't have this cap and I 
have do a test to verify this

#include <stdio.h>
#include <unistd.h>
#include <sys/types.h>
#include <errno.h>
#include "tst_test.h"
#include "tst_capability.h"

static void run(void)
{
         FILE *f;

         f = fopen("/proc/self/oom_score_adj", "w");
         if (!f)
                 tst_res(TFAIL, "non-exist");

         if (fprintf(f, "%d", -200) <= 0)
                 tst_res(TFAIL, "write failed");

         if (fclose(f))
                 tst_res(TFAIL, "close %d",errno);

         tst_res(TPASS, "write pass");
}

static struct tst_test test = {
         .test_all = run,
         .caps = (struct tst_cap []) {
                 TST_CAP(TST_CAP_REQ, CAP_SYS_ADMIN),
                 TST_CAP(TST_CAP_DROP, CAP_SYS_RESOURCE),
                 {}
         },
};

It still fails if we only have CAP_SYS_ADMIN.

Best Regards
Yang Xu

  However setting the library process score is a best
>> effort operation, so let's skip it silently when the user is not
>> privileged to do so.
> +1
>
> LGTM, thanks for this version, I'm for merging it.
>
> Also tested locally and on CI, working as expected.
> https://github.com/pevik/ltp/runs/4607602484?check_suite_focus=true
>
> Kind regards,
> Petr

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

^ permalink raw reply	[flat|nested] 8+ messages in thread

* Re: [LTP] [PATCH v2] lib: Do not fail a test if oom score cannot be adjusted.
  2021-12-23  1:17   ` xuyang2018.jy
@ 2021-12-23  2:13     ` Li Wang
  2021-12-23  2:48       ` [LTP] [PATCH v3] " Li Wang
  0 siblings, 1 reply; 8+ messages in thread
From: Li Wang @ 2021-12-23  2:13 UTC (permalink / raw)
  To: xuyang2018.jy; +Cc: ltp

 Hi All,

On Thu, Dec 23, 2021 at 9:17 AM xuyang2018.jy@fujitsu.com
<xuyang2018.jy@fujitsu.com> wrote:

> >> Setting value<  0 in /proc/*/oom_score_adj requires CAP_SYS_RESOURCE or
> >> CAP_SYS_ADMIN.
> Here will mislead user.
> Since the default oom_score_adj value is 0, so we can not set a value < 0.
>
> The value of /proc/<pid>/oom_score_adj may be reduced no lower than the
> last value set by a CAP_SYS_RESOURCE process. To reduce the value any
> lower requires CAP_SYS_RESOURCE.

Good to know this, thank you Xu!

From the man page, CAP_SYS_RESOURCE gives:

"
        CAP_SYS_RESOURCE
              ...
              * set /proc/[pid]/oom_score_adj to a value lower than
the value last set by a process with CAP_SYS_RESOURCE.
"

I will send a V3 base on Petr and Cyril's work to speed up merging this fix.

-- 
Regards,
Li Wang


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

^ permalink raw reply	[flat|nested] 8+ messages in thread

* [LTP] [PATCH v3] lib: Do not fail a test if oom score cannot be adjusted.
  2021-12-23  2:13     ` Li Wang
@ 2021-12-23  2:48       ` Li Wang
  2021-12-23  3:03         ` xuyang2018.jy
  2021-12-23  7:57         ` Petr Vorel
  0 siblings, 2 replies; 8+ messages in thread
From: Li Wang @ 2021-12-23  2:48 UTC (permalink / raw)
  To: ltp

From: Petr Vorel <pvorel@suse.cz>

Setting value < 0 in /proc/<pid>/oom_score_adj requires CAP_SYS_RESOURCE.
However setting the library process score is a best effort operation,
so let's skip it silently when the user is not privileged to do so.

Fixes: 8a0827766d ("lib: add functions to adjust oom score")
Signed-off-by: Petr Vorel <pvorel@suse.cz>
Signed-off-by: Cyril Hrubis <chrubis@suse.cz>
Signed-off-by: Li Wang <liwang@redhat.com>
---

Notes:
    v2 --> v3
    	* drop the useless CAP_SYS_ADMIN
    	* add some new commments
    CI job: https://github.com/wangli5665/ltp/actions/runs/1613852940

 include/tst_memutils.h | 10 +++++++-
 lib/tst_memutils.c     | 54 +++++++++++++++++++++++++++++++++++-------
 2 files changed, 55 insertions(+), 9 deletions(-)

diff --git a/include/tst_memutils.h b/include/tst_memutils.h
index 68a6e3771..45dec55bc 100644
--- a/include/tst_memutils.h
+++ b/include/tst_memutils.h
@@ -28,13 +28,21 @@ long long tst_available_mem(void);
 /*
  * Enable OOM protection to prevent process($PID) being killed by OOM Killer.
  *   echo -1000 >/proc/$PID/oom_score_adj
+ *
  * If the pid is 0 which means it will set on current(self) process.
  *
+ * Unless the process has CAP_SYS_RESOURCE this call will be no-op because
+ * setting adj value < 0 requires it.
+ *
+ * CAP_SYS_RESOURCE:
+ *   set /proc/[pid]/oom_score_adj to a value lower than the value last set
+ *   by a process with CAP_SYS_RESOURCE.
+ *
  * Note:
  *  This exported tst_enable_oom_protection function can be used at anywhere
  *  you want to protect, but please remember that if you do enable protection
  *  on a process($PID) that all the children will inherit its score and be
- *  ignored by OOM Killer as well. So that's why tst_disable_oom_protection
+ *  ignored by OOM Killer as well. So that's why tst_disable_oom_protection()
  *  to be used in combination.
  */
 void tst_enable_oom_protection(pid_t pid);
diff --git a/lib/tst_memutils.c b/lib/tst_memutils.c
index 4dea30330..4a4974761 100644
--- a/lib/tst_memutils.c
+++ b/lib/tst_memutils.c
@@ -11,6 +11,8 @@
 
 #define TST_NO_DEFAULT_MAIN
 #include "tst_test.h"
+#include "tst_capability.h"
+#include "lapi/syscalls.h"
 
 #define BLOCKSIZE (16 * 1024 * 1024)
 
@@ -93,6 +95,41 @@ long long tst_available_mem(void)
 	return mem_available;
 }
 
+static int has_caps(void)
+{
+	struct tst_cap_user_header hdr = {
+		.version = 0x20080522,
+		.pid = tst_syscall(__NR_gettid),
+	};
+
+	struct tst_cap_user_data caps[2];
+
+	if (tst_capget(&hdr, caps))
+		tst_brk(TBROK | TERRNO, "tst_capget()");
+
+	if (caps[0].effective & (1U << CAP_SYS_RESOURCE))
+		return 1;
+
+	return 0;
+}
+
+static int write_score(const char *path, int score)
+{
+	FILE *f;
+
+	f = fopen(path, "w");
+	if (!f)
+		return 1;
+
+	if (fprintf(f, "%d", score) <= 0)
+		return 1;
+
+	if (fclose(f))
+		return 1;
+
+	return 0;
+}
+
 static void set_oom_score_adj(pid_t pid, int value)
 {
 	int val;
@@ -111,17 +148,18 @@ static void set_oom_score_adj(pid_t pid, int value)
 			tst_brk(TBROK, "%s does not exist, please check if PID is valid", score_path);
 	}
 
-	FILE_PRINTF(score_path, "%d", value);
+	if (write_score(score_path, value)) {
+		if (!has_caps())
+			return;
+
+		tst_res(TWARN, "Can't adjust score, even with capabilities!?");
+		return;
+	}
+
 	FILE_SCANF(score_path, "%d", &val);
 
-	if (val != value) {
-		if (value < 0) {
-			tst_res(TWARN, "'%s' cannot be set to %i, are you root?",
-				score_path, value);
-			return;
-		}
+	if (val != value)
 		tst_brk(TBROK, "oom_score_adj = %d, but expect %d.", val, value);
-	}
 }
 
 void tst_enable_oom_protection(pid_t pid)
-- 
2.31.1


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

^ permalink raw reply related	[flat|nested] 8+ messages in thread

* Re: [LTP] [PATCH v3] lib: Do not fail a test if oom score cannot be adjusted.
  2021-12-23  2:48       ` [LTP] [PATCH v3] " Li Wang
@ 2021-12-23  3:03         ` xuyang2018.jy
  2021-12-23  7:57         ` Petr Vorel
  1 sibling, 0 replies; 8+ messages in thread
From: xuyang2018.jy @ 2021-12-23  3:03 UTC (permalink / raw)
  To: Li Wang; +Cc: ltp

Hi Li
> From: Petr Vorel<pvorel@suse.cz>
>
> Setting value<  0 in/proc/<pid>/oom_score_adj requires CAP_SYS_RESOURCE.
> However setting the library process score is a best effort operation,
> so let's skip it silently when the user is not privileged to do so.
>
> Fixes: 8a0827766d ("lib: add functions to adjust oom score")
> Signed-off-by: Petr Vorel<pvorel@suse.cz>
> Signed-off-by: Cyril Hrubis<chrubis@suse.cz>
> Signed-off-by: Li Wang<liwang@redhat.com>
Now, LGTM,
Reviewed-by: Yang Xu <xuyang2018.jy@fujitsu.com>

ps: ubuntun:groxy fails because of other thing..., it is another problem.

Best Regards
Yang Xu
> ---
>
> Notes:
>      v2 -->  v3
>      	* drop the useless CAP_SYS_ADMIN
>      	* add some new commments
>      CI job:https://github.com/wangli5665/ltp/actions/runs/1613852940
>

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

^ permalink raw reply	[flat|nested] 8+ messages in thread

* Re: [LTP] [PATCH v3] lib: Do not fail a test if oom score cannot be adjusted.
  2021-12-23  2:48       ` [LTP] [PATCH v3] " Li Wang
  2021-12-23  3:03         ` xuyang2018.jy
@ 2021-12-23  7:57         ` Petr Vorel
  2021-12-23  8:29           ` Petr Vorel
  1 sibling, 1 reply; 8+ messages in thread
From: Petr Vorel @ 2021-12-23  7:57 UTC (permalink / raw)
  To: Li Wang; +Cc: ltp

Hi all,

> Notes:
>     v2 --> v3
>     	* drop the useless CAP_SYS_ADMIN
>     	* add some new commments
Yes, CAP_SYS_ADMIN was redundant, thus I merged it.
Thanks a lot all.

>     CI job: https://github.com/wangli5665/ltp/actions/runs/1613852940

Yes, ubuntu cannot reach repos:

    Ign:1 http://security.ubuntu.com/ubuntu groovy-security InRelease
    Ign:2 http://archive.ubuntu.com/ubuntu groovy InRelease
    Err:3 http://security.ubuntu.com/ubuntu groovy-security Release
      404  Not Found [IP: 91.189.91.39 80]
    Ign:4 http://archive.ubuntu.com/ubuntu groovy-updates InRelease
    Ign:5 http://archive.ubuntu.com/ubuntu groovy-backports InRelease
    Err:6 http://archive.ubuntu.com/ubuntu groovy Release
      404  Not Found [IP: 91.189.88.142 80]
    Err:7 http://archive.ubuntu.com/ubuntu groovy-updates Release
      404  Not Found [IP: 91.189.88.142 80]
    Err:8 http://archive.ubuntu.com/ubuntu groovy-backports Release
      404  Not Found [IP: 91.189.88.142 80]
    Reading package lists...
    E: The repository 'http://security.ubuntu.com/ubuntu groovy-security Release' does not have a Release file.
    E: The repository 'http://archive.ubuntu.com/ubuntu groovy Release' does not have a Release file.
    E: The repository 'http://archive.ubuntu.com/ubuntu groovy-updates Release' does not have a Release file.
    E: The repository 'http://archive.ubuntu.com/ubuntu groovy-backports Release' does not have a Release file.

I'm testing upgrade to impish:

https://github.com/pevik/ltp/actions/runs/1614667613

Kind regards,
Petr

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

^ permalink raw reply	[flat|nested] 8+ messages in thread

* Re: [LTP] [PATCH v3] lib: Do not fail a test if oom score cannot be adjusted.
  2021-12-23  7:57         ` Petr Vorel
@ 2021-12-23  8:29           ` Petr Vorel
  0 siblings, 0 replies; 8+ messages in thread
From: Petr Vorel @ 2021-12-23  8:29 UTC (permalink / raw)
  To: Li Wang, ltp

Hi all,

> I'm testing upgrade to impish:

> https://github.com/pevik/ltp/actions/runs/1614667613

Hm, impish is also broken. Hope it's just temporary issue.

Kind regards,
Petr

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

^ permalink raw reply	[flat|nested] 8+ messages in thread

end of thread, other threads:[~2021-12-23  8:29 UTC | newest]

Thread overview: 8+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-12-22 13:52 [LTP] [PATCH v2] lib: Do not fail a test if oom score cannot be adjusted Cyril Hrubis
2021-12-22 14:03 ` Petr Vorel
2021-12-23  1:17   ` xuyang2018.jy
2021-12-23  2:13     ` Li Wang
2021-12-23  2:48       ` [LTP] [PATCH v3] " Li Wang
2021-12-23  3:03         ` xuyang2018.jy
2021-12-23  7:57         ` Petr Vorel
2021-12-23  8:29           ` Petr Vorel

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.