Skip to contentSkip to conversation

mem/oom: fork may fail before test

PATCH New Jan Stancek 2 participants 3 messages May 19, 2016 – May 23, 2016

lore
Share & tools
lore

Jan Stancek

author
#0 lore raw

PATCH 1/1 mem/oom: fork may fail before test

On systems with small RAM and vm.overcommit_memory == 2, the system may already be in state where Committed_AS is larger than CommitLimit due to various backround daemons (NM, tuned, polkitd,...) asking for big allocations.

In such instances oom testcases report TBROK on very first fork. This patch allows this to be a valid outcome (TPASS).

Signed-off-by Jan Stancek <[email protected]>
---
 testcases/kernel/mem/lib/mem.c | 4 ++++
 1 file changed, 4 insertions(+)
testcases/kernel/mem/lib/mem.c +4 -0
index c55a55d57915..3d853a393d28 100644
@@ -127,6 +127,10 @@ void oom(int testcase, int lite, int retcode, int allow_sigkill)
 
 	switch (pid = fork()) {
 	case -1:
+		if (errno == retcode) {
+			tst_resm(TPASS | TERRNO, "fork");
+			return;
+		}
 		tst_brkm(TBROK | TERRNO, cleanup, "fork");
 	case 0:
 		threads = MAX(1, tst_ncpus() - 1);
Signature · 1 line
1.8.3.1
#1 lore raw

PATCH 1/1 mem/oom: fork may fail before test

Hi!

Quoted 7 lines

On systems with small RAM and vm.overcommit_memory == 2, the system may already be in state where Committed_AS is larger than CommitLimit due to various backround daemons (NM, tuned, polkitd,...) asking for big allocations.

In such instances oom testcases report TBROK on very first fork. This patch allows this to be a valid outcome (TPASS).

Looks good to me, acked.

Signature · 2 lines
Cyril Hrubis
[email protected]
author
#2 lore raw

PATCH 1/1 mem/oom: fork may fail before test

----- Original Message -----

Quoted 12 lines

From: "Cyril Hrubis" <[email protected]>
To: "Jan Stancek" <[email protected]>
Cc: [email protected]
Sent: Monday, 23 May, 2016 4:36:11 PM
Subject: Re: [LTP] [PATCH] mem/oom: fork may fail before test

Hi!

Quoted 5 lines

On systems with small RAM and vm.overcommit_memory == 2, the system may already be in state where Committed_AS is larger than CommitLimit due to various backround daemons (NM, tuned, polkitd,...) asking for big allocations.

Quoted 4 lines
Quoted 2 lines

In such instances oom testcases report TBROK on very first fork. This patch allows this to be a valid outcome (TPASS).

Looks good to me, acked.

Pushed.

Regards,
Jan

Quoted 5 lines

--
Cyril Hrubis
[email protected]