linux-kernel - [PATCH 1/3] selftests: vm: check numa_available() before operating "merge_across_nodes" in ksm

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-Id: <20220521083825.319654-2-patrick.wang.shcn@gmail.com>
Date:   Sat, 21 May 2022 16:38:23 +0800
From:   Patrick Wang <patrick.wang.shcn@...il.com>
To:     akpm@...ux-foundation.org, shuah@...nel.org
Cc:     linux-mm@...ck.org, linux-kselftest@...r.kernel.org,
        linux-kernel@...r.kernel.org, patrick.wang.shcn@...il.com
Subject: [PATCH 1/3] selftests: vm: check numa_available() before operating "merge_across_nodes" in ksm_tests

Currently, ksm_tests operates "merge_across_nodes" with NUMA either
enabled or disabled. In a system with NUMA disabled, these operations
will fail and output a misleading report given "merge_across_nodes" does
not exist in sysfs:

  ----------------------------
  running ./ksm_tests -M -p 10
  ----------------------------
  f /sys/kernel/mm/ksm/merge_across_nodes
  fopen: No such file or directory
  Cannot save default tunables
  [FAIL]
  ----------------------

So check numa_available() before those operations to skip them if NUMA
is disabled.

Signed-off-by: Patrick Wang <patrick.wang.shcn@...il.com>
---
 tools/testing/selftests/vm/ksm_tests.c | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

diff --git a/tools/testing/selftests/vm/ksm_tests.c b/tools/testing/selftests/vm/ksm_tests.c
index fd85f15869d1..2fcf24312da8 100644
--- a/tools/testing/selftests/vm/ksm_tests.c
+++ b/tools/testing/selftests/vm/ksm_tests.c
@@ -221,7 +221,8 @@ static bool assert_ksm_pages_count(long dupl_page_count)
 static int ksm_save_def(struct ksm_sysfs *ksm_sysfs)
 {
 	if (ksm_read_sysfs(KSM_FP("max_page_sharing"), &ksm_sysfs->max_page_sharing) ||
-	    ksm_read_sysfs(KSM_FP("merge_across_nodes"), &ksm_sysfs->merge_across_nodes) ||
+	    numa_available() ? 0 :
+		ksm_read_sysfs(KSM_FP("merge_across_nodes"), &ksm_sysfs->merge_across_nodes) ||
 	    ksm_read_sysfs(KSM_FP("sleep_millisecs"), &ksm_sysfs->sleep_millisecs) ||
 	    ksm_read_sysfs(KSM_FP("pages_to_scan"), &ksm_sysfs->pages_to_scan) ||
 	    ksm_read_sysfs(KSM_FP("run"), &ksm_sysfs->run) ||
@@ -236,7 +237,8 @@ static int ksm_save_def(struct ksm_sysfs *ksm_sysfs)
 static int ksm_restore(struct ksm_sysfs *ksm_sysfs)
 {
 	if (ksm_write_sysfs(KSM_FP("max_page_sharing"), ksm_sysfs->max_page_sharing) ||
-	    ksm_write_sysfs(KSM_FP("merge_across_nodes"), ksm_sysfs->merge_across_nodes) ||
+	    numa_available() ? 0 :
+		ksm_write_sysfs(KSM_FP("merge_across_nodes"), ksm_sysfs->merge_across_nodes) ||
 	    ksm_write_sysfs(KSM_FP("pages_to_scan"), ksm_sysfs->pages_to_scan) ||
 	    ksm_write_sysfs(KSM_FP("run"), ksm_sysfs->run) ||
 	    ksm_write_sysfs(KSM_FP("sleep_millisecs"), ksm_sysfs->sleep_millisecs) ||
@@ -720,7 +722,8 @@ int main(int argc, char *argv[])
 
 	if (ksm_write_sysfs(KSM_FP("run"), 2) ||
 	    ksm_write_sysfs(KSM_FP("sleep_millisecs"), 0) ||
-	    ksm_write_sysfs(KSM_FP("merge_across_nodes"), 1) ||
+	    numa_available() ? 0 :
+		ksm_write_sysfs(KSM_FP("merge_across_nodes"), 1) ||
 	    ksm_write_sysfs(KSM_FP("pages_to_scan"), page_count))
 		return KSFT_FAIL;
 
-- 
2.25.1