Skip to contentSkip to conversation

Hugetlb: Migrating libhugetlbfs shm-perms

PATCH Superseded Spoorthy 2 participants 2 messages Sep 14, 2023 – Oct 30, 2023

v1 of 3 Superseded Latest is v3 →

lore
Share & tools
lore

Spoorthy

author
#0 lore raw

PATCH 1/1 Hugetlb: Migrating libhugetlbfs shm-perms

Migrating the libhugetlbfs/tests/shm-perms.c test

Test Description: Test shared memory behavior when multiple threads are attached to a segment with different permissions. A segment is created
and children attach read-only to check reservation accounting.

---
 runtest/hugetlb                               |  1 +
 testcases/kernel/mem/.gitignore               |  1 +
 .../mem/hugetlb/hugeshmat/hugeshmat06.c       | 86 +++++++++++++++++++
 3 files changed, 88 insertions(+)
 create mode 100644 testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat06.c
runtest/hugetlb +1 -0
index 299c07ac9..240701b2b 100644
@@ -44,6 +44,7 @@ hugeshmat02 hugeshmat02 -i 5
 hugeshmat03 hugeshmat03 -i 5
 hugeshmat04 hugeshmat04 -i 5
 hugeshmat05 hugeshmat05 -i 5
+hugeshmat06 hugeshmat06
 
 hugeshmctl01 hugeshmctl01 -i 5
 hugeshmctl02 hugeshmctl02 -i 5
testcases/kernel/mem/.gitignore +1 -0
index 7258489ed..9f7fb1e76 100644
@@ -39,6 +39,7 @@
 /hugetlb/hugeshmat/hugeshmat03
 /hugetlb/hugeshmat/hugeshmat04
 /hugetlb/hugeshmat/hugeshmat05
+/hugetlb/hugeshmat/hugeshmat06
 /hugetlb/hugeshmctl/hugeshmctl01
 /hugetlb/hugeshmctl/hugeshmctl02
 /hugetlb/hugeshmctl/hugeshmctl03
testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat06.c new file +86 -0

86 lines not on this page · raw message

Signature · 5 lines
2.39.3


-- 
Mailing list info: https://lists.linux.it/listinfo/ltp
#1 lore raw

Hi!

Quoted 5 lines

Migrating the libhugetlbfs/tests/shm-perms.c test

Test Description: Test shared memory behavior when multiple threads are attached to a segment with different permissions. A segment is created
and children attach read-only to check reservation accounting.

First of all there are style problems in the patchset, please run 'make check' and fix all reported problems.

Quoted 54 lines
---
 runtest/hugetlb                               |  1 +
 testcases/kernel/mem/.gitignore               |  1 +
 .../mem/hugetlb/hugeshmat/hugeshmat06.c       | 86 +++++++++++++++++++
 3 files changed, 88 insertions(+)
 create mode 100644 testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat06.c
runtest/hugetlb +1 -0
index 299c07ac9..240701b2b 100644
@@ -44,6 +44,7 @@ hugeshmat02 hugeshmat02 -i 5
 hugeshmat03 hugeshmat03 -i 5
 hugeshmat04 hugeshmat04 -i 5
 hugeshmat05 hugeshmat05 -i 5
+hugeshmat06 hugeshmat06
 
 hugeshmctl01 hugeshmctl01 -i 5
 hugeshmctl02 hugeshmctl02 -i 5
testcases/kernel/mem/.gitignore +1 -0
index 7258489ed..9f7fb1e76 100644
@@ -39,6 +39,7 @@
 /hugetlb/hugeshmat/hugeshmat03
 /hugetlb/hugeshmat/hugeshmat04
 /hugetlb/hugeshmat/hugeshmat05
+/hugetlb/hugeshmat/hugeshmat06
 /hugetlb/hugeshmctl/hugeshmctl01
 /hugetlb/hugeshmctl/hugeshmctl02
 /hugetlb/hugeshmctl/hugeshmctl03
testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat06.c new file +16 -0
new file mode 100644
index 000000000..ca36ae2f4
@@ -0,0 +1,86 @@
+// SPDX-License-Identifier: LGPL-2.1-or-later
+/*
+ * Copyright (C) 2005-2006 IBM Corporation.
+ * Author: David Gibson & Adam Litke
+ */
+
+/*\
+ * [Description]
+ *
+ * Test shared memory behavior when multiple threads are attached  
+ * to a segment with different permissions.  A segment is created  
+ * and children attach read-only to check reservation accounting.  
+ */
+
+
+#include "tst_safe_sysv_ipc.h"

testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat06.c line 26

Quoted 10 lines
testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat06.c new file +10 -0
+#include "hugetlb.h"
+
+#define SEGMENT_SIZE ((size_t) 0x4000000)
+#define SEGMENT_KEY (0x82ba15ff)
+#define STRIDE (0x200000)
+#define MNTPOINT "hugetlbfs/"
+
+static int global_shmid = -1;
+
+void *shm_addr = NULL;

should be static

testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat06.c line 29

Quoted 3 lines
testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat06.c new file +3 -0
+static long hpage_size;
+
+int attach_segment(size_t segsize, int shmflags, int shmperms)

here as well.

testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat06.c line 34

Quoted 5 lines
testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat06.c new file +5 -0
+{
+    int shmid;
+    shmid = SAFE_SHMGET(SEGMENT_KEY, segsize, shmflags);
+    shm_addr = SAFE_SHMAT(shmid, shm_addr, shmperms); 
+    global_shmid = shmid;

I do not thint that this works as intended. The test cleanup() is not executed for child processes.

Quoted 5 lines
testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat06.c new file +5 -0
+    return shmid;
+}
+
+static void setup(void)
+{

testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat06.c line 49

Quoted 10 lines
testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat06.c new file +10 -0
+    hpage_size = tst_get_hugepage_size();
+}
+
+static void run_test(void)
+{
+    char *p;
+    int i, iterations;
+    long total_hpages = SAFE_READ_MEMINFO(MEMINFO_HPAGE_TOTAL);
+    if (hpage_size > SEGMENT_SIZE)
+        tst_res(TCONF,"Page size is too large for configured SEGMENT_SIZE");

Why do we set static sement size? Why can't we just use multiple of hugepage size instead?

testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat06.c line 63

Quoted 13 lines
testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat06.c new file +13 -0
+    iterations = (total_hpages * hpage_size) / SEGMENT_SIZE + 1;
+    SAFE_MALLOC(sizeof(pid_t) * iterations);
+    attach_segment(SEGMENT_SIZE, IPC_CREAT | SHM_HUGETLB | 0640, 0);
+    p = (char *)shm_addr;
+    for (i = 0; i < 4; i++, p += STRIDE)
+        memset(p, 0x55, STRIDE);
+    
+    SAFE_SHMDT((const void *)shm_addr);
+    for (i = 0; i < iterations; i++)
+    {
+        SAFE_FORK();
+        attach_segment(0, 0, SHM_RDONLY);
+        SAFE_SHMDT((const void *)shm_addr);

This code runs both for child and parent, this looks like a mistake.

Also at which point should the child exit?

testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat06.c line 64

Quoted 2 lines
testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat06.c new file +2 -0
+    }
+    tst_reap_children();

Here as well all children gets here.

testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat06.c line 66

Quoted 1 line
testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat06.c new file +1 -0
+    tst_res(TPASS, "Successfully tested shared memory behavior when multiple threads are attached");

What did we actually validated here? What is the purpose of the test?

My guess is that we are trying to check that we can attach shm hugepages several times and the number of hugepages in use does not grow, but the code does not seem to match that expecation.

Quoted 19 lines
testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat06.c new file +19 -0
+}
+
+static void cleanup(void)
+{
+    if (global_shmid >= 0)
+        SAFE_SHMCTL(global_shmid, IPC_RMID, NULL);
+}
+
+static struct tst_test test = {
+	.needs_root = 1,
+	.mntpoint = MNTPOINT,
+	.needs_hugetlbfs = 1,
+	.needs_tmpdir = 1,
+	.forks_child = 1,
+	.setup = setup,
+	.cleanup = cleanup,
+	.test_all = run_test,
+	.hugepages = {32, TST_NEEDS},
+};
Signature · 5 lines
Cyril Hrubis
[email protected]

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