On 27.03.24 07:09, Jinjiang Tu wrote:
In order to extend test_prctl_fork() and test_prctl_fork_exec() to make
sure that deduplication really happens, mmap_and_merge_range() needs to be
refactored.
Firstly, mmap_and_merge_range() will be called with no need to call enable
KSM by madvise or prctl. So, switch the 'bool use_prctl' parameter to enum
ksm_merge_mode.
Secondly, mmap_and_merge_range() will be called in child process in the
two testcases, it isn't appropriate to call ksft_test_result_{fail, skip},
because the global variables ksft_{fail, skip} aren't consistent with the
parent process. Thus, convert calls of ksft_test_result_{fail, skip} to
ksft_print_msg() and return differrent error according to the two cases.
The callers of mmap_and_merge_range() then call ksft_test_result_{fail,
skip} according to the return value. Introduce mmap_and_merge_err_print()
helper to make it easier.
Suggested-by: David Hildenbrand <david@xxxxxxxxxx>
Signed-off-by: Jinjiang Tu <tujinjiang@xxxxxxxxxx>
---
[...]
@@ -209,14 +228,28 @@ static char *mmap_and_merge_range(char val, unsigned long size, int prot,
* accounted differently (depending on kernel support).
*/
if (val && !get_my_merging_pages()) {
- ksft_test_result_fail("No pages got merged\n");
+ ksft_print_msg("No pages got merged\n");
goto unmap;
}
return map;
unmap:
munmap(map, size);
- return MAP_FAILED;
+ return err_map;
+}
+
+static inline int mmap_and_merge_err_print(char *map)
+{
+ int ret = -1;
+
+ if (map == MAP_MERGE_FAIL)
+ ksft_test_result_fail("Merging memory failed");
+ else if (map == MAP_MERGE_FAIL)
MAP_MERGE_SKIP
+ ksft_test_result_skip("Merging memory skiped");
"skipped"
+ else
+ ret = 0;
+
+ return ret;
}
static void test_unmerge(void)
@@ -226,8 +259,8 @@ static void test_unmerge(void)
ksft_print_msg("[RUN] %s\n", __func__);
- map = mmap_and_merge_range(0xcf, size, PROT_READ | PROT_WRITE, false);
- if (map == MAP_FAILED)
+ map = mmap_and_merge_range(0xcf, size, PROT_READ | PROT_WRITE, KSM_MERGE_MADVISE);
+ if (mmap_and_merge_err_print(map))
I like the concept.
The following would be cleaner and result in less churn:
1) Rename mmap_and_merge_range() to "__mmap_and_merge_range()"
2) Introduce a new mmap_and_merge_range()
static inline char *mmap_and_merge_range(char val, unsigned long size,
int prot, enum ksm_merge_mode mode)
{
char *map = __mmap_and_merge_range(val, size, prot, mode);
char *ret = MAP_FAILED;
if (map == MAP_MERGE_FAIL)
ksft_test_result_fail("Merging memory failed");
else if (map == MAP_MERGE_SKIP)
ksft_test_result_skip("Merging memory skipped");
else
ret = map;
return ret;
}
3) Use __mmap_and_merge_range() in patch #3.
No need to adjust any existing callers of mmap_and_merge_range().
--
Cheers,
David / dhildenb
