add anon mTHP collapse test cases

To support checking for various sized mTHPs during mTHP collapse, extend the check_huge() function prototype to accept two new parameters specifying the address range and mTHP size, in preparation for the following patches.

No functional changes.

Signed-off-by Baolin Wang <[email protected]>

To support checking for various sized mTHPs during mTHP collapse, extend the check_huge() function prototype to accept two new parameters specifying the address range and mTHP size, in preparation for the following patches.

No functional changes.

Signed-off-by Baolin Wang <[email protected]>

---
 .../selftests/mm/folio_split_race_test.c      |  2 +-
 tools/testing/selftests/mm/khugepaged.c       | 66 ++++++++++---------
 tools/testing/selftests/mm/pagemap_ioctl.c    |  2 +-
 .../testing/selftests/mm/prctl_thp_disable.c  |  2 +-
 tools/testing/selftests/mm/soft-dirty.c       |  2 +-
 .../selftests/mm/split_huge_page_test.c       | 14 ++--
 tools/testing/selftests/mm/uffd-common.c      |  4 +-
 tools/testing/selftests/mm/vm_util.c          |  6 +-
 tools/testing/selftests/mm/vm_util.h          |  6 +-
 9 files changed, 56 insertions(+), 48 deletions(-)

tools/testing/selftests/mm/vm_util.h +0 -0

index 7799154b67ee..c7ae28dfda63 100644
@@ -90,9 +90,9 @@ void clear_softdirty(void);
 bool check_for_pattern(FILE *fp, const char *pattern, char *buf, size_t len);
 uint64_t read_pmd_pagesize(void);

Quoted 19 lines · Baolin Wang

To support checking for various sized mTHPs during mTHP collapse, extend the check_huge() function prototype to accept two new parameters specifying the address range and mTHP size, in preparation for the following patches.

No functional changes.

Signed-off-by Baolin Wang <[email protected]>

---
  .../selftests/mm/folio_split_race_test.c      |  2 +-
  tools/testing/selftests/mm/khugepaged.c       | 66 ++++++++++---------
  tools/testing/selftests/mm/pagemap_ioctl.c    |  2 +-
  .../testing/selftests/mm/prctl_thp_disable.c  |  2 +-
  tools/testing/selftests/mm/soft-dirty.c       |  2 +-
  .../selftests/mm/split_huge_page_test.c       | 14 ++--
  tools/testing/selftests/mm/uffd-common.c      |  4 +-
  tools/testing/selftests/mm/vm_util.c          |  6 +-
  tools/testing/selftests/mm/vm_util.h          |  6 +-
  9 files changed, 56 insertions(+), 48 deletions(-)

<snip>

Quoted 15 lines

tools/testing/selftests/mm/vm_util.h +3 -3

index 7799154b67ee..c7ae28dfda63 100644
@@ -90,9 +90,9 @@ void clear_softdirty(void);
  bool check_for_pattern(FILE *fp, const char *pattern, char *buf, size_t len);
  uint64_t read_pmd_pagesize(void);
  unsigned long rss_anon(void);
-bool check_huge_anon(void *addr, int nr_hpages, uint64_t hpage_size);
-bool check_huge_file(void *addr, int nr_hpages, uint64_t hpage_size);
-bool check_huge_shmem(void *addr, int nr_hpages, uint64_t hpage_size);
+bool check_huge_anon(void *addr, unsigned long size, int nr_hpages, uint64_t hpage_size);
+bool check_huge_file(void *addr, unsigned long size, int nr_hpages, uint64_t hpage_size);
+bool check_huge_shmem(void *addr, unsigned long size, int nr_hpages, uint64_t hpage_size);
  int64_t allocate_transhuge(void *ptr, int pagemap_fd);

tools/testing/selftests/mm/vm_util.h void clear_softdirty(void); · line 98

Quoted 2 lines

tools/testing/selftests/mm/vm_util.h +0 -0

  int pageflags_get(unsigned long pfn, int kpageflags_fd, uint64_t *flags);
 

I get what "size" means only after I see Patch 3 where gather_after_split_folio_orders() is used. I think len is a better name if not addr_len and its type should be size_t to match gather_after_split_folio_orders()'s signature. The rest patches should have the same rename and type change. Thanks.

Otherwise, LGTM.

Acked-by Zi Yan <[email protected]>

Move gather_after_split_folio_orders() to vm_util.c as a helper function in preparation for implementing checks for mTHP collapse. While we are at it, rename this function to indicate that it is not only used for large folio splits.

No functional changes.

Signed-off-by Baolin Wang <[email protected]>

Move gather_after_split_folio_orders() to vm_util.c as a helper function in preparation for implementing checks for mTHP collapse. While we are at it, rename this function to indicate that it is not only used for large folio splits.

No functional changes.

Signed-off-by Baolin Wang <[email protected]>

---
 .../selftests/mm/split_huge_page_test.c       | 125 +-----------------
 tools/testing/selftests/mm/vm_util.c          | 119 +++++++++++++++++
 tools/testing/selftests/mm/vm_util.h          |   2 +
 3 files changed, 122 insertions(+), 124 deletions(-)

Implement mTHP-sized hugepage checking helpers using gather_folio_orders(). Also rename the existing PMD-sized huge page check function to __check_pmd_huge() for clarity.

Signed-off-by Baolin Wang <[email protected]>

Implement mTHP-sized hugepage checking helpers using gather_folio_orders(). Also rename the existing PMD-sized huge page check function to __check_pmd_huge() for clarity.

Signed-off-by Baolin Wang <[email protected]>

---
 tools/testing/selftests/mm/vm_util.c | 60 ++++++++++++++++++++++++++--
 1 file changed, 56 insertions(+), 4 deletions(-)

tools/testing/selftests/mm/vm_util.c +0 -0

index 09e5d5cabe21..1240d783669c 100644

tools/testing/selftests/mm/vm_util.c +1 -1

 unsigned int __page_size;
 unsigned int __page_shift;

@@ -348,7 +352,7 @@ char *__get_smap_entry(void *addr, const char *pattern, char *buf, size_t len)
 	return entry;
 }
 
-bool __check_huge(void *addr, char *pattern, int nr_hpages,
+static bool __check_pmd_huge(void *addr, char *pattern, int nr_hpages,
 		  uint64_t hpage_size)
 {
 	char buffer[MAX_LINE_LENGTH];

tools/testing/selftests/mm/vm_util.c +20 -0

+
+	memset(orders, 0, sizeof(int) * NR_ORDERS);
+	pagemap_fd = open(PAGEMAP_PATH, O_RDONLY);
+	if (pagemap_fd == -1)
+		ksft_exit_fail_msg("read pagemap fail\n");
+
+	kpageflags_fd = open(KPAGEFLAGS_PATH, O_RDONLY);
+	if (kpageflags_fd == -1) {
+		close(pagemap_fd);
+		ksft_exit_fail_msg("read kpageflags fail\n");
+	}
+
+	status = gather_folio_orders(addr, size, pagemap_fd,
+			kpageflags_fd, orders, NR_ORDERS);
+	if (status)
+		goto out;
+
+	if (orders[order] == nr_hpages)
+		ret = true;
+

tools/testing/selftests/mm/vm_util.c +7 -1

+		return __check_pmd_huge(addr, "AnonHugePages: ", nr_hpages, hpage_size);
+
+	return check_large_folios(addr, size, nr_hpages, hpage_size);
 }
 
 bool check_huge_file(void *addr, unsigned long size, int nr_hpages, uint64_t hpage_size)
 {
-	return __check_huge(addr, "FilePmdMapped:", nr_hpages, hpage_size);
+	if (hpage_size == read_pmd_pagesize())
+		return __check_pmd_huge(addr, "FilePmdMapped:", nr_hpages, hpage_size);
+
+	return check_large_folios(addr, size, nr_hpages, hpage_size);
 }
 

Quoted 14 lines · Baolin Wang

Implement mTHP-sized hugepage checking helpers using gather_folio_orders(). Also rename the existing PMD-sized huge page check function to __check_pmd_huge() for clarity.

Signed-off-by Baolin Wang <[email protected]>

---
  tools/testing/selftests/mm/vm_util.c | 60 ++++++++++++++++++++++++++--
  1 file changed, 56 insertions(+), 4 deletions(-)

tools/testing/selftests/mm/vm_util.c +0 -0

index 09e5d5cabe21..1240d783669c 100644

tools/testing/selftests/mm/vm_util.c line 20

Quoted 6 lines

tools/testing/selftests/mm/vm_util.c +3 -0

@@ -15,6 +15,10 @@
  #define SMAP_FILE_PATH "/proc/self/smaps"
  #define STATUS_FILE_PATH "/proc/self/status"
  #define MAX_LINE_LENGTH 500
+#define PAGEMAP_PATH "/proc/self/pagemap"
+#define KPAGEFLAGS_PATH "/proc/kpageflags"
+#define GET_ORDER(nr_pages)    (31 - __builtin_clz(nr_pages))

This is a fast way of getting log2, but there is no check of the type of nr_pages. What prevents one using it on a long variable?

At least you can convert nr_pages to unsigned int and add a comment.

tools/testing/selftests/mm/vm_util.c line 21

Quoted 1 line

tools/testing/selftests/mm/vm_util.c +1 -0

+#define NR_ORDERS 20

Why 20? mTHP orders can only be in [1, pmd_order - 1] and different arch has different pmd_orders. check_large_folios() probably should just cap NR_ORDERS at pmd_order.

Quoted 15 lines

tools/testing/selftests/mm/vm_util.c +1 -1

  unsigned int __page_size;
  unsigned int __page_shift;

@@ -348,7 +352,7 @@ char *__get_smap_entry(void *addr, const char *pattern, char *buf, size_t len)
  	return entry;
  }
 
-bool __check_huge(void *addr, char *pattern, int nr_hpages,
+static bool __check_pmd_huge(void *addr, char *pattern, int nr_hpages,
  		  uint64_t hpage_size)
  {
  	char buffer[MAX_LINE_LENGTH];

@@ -366,19 +370,67 @@ bool __check_huge(void *addr, char *pattern, int nr_hpages,
  	return thp == (nr_hpages * (hpage_size >> 10));
  }
 

tools/testing/selftests/mm/vm_util.c bool __check_huge(void *addr, char *pattern, int nr_hpages, · line 378

Quoted 6 lines

tools/testing/selftests/mm/vm_util.c +6 -0

+static bool check_large_folios(void *addr, unsigned long size, int nr_hpages, uint64_t hpage_size)
+{
+	int order = 0, pagesize = getpagesize();
+	int nr_pages = hpage_size / pagesize;
+	int pagemap_fd, kpageflags_fd;
+	int orders[NR_ORDERS], status;

NR_ORDERS is a constant, so you can put orders on stack. I think you can rename it to MAX_NR_ORDERS, use 20,

than MAX_NR_ORDERS.

tools/testing/selftests/mm/vm_util.c bool __check_huge(void *addr, char *pattern, int nr_hpages, · line 382

Quoted 4 lines

tools/testing/selftests/mm/vm_util.c +4 -0

+	bool ret = false;
+
+	if (nr_pages > 0)
+		order = GET_ORDER(nr_pages);

This funciton can fail early if nr_pages is 0.

tools/testing/selftests/mm/vm_util.c bool __check_huge(void *addr, char *pattern, int nr_hpages, · line 385

Quoted 3 lines

tools/testing/selftests/mm/vm_util.c +3 -0

+
+	if (!order || order >= NR_ORDERS)
+		ksft_exit_fail_msg("invalid order\n");

It should check against pmd_order.

Quoted 27 lines

tools/testing/selftests/mm/vm_util.c +26 -0

+
+	memset(orders, 0, sizeof(int) * NR_ORDERS);
+	pagemap_fd = open(PAGEMAP_PATH, O_RDONLY);
+	if (pagemap_fd == -1)
+		ksft_exit_fail_msg("read pagemap fail\n");
+
+	kpageflags_fd = open(KPAGEFLAGS_PATH, O_RDONLY);
+	if (kpageflags_fd == -1) {
+		close(pagemap_fd);
+		ksft_exit_fail_msg("read kpageflags fail\n");
+	}
+
+	status = gather_folio_orders(addr, size, pagemap_fd,
+			kpageflags_fd, orders, NR_ORDERS);
+	if (status)
+		goto out;
+
+	if (orders[order] == nr_hpages)
+		ret = true;
+
+out:
+	close(pagemap_fd);
+	close(kpageflags_fd);
+	return ret;
+}
+

tools/testing/selftests/mm/vm_util.c bool __check_huge(void *addr, char *pattern, int nr_hpages, · line 414

Quoted 4 lines

tools/testing/selftests/mm/vm_util.c +1 -1

  bool check_huge_anon(void *addr, unsigned long size, int nr_hpages, uint64_t hpage_size)
  {
-	return __check_huge(addr, "AnonHugePages: ", nr_hpages, hpage_size);
+	if (hpage_size == read_pmd_pagesize())

read_pmd_pagesize() can fail and return 0. Either you need a check here or add a ksft_exit_fail_msg() in read_pmd_pagesize() to remove the burden from all callers.

Quoted 24 lines

tools/testing/selftests/mm/vm_util.c +11 -2

+		return __check_pmd_huge(addr, "AnonHugePages: ", nr_hpages, hpage_size);
+
+	return check_large_folios(addr, size, nr_hpages, hpage_size);
  }
 
  bool check_huge_file(void *addr, unsigned long size, int nr_hpages, uint64_t hpage_size)
  {
-	return __check_huge(addr, "FilePmdMapped:", nr_hpages, hpage_size);
+	if (hpage_size == read_pmd_pagesize())
+		return __check_pmd_huge(addr, "FilePmdMapped:", nr_hpages, hpage_size);
+
+	return check_large_folios(addr, size, nr_hpages, hpage_size);
  }
 
  bool check_huge_shmem(void *addr, unsigned long size, int nr_hpages, uint64_t hpage_size)
  {
-	return __check_huge(addr, "ShmemPmdMapped:", nr_hpages, hpage_size);
+	if (hpage_size == read_pmd_pagesize())
+		return __check_pmd_huge(addr, "ShmemPmdMapped:", nr_hpages, hpage_size);
+
+	return check_large_folios(addr, size, nr_hpages, hpage_size);
  }
 

tools/testing/selftests/mm/vm_util.c bool __check_huge(void *addr, char *pattern, int nr_hpages, · line 436

Quoted 1 line

tools/testing/selftests/mm/vm_util.c +0 -0

  int64_t allocate_transhuge(void *ptr, int pagemap_fd)

All read_pmd_pagesize()s above need to be handled.

Added a new command 'mthp_khugepaged' for mTHP collapse, along with the '-c' parameter to specify the collapse order. Additionally, added mTHP collapse test cases for 'collapse_full', 'collapse_empty', and 'collapse_single_mthp' for anonymous folios. All khugepaged test cases passed.

Signed-off-by Baolin Wang <[email protected]>

Added a new command 'mthp_khugepaged' for mTHP collapse, along with the '-c' parameter to specify the collapse order. Additionally, added mTHP collapse test cases for 'collapse_full', 'collapse_empty', and 'collapse_single_mthp' for anonymous folios. All khugepaged test cases passed.

Signed-off-by Baolin Wang <[email protected]>

---
 tools/testing/selftests/mm/khugepaged.c   | 130 ++++++++++++++++++----
 tools/testing/selftests/mm/run_vmtests.sh |   2 +
 2 files changed, 113 insertions(+), 19 deletions(-)

tools/testing/selftests/mm/khugepaged.c +2 -0

index f69be6be0ecd..64d6da6aa90d 100644
@@ -26,9 +26,11 @@
 
 #define BASE_ADDR ((void *)(1UL << 30))
 static unsigned long hpage_pmd_size;
+static int hpage_pmd_order;
 static unsigned long page_size;
 static int hpage_pmd_nr;
 static int anon_order;
+static int collapse_order;

tools/testing/selftests/mm/run_vmtests.sh +1 -0

+
 # Try to create XFS if not provided
 if [ -z "${SPLIT_HUGE_PAGE_TEST_XFS_PATH}" ]; then
     if test_selected "thp"; then

Quoted 24 lines · Baolin Wang

Added a new command 'mthp_khugepaged' for mTHP collapse, along with the '-c' parameter to specify the collapse order. Additionally, added mTHP collapse test cases for 'collapse_full', 'collapse_empty', and 'collapse_single_mthp' for anonymous folios. All khugepaged test cases passed.

Signed-off-by Baolin Wang <[email protected]>

---
  tools/testing/selftests/mm/khugepaged.c   | 130 ++++++++++++++++++----
  tools/testing/selftests/mm/run_vmtests.sh |   2 +
  2 files changed, 113 insertions(+), 19 deletions(-)

tools/testing/selftests/mm/khugepaged.c +2 -0

index f69be6be0ecd..64d6da6aa90d 100644
@@ -26,9 +26,11 @@
 
  #define BASE_ADDR ((void *)(1UL << 30))
  static unsigned long hpage_pmd_size;
+static int hpage_pmd_order;
  static unsigned long page_size;
  static int hpage_pmd_nr;
  static int anon_order;
+static int collapse_order;

<snip>

Quoted 7 lines

tools/testing/selftests/mm/run_vmtests.sh +0 -0

index 687d115e3bd8..d09f9f6a384e 100755
@@ -410,6 +410,8 @@ CATEGORY="thp" run_test ./khugepaged all:shmem
 

tools/testing/selftests/mm/run_vmtests.sh CATEGORY="thp" run_test ./khugepaged all:shmem · line 413

Quoted 3 lines

tools/testing/selftests/mm/run_vmtests.sh +1 -0

  CATEGORY="thp" run_test ./khugepaged -s 4 all:shmem
 
+CATEGORY="thp" run_test ./khugepaged -c 4 mthp_khugepaged:anon

Why only testing collapse_order == 4? It is more useful to test all possible mTHP orders. -c can accept common separated order and a range.

Why only testing mthp_khugepaged:anon? IIRC, file and shmem are not supported, so khugepaged should reject them, right?

Quoted 27 lines · Zi Yan

Quoted 23 lines · Baolin Wang

Added a new command 'mthp_khugepaged' for mTHP collapse, along with the '-c' parameter to specify the collapse order. Additionally, added mTHP collapse test cases for 'collapse_full', 'collapse_empty', and 'collapse_single_mthp' for anonymous folios. All khugepaged test cases passed.

Signed-off-by Baolin Wang <[email protected]>

---
  tools/testing/selftests/mm/khugepaged.c   | 130 ++++++++++++++++++----
  tools/testing/selftests/mm/run_vmtests.sh |   2 +
  2 files changed, 113 insertions(+), 19 deletions(-)

tools/testing/selftests/mm/khugepaged.c +2 -0

index f69be6be0ecd..64d6da6aa90d 100644
@@ -26,9 +26,11 @@
   #define BASE_ADDR ((void *)(1UL << 30))
  static unsigned long hpage_pmd_size;
+static int hpage_pmd_order;
  static unsigned long page_size;
  static int hpage_pmd_nr;
  static int anon_order;
+static int collapse_order;

<snip>

Quoted 10 lines

tools/testing/selftests/mm/run_vmtests.sh CATEGORY="thp" run_test ./khugepaged all:shmem · line 411

Quoted 7 lines

diff --git a/tools/testing/selftests/mm/run_vmtests.sh b/tools/testing/selftests/mm/run_vmtests.sh
index 687d115e3bd8..d09f9f6a384e 100755

tools/testing/selftests/mm/run_vmtests.sh +0 -0

@@ -410,6 +410,8 @@ CATEGORY="thp" run_test ./khugepaged all:shmem
   CATEGORY="thp" run_test ./khugepaged -s 4 all:shmem
 +CATEGORY="thp" run_test ./khugepaged -c 4 mthp_khugepaged:anon

Why only testing collapse_order == 4? It is more useful to test all possible mTHP orders. -c can accept common separated order and a range.

This also follows the previous '-s' parameter configuration, which is used to specify the mTHP order allocation.

I think specifying a single order size already tests the overall mTHP collapse

flow, so I'm not sure whether testing all orders is worthwhile. Maybe I can add a few more orders? ./khugepaged -c 2 mthp_khugepaged:anon
./khugepaged -c 4 mthp_khugepaged:anon
./khugepaged -c 8 mthp_khugepaged:anon

Quoted 2 lines

Why only testing mthp_khugepaged:anon? IIRC, file and shmem are not supported, so khugepaged should reject them, right?

Currently, I have not added test cases for shmem mTHP collapse, so adding file or shmem would be no-ops and might mislead others. I will add the shmem/file mTHP collapse feature and test cases in a following patchset.

Quoted 39 lines · Baolin Wang

Quoted 27 lines · Zi Yan

Quoted 23 lines · Baolin Wang

Added a new command 'mthp_khugepaged' for mTHP collapse, along with the '-c' parameter to specify the collapse order. Additionally, added mTHP collapse test cases for 'collapse_full', 'collapse_empty', and 'collapse_single_mthp' for anonymous folios. All khugepaged test cases passed.

Signed-off-by Baolin Wang <[email protected]>

---
   tools/testing/selftests/mm/khugepaged.c   | 130 ++++++++++++++++++----
   tools/testing/selftests/mm/run_vmtests.sh |   2 +
   2 files changed, 113 insertions(+), 19 deletions(-)

tools/testing/selftests/mm/khugepaged.c +2 -0

index f69be6be0ecd..64d6da6aa90d 100644
@@ -26,9 +26,11 @@
    #define BASE_ADDR ((void *)(1UL << 30))
   static unsigned long hpage_pmd_size;
+static int hpage_pmd_order;
   static unsigned long page_size;
   static int hpage_pmd_nr;
   static int anon_order;
+static int collapse_order;

<snip>

Quoted 10 lines

tools/testing/selftests/mm/run_vmtests.sh CATEGORY="thp" run_test ./khugepaged all:shmem · line 411

Quoted 7 lines

diff --git a/tools/testing/selftests/mm/run_vmtests.sh b/tools/testing/selftests/mm/run_vmtests.sh
index 687d115e3bd8..d09f9f6a384e 100755

tools/testing/selftests/mm/run_vmtests.sh +0 -0

@@ -410,6 +410,8 @@ CATEGORY="thp" run_test ./khugepaged all:shmem
    CATEGORY="thp" run_test ./khugepaged -s 4 all:shmem
  +CATEGORY="thp" run_test ./khugepaged -c 4 mthp_khugepaged:anon

Why only testing collapse_order == 4? It is more useful to test all possible mTHP orders. -c can accept common separated order and a range.

Quoted 3 lines

This also follows the previous '-s' parameter configuration, which is used to specify the mTHP order allocation.

I think specifying a single order size already tests the overall mTHP collapse

Got it. If a single order covers the flow, I am fine with it.

Quoted 5 lines

flow, so I'm not sure whether testing all orders is worthwhile. Maybe I can add a few more orders? ./khugepaged -c 2 mthp_khugepaged:anon
./khugepaged -c 4 mthp_khugepaged:anon
./khugepaged -c 8 mthp_khugepaged:anon

Quoted 4 lines

Quoted 2 lines

Why only testing mthp_khugepaged:anon? IIRC, file and shmem are not supported, so khugepaged should reject them, right?

Currently, I have not added test cases for shmem mTHP collapse, so adding file or shmem would be no-ops and might mislead others. I will add the shmem/file mTHP collapse feature and test cases in a following patchset.

Right. I mean when one specifies ./khugepaged mthp_khugepaged:file/shmem, the program should show some error like “file/shmem not supported”.

Quoted 55 lines · Zi Yan

Quoted 39 lines · Baolin Wang

Quoted 27 lines · Zi Yan

Quoted 23 lines · Baolin Wang

Added a new command 'mthp_khugepaged' for mTHP collapse, along with the '-c' parameter to specify the collapse order. Additionally, added mTHP collapse test cases for 'collapse_full', 'collapse_empty', and 'collapse_single_mthp' for anonymous folios. All khugepaged test cases passed.

Signed-off-by Baolin Wang <[email protected]>

---
   tools/testing/selftests/mm/khugepaged.c   | 130 ++++++++++++++++++----
   tools/testing/selftests/mm/run_vmtests.sh |   2 +
   2 files changed, 113 insertions(+), 19 deletions(-)

tools/testing/selftests/mm/khugepaged.c +2 -0

index f69be6be0ecd..64d6da6aa90d 100644
@@ -26,9 +26,11 @@
    #define BASE_ADDR ((void *)(1UL << 30))
   static unsigned long hpage_pmd_size;
+static int hpage_pmd_order;
   static unsigned long page_size;
   static int hpage_pmd_nr;
   static int anon_order;
+static int collapse_order;

<snip>

Quoted 10 lines

tools/testing/selftests/mm/run_vmtests.sh CATEGORY="thp" run_test ./khugepaged all:shmem · line 411

Quoted 7 lines

diff --git a/tools/testing/selftests/mm/run_vmtests.sh b/tools/testing/selftests/mm/run_vmtests.sh
index 687d115e3bd8..d09f9f6a384e 100755

tools/testing/selftests/mm/run_vmtests.sh +0 -0

@@ -410,6 +410,8 @@ CATEGORY="thp" run_test ./khugepaged all:shmem
    CATEGORY="thp" run_test ./khugepaged -s 4 all:shmem
  +CATEGORY="thp" run_test ./khugepaged -c 4 mthp_khugepaged:anon

Why only testing collapse_order == 4? It is more useful to test all possible mTHP orders. -c can accept common separated order and a range.

Quoted 3 lines

This also follows the previous '-s' parameter configuration, which is used to specify the mTHP order allocation.

I think specifying a single order size already tests the overall mTHP collapse

Got it. If a single order covers the flow, I am fine with it.

Quoted 8 lines

flow, so I'm not sure whether testing all orders is worthwhile. Maybe I can add a few more orders? ./khugepaged -c 2 mthp_khugepaged:anon
./khugepaged -c 4 mthp_khugepaged:anon
./khugepaged -c 8 mthp_khugepaged:anon

Quoted 2 lines

Why only testing mthp_khugepaged:anon? IIRC, file and shmem are not supported, so khugepaged should reject them, right?

Quoted 4 lines

Quoted 1 line

Currently, I have not added test cases for shmem mTHP collapse, so adding file or shmem would be no-ops and might mislead others. I will add the shmem/file mTHP collapse feature and test cases in a following patchset.

Right. I mean when one specifies ./khugepaged mthp_khugepaged:file/shmem, the program should show some error like “file/shmem not supported”.

Yes, that works too. I still prefer to add the corresponding test cases after I add shmem mTHP collapse support:)

But if you have a strong opinion on this, I can also add a "not supported" message first.

Quoted 58 lines · Baolin Wang

Quoted 57 lines · Zi Yan

Quoted 39 lines · Baolin Wang

Quoted 27 lines · Zi Yan

Quoted 23 lines · Baolin Wang

Added a new command 'mthp_khugepaged' for mTHP collapse, along with the '-c' parameter to specify the collapse order. Additionally, added mTHP collapse test cases for 'collapse_full', 'collapse_empty', and 'collapse_single_mthp' for anonymous folios. All khugepaged test cases passed.

Signed-off-by Baolin Wang <[email protected]>

---
    tools/testing/selftests/mm/khugepaged.c   | 130 ++++++++++++++++++----
    tools/testing/selftests/mm/run_vmtests.sh |   2 +
    2 files changed, 113 insertions(+), 19 deletions(-)

tools/testing/selftests/mm/khugepaged.c +2 -0

index f69be6be0ecd..64d6da6aa90d 100644
@@ -26,9 +26,11 @@
     #define BASE_ADDR ((void *)(1UL << 30))
    static unsigned long hpage_pmd_size;
+static int hpage_pmd_order;
    static unsigned long page_size;
    static int hpage_pmd_nr;
    static int anon_order;
+static int collapse_order;

<snip>

Quoted 10 lines

tools/testing/selftests/mm/run_vmtests.sh CATEGORY="thp" run_test ./khugepaged all:shmem · line 411

Quoted 7 lines

   diff --git a/tools/testing/selftests/mm/run_vmtests.sh b/tools/testing/selftests/mm/run_vmtests.sh

index 687d115e3bd8..d09f9f6a384e 100755

tools/testing/selftests/mm/run_vmtests.sh +0 -0

@@ -410,6 +410,8 @@ CATEGORY="thp" run_test ./khugepaged all:shmem
     CATEGORY="thp" run_test ./khugepaged -s 4 all:shmem
   +CATEGORY="thp" run_test ./khugepaged -c 4 mthp_khugepaged:anon

Why only testing collapse_order == 4? It is more useful to test all possible mTHP orders. -c can accept common separated order and a range.

Quoted 3 lines

This also follows the previous '-s' parameter configuration, which is used to specify the mTHP order allocation.

I think specifying a single order size already tests the overall mTHP collapse

Got it. If a single order covers the flow, I am fine with it.

Quoted 9 lines

flow, so I'm not sure whether testing all orders is worthwhile. Maybe I can add a few more orders? ./khugepaged -c 2 mthp_khugepaged:anon
./khugepaged -c 4 mthp_khugepaged:anon
./khugepaged -c 8 mthp_khugepaged:anon

Quoted 2 lines

Why only testing mthp_khugepaged:anon? IIRC, file and shmem are not supported, so khugepaged should reject them, right?

Currently, I have not added test cases for shmem mTHP collapse, so adding file or shmem would be no-ops and might mislead others. I will add the shmem/file mTHP collapse feature and test cases in a following patchset.

Quoted 4 lines

Quoted 2 lines

Right. I mean when one specifies ./khugepaged mthp_khugepaged:file/shmem, the program should show some error like “file/shmem not supported”.

Yes, that works too. I still prefer to add the corresponding test cases after I add shmem mTHP collapse support:)

Sure. Or update usage() to print “mthp_khugepaged only supports anon”. You can remove that after you add other support types.

This patchset is split from the previous shmem mTHP collapse patchset[1], which only adds anon mTHP collapse selftests. All khugepaged test cases can pass.

Quoted 3 lines

This patchset is split from the previous shmem mTHP collapse patchset[1], which only adds anon mTHP collapse selftests. All khugepaged test cases can pass.

Thank you for sending this out :) I played around with it and it seems to be working. This lays good groundwork to expand khugepaged mTHP testing in the future too.

Tested-by Nico Pache (Red Hat) <[email protected]>

This patchset is split from the previous shmem mTHP collapse patchset[1], which only adds anon mTHP collapse selftests. All khugepaged test cases can pass.

Quoted 3 lines · Baolin Wang

This patchset is split from the previous shmem mTHP collapse patchset[1], which only adds anon mTHP collapse selftests. All khugepaged test cases can pass.

The patchset looks good to me:

Acked-by Kiryl Shutsemau (Meta) <[email protected]>