Skip to contentSkip to conversation

selftests/mm: FOLL_LONGTERM need to be updated to 0x100

PATCH Applied Ayush Jain 3 participants 3 messages Aug 8, 2023 – Aug 9, 2023

Subsystems (4)
lore
Share & tools
lore

Ayush Jain

author
#0 lore raw

PATCH 1/1 selftests/mm: FOLL_LONGTERM need to be updated to 0x100

After commit 2c2241081f7d ("mm/gup: move private gup FOLL_ flags to internal.h") FOLL_LONGTERM flag value got updated from 0x10000 to 0x100 at include/linux/mm_types.h.
As hmm.hmm_device_private.hmm_gup_test uses FOLL_LONGTERM
Updating same here as well.

Before this change test goes in an infinite assert loop in hmm.hmm_device_private.hmm_gup_test
==========================================================
RUN hmm.hmm_device_private.hmm_gup_test ...
hmm-tests.c:1962:hmm_gup_test:Expected HMM_DMIRROR_PROT_WRITE.. ..(2) == m[2] (34) hmm-tests.c:157:hmm_gup_test:Expected ret (-1) == 0 (0) hmm-tests.c:157:hmm_gup_test:Expected ret (-1) == 0 (0)
...
==========================================================

 Call Trace:
 <TASK>
 ? sched_clock+0xd/0x20
 ? __lock_acquire.constprop.0+0x120/0x6c0
 ? ktime_get+0x2c/0xd0
 ? sched_clock+0xd/0x20
 ? local_clock+0x12/0xd0
 ? lock_release+0x26e/0x3b0
 pin_user_pages_fast+0x4c/0x70
 gup_test_ioctl+0x4ff/0xbb0
 ? gup_test_ioctl+0x68c/0xbb0
 __x64_sys_ioctl+0x99/0xd0
 do_syscall_64+0x60/0x90
 ? syscall_exit_to_user_mode+0x2a/0x50
 ? do_syscall_64+0x6d/0x90
 ? syscall_exit_to_user_mode+0x2a/0x50
 ? do_syscall_64+0x6d/0x90
 ? irqentry_exit_to_user_mode+0xd/0x20
 ? irqentry_exit+0x3f/0x50
 ? exc_page_fault+0x96/0x200
 entry_SYSCALL_64_after_hwframe+0x72/0xdc
 RIP: 0033:0x7f6aaa31aaff

After this change test is able to pass successfully.

---
 tools/testing/selftests/mm/hmm-tests.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)
tools/testing/selftests/mm/hmm-tests.c +6 -1
index 4adaad1b822f..20294553a5dd 100644
@@ -57,9 +57,14 @@ enum {
 
 #define ALIGN(x, a) (((x) + (a - 1)) & (~((a) - 1)))
 /* Just the flags we need, copied from mm.h: */
+
+#ifndef FOLL_WRITE
 #define FOLL_WRITE	0x01	/* check pte is writable */
-#define FOLL_LONGTERM   0x10000 /* mapping lifetime is indefinite */
+#endif
 
+#ifndef FOLL_LONGTERM
+#define FOLL_LONGTERM   0x100 /* mapping lifetime is indefinite */
+#endif
 FIXTURE(hmm)
 {
 	int		fd;
Signature · 1 line
2.39.3
Reviewed-by on v1 1/1
#1 lore raw

On 8/8/23 05:43, Ayush Jain wrote:
...

Quoted 10 lines
tools/testing/selftests/mm/hmm-tests.c +2 -0
index 4adaad1b822f..20294553a5dd 100644
@@ -57,9 +57,14 @@ enum {
 
  #define ALIGN(x, a) (((x) + (a - 1)) & (~((a) - 1)))
  /* Just the flags we need, copied from mm.h: */
+
+#ifndef FOLL_WRITE

tools/testing/selftests/mm/hmm-tests.c enum { · line 70

Quoted 10 lines
tools/testing/selftests/mm/hmm-tests.c +4 -1
  #define FOLL_WRITE	0x01	/* check pte is writable */
-#define FOLL_LONGTERM   0x10000 /* mapping lifetime is indefinite */
+#endif
 
+#ifndef FOLL_LONGTERM
+#define FOLL_LONGTERM   0x100 /* mapping lifetime is indefinite */
+#endif
  FIXTURE(hmm)
  {
  	int		fd;

Yes, it's unfortunate that this selftest framework prevents easily including those FOLL_* flags directly.

Anyway, thanks for fixing this,

thanks,

Signature · 2 lines
John Hubbard
NVIDIA
Acked-by on v1 1/1
#2 lore raw
In reply to Ayush Jain
Quoted 58 lines · Ayush Jain

After commit 2c2241081f7d ("mm/gup: move private gup FOLL_ flags to internal.h") FOLL_LONGTERM flag value got updated from 0x10000 to 0x100 at include/linux/mm_types.h.
As hmm.hmm_device_private.hmm_gup_test uses FOLL_LONGTERM
Updating same here as well.

Before this change test goes in an infinite assert loop in hmm.hmm_device_private.hmm_gup_test
==========================================================
RUN hmm.hmm_device_private.hmm_gup_test ...
hmm-tests.c:1962:hmm_gup_test:Expected HMM_DMIRROR_PROT_WRITE.. ..(2) == m[2] (34) hmm-tests.c:157:hmm_gup_test:Expected ret (-1) == 0 (0) hmm-tests.c:157:hmm_gup_test:Expected ret (-1) == 0 (0)
...
==========================================================

  Call Trace:
  <TASK>
  ? sched_clock+0xd/0x20
  ? __lock_acquire.constprop.0+0x120/0x6c0
  ? ktime_get+0x2c/0xd0
  ? sched_clock+0xd/0x20
  ? local_clock+0x12/0xd0
  ? lock_release+0x26e/0x3b0
  pin_user_pages_fast+0x4c/0x70
  gup_test_ioctl+0x4ff/0xbb0
  ? gup_test_ioctl+0x68c/0xbb0
  __x64_sys_ioctl+0x99/0xd0
  do_syscall_64+0x60/0x90
  ? syscall_exit_to_user_mode+0x2a/0x50
  ? do_syscall_64+0x6d/0x90
  ? syscall_exit_to_user_mode+0x2a/0x50
  ? do_syscall_64+0x6d/0x90
  ? irqentry_exit_to_user_mode+0xd/0x20
  ? irqentry_exit+0x3f/0x50
  ? exc_page_fault+0x96/0x200
  entry_SYSCALL_64_after_hwframe+0x72/0xdc
  RIP: 0033:0x7f6aaa31aaff

After this change test is able to pass successfully.

---
  tools/testing/selftests/mm/hmm-tests.c | 7 ++++++-
  1 file changed, 6 insertions(+), 1 deletion(-)
tools/testing/selftests/mm/hmm-tests.c +2 -0
index 4adaad1b822f..20294553a5dd 100644
@@ -57,9 +57,14 @@ enum {
 
  #define ALIGN(x, a) (((x) + (a - 1)) & (~((a) - 1)))
  /* Just the flags we need, copied from mm.h: */
+
+#ifndef FOLL_WRITE

tools/testing/selftests/mm/hmm-tests.c enum { · line 70

Quoted 10 lines
tools/testing/selftests/mm/hmm-tests.c +4 -1
  #define FOLL_WRITE	0x01	/* check pte is writable */
-#define FOLL_LONGTERM   0x10000 /* mapping lifetime is indefinite */
+#endif
 
+#ifndef FOLL_LONGTERM
+#define FOLL_LONGTERM   0x100 /* mapping lifetime is indefinite */
+#endif
  FIXTURE(hmm)
  {
  	int		fd;

Kind of suboptimal that we are passing in FOLL_ flags. mm/gup_test.c should just have provided an abstraction for that.

Signature · 3 lines
Cheers,

David / dhildenb