Skip to contentSkip to conversation

mm: Remove stub for non_swap_entry()

PATCH Applied Peter Xu 2 participants 3 messages Apr 13, 2022 – Apr 14, 2022

Subsystems (2)
lore
Share & tools
lore

Peter Xu

author
#0 lore raw

PATCH 1/1 mm: Remove stub for non_swap_entry()

The stub for non_swap_entry() may not help much, because MAX_SWAPFILES has already contained all the information to decide whether a swap entry is real swap entry of pesudo ones (migrations, ...).

There can be some performance influences on non_swap_entry() with below conditions all met:

  !CONFIG_MIGRATION && !CONFIG_MEMORY_FAILURE && !CONFIG_DEVICE_PRIVATE

But that's definitely not the major config most machines will use, at the meantime it's already in a slow path of swap entry (being parsed from a swap pte), so IMHO it shouldn't be a major issue. Also according to the analysis from Alistair, somehow the stub didn't do the job right [1].

To make the code cleaner, let's drop the stub.

[1] https://lore.kernel.org/lkml/[email protected]/

Cc Alistair Popple <[email protected]>

---

Note: the uffd-wp shmem & hugetlbfs series will need this patch to make sure swap entries work as expected with below config as spotted by Alistair:

!CONFIG_MIGRATION &&
!CONFIG_MEMORY_FAILURE &&
!CONFIG_DEVICE_PRIVATE &&
CONFIG_PTE_MARKER

(PS: this config should mostly never gonna happen, though, afaict..)

Quoting the same thread [1] as above.

Please review, thanks.

---
 include/linux/swapops.h | 8 --------
 1 file changed, 8 deletions(-)
include/linux/swapops.h +0 -8
index fffbba0036f6..a291f210e7f8 100644
@@ -493,18 +493,10 @@ static inline void num_poisoned_pages_inc(void)
 }
 #endif
 
-#if defined(CONFIG_MEMORY_FAILURE) || defined(CONFIG_MIGRATION) || \
-    defined(CONFIG_DEVICE_PRIVATE)
 static inline int non_swap_entry(swp_entry_t entry)
 {
 	return swp_type(entry) >= MAX_SWAPFILES;
 }
-#else
-static inline int non_swap_entry(swp_entry_t entry)
-{
-	return 0;
-}
-#endif
 
 #endif /* CONFIG_MMU */
 #endif /* _LINUX_SWAPOPS_H */
Signature · 1 line
2.32.0
Reviewed-by on v1 1/1
#1 lore raw
Quoted 7 lines · Peter Xu

The stub for non_swap_entry() may not help much, because MAX_SWAPFILES has already contained all the information to decide whether a swap entry is real swap entry of pesudo ones (migrations, ...).

There can be some performance influences on non_swap_entry() with below conditions all met:

Quoted 6 lines
  !CONFIG_MIGRATION && !CONFIG_MEMORY_FAILURE && !CONFIG_DEVICE_PRIVATE

But that's definitely not the major config most machines will use, at the meantime it's already in a slow path of swap entry (being parsed from a swap pte), so IMHO it shouldn't be a major issue. Also according to the analysis from Alistair, somehow the stub didn't do the job right [1].

I wasn't so much concerned about execution speed given it's on the slow path anyway but overall code size, which is one reason all those config options might be disabled. However in practice it made little to no difference as those config options already remove most of the extra code so I agree we can drop the stub.

Quoted 49 lines

To make the code cleaner, let's drop the stub.

[1] <https://lore.kernel.org/lkml/[email protected]/>

Cc Alistair Popple <[email protected]>

---

Note: the uffd-wp shmem & hugetlbfs series will need this patch to make sure swap entries work as expected with below config as spotted by Alistair:

!CONFIG_MIGRATION &&
!CONFIG_MEMORY_FAILURE &&
!CONFIG_DEVICE_PRIVATE &&
CONFIG_PTE_MARKER

(PS: this config should mostly never gonna happen, though, afaict..)

Quoting the same thread [1] as above.

Please review, thanks.

---
 include/linux/swapops.h | 8 --------
 1 file changed, 8 deletions(-)
include/linux/swapops.h +0 -8
index fffbba0036f6..a291f210e7f8 100644
@@ -493,18 +493,10 @@ static inline void num_poisoned_pages_inc(void)
 }
 #endif
 
-#if defined(CONFIG_MEMORY_FAILURE) || defined(CONFIG_MIGRATION) || \
-    defined(CONFIG_DEVICE_PRIVATE)
 static inline int non_swap_entry(swp_entry_t entry)
 {
 	return swp_type(entry) >= MAX_SWAPFILES;
 }
-#else
-static inline int non_swap_entry(swp_entry_t entry)
-{
-	return 0;
-}
-#endif
 
 #endif /* CONFIG_MMU */
 #endif /* _LINUX_SWAPOPS_H */
author
#2 lore raw
Quoted 16 lines · Alistair Popple
Quoted 7 lines · Peter Xu

The stub for non_swap_entry() may not help much, because MAX_SWAPFILES has already contained all the information to decide whether a swap entry is real swap entry of pesudo ones (migrations, ...).

There can be some performance influences on non_swap_entry() with below conditions all met:

Quoted 6 lines
  !CONFIG_MIGRATION && !CONFIG_MEMORY_FAILURE && !CONFIG_DEVICE_PRIVATE

But that's definitely not the major config most machines will use, at the meantime it's already in a slow path of swap entry (being parsed from a swap pte), so IMHO it shouldn't be a major issue. Also according to the analysis from Alistair, somehow the stub didn't do the job right [1].

Quoted 4 lines

I wasn't so much concerned about execution speed given it's on the slow path anyway but overall code size, which is one reason all those config options might be disabled. However in practice it made little to no difference as those config options already remove most of the extra code so I agree we can drop the stub.

I see, yeah that's a good point.

I'd wildly guess a minumum set of Linux build could still like that, but not strongly, as I'd first think about not having CONFIG_SWAP at all if so.

Quoted 2 lines

Thanks!

Signature · 1 line
Peter Xu