Skip to contentSkip to conversation

mm/huge_memory: improve split_huge_page_to_list_to_order() return value documentation

PATCH Superseded David Hildenbrand 3 participants 5 messages Apr 18, 2024 – Apr 22, 2024

Subsystems (2)

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 mm/huge_memory: improve split_huge_page_to_list_to_order() return value documentation

The documentation is wrong and relying on it almost resulted in BUGs in new callers: we return -EAGAIN on unexpected folio references, not -EBUSY.

Let's fix that and also document which other return values we can currently see and why they could happen.

Cc John Hubbard <[email protected]>
Cc Zi Yan <[email protected]>
Cc Matthew Wilcox <[email protected]>
Cc Andrew Morton <[email protected]>
---
 mm/huge_memory.c | 13 ++++++++++---
 1 file changed, 10 insertions(+), 3 deletions(-)
mm/huge_memory.c +10 -3
index ee12726291f1b..824eff9211db8 100644
@@ -2956,7 +2956,7 @@ bool can_split_folio(struct folio *folio, int *pextra_pins)
  *
  * 3) The folio must not be pinned. Any unexpected folio references, including
  *    GUP pins, will result in the folio not getting split; instead, the caller
- *    will receive an -EBUSY.
+ *    will receive an -EAGAIN.
  *
  * 4) @new_order > 1, usually. Splitting to order-1 anonymous folios is not
  *    supported for non-file-backed folios, because folio->_deferred_list, which
@@ -2975,8 +2975,15 @@ bool can_split_folio(struct folio *folio, int *pextra_pins)
  *
  * Returns 0 if the huge page was split successfully.
  *
- * Returns -EBUSY if @page's folio is pinned, or if the anon_vma disappeared
- * from under us.
+ * Returns -EAGAIN if the folio has unexpected reference (e.g., GUP).
+ *
+ * Returns -EBUSY when trying to split the huge zeropage, if the folio is
+ * under writeback, if fs-specific folio metadata cannot currently be
+ * released, or if some unexpected race happened (e.g., anon VMA disappeared,
+ * truncation).
+ *
+ * Returns -EINVAL when trying to split to an order that is incompatible
+ * with the folio. Splitting to order 0 is compatible with all folios.
  */
 int split_huge_page_to_list_to_order(struct page *page, struct list_head *list,
 				     unsigned int new_order)
Signature · 1 line
2.44.0
#1 lore raw

mm/huge_memory.c bool can_split_folio(struct folio *folio, int *pextra_pins)

Quoted 30 lines · David Hildenbrand

The documentation is wrong and relying on it almost resulted in BUGs in new callers: we return -EAGAIN on unexpected folio references, not -EBUSY.

Let's fix that and also document which other return values we can currently see and why they could happen.

Cc John Hubbard <[email protected]>
Cc Zi Yan <[email protected]>
Cc Matthew Wilcox <[email protected]>
Cc Andrew Morton <[email protected]>
---
  mm/huge_memory.c | 13 ++++++++++---
  1 file changed, 10 insertions(+), 3 deletions(-)
mm/huge_memory.c +1 -1
index ee12726291f1b..824eff9211db8 100644
@@ -2956,7 +2956,7 @@ bool can_split_folio(struct folio *folio, int *pextra_pins)
   *
   * 3) The folio must not be pinned. Any unexpected folio references, including
   *    GUP pins, will result in the folio not getting split; instead, the caller
- *    will receive an -EBUSY.
+ *    will receive an -EAGAIN.
   *
   * 4) @new_order > 1, usually. Splitting to order-1 anonymous folios is not
   *    supported for non-file-backed folios, because folio->_deferred_list, which
@@ -2975,8 +2975,15 @@ bool can_split_folio(struct folio *folio, int *pextra_pins)

As an aside, the use of unconditional local_irq_disable() / local_irq_enable() calls in this routine almost makes me believe that we should have:

5) Local IRQs should be enabled. Because this routine may enable them.

...but I can't imagine a way to end up calling this with interrupts disabled, so it seems like documentation overkill. Just thought I'd mention it, though.

mm/huge_memory.c bool can_split_folio(struct folio *folio, int *pextra_pins) · line 2978

Quoted 6 lines
mm/huge_memory.c +1 -2
   *
   * Returns 0 if the huge page was split successfully.
   *
- * Returns -EBUSY if @page's folio is pinned, or if the anon_vma disappeared
- * from under us.
+ * Returns -EAGAIN if the folio has unexpected reference (e.g., GUP).

...or if the folio was removed from the page cache before this routine got a chance to lock it, right? (See the "fail:" path.)

mm/huge_memory.c bool can_split_folio(struct folio *folio, int *pextra_pins) · line 2989

Quoted 11 lines
mm/huge_memory.c +8 -0
+ *
+ * Returns -EBUSY when trying to split the huge zeropage, if the folio is
+ * under writeback, if fs-specific folio metadata cannot currently be
+ * released, or if some unexpected race happened (e.g., anon VMA disappeared,
+ * truncation).
+ *
+ * Returns -EINVAL when trying to split to an order that is incompatible
+ * with the folio. Splitting to order 0 is compatible with all folios.
   */
  int split_huge_page_to_list_to_order(struct page *page, struct list_head *list,
  				     unsigned int new_order)

Otherwise, looks good.

thanks,

Signature · 2 lines
John Hubbard
NVIDIA
#2 lore raw
Quoted 35 lines · John Hubbard

mm/huge_memory.c bool can_split_folio(struct folio *folio, int *pextra_pins)

Quoted 30 lines · David Hildenbrand

The documentation is wrong and relying on it almost resulted in BUGs in new callers: we return -EAGAIN on unexpected folio references, not -EBUSY.

Let's fix that and also document which other return values we can currently see and why they could happen.

Cc John Hubbard <[email protected]>
Cc Zi Yan <[email protected]>
Cc Matthew Wilcox <[email protected]>
Cc Andrew Morton <[email protected]>
---
   mm/huge_memory.c | 13 ++++++++++---
   1 file changed, 10 insertions(+), 3 deletions(-)
mm/huge_memory.c +1 -1
index ee12726291f1b..824eff9211db8 100644
@@ -2956,7 +2956,7 @@ bool can_split_folio(struct folio *folio, int *pextra_pins)
    *
    * 3) The folio must not be pinned. Any unexpected folio references, including
    *    GUP pins, will result in the folio not getting split; instead, the caller
- *    will receive an -EBUSY.
+ *    will receive an -EAGAIN.
    *
    * 4) @new_order > 1, usually. Splitting to order-1 anonymous folios is not
    *    supported for non-file-backed folios, because folio->_deferred_list, which
@@ -2975,8 +2975,15 @@ bool can_split_folio(struct folio *folio, int *pextra_pins)

As an aside, the use of unconditional local_irq_disable() / local_irq_enable() calls in this routine almost makes me believe that we should have:

Quoted 5 lines

5) Local IRQs should be enabled. Because this routine may enable them.

...but I can't imagine a way to end up calling this with interrupts disabled, so it seems like documentation overkill. Just thought I'd mention it, though.

Yes, I think there might be more issues lurking with disabled interrupts.

anon_vma_lock_write() and i_mmap_lock_read() might even sleep ... so we must not be in any atomic context. that's why relevant page table walkers drop the PTL.

Quoted 11 lines

mm/huge_memory.c bool can_split_folio(struct folio *folio, int *pextra_pins) · line 2978

Quoted 8 lines
mm/huge_memory.c +1 -2
    *
    * Returns 0 if the huge page was split successfully.
    *
- * Returns -EBUSY if @page's folio is pinned, or if the anon_vma disappeared
- * from under us.
+ * Returns -EAGAIN if the folio has unexpected reference (e.g., GUP).

...or if the folio was removed from the page cache before this routine got a chance to lock it, right? (See the "fail:" path.)

Right, that is sneaky. Let me extend to cover that case as well.

mm/huge_memory.c +2 -1
index 824eff9211db8..a7406267323ed 100644
@@ -2975,7 +2975,8 @@ bool can_split_folio(struct folio *folio, int *pextra_pins)
  *
  * Returns 0 if the huge page was split successfully.
  *
- * Returns -EAGAIN if the folio has unexpected reference (e.g., GUP).
+ * Returns -EAGAIN if the folio has unexpected reference (e.g., GUP) or if
+ * the folio was concurrently removed from the page cache.
  *
  * Returns -EBUSY when trying to split the huge zeropage, if the folio is
  * under writeback, if fs-specific folio metadata cannot currently be

Naive me would assume that this happens rarely ... but not an expert :)

Quoted 13 lines

mm/huge_memory.c bool can_split_folio(struct folio *folio, int *pextra_pins) · line 2989

Quoted 12 lines
mm/huge_memory.c +8 -0
+ *
+ * Returns -EBUSY when trying to split the huge zeropage, if the folio is
+ * under writeback, if fs-specific folio metadata cannot currently be
+ * released, or if some unexpected race happened (e.g., anon VMA disappeared,
+ * truncation).
+ *
+ * Returns -EINVAL when trying to split to an order that is incompatible
+ * with the folio. Splitting to order 0 is compatible with all folios.
    */
   int split_huge_page_to_list_to_order(struct page *page, struct list_head *list,
   				     unsigned int new_order)
Quoted 1 line

Otherwise, looks good.

Thanks!

Signature · 3 lines
Cheers,

David / dhildenb
Reviewed-by
#3 lore raw
Quoted 2 lines · David Hildenbrand
Quoted 1 line · John Hubbard

On 4/18/24 8:18 AM, David Hildenbrand wrote:

...

Quoted 17 lines
Quoted 9 lines
Quoted 6 lines

    *
    * Returns 0 if the huge page was split successfully.
    *
- * Returns -EBUSY if @page's folio is pinned, or if the anon_vma disappeared
- * from under us.
+ * Returns -EAGAIN if the folio has unexpected reference (e.g., GUP).

...or if the folio was removed from the page cache before this routine got a chance to lock it, right? (See the "fail:" path.)

Right, that is sneaky. Let me extend to cover that case as well.

mm/huge_memory.c +0 -0
index 824eff9211db8..a7406267323ed 100644

mm/huge_memory.c bool can_split_folio(struct folio *folio, int *pextra_pins) · line 2979

Quoted 6 lines
mm/huge_memory.c +2 -1
@@ -2975,7 +2975,8 @@ bool can_split_folio(struct folio *folio, int *pextra_pins)
   *
   * Returns 0 if the huge page was split successfully.
   *
- * Returns -EAGAIN if the folio has unexpected reference (e.g., GUP).
+ * Returns -EAGAIN if the folio has unexpected reference (e.g., GUP) or if
+ * the folio was concurrently removed from the page cache.

Looks good,

thanks,

Signature · 2 lines
John Hubbard
NVIDIA
reviewer
#4 lore raw
In reply to David Hildenbrand
Quoted 3 lines · David Hildenbrand

The documentation is wrong and relying on it almost resulted in BUGs in new callers: we return -EAGAIN on unexpected folio references, not -EBUSY.

+Baolin

The code was changed at the commit fd4a7ac32918 ("mm: migrate: try again if THP split is failed due to page refcnt") without changing the comment.

Quoted 4 lines

Let's fix that and also document which other return values we can currently see and why they could happen.

Quoted 8 lines
Cc John Hubbard <[email protected]>
Cc Zi Yan <[email protected]>
Cc Matthew Wilcox <[email protected]>
Cc Andrew Morton <[email protected]>
---
 mm/huge_memory.c | 13 ++++++++++---
 1 file changed, 10 insertions(+), 3 deletions(-)

The changes look good to me. Thanks. Reviewed-by: Zi Yan <[email protected]>

--
Best Regards,
Yan, Zi