mm/swap: fix race when skipping swapcache

PATCH Superseded Kairui Song 5 participants 20 messages Feb 16, 2024 – Feb 19, 2024

Subsystems (4)

Stable branch Memory management MM core MM swap

lore

Share & tools

lore

Kairui Song

author

2024-02-16 09:51 UTC #0 lore raw

PATCH v3 1/1 mm/swap: fix race when skipping swapcache

From: Kairui Song <[email protected]>

When skipping swapcache for SWP_SYNCHRONOUS_IO, if two or more threads swapin the same entry at the same time, they get different pages (A, B). Before one thread (T0) finishes the swapin and installs page (A) to the PTE, another thread (T1) could finish swapin of page (B), swap_free the entry, then swap out the possibly modified page reusing the same entry. It breaks the pte_same check in (T0) because PTE value is unchanged, causing ABA problem. Thread (T0) will install a stalled page (A) into the PTE and cause data corruption.

One possible callstack is like this:

CPU0                                 CPU1
----                                 ----
do_swap_page()                       do_swap_page() with same entry
<direct swapin path>                 <direct swapin path>
<alloc page A>                       <alloc page B>
swap_read_folio() <- read to page A  swap_read_folio() <- read to page B
<slow on later locks or interrupt>   <finished swapin first>
...                                  set_pte_at()

swap_free() <- entry is free <write to page B, now page A stalled> <swap out page B to same swap entry>
pte_same() <- Check pass, PTE seems
unchanged, but page A
is stalled!
swap_free() <- page B content lost!
set_pte_at() <- staled page A installed!

And besides, for ZRAM, swap_free() allows the swap device to discard the entry content, so even if page (B) is not modified, if swap_read_folio() on CPU0 happens later than swap_free() on CPU1, it may also cause data loss.

To fix this, reuse swapcache_prepare which will pin the swap entry using the cache flag, and allow only one thread to pin it. Release the pin after PT unlocked. Racers will simply wait since it's a rare and very short event. A schedule() call is added to avoid wasting too much CPU or adding too much noise to perf statistics

Other methods like increasing the swap count don't seem to be a good idea after some tests, that will cause racers to fall back to use the swap cache again. Parallel swapin using different methods leads to a much more complex scenario.

Reproducer:

This race issue can be triggered easily using a well constructed reproducer and patched brd (with a delay in read path) [1]:

With latest 6.8 mainline, race caused data loss can be observed easily:

$ gcc -g -lpthread test-thread-swap-race.c && ./a.out
  Polulating 32MB of memory region...
  Keep swapping out...
  Starting round 0...
  Spawning 65536 workers...
  32746 workers spawned, wait for done...
  Round 0: Error on 0x5aa00, expected 32746, got 32743, 3 data loss!
  Round 0: Error on 0x395200, expected 32746, got 32743, 3 data loss!
  Round 0: Error on 0x3fd000, expected 32746, got 32737, 9 data loss!
  Round 0 Failed, 15 data loss!

This reproducer spawns multiple threads sharing the same memory region using a small swap device. Every two threads updates mapped pages one by one in opposite direction trying to create a race, with one dedicated thread keep swapping out the data out using madvise.

The reproducer created a reproduce rate of about once every 5 minutes, so the race should be totally possible in production.

After this patch, I ran the reproducer for over a few hundred rounds and no data loss observed.

Performance overhead is minimal, microbenchmark swapin 10G from 32G zram:

Before:     10934698 us
After:      11157121 us
Non-direct: 13155355 us (Dropping SWP_SYNCHRONOUS_IO flag)

Fixes 0bcac06f27d7 ("mm, swap: skip swapcache for swapin of synchronous device")

Link https://github.com/ryncsn/emm-test-project/tree/master/swap-stress-race [1]

Reported-by "Huang, Ying" <[email protected]>

Closes https://lore.kernel.org/lkml/[email protected]/ open here

Signed-off-by Kairui Song <[email protected]>

Cc [email protected]

---
Update from V2:
- Add a schedule() if raced to prevent repeated page faults wasting CPU and add noise to perf statistics.
- Use a bool to state the special case instead of reusing existing variables fixing error handling [Minchan Kim].

V2: https://lore.kernel.org/all/[email protected]/

Update from V1:
- Add some words on ZRAM case, it will discard swap content on swap_free so the race window is a bit different but cure is the same. [Barry Song]
- Update comments make it cleaner [Huang, Ying]
- Add a function place holder to fix CONFIG_SWAP=n built [SeongJae Park]
- Update the commit message and summary, refer to SWP_SYNCHRONOUS_IO instead of "direct swapin path" [Yu Zhao]
- Update commit message.
- Collect Review and Acks.

V1: https://lore.kernel.org/all/[email protected]/

 include/linux/swap.h |  5 +++++
 mm/memory.c          | 20 ++++++++++++++++++++
 mm/swap.h            |  5 +++++
 mm/swapfile.c        | 13 +++++++++++++
 4 files changed, 43 insertions(+)

include/linux/swap.h +5 -0

index 4db00ddad261..8d28f6091a32 100644
@@ -549,6 +549,11 @@ static inline int swap_duplicate(swp_entry_t swp)
 	return 0;
 }
 
+static inline int swapcache_prepare(swp_entry_t swp)
+{
+	return 0;
+}
+
 static inline void swap_free(swp_entry_t swp)
 {
 }

mm/memory.c +20 -0

index 7e1f4849463a..7059230d0a54 100644
@@ -3799,6 +3799,7 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
 	struct page *page;
 	struct swap_info_struct *si = NULL;
 	rmap_t rmap_flags = RMAP_NONE;
+	bool need_clear_cache = false;
 	bool exclusive = false;
 	swp_entry_t entry;
 	pte_t pte;

@@ -3867,6 +3868,20 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
 	if (!folio) {
 		if (data_race(si->flags & SWP_SYNCHRONOUS_IO) &&
 		    __swap_count(entry) == 1) {
+			/*
+			 * Prevent parallel swapin from proceeding with
+			 * the cache flag. Otherwise, another thread may
+			 * finish swapin first, free the entry, and swapout
+			 * reusing the same entry. It's undetectable as
+			 * pte_same() returns true due to entry reuse.
+			 */
+			if (swapcache_prepare(entry)) {
+				/* Relax a bit to prevent rapid repeated page faults */
+				schedule();
+				goto out;
+			}
+			need_clear_cache = true;
+
 			/* skip swapcache */
 			folio = vma_alloc_folio(GFP_HIGHUSER_MOVABLE, 0,
 						vma, vmf->address, false);

@@ -4117,6 +4132,9 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
 	if (vmf->pte)
 		pte_unmap_unlock(vmf->pte, vmf->ptl);
 out:
+	/* Clear the swap cache pin for direct swapin after PTL unlock */
+	if (need_clear_cache)
+		swapcache_clear(si, entry);
 	if (si)
 		put_swap_device(si);
 	return ret;

@@ -4131,6 +4149,8 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
 		folio_unlock(swapcache);
 		folio_put(swapcache);
 	}
+	if (need_clear_cache)
+		swapcache_clear(si, entry);
 	if (si)
 		put_swap_device(si);
 	return ret;

mm/swap.h +5 -0

index 758c46ca671e..fc2f6ade7f80 100644
@@ -41,6 +41,7 @@ void __delete_from_swap_cache(struct folio *folio,
 void delete_from_swap_cache(struct folio *folio);
 void clear_shadow_from_swap_cache(int type, unsigned long begin,
 				  unsigned long end);
+void swapcache_clear(struct swap_info_struct *si, swp_entry_t entry);
 struct folio *swap_cache_get_folio(swp_entry_t entry,
 		struct vm_area_struct *vma, unsigned long addr);
 struct folio *filemap_get_incore_folio(struct address_space *mapping,

@@ -97,6 +98,10 @@ static inline int swap_writepage(struct page *p, struct writeback_control *wbc)
 	return 0;
 }
 
+static inline void swapcache_clear(struct swap_info_struct *si, swp_entry_t entry)
+{
+}
+
 static inline struct folio *swap_cache_get_folio(swp_entry_t entry,
 		struct vm_area_struct *vma, unsigned long addr)
 {

mm/swapfile.c +13 -0

index 556ff7347d5f..746aa9da5302 100644
@@ -3365,6 +3365,19 @@ int swapcache_prepare(swp_entry_t entry)
 	return __swap_duplicate(entry, SWAP_HAS_CACHE);
 }
 
+void swapcache_clear(struct swap_info_struct *si, swp_entry_t entry)
+{
+	struct swap_cluster_info *ci;
+	unsigned long offset = swp_offset(entry);
+	unsigned char usage;
+
+	ci = lock_cluster_or_swap_info(si, offset);
+	usage = __swap_entry_free_locked(si, offset, SWAP_HAS_CACHE);
+	unlock_cluster_or_swap_info(si, ci);
+	if (!usage)
+		free_swap_slot(entry);
+}
+
 struct swap_info_struct *swp_swap_info(swp_entry_t entry)
 {
 	return swap_type_to_swap_info(swp_type(entry));

Signature · 1 line

2.43.0

Barry Song

2024-02-16 10:14 UTC #1 lore raw

Quoted 149 lines · Kairui Song

From: Kairui Song <[email protected]>
When skipping swapcache for SWP_SYNCHRONOUS_IO, if two or more threads swapin the same entry at the same time, they get different pages (A, B). Before one thread (T0) finishes the swapin and installs page (A) to the PTE, another thread (T1) could finish swapin of page (B), swap_free the entry, then swap out the possibly modified page reusing the same entry. It breaks the pte_same check in (T0) because PTE value is unchanged, causing ABA problem. Thread (T0) will install a stalled page (A) into the PTE and cause data corruption.
One possible callstack is like this:
CPU0                                 CPU1
----                                 ----
do_swap_page()                       do_swap_page() with same entry
<direct swapin path>                 <direct swapin path>
<alloc page A>                       <alloc page B>
swap_read_folio() <- read to page A  swap_read_folio() <- read to page B
<slow on later locks or interrupt>   <finished swapin first>
.. set_pte_at()
swap_free() <- entry is free <write to page B, now page A stalled> <swap out page B to same swap entry>
pte_same() <- Check pass, PTE seems
unchanged, but page A
is stalled!
swap_free() <- page B content lost!
set_pte_at() <- staled page A installed!
And besides, for ZRAM, swap_free() allows the swap device to discard the entry content, so even if page (B) is not modified, if swap_read_folio() on CPU0 happens later than swap_free() on CPU1, it may also cause data loss.
To fix this, reuse swapcache_prepare which will pin the swap entry using the cache flag, and allow only one thread to pin it. Release the pin after PT unlocked. Racers will simply wait since it's a rare and very short event. A schedule() call is added to avoid wasting too much CPU or adding too much noise to perf statistics
Other methods like increasing the swap count don't seem to be a good idea after some tests, that will cause racers to fall back to use the swap cache again. Parallel swapin using different methods leads to a much more complex scenario.
Reproducer:
This race issue can be triggered easily using a well constructed reproducer and patched brd (with a delay in read path) [1]:
With latest 6.8 mainline, race caused data loss can be observed easily:
$ gcc -g -lpthread test-thread-swap-race.c && ./a.out
  Polulating 32MB of memory region...
  Keep swapping out...
  Starting round 0...
  Spawning 65536 workers...
  32746 workers spawned, wait for done...
  Round 0: Error on 0x5aa00, expected 32746, got 32743, 3 data loss!
  Round 0: Error on 0x395200, expected 32746, got 32743, 3 data loss!
  Round 0: Error on 0x3fd000, expected 32746, got 32737, 9 data loss!
  Round 0 Failed, 15 data loss!
This reproducer spawns multiple threads sharing the same memory region using a small swap device. Every two threads updates mapped pages one by one in opposite direction trying to create a race, with one dedicated thread keep swapping out the data out using madvise.
The reproducer created a reproduce rate of about once every 5 minutes, so the race should be totally possible in production.
After this patch, I ran the reproducer for over a few hundred rounds and no data loss observed.
Performance overhead is minimal, microbenchmark swapin 10G from 32G zram:
Before:     10934698 us
After:      11157121 us
Non-direct: 13155355 us (Dropping SWP_SYNCHRONOUS_IO flag)
Fixes 0bcac06f27d7 ("mm, swap: skip swapcache for swapin of synchronous device")
Link https://github.com/ryncsn/emm-test-project/tree/master/swap-stress-race [1]
Reported-by "Huang, Ying" <[email protected]>
Closes https://lore.kernel.org/lkml/[email protected]/ open here
Signed-off-by Kairui Song <[email protected]>
Cc [email protected]
---
Update from V2:
- Add a schedule() if raced to prevent repeated page faults wasting CPU and add noise to perf statistics.
- Use a bool to state the special case instead of reusing existing variables fixing error handling [Minchan Kim].
V2: https://lore.kernel.org/all/[email protected]/
Update from V1:
- Add some words on ZRAM case, it will discard swap content on swap_free so the race window is a bit different but cure is the same. [Barry Song]
- Update comments make it cleaner [Huang, Ying]
- Add a function place holder to fix CONFIG_SWAP=n built [SeongJae Park]
- Update the commit message and summary, refer to SWP_SYNCHRONOUS_IO instead of "direct swapin path" [Yu Zhao]
- Update commit message.
- Collect Review and Acks.
V1: https://lore.kernel.org/all/[email protected]/
 include/linux/swap.h |  5 +++++
 mm/memory.c          | 20 ++++++++++++++++++++
 mm/swap.h            |  5 +++++
 mm/swapfile.c        | 13 +++++++++++++
 4 files changed, 43 insertions(+)
include/linux/swap.h +5 -0
index 4db00ddad261..8d28f6091a32 100644
@@ -549,6 +549,11 @@ static inline int swap_duplicate(swp_entry_t swp)
        return 0;
 }
 
+static inline int swapcache_prepare(swp_entry_t swp)
+{
+       return 0;
+}
+
 static inline void swap_free(swp_entry_t swp)
 {
 }
mm/memory.c +4 -0
index 7e1f4849463a..7059230d0a54 100644
@@ -3799,6 +3799,7 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
        struct page *page;
        struct swap_info_struct *si = NULL;
        rmap_t rmap_flags = RMAP_NONE;
+       bool need_clear_cache = false;
        bool exclusive = false;
        swp_entry_t entry;
        pte_t pte;
@@ -3867,6 +3868,20 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
        if (!folio) {
                if (data_race(si->flags & SWP_SYNCHRONOUS_IO) &&
                    __swap_count(entry) == 1) {
+                       /*
+                        * Prevent parallel swapin from proceeding with
+                        * the cache flag. Otherwise, another thread may

mm/memory.c vm_fault_t do_swap_page(struct vm_fault *vmf) · line 3883

Quoted 10 lines

mm/memory.c +10 -0

+                        * finish swapin first, free the entry, and swapout
+                        * reusing the same entry. It's undetectable as
+                        * pte_same() returns true due to entry reuse.
+                        */
+                       if (swapcache_prepare(entry)) {
+                               /* Relax a bit to prevent rapid repeated page faults */
+                               schedule();
+                               goto out;
+                       }
+                       need_clear_cache = true;

Hi Kairui,
i remember Ying had a suggestion to move swapcache_prepare after swap_read_folio() to decrease race window. does that one even help more to somehow "fix" the counting issue?

Quoted 71 lines

mm/memory.c +6 -0

+
                        /* skip swapcache */
                        folio = vma_alloc_folio(GFP_HIGHUSER_MOVABLE, 0,
                                                vma, vmf->address, false);

@@ -4117,6 +4132,9 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
        if (vmf->pte)
                pte_unmap_unlock(vmf->pte, vmf->ptl);
 out:
+       /* Clear the swap cache pin for direct swapin after PTL unlock */
+       if (need_clear_cache)
+               swapcache_clear(si, entry);
        if (si)
                put_swap_device(si);
        return ret;

@@ -4131,6 +4149,8 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
                folio_unlock(swapcache);
                folio_put(swapcache);
        }
+       if (need_clear_cache)
+               swapcache_clear(si, entry);
        if (si)
                put_swap_device(si);
        return ret;

mm/swap.h +5 -0

index 758c46ca671e..fc2f6ade7f80 100644
@@ -41,6 +41,7 @@ void __delete_from_swap_cache(struct folio *folio,
 void delete_from_swap_cache(struct folio *folio);
 void clear_shadow_from_swap_cache(int type, unsigned long begin,
                                  unsigned long end);
+void swapcache_clear(struct swap_info_struct *si, swp_entry_t entry);
 struct folio *swap_cache_get_folio(swp_entry_t entry,
                struct vm_area_struct *vma, unsigned long addr);
 struct folio *filemap_get_incore_folio(struct address_space *mapping,

@@ -97,6 +98,10 @@ static inline int swap_writepage(struct page *p, struct writeback_control *wbc)
        return 0;
 }
 
+static inline void swapcache_clear(struct swap_info_struct *si, swp_entry_t entry)
+{
+}
+
 static inline struct folio *swap_cache_get_folio(swp_entry_t entry,
                struct vm_area_struct *vma, unsigned long addr)
 {

mm/swapfile.c +13 -1

index 556ff7347d5f..746aa9da5302 100644
@@ -3365,6 +3365,19 @@ int swapcache_prepare(swp_entry_t entry)
        return __swap_duplicate(entry, SWAP_HAS_CACHE);
 }
 
+void swapcache_clear(struct swap_info_struct *si, swp_entry_t entry)
+{
+       struct swap_cluster_info *ci;
+       unsigned long offset = swp_offset(entry);
+       unsigned char usage;
+
+       ci = lock_cluster_or_swap_info(si, offset);
+       usage = __swap_entry_free_locked(si, offset, SWAP_HAS_CACHE);
+       unlock_cluster_or_swap_info(si, ci);
+       if (!usage)
+               free_swap_slot(entry);
+}
+
 struct swap_info_struct *swp_swap_info(swp_entry_t entry)
 {
        return swap_type_to_swap_info(swp_type(entry));
--

Quoted 3 lines

2.43.0

Thanks
Barry

Kairui Song

author

2024-02-16 10:56 UTC #2 lore raw

Quoted 158 lines · Barry Song

Quoted 156 lines · Kairui Song
From: Kairui Song <[email protected]>
When skipping swapcache for SWP_SYNCHRONOUS_IO, if two or more threads swapin the same entry at the same time, they get different pages (A, B). Before one thread (T0) finishes the swapin and installs page (A) to the PTE, another thread (T1) could finish swapin of page (B), swap_free the entry, then swap out the possibly modified page reusing the same entry. It breaks the pte_same check in (T0) because PTE value is unchanged, causing ABA problem. Thread (T0) will install a stalled page (A) into the PTE and cause data corruption.
One possible callstack is like this:
CPU0                                 CPU1
----                                 ----
do_swap_page()                       do_swap_page() with same entry
<direct swapin path>                 <direct swapin path>
<alloc page A>                       <alloc page B>
swap_read_folio() <- read to page A  swap_read_folio() <- read to page B
<slow on later locks or interrupt>   <finished swapin first>
.. set_pte_at()
swap_free() <- entry is free <write to page B, now page A stalled> <swap out page B to same swap entry>
pte_same() <- Check pass, PTE seems
unchanged, but page A
is stalled!
swap_free() <- page B content lost!
set_pte_at() <- staled page A installed!
And besides, for ZRAM, swap_free() allows the swap device to discard the entry content, so even if page (B) is not modified, if swap_read_folio() on CPU0 happens later than swap_free() on CPU1, it may also cause data loss.
To fix this, reuse swapcache_prepare which will pin the swap entry using the cache flag, and allow only one thread to pin it. Release the pin after PT unlocked. Racers will simply wait since it's a rare and very short event. A schedule() call is added to avoid wasting too much CPU or adding too much noise to perf statistics
Other methods like increasing the swap count don't seem to be a good idea after some tests, that will cause racers to fall back to use the swap cache again. Parallel swapin using different methods leads to a much more complex scenario.
Reproducer:
This race issue can be triggered easily using a well constructed reproducer and patched brd (with a delay in read path) [1]:
With latest 6.8 mainline, race caused data loss can be observed easily:
$ gcc -g -lpthread test-thread-swap-race.c && ./a.out
  Polulating 32MB of memory region...
  Keep swapping out...
  Starting round 0...
  Spawning 65536 workers...
  32746 workers spawned, wait for done...
  Round 0: Error on 0x5aa00, expected 32746, got 32743, 3 data loss!
  Round 0: Error on 0x395200, expected 32746, got 32743, 3 data loss!
  Round 0: Error on 0x3fd000, expected 32746, got 32737, 9 data loss!
  Round 0 Failed, 15 data loss!
This reproducer spawns multiple threads sharing the same memory region using a small swap device. Every two threads updates mapped pages one by one in opposite direction trying to create a race, with one dedicated thread keep swapping out the data out using madvise.
The reproducer created a reproduce rate of about once every 5 minutes, so the race should be totally possible in production.
After this patch, I ran the reproducer for over a few hundred rounds and no data loss observed.
Performance overhead is minimal, microbenchmark swapin 10G from 32G zram:
Before:     10934698 us
After:      11157121 us
Non-direct: 13155355 us (Dropping SWP_SYNCHRONOUS_IO flag)
Fixes 0bcac06f27d7 ("mm, swap: skip swapcache for swapin of synchronous device")
Link https://github.com/ryncsn/emm-test-project/tree/master/swap-stress-race [1]
Reported-by "Huang, Ying" <[email protected]>
Closes https://lore.kernel.org/lkml/[email protected]/ open here
Signed-off-by Kairui Song <[email protected]>
Cc [email protected]
---
Update from V2:
- Add a schedule() if raced to prevent repeated page faults wasting CPU and add noise to perf statistics.
- Use a bool to state the special case instead of reusing existing variables fixing error handling [Minchan Kim].
V2: https://lore.kernel.org/all/[email protected]/
Update from V1:
- Add some words on ZRAM case, it will discard swap content on swap_free so the race window is a bit different but cure is the same. [Barry Song]
- Update comments make it cleaner [Huang, Ying]
- Add a function place holder to fix CONFIG_SWAP=n built [SeongJae Park]
- Update the commit message and summary, refer to SWP_SYNCHRONOUS_IO instead of "direct swapin path" [Yu Zhao]
- Update commit message.
- Collect Review and Acks.
V1: https://lore.kernel.org/all/[email protected]/
 include/linux/swap.h |  5 +++++
 mm/memory.c          | 20 ++++++++++++++++++++
 mm/swap.h            |  5 +++++
 mm/swapfile.c        | 13 +++++++++++++
 4 files changed, 43 insertions(+)
include/linux/swap.h +5 -0
index 4db00ddad261..8d28f6091a32 100644
@@ -549,6 +549,11 @@ static inline int swap_duplicate(swp_entry_t swp)
        return 0;
 }
 
+static inline int swapcache_prepare(swp_entry_t swp)
+{
+       return 0;
+}
+
 static inline void swap_free(swp_entry_t swp)
 {
 }
mm/memory.c +11 -0
index 7e1f4849463a..7059230d0a54 100644
@@ -3799,6 +3799,7 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
        struct page *page;
        struct swap_info_struct *si = NULL;
        rmap_t rmap_flags = RMAP_NONE;
+       bool need_clear_cache = false;
        bool exclusive = false;
        swp_entry_t entry;
        pte_t pte;
@@ -3867,6 +3868,20 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
        if (!folio) {
                if (data_race(si->flags & SWP_SYNCHRONOUS_IO) &&
                    __swap_count(entry) == 1) {
+                       /*
+                        * Prevent parallel swapin from proceeding with
+                        * the cache flag. Otherwise, another thread may
+                        * finish swapin first, free the entry, and swapout
+                        * reusing the same entry. It's undetectable as
+                        * pte_same() returns true due to entry reuse.
+                        */
+                       if (swapcache_prepare(entry)) {
+                               /* Relax a bit to prevent rapid repeated page faults */
+                               schedule();

Quoted 9 lines

mm/memory.c vm_fault_t do_swap_page(struct vm_fault *vmf) · line 3883
Quoted 3 lines
mm/memory.c +3 -0
+                               goto out;
+                       }
+                       need_clear_cache = true;
Hi Kairui,
i remember Ying had a suggestion to move swapcache_prepare after swap_read_folio() to decrease race window. does that one even help more to somehow "fix" the counting issue?

Hi Barry,

Thanks for the comments!

Yes, that's one of the suggestions, I found the result already looks good enough in test scenario I've posted after adding scheduler() change. After moving swapcache_prepare actually waste more CPU as it need to alloc/free extra folios now, and test result for the counting issue barely changed, so I kept the original code.

Huang, Ying

2024-02-18 08:44 UTC #3 lore raw

Quoted 171 lines · Kairui Song

Quoted 158 lines · Barry Song
Quoted 156 lines · Kairui Song
From: Kairui Song <[email protected]>
When skipping swapcache for SWP_SYNCHRONOUS_IO, if two or more threads swapin the same entry at the same time, they get different pages (A, B). Before one thread (T0) finishes the swapin and installs page (A) to the PTE, another thread (T1) could finish swapin of page (B), swap_free the entry, then swap out the possibly modified page reusing the same entry. It breaks the pte_same check in (T0) because PTE value is unchanged, causing ABA problem. Thread (T0) will install a stalled page (A) into the PTE and cause data corruption.
One possible callstack is like this:
CPU0                                 CPU1
----                                 ----
do_swap_page()                       do_swap_page() with same entry
<direct swapin path>                 <direct swapin path>
<alloc page A>                       <alloc page B>
swap_read_folio() <- read to page A  swap_read_folio() <- read to page B
<slow on later locks or interrupt>   <finished swapin first>
.. set_pte_at()
swap_free() <- entry is free <write to page B, now page A stalled> <swap out page B to same swap entry>
pte_same() <- Check pass, PTE seems
unchanged, but page A
is stalled!
swap_free() <- page B content lost!
set_pte_at() <- staled page A installed!
And besides, for ZRAM, swap_free() allows the swap device to discard the entry content, so even if page (B) is not modified, if swap_read_folio() on CPU0 happens later than swap_free() on CPU1, it may also cause data loss.
To fix this, reuse swapcache_prepare which will pin the swap entry using the cache flag, and allow only one thread to pin it. Release the pin after PT unlocked. Racers will simply wait since it's a rare and very short event. A schedule() call is added to avoid wasting too much CPU or adding too much noise to perf statistics
Other methods like increasing the swap count don't seem to be a good idea after some tests, that will cause racers to fall back to use the swap cache again. Parallel swapin using different methods leads to a much more complex scenario.
Reproducer:
This race issue can be triggered easily using a well constructed reproducer and patched brd (with a delay in read path) [1]:
With latest 6.8 mainline, race caused data loss can be observed easily:
$ gcc -g -lpthread test-thread-swap-race.c && ./a.out
  Polulating 32MB of memory region...
  Keep swapping out...
  Starting round 0...
  Spawning 65536 workers...
  32746 workers spawned, wait for done...
  Round 0: Error on 0x5aa00, expected 32746, got 32743, 3 data loss!
  Round 0: Error on 0x395200, expected 32746, got 32743, 3 data loss!
  Round 0: Error on 0x3fd000, expected 32746, got 32737, 9 data loss!
  Round 0 Failed, 15 data loss!
This reproducer spawns multiple threads sharing the same memory region using a small swap device. Every two threads updates mapped pages one by one in opposite direction trying to create a race, with one dedicated thread keep swapping out the data out using madvise.
The reproducer created a reproduce rate of about once every 5 minutes, so the race should be totally possible in production.
After this patch, I ran the reproducer for over a few hundred rounds and no data loss observed.
Performance overhead is minimal, microbenchmark swapin 10G from 32G zram:
Before:     10934698 us
After:      11157121 us
Non-direct: 13155355 us (Dropping SWP_SYNCHRONOUS_IO flag)
Fixes 0bcac06f27d7 ("mm, swap: skip swapcache for swapin of synchronous device")
Link https://github.com/ryncsn/emm-test-project/tree/master/swap-stress-race [1]
Reported-by "Huang, Ying" <[email protected]>
Closes https://lore.kernel.org/lkml/[email protected]/ open here
Signed-off-by Kairui Song <[email protected]>
Cc [email protected]
---
Update from V2:
- Add a schedule() if raced to prevent repeated page faults wasting CPU and add noise to perf statistics.
- Use a bool to state the special case instead of reusing existing variables fixing error handling [Minchan Kim].
V2: https://lore.kernel.org/all/[email protected]/
Update from V1:
- Add some words on ZRAM case, it will discard swap content on swap_free so the race window is a bit different but cure is the same. [Barry Song]
- Update comments make it cleaner [Huang, Ying]
- Add a function place holder to fix CONFIG_SWAP=n built [SeongJae Park]
- Update the commit message and summary, refer to SWP_SYNCHRONOUS_IO instead of "direct swapin path" [Yu Zhao]
- Update commit message.
- Collect Review and Acks.
V1: https://lore.kernel.org/all/[email protected]/
 include/linux/swap.h |  5 +++++
 mm/memory.c          | 20 ++++++++++++++++++++
 mm/swap.h            |  5 +++++
 mm/swapfile.c        | 13 +++++++++++++
 4 files changed, 43 insertions(+)
include/linux/swap.h +5 -0
index 4db00ddad261..8d28f6091a32 100644
@@ -549,6 +549,11 @@ static inline int swap_duplicate(swp_entry_t swp)
        return 0;
 }
 
+static inline int swapcache_prepare(swp_entry_t swp)
+{
+       return 0;
+}
+
 static inline void swap_free(swp_entry_t swp)
 {
 }
mm/memory.c +11 -0
index 7e1f4849463a..7059230d0a54 100644
@@ -3799,6 +3799,7 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
        struct page *page;
        struct swap_info_struct *si = NULL;
        rmap_t rmap_flags = RMAP_NONE;
+       bool need_clear_cache = false;
        bool exclusive = false;
        swp_entry_t entry;
        pte_t pte;
@@ -3867,6 +3868,20 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
        if (!folio) {
                if (data_race(si->flags & SWP_SYNCHRONOUS_IO) &&
                    __swap_count(entry) == 1) {
+                       /*
+                        * Prevent parallel swapin from proceeding with
+                        * the cache flag. Otherwise, another thread may
+                        * finish swapin first, free the entry, and swapout
+                        * reusing the same entry. It's undetectable as
+                        * pte_same() returns true due to entry reuse.
+                        */
+                       if (swapcache_prepare(entry)) {
+                               /* Relax a bit to prevent rapid repeated page faults */
+                               schedule();
Quoted 9 lines
mm/memory.c vm_fault_t do_swap_page(struct vm_fault *vmf) · line 3883
Quoted 3 lines
mm/memory.c +3 -0
+                               goto out;
+                       }
+                       need_clear_cache = true;
Hi Kairui,
i remember Ying had a suggestion to move swapcache_prepare after swap_read_folio() to decrease race window. does that one even help more to somehow "fix" the counting issue?
Hi Barry,

Quoted 7 lines

Thanks for the comments!
Yes, that's one of the suggestions, I found the result already looks good enough in test scenario I've posted after adding scheduler() change. After moving swapcache_prepare actually waste more CPU as it need to alloc/free extra folios now, and test result for the counting issue barely changed, so I kept the original code.

The time to allocate folios is unbounded, especially when direct reclaiming is triggered. And we can clear swap cache flag earlier to reduce the race window too.

--
Best Regards,
Huang, Ying

David Hildenbrand

2024-02-16 16:53 UTC #4 lore raw

In reply to Kairui Song

Quoted 149 lines · Kairui Song

From: Kairui Song <[email protected]>
When skipping swapcache for SWP_SYNCHRONOUS_IO, if two or more threads swapin the same entry at the same time, they get different pages (A, B). Before one thread (T0) finishes the swapin and installs page (A) to the PTE, another thread (T1) could finish swapin of page (B), swap_free the entry, then swap out the possibly modified page reusing the same entry. It breaks the pte_same check in (T0) because PTE value is unchanged, causing ABA problem. Thread (T0) will install a stalled page (A) into the PTE and cause data corruption.
One possible callstack is like this:
CPU0                                 CPU1
----                                 ----
do_swap_page()                       do_swap_page() with same entry
<direct swapin path>                 <direct swapin path>
<alloc page A>                       <alloc page B>
swap_read_folio() <- read to page A  swap_read_folio() <- read to page B
<slow on later locks or interrupt>   <finished swapin first>
...                                  set_pte_at()
swap_free() <- entry is free <write to page B, now page A stalled> <swap out page B to same swap entry>
pte_same() <- Check pass, PTE seems
unchanged, but page A
is stalled!
swap_free() <- page B content lost!
set_pte_at() <- staled page A installed!
And besides, for ZRAM, swap_free() allows the swap device to discard the entry content, so even if page (B) is not modified, if swap_read_folio() on CPU0 happens later than swap_free() on CPU1, it may also cause data loss.
To fix this, reuse swapcache_prepare which will pin the swap entry using the cache flag, and allow only one thread to pin it. Release the pin after PT unlocked. Racers will simply wait since it's a rare and very short event. A schedule() call is added to avoid wasting too much CPU or adding too much noise to perf statistics
Other methods like increasing the swap count don't seem to be a good idea after some tests, that will cause racers to fall back to use the swap cache again. Parallel swapin using different methods leads to a much more complex scenario.
Reproducer:
This race issue can be triggered easily using a well constructed reproducer and patched brd (with a delay in read path) [1]:
With latest 6.8 mainline, race caused data loss can be observed easily:
$ gcc -g -lpthread test-thread-swap-race.c && ./a.out
   Polulating 32MB of memory region...
   Keep swapping out...
   Starting round 0...
   Spawning 65536 workers...
   32746 workers spawned, wait for done...
   Round 0: Error on 0x5aa00, expected 32746, got 32743, 3 data loss!
   Round 0: Error on 0x395200, expected 32746, got 32743, 3 data loss!
   Round 0: Error on 0x3fd000, expected 32746, got 32737, 9 data loss!
   Round 0 Failed, 15 data loss!
This reproducer spawns multiple threads sharing the same memory region using a small swap device. Every two threads updates mapped pages one by one in opposite direction trying to create a race, with one dedicated thread keep swapping out the data out using madvise.
The reproducer created a reproduce rate of about once every 5 minutes, so the race should be totally possible in production.
After this patch, I ran the reproducer for over a few hundred rounds and no data loss observed.
Performance overhead is minimal, microbenchmark swapin 10G from 32G zram:
Before:     10934698 us
After:      11157121 us
Non-direct: 13155355 us (Dropping SWP_SYNCHRONOUS_IO flag)
Fixes 0bcac06f27d7 ("mm, swap: skip swapcache for swapin of synchronous device")
Link https://github.com/ryncsn/emm-test-project/tree/master/swap-stress-race [1]
Reported-by "Huang, Ying" <[email protected]>
Closes https://lore.kernel.org/lkml/[email protected]/ open here
Signed-off-by Kairui Song <[email protected]>
Cc [email protected]
---
Update from V2:
- Add a schedule() if raced to prevent repeated page faults wasting CPU and add noise to perf statistics.
- Use a bool to state the special case instead of reusing existing variables fixing error handling [Minchan Kim].
V2: https://lore.kernel.org/all/[email protected]/
Update from V1:
- Add some words on ZRAM case, it will discard swap content on swap_free so the race window is a bit different but cure is the same. [Barry Song]
- Update comments make it cleaner [Huang, Ying]
- Add a function place holder to fix CONFIG_SWAP=n built [SeongJae Park]
- Update the commit message and summary, refer to SWP_SYNCHRONOUS_IO instead of "direct swapin path" [Yu Zhao]
- Update commit message.
- Collect Review and Acks.
V1: https://lore.kernel.org/all/[email protected]/
  include/linux/swap.h |  5 +++++
  mm/memory.c          | 20 ++++++++++++++++++++
  mm/swap.h            |  5 +++++
  mm/swapfile.c        | 13 +++++++++++++
  4 files changed, 43 insertions(+)
include/linux/swap.h +5 -0
index 4db00ddad261..8d28f6091a32 100644
@@ -549,6 +549,11 @@ static inline int swap_duplicate(swp_entry_t swp)
  	return 0;
  }
 
+static inline int swapcache_prepare(swp_entry_t swp)
+{
+	return 0;
+}
+
  static inline void swap_free(swp_entry_t swp)
  {
  }
mm/memory.c +5 -0
index 7e1f4849463a..7059230d0a54 100644
@@ -3799,6 +3799,7 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
  	struct page *page;
  	struct swap_info_struct *si = NULL;
  	rmap_t rmap_flags = RMAP_NONE;
+	bool need_clear_cache = false;
  	bool exclusive = false;
  	swp_entry_t entry;
  	pte_t pte;
@@ -3867,6 +3868,20 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
  	if (!folio) {
  		if (data_race(si->flags & SWP_SYNCHRONOUS_IO) &&
  		    __swap_count(entry) == 1) {
+			/*
+			 * Prevent parallel swapin from proceeding with
+			 * the cache flag. Otherwise, another thread may
+			 * finish swapin first, free the entry, and swapout

mm/memory.c vm_fault_t do_swap_page(struct vm_fault *vmf) · line 3884

Quoted 10 lines

mm/memory.c +10 -0

+			 * reusing the same entry. It's undetectable as
+			 * pte_same() returns true due to entry reuse.
+			 */
+			if (swapcache_prepare(entry)) {
+				/* Relax a bit to prevent rapid repeated page faults */
+				schedule();
+				goto out;
+			}
+			need_clear_cache = true;
+

I took a closer look at __read_swap_cache_async() and it essentially does something similar.

Instead of returning, it keeps retrying until it finds that swapcache_prepare() fails for another reason than -EEXISTS (e.g., freed concurrently) or it finds the entry in the swapcache.

So if you would succeed here on a freed+reused swap entry, __read_swap_cache_async() would simply retry.

It spells that out:

/*
* We might race against __delete_from_swap_cache(), and
* stumble across a swap_map entry whose SWAP_HAS_CACHE
* has not yet been cleared. Or race against another
* __read_swap_cache_async(), which has set SWAP_HAS_CACHE
* in swap_map, but not yet added its folio to swap cache. */

Whereby we could not race against this code here as well where we speculatively set SWAP_HAS_CACHE and might never add something to the swap cache.

I'd probably avoid the wrong returns and do something even closer to __read_swap_cache_async().

while (true) {
	/*
	 * Fake that we are trying to insert a page into the swapcache, to
	 * serialize against concurrent threads wanting to do the same.
	 * [more from your description]
	 */
	ret = swapcache_prepare(entry);
	if (likely(!ret)
		/*
		 * Move forward with swapin, we'll recheck if the PTE hasn't
		 * changed later.
		 */
		break;
	else if (ret != -EEXIST)
		goto out;
	
	/*
         * See __read_swap_cache_async(). We might either have raced against
         * another thread, or the entry could have been freed and reused in the
	 * meantime. Make sure that the PTE did not change, to detect freeing.
	 */
	vmf->pte = pte_offset_map_lock(vma->vm_mm, vmf->pmd,
				       vmf->address, &vmf->ptl);
	if (!vmf->pte || !pte_same(ptep_get(vmf->pte), vmf->orig_pte))
		goto unlock;


	schedule();
}

I was skeptical about the schedule(), but __read_swap_cache_async() does it already because there is no better way to wait for the event to happen.

With something like above you would no longer depend on the speed of schedule() to determine how often you would retry the fault, which would likely make sense.

I do wonder about the schedule() vs. schedule_timeout_uninterruptible(), though. No expert on that area, do you have any idea?

Signature · 3 lines

Cheers,

David / dhildenb

David Hildenbrand

Acked-by

2024-02-16 18:01 UTC #5 lore raw

Quoted 221 lines · David Hildenbrand

Quoted 149 lines · Kairui Song
From: Kairui Song <[email protected]>
When skipping swapcache for SWP_SYNCHRONOUS_IO, if two or more threads swapin the same entry at the same time, they get different pages (A, B). Before one thread (T0) finishes the swapin and installs page (A) to the PTE, another thread (T1) could finish swapin of page (B), swap_free the entry, then swap out the possibly modified page reusing the same entry. It breaks the pte_same check in (T0) because PTE value is unchanged, causing ABA problem. Thread (T0) will install a stalled page (A) into the PTE and cause data corruption.
One possible callstack is like this:
CPU0                                 CPU1
----                                 ----
do_swap_page()                       do_swap_page() with same entry
<direct swapin path>                 <direct swapin path>
<alloc page A>                       <alloc page B>
swap_read_folio() <- read to page A  swap_read_folio() <- read to page B
<slow on later locks or interrupt>   <finished swapin first>
...                                  set_pte_at()
swap_free() <- entry is free <write to page B, now page A stalled> <swap out page B to same swap entry>
pte_same() <- Check pass, PTE seems
unchanged, but page A
is stalled!
swap_free() <- page B content lost!
set_pte_at() <- staled page A installed!
And besides, for ZRAM, swap_free() allows the swap device to discard the entry content, so even if page (B) is not modified, if swap_read_folio() on CPU0 happens later than swap_free() on CPU1, it may also cause data loss.
To fix this, reuse swapcache_prepare which will pin the swap entry using the cache flag, and allow only one thread to pin it. Release the pin after PT unlocked. Racers will simply wait since it's a rare and very short event. A schedule() call is added to avoid wasting too much CPU or adding too much noise to perf statistics
Other methods like increasing the swap count don't seem to be a good idea after some tests, that will cause racers to fall back to use the swap cache again. Parallel swapin using different methods leads to a much more complex scenario.
Reproducer:
This race issue can be triggered easily using a well constructed reproducer and patched brd (with a delay in read path) [1]:
With latest 6.8 mainline, race caused data loss can be observed easily:
$ gcc -g -lpthread test-thread-swap-race.c && ./a.out
    Polulating 32MB of memory region...
    Keep swapping out...
    Starting round 0...
    Spawning 65536 workers...
    32746 workers spawned, wait for done...
    Round 0: Error on 0x5aa00, expected 32746, got 32743, 3 data loss!
    Round 0: Error on 0x395200, expected 32746, got 32743, 3 data loss!
    Round 0: Error on 0x3fd000, expected 32746, got 32737, 9 data loss!
    Round 0 Failed, 15 data loss!
This reproducer spawns multiple threads sharing the same memory region using a small swap device. Every two threads updates mapped pages one by one in opposite direction trying to create a race, with one dedicated thread keep swapping out the data out using madvise.
The reproducer created a reproduce rate of about once every 5 minutes, so the race should be totally possible in production.
After this patch, I ran the reproducer for over a few hundred rounds and no data loss observed.
Performance overhead is minimal, microbenchmark swapin 10G from 32G zram:
Before:     10934698 us
After:      11157121 us
Non-direct: 13155355 us (Dropping SWP_SYNCHRONOUS_IO flag)
Fixes 0bcac06f27d7 ("mm, swap: skip swapcache for swapin of synchronous device")
Link https://github.com/ryncsn/emm-test-project/tree/master/swap-stress-race [1]
Reported-by "Huang, Ying" <[email protected]>
Closes https://lore.kernel.org/lkml/[email protected]/ open here
Signed-off-by Kairui Song <[email protected]>
Cc [email protected]
---
Update from V2:
- Add a schedule() if raced to prevent repeated page faults wasting CPU and add noise to perf statistics.
- Use a bool to state the special case instead of reusing existing variables fixing error handling [Minchan Kim].
V2: https://lore.kernel.org/all/[email protected]/
Update from V1:
- Add some words on ZRAM case, it will discard swap content on swap_free so the race window is a bit different but cure is the same. [Barry Song]
- Update comments make it cleaner [Huang, Ying]
- Add a function place holder to fix CONFIG_SWAP=n built [SeongJae Park]
- Update the commit message and summary, refer to SWP_SYNCHRONOUS_IO instead of "direct swapin path" [Yu Zhao]
- Update commit message.
- Collect Review and Acks.
V1: https://lore.kernel.org/all/[email protected]/
   include/linux/swap.h |  5 +++++
   mm/memory.c          | 20 ++++++++++++++++++++
   mm/swap.h            |  5 +++++
   mm/swapfile.c        | 13 +++++++++++++
   4 files changed, 43 insertions(+)
include/linux/swap.h +5 -0
index 4db00ddad261..8d28f6091a32 100644
@@ -549,6 +549,11 @@ static inline int swap_duplicate(swp_entry_t swp)
   	return 0;
   }
 
+static inline int swapcache_prepare(swp_entry_t swp)
+{
+	return 0;
+}
+
   static inline void swap_free(swp_entry_t swp)
   {
   }
mm/memory.c +5 -0
index 7e1f4849463a..7059230d0a54 100644
@@ -3799,6 +3799,7 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
   	struct page *page;
   	struct swap_info_struct *si = NULL;
   	rmap_t rmap_flags = RMAP_NONE;
+	bool need_clear_cache = false;
   	bool exclusive = false;
   	swp_entry_t entry;
   	pte_t pte;
@@ -3867,6 +3868,20 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
   	if (!folio) {
   		if (data_race(si->flags & SWP_SYNCHRONOUS_IO) &&
   		    __swap_count(entry) == 1) {
+			/*
+			 * Prevent parallel swapin from proceeding with
+			 * the cache flag. Otherwise, another thread may
+			 * finish swapin first, free the entry, and swapout
mm/memory.c vm_fault_t do_swap_page(struct vm_fault *vmf) · line 3884
Quoted 10 lines
mm/memory.c +10 -0
+			 * reusing the same entry. It's undetectable as
+			 * pte_same() returns true due to entry reuse.
+			 */
+			if (swapcache_prepare(entry)) {
+				/* Relax a bit to prevent rapid repeated page faults */
+				schedule();
+				goto out;
+			}
+			need_clear_cache = true;
+
I took a closer look at __read_swap_cache_async() and it essentially does something similar.
Instead of returning, it keeps retrying until it finds that swapcache_prepare() fails for another reason than -EEXISTS (e.g., freed concurrently) or it finds the entry in the swapcache.
So if you would succeed here on a freed+reused swap entry, __read_swap_cache_async() would simply retry.
It spells that out:
/*
* We might race against __delete_from_swap_cache(), and
* stumble across a swap_map entry whose SWAP_HAS_CACHE
* has not yet been cleared. Or race against another
* __read_swap_cache_async(), which has set SWAP_HAS_CACHE
* in swap_map, but not yet added its folio to swap cache. */
Whereby we could not race against this code here as well where we speculatively set SWAP_HAS_CACHE and might never add something to the swap cache.
I'd probably avoid the wrong returns and do something even closer to __read_swap_cache_async().
while (true) {
	/*
	 * Fake that we are trying to insert a page into the swapcache, to
	 * serialize against concurrent threads wanting to do the same.
	 * [more from your description]
	 */
	ret = swapcache_prepare(entry);
	if (likely(!ret)
		/*
		 * Move forward with swapin, we'll recheck if the PTE hasn't
		 * changed later.
		 */
		break;
	else if (ret != -EEXIST)
		goto out;
	
	/*
           * See __read_swap_cache_async(). We might either have raced against
           * another thread, or the entry could have been freed and reused in the
	 * meantime. Make sure that the PTE did not change, to detect freeing.
	 */
	vmf->pte = pte_offset_map_lock(vma->vm_mm, vmf->pmd,
				       vmf->address, &vmf->ptl);
	if (!vmf->pte || !pte_same(ptep_get(vmf->pte), vmf->orig_pte))
		goto unlock;


	schedule();
}

Quoted 9 lines

I was skeptical about the schedule(), but __read_swap_cache_async() does it already because there is no better way to wait for the event to happen.
With something like above you would no longer depend on the speed of schedule() to determine how often you would retry the fault, which would likely make sense.
I do wonder about the schedule() vs. schedule_timeout_uninterruptible(), though. No expert on that area, do you have any idea?

Forgot to add

Acked-by David Hildenbrand <[email protected]>

But I suspect we do not want to not rely on schedule() to actually sleep, and instead keep retrying until the other thread finished, similar to above.

Signature · 3 lines

Cheers,

David / dhildenb

Kairui Song

author

2024-02-18 08:29 UTC #6 lore raw

Quoted 153 lines · David Hildenbrand

Quoted 152 lines · David Hildenbrand
Quoted 151 lines · Kairui Song
From: Kairui Song <[email protected]>
When skipping swapcache for SWP_SYNCHRONOUS_IO, if two or more threads swapin the same entry at the same time, they get different pages (A, B). Before one thread (T0) finishes the swapin and installs page (A) to the PTE, another thread (T1) could finish swapin of page (B), swap_free the entry, then swap out the possibly modified page reusing the same entry. It breaks the pte_same check in (T0) because PTE value is unchanged, causing ABA problem. Thread (T0) will install a stalled page (A) into the PTE and cause data corruption.
One possible callstack is like this:
CPU0                                 CPU1
----                                 ----
do_swap_page()                       do_swap_page() with same entry
<direct swapin path>                 <direct swapin path>
<alloc page A>                       <alloc page B>
swap_read_folio() <- read to page A  swap_read_folio() <- read to page B
<slow on later locks or interrupt>   <finished swapin first>
...                                  set_pte_at()
swap_free() <- entry is free <write to page B, now page A stalled> <swap out page B to same swap entry>
pte_same() <- Check pass, PTE seems
unchanged, but page A
is stalled!
swap_free() <- page B content lost!
set_pte_at() <- staled page A installed!
And besides, for ZRAM, swap_free() allows the swap device to discard the entry content, so even if page (B) is not modified, if swap_read_folio() on CPU0 happens later than swap_free() on CPU1, it may also cause data loss.
To fix this, reuse swapcache_prepare which will pin the swap entry using the cache flag, and allow only one thread to pin it. Release the pin after PT unlocked. Racers will simply wait since it's a rare and very short event. A schedule() call is added to avoid wasting too much CPU or adding too much noise to perf statistics
Other methods like increasing the swap count don't seem to be a good idea after some tests, that will cause racers to fall back to use the swap cache again. Parallel swapin using different methods leads to a much more complex scenario.
Reproducer:
This race issue can be triggered easily using a well constructed reproducer and patched brd (with a delay in read path) [1]:
With latest 6.8 mainline, race caused data loss can be observed easily:
$ gcc -g -lpthread test-thread-swap-race.c && ./a.out
    Polulating 32MB of memory region...
    Keep swapping out...
    Starting round 0...
    Spawning 65536 workers...
    32746 workers spawned, wait for done...
    Round 0: Error on 0x5aa00, expected 32746, got 32743, 3 data loss!
    Round 0: Error on 0x395200, expected 32746, got 32743, 3 data loss!
    Round 0: Error on 0x3fd000, expected 32746, got 32737, 9 data loss!
    Round 0 Failed, 15 data loss!
This reproducer spawns multiple threads sharing the same memory region using a small swap device. Every two threads updates mapped pages one by one in opposite direction trying to create a race, with one dedicated thread keep swapping out the data out using madvise.
The reproducer created a reproduce rate of about once every 5 minutes, so the race should be totally possible in production.
After this patch, I ran the reproducer for over a few hundred rounds and no data loss observed.
Performance overhead is minimal, microbenchmark swapin 10G from 32G zram:
Before:     10934698 us
After:      11157121 us
Non-direct: 13155355 us (Dropping SWP_SYNCHRONOUS_IO flag)
Fixes 0bcac06f27d7 ("mm, swap: skip swapcache for swapin of synchronous device")
Link https://github.com/ryncsn/emm-test-project/tree/master/swap-stress-race [1]
Reported-by "Huang, Ying" <[email protected]>
Closes https://lore.kernel.org/lkml/[email protected]/ open here
Signed-off-by Kairui Song <[email protected]>
Cc [email protected]
---
Update from V2:
- Add a schedule() if raced to prevent repeated page faults wasting CPU and add noise to perf statistics.
- Use a bool to state the special case instead of reusing existing variables fixing error handling [Minchan Kim].
V2: https://lore.kernel.org/all/[email protected]/
Update from V1:
- Add some words on ZRAM case, it will discard swap content on swap_free so the race window is a bit different but cure is the same. [Barry Song]
- Update comments make it cleaner [Huang, Ying]
- Add a function place holder to fix CONFIG_SWAP=n built [SeongJae Park]
- Update the commit message and summary, refer to SWP_SYNCHRONOUS_IO instead of "direct swapin path" [Yu Zhao]
- Update commit message.
- Collect Review and Acks.
V1: https://lore.kernel.org/all/[email protected]/
   include/linux/swap.h |  5 +++++
   mm/memory.c          | 20 ++++++++++++++++++++
   mm/swap.h            |  5 +++++
   mm/swapfile.c        | 13 +++++++++++++
   4 files changed, 43 insertions(+)
include/linux/swap.h +5 -0
index 4db00ddad261..8d28f6091a32 100644
@@ -549,6 +549,11 @@ static inline int swap_duplicate(swp_entry_t swp)
     return 0;
   }
 
+static inline int swapcache_prepare(swp_entry_t swp)
+{
+    return 0;
+}
+
   static inline void swap_free(swp_entry_t swp)
   {
   }
mm/memory.c +7 -0
index 7e1f4849463a..7059230d0a54 100644
@@ -3799,6 +3799,7 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
     struct page *page;
     struct swap_info_struct *si = NULL;
     rmap_t rmap_flags = RMAP_NONE;
+    bool need_clear_cache = false;
     bool exclusive = false;
     swp_entry_t entry;
     pte_t pte;
@@ -3867,6 +3868,20 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
     if (!folio) {
             if (data_race(si->flags & SWP_SYNCHRONOUS_IO) &&
                 __swap_count(entry) == 1) {
+                    /*
+                     * Prevent parallel swapin from proceeding with
+                     * the cache flag. Otherwise, another thread may
+                     * finish swapin first, free the entry, and swapout
+                     * reusing the same entry. It's undetectable as
+                     * pte_same() returns true due to entry reuse.

Quoted 8 lines

mm/memory.c vm_fault_t do_swap_page(struct vm_fault *vmf) · line 3884

Quoted 8 lines

mm/memory.c +8 -0

+                     */
+                    if (swapcache_prepare(entry)) {
+                            /* Relax a bit to prevent rapid repeated page faults */
+                            schedule();
+                            goto out;
+                    }
+                    need_clear_cache = true;
+

Hi David

Thanks for the review! I saw you added more replies so I'll just post reply on your last mail.

Quoted 56 lines

Quoted 56 lines
I took a closer look at __read_swap_cache_async() and it essentially does something similar.
Instead of returning, it keeps retrying until it finds that swapcache_prepare() fails for another reason than -EEXISTS (e.g., freed concurrently) or it finds the entry in the swapcache.
So if you would succeed here on a freed+reused swap entry, __read_swap_cache_async() would simply retry.
It spells that out:
/*
* We might race against __delete_from_swap_cache(), and
* stumble across a swap_map entry whose SWAP_HAS_CACHE
* has not yet been cleared. Or race against another
* __read_swap_cache_async(), which has set SWAP_HAS_CACHE
* in swap_map, but not yet added its folio to swap cache. */
Whereby we could not race against this code here as well where we speculatively set SWAP_HAS_CACHE and might never add something to the swap cache.
I'd probably avoid the wrong returns and do something even closer to __read_swap_cache_async().
while (true) {
      /*
       * Fake that we are trying to insert a page into the swapcache, to
       * serialize against concurrent threads wanting to do the same.
       * [more from your description]
       */
      ret = swapcache_prepare(entry);
      if (likely(!ret)
              /*
               * Move forward with swapin, we'll recheck if the PTE hasn't
               * changed later.
               */
              break;
      else if (ret != -EEXIST)
              goto out;

      /*
           * See __read_swap_cache_async(). We might either have raced against
           * another thread, or the entry could have been freed and reused in the
       * meantime. Make sure that the PTE did not change, to detect freeing.
       */
      vmf->pte = pte_offset_map_lock(vma->vm_mm, vmf->pmd,
                                     vmf->address, &vmf->ptl);
      if (!vmf->pte || !pte_same(ptep_get(vmf->pte), vmf->orig_pte))
              goto unlock;

Quoted 2 lines

      schedule();
}

It was discussed earlier about looping in the page fault. One issue is about swap entry may stuck in swapcache for a long time, so at lease an extra cache loop up is need. To be safe we need to implement a similar loop like the one in mm/swap_state.c, which I doubt is necessary...

Quoted 10 lines

Quoted 10 lines
I was skeptical about the schedule(), but __read_swap_cache_async() does it already because there is no better way to wait for the event to happen.
With something like above you would no longer depend on the speed of schedule() to determine how often you would retry the fault, which would likely make sense.
I do wonder about the schedule() vs. schedule_timeout_uninterruptible(), though. No expert on that area, do you have any idea?

schedule_timeout_uninterruptible seems more reasonable here from its name (delay a bit to wait). My idea here is that SWP_SYNCHRONOUS_IO devices are supposed to be super fast, so usually a second try will just work (when tested with a less stressed test case and that seems to be always true), and the race itself is rare enough to be ignore for 7 years.

But when system is really stressed (eg. the reproducer I provided), it may take longer to finish (SWP_SYNCHRONOUS_IO devices are CPU bound). So a schedule() can help to avoid one task from looping page fault, for better statistic and CPU usage.

Previous test results:
https://lore.kernel.org/lkml/CAMgjq7BvTJmxrWQOJvkLt4g_jnvmx07NdU63sGeRMGde4Ov=gA@mail.gmail.com/
It showed schedule() works fine here.

Huang, Ying

2024-02-18 08:01 UTC #7 lore raw

In reply to David Hildenbrand

Quoted 175 lines · David Hildenbrand

Quoted 135 lines · Kairui Song
From: Kairui Song <[email protected]>
When skipping swapcache for SWP_SYNCHRONOUS_IO, if two or more threads
swapin the same entry at the same time, they get different pages (A, B). Before one thread (T0) finishes the swapin and installs page (A) to the PTE, another thread (T1) could finish swapin of page (B), swap_free the entry, then swap out the possibly modified page reusing the same entry. It breaks the pte_same check in (T0) because PTE value is unchanged, causing ABA problem. Thread (T0) will install a stalled page (A) into the PTE and cause data corruption. One possible callstack is like this:
CPU0                                 CPU1
----                                 ----
do_swap_page()                       do_swap_page() with same entry
<direct swapin path>                 <direct swapin path>
<alloc page A>                       <alloc page B>
swap_read_folio() <- read to page A  swap_read_folio() <- read to page B
<slow on later locks or interrupt>   <finished swapin first>
...                                  set_pte_at()
swap_free() <- entry is free <write to page B, now page A stalled> <swap out page B to same swap entry>
pte_same() <- Check pass, PTE seems
unchanged, but page A
is stalled!
swap_free() <- page B content lost!
set_pte_at() <- staled page A installed!
And besides, for ZRAM, swap_free() allows the swap device to discard the entry content, so even if page (B) is not modified, if swap_read_folio() on CPU0 happens later than swap_free() on CPU1, it may also cause data loss.
To fix this, reuse swapcache_prepare which will pin the swap entry using
the cache flag, and allow only one thread to pin it. Release the pin after PT unlocked. Racers will simply wait since it's a rare and very short event. A schedule() call is added to avoid wasting too much CPU or adding too much noise to perf statistics
Other methods like increasing the swap count don't seem to be a good idea after some tests, that will cause racers to fall back to use the swap cache again. Parallel swapin using different methods leads to a much more complex scenario.
Reproducer:
This race issue can be triggered easily using a well constructed reproducer and patched brd (with a delay in read path) [1]:
With latest 6.8 mainline, race caused data loss can be observed easily:
$ gcc -g -lpthread test-thread-swap-race.c && ./a.out
   Polulating 32MB of memory region...
   Keep swapping out...
   Starting round 0...
   Spawning 65536 workers...
   32746 workers spawned, wait for done...
   Round 0: Error on 0x5aa00, expected 32746, got 32743, 3 data loss!
   Round 0: Error on 0x395200, expected 32746, got 32743, 3 data loss!
   Round 0: Error on 0x3fd000, expected 32746, got 32737, 9 data loss!
   Round 0 Failed, 15 data loss!
This reproducer spawns multiple threads sharing the same memory
region
using a small swap device. Every two threads updates mapped pages one by
one in opposite direction trying to create a race, with one dedicated
thread keep swapping out the data out using madvise.
The reproducer created a reproduce rate of about once every 5
minutes,
so the race should be totally possible in production.
After this patch, I ran the reproducer for over a few hundred rounds
and no data loss observed.
Performance overhead is minimal, microbenchmark swapin 10G from 32G
zram:
Before:     10934698 us
After:      11157121 us
Non-direct: 13155355 us (Dropping SWP_SYNCHRONOUS_IO flag)
Fixes: 0bcac06f27d7 ("mm, swap: skip swapcache for swapin of
synchronous device")
Link: https://github.com/ryncsn/emm-test-project/tree/master/swap-stress-race [1]
Reported-by: "Huang, Ying" <[email protected]>
Closes: https://lore.kernel.org/lkml/[email protected]/
Signed-off-by: Kairui Song <[email protected]>
Cc: [email protected]
---
Update from V2:
- Add a schedule() if raced to prevent repeated page faults wasting CPU
   and add noise to perf statistics.
- Use a bool to state the special case instead of reusing existing
   variables fixing error handling [Minchan Kim].
V2:
https://lore.kernel.org/all/[email protected]/
Update from V1:
- Add some words on ZRAM case, it will discard swap content on swap_free so the race window is a bit different but cure is the same. [Barry Song]
- Update comments make it cleaner [Huang, Ying]
- Add a function place holder to fix CONFIG_SWAP=n built [SeongJae Park]
- Update the commit message and summary, refer to SWP_SYNCHRONOUS_IO instead of "direct swapin path" [Yu Zhao]
- Update commit message.
- Collect Review and Acks.
V1:
https://lore.kernel.org/all/[email protected]/
  include/linux/swap.h |  5 +++++
  mm/memory.c          | 20 ++++++++++++++++++++
  mm/swap.h            |  5 +++++
  mm/swapfile.c        | 13 +++++++++++++
  4 files changed, 43 insertions(+)
include/linux/swap.h +4 -0
index 4db00ddad261..8d28f6091a32 100644
@@ -549,6 +549,11 @@ static inline int swap_duplicate(swp_entry_t swp)
  	return 0;
  }
  +static inline int swapcache_prepare(swp_entry_t swp)
+{
+	return 0;
+}
+
  static inline void swap_free(swp_entry_t swp)
  {
  }
mm/memory.c +5 -0
index 7e1f4849463a..7059230d0a54 100644
@@ -3799,6 +3799,7 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
  	struct page *page;
  	struct swap_info_struct *si = NULL;
  	rmap_t rmap_flags = RMAP_NONE;
+	bool need_clear_cache = false;
  	bool exclusive = false;
  	swp_entry_t entry;
  	pte_t pte;
@@ -3867,6 +3868,20 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
  	if (!folio) {
  		if (data_race(si->flags & SWP_SYNCHRONOUS_IO) &&
  		    __swap_count(entry) == 1) {
+			/*
+			 * Prevent parallel swapin from proceeding with
+			 * the cache flag. Otherwise, another thread may
+			 * finish swapin first, free the entry, and swapout
mm/memory.c vm_fault_t do_swap_page(struct vm_fault *vmf) · line 3884
Quoted 10 lines
mm/memory.c +10 -0
+			 * reusing the same entry. It's undetectable as
+			 * pte_same() returns true due to entry reuse.
+			 */
+			if (swapcache_prepare(entry)) {
+				/* Relax a bit to prevent rapid repeated page faults */
+				schedule();
+				goto out;
+			}
+			need_clear_cache = true;
+
I took a closer look at __read_swap_cache_async() and it essentially does something similar.
Instead of returning, it keeps retrying until it finds that swapcache_prepare() fails for another reason than -EEXISTS (e.g., freed concurrently) or it finds the entry in the swapcache.
So if you would succeed here on a freed+reused swap entry, __read_swap_cache_async() would simply retry.
It spells that out:
/*
* We might race against __delete_from_swap_cache(), and
* stumble across a swap_map entry whose SWAP_HAS_CACHE
* has not yet been cleared. Or race against another
* __read_swap_cache_async(), which has set SWAP_HAS_CACHE
* in swap_map, but not yet added its folio to swap cache. */
Whereby we could not race against this code here as well where we speculatively set SWAP_HAS_CACHE and might never add something to the swap cache.
I'd probably avoid the wrong returns and do something even closer to __read_swap_cache_async().

Quoted 15 lines

while (true) {
	/*
	 * Fake that we are trying to insert a page into the swapcache, to
	 * serialize against concurrent threads wanting to do the same.
	 * [more from your description]
	 */
	ret = swapcache_prepare(entry);
	if (likely(!ret)
		/*
		 * Move forward with swapin, we'll recheck if the PTE hasn't
		 * changed later.
		 */
		break;
	else if (ret != -EEXIST)
		goto out;

The swap entry may be kept in swap cache for long time. For example, it may be read into swap cache via MADV_WILLNEED.

--
Best Regards,
Huang, Ying

Quoted 25 lines

	
	/*
         * See __read_swap_cache_async(). We might either have raced against
         * another thread, or the entry could have been freed and reused in the
	 * meantime. Make sure that the PTE did not change, to detect freeing.
	 */
	vmf->pte = pte_offset_map_lock(vma->vm_mm, vmf->pmd,
				       vmf->address, &vmf->ptl);
	if (!vmf->pte || !pte_same(ptep_get(vmf->pte), vmf->orig_pte))
		goto unlock;


	schedule();
}
I was skeptical about the schedule(), but __read_swap_cache_async() does it already because there is no better way to wait for the event to happen.
With something like above you would no longer depend on the speed of schedule() to determine how often you would retry the fault, which would likely make sense.
I do wonder about the schedule() vs. schedule_timeout_uninterruptible(), though. No expert on that area, do you have any idea?

Barry Song

2024-02-18 08:19 UTC #8 lore raw

Quoted 194 lines · Ying

Quoted 175 lines · David Hildenbrand
Quoted 135 lines · Kairui Song
From: Kairui Song <[email protected]>
When skipping swapcache for SWP_SYNCHRONOUS_IO, if two or more threads
swapin the same entry at the same time, they get different pages (A, B). Before one thread (T0) finishes the swapin and installs page (A) to the PTE, another thread (T1) could finish swapin of page (B), swap_free the entry, then swap out the possibly modified page reusing the same entry. It breaks the pte_same check in (T0) because PTE value is unchanged, causing ABA problem. Thread (T0) will install a stalled page (A) into the PTE and cause data corruption. One possible callstack is like this:
CPU0                                 CPU1
----                                 ----
do_swap_page()                       do_swap_page() with same entry
<direct swapin path>                 <direct swapin path>
<alloc page A>                       <alloc page B>
swap_read_folio() <- read to page A  swap_read_folio() <- read to page B
<slow on later locks or interrupt>   <finished swapin first>
...                                  set_pte_at()
swap_free() <- entry is free <write to page B, now page A stalled> <swap out page B to same swap entry>
pte_same() <- Check pass, PTE seems
unchanged, but page A
is stalled!
swap_free() <- page B content lost!
set_pte_at() <- staled page A installed!
And besides, for ZRAM, swap_free() allows the swap device to discard the entry content, so even if page (B) is not modified, if swap_read_folio() on CPU0 happens later than swap_free() on CPU1, it may also cause data loss.
To fix this, reuse swapcache_prepare which will pin the swap entry using
the cache flag, and allow only one thread to pin it. Release the pin after PT unlocked. Racers will simply wait since it's a rare and very short event. A schedule() call is added to avoid wasting too much CPU or adding too much noise to perf statistics
Other methods like increasing the swap count don't seem to be a good idea after some tests, that will cause racers to fall back to use the swap cache again. Parallel swapin using different methods leads to a much more complex scenario.
Reproducer:
This race issue can be triggered easily using a well constructed reproducer and patched brd (with a delay in read path) [1]:
With latest 6.8 mainline, race caused data loss can be observed easily:
$ gcc -g -lpthread test-thread-swap-race.c && ./a.out
   Polulating 32MB of memory region...
   Keep swapping out...
   Starting round 0...
   Spawning 65536 workers...
   32746 workers spawned, wait for done...
   Round 0: Error on 0x5aa00, expected 32746, got 32743, 3 data loss!
   Round 0: Error on 0x395200, expected 32746, got 32743, 3 data loss!
   Round 0: Error on 0x3fd000, expected 32746, got 32737, 9 data loss!
   Round 0 Failed, 15 data loss!
This reproducer spawns multiple threads sharing the same memory
region
using a small swap device. Every two threads updates mapped pages one by
one in opposite direction trying to create a race, with one dedicated
thread keep swapping out the data out using madvise.
The reproducer created a reproduce rate of about once every 5
minutes,
so the race should be totally possible in production.
After this patch, I ran the reproducer for over a few hundred rounds
and no data loss observed.
Performance overhead is minimal, microbenchmark swapin 10G from 32G
zram:
Before:     10934698 us
After:      11157121 us
Non-direct: 13155355 us (Dropping SWP_SYNCHRONOUS_IO flag)
Fixes: 0bcac06f27d7 ("mm, swap: skip swapcache for swapin of
synchronous device")
Link: https://github.com/ryncsn/emm-test-project/tree/master/swap-stress-race [1]
Reported-by: "Huang, Ying" <[email protected]>
Closes: https://lore.kernel.org/lkml/[email protected]/
Signed-off-by: Kairui Song <[email protected]>
Cc: [email protected]
---
Update from V2:
- Add a schedule() if raced to prevent repeated page faults wasting CPU
   and add noise to perf statistics.
- Use a bool to state the special case instead of reusing existing
   variables fixing error handling [Minchan Kim].
V2:
https://lore.kernel.org/all/[email protected]/
Update from V1:
- Add some words on ZRAM case, it will discard swap content on swap_free so the race window is a bit different but cure is the same. [Barry Song]
- Update comments make it cleaner [Huang, Ying]
- Add a function place holder to fix CONFIG_SWAP=n built [SeongJae Park]
- Update the commit message and summary, refer to SWP_SYNCHRONOUS_IO instead of "direct swapin path" [Yu Zhao]
- Update commit message.
- Collect Review and Acks.
V1:
https://lore.kernel.org/all/[email protected]/
  include/linux/swap.h |  5 +++++
  mm/memory.c          | 20 ++++++++++++++++++++
  mm/swap.h            |  5 +++++
  mm/swapfile.c        | 13 +++++++++++++
  4 files changed, 43 insertions(+)
include/linux/swap.h +4 -0
index 4db00ddad261..8d28f6091a32 100644
@@ -549,6 +549,11 @@ static inline int swap_duplicate(swp_entry_t swp)
     return 0;
  }
  +static inline int swapcache_prepare(swp_entry_t swp)
+{
+    return 0;
+}
+
  static inline void swap_free(swp_entry_t swp)
  {
  }
mm/memory.c +5 -0
index 7e1f4849463a..7059230d0a54 100644
@@ -3799,6 +3799,7 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
     struct page *page;
     struct swap_info_struct *si = NULL;
     rmap_t rmap_flags = RMAP_NONE;
+    bool need_clear_cache = false;
     bool exclusive = false;
     swp_entry_t entry;
     pte_t pte;
@@ -3867,6 +3868,20 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
     if (!folio) {
             if (data_race(si->flags & SWP_SYNCHRONOUS_IO) &&
                 __swap_count(entry) == 1) {
+                    /*
+                     * Prevent parallel swapin from proceeding with
+                     * the cache flag. Otherwise, another thread may
+                     * finish swapin first, free the entry, and swapout
mm/memory.c vm_fault_t do_swap_page(struct vm_fault *vmf) · line 3884
Quoted 10 lines
mm/memory.c +10 -0
+                     * reusing the same entry. It's undetectable as
+                     * pte_same() returns true due to entry reuse.
+                     */
+                    if (swapcache_prepare(entry)) {
+                            /* Relax a bit to prevent rapid repeated page faults */
+                            schedule();
+                            goto out;
+                    }
+                    need_clear_cache = true;
+
I took a closer look at __read_swap_cache_async() and it essentially does something similar.
Instead of returning, it keeps retrying until it finds that swapcache_prepare() fails for another reason than -EEXISTS (e.g., freed concurrently) or it finds the entry in the swapcache.
So if you would succeed here on a freed+reused swap entry, __read_swap_cache_async() would simply retry.
It spells that out:
/*
* We might race against __delete_from_swap_cache(), and
* stumble across a swap_map entry whose SWAP_HAS_CACHE
* has not yet been cleared. Or race against another
* __read_swap_cache_async(), which has set SWAP_HAS_CACHE
* in swap_map, but not yet added its folio to swap cache. */
Whereby we could not race against this code here as well where we speculatively set SWAP_HAS_CACHE and might never add something to the swap cache.
I'd probably avoid the wrong returns and do something even closer to __read_swap_cache_async().
Quoted 15 lines
while (true) {
      /*
       * Fake that we are trying to insert a page into the swapcache, to
       * serialize against concurrent threads wanting to do the same.
       * [more from your description]
       */
      ret = swapcache_prepare(entry);
      if (likely(!ret)
              /*
               * Move forward with swapin, we'll recheck if the PTE hasn't
               * changed later.
               */
              break;
      else if (ret != -EEXIST)
              goto out;

Quoted 2 lines

The swap entry may be kept in swap cache for long time. For example, it may be read into swap cache via MADV_WILLNEED.

This seems fine.

if swapcache has data from WILLNEED, the new page fault will hit it. Thus, we won't go into the SYNC_IO path any more?

Quoted 31 lines

--
Best Regards,
Huang, Ying
Quoted 25 lines
      /*
         * See __read_swap_cache_async(). We might either have raced against
         * another thread, or the entry could have been freed and reused in the
       * meantime. Make sure that the PTE did not change, to detect freeing.
       */
      vmf->pte = pte_offset_map_lock(vma->vm_mm, vmf->pmd,
                                     vmf->address, &vmf->ptl);
      if (!vmf->pte || !pte_same(ptep_get(vmf->pte), vmf->orig_pte))
              goto unlock;


      schedule();
}
I was skeptical about the schedule(), but __read_swap_cache_async() does it already because there is no better way to wait for the event to happen.
With something like above you would no longer depend on the speed of schedule() to determine how often you would retry the fault, which would likely make sense.
I do wonder about the schedule() vs. schedule_timeout_uninterruptible(), though. No expert on that area, do you have any idea?

Huang, Ying

2024-02-18 08:41 UTC #9 lore raw

Quoted 195 lines · Barry Song

Quoted 194 lines · Ying
Quoted 190 lines · David Hildenbrand
Quoted 135 lines · Kairui Song
From: Kairui Song <[email protected]>
When skipping swapcache for SWP_SYNCHRONOUS_IO, if two or more threads
swapin the same entry at the same time, they get different pages (A, B). Before one thread (T0) finishes the swapin and installs page (A) to the PTE, another thread (T1) could finish swapin of page (B), swap_free the entry, then swap out the possibly modified page reusing the same entry. It breaks the pte_same check in (T0) because PTE value is unchanged, causing ABA problem. Thread (T0) will install a stalled page (A) into the PTE and cause data corruption. One possible callstack is like this:
CPU0                                 CPU1
----                                 ----
do_swap_page()                       do_swap_page() with same entry
<direct swapin path>                 <direct swapin path>
<alloc page A>                       <alloc page B>
swap_read_folio() <- read to page A  swap_read_folio() <- read to page B
<slow on later locks or interrupt>   <finished swapin first>
...                                  set_pte_at()
swap_free() <- entry is free <write to page B, now page A stalled> <swap out page B to same swap entry>
pte_same() <- Check pass, PTE seems
unchanged, but page A
is stalled!
swap_free() <- page B content lost!
set_pte_at() <- staled page A installed!
And besides, for ZRAM, swap_free() allows the swap device to discard the entry content, so even if page (B) is not modified, if swap_read_folio() on CPU0 happens later than swap_free() on CPU1, it may also cause data loss.
To fix this, reuse swapcache_prepare which will pin the swap entry using
the cache flag, and allow only one thread to pin it. Release the pin after PT unlocked. Racers will simply wait since it's a rare and very short event. A schedule() call is added to avoid wasting too much CPU or adding too much noise to perf statistics
Other methods like increasing the swap count don't seem to be a good idea after some tests, that will cause racers to fall back to use the swap cache again. Parallel swapin using different methods leads to a much more complex scenario.
Reproducer:
This race issue can be triggered easily using a well constructed reproducer and patched brd (with a delay in read path) [1]:
With latest 6.8 mainline, race caused data loss can be observed easily:
$ gcc -g -lpthread test-thread-swap-race.c && ./a.out
   Polulating 32MB of memory region...
   Keep swapping out...
   Starting round 0...
   Spawning 65536 workers...
   32746 workers spawned, wait for done...
   Round 0: Error on 0x5aa00, expected 32746, got 32743, 3 data loss!
   Round 0: Error on 0x395200, expected 32746, got 32743, 3 data loss!
   Round 0: Error on 0x3fd000, expected 32746, got 32737, 9 data loss!
   Round 0 Failed, 15 data loss!
This reproducer spawns multiple threads sharing the same memory
region
using a small swap device. Every two threads updates mapped pages one by
one in opposite direction trying to create a race, with one dedicated
thread keep swapping out the data out using madvise.
The reproducer created a reproduce rate of about once every 5
minutes,
so the race should be totally possible in production.
After this patch, I ran the reproducer for over a few hundred rounds
and no data loss observed.
Performance overhead is minimal, microbenchmark swapin 10G from 32G
zram:
Before:     10934698 us
After:      11157121 us
Non-direct: 13155355 us (Dropping SWP_SYNCHRONOUS_IO flag)
Fixes: 0bcac06f27d7 ("mm, swap: skip swapcache for swapin of
synchronous device")
Link: https://github.com/ryncsn/emm-test-project/tree/master/swap-stress-race [1]
Reported-by: "Huang, Ying" <[email protected]>
Closes: https://lore.kernel.org/lkml/[email protected]/
Signed-off-by: Kairui Song <[email protected]>
Cc: [email protected]
---
Update from V2:
- Add a schedule() if raced to prevent repeated page faults wasting CPU
   and add noise to perf statistics.
- Use a bool to state the special case instead of reusing existing
   variables fixing error handling [Minchan Kim].
V2:
https://lore.kernel.org/all/[email protected]/
Update from V1:
- Add some words on ZRAM case, it will discard swap content on swap_free so the race window is a bit different but cure is the same. [Barry Song]
- Update comments make it cleaner [Huang, Ying]
- Add a function place holder to fix CONFIG_SWAP=n built [SeongJae Park]
- Update the commit message and summary, refer to SWP_SYNCHRONOUS_IO instead of "direct swapin path" [Yu Zhao]
- Update commit message.
- Collect Review and Acks.
V1:
https://lore.kernel.org/all/[email protected]/
  include/linux/swap.h |  5 +++++
  mm/memory.c          | 20 ++++++++++++++++++++
  mm/swap.h            |  5 +++++
  mm/swapfile.c        | 13 +++++++++++++
  4 files changed, 43 insertions(+)
include/linux/swap.h +4 -0
index 4db00ddad261..8d28f6091a32 100644
@@ -549,6 +549,11 @@ static inline int swap_duplicate(swp_entry_t swp)
     return 0;
  }
  +static inline int swapcache_prepare(swp_entry_t swp)
+{
+    return 0;
+}
+
  static inline void swap_free(swp_entry_t swp)
  {
  }
mm/memory.c +5 -0
index 7e1f4849463a..7059230d0a54 100644
@@ -3799,6 +3799,7 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
     struct page *page;
     struct swap_info_struct *si = NULL;
     rmap_t rmap_flags = RMAP_NONE;
+    bool need_clear_cache = false;
     bool exclusive = false;
     swp_entry_t entry;
     pte_t pte;
@@ -3867,6 +3868,20 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
     if (!folio) {
             if (data_race(si->flags & SWP_SYNCHRONOUS_IO) &&
                 __swap_count(entry) == 1) {
+                    /*
+                     * Prevent parallel swapin from proceeding with
+                     * the cache flag. Otherwise, another thread may
+                     * finish swapin first, free the entry, and swapout
mm/memory.c vm_fault_t do_swap_page(struct vm_fault *vmf) · line 3884
Quoted 10 lines
mm/memory.c +10 -0
+                     * reusing the same entry. It's undetectable as
+                     * pte_same() returns true due to entry reuse.
+                     */
+                    if (swapcache_prepare(entry)) {
+                            /* Relax a bit to prevent rapid repeated page faults */
+                            schedule();
+                            goto out;
+                    }
+                    need_clear_cache = true;
+
I took a closer look at __read_swap_cache_async() and it essentially does something similar.
Instead of returning, it keeps retrying until it finds that swapcache_prepare() fails for another reason than -EEXISTS (e.g., freed concurrently) or it finds the entry in the swapcache.
So if you would succeed here on a freed+reused swap entry, __read_swap_cache_async() would simply retry.
It spells that out:
/*
* We might race against __delete_from_swap_cache(), and
* stumble across a swap_map entry whose SWAP_HAS_CACHE
* has not yet been cleared. Or race against another
* __read_swap_cache_async(), which has set SWAP_HAS_CACHE
* in swap_map, but not yet added its folio to swap cache. */
Whereby we could not race against this code here as well where we speculatively set SWAP_HAS_CACHE and might never add something to the swap cache.
I'd probably avoid the wrong returns and do something even closer to __read_swap_cache_async().
while (true) {
      /*
       * Fake that we are trying to insert a page into the swapcache, to
       * serialize against concurrent threads wanting to do the same.
       * [more from your description]
       */
      ret = swapcache_prepare(entry);
      if (likely(!ret)
              /*
               * Move forward with swapin, we'll recheck if the PTE hasn't
               * changed later.
               */
              break;
      else if (ret != -EEXIST)
              goto out;

Quoted 7 lines

Quoted 2 lines
The swap entry may be kept in swap cache for long time. For example, it may be read into swap cache via MADV_WILLNEED.
This seems fine.
if swapcache has data from WILLNEED, the new page fault will hit it. Thus, we won't go into the SYNC_IO path any more?

They may happen in parallel. That is, one task is busy looping, while another task read the swap entry into swap cache.

--
Best Regards,
Huang, Ying

Quoted 31 lines

Quoted 31 lines
--
Best Regards,
Huang, Ying
Quoted 25 lines
      /*
         * See __read_swap_cache_async(). We might either have raced against
         * another thread, or the entry could have been freed and reused in the
       * meantime. Make sure that the PTE did not change, to detect freeing.
       */
      vmf->pte = pte_offset_map_lock(vma->vm_mm, vmf->pmd,
                                     vmf->address, &vmf->ptl);
      if (!vmf->pte || !pte_same(ptep_get(vmf->pte), vmf->orig_pte))
              goto unlock;


      schedule();
}
I was skeptical about the schedule(), but __read_swap_cache_async() does it already because there is no better way to wait for the event to happen.
With something like above you would no longer depend on the speed of schedule() to determine how often you would retry the fault, which would likely make sense.
I do wonder about the schedule() vs. schedule_timeout_uninterruptible(), though. No expert on that area, do you have any idea?

Barry Song

2024-02-18 08:47 UTC #10 lore raw

Quoted 201 lines · Ying

Quoted 198 lines · Barry Song
Quoted 196 lines · Ying
Quoted 175 lines · David Hildenbrand
Quoted 135 lines · Kairui Song
From: Kairui Song <[email protected]>
When skipping swapcache for SWP_SYNCHRONOUS_IO, if two or more threads
swapin the same entry at the same time, they get different pages (A, B). Before one thread (T0) finishes the swapin and installs page (A) to the PTE, another thread (T1) could finish swapin of page (B), swap_free the entry, then swap out the possibly modified page reusing the same entry. It breaks the pte_same check in (T0) because PTE value is unchanged, causing ABA problem. Thread (T0) will install a stalled page (A) into the PTE and cause data corruption. One possible callstack is like this:
CPU0                                 CPU1
----                                 ----
do_swap_page()                       do_swap_page() with same entry
<direct swapin path>                 <direct swapin path>
<alloc page A>                       <alloc page B>
swap_read_folio() <- read to page A  swap_read_folio() <- read to page B
<slow on later locks or interrupt>   <finished swapin first>
...                                  set_pte_at()
swap_free() <- entry is free <write to page B, now page A stalled> <swap out page B to same swap entry>
pte_same() <- Check pass, PTE seems
unchanged, but page A
is stalled!
swap_free() <- page B content lost!
set_pte_at() <- staled page A installed!
And besides, for ZRAM, swap_free() allows the swap device to discard the entry content, so even if page (B) is not modified, if swap_read_folio() on CPU0 happens later than swap_free() on CPU1, it may also cause data loss.
To fix this, reuse swapcache_prepare which will pin the swap entry using
the cache flag, and allow only one thread to pin it. Release the pin after PT unlocked. Racers will simply wait since it's a rare and very short event. A schedule() call is added to avoid wasting too much CPU or adding too much noise to perf statistics
Other methods like increasing the swap count don't seem to be a good idea after some tests, that will cause racers to fall back to use the swap cache again. Parallel swapin using different methods leads to a much more complex scenario.
Reproducer:
This race issue can be triggered easily using a well constructed reproducer and patched brd (with a delay in read path) [1]:
With latest 6.8 mainline, race caused data loss can be observed easily:
$ gcc -g -lpthread test-thread-swap-race.c && ./a.out
   Polulating 32MB of memory region...
   Keep swapping out...
   Starting round 0...
   Spawning 65536 workers...
   32746 workers spawned, wait for done...
   Round 0: Error on 0x5aa00, expected 32746, got 32743, 3 data loss!
   Round 0: Error on 0x395200, expected 32746, got 32743, 3 data loss!
   Round 0: Error on 0x3fd000, expected 32746, got 32737, 9 data loss!
   Round 0 Failed, 15 data loss!
This reproducer spawns multiple threads sharing the same memory
region
using a small swap device. Every two threads updates mapped pages one by
one in opposite direction trying to create a race, with one dedicated
thread keep swapping out the data out using madvise.
The reproducer created a reproduce rate of about once every 5
minutes,
so the race should be totally possible in production.
After this patch, I ran the reproducer for over a few hundred rounds
and no data loss observed.
Performance overhead is minimal, microbenchmark swapin 10G from 32G
zram:
Before:     10934698 us
After:      11157121 us
Non-direct: 13155355 us (Dropping SWP_SYNCHRONOUS_IO flag)
Fixes: 0bcac06f27d7 ("mm, swap: skip swapcache for swapin of
synchronous device")
Link: https://github.com/ryncsn/emm-test-project/tree/master/swap-stress-race [1]
Reported-by: "Huang, Ying" <[email protected]>
Closes: https://lore.kernel.org/lkml/[email protected]/
Signed-off-by: Kairui Song <[email protected]>
Cc: [email protected]
---
Update from V2:
- Add a schedule() if raced to prevent repeated page faults wasting CPU
   and add noise to perf statistics.
- Use a bool to state the special case instead of reusing existing
   variables fixing error handling [Minchan Kim].
V2:
https://lore.kernel.org/all/[email protected]/
Update from V1:
- Add some words on ZRAM case, it will discard swap content on swap_free so the race window is a bit different but cure is the same. [Barry Song]
- Update comments make it cleaner [Huang, Ying]
- Add a function place holder to fix CONFIG_SWAP=n built [SeongJae Park]
- Update the commit message and summary, refer to SWP_SYNCHRONOUS_IO instead of "direct swapin path" [Yu Zhao]
- Update commit message.
- Collect Review and Acks.
V1:
https://lore.kernel.org/all/[email protected]/
  include/linux/swap.h |  5 +++++
  mm/memory.c          | 20 ++++++++++++++++++++
  mm/swap.h            |  5 +++++
  mm/swapfile.c        | 13 +++++++++++++
  4 files changed, 43 insertions(+)
include/linux/swap.h +4 -0
index 4db00ddad261..8d28f6091a32 100644
@@ -549,6 +549,11 @@ static inline int swap_duplicate(swp_entry_t swp)
     return 0;
  }
  +static inline int swapcache_prepare(swp_entry_t swp)
+{
+    return 0;
+}
+
  static inline void swap_free(swp_entry_t swp)
  {
  }
mm/memory.c +5 -0
index 7e1f4849463a..7059230d0a54 100644
@@ -3799,6 +3799,7 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
     struct page *page;
     struct swap_info_struct *si = NULL;
     rmap_t rmap_flags = RMAP_NONE;
+    bool need_clear_cache = false;
     bool exclusive = false;
     swp_entry_t entry;
     pte_t pte;
@@ -3867,6 +3868,20 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
     if (!folio) {
             if (data_race(si->flags & SWP_SYNCHRONOUS_IO) &&
                 __swap_count(entry) == 1) {
+                    /*
+                     * Prevent parallel swapin from proceeding with
+                     * the cache flag. Otherwise, another thread may
+                     * finish swapin first, free the entry, and swapout
mm/memory.c vm_fault_t do_swap_page(struct vm_fault *vmf) · line 3884
Quoted 10 lines
mm/memory.c +10 -0
+                     * reusing the same entry. It's undetectable as
+                     * pte_same() returns true due to entry reuse.
+                     */
+                    if (swapcache_prepare(entry)) {
+                            /* Relax a bit to prevent rapid repeated page faults */
+                            schedule();
+                            goto out;
+                    }
+                    need_clear_cache = true;
+
I took a closer look at __read_swap_cache_async() and it essentially does something similar.
Instead of returning, it keeps retrying until it finds that swapcache_prepare() fails for another reason than -EEXISTS (e.g., freed concurrently) or it finds the entry in the swapcache.
So if you would succeed here on a freed+reused swap entry, __read_swap_cache_async() would simply retry.
It spells that out:
/*
* We might race against __delete_from_swap_cache(), and
* stumble across a swap_map entry whose SWAP_HAS_CACHE
* has not yet been cleared. Or race against another
* __read_swap_cache_async(), which has set SWAP_HAS_CACHE
* in swap_map, but not yet added its folio to swap cache. */
Whereby we could not race against this code here as well where we speculatively set SWAP_HAS_CACHE and might never add something to the swap cache.
I'd probably avoid the wrong returns and do something even closer to __read_swap_cache_async().
Quoted 15 lines
while (true) {
      /*
       * Fake that we are trying to insert a page into the swapcache, to
       * serialize against concurrent threads wanting to do the same.
       * [more from your description]
       */
      ret = swapcache_prepare(entry);
      if (likely(!ret)
              /*
               * Move forward with swapin, we'll recheck if the PTE hasn't
               * changed later.
               */
              break;
      else if (ret != -EEXIST)
              goto out;
The swap entry may be kept in swap cache for long time. For example, it may be read into swap cache via MADV_WILLNEED.

Quoted 7 lines

Quoted 4 lines
This seems fine.
if swapcache has data from WILLNEED, the new page fault will hit it. Thus, we won't go into the SYNC_IO path any more?
They may happen in parallel. That is, one task is busy looping, while another task read the swap entry into swap cache.

do_swap_page isn't busy looping swapcache_prepare, if it fails, it exits, then we have a completely new page fault. this new page fault will lookup swapcache and find it, going into the path to set swapcache to ptes. so the new page fault won't do swapcache_prepare any more.

Quoted 36 lines

--
Best Regards,
Huang, Ying
Quoted 31 lines
Quoted 31 lines
--
Best Regards,
Huang, Ying
Quoted 25 lines
      /*
         * See __read_swap_cache_async(). We might either have raced against
         * another thread, or the entry could have been freed and reused in the
       * meantime. Make sure that the PTE did not change, to detect freeing.
       */
      vmf->pte = pte_offset_map_lock(vma->vm_mm, vmf->pmd,
                                     vmf->address, &vmf->ptl);
      if (!vmf->pte || !pte_same(ptep_get(vmf->pte), vmf->orig_pte))
              goto unlock;


      schedule();
}
I was skeptical about the schedule(), but __read_swap_cache_async() does it already because there is no better way to wait for the event to happen.
With something like above you would no longer depend on the speed of schedule() to determine how often you would retry the fault, which would likely make sense.
I do wonder about the schedule() vs. schedule_timeout_uninterruptible(), though. No expert on that area, do you have any idea?

Kairui Song

author

2024-02-18 09:30 UTC #11 lore raw

Quoted 208 lines · Barry Song

Quoted 206 lines · Ying
Quoted 202 lines · Barry Song
Quoted 194 lines · Ying
Quoted 175 lines · David Hildenbrand
Quoted 135 lines · Kairui Song
From: Kairui Song <[email protected]>
When skipping swapcache for SWP_SYNCHRONOUS_IO, if two or more threads
swapin the same entry at the same time, they get different pages (A, B). Before one thread (T0) finishes the swapin and installs page (A) to the PTE, another thread (T1) could finish swapin of page (B), swap_free the entry, then swap out the possibly modified page reusing the same entry. It breaks the pte_same check in (T0) because PTE value is unchanged, causing ABA problem. Thread (T0) will install a stalled page (A) into the PTE and cause data corruption. One possible callstack is like this:
CPU0                                 CPU1
----                                 ----
do_swap_page()                       do_swap_page() with same entry
<direct swapin path>                 <direct swapin path>
<alloc page A>                       <alloc page B>
swap_read_folio() <- read to page A  swap_read_folio() <- read to page B
<slow on later locks or interrupt>   <finished swapin first>
...                                  set_pte_at()
swap_free() <- entry is free <write to page B, now page A stalled> <swap out page B to same swap entry>
pte_same() <- Check pass, PTE seems
unchanged, but page A
is stalled!
swap_free() <- page B content lost!
set_pte_at() <- staled page A installed!
And besides, for ZRAM, swap_free() allows the swap device to discard the entry content, so even if page (B) is not modified, if swap_read_folio() on CPU0 happens later than swap_free() on CPU1, it may also cause data loss.
To fix this, reuse swapcache_prepare which will pin the swap entry using
the cache flag, and allow only one thread to pin it. Release the pin after PT unlocked. Racers will simply wait since it's a rare and very short event. A schedule() call is added to avoid wasting too much CPU or adding too much noise to perf statistics
Other methods like increasing the swap count don't seem to be a good idea after some tests, that will cause racers to fall back to use the swap cache again. Parallel swapin using different methods leads to a much more complex scenario.
Reproducer:
This race issue can be triggered easily using a well constructed reproducer and patched brd (with a delay in read path) [1]:
With latest 6.8 mainline, race caused data loss can be observed easily:
$ gcc -g -lpthread test-thread-swap-race.c && ./a.out
   Polulating 32MB of memory region...
   Keep swapping out...
   Starting round 0...
   Spawning 65536 workers...
   32746 workers spawned, wait for done...
   Round 0: Error on 0x5aa00, expected 32746, got 32743, 3 data loss!
   Round 0: Error on 0x395200, expected 32746, got 32743, 3 data loss!
   Round 0: Error on 0x3fd000, expected 32746, got 32737, 9 data loss!
   Round 0 Failed, 15 data loss!
This reproducer spawns multiple threads sharing the same memory
region
using a small swap device. Every two threads updates mapped pages one by
one in opposite direction trying to create a race, with one dedicated
thread keep swapping out the data out using madvise.
The reproducer created a reproduce rate of about once every 5
minutes,
so the race should be totally possible in production.
After this patch, I ran the reproducer for over a few hundred rounds
and no data loss observed.
Performance overhead is minimal, microbenchmark swapin 10G from 32G
zram:
Before:     10934698 us
After:      11157121 us
Non-direct: 13155355 us (Dropping SWP_SYNCHRONOUS_IO flag)
Fixes: 0bcac06f27d7 ("mm, swap: skip swapcache for swapin of
synchronous device")
Link: https://github.com/ryncsn/emm-test-project/tree/master/swap-stress-race [1]
Reported-by: "Huang, Ying" <[email protected]>
Closes: https://lore.kernel.org/lkml/[email protected]/
Signed-off-by: Kairui Song <[email protected]>
Cc: [email protected]
---
Update from V2:
- Add a schedule() if raced to prevent repeated page faults wasting CPU
   and add noise to perf statistics.
- Use a bool to state the special case instead of reusing existing
   variables fixing error handling [Minchan Kim].
V2:
https://lore.kernel.org/all/[email protected]/
Update from V1:
- Add some words on ZRAM case, it will discard swap content on swap_free so the race window is a bit different but cure is the same. [Barry Song]
- Update comments make it cleaner [Huang, Ying]
- Add a function place holder to fix CONFIG_SWAP=n built [SeongJae Park]
- Update the commit message and summary, refer to SWP_SYNCHRONOUS_IO instead of "direct swapin path" [Yu Zhao]
- Update commit message.
- Collect Review and Acks.
V1:
https://lore.kernel.org/all/[email protected]/
  include/linux/swap.h |  5 +++++
  mm/memory.c          | 20 ++++++++++++++++++++
  mm/swap.h            |  5 +++++
  mm/swapfile.c        | 13 +++++++++++++
  4 files changed, 43 insertions(+)
include/linux/swap.h +4 -0
index 4db00ddad261..8d28f6091a32 100644
@@ -549,6 +549,11 @@ static inline int swap_duplicate(swp_entry_t swp)
     return 0;
  }
  +static inline int swapcache_prepare(swp_entry_t swp)
+{
+    return 0;
+}
+
  static inline void swap_free(swp_entry_t swp)
  {
  }
mm/memory.c +5 -0
index 7e1f4849463a..7059230d0a54 100644
@@ -3799,6 +3799,7 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
     struct page *page;
     struct swap_info_struct *si = NULL;
     rmap_t rmap_flags = RMAP_NONE;
+    bool need_clear_cache = false;
     bool exclusive = false;
     swp_entry_t entry;
     pte_t pte;
@@ -3867,6 +3868,20 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
     if (!folio) {
             if (data_race(si->flags & SWP_SYNCHRONOUS_IO) &&
                 __swap_count(entry) == 1) {
+                    /*
+                     * Prevent parallel swapin from proceeding with
+                     * the cache flag. Otherwise, another thread may
+                     * finish swapin first, free the entry, and swapout
mm/memory.c vm_fault_t do_swap_page(struct vm_fault *vmf) · line 3884
Quoted 10 lines
mm/memory.c +10 -0
+                     * reusing the same entry. It's undetectable as
+                     * pte_same() returns true due to entry reuse.
+                     */
+                    if (swapcache_prepare(entry)) {
+                            /* Relax a bit to prevent rapid repeated page faults */
+                            schedule();
+                            goto out;
+                    }
+                    need_clear_cache = true;
+
I took a closer look at __read_swap_cache_async() and it essentially does something similar.
Instead of returning, it keeps retrying until it finds that swapcache_prepare() fails for another reason than -EEXISTS (e.g., freed concurrently) or it finds the entry in the swapcache.
So if you would succeed here on a freed+reused swap entry, __read_swap_cache_async() would simply retry.
It spells that out:
/*
* We might race against __delete_from_swap_cache(), and
* stumble across a swap_map entry whose SWAP_HAS_CACHE
* has not yet been cleared. Or race against another
* __read_swap_cache_async(), which has set SWAP_HAS_CACHE
* in swap_map, but not yet added its folio to swap cache. */
Whereby we could not race against this code here as well where we speculatively set SWAP_HAS_CACHE and might never add something to the swap cache.
I'd probably avoid the wrong returns and do something even closer to __read_swap_cache_async().
Quoted 15 lines
while (true) {
      /*
       * Fake that we are trying to insert a page into the swapcache, to
       * serialize against concurrent threads wanting to do the same.
       * [more from your description]
       */
      ret = swapcache_prepare(entry);
      if (likely(!ret)
              /*
               * Move forward with swapin, we'll recheck if the PTE hasn't
               * changed later.
               */
              break;
      else if (ret != -EEXIST)
              goto out;
Quoted 2 lines
The swap entry may be kept in swap cache for long time. For example, it may be read into swap cache via MADV_WILLNEED.
This seems fine.
if swapcache has data from WILLNEED, the new page fault will hit it. Thus, we won't go into the SYNC_IO path any more?

Quoted 8 lines

Quoted 2 lines
They may happen in parallel. That is, one task is busy looping, while another task read the swap entry into swap cache.
do_swap_page isn't busy looping swapcache_prepare, if it fails, it exits, then we have a completely new page fault. this new page fault will lookup swapcache and find it, going into the path to set swapcache to ptes. so the new page fault won't do swapcache_prepare any more.

Hi Barry

The issue here with this code snip is that we could have swapcache_prepare(entry) == -EEXIST and pte_same == true, not necessarily WILLNEED, a concurrent fault and swapout could cause that. Then we are stuck with while(true) here.

It can be fixed, but still there are other potential issues, we end up with a similar loop in swap_state.c, may still need to bail out the fault for some cases, things seem not improved here.

Chris Li

maintainer

2024-02-18 18:37 UTC #12 lore raw

In reply to Huang, Ying

Quoted 198 lines · Ying

Quoted 195 lines · Barry Song
Quoted 194 lines · Ying
Quoted 190 lines · David Hildenbrand
Quoted 135 lines · Kairui Song
From: Kairui Song <[email protected]>
When skipping swapcache for SWP_SYNCHRONOUS_IO, if two or more threads
swapin the same entry at the same time, they get different pages (A, B). Before one thread (T0) finishes the swapin and installs page (A) to the PTE, another thread (T1) could finish swapin of page (B), swap_free the entry, then swap out the possibly modified page reusing the same entry. It breaks the pte_same check in (T0) because PTE value is unchanged, causing ABA problem. Thread (T0) will install a stalled page (A) into the PTE and cause data corruption. One possible callstack is like this:
CPU0                                 CPU1
----                                 ----
do_swap_page()                       do_swap_page() with same entry
<direct swapin path>                 <direct swapin path>
<alloc page A>                       <alloc page B>
swap_read_folio() <- read to page A  swap_read_folio() <- read to page B
<slow on later locks or interrupt>   <finished swapin first>
...                                  set_pte_at()
swap_free() <- entry is free <write to page B, now page A stalled> <swap out page B to same swap entry>
pte_same() <- Check pass, PTE seems
unchanged, but page A
is stalled!
swap_free() <- page B content lost!
set_pte_at() <- staled page A installed!
And besides, for ZRAM, swap_free() allows the swap device to discard the entry content, so even if page (B) is not modified, if swap_read_folio() on CPU0 happens later than swap_free() on CPU1, it may also cause data loss.
To fix this, reuse swapcache_prepare which will pin the swap entry using
the cache flag, and allow only one thread to pin it. Release the pin after PT unlocked. Racers will simply wait since it's a rare and very short event. A schedule() call is added to avoid wasting too much CPU or adding too much noise to perf statistics
Other methods like increasing the swap count don't seem to be a good idea after some tests, that will cause racers to fall back to use the swap cache again. Parallel swapin using different methods leads to a much more complex scenario.
Reproducer:
This race issue can be triggered easily using a well constructed reproducer and patched brd (with a delay in read path) [1]:
With latest 6.8 mainline, race caused data loss can be observed easily:
$ gcc -g -lpthread test-thread-swap-race.c && ./a.out
   Polulating 32MB of memory region...
   Keep swapping out...
   Starting round 0...
   Spawning 65536 workers...
   32746 workers spawned, wait for done...
   Round 0: Error on 0x5aa00, expected 32746, got 32743, 3 data loss!
   Round 0: Error on 0x395200, expected 32746, got 32743, 3 data loss!
   Round 0: Error on 0x3fd000, expected 32746, got 32737, 9 data loss!
   Round 0 Failed, 15 data loss!
This reproducer spawns multiple threads sharing the same memory
region
using a small swap device. Every two threads updates mapped pages one by
one in opposite direction trying to create a race, with one dedicated
thread keep swapping out the data out using madvise.
The reproducer created a reproduce rate of about once every 5
minutes,
so the race should be totally possible in production.
After this patch, I ran the reproducer for over a few hundred rounds
and no data loss observed.
Performance overhead is minimal, microbenchmark swapin 10G from 32G
zram:
Before:     10934698 us
After:      11157121 us
Non-direct: 13155355 us (Dropping SWP_SYNCHRONOUS_IO flag)
Fixes: 0bcac06f27d7 ("mm, swap: skip swapcache for swapin of
synchronous device")
Link: https://github.com/ryncsn/emm-test-project/tree/master/swap-stress-race [1]
Reported-by: "Huang, Ying" <[email protected]>
Closes: https://lore.kernel.org/lkml/[email protected]/
Signed-off-by: Kairui Song <[email protected]>
Cc: [email protected]
---
Update from V2:
- Add a schedule() if raced to prevent repeated page faults wasting CPU
   and add noise to perf statistics.
- Use a bool to state the special case instead of reusing existing
   variables fixing error handling [Minchan Kim].
V2:
https://lore.kernel.org/all/[email protected]/
Update from V1:
- Add some words on ZRAM case, it will discard swap content on swap_free so the race window is a bit different but cure is the same. [Barry Song]
- Update comments make it cleaner [Huang, Ying]
- Add a function place holder to fix CONFIG_SWAP=n built [SeongJae Park]
- Update the commit message and summary, refer to SWP_SYNCHRONOUS_IO instead of "direct swapin path" [Yu Zhao]
- Update commit message.
- Collect Review and Acks.
V1:
https://lore.kernel.org/all/[email protected]/
  include/linux/swap.h |  5 +++++
  mm/memory.c          | 20 ++++++++++++++++++++
  mm/swap.h            |  5 +++++
  mm/swapfile.c        | 13 +++++++++++++
  4 files changed, 43 insertions(+)
include/linux/swap.h +4 -0
index 4db00ddad261..8d28f6091a32 100644
@@ -549,6 +549,11 @@ static inline int swap_duplicate(swp_entry_t swp)
     return 0;
  }
  +static inline int swapcache_prepare(swp_entry_t swp)
+{
+    return 0;
+}
+
  static inline void swap_free(swp_entry_t swp)
  {
  }
mm/memory.c +5 -0
index 7e1f4849463a..7059230d0a54 100644
@@ -3799,6 +3799,7 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
     struct page *page;
     struct swap_info_struct *si = NULL;
     rmap_t rmap_flags = RMAP_NONE;
+    bool need_clear_cache = false;
     bool exclusive = false;
     swp_entry_t entry;
     pte_t pte;
@@ -3867,6 +3868,20 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
     if (!folio) {
             if (data_race(si->flags & SWP_SYNCHRONOUS_IO) &&
                 __swap_count(entry) == 1) {
+                    /*
+                     * Prevent parallel swapin from proceeding with
+                     * the cache flag. Otherwise, another thread may
+                     * finish swapin first, free the entry, and swapout
mm/memory.c vm_fault_t do_swap_page(struct vm_fault *vmf) · line 3884
Quoted 10 lines
mm/memory.c +10 -0
+                     * reusing the same entry. It's undetectable as
+                     * pte_same() returns true due to entry reuse.
+                     */
+                    if (swapcache_prepare(entry)) {
+                            /* Relax a bit to prevent rapid repeated page faults */
+                            schedule();
+                            goto out;
+                    }
+                    need_clear_cache = true;
+
I took a closer look at __read_swap_cache_async() and it essentially does something similar.
Instead of returning, it keeps retrying until it finds that swapcache_prepare() fails for another reason than -EEXISTS (e.g., freed concurrently) or it finds the entry in the swapcache.
So if you would succeed here on a freed+reused swap entry, __read_swap_cache_async() would simply retry.
It spells that out:
/*
* We might race against __delete_from_swap_cache(), and
* stumble across a swap_map entry whose SWAP_HAS_CACHE
* has not yet been cleared. Or race against another
* __read_swap_cache_async(), which has set SWAP_HAS_CACHE
* in swap_map, but not yet added its folio to swap cache. */
Whereby we could not race against this code here as well where we speculatively set SWAP_HAS_CACHE and might never add something to the swap cache.
I'd probably avoid the wrong returns and do something even closer to __read_swap_cache_async().
while (true) {
      /*
       * Fake that we are trying to insert a page into the swapcache, to
       * serialize against concurrent threads wanting to do the same.
       * [more from your description]
       */
      ret = swapcache_prepare(entry);
      if (likely(!ret)
              /*
               * Move forward with swapin, we'll recheck if the PTE hasn't
               * changed later.
               */
              break;
      else if (ret != -EEXIST)
              goto out;

Quoted 2 lines

Quoted 2 lines
Quoted 2 lines
The swap entry may be kept in swap cache for long time. For example, it may be read into swap cache via MADV_WILLNEED.

I am trying to find the alternative path which can cause the swap fault to bring in the swap cache in page fault while the SYNC IO is looping for HAS_SWAP_CACHE. Kairui was able to identify the in the current code in do_page_fault() path, the rmap and fork case wouldn't be able to modify the swap cache causing a problem. The MADV_WILLNEED is an excellent example. Thank you for finding this example.

Quoted 8 lines

Quoted 5 lines
This seems fine.
if swapcache has data from WILLNEED, the new page fault will hit it. Thus, we won't go into the SYNC_IO path any more?
They may happen in parallel. That is, one task is busy looping, while another task read the swap entry into swap cache.

Agree.

Chris

David Hildenbrand

2024-02-19 09:01 UTC #13 lore raw

In reply to Huang, Ying

Quoted 193 lines · Ying

Quoted 175 lines · David Hildenbrand
Quoted 135 lines · Kairui Song
From: Kairui Song <[email protected]>
When skipping swapcache for SWP_SYNCHRONOUS_IO, if two or more threads
swapin the same entry at the same time, they get different pages (A, B). Before one thread (T0) finishes the swapin and installs page (A) to the PTE, another thread (T1) could finish swapin of page (B), swap_free the entry, then swap out the possibly modified page reusing the same entry. It breaks the pte_same check in (T0) because PTE value is unchanged, causing ABA problem. Thread (T0) will install a stalled page (A) into the PTE and cause data corruption. One possible callstack is like this:
CPU0                                 CPU1
----                                 ----
do_swap_page()                       do_swap_page() with same entry
<direct swapin path>                 <direct swapin path>
<alloc page A>                       <alloc page B>
swap_read_folio() <- read to page A  swap_read_folio() <- read to page B
<slow on later locks or interrupt>   <finished swapin first>
...                                  set_pte_at()
swap_free() <- entry is free <write to page B, now page A stalled> <swap out page B to same swap entry>
pte_same() <- Check pass, PTE seems
unchanged, but page A
is stalled!
swap_free() <- page B content lost!
set_pte_at() <- staled page A installed!
And besides, for ZRAM, swap_free() allows the swap device to discard the entry content, so even if page (B) is not modified, if swap_read_folio() on CPU0 happens later than swap_free() on CPU1, it may also cause data loss.
To fix this, reuse swapcache_prepare which will pin the swap entry using
the cache flag, and allow only one thread to pin it. Release the pin after PT unlocked. Racers will simply wait since it's a rare and very short event. A schedule() call is added to avoid wasting too much CPU or adding too much noise to perf statistics
Other methods like increasing the swap count don't seem to be a good idea after some tests, that will cause racers to fall back to use the swap cache again. Parallel swapin using different methods leads to a much more complex scenario.
Reproducer:
This race issue can be triggered easily using a well constructed reproducer and patched brd (with a delay in read path) [1]:
With latest 6.8 mainline, race caused data loss can be observed easily:
$ gcc -g -lpthread test-thread-swap-race.c && ./a.out
    Polulating 32MB of memory region...
    Keep swapping out...
    Starting round 0...
    Spawning 65536 workers...
    32746 workers spawned, wait for done...
    Round 0: Error on 0x5aa00, expected 32746, got 32743, 3 data loss!
    Round 0: Error on 0x395200, expected 32746, got 32743, 3 data loss!
    Round 0: Error on 0x3fd000, expected 32746, got 32737, 9 data loss!
    Round 0 Failed, 15 data loss!
This reproducer spawns multiple threads sharing the same memory
region
using a small swap device. Every two threads updates mapped pages one by
one in opposite direction trying to create a race, with one dedicated
thread keep swapping out the data out using madvise.
The reproducer created a reproduce rate of about once every 5
minutes,
so the race should be totally possible in production.
After this patch, I ran the reproducer for over a few hundred rounds
and no data loss observed.
Performance overhead is minimal, microbenchmark swapin 10G from 32G
zram:
Before:     10934698 us
After:      11157121 us
Non-direct: 13155355 us (Dropping SWP_SYNCHRONOUS_IO flag)
Fixes: 0bcac06f27d7 ("mm, swap: skip swapcache for swapin of
synchronous device")
Link: https://github.com/ryncsn/emm-test-project/tree/master/swap-stress-race [1]
Reported-by: "Huang, Ying" <[email protected]>
Closes: https://lore.kernel.org/lkml/[email protected]/
Signed-off-by: Kairui Song <[email protected]>
Cc: [email protected]
---
Update from V2:
- Add a schedule() if raced to prevent repeated page faults wasting CPU
    and add noise to perf statistics.
- Use a bool to state the special case instead of reusing existing
    variables fixing error handling [Minchan Kim].
V2:
https://lore.kernel.org/all/[email protected]/
Update from V1:
- Add some words on ZRAM case, it will discard swap content on swap_free so the race window is a bit different but cure is the same. [Barry Song]
- Update comments make it cleaner [Huang, Ying]
- Add a function place holder to fix CONFIG_SWAP=n built [SeongJae Park]
- Update the commit message and summary, refer to SWP_SYNCHRONOUS_IO instead of "direct swapin path" [Yu Zhao]
- Update commit message.
- Collect Review and Acks.
V1:
https://lore.kernel.org/all/[email protected]/
   include/linux/swap.h |  5 +++++
   mm/memory.c          | 20 ++++++++++++++++++++
   mm/swap.h            |  5 +++++
   mm/swapfile.c        | 13 +++++++++++++
   4 files changed, 43 insertions(+)
include/linux/swap.h +4 -0
index 4db00ddad261..8d28f6091a32 100644
@@ -549,6 +549,11 @@ static inline int swap_duplicate(swp_entry_t swp)
   	return 0;
   }
   +static inline int swapcache_prepare(swp_entry_t swp)
+{
+	return 0;
+}
+
   static inline void swap_free(swp_entry_t swp)
   {
   }
mm/memory.c +5 -0
index 7e1f4849463a..7059230d0a54 100644
@@ -3799,6 +3799,7 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
   	struct page *page;
   	struct swap_info_struct *si = NULL;
   	rmap_t rmap_flags = RMAP_NONE;
+	bool need_clear_cache = false;
   	bool exclusive = false;
   	swp_entry_t entry;
   	pte_t pte;
@@ -3867,6 +3868,20 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
   	if (!folio) {
   		if (data_race(si->flags & SWP_SYNCHRONOUS_IO) &&
   		    __swap_count(entry) == 1) {
+			/*
+			 * Prevent parallel swapin from proceeding with
+			 * the cache flag. Otherwise, another thread may
+			 * finish swapin first, free the entry, and swapout
mm/memory.c vm_fault_t do_swap_page(struct vm_fault *vmf) · line 3884
Quoted 10 lines
mm/memory.c +10 -0
+			 * reusing the same entry. It's undetectable as
+			 * pte_same() returns true due to entry reuse.
+			 */
+			if (swapcache_prepare(entry)) {
+				/* Relax a bit to prevent rapid repeated page faults */
+				schedule();
+				goto out;
+			}
+			need_clear_cache = true;
+
I took a closer look at __read_swap_cache_async() and it essentially does something similar.
Instead of returning, it keeps retrying until it finds that swapcache_prepare() fails for another reason than -EEXISTS (e.g., freed concurrently) or it finds the entry in the swapcache.
So if you would succeed here on a freed+reused swap entry, __read_swap_cache_async() would simply retry.
It spells that out:
/*
* We might race against __delete_from_swap_cache(), and
* stumble across a swap_map entry whose SWAP_HAS_CACHE
* has not yet been cleared. Or race against another
* __read_swap_cache_async(), which has set SWAP_HAS_CACHE
* in swap_map, but not yet added its folio to swap cache. */
Whereby we could not race against this code here as well where we speculatively set SWAP_HAS_CACHE and might never add something to the swap cache.
I'd probably avoid the wrong returns and do something even closer to __read_swap_cache_async().
Quoted 15 lines
while (true) {
	/*
	 * Fake that we are trying to insert a page into the swapcache, to
	 * serialize against concurrent threads wanting to do the same.
	 * [more from your description]
	 */
	ret = swapcache_prepare(entry);
	if (likely(!ret)
		/*
		 * Move forward with swapin, we'll recheck if the PTE hasn't
		 * changed later.
		 */
		break;
	else if (ret != -EEXIST)
		goto out;

Quoted 2 lines

The swap entry may be kept in swap cache for long time. For example, it may be read into swap cache via MADV_WILLNEED.

Right, we'd have to check for the swapcache.

I briefly thought about just factoring out what we have in __read_swap_cache_async() and reusing here. Similar problem to solve, and quite a lot of duplicate code.

But not worth the churn in a simple fix. We could explore that option as a cleanup on top.

Signature · 3 lines

Cheers,

David / dhildenb

Huang, Ying

2024-02-18 08:34 UTC #14 lore raw

In reply to Kairui Song

Quoted 41 lines · Kairui Song

From: Kairui Song <[email protected]>
When skipping swapcache for SWP_SYNCHRONOUS_IO, if two or more threads swapin the same entry at the same time, they get different pages (A, B). Before one thread (T0) finishes the swapin and installs page (A) to the PTE, another thread (T1) could finish swapin of page (B), swap_free the entry, then swap out the possibly modified page reusing the same entry. It breaks the pte_same check in (T0) because PTE value is unchanged, causing ABA problem. Thread (T0) will install a stalled page (A) into the PTE and cause data corruption.
One possible callstack is like this:
CPU0                                 CPU1
----                                 ----
do_swap_page()                       do_swap_page() with same entry
<direct swapin path>                 <direct swapin path>
<alloc page A>                       <alloc page B>
swap_read_folio() <- read to page A  swap_read_folio() <- read to page B
<slow on later locks or interrupt>   <finished swapin first>
...                                  set_pte_at()
swap_free() <- entry is free <write to page B, now page A stalled> <swap out page B to same swap entry>
pte_same() <- Check pass, PTE seems
unchanged, but page A
is stalled!
swap_free() <- page B content lost!
set_pte_at() <- staled page A installed!
And besides, for ZRAM, swap_free() allows the swap device to discard the entry content, so even if page (B) is not modified, if swap_read_folio() on CPU0 happens later than swap_free() on CPU1, it may also cause data loss.
To fix this, reuse swapcache_prepare which will pin the swap entry using the cache flag, and allow only one thread to pin it. Release the pin after PT unlocked. Racers will simply wait since it's a rare and very short event. A schedule() call is added to avoid wasting too much CPU or adding too much noise to perf statistics

Quoted 4 lines

Other methods like increasing the swap count don't seem to be a good idea after some tests, that will cause racers to fall back to use the swap cache again. Parallel swapin using different methods leads to a much more complex scenario.

The swap entry may be put in swap cache by some parallel code path anyway. So, we always need to consider that when reasoning the code.

Quoted 99 lines

Reproducer:
This race issue can be triggered easily using a well constructed reproducer and patched brd (with a delay in read path) [1]:
With latest 6.8 mainline, race caused data loss can be observed easily:
$ gcc -g -lpthread test-thread-swap-race.c && ./a.out
  Polulating 32MB of memory region...
  Keep swapping out...
  Starting round 0...
  Spawning 65536 workers...
  32746 workers spawned, wait for done...
  Round 0: Error on 0x5aa00, expected 32746, got 32743, 3 data loss!
  Round 0: Error on 0x395200, expected 32746, got 32743, 3 data loss!
  Round 0: Error on 0x3fd000, expected 32746, got 32737, 9 data loss!
  Round 0 Failed, 15 data loss!
This reproducer spawns multiple threads sharing the same memory region using a small swap device. Every two threads updates mapped pages one by one in opposite direction trying to create a race, with one dedicated thread keep swapping out the data out using madvise.
The reproducer created a reproduce rate of about once every 5 minutes, so the race should be totally possible in production.
After this patch, I ran the reproducer for over a few hundred rounds and no data loss observed.
Performance overhead is minimal, microbenchmark swapin 10G from 32G zram:
Before:     10934698 us
After:      11157121 us
Non-direct: 13155355 us (Dropping SWP_SYNCHRONOUS_IO flag)
Fixes 0bcac06f27d7 ("mm, swap: skip swapcache for swapin of synchronous device")
Link https://github.com/ryncsn/emm-test-project/tree/master/swap-stress-race [1]
Reported-by "Huang, Ying" <[email protected]>
Closes https://lore.kernel.org/lkml/[email protected]/ open here
Signed-off-by Kairui Song <[email protected]>
Cc [email protected]
---
Update from V2:
- Add a schedule() if raced to prevent repeated page faults wasting CPU and add noise to perf statistics.
- Use a bool to state the special case instead of reusing existing variables fixing error handling [Minchan Kim].
V2: https://lore.kernel.org/all/[email protected]/
Update from V1:
- Add some words on ZRAM case, it will discard swap content on swap_free so the race window is a bit different but cure is the same. [Barry Song]
- Update comments make it cleaner [Huang, Ying]
- Add a function place holder to fix CONFIG_SWAP=n built [SeongJae Park]
- Update the commit message and summary, refer to SWP_SYNCHRONOUS_IO instead of "direct swapin path" [Yu Zhao]
- Update commit message.
- Collect Review and Acks.
V1: https://lore.kernel.org/all/[email protected]/
 include/linux/swap.h |  5 +++++
 mm/memory.c          | 20 ++++++++++++++++++++
 mm/swap.h            |  5 +++++
 mm/swapfile.c        | 13 +++++++++++++
 4 files changed, 43 insertions(+)
include/linux/swap.h +5 -0
index 4db00ddad261..8d28f6091a32 100644
@@ -549,6 +549,11 @@ static inline int swap_duplicate(swp_entry_t swp)
 	return 0;
 }
 
+static inline int swapcache_prepare(swp_entry_t swp)
+{
+	return 0;
+}
+
 static inline void swap_free(swp_entry_t swp)
 {
 }
mm/memory.c +1 -0
index 7e1f4849463a..7059230d0a54 100644
@@ -3799,6 +3799,7 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
 	struct page *page;
 	struct swap_info_struct *si = NULL;
 	rmap_t rmap_flags = RMAP_NONE;
+	bool need_clear_cache = false;
 	bool exclusive = false;
 	swp_entry_t entry;
 	pte_t pte;
@@ -3867,6 +3868,20 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
 	if (!folio) {
 		if (data_race(si->flags & SWP_SYNCHRONOUS_IO) &&
 		    __swap_count(entry) == 1) {

mm/memory.c vm_fault_t do_swap_page(struct vm_fault *vmf) · line 3880

Quoted 10 lines

mm/memory.c +10 -0

+			/*
+			 * Prevent parallel swapin from proceeding with
+			 * the cache flag. Otherwise, another thread may
+			 * finish swapin first, free the entry, and swapout
+			 * reusing the same entry. It's undetectable as
+			 * pte_same() returns true due to entry reuse.
+			 */
+			if (swapcache_prepare(entry)) {
+				/* Relax a bit to prevent rapid repeated page faults */
+				schedule();

The current task may be chosen in schedule(). So, I think that we should use cond_resched() here.

--
Best Regards,
Huang, Ying

Quoted 73 lines

mm/memory.c +9 -0

+				goto out;
+			}
+			need_clear_cache = true;
+
 			/* skip swapcache */
 			folio = vma_alloc_folio(GFP_HIGHUSER_MOVABLE, 0,
 						vma, vmf->address, false);

@@ -4117,6 +4132,9 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
 	if (vmf->pte)
 		pte_unmap_unlock(vmf->pte, vmf->ptl);
 out:
+	/* Clear the swap cache pin for direct swapin after PTL unlock */
+	if (need_clear_cache)
+		swapcache_clear(si, entry);
 	if (si)
 		put_swap_device(si);
 	return ret;

@@ -4131,6 +4149,8 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
 		folio_unlock(swapcache);
 		folio_put(swapcache);
 	}
+	if (need_clear_cache)
+		swapcache_clear(si, entry);
 	if (si)
 		put_swap_device(si);
 	return ret;

mm/swap.h +5 -0

index 758c46ca671e..fc2f6ade7f80 100644
@@ -41,6 +41,7 @@ void __delete_from_swap_cache(struct folio *folio,
 void delete_from_swap_cache(struct folio *folio);
 void clear_shadow_from_swap_cache(int type, unsigned long begin,
 				  unsigned long end);
+void swapcache_clear(struct swap_info_struct *si, swp_entry_t entry);
 struct folio *swap_cache_get_folio(swp_entry_t entry,
 		struct vm_area_struct *vma, unsigned long addr);
 struct folio *filemap_get_incore_folio(struct address_space *mapping,

@@ -97,6 +98,10 @@ static inline int swap_writepage(struct page *p, struct writeback_control *wbc)
 	return 0;
 }
 
+static inline void swapcache_clear(struct swap_info_struct *si, swp_entry_t entry)
+{
+}
+
 static inline struct folio *swap_cache_get_folio(swp_entry_t entry,
 		struct vm_area_struct *vma, unsigned long addr)
 {

mm/swapfile.c +13 -0

index 556ff7347d5f..746aa9da5302 100644
@@ -3365,6 +3365,19 @@ int swapcache_prepare(swp_entry_t entry)
 	return __swap_duplicate(entry, SWAP_HAS_CACHE);
 }
 
+void swapcache_clear(struct swap_info_struct *si, swp_entry_t entry)
+{
+	struct swap_cluster_info *ci;
+	unsigned long offset = swp_offset(entry);
+	unsigned char usage;
+
+	ci = lock_cluster_or_swap_info(si, offset);
+	usage = __swap_entry_free_locked(si, offset, SWAP_HAS_CACHE);
+	unlock_cluster_or_swap_info(si, ci);
+	if (!usage)
+		free_swap_slot(entry);
+}
+
 struct swap_info_struct *swp_swap_info(swp_entry_t entry)
 {
 	return swap_type_to_swap_info(swp_type(entry));

Kairui Song

author

2024-02-18 09:20 UTC #15 lore raw

Quoted 156 lines · Ying

Quoted 41 lines · Kairui Song
From: Kairui Song <[email protected]>
When skipping swapcache for SWP_SYNCHRONOUS_IO, if two or more threads swapin the same entry at the same time, they get different pages (A, B). Before one thread (T0) finishes the swapin and installs page (A) to the PTE, another thread (T1) could finish swapin of page (B), swap_free the entry, then swap out the possibly modified page reusing the same entry. It breaks the pte_same check in (T0) because PTE value is unchanged, causing ABA problem. Thread (T0) will install a stalled page (A) into the PTE and cause data corruption.
One possible callstack is like this:
CPU0                                 CPU1
----                                 ----
do_swap_page()                       do_swap_page() with same entry
<direct swapin path>                 <direct swapin path>
<alloc page A>                       <alloc page B>
swap_read_folio() <- read to page A  swap_read_folio() <- read to page B
<slow on later locks or interrupt>   <finished swapin first>
...                                  set_pte_at()
swap_free() <- entry is free <write to page B, now page A stalled> <swap out page B to same swap entry>
pte_same() <- Check pass, PTE seems
unchanged, but page A
is stalled!
swap_free() <- page B content lost!
set_pte_at() <- staled page A installed!
And besides, for ZRAM, swap_free() allows the swap device to discard the entry content, so even if page (B) is not modified, if swap_read_folio() on CPU0 happens later than swap_free() on CPU1, it may also cause data loss.
To fix this, reuse swapcache_prepare which will pin the swap entry using the cache flag, and allow only one thread to pin it. Release the pin after PT unlocked. Racers will simply wait since it's a rare and very short event. A schedule() call is added to avoid wasting too much CPU or adding too much noise to perf statistics
Quoted 4 lines
Other methods like increasing the swap count don't seem to be a good idea after some tests, that will cause racers to fall back to use the swap cache again. Parallel swapin using different methods leads to a much more complex scenario.
The swap entry may be put in swap cache by some parallel code path anyway. So, we always need to consider that when reasoning the code.
Quoted 104 lines
Reproducer:
This race issue can be triggered easily using a well constructed reproducer and patched brd (with a delay in read path) [1]:
With latest 6.8 mainline, race caused data loss can be observed easily:
$ gcc -g -lpthread test-thread-swap-race.c && ./a.out
  Polulating 32MB of memory region...
  Keep swapping out...
  Starting round 0...
  Spawning 65536 workers...
  32746 workers spawned, wait for done...
  Round 0: Error on 0x5aa00, expected 32746, got 32743, 3 data loss!
  Round 0: Error on 0x395200, expected 32746, got 32743, 3 data loss!
  Round 0: Error on 0x3fd000, expected 32746, got 32737, 9 data loss!
  Round 0 Failed, 15 data loss!
This reproducer spawns multiple threads sharing the same memory region using a small swap device. Every two threads updates mapped pages one by one in opposite direction trying to create a race, with one dedicated thread keep swapping out the data out using madvise.
The reproducer created a reproduce rate of about once every 5 minutes, so the race should be totally possible in production.
After this patch, I ran the reproducer for over a few hundred rounds and no data loss observed.
Performance overhead is minimal, microbenchmark swapin 10G from 32G zram:
Before:     10934698 us
After:      11157121 us
Non-direct: 13155355 us (Dropping SWP_SYNCHRONOUS_IO flag)
Fixes 0bcac06f27d7 ("mm, swap: skip swapcache for swapin of synchronous device")
Link https://github.com/ryncsn/emm-test-project/tree/master/swap-stress-race [1]
Reported-by "Huang, Ying" <[email protected]>
Closes https://lore.kernel.org/lkml/[email protected]/ open here
Signed-off-by Kairui Song <[email protected]>
Cc [email protected]
---
Update from V2:
- Add a schedule() if raced to prevent repeated page faults wasting CPU and add noise to perf statistics.
- Use a bool to state the special case instead of reusing existing variables fixing error handling [Minchan Kim].
V2: https://lore.kernel.org/all/[email protected]/
Update from V1:
- Add some words on ZRAM case, it will discard swap content on swap_free so the race window is a bit different but cure is the same. [Barry Song]
- Update comments make it cleaner [Huang, Ying]
- Add a function place holder to fix CONFIG_SWAP=n built [SeongJae Park]
- Update the commit message and summary, refer to SWP_SYNCHRONOUS_IO instead of "direct swapin path" [Yu Zhao]
- Update commit message.
- Collect Review and Acks.
V1: https://lore.kernel.org/all/[email protected]/
 include/linux/swap.h |  5 +++++
 mm/memory.c          | 20 ++++++++++++++++++++
 mm/swap.h            |  5 +++++
 mm/swapfile.c        | 13 +++++++++++++
 4 files changed, 43 insertions(+)
include/linux/swap.h +5 -0
index 4db00ddad261..8d28f6091a32 100644
@@ -549,6 +549,11 @@ static inline int swap_duplicate(swp_entry_t swp)
      return 0;
 }
 
+static inline int swapcache_prepare(swp_entry_t swp)
+{
+     return 0;
+}
+
 static inline void swap_free(swp_entry_t swp)
 {
 }
mm/memory.c +6 -0
index 7e1f4849463a..7059230d0a54 100644
@@ -3799,6 +3799,7 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
      struct page *page;
      struct swap_info_struct *si = NULL;
      rmap_t rmap_flags = RMAP_NONE;
+     bool need_clear_cache = false;
      bool exclusive = false;
      swp_entry_t entry;
      pte_t pte;
@@ -3867,6 +3868,20 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
      if (!folio) {
              if (data_race(si->flags & SWP_SYNCHRONOUS_IO) &&
                  __swap_count(entry) == 1) {
+                     /*
+                      * Prevent parallel swapin from proceeding with
+                      * the cache flag. Otherwise, another thread may
+                      * finish swapin first, free the entry, and swapout
+                      * reusing the same entry. It's undetectable as

Quoted 9 lines

mm/memory.c vm_fault_t do_swap_page(struct vm_fault *vmf) · line 3880

Quoted 5 lines

mm/memory.c +5 -0

+                      * pte_same() returns true due to entry reuse.
+                      */
+                     if (swapcache_prepare(entry)) {
+                             /* Relax a bit to prevent rapid repeated page faults */
+                             schedule();

The current task may be chosen in schedule(). So, I think that we should use cond_resched() here.

I think if we are worried about current task got chosen again we can use schedule_timeout_uninterruptible(1) here. Isn't cond_resched still __schedule() and and it can even get omitted, so it should be "weaker" IIUC.

Huang, Ying

2024-02-19 00:41 UTC #16 lore raw

Quoted 167 lines · Kairui Song

Quoted 156 lines · Ying
Quoted 41 lines · Kairui Song
From: Kairui Song <[email protected]>
When skipping swapcache for SWP_SYNCHRONOUS_IO, if two or more threads swapin the same entry at the same time, they get different pages (A, B). Before one thread (T0) finishes the swapin and installs page (A) to the PTE, another thread (T1) could finish swapin of page (B), swap_free the entry, then swap out the possibly modified page reusing the same entry. It breaks the pte_same check in (T0) because PTE value is unchanged, causing ABA problem. Thread (T0) will install a stalled page (A) into the PTE and cause data corruption.
One possible callstack is like this:
CPU0                                 CPU1
----                                 ----
do_swap_page()                       do_swap_page() with same entry
<direct swapin path>                 <direct swapin path>
<alloc page A>                       <alloc page B>
swap_read_folio() <- read to page A  swap_read_folio() <- read to page B
<slow on later locks or interrupt>   <finished swapin first>
...                                  set_pte_at()
swap_free() <- entry is free <write to page B, now page A stalled> <swap out page B to same swap entry>
pte_same() <- Check pass, PTE seems
unchanged, but page A
is stalled!
swap_free() <- page B content lost!
set_pte_at() <- staled page A installed!
And besides, for ZRAM, swap_free() allows the swap device to discard the entry content, so even if page (B) is not modified, if swap_read_folio() on CPU0 happens later than swap_free() on CPU1, it may also cause data loss.
To fix this, reuse swapcache_prepare which will pin the swap entry using the cache flag, and allow only one thread to pin it. Release the pin after PT unlocked. Racers will simply wait since it's a rare and very short event. A schedule() call is added to avoid wasting too much CPU or adding too much noise to perf statistics
Quoted 4 lines
Other methods like increasing the swap count don't seem to be a good idea after some tests, that will cause racers to fall back to use the swap cache again. Parallel swapin using different methods leads to a much more complex scenario.
The swap entry may be put in swap cache by some parallel code path anyway. So, we always need to consider that when reasoning the code.
Quoted 104 lines
Reproducer:
This race issue can be triggered easily using a well constructed reproducer and patched brd (with a delay in read path) [1]:
With latest 6.8 mainline, race caused data loss can be observed easily:
$ gcc -g -lpthread test-thread-swap-race.c && ./a.out
  Polulating 32MB of memory region...
  Keep swapping out...
  Starting round 0...
  Spawning 65536 workers...
  32746 workers spawned, wait for done...
  Round 0: Error on 0x5aa00, expected 32746, got 32743, 3 data loss!
  Round 0: Error on 0x395200, expected 32746, got 32743, 3 data loss!
  Round 0: Error on 0x3fd000, expected 32746, got 32737, 9 data loss!
  Round 0 Failed, 15 data loss!
This reproducer spawns multiple threads sharing the same memory region using a small swap device. Every two threads updates mapped pages one by one in opposite direction trying to create a race, with one dedicated thread keep swapping out the data out using madvise.
The reproducer created a reproduce rate of about once every 5 minutes, so the race should be totally possible in production.
After this patch, I ran the reproducer for over a few hundred rounds and no data loss observed.
Performance overhead is minimal, microbenchmark swapin 10G from 32G zram:
Before:     10934698 us
After:      11157121 us
Non-direct: 13155355 us (Dropping SWP_SYNCHRONOUS_IO flag)
Fixes 0bcac06f27d7 ("mm, swap: skip swapcache for swapin of synchronous device")
Link https://github.com/ryncsn/emm-test-project/tree/master/swap-stress-race [1]
Reported-by "Huang, Ying" <[email protected]>
Closes https://lore.kernel.org/lkml/[email protected]/ open here
Signed-off-by Kairui Song <[email protected]>
Cc [email protected]
---
Update from V2:
- Add a schedule() if raced to prevent repeated page faults wasting CPU and add noise to perf statistics.
- Use a bool to state the special case instead of reusing existing variables fixing error handling [Minchan Kim].
V2: https://lore.kernel.org/all/[email protected]/
Update from V1:
- Add some words on ZRAM case, it will discard swap content on swap_free so the race window is a bit different but cure is the same. [Barry Song]
- Update comments make it cleaner [Huang, Ying]
- Add a function place holder to fix CONFIG_SWAP=n built [SeongJae Park]
- Update the commit message and summary, refer to SWP_SYNCHRONOUS_IO instead of "direct swapin path" [Yu Zhao]
- Update commit message.
- Collect Review and Acks.
V1: https://lore.kernel.org/all/[email protected]/
 include/linux/swap.h |  5 +++++
 mm/memory.c          | 20 ++++++++++++++++++++
 mm/swap.h            |  5 +++++
 mm/swapfile.c        | 13 +++++++++++++
 4 files changed, 43 insertions(+)
include/linux/swap.h +5 -0
index 4db00ddad261..8d28f6091a32 100644
@@ -549,6 +549,11 @@ static inline int swap_duplicate(swp_entry_t swp)
      return 0;
 }
 
+static inline int swapcache_prepare(swp_entry_t swp)
+{
+     return 0;
+}
+
 static inline void swap_free(swp_entry_t swp)
 {
 }
mm/memory.c +6 -0
index 7e1f4849463a..7059230d0a54 100644
@@ -3799,6 +3799,7 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
      struct page *page;
      struct swap_info_struct *si = NULL;
      rmap_t rmap_flags = RMAP_NONE;
+     bool need_clear_cache = false;
      bool exclusive = false;
      swp_entry_t entry;
      pte_t pte;
@@ -3867,6 +3868,20 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
      if (!folio) {
              if (data_race(si->flags & SWP_SYNCHRONOUS_IO) &&
                  __swap_count(entry) == 1) {
+                     /*
+                      * Prevent parallel swapin from proceeding with
+                      * the cache flag. Otherwise, another thread may
+                      * finish swapin first, free the entry, and swapout
+                      * reusing the same entry. It's undetectable as
Quoted 9 lines
mm/memory.c vm_fault_t do_swap_page(struct vm_fault *vmf) · line 3880
Quoted 5 lines
mm/memory.c +5 -0
+                      * pte_same() returns true due to entry reuse.
+                      */
+                     if (swapcache_prepare(entry)) {
+                             /* Relax a bit to prevent rapid repeated page faults */
+                             schedule();
The current task may be chosen in schedule(). So, I think that we should use cond_resched() here.

Quoted 4 lines

I think if we are worried about current task got chosen again we can use schedule_timeout_uninterruptible(1) here. Isn't cond_resched still __schedule() and and it can even get omitted, so it should be "weaker" IIUC.

schedule_timeout_uninterruptible(1) will introduce 1ms latency for the second task. That may kill performance of some workloads.

--
Best Regards,
Huang, Ying

Huang, Ying

2024-02-19 02:35 UTC #17 lore raw

Quoted 169 lines · Ying

Quoted 167 lines · Kairui Song
Quoted 165 lines · Ying
Quoted 41 lines · Kairui Song
From: Kairui Song <[email protected]>
When skipping swapcache for SWP_SYNCHRONOUS_IO, if two or more threads swapin the same entry at the same time, they get different pages (A, B). Before one thread (T0) finishes the swapin and installs page (A) to the PTE, another thread (T1) could finish swapin of page (B), swap_free the entry, then swap out the possibly modified page reusing the same entry. It breaks the pte_same check in (T0) because PTE value is unchanged, causing ABA problem. Thread (T0) will install a stalled page (A) into the PTE and cause data corruption.
One possible callstack is like this:
CPU0                                 CPU1
----                                 ----
do_swap_page()                       do_swap_page() with same entry
<direct swapin path>                 <direct swapin path>
<alloc page A>                       <alloc page B>
swap_read_folio() <- read to page A  swap_read_folio() <- read to page B
<slow on later locks or interrupt>   <finished swapin first>
...                                  set_pte_at()
swap_free() <- entry is free <write to page B, now page A stalled> <swap out page B to same swap entry>
pte_same() <- Check pass, PTE seems
unchanged, but page A
is stalled!
swap_free() <- page B content lost!
set_pte_at() <- staled page A installed!
And besides, for ZRAM, swap_free() allows the swap device to discard the entry content, so even if page (B) is not modified, if swap_read_folio() on CPU0 happens later than swap_free() on CPU1, it may also cause data loss.
To fix this, reuse swapcache_prepare which will pin the swap entry using the cache flag, and allow only one thread to pin it. Release the pin after PT unlocked. Racers will simply wait since it's a rare and very short event. A schedule() call is added to avoid wasting too much CPU or adding too much noise to perf statistics
Quoted 4 lines
Other methods like increasing the swap count don't seem to be a good idea after some tests, that will cause racers to fall back to use the swap cache again. Parallel swapin using different methods leads to a much more complex scenario.
The swap entry may be put in swap cache by some parallel code path anyway. So, we always need to consider that when reasoning the code.
Quoted 99 lines
Reproducer:
This race issue can be triggered easily using a well constructed reproducer and patched brd (with a delay in read path) [1]:
With latest 6.8 mainline, race caused data loss can be observed easily:
$ gcc -g -lpthread test-thread-swap-race.c && ./a.out
  Polulating 32MB of memory region...
  Keep swapping out...
  Starting round 0...
  Spawning 65536 workers...
  32746 workers spawned, wait for done...
  Round 0: Error on 0x5aa00, expected 32746, got 32743, 3 data loss!
  Round 0: Error on 0x395200, expected 32746, got 32743, 3 data loss!
  Round 0: Error on 0x3fd000, expected 32746, got 32737, 9 data loss!
  Round 0 Failed, 15 data loss!
This reproducer spawns multiple threads sharing the same memory region using a small swap device. Every two threads updates mapped pages one by one in opposite direction trying to create a race, with one dedicated thread keep swapping out the data out using madvise.
The reproducer created a reproduce rate of about once every 5 minutes, so the race should be totally possible in production.
After this patch, I ran the reproducer for over a few hundred rounds and no data loss observed.
Performance overhead is minimal, microbenchmark swapin 10G from 32G zram:
Before:     10934698 us
After:      11157121 us
Non-direct: 13155355 us (Dropping SWP_SYNCHRONOUS_IO flag)
Fixes 0bcac06f27d7 ("mm, swap: skip swapcache for swapin of synchronous device")
Link https://github.com/ryncsn/emm-test-project/tree/master/swap-stress-race [1]
Reported-by "Huang, Ying" <[email protected]>
Closes https://lore.kernel.org/lkml/[email protected]/ open here
Signed-off-by Kairui Song <[email protected]>
Cc [email protected]
---
Update from V2:
- Add a schedule() if raced to prevent repeated page faults wasting CPU and add noise to perf statistics.
- Use a bool to state the special case instead of reusing existing variables fixing error handling [Minchan Kim].
V2: https://lore.kernel.org/all/[email protected]/
Update from V1:
- Add some words on ZRAM case, it will discard swap content on swap_free so the race window is a bit different but cure is the same. [Barry Song]
- Update comments make it cleaner [Huang, Ying]
- Add a function place holder to fix CONFIG_SWAP=n built [SeongJae Park]
- Update the commit message and summary, refer to SWP_SYNCHRONOUS_IO instead of "direct swapin path" [Yu Zhao]
- Update commit message.
- Collect Review and Acks.
V1: https://lore.kernel.org/all/[email protected]/
 include/linux/swap.h |  5 +++++
 mm/memory.c          | 20 ++++++++++++++++++++
 mm/swap.h            |  5 +++++
 mm/swapfile.c        | 13 +++++++++++++
 4 files changed, 43 insertions(+)
include/linux/swap.h +5 -0
index 4db00ddad261..8d28f6091a32 100644
@@ -549,6 +549,11 @@ static inline int swap_duplicate(swp_entry_t swp)
      return 0;
 }
 
+static inline int swapcache_prepare(swp_entry_t swp)
+{
+     return 0;
+}
+
 static inline void swap_free(swp_entry_t swp)
 {
 }
mm/memory.c +1 -0
index 7e1f4849463a..7059230d0a54 100644
@@ -3799,6 +3799,7 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
      struct page *page;
      struct swap_info_struct *si = NULL;
      rmap_t rmap_flags = RMAP_NONE;
+     bool need_clear_cache = false;
      bool exclusive = false;
      swp_entry_t entry;
      pte_t pte;
@@ -3867,6 +3868,20 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
      if (!folio) {
              if (data_race(si->flags & SWP_SYNCHRONOUS_IO) &&
                  __swap_count(entry) == 1) {
mm/memory.c vm_fault_t do_swap_page(struct vm_fault *vmf) · line 3880
Quoted 10 lines
mm/memory.c +10 -0
+                     /*
+                      * Prevent parallel swapin from proceeding with
+                      * the cache flag. Otherwise, another thread may
+                      * finish swapin first, free the entry, and swapout
+                      * reusing the same entry. It's undetectable as
+                      * pte_same() returns true due to entry reuse.
+                      */
+                     if (swapcache_prepare(entry)) {
+                             /* Relax a bit to prevent rapid repeated page faults */
+                             schedule();
The current task may be chosen in schedule(). So, I think that we should use cond_resched() here.

Quoted 7 lines

Quoted 4 lines
I think if we are worried about current task got chosen again we can use schedule_timeout_uninterruptible(1) here. Isn't cond_resched still __schedule() and and it can even get omitted, so it should be "weaker" IIUC.
schedule_timeout_uninterruptible(1) will introduce 1ms latency for the second task. That may kill performance of some workloads.

Just found that the cond_sched() in __read_swap_cache_async() has been changed to schedule_timeout_uninterruptible(1) to fix some live lock. Details are in the description of commit 029c4628b2eb ("mm: swap: get rid of livelock in swapin readahead"). I think the similar issue may happen here too. So, we must use schedule_timeout_uninterruptible(1) here until some other better idea becomes available.

--
Best Regards,
Huang, Ying

Kairui Song

author

2024-02-19 03:10 UTC #18 lore raw

Quoted 170 lines · Ying

Quoted 169 lines · Ying
Quoted 167 lines · Kairui Song
Quoted 165 lines · Ying
Quoted 41 lines · Kairui Song
From: Kairui Song <[email protected]>
When skipping swapcache for SWP_SYNCHRONOUS_IO, if two or more threads swapin the same entry at the same time, they get different pages (A, B). Before one thread (T0) finishes the swapin and installs page (A) to the PTE, another thread (T1) could finish swapin of page (B), swap_free the entry, then swap out the possibly modified page reusing the same entry. It breaks the pte_same check in (T0) because PTE value is unchanged, causing ABA problem. Thread (T0) will install a stalled page (A) into the PTE and cause data corruption.
One possible callstack is like this:
CPU0                                 CPU1
----                                 ----
do_swap_page()                       do_swap_page() with same entry
<direct swapin path>                 <direct swapin path>
<alloc page A>                       <alloc page B>
swap_read_folio() <- read to page A  swap_read_folio() <- read to page B
<slow on later locks or interrupt>   <finished swapin first>
...                                  set_pte_at()
swap_free() <- entry is free <write to page B, now page A stalled> <swap out page B to same swap entry>
pte_same() <- Check pass, PTE seems
unchanged, but page A
is stalled!
swap_free() <- page B content lost!
set_pte_at() <- staled page A installed!
And besides, for ZRAM, swap_free() allows the swap device to discard the entry content, so even if page (B) is not modified, if swap_read_folio() on CPU0 happens later than swap_free() on CPU1, it may also cause data loss.
To fix this, reuse swapcache_prepare which will pin the swap entry using the cache flag, and allow only one thread to pin it. Release the pin after PT unlocked. Racers will simply wait since it's a rare and very short event. A schedule() call is added to avoid wasting too much CPU or adding too much noise to perf statistics
Quoted 4 lines
Other methods like increasing the swap count don't seem to be a good idea after some tests, that will cause racers to fall back to use the swap cache again. Parallel swapin using different methods leads to a much more complex scenario.
The swap entry may be put in swap cache by some parallel code path anyway. So, we always need to consider that when reasoning the code.
Quoted 99 lines
Reproducer:
This race issue can be triggered easily using a well constructed reproducer and patched brd (with a delay in read path) [1]:
With latest 6.8 mainline, race caused data loss can be observed easily:
$ gcc -g -lpthread test-thread-swap-race.c && ./a.out
  Polulating 32MB of memory region...
  Keep swapping out...
  Starting round 0...
  Spawning 65536 workers...
  32746 workers spawned, wait for done...
  Round 0: Error on 0x5aa00, expected 32746, got 32743, 3 data loss!
  Round 0: Error on 0x395200, expected 32746, got 32743, 3 data loss!
  Round 0: Error on 0x3fd000, expected 32746, got 32737, 9 data loss!
  Round 0 Failed, 15 data loss!
This reproducer spawns multiple threads sharing the same memory region using a small swap device. Every two threads updates mapped pages one by one in opposite direction trying to create a race, with one dedicated thread keep swapping out the data out using madvise.
The reproducer created a reproduce rate of about once every 5 minutes, so the race should be totally possible in production.
After this patch, I ran the reproducer for over a few hundred rounds and no data loss observed.
Performance overhead is minimal, microbenchmark swapin 10G from 32G zram:
Before:     10934698 us
After:      11157121 us
Non-direct: 13155355 us (Dropping SWP_SYNCHRONOUS_IO flag)
Fixes 0bcac06f27d7 ("mm, swap: skip swapcache for swapin of synchronous device")
Link https://github.com/ryncsn/emm-test-project/tree/master/swap-stress-race [1]
Reported-by "Huang, Ying" <[email protected]>
Closes https://lore.kernel.org/lkml/[email protected]/ open here
Signed-off-by Kairui Song <[email protected]>
Cc [email protected]
---
Update from V2:
- Add a schedule() if raced to prevent repeated page faults wasting CPU and add noise to perf statistics.
- Use a bool to state the special case instead of reusing existing variables fixing error handling [Minchan Kim].
V2: https://lore.kernel.org/all/[email protected]/
Update from V1:
- Add some words on ZRAM case, it will discard swap content on swap_free so the race window is a bit different but cure is the same. [Barry Song]
- Update comments make it cleaner [Huang, Ying]
- Add a function place holder to fix CONFIG_SWAP=n built [SeongJae Park]
- Update the commit message and summary, refer to SWP_SYNCHRONOUS_IO instead of "direct swapin path" [Yu Zhao]
- Update commit message.
- Collect Review and Acks.
V1: https://lore.kernel.org/all/[email protected]/
 include/linux/swap.h |  5 +++++
 mm/memory.c          | 20 ++++++++++++++++++++
 mm/swap.h            |  5 +++++
 mm/swapfile.c        | 13 +++++++++++++
 4 files changed, 43 insertions(+)
include/linux/swap.h +5 -0
index 4db00ddad261..8d28f6091a32 100644
@@ -549,6 +549,11 @@ static inline int swap_duplicate(swp_entry_t swp)
      return 0;
 }
 
+static inline int swapcache_prepare(swp_entry_t swp)
+{
+     return 0;
+}
+
 static inline void swap_free(swp_entry_t swp)
 {
 }
mm/memory.c +1 -0
index 7e1f4849463a..7059230d0a54 100644
@@ -3799,6 +3799,7 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
      struct page *page;
      struct swap_info_struct *si = NULL;
      rmap_t rmap_flags = RMAP_NONE;
+     bool need_clear_cache = false;
      bool exclusive = false;
      swp_entry_t entry;
      pte_t pte;
@@ -3867,6 +3868,20 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
      if (!folio) {
              if (data_race(si->flags & SWP_SYNCHRONOUS_IO) &&
                  __swap_count(entry) == 1) {
mm/memory.c vm_fault_t do_swap_page(struct vm_fault *vmf) · line 3880
Quoted 10 lines
mm/memory.c +10 -0
+                     /*
+                      * Prevent parallel swapin from proceeding with
+                      * the cache flag. Otherwise, another thread may
+                      * finish swapin first, free the entry, and swapout
+                      * reusing the same entry. It's undetectable as
+                      * pte_same() returns true due to entry reuse.
+                      */
+                     if (swapcache_prepare(entry)) {
+                             /* Relax a bit to prevent rapid repeated page faults */
+                             schedule();
The current task may be chosen in schedule(). So, I think that we should use cond_resched() here.

Quoted 7 lines

Quoted 7 lines
Quoted 4 lines
I think if we are worried about current task got chosen again we can use schedule_timeout_uninterruptible(1) here. Isn't cond_resched still __schedule() and and it can even get omitted, so it should be "weaker" IIUC.
schedule_timeout_uninterruptible(1) will introduce 1ms latency for the second task. That may kill performance of some workloads.

It actually calls schedule_timeout so it should be a 1 jiffy latency, not 1ms, right?

/**
* schedule_timeout - sleep until timeout
* @timeout: timeout value in jiffies
...

But I think what we really want here is actually the set_current_state to force yield CPU for a short period. The latency should be mild.

Quoted 7 lines

Just found that the cond_sched() in __read_swap_cache_async() has been changed to schedule_timeout_uninterruptible(1) to fix some live lock. Details are in the description of commit 029c4628b2eb ("mm: swap: get rid of livelock in swapin readahead"). I think the similar issue may happen here too. So, we must use schedule_timeout_uninterruptible(1) here until some other better idea becomes available.

Indeed, I'll switch to schedule_timeout_uninterruptible(1). I've tested and posted the result with schedule_timeout_uninterruptible(1) before, it looked fine, or even better.

Kairui Song

author

2024-02-19 07:29 UTC #19 lore raw

Quoted 183 lines · Kairui Song

Quoted 170 lines · Ying
Quoted 169 lines · Ying
Quoted 167 lines · Kairui Song
Quoted 165 lines · Ying
Quoted 41 lines · Kairui Song
From: Kairui Song <[email protected]>
When skipping swapcache for SWP_SYNCHRONOUS_IO, if two or more threads swapin the same entry at the same time, they get different pages (A, B). Before one thread (T0) finishes the swapin and installs page (A) to the PTE, another thread (T1) could finish swapin of page (B), swap_free the entry, then swap out the possibly modified page reusing the same entry. It breaks the pte_same check in (T0) because PTE value is unchanged, causing ABA problem. Thread (T0) will install a stalled page (A) into the PTE and cause data corruption.
One possible callstack is like this:
CPU0                                 CPU1
----                                 ----
do_swap_page()                       do_swap_page() with same entry
<direct swapin path>                 <direct swapin path>
<alloc page A>                       <alloc page B>
swap_read_folio() <- read to page A  swap_read_folio() <- read to page B
<slow on later locks or interrupt>   <finished swapin first>
...                                  set_pte_at()
swap_free() <- entry is free <write to page B, now page A stalled> <swap out page B to same swap entry>
pte_same() <- Check pass, PTE seems
unchanged, but page A
is stalled!
swap_free() <- page B content lost!
set_pte_at() <- staled page A installed!
And besides, for ZRAM, swap_free() allows the swap device to discard the entry content, so even if page (B) is not modified, if swap_read_folio() on CPU0 happens later than swap_free() on CPU1, it may also cause data loss.
To fix this, reuse swapcache_prepare which will pin the swap entry using the cache flag, and allow only one thread to pin it. Release the pin after PT unlocked. Racers will simply wait since it's a rare and very short event. A schedule() call is added to avoid wasting too much CPU or adding too much noise to perf statistics
Quoted 4 lines
Other methods like increasing the swap count don't seem to be a good idea after some tests, that will cause racers to fall back to use the swap cache again. Parallel swapin using different methods leads to a much more complex scenario.
The swap entry may be put in swap cache by some parallel code path anyway. So, we always need to consider that when reasoning the code.
Quoted 99 lines
Reproducer:
This race issue can be triggered easily using a well constructed reproducer and patched brd (with a delay in read path) [1]:
With latest 6.8 mainline, race caused data loss can be observed easily:
$ gcc -g -lpthread test-thread-swap-race.c && ./a.out
  Polulating 32MB of memory region...
  Keep swapping out...
  Starting round 0...
  Spawning 65536 workers...
  32746 workers spawned, wait for done...
  Round 0: Error on 0x5aa00, expected 32746, got 32743, 3 data loss!
  Round 0: Error on 0x395200, expected 32746, got 32743, 3 data loss!
  Round 0: Error on 0x3fd000, expected 32746, got 32737, 9 data loss!
  Round 0 Failed, 15 data loss!
This reproducer spawns multiple threads sharing the same memory region using a small swap device. Every two threads updates mapped pages one by one in opposite direction trying to create a race, with one dedicated thread keep swapping out the data out using madvise.
The reproducer created a reproduce rate of about once every 5 minutes, so the race should be totally possible in production.
After this patch, I ran the reproducer for over a few hundred rounds and no data loss observed.
Performance overhead is minimal, microbenchmark swapin 10G from 32G zram:
Before:     10934698 us
After:      11157121 us
Non-direct: 13155355 us (Dropping SWP_SYNCHRONOUS_IO flag)
Fixes 0bcac06f27d7 ("mm, swap: skip swapcache for swapin of synchronous device")
Link https://github.com/ryncsn/emm-test-project/tree/master/swap-stress-race [1]
Reported-by "Huang, Ying" <[email protected]>
Closes https://lore.kernel.org/lkml/[email protected]/ open here
Signed-off-by Kairui Song <[email protected]>
Cc [email protected]
---
Update from V2:
- Add a schedule() if raced to prevent repeated page faults wasting CPU and add noise to perf statistics.
- Use a bool to state the special case instead of reusing existing variables fixing error handling [Minchan Kim].
V2: https://lore.kernel.org/all/[email protected]/
Update from V1:
- Add some words on ZRAM case, it will discard swap content on swap_free so the race window is a bit different but cure is the same. [Barry Song]
- Update comments make it cleaner [Huang, Ying]
- Add a function place holder to fix CONFIG_SWAP=n built [SeongJae Park]
- Update the commit message and summary, refer to SWP_SYNCHRONOUS_IO instead of "direct swapin path" [Yu Zhao]
- Update commit message.
- Collect Review and Acks.
V1: https://lore.kernel.org/all/[email protected]/
 include/linux/swap.h |  5 +++++
 mm/memory.c          | 20 ++++++++++++++++++++
 mm/swap.h            |  5 +++++
 mm/swapfile.c        | 13 +++++++++++++
 4 files changed, 43 insertions(+)
include/linux/swap.h +5 -0
index 4db00ddad261..8d28f6091a32 100644
@@ -549,6 +549,11 @@ static inline int swap_duplicate(swp_entry_t swp)
      return 0;
 }
 
+static inline int swapcache_prepare(swp_entry_t swp)
+{
+     return 0;
+}
+
 static inline void swap_free(swp_entry_t swp)
 {
 }
mm/memory.c +1 -0
index 7e1f4849463a..7059230d0a54 100644
@@ -3799,6 +3799,7 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
      struct page *page;
      struct swap_info_struct *si = NULL;
      rmap_t rmap_flags = RMAP_NONE;
+     bool need_clear_cache = false;
      bool exclusive = false;
      swp_entry_t entry;
      pte_t pte;
@@ -3867,6 +3868,20 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
      if (!folio) {
              if (data_race(si->flags & SWP_SYNCHRONOUS_IO) &&
                  __swap_count(entry) == 1) {
mm/memory.c vm_fault_t do_swap_page(struct vm_fault *vmf) · line 3880
Quoted 10 lines
mm/memory.c +10 -0
+                     /*
+                      * Prevent parallel swapin from proceeding with
+                      * the cache flag. Otherwise, another thread may
+                      * finish swapin first, free the entry, and swapout
+                      * reusing the same entry. It's undetectable as
+                      * pte_same() returns true due to entry reuse.
+                      */
+                     if (swapcache_prepare(entry)) {
+                             /* Relax a bit to prevent rapid repeated page faults */
+                             schedule();
The current task may be chosen in schedule(). So, I think that we should use cond_resched() here.
Quoted 7 lines
Quoted 7 lines
Quoted 4 lines
I think if we are worried about current task got chosen again we can use schedule_timeout_uninterruptible(1) here. Isn't cond_resched still __schedule() and and it can even get omitted, so it should be "weaker" IIUC.
schedule_timeout_uninterruptible(1) will introduce 1ms latency for the second task. That may kill performance of some workloads.
It actually calls schedule_timeout so it should be a 1 jiffy latency, not 1ms, right?

Quoted 7 lines

/**
* schedule_timeout - sleep until timeout
* @timeout: timeout value in jiffies
...
But I think what we really want here is actually the set_current_state to force yield CPU for a short period. The latency should be mild.

I just forgot 1 jiffy >= 1 ms here, and uninterruptible should make it unable to wakeup until timeout...

Quoted 10 lines

Quoted 6 lines
Just found that the cond_sched() in __read_swap_cache_async() has been changed to schedule_timeout_uninterruptible(1) to fix some live lock. Details are in the description of commit 029c4628b2eb ("mm: swap: get rid of livelock in swapin readahead"). I think the similar issue may happen here too. So, we must use schedule_timeout_uninterruptible(1) here until some other better idea becomes available.
Indeed, I'll switch to schedule_timeout_uninterruptible(1). I've tested and posted the result with schedule_timeout_uninterruptible(1) before, it looked fine, or even better.

But this should be still the same though, the minor/major fault ratio in previous test result [1] shows the race on ZRAM even with threads set to race on purpose, the chance is low, and thanks for the info on mentioning another commit!

[1] https://lore.kernel.org/all/CAMgjq7BvTJmxrWQOJvkLt4g_jnvmx07NdU63sGeRMGde4Ov=gA@mail.gmail.com/