mm/gup: stop leaking pinned pages in low memory conditions
Subsystems (2)
John Hubbard
reviewerv2 cover mm/gup: stop leaking pinned pages in low memory conditions
Changes since v1 [1]:
1) Thanks to David Hildenbrand for this part: added a second patch to fix the same issue (incomplete error handling for the return value from check_and_migrate_movable_folios()), but this time in memfd_pin_folios().
Please note that I am not set up to test memfd things, so at this point all I can claim is that patch #2 doesn't prevent my test machine from booting. :)
2) Fixed a typo in the comment, in the first patch.
3) Added review and ack tags.
[1] https://lore.kernel.org/[email protected]
John Hubbard (2):
mm/gup: stop leaking pinned pages in low memory conditions mm/gup: memfd: stop leaking pinned pages in low memory conditions
mm/gup.c | 19 ++++++++++++++----- 1 file changed, 14 insertions(+), 5 deletions(-)
base-commit: 274995ffe748bbee39f1ca65c6a31a1800140b27
Signature · 1 line
2.47.0
John Hubbard
reviewerPATCH v2 1/2 mm/gup: stop leaking pinned pages in low memory conditions
If a driver tries to call any of the pin_user_pages*(FOLL_LONGTERM) family of functions, and requests "too many" pages, then the call will erroneously leave pages pinned. This is visible in user space as an actual memory leak.
Repro is trivial: just make enough pin_user_pages(FOLL_LONGTERM) calls to exhaust memory.
The root cause of the problem is this sequence, within __gup_longterm_locked():
__get_user_pages_locked()
rc = check_and_migrate_movable_pages()
...which gets retried in a loop. The loop error handling is incomplete, clearly due to a somewhat unusual and complicated tri-state error API. But anyway, if -ENOMEM, or in fact, any unexpected error is returned from check_and_migrate_movable_pages(), then __gup_longterm_locked() happily returns the error, while leaving the pages pinned.
In the failed case, which is an app that requests (via a device driver) 30720000000 bytes to be pinned, and then exits, I see this:
$ grep foll /proc/vmstat
nr_foll_pin_acquired 7502048
nr_foll_pin_released 2048And after applying this patch, it returns to balanced pins:
$ grep foll /proc/vmstat
nr_foll_pin_acquired 7502048
nr_foll_pin_released 7502048Fix this by unpinning the pages that __get_user_pages_locked() has pinned, in such error cases.
--- mm/gup.c | 11 +++++++++++ 1 file changed, 11 insertions(+)
mm/gup.c +11 -0
@@ -2492,6 +2492,17 @@ static long __gup_longterm_locked(struct mm_struct *mm,/* FOLL_LONGTERM implies FOLL_PIN */ rc =check_and_migrate_movable_pages (nr_pinned_pages, pages); + +/* +* The __get_user_pages_locked() call happens before we know if +* it's possible to successfully complete the whole operation. +* To compensate for this, if we get an unexpected error (such +* as -ENOMEM) then we must unpin everything, before erroring +* out. +*/ +if (rc != -EAGAIN && rc !=0 ) +unpin_user_pages (pages, nr_pinned_pages); + }while (rc == -EAGAIN );memalloc_pin_restore (flags);return rc ? rc : nr_pinned_pages;
Signature · 1 line
2.47.0
Quoted 11 lines · John Hubbard
If a driver tries to call any of the pin_user_pages*(FOLL_LONGTERM) family of functions, and requests "too many" pages, then the call will erroneously leave pages pinned. This is visible in user space as an actual memory leak.
Repro is trivial: just make enough
pin_user_pages(FOLL_LONGTERM)calls to exhaust memory.The root cause of the problem is this sequence, within
__gup_longterm_locked():
Quoted 8 lines
__get_user_pages_locked()
rc =check_and_migrate_movable_pages()...which gets retried in a loop. The loop error handling is incomplete, clearly due to a somewhat unusual and complicated tri-state error API. But anyway, if
-ENOMEM, or in fact, any unexpected error is returned fromcheck_and_migrate_movable_pages(), then__gup_longterm_locked()happily returns the error, while leaving the pages pinned.
Sorry for another comment, I am taking my time to look into the code again in more detail ...
migrate_longterm_unpinnable_folios() will always unpin all pages: no matter which error it returns.
a) If it returns -EAGAIN, it unpinned all folios
b) If it returns any error it first calls unpin_folios().
So shouldn't the fix just be in check_and_migrate_movable_pages()?
mm/gup.c +6 -3
@@ -2403,8 +2403,9 @@ static int migrate_longterm_unpinnable_folios(* -EAGAIN. The caller should re-pin the entire range with FOLL_PIN and then * call this routine again. * -* If an error other than -EAGAIN occurs, this indicates a migration failure. -* The caller should give up, and propagate the error back up the call stack. +* If an error occurs, all folios are unpinned. If an error other than +* -EAGAIN occurs, this indicates a migration failure. The caller should give +* up, and propagate the error back up the call stack. * * If everything is OK and all folios in the range are allowed to be pinned, * then this routine leaves all folios pinned and returns zero for success.
@@ -2437,8 +2438,10 @@ static long check_and_migrate_movable_pages(unsigned long nr_pages,long i, ret; folios =kmalloc_array (nr_pages,sizeof (*folios),GFP_KERNEL ); -if (!folios) +if (!folios) { +unpin_user_pages (pages, nr_pages);return -ENOMEM ; + }for (i =0 ; i < nr_pages; i++) folios[i] =page_folio (pages[i]);
Then, check_and_migrate_movable_pages() will never return with an error and having folios pinned.
If check_and_migrate_movable_pages() -> check_and_migrate_movable_folios() returns "0", all folios remain pinned an no harm is done.
Consequently, I think patch #2 is not really required, because it doesn't perform the temporary allocation that could fail with -ENOMEM.
Sorry for taking a closer look only now ...
Signature · 3 lines
Cheers, David / dhildenb
John Hubbard
reviewerQuoted 21 lines · David Hildenbrand
Quoted 11 lines · John Hubbard
If a driver tries to call any of the pin_user_pages*(FOLL_LONGTERM) family of functions, and requests "too many" pages, then the call will erroneously leave pages pinned. This is visible in user space as an actual memory leak.
Repro is trivial: just make enough
pin_user_pages(FOLL_LONGTERM)calls to exhaust memory.The root cause of the problem is this sequence, within
__gup_longterm_locked():Quoted 8 lines
__get_user_pages_locked()
rc =check_and_migrate_movable_pages()...which gets retried in a loop. The loop error handling is incomplete, clearly due to a somewhat unusual and complicated tri-state error API. But anyway, if
-ENOMEM, or in fact, any unexpected error is returned fromcheck_and_migrate_movable_pages(), then__gup_longterm_locked()happily returns the error, while leaving the pages pinned.
Quoted 8 lines
Sorry for another comment, I am taking my time to look into the code again in more detail ...
migrate_longterm_unpinnable_folios()will always unpin all pages: no matter which error it returns.a) If it returns
-EAGAIN, it unpinned all folios
b) If it returns any error it first callsunpin_folios().So shouldn't the fix just be in
check_and_migrate_movable_pages()?
OK, sure. It's a little odd from a layering point of view, because the callee "helpfully" unpins the pages for you (wheee!), but the updated comment highlights that, at least.
And actually this whole thing of "pin the pages, just for a short time, even though you're not allowed to" is partly why this area is so entertaining.
Quoted 43 lines
mm/gup.c +0
-0@@ -2403,8 +2403,9 @@ static int migrate_longterm_unpinnable_folios( * -EAGAIN. The caller should re-pin the entire range with FOLL_PINand then
* call this routine again.
*
- * If an error other than-EAGAINoccurs, this indicates a migration failure.
- * The caller should give up, and propagate the error back up the call stack.
+ * If an error occurs, all folios are unpinned. If an error other than
+ *-EAGAINoccurs, this indicates a migration failure. The caller should give
+ * up, and propagate the error back up the call stack.
*
* If everything is OK and all folios in the range are allowed to be
pinned,
* then this routine leaves all folios pinned and returns zero for
success.mm/gup.c +3
-1@@ -2437,8 +2438,10 @@ static long check_and_migrate_movable_pages(unsigned long nr_pages, long i, ret; folios = kmalloc_array(nr_pages, sizeof(*folios), GFP_KERNEL); - if (!folios) + if (!folios) { + unpin_user_pages(pages, nr_pages); return -ENOMEM; + } for (i = 0; i < nr_pages; i++)folios[i] =
page_folio(pages[i]);Then,
check_and_migrate_movable_pages()will never return with an error and having folios pinned.
Quoted 7 lines
If
check_and_migrate_movable_pages()->check_and_migrate_movable_folios()returns "0", all folios remain pinned an no harm is done.Consequently, I think patch #2 is not really required, because it doesn't perform the temporary allocation that could fail with
-ENOMEM.
Yes!
Quoted 3 lines
Sorry for taking a closer look only now ...
It's all still in review, so the timing is perfectly fine. I really appreciate the closer look, it's definitely making things better.
thanks,
Signature · 1 line
John Hubbard
Quoted 2 lines · John Hubbard
Quoted 1 line · David Hildenbrand
On 18.10.24 03:17, John Hubbard wrote:
[...]
Quoted 2 lines
And actually this whole thing of "pin the pages, just for a short time, even though you're not allowed to" is partly why this area is so entertaining.
I'm looking at your v3 but as an aside I disagree with this statement. AFAIK you're always allowed to pin the pages for a short time (ie. !FOLL_LONGTERM), or did I misunderstand your comment?
Quoted 55 lines
Quoted 35 lines
mm/gup.c +0
-0@@ -2403,8 +2403,9 @@ static int migrate_longterm_unpinnable_folios( * -EAGAIN. The caller should re-pin the entire range with
FOLL_PINand then
* call this routine again.
*
- * If an error other than-EAGAINoccurs, this indicates a migration failure.
- * The caller should give up, and propagate the error back up the call stack.
+ * If an error occurs, all folios are unpinned. If an error other than
+ *-EAGAINoccurs, this indicates a migration failure. The caller should give
+ * up, and propagate the error back up the call stack.
*
* If everything is OK and all folios in the range are allowed to
be pinned,
* then this routine leaves all folios pinned and returns zero for
success.mm/gup.c +3
-1@@ -2437,8 +2438,10 @@ static long check_and_migrate_movable_pages(unsigned long nr_pages, long i, ret; folios = kmalloc_array(nr_pages, sizeof(*folios),GFP_KERNEL ); - if (!folios) + if (!folios) { + unpin_user_pages(pages, nr_pages); return -ENOMEM; + } for (i = 0; i < nr_pages; i++) folios[i] = page_folio(pages[i]);Then,
check_and_migrate_movable_pages()will never return with anQuoted 9 lines
error and
having folios pinned.
Ifcheck_and_migrate_movable_pages()->check_and_migrate_movable_folios()
returns "0", all folios remain pinned an no harm is done. Consequently, I think patch #2 is not really required, because it doesn't
perform the temporary allocation that could fail with-ENOMEM.Yes!
Quoted 2 lines
Sorry for taking a closer look only now ...
It's all still in review, so the timing is perfectly fine. I really appreciate the closer look, it's definitely making things better.
thanks,
John Hubbard
reviewerQuoted 10 lines · Alistair Popple
Quoted 2 lines · John Hubbard
Quoted 1 line · David Hildenbrand
On 18.10.24 03:17, John Hubbard wrote:
[...]
Quoted 2 lines
And actually this whole thing of "pin the pages, just for a short time, even though you're not allowed to" is partly why this area is so entertaining.
I'm looking at your v3 but as an aside I disagree with this statement. AFAIK you're always allowed to pin the pages for a short time (ie. !FOLL_LONGTERM), or did I misunderstand your comment?
Sort of: short term pins are allowed, but at this point in the code, here:
pin_user_pages(FOLL_PIN | FOLL_LONGTERM)
__gup_longterm_locked()
__get_user_pages_locked(FOLL_PIN | FOLL_LONGTERM)
, just before calling check_and_migrate_movable_pages(), we have already filtered out any cases other than (FOLL_PIN | FOLL_LONGTERM).
And that means that code has taken a *longterm* pin of presumably short duration (this incongruity bothers me), on pages that are not actually allowed to be long term pinned. That also feels imperfect, even though it is supposedly short duration...except that page migration is only sort of short...hmmm.
I'm starting to think that migrating any ZONE_MOVABLE pages away first might be better.
Since I'm already preparing that "wait for folio refcount" idea for migration, which is almost related, I'll take a closer look at this idea while I'm at it.
thanks,
Signature · 1 line
John Hubbard
John Hubbard
reviewerPATCH v2 2/2 mm/gup: memfd: stop leaking pinned pages in low memory conditions
In reply to John HubbardIf check_and_migrate_movable_pages() fails, typically with -ENOMEM, then memfd_pin_folios() will leave pages pinned. Those are leaked forever, and are visible to user space as a memory leak.
Fix this by unpinning the folios that try_grab_folio(FOLL_PIN) has pinned, in such error cases.
--- mm/gup.c | 8 +++----- 1 file changed, 3 insertions(+), 5 deletions(-)
mm/gup.c +3 -5
@@ -3719,12 +3719,10 @@ long memfd_pin_folios(struct file *memfd, loff_t start, loff_t end, ret =check_and_migrate_movable_folios (nr_folios, folios); }while (ret == -EAGAIN ); -memalloc_pin_restore (flags); -return ret ? ret : nr_folios; err:memalloc_pin_restore (flags); -unpin_folios (folios, nr_folios); - -return ret; +if (ret) +unpin_folios (folios, nr_folios); +return ret ? ret : nr_folios; }EXPORT_SYMBOL_GPL (memfd_pin_folios);
Signature · 1 line
2.47.0
Andrew Morton
maintainerOn Thu, 17 Oct 2024 18:17:09 -0700 John Hubbard <[email protected]> wrote:
I added cc:stable to both of these. Which might be inappropriate since "patch #2 is not really required".
Quoted 1 line
mm/gup: stop leaking pinned pages in low memory conditions
In mainline since v6.1!
Quoted 1 line
mm/gup: memfd: stop leaking pinned pages in low memory conditions
Since v6.11.
So these are quite independent fixes. Kernels 6.1.x ... 6.10.x will have the first patch and not the second. That's presumably an untested combination, fingers crossed.
John Hubbard
reviewerQuoted 4 lines · Andrew Morton
On Thu, 17 Oct 2024 18:17:09 -0700 John Hubbard <[email protected]> wrote:
I added cc:stable to both of these. Which might be inappropriate since "patch #2 is not really required".
Right.
Quoted 9 lines
Quoted 2 lines
mm/gup: stop leaking pinned pages in low memory conditions
Fixes 24a95998e9ba ("mm/gup.c: simplify and fix check_and_migrate_movable_pages() return codes")In mainline since v6.1!
Quoted 1 line
mm/gup: memfd: stop leaking pinned pages in low memory conditions
Quoted 9 lines
Fixes 89c1905d9c14 ("mm/gup: introduce memfd_pin_folios() for pinning memfd folios")Since v6.11.
So these are quite independent fixes. Kernels 6.1.x ... 6.10.x will have the first patch and not the second. That's presumably an untested combination, fingers crossed.
Probably fine.
Quoted 1 line
Ah, I'm actually about to send out v3 in a moment, which only has one patch, whose diffs are just comment changes, plus David's latest suggestion:
static long check_and_migrate_movable_pages(unsigned long nr_pages,
struct page **pages)
@@ -2437,8 +2440,10 @@ static long check_and_migrate_movable_pages(unsigned long nr_pages,
long i, ret;
folios = kmalloc_array(nr_pages, sizeof(*folios), GFP_KERNEL);
- if (!folios)
+ if (!folios) {
+ unpin_user_pages(pages, nr_pages);
return -ENOMEM;+ }
for (i = 0; i < nr_pages; i++)
folios[i] = page_folio(pages[i]);thanks,
Signature · 1 line
John Hubbard