The VMA count limit check in do_mmap() and do_brk_flags() uses a strict inequality (>), which allows a process's VMA count to exceed the configured sysctl_max_map_count limit by one.
A process with mm->map_count == sysctl_max_map_count will incorrectly pass this check and then exceed the limit upon allocation of a new VMA when its map_count is incremented.
Other VMA allocation paths, such as split_vma(), already use the correct, inclusive (>=) comparison.
Fix this bug by changing the comparison to be inclusive in do_mmap() and do_brk_flags(), bringing them in line with the correct behavior of other allocation paths.
Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Cc: stable@vger.kernel.org Cc: Andrew Morton akpm@linux-foundation.org Cc: David Hildenbrand david@redhat.com Cc: "Liam R. Howlett" Liam.Howlett@oracle.com Cc: Lorenzo Stoakes lorenzo.stoakes@oracle.com Cc: Mike Rapoport rppt@kernel.org Cc: Minchan Kim minchan@kernel.org Cc: Pedro Falcato pfalcato@suse.de Reviewed-by: David Hildenbrand david@redhat.com Reviewed-by: Lorenzo Stoakes lorenzo.stoakes@oracle.com Reviewed-by: Pedro Falcato pfalcato@suse.de Acked-by: SeongJae Park sj@kernel.org Signed-off-by: Kalesh Singh kaleshsingh@google.com ---
Changes in v3: - Collect Reviewed-by and Acked-by tags.
Changes in v2: - Fix mmap check, per Pedro
mm/mmap.c | 2 +- mm/vma.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/mm/mmap.c b/mm/mmap.c index 644f02071a41..da2cbdc0f87b 100644 --- a/mm/mmap.c +++ b/mm/mmap.c @@ -374,7 +374,7 @@ unsigned long do_mmap(struct file *file, unsigned long addr, return -EOVERFLOW;
/* Too many mappings? */ - if (mm->map_count > sysctl_max_map_count) + if (mm->map_count >= sysctl_max_map_count) return -ENOMEM;
/* diff --git a/mm/vma.c b/mm/vma.c index a2e1ae954662..fba68f13e628 100644 --- a/mm/vma.c +++ b/mm/vma.c @@ -2797,7 +2797,7 @@ int do_brk_flags(struct vma_iterator *vmi, struct vm_area_struct *vma, if (!may_expand_vm(mm, vm_flags, len >> PAGE_SHIFT)) return -ENOMEM;
- if (mm->map_count > sysctl_max_map_count) + if (mm->map_count >= sysctl_max_map_count) return -ENOMEM;
if (security_vm_enough_memory_mm(mm, len >> PAGE_SHIFT))
On Mon, 13 Oct 2025, Kalesh Singh wrote:
The VMA count limit check in do_mmap() and do_brk_flags() uses a strict inequality (>), which allows a process's VMA count to exceed the configured sysctl_max_map_count limit by one.
A process with mm->map_count == sysctl_max_map_count will incorrectly pass this check and then exceed the limit upon allocation of a new VMA when its map_count is incremented.
Other VMA allocation paths, such as split_vma(), already use the correct, inclusive (>=) comparison.
Fix this bug by changing the comparison to be inclusive in do_mmap() and do_brk_flags(), bringing them in line with the correct behavior of other allocation paths.
Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Cc: stable@vger.kernel.org Cc: Andrew Morton akpm@linux-foundation.org Cc: David Hildenbrand david@redhat.com Cc: "Liam R. Howlett" Liam.Howlett@oracle.com Cc: Lorenzo Stoakes lorenzo.stoakes@oracle.com Cc: Mike Rapoport rppt@kernel.org Cc: Minchan Kim minchan@kernel.org Cc: Pedro Falcato pfalcato@suse.de Reviewed-by: David Hildenbrand david@redhat.com Reviewed-by: Lorenzo Stoakes lorenzo.stoakes@oracle.com Reviewed-by: Pedro Falcato pfalcato@suse.de Acked-by: SeongJae Park sj@kernel.org Signed-off-by: Kalesh Singh kaleshsingh@google.com
Changes in v3:
- Collect Reviewed-by and Acked-by tags.
Changes in v2:
- Fix mmap check, per Pedro
mm/mmap.c | 2 +- mm/vma.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/mm/mmap.c b/mm/mmap.c index 644f02071a41..da2cbdc0f87b 100644 --- a/mm/mmap.c +++ b/mm/mmap.c @@ -374,7 +374,7 @@ unsigned long do_mmap(struct file *file, unsigned long addr, return -EOVERFLOW; /* Too many mappings? */
- if (mm->map_count > sysctl_max_map_count)
- if (mm->map_count >= sysctl_max_map_count) return -ENOMEM;
/* diff --git a/mm/vma.c b/mm/vma.c index a2e1ae954662..fba68f13e628 100644 --- a/mm/vma.c +++ b/mm/vma.c @@ -2797,7 +2797,7 @@ int do_brk_flags(struct vma_iterator *vmi, struct vm_area_struct *vma, if (!may_expand_vm(mm, vm_flags, len >> PAGE_SHIFT)) return -ENOMEM;
- if (mm->map_count > sysctl_max_map_count)
- if (mm->map_count >= sysctl_max_map_count) return -ENOMEM;
if (security_vm_enough_memory_mm(mm, len >> PAGE_SHIFT)) -- 2.51.0.760.g7b8bcc2412-goog
Sorry for letting you go so far before speaking up (I had to test what I believed to be true, and had hoped that meanwhile one of your many illustrious reviewers would say so first, but no): it's a NAK from me.
These are not off-by-ones: at the point of these checks, it is not known whether an additional map/vma will have to be added, or the addition will be merged into an existing map/vma. So the checks err on the lenient side, letting you get perhaps one more than the sysctl said, but not allowing any more than that.
Which is all that matters, isn't it? Limiting unrestrained growth.
In this patch you're proposing to change it from erring on the lenient side to erring on the strict side - prohibiting merges at the limit which have been allowed for many years.
Whatever one thinks about the merits of erring on the lenient versus erring on the strict side, I see no reason to make this change now, and most certainly not with a Fixes Cc: stable. There is no danger in the current behaviour; there is danger in prohibiting what was allowed before.
As to the remainder of your series: I have to commend you for doing a thorough and well-presented job, but I cannot myself see the point in changing 21 files for what almost amounts to a max_map_count subsystem. I call it misdirected effort, not at all to my taste, which prefers the straightforward checks already there; but accept that my taste may be out of fashion, so won't stand in the way if others think it worthwhile.
Hugh
* Hugh Dickins hughd@google.com [251014 02:28]:
On Mon, 13 Oct 2025, Kalesh Singh wrote:
The VMA count limit check in do_mmap() and do_brk_flags() uses a strict inequality (>), which allows a process's VMA count to exceed the configured sysctl_max_map_count limit by one.
...
/* Too many mappings? */
- if (mm->map_count > sysctl_max_map_count)
- if (mm->map_count >= sysctl_max_map_count) return -ENOMEM;
/* diff --git a/mm/vma.c b/mm/vma.c index a2e1ae954662..fba68f13e628 100644 --- a/mm/vma.c +++ b/mm/vma.c @@ -2797,7 +2797,7 @@ int do_brk_flags(struct vma_iterator *vmi, struct vm_area_struct *vma, if (!may_expand_vm(mm, vm_flags, len >> PAGE_SHIFT)) return -ENOMEM;
- if (mm->map_count > sysctl_max_map_count)
- if (mm->map_count >= sysctl_max_map_count) return -ENOMEM;
if (security_vm_enough_memory_mm(mm, len >> PAGE_SHIFT)) -- 2.51.0.760.g7b8bcc2412-goog
Sorry for letting you go so far before speaking up (I had to test what I believed to be true, and had hoped that meanwhile one of your many illustrious reviewers would say so first, but no): it's a NAK from me.
These are not off-by-ones: at the point of these checks, it is not known whether an additional map/vma will have to be added, or the addition will be merged into an existing map/vma. So the checks err on the lenient side, letting you get perhaps one more than the sysctl said, but not allowing any more than that.
Which is all that matters, isn't it? Limiting unrestrained growth.
In this patch you're proposing to change it from erring on the lenient side to erring on the strict side - prohibiting merges at the limit which have been allowed for many years.
Whatever one thinks about the merits of erring on the lenient versus erring on the strict side, I see no reason to make this change now, and most certainly not with a Fixes Cc: stable. There is no danger in the current behaviour; there is danger in prohibiting what was allowed before.
Thanks Hugh.
I'm left wondering if the issue is that we are checking in the wrong location. That is, should we be checking so early in the process or later when we know the count adjustment?
But then again, later we may be in mid-operation and find out we're out of room. Other places are even more lenient and allow us to exceed the count for a potential limited time, and we really don't know what's going to happen until we examine what already exists.. So it seems like a lot of effort to avoid one extra vma.
As to the remainder of your series: I have to commend you for doing a thorough and well-presented job, but I cannot myself see the point in changing 21 files for what almost amounts to a max_map_count subsystem. I call it misdirected effort, not at all to my taste, which prefers the straightforward checks already there; but accept that my taste may be out of fashion, so won't stand in the way if others think it worthwhile.
I'm not sure which way I favour, it does seem like a large change to avoid an issue that never existed.
In either case, it seems like a good idea to adjust the comments to state that the count may not change.
Thanks, Liam
On Tue, Oct 14, 2025 at 01:51:31PM -0400, Liam R. Howlett wrote:
In either case, it seems like a good idea to adjust the comments to state that the count may not change.
Yes, I am busy trying to catch up with my backlog (do intend to look at this more closely) but wanted to say - let's please document what we're doing here in a comment.
It's my firm conviction that we in mm need to eliminate as many instances of 'just have to know' or implicit knowledge that is either undocumented or tricky to find (buried in a commit message from X yrs ago for instance).
Cheers, Lorenzo
On Mon, Oct 13, 2025 at 11:28 PM Hugh Dickins hughd@google.com wrote:
On Mon, 13 Oct 2025, Kalesh Singh wrote:
The VMA count limit check in do_mmap() and do_brk_flags() uses a strict inequality (>), which allows a process's VMA count to exceed the configured sysctl_max_map_count limit by one.
A process with mm->map_count == sysctl_max_map_count will incorrectly pass this check and then exceed the limit upon allocation of a new VMA when its map_count is incremented.
Other VMA allocation paths, such as split_vma(), already use the correct, inclusive (>=) comparison.
Fix this bug by changing the comparison to be inclusive in do_mmap() and do_brk_flags(), bringing them in line with the correct behavior of other allocation paths.
Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Cc: stable@vger.kernel.org Cc: Andrew Morton akpm@linux-foundation.org Cc: David Hildenbrand david@redhat.com Cc: "Liam R. Howlett" Liam.Howlett@oracle.com Cc: Lorenzo Stoakes lorenzo.stoakes@oracle.com Cc: Mike Rapoport rppt@kernel.org Cc: Minchan Kim minchan@kernel.org Cc: Pedro Falcato pfalcato@suse.de Reviewed-by: David Hildenbrand david@redhat.com Reviewed-by: Lorenzo Stoakes lorenzo.stoakes@oracle.com Reviewed-by: Pedro Falcato pfalcato@suse.de Acked-by: SeongJae Park sj@kernel.org Signed-off-by: Kalesh Singh kaleshsingh@google.com
Changes in v3:
- Collect Reviewed-by and Acked-by tags.
Changes in v2:
- Fix mmap check, per Pedro
mm/mmap.c | 2 +- mm/vma.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/mm/mmap.c b/mm/mmap.c index 644f02071a41..da2cbdc0f87b 100644 --- a/mm/mmap.c +++ b/mm/mmap.c @@ -374,7 +374,7 @@ unsigned long do_mmap(struct file *file, unsigned long addr, return -EOVERFLOW;
/* Too many mappings? */
if (mm->map_count > sysctl_max_map_count)
if (mm->map_count >= sysctl_max_map_count) return -ENOMEM; /*diff --git a/mm/vma.c b/mm/vma.c index a2e1ae954662..fba68f13e628 100644 --- a/mm/vma.c +++ b/mm/vma.c @@ -2797,7 +2797,7 @@ int do_brk_flags(struct vma_iterator *vmi, struct vm_area_struct *vma, if (!may_expand_vm(mm, vm_flags, len >> PAGE_SHIFT)) return -ENOMEM;
if (mm->map_count > sysctl_max_map_count)
if (mm->map_count >= sysctl_max_map_count) return -ENOMEM; if (security_vm_enough_memory_mm(mm, len >> PAGE_SHIFT))-- 2.51.0.760.g7b8bcc2412-goog
Sorry for letting you go so far before speaking up (I had to test what I believed to be true, and had hoped that meanwhile one of your many illustrious reviewers would say so first, but no): it's a NAK from me.
These are not off-by-ones: at the point of these checks, it is not known whether an additional map/vma will have to be added, or the addition will be merged into an existing map/vma. So the checks err on the lenient side, letting you get perhaps one more than the sysctl said, but not allowing any more than that.
Which is all that matters, isn't it? Limiting unrestrained growth.
In this patch you're proposing to change it from erring on the lenient side to erring on the strict side - prohibiting merges at the limit which have been allowed for many years.
Whatever one thinks about the merits of erring on the lenient versus erring on the strict side, I see no reason to make this change now, and most certainly not with a Fixes Cc: stable. There is no danger in the current behaviour; there is danger in prohibiting what was allowed before.
As to the remainder of your series: I have to commend you for doing a thorough and well-presented job, but I cannot myself see the point in changing 21 files for what almost amounts to a max_map_count subsystem. I call it misdirected effort, not at all to my taste, which prefers the straightforward checks already there; but accept that my taste may be out of fashion, so won't stand in the way if others think it worthwhile.
Hi Hugh,
Thanks for the detailed review and for taking the time to test the behavior.
You've raised a valid point. I wasn't aware of the history behind the lenient check for merges. The lack of a comment, like the one that exists for exceeding the limit in munmap(), led me to misinterpret this as an off-by-one bug. The convention makes sense if we consider potential merges.
If it was in-fact the intended behavior, then I agree we should keep it lenient. It would mean though, that munmap() being able to free a VMA if a split is required (by permitting exceeding the limit by 1) would not work in the case where we have already exceeded the limit. I find this to be inconsistent but this is also the current behavior ...
I will drop this patch and the patch that introduces the vma_count_remaining() helper, as I see your point about it potentially being unnecessary overhead.
Regarding your feedback on the rest of the series, I believe the 3 remaining patches are still valuable on their own.
- The selftest adds a comprehensive tests for VMA operations at the sysctl_max_map_count limit. This will self-document the exact behavior expected, including the leniency for potential merges that you highlighted, preventing the kind of misunderstanding that led to my initial patch.
- The rename of mm_struct->map_count to vma_count, is a straightforward cleanup for code clarity that makes the purpose of the field more explicit.
- The tracepoint adds needed observability for telemetry, allowing us to see when processes are failing in the field due to VMA count limit.
The selftest, is what makes up a large portion of the diff you sited, and with vma_count_remaining() gone the series will not touch nearly as many files.
Would this be an acceptable path forward?
Thanks, Kalesh
Hugh
On Tue, 14 Oct 2025, Kalesh Singh wrote:
On Mon, Oct 13, 2025 at 11:28 PM Hugh Dickins hughd@google.com wrote:
Sorry for letting you go so far before speaking up (I had to test what I believed to be true, and had hoped that meanwhile one of your many illustrious reviewers would say so first, but no): it's a NAK from me.
These are not off-by-ones: at the point of these checks, it is not known whether an additional map/vma will have to be added, or the addition will be merged into an existing map/vma. So the checks err on the lenient side, letting you get perhaps one more than the sysctl said, but not allowing any more than that.
Which is all that matters, isn't it? Limiting unrestrained growth.
In this patch you're proposing to change it from erring on the lenient side to erring on the strict side - prohibiting merges at the limit which have been allowed for many years.
Whatever one thinks about the merits of erring on the lenient versus erring on the strict side, I see no reason to make this change now, and most certainly not with a Fixes Cc: stable. There is no danger in the current behaviour; there is danger in prohibiting what was allowed before.
As to the remainder of your series: I have to commend you for doing a thorough and well-presented job, but I cannot myself see the point in changing 21 files for what almost amounts to a max_map_count subsystem. I call it misdirected effort, not at all to my taste, which prefers the straightforward checks already there; but accept that my taste may be out of fashion, so won't stand in the way if others think it worthwhile.
Hi Hugh,
Thanks for the detailed review and for taking the time to test the behavior.
You've raised a valid point. I wasn't aware of the history behind the lenient check for merges. The lack of a comment, like the one that exists for exceeding the limit in munmap(), led me to misinterpret this as an off-by-one bug. The convention makes sense if we consider potential merges.
Yes, a comment there would be helpful (and I doubt it's worth more than adding a comment); but I did not understand at all, Liam's suggestion for the comment "to state that the count may not change".
If it was in-fact the intended behavior, then I agree we should keep it lenient. It would mean though, that munmap() being able to free a VMA if a split is required (by permitting exceeding the limit by 1) would not work in the case where we have already exceeded the limit. I find this to be inconsistent but this is also the current behavior ...
You're saying that once we go one over the limit, say with a new mmap, an munmap check makes it impossible to munmap that or any other vma?
If that's so, I do agree with you, that's nasty, and I would hate any new code to behave that way. In code that's survived as long as this without troubling anyone, I'm not so sure: but if it's easily fixed (a more lenient check at the munmap end?) that would seem worthwhile.
Ah, but reading again, you say "if a split is required": I guess munmapping the whole vma has no problem; and it's fine for a middle munmap, splitting into three before munmapping the middle, to fail. I suppose it would be nicer if munmaping start or end succeeeded, but I don't think that matters very much in this case.
I will drop this patch and the patch that introduces the vma_count_remaining() helper, as I see your point about it potentially being unnecessary overhead.
Regarding your feedback on the rest of the series, I believe the 3 remaining patches are still valuable on their own.
- The selftest adds a comprehensive tests for VMA operations at the
sysctl_max_map_count limit. This will self-document the exact behavior expected, including the leniency for potential merges that you highlighted, preventing the kind of misunderstanding that led to my initial patch.
- The rename of mm_struct->map_count to vma_count, is a
straightforward cleanup for code clarity that makes the purpose of the field more explicit.
- The tracepoint adds needed observability for telemetry, allowing us
to see when processes are failing in the field due to VMA count limit.
The selftest, is what makes up a large portion of the diff you sited, and with vma_count_remaining() gone the series will not touch nearly as many files.
Would this be an acceptable path forward?
Possibly, if others like it: my concern was to end a misunderstanding (I'm generally much too slow to get involved in cleanups).
Though given that the sysctl is named "max_map_count", I'm not very keen on renaming everything else from map_count to vma_count (and of course I'm not suggesting to rename the sysctl).
Hugh
On Wed, Oct 15, 2025 at 10:05 PM Hugh Dickins hughd@google.com wrote:
On Tue, 14 Oct 2025, Kalesh Singh wrote:
On Mon, Oct 13, 2025 at 11:28 PM Hugh Dickins hughd@google.com wrote:
Sorry for letting you go so far before speaking up (I had to test what I believed to be true, and had hoped that meanwhile one of your many illustrious reviewers would say so first, but no): it's a NAK from me.
These are not off-by-ones: at the point of these checks, it is not known whether an additional map/vma will have to be added, or the addition will be merged into an existing map/vma. So the checks err on the lenient side, letting you get perhaps one more than the sysctl said, but not allowing any more than that.
Which is all that matters, isn't it? Limiting unrestrained growth.
In this patch you're proposing to change it from erring on the lenient side to erring on the strict side - prohibiting merges at the limit which have been allowed for many years.
Whatever one thinks about the merits of erring on the lenient versus erring on the strict side, I see no reason to make this change now, and most certainly not with a Fixes Cc: stable. There is no danger in the current behaviour; there is danger in prohibiting what was allowed before.
As to the remainder of your series: I have to commend you for doing a thorough and well-presented job, but I cannot myself see the point in changing 21 files for what almost amounts to a max_map_count subsystem. I call it misdirected effort, not at all to my taste, which prefers the straightforward checks already there; but accept that my taste may be out of fashion, so won't stand in the way if others think it worthwhile.
Hi Hugh,
Thanks for the detailed review and for taking the time to test the behavior.
You've raised a valid point. I wasn't aware of the history behind the lenient check for merges. The lack of a comment, like the one that exists for exceeding the limit in munmap(), led me to misinterpret this as an off-by-one bug. The convention makes sense if we consider potential merges.
Yes, a comment there would be helpful (and I doubt it's worth more than adding a comment); but I did not understand at all, Liam's suggestion for the comment "to state that the count may not change".
If it was in-fact the intended behavior, then I agree we should keep it lenient. It would mean though, that munmap() being able to free a VMA if a split is required (by permitting exceeding the limit by 1) would not work in the case where we have already exceeded the limit. I find this to be inconsistent but this is also the current behavior ...
You're saying that once we go one over the limit, say with a new mmap, an munmap check makes it impossible to munmap that or any other vma?
If that's so, I do agree with you, that's nasty, and I would hate any new code to behave that way. In code that's survived as long as this without troubling anyone, I'm not so sure: but if it's easily fixed (a more lenient check at the munmap end?) that would seem worthwhile.
Ah, but reading again, you say "if a split is required": I guess munmapping the whole vma has no problem; and it's fine for a middle munmap, splitting into three before munmapping the middle, to fail. I suppose it would be nicer if munmaping start or end succeeeded, but I don't think that matters very much in this case.
Yes, your understanding is correct. I meant that currently, we allow for an munmap() requiring a single split to succeed even if it will temporarily exceed the limit by one, as immediately after we will be removing one of those VMAs. However, if the process has already exceeded the limit, say, due to a non-merging mmap(), then an munmap() requiring a split will fail. It's not a big issue, but I found it inconsistent that this succeeds in some cases and not in others.
I will drop this patch and the patch that introduces the vma_count_remaining() helper, as I see your point about it potentially being unnecessary overhead.
Regarding your feedback on the rest of the series, I believe the 3 remaining patches are still valuable on their own.
- The selftest adds a comprehensive tests for VMA operations at the
sysctl_max_map_count limit. This will self-document the exact behavior expected, including the leniency for potential merges that you highlighted, preventing the kind of misunderstanding that led to my initial patch.
- The rename of mm_struct->map_count to vma_count, is a
straightforward cleanup for code clarity that makes the purpose of the field more explicit.
- The tracepoint adds needed observability for telemetry, allowing us
to see when processes are failing in the field due to VMA count limit.
The selftest, is what makes up a large portion of the diff you sited, and with vma_count_remaining() gone the series will not touch nearly as many files.
Would this be an acceptable path forward?
Possibly, if others like it: my concern was to end a misunderstanding (I'm generally much too slow to get involved in cleanups).
Though given that the sysctl is named "max_map_count", I'm not very keen on renaming everything else from map_count to vma_count (and of course I'm not suggesting to rename the sysctl).
I still believe vma_count is a clearer name for the field, given some existing comments already refer to it as vma count. The inconsistency between vma_count and sysctl_max_map_count can be abstracted away; and the sysctl made non-global. I'll wait for feedback form others on how to proceed.
Thanks for the thorough review and discussion.
-- Kalesh
Hugh
linux-stable-mirror@lists.linaro.org
-
Hugh Dickins -
Kalesh Singh -
Liam R. Howlett -
Lorenzo Stoakes