When forking a child process, parent write-protects an anonymous page
and COW-shares it with the child being forked using copy_present_pte().
Parent's TLB is flushed right before we drop the parent's mmap_lock in
dup_mmap(). If we get a write-fault before that TLB flush in the parent,
and we end up replacing that anonymous page in the parent process in
do_wp_page() (because, COW-shared with the child), this might lead to
some stale writable TLB entries targeting the wrong (old) page.
Similar issue happened in the past with userfaultfd (see flush_tlb_page()
call inside do_wp_page()).
Lock VMAs of the parent process when forking a child, which prevents
concurrent page faults during fork operation and avoids this issue.
This fix can potentially regress some fork-heavy workloads. Kernel build
time did not show noticeable regression on a 56-core machine while a
stress test mapping 10000 VMAs and forking 5000 times in a tight loop
shows ~7% regression. If such fork time regression is unacceptable,
disabling CONFIG_PER_VMA_LOCK should restore its performance. Further
optimizations are possible if this regression proves to be problematic.
Suggested-by: David Hildenbrand <[email protected]>
Reported-by: Jiri Slaby <[email protected]>
Closes: https://lore.kernel.org/all/[email protected]/
Reported-by: Holger Hoffstätte <[email protected]>
Closes: https://lore.kernel.org/all/[email protected]/
Reported-by: Jacob Young <[email protected]>
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=217624
Fixes: 0bff0aaea03e ("x86/mm: try VMA lock-based page fault handling first")
Cc: [email protected]
Signed-off-by: Suren Baghdasaryan <[email protected]>
---
kernel/fork.c | 6 ++++++
1 file changed, 6 insertions(+)
diff --git a/kernel/fork.c b/kernel/fork.c
index b85814e614a5..2ba918f83bde 100644
--- a/kernel/fork.c
+++ b/kernel/fork.c
@@ -658,6 +658,12 @@ static __latent_entropy int dup_mmap(struct mm_struct *mm,
retval = -EINTR;
goto fail_uprobe_end;
}
+#ifdef CONFIG_PER_VMA_LOCK
+ /* Disallow any page faults before calling flush_cache_dup_mm */
+ for_each_vma(old_vmi, mpnt)
+ vma_start_write(mpnt);
+ vma_iter_set(&old_vmi, 0);
+#endif
flush_cache_dup_mm(oldmm);
uprobe_dup_mmap(oldmm, mm);
/*
--
2.41.0.255.g8b1d071c50-goog
* Suren Baghdasaryan <[email protected]> [230705 21:14]:
> When forking a child process, parent write-protects an anonymous page
> and COW-shares it with the child being forked using copy_present_pte().
> Parent's TLB is flushed right before we drop the parent's mmap_lock in
> dup_mmap(). If we get a write-fault before that TLB flush in the parent,
> and we end up replacing that anonymous page in the parent process in
> do_wp_page() (because, COW-shared with the child), this might lead to
> some stale writable TLB entries targeting the wrong (old) page.
> Similar issue happened in the past with userfaultfd (see flush_tlb_page()
> call inside do_wp_page()).
> Lock VMAs of the parent process when forking a child, which prevents
> concurrent page faults during fork operation and avoids this issue.
> This fix can potentially regress some fork-heavy workloads. Kernel build
> time did not show noticeable regression on a 56-core machine while a
> stress test mapping 10000 VMAs and forking 5000 times in a tight loop
> shows ~7% regression. If such fork time regression is unacceptable,
> disabling CONFIG_PER_VMA_LOCK should restore its performance. Further
> optimizations are possible if this regression proves to be problematic.
>
> Suggested-by: David Hildenbrand <[email protected]>
> Reported-by: Jiri Slaby <[email protected]>
> Closes: https://lore.kernel.org/all/[email protected]/
> Reported-by: Holger Hoffst?tte <[email protected]>
> Closes: https://lore.kernel.org/all/[email protected]/
> Reported-by: Jacob Young <[email protected]>
> Closes: https://bugzilla.kernel.org/show_bug.cgi?id=217624
> Fixes: 0bff0aaea03e ("x86/mm: try VMA lock-based page fault handling first")
> Cc: [email protected]
> Signed-off-by: Suren Baghdasaryan <[email protected]>
Reviewed-by: Liam R. Howlett <[email protected]>
> ---
> kernel/fork.c | 6 ++++++
> 1 file changed, 6 insertions(+)
>
> diff --git a/kernel/fork.c b/kernel/fork.c
> index b85814e614a5..2ba918f83bde 100644
> --- a/kernel/fork.c
> +++ b/kernel/fork.c
> @@ -658,6 +658,12 @@ static __latent_entropy int dup_mmap(struct mm_struct *mm,
> retval = -EINTR;
> goto fail_uprobe_end;
> }
> +#ifdef CONFIG_PER_VMA_LOCK
> + /* Disallow any page faults before calling flush_cache_dup_mm */
> + for_each_vma(old_vmi, mpnt)
> + vma_start_write(mpnt);
> + vma_iter_set(&old_vmi, 0);
> +#endif
> flush_cache_dup_mm(oldmm);
> uprobe_dup_mmap(oldmm, mm);
> /*
> --
> 2.41.0.255.g8b1d071c50-goog
>
On 06.07.23 03:13, Suren Baghdasaryan wrote:
> When forking a child process, parent write-protects an anonymous page
> and COW-shares it with the child being forked using copy_present_pte().
> Parent's TLB is flushed right before we drop the parent's mmap_lock in
> dup_mmap(). If we get a write-fault before that TLB flush in the parent,
> and we end up replacing that anonymous page in the parent process in
> do_wp_page() (because, COW-shared with the child), this might lead to
> some stale writable TLB entries targeting the wrong (old) page.
> Similar issue happened in the past with userfaultfd (see flush_tlb_page()
> call inside do_wp_page()).
> Lock VMAs of the parent process when forking a child, which prevents
> concurrent page faults during fork operation and avoids this issue.
> This fix can potentially regress some fork-heavy workloads. Kernel build
> time did not show noticeable regression on a 56-core machine while a
> stress test mapping 10000 VMAs and forking 5000 times in a tight loop
> shows ~7% regression. If such fork time regression is unacceptable,
> disabling CONFIG_PER_VMA_LOCK should restore its performance. Further
> optimizations are possible if this regression proves to be problematic.
>
> Suggested-by: David Hildenbrand <[email protected]>
> Reported-by: Jiri Slaby <[email protected]>
> Closes: https://lore.kernel.org/all/[email protected]/
> Reported-by: Holger Hoffstätte <[email protected]>
> Closes: https://lore.kernel.org/all/[email protected]/
> Reported-by: Jacob Young <[email protected]>
> Closes: https://bugzilla.kernel.org/show_bug.cgi?id=217624
> Fixes: 0bff0aaea03e ("x86/mm: try VMA lock-based page fault handling first")
> Cc: [email protected]
> Signed-off-by: Suren Baghdasaryan <[email protected]>
Acked-by: David Hildenbrand <[email protected]>
Feel free to keep my ACK on minor changes.
--
Cheers,
David / dhildenb