* [PATCH] KVM: Fix an issue in non-preemptible kernel.
@ 2019-09-02 9:02 Gary Fu
2019-09-04 14:02 ` Paul Burton
0 siblings, 1 reply; 4+ messages in thread
From: Gary Fu @ 2019-09-02 9:02 UTC (permalink / raw)
To: linux-mips; +Cc: Paul Burton, Archer Yan, Gary Fu
Add a cond_resched() to give the scheduler a chance to run madvise
task to avoid endless loop here in non-preemptible kernel.
Otherwise, the kvm_mmu_notifier would have no chance to be descreased
to 0 by madvise task -> syscall -> zap_page_range ->
mmu_notifier_invalidate_range_end ->
__mmu_notifier_invalidate_range_end -> invalidate_range_end ->
kvm_mmu_notifier_invalidate_range_end, as the madvise task would be
scheduled when running unmap_single_vma -> unmap_page_range ->
zap_p4d_range -> zap_pud_range -> zap_pmd_range -> cond_resched which
is called before mmu_notifier_invalidate_range_end in zap_page_range.
Signed-off-by: Gary Fu <qfu@wavecomp.com>
---
arch/mips/kvm/mmu.c | 16 ++++++++++++++++
1 file changed, 16 insertions(+)
diff --git a/arch/mips/kvm/mmu.c b/arch/mips/kvm/mmu.c
index 97e538a8c1be..e52e63d225f4 100644
--- a/arch/mips/kvm/mmu.c
+++ b/arch/mips/kvm/mmu.c
@@ -746,6 +746,22 @@ static int kvm_mips_map_page(struct kvm_vcpu *vcpu, unsigned long gpa,
*/
spin_unlock(&kvm->mmu_lock);
kvm_release_pfn_clean(pfn);
+ /*
+ * Add a cond_resched() to give the scheduler a chance to run
+ * madvise task to avoid endless loop here in non-preemptible
+ * kernel.
+ * Otherwise, the kvm_mmu_notifier would have no chance to be
+ * descreased to 0 by madvise task -> syscall -> zap_page_range
+ * -> mmu_notifier_invalidate_range_end ->
+ * __mmu_notifier_invalidate_range_end -> invalidate_range_end
+ * -> kvm_mmu_notifier_invalidate_range_end, as the madvise task
+ * would be scheduled when running unmap_single_vma ->
+ * unmap_page_range -> zap_p4d_range -> zap_pud_range ->
+ * zap_pmd_range -> cond_resched which is called before
+ * mmu_notifier_invalidate_range_end in zap_page_range.
+ */
+ if (need_resched())
+ cond_resched();
goto retry;
}
--
2.17.1
^ permalink raw reply related [flat|nested] 4+ messages in thread
* Re: [PATCH] KVM: Fix an issue in non-preemptible kernel.
2019-09-02 9:02 [PATCH] KVM: Fix an issue in non-preemptible kernel Gary Fu
@ 2019-09-04 14:02 ` Paul Burton
2019-09-05 13:54 ` Gary Fu
0 siblings, 1 reply; 4+ messages in thread
From: Paul Burton @ 2019-09-04 14:02 UTC (permalink / raw)
To: Gary Fu; +Cc: linux-mips, Paul Burton, Archer Yan, James Hogan
Hi Gary,
On Mon, Sep 02, 2019 at 09:02:32AM +0000, Gary Fu wrote:
> Add a cond_resched() to give the scheduler a chance to run madvise
> task to avoid endless loop here in non-preemptible kernel.
Thanks for the patch!
> Otherwise, the kvm_mmu_notifier would have no chance to be descreased
s/descreased/decreased/
(and in the comment too)
> to 0 by madvise task -> syscall -> zap_page_range ->
> mmu_notifier_invalidate_range_end ->
> __mmu_notifier_invalidate_range_end -> invalidate_range_end ->
> kvm_mmu_notifier_invalidate_range_end, as the madvise task would be
> scheduled when running unmap_single_vma -> unmap_page_range ->
> zap_p4d_range -> zap_pud_range -> zap_pmd_range -> cond_resched which
> is called before mmu_notifier_invalidate_range_end in zap_page_range.
I'm not entirely sure I follow - could you clarify whether the task
invoking the madvise syscall is related to the task using KVM?
> Signed-off-by: Gary Fu <qfu@wavecomp.com>
> ---
> arch/mips/kvm/mmu.c | 16 ++++++++++++++++
> 1 file changed, 16 insertions(+)
>
> diff --git a/arch/mips/kvm/mmu.c b/arch/mips/kvm/mmu.c
> index 97e538a8c1be..e52e63d225f4 100644
> --- a/arch/mips/kvm/mmu.c
> +++ b/arch/mips/kvm/mmu.c
> @@ -746,6 +746,22 @@ static int kvm_mips_map_page(struct kvm_vcpu *vcpu, unsigned long gpa,
> */
> spin_unlock(&kvm->mmu_lock);
> kvm_release_pfn_clean(pfn);
> + /*
> + * Add a cond_resched() to give the scheduler a chance to run
> + * madvise task to avoid endless loop here in non-preemptible
> + * kernel.
> + * Otherwise, the kvm_mmu_notifier would have no chance to be
> + * descreased to 0 by madvise task -> syscall -> zap_page_range
> + * -> mmu_notifier_invalidate_range_end ->
> + * __mmu_notifier_invalidate_range_end -> invalidate_range_end
> + * -> kvm_mmu_notifier_invalidate_range_end, as the madvise task
> + * would be scheduled when running unmap_single_vma ->
> + * unmap_page_range -> zap_p4d_range -> zap_pud_range ->
> + * zap_pmd_range -> cond_resched which is called before
> + * mmu_notifier_invalidate_range_end in zap_page_range.
> + */
> + if (need_resched())
> + cond_resched();
Can we remove the need_resched() check here? cond_resched() already
checks should_resched(0) which tests the same thread-info flag as
need_resched(). So we should be fine to just call cond_resched()
unconditionally.
Thanks,
Paul
> goto retry;
> }
>
> --
> 2.17.1
>
^ permalink raw reply [flat|nested] 4+ messages in thread* RE: [PATCH] KVM: Fix an issue in non-preemptible kernel.
2019-09-04 14:02 ` Paul Burton
@ 2019-09-05 13:54 ` Gary Fu
0 siblings, 0 replies; 4+ messages in thread
From: Gary Fu @ 2019-09-05 13:54 UTC (permalink / raw)
To: Paul Burton; +Cc: linux-mips, Archer Yan, James Hogan
Hi Paul,
Thanks for your review.
Please see my comments below.
-----Original Message-----
From: Paul Burton <pburton@wavecomp.com>
Sent: Wednesday, September 4, 2019 10:03 PM
To: Gary Fu <qfu@wavecomp.com>
Cc: linux-mips@vger.kernel.org; Paul Burton <pburton@wavecomp.com>; Archer Yan <ayan@wavecomp.com>; James Hogan <jhogan@kernel.org>
Subject: Re: [PATCH] KVM: Fix an issue in non-preemptible kernel.
Hi Gary,
On Mon, Sep 02, 2019 at 09:02:32AM +0000, Gary Fu wrote:
> Add a cond_resched() to give the scheduler a chance to run madvise
> task to avoid endless loop here in non-preemptible kernel.
Thanks for the patch!
> Otherwise, the kvm_mmu_notifier would have no chance to be descreased
s/descreased/decreased/
(and in the comment too)
Thank you for your comments. I'll fix these typo errors.
> to 0 by madvise task -> syscall -> zap_page_range ->
> mmu_notifier_invalidate_range_end ->
> __mmu_notifier_invalidate_range_end -> invalidate_range_end ->
> kvm_mmu_notifier_invalidate_range_end, as the madvise task would be
> scheduled when running unmap_single_vma -> unmap_page_range ->
> zap_p4d_range -> zap_pud_range -> zap_pmd_range -> cond_resched which
> is called before mmu_notifier_invalidate_range_end in zap_page_range.
I'm not entirely sure I follow - could you clarify whether the task invoking the madvise syscall is related to the task using KVM?
Yes, the QEMU application invokes the madvise syscall with behavior param MADV_DONTNEED.
When handling GPA faults by creating a new GPA mapping in kvm_mips_map_page, it will be retrying to get available page. In the low memory case, it is waiting for the memory resources freed by madvise syscall with MADV_DONTNEED (QEMU application -> madvise with MADV_DONTNEED -> syscall -> madvise_vma -> madvise_dontneed_free -> madvise_dontneed_single_vma -> zap_page_range). In zap_page_range, after the TLB of given address range is cleared by unmap_single_vma, it will call __mmu_notifier_invalidate_range_end which finally calls kvm_mmu_notifier_invalidate_range_end to decrease mmu_notifier_count to 0. The retrying loop in kvm_mips_map_page checks the mmu_notifier_count and if the value is 0 which indicates that some new page is available for mapping, it will jump out the retrying loop and set up PTE for a new GPA mapping.
During the TLB clearing ( in unmap_single_vma in madvise syscall) mentioned above, it will call cond_resched() per PMD for avoiding occupying CPU for a long time (in case of huge page range zapping). When this happened in the non-preemptible kernel, the retrying loop in kvm_mips_map_page will be running endlessly as there is no chance to reschedule back to madvise syscall to run __mmu_notifier_invalidate_range_end to decrease mmu_notifier_count so that the value of mmu_notifier_count is always 1.
Adding a scheduling point before every retry in kvm_mips_map_page will give the madvise syscall (invoked by QEMU) a chance to be re-scheduled back to zap all the given pages and clear mmu_notifier_count value to let kvm_mips_map_page task jump out the loop.
> Signed-off-by: Gary Fu <qfu@wavecomp.com>
> ---
> arch/mips/kvm/mmu.c | 16 ++++++++++++++++
> 1 file changed, 16 insertions(+)
>
> diff --git a/arch/mips/kvm/mmu.c b/arch/mips/kvm/mmu.c index
> 97e538a8c1be..e52e63d225f4 100644
> --- a/arch/mips/kvm/mmu.c
> +++ b/arch/mips/kvm/mmu.c
> @@ -746,6 +746,22 @@ static int kvm_mips_map_page(struct kvm_vcpu *vcpu, unsigned long gpa,
> */
> spin_unlock(&kvm->mmu_lock);
> kvm_release_pfn_clean(pfn);
> + /*
> + * Add a cond_resched() to give the scheduler a chance to run
> + * madvise task to avoid endless loop here in non-preemptible
> + * kernel.
> + * Otherwise, the kvm_mmu_notifier would have no chance to be
> + * descreased to 0 by madvise task -> syscall -> zap_page_range
> + * -> mmu_notifier_invalidate_range_end ->
> + * __mmu_notifier_invalidate_range_end -> invalidate_range_end
> + * -> kvm_mmu_notifier_invalidate_range_end, as the madvise task
> + * would be scheduled when running unmap_single_vma ->
> + * unmap_page_range -> zap_p4d_range -> zap_pud_range ->
> + * zap_pmd_range -> cond_resched which is called before
> + * mmu_notifier_invalidate_range_end in zap_page_range.
> + */
> + if (need_resched())
> + cond_resched();
Can we remove the need_resched() check here? cond_resched() already checks should_resched(0) which tests the same thread-info flag as need_resched(). So we should be fine to just call cond_resched() unconditionally.
Yes, you're right. Just calling cond_resched() is enough. I'll update the patch.
Thanks,
Paul
> goto retry;
> }
>
> --
> 2.17.1
>
^ permalink raw reply [flat|nested] 4+ messages in thread
* [PATCH] KVM: Fix an issue in non-preemptible kernel.
@ 2019-09-09 2:49 Gary Fu
0 siblings, 0 replies; 4+ messages in thread
From: Gary Fu @ 2019-09-09 2:49 UTC (permalink / raw)
To: linux-mips; +Cc: Paul Burton, jhogan, Archer Yan, Gary Fu
Add a cond_resched() to give the scheduler a chance to run madvise
task to avoid endless loop here in non-preemptible kernel.
Otherwise, the kvm_mmu_notifier would have no chance to be decreased
to 0 by madvise task -> syscall -> zap_page_range ->
mmu_notifier_invalidate_range_end ->
__mmu_notifier_invalidate_range_end -> invalidate_range_end ->
kvm_mmu_notifier_invalidate_range_end, as the madvise task would be
scheduled when running unmap_single_vma -> unmap_page_range ->
zap_p4d_range -> zap_pud_range -> zap_pmd_range -> cond_resched which
is called before mmu_notifier_invalidate_range_end in zap_page_range.
When handling GPA faults by creating a new GPA mapping in
kvm_mips_map_page, it will be retrying to get available page. In the
low memory case, it is waiting for the memory resources freed by
madvise syscall with MADV_DONTNEED (QEMU application -> madvise with
MADV_DONTNEED -> syscall -> madvise_vma -> madvise_dontneed_free ->
madvise_dontneed_single_vma -> zap_page_range). In zap_page_range,
after the TLB of given address range is cleared by unmap_single_vma,
it will call __mmu_notifier_invalidate_range_end which finally calls
kvm_mmu_notifier_invalidate_range_end to decrease mmu_notifier_count
to 0. The retrying loop in kvm_mips_map_page checks the
mmu_notifier_count and if the value is 0 which indicates that some
new page is available for mapping, it will jump out the retrying loop
and set up PTE for a new GPA mapping.
During the TLB clearing (in unmap_single_vma in madvise syscall)
mentioned above, it will call cond_resched() per PMD for avoiding
occupying CPU for a long time (in case of huge page range zapping).
When this happens in the non-preemptible kernel, the retrying loop in
kvm_mips_map_page will be running endlessly as there is no chance to
reschedule back to madvise syscall to run
__mmu_notifier_invalidate_range_end to decrease mmu_notifier_count so
that the value of mmu_notifier_count is always 1.
Adding a scheduling point before every retry in kvm_mips_map_page will
give the madvise syscall (invoked by QEMU) a chance to be re-scheduled
back to zap pages in the given range and clear mmu_notifier_count
value to let kvm_mips_map_page task jump out the loop.
Signed-off-by: Gary Fu <qfu@wavecomp.com>
---
arch/mips/kvm/mmu.c | 48 +++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 48 insertions(+)
diff --git a/arch/mips/kvm/mmu.c b/arch/mips/kvm/mmu.c
index 97e538a8c1be..26bac7e1ea85 100644
--- a/arch/mips/kvm/mmu.c
+++ b/arch/mips/kvm/mmu.c
@@ -746,6 +746,54 @@ static int kvm_mips_map_page(struct kvm_vcpu *vcpu, unsigned long gpa,
*/
spin_unlock(&kvm->mmu_lock);
kvm_release_pfn_clean(pfn);
+ /*
+ * Add a cond_resched() to give the scheduler a chance to run
+ * madvise task to avoid endless loop here in non-preemptible
+ * kernel.
+ * Otherwise, the kvm_mmu_notifier would have no chance to be
+ * decreased to 0 by madvise task -> syscall -> zap_page_range
+ * -> mmu_notifier_invalidate_range_end ->
+ * __mmu_notifier_invalidate_range_end -> invalidate_range_end
+ * -> kvm_mmu_notifier_invalidate_range_end, as the madvise task
+ * would be scheduled when running unmap_single_vma ->
+ * unmap_page_range -> zap_p4d_range -> zap_pud_range ->
+ * zap_pmd_range -> cond_resched which is called before
+ * mmu_notifier_invalidate_range_end in zap_page_range.
+ *
+ * When handling GPA faults by creating a new GPA mapping in
+ * kvm_mips_map_page, it will be retrying to get available
+ * pages.
+ * In the low memory case, it is waiting for the memory
+ * resources freed by madvise syscall with MADV_DONTNEED (QEMU
+ * application -> madvise with MADV_DONTNEED -> syscall ->
+ * madvise_vma -> madvise_dontneed_free ->
+ * madvise_dontneed_single_vma -> zap_page_range). In
+ * zap_page_range, after the TLB of given address range is
+ * cleared by unmap_single_vma, it will call
+ * __mmu_notifier_invalidate_range_end which finally calls
+ * kvm_mmu_notifier_invalidate_range_end to decrease
+ * mmu_notifier_count to 0. The retrying loop in
+ * kvm_mips_map_page checks the mmu_notifier_count and if the
+ * value is 0 which indicates that some new page is available
+ * for mapping, it will jump out the retrying loop and set up
+ * PTE for a new GPA mapping.
+ * During the TLB clearing (in unmap_single_vma in madvise
+ * syscall) mentioned above, it will call cond_resched() per
+ * PMD for avoiding occupying CPU for a long time (in case of
+ * huge page range zapping). When this happens in the
+ * non-preemptible kernel, the retrying loop in
+ * kvm_mips_map_page will be running endlessly as there is no
+ * chance to reschedule back to madvise syscall to run
+ * __mmu_notifier_invalidate_range_end to decrease
+ * mmu_notifier_count so that the value of mmu_notifier_count
+ * is always 1.
+ * Adding a scheduling point before every retry in
+ * kvm_mips_map_page will give the madvise syscall (invoked by
+ * QEMU) a chance to be re-scheduled back to zap pages in the
+ * given range and clear mmu_notifier_count value to let
+ * kvm_mips_map_page task jump out the loop.
+ */
+ cond_resched();
goto retry;
}
--
2.17.1
^ permalink raw reply related [flat|nested] 4+ messages in thread
end of thread, other threads:[~2019-09-09 2:49 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-09-02 9:02 [PATCH] KVM: Fix an issue in non-preemptible kernel Gary Fu
2019-09-04 14:02 ` Paul Burton
2019-09-05 13:54 ` Gary Fu
2019-09-09 2:49 Gary Fu
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).