From: sashiko-bot@kernel.org
To: "Amit Machhiwal" <amachhiw@linux.ibm.com>
Cc: kvm@vger.kernel.org
Subject: Re: [PATCH] KVM: PPC: Book3S HV: Disable preemption around host-side HPT lock users
Date: Wed, 14 Oct 2026 19:09:57 +0000 [thread overview]
Message-ID: <sashiko-outbox-171754@kernel.org> (raw)
In-Reply-To: <20261014185215.71832-1-amachhiw@linux.ibm.com>
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] kvm: missing preempt_disable() in kvm_test_age_rmapp() leading to deadlock
--
commit dbd9d3f82d3dd2c11f69c26045433b125bc70408
Author: Amit Machhiwal <amachhiw@linux.ibm.com>
KVM: PPC: Book3S HV: Disable preemption around host-side HPT lock users
This commit disables preemption around host-side virtual-mode paths that
acquire HPTE_V_HVLOCK and lock_rmap(). This prevents a permanent CPU hang
where a preempted host thread holding either bit-lock blocks a spinning
guest vCPU thread.
> diff --git a/arch/powerpc/kvm/book3s_64_mmu_hv.c b/arch/powerpc/kvm/book3s_64_mmu_hv.c
> index 2ccb3d138f46c..8d85ce3440ae3 100644
> --- a/arch/powerpc/kvm/book3s_64_mmu_hv.c
> +++ b/arch/powerpc/kvm/book3s_64_mmu_hv.c
[ ... ]
> @@ -890,6 +913,7 @@ static bool kvm_age_rmapp(struct kvm *kvm, struct kvm_memory_slot *memslot,
>
> rmapp = &memslot->arch.rmap[gfn - memslot->base_gfn];
> retry:
> + preempt_disable();
> lock_rmap(rmapp);
[Severity: High]
Does this same deadlock risk apply to kvm_test_age_rmapp()?
Looking at kvm_test_age_rmapp() in the same file, it also acquires
the spinning rmap bit-lock without explicitly disabling preemption:
arch/powerpc/kvm/book3s_64_mmu_hv.c:kvm_test_age_rmapp() {
...
lock_rmap(rmapp);
if (*rmapp & KVMPPC_RMAP_REFERENCED)
goto out;
...
}
kvm_test_age_rmapp() is invoked via MMU notifiers (like
mmu_notifier_test_young) holding the PTE lock. Under PREEMPT_RT, the PTE
lock is sleepable and does not disable preemption, so lock_rmap() is
acquired in a preemptible context.
If a host thread holding this bit-lock is preempted, and a guest vCPU
thread on the same CPU executes a virtual-mode path like
kvmppc_virtmode_do_h_remove() (which disables preemption) and attempts to
acquire the same rmap lock via remove_revmap_chain(), could it spin
indefinitely without yielding?
> if (*rmapp & KVMPPC_RMAP_REFERENCED) {
> *rmapp &= ~KVMPPC_RMAP_REFERENCED;
[ ... ]
--
Sashiko AI review · https://sashiko.dev/#/patchset/20261014185215.71832-1-amachhiw@linux.ibm.com?part=1
next prev parent reply other threads:[~2026-10-14 19:09 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-10-14 18:52 [PATCH] KVM: PPC: Book3S HV: Disable preemption around host-side HPT lock users Amit Machhiwal
2026-10-14 19:09 ` sashiko-bot [this message]
2026-10-15 3:44 ` Ritesh Harjani
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=sashiko-outbox-171754@kernel.org \
--to=sashiko-bot@kernel.org \
--cc=amachhiw@linux.ibm.com \
--cc=kvm@vger.kernel.org \
--cc=sashiko-reviews@lists.linux.dev \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox