Kernel KVM virtualization development
 help / color / mirror / Atom feed
From: sashiko-bot@kernel.org
To: "Bibo Mao" <maobibo@loongson.cn>
Cc: kvm@vger.kernel.org
Subject: Re: [PATCH v6 4/6] LoongArch: KVM: Inject interrupt with batch method
Date: Thu, 11 Jun 2026 07:17:21 +0000	[thread overview]
Message-ID: <20260611071722.41BF31F00893@smtp.kernel.org> (raw)
In-Reply-To: <20260611065815.1439154-5-maobibo@loongson.cn>

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:

Pre-existing issues:
- [High] Uninitialized stack array `vector` is accessed conditionally in `dmsintc_inject_irq`, potentially leading to hardware state corruption.
--

commit 1bbf844f12dc29aa2ddec056c816e1245e63e5fb
Author: Bibo Mao <maobibo@loongson.cn>

    LoongArch: KVM: Inject interrupt with batch method

This commit converts interrupt injection and clearing to process multiple
interrupts simultaneously using a bitmask instead of one-by-one based on
priorities. It also removes the unused priority_to_irq array.

> diff --git a/arch/loongarch/kvm/interrupt.c b/arch/loongarch/kvm/interrupt.c
> index 48dd56aa4dc56..c1afeb058c5b3 100644
> --- a/arch/loongarch/kvm/interrupt.c
> +++ b/arch/loongarch/kvm/interrupt.c
[ ... ]
> -static int kvm_irq_deliver(struct kvm_vcpu *vcpu, unsigned int priority)
> +static void kvm_irq_deliver(struct kvm_vcpu *vcpu, unsigned long mask)
>  {
> -	unsigned int irq = 0;
> +	unsigned long irq;
>  	unsigned long old, new;
>  
> -	clear_bit(priority, &vcpu->arch.irq_pending);
> -	if (priority < EXCCODE_INT_NUM)
> -		irq = priority_to_irq[priority];
> -
> -	switch (priority) {
> -	case INT_AVEC:
> -		dmsintc_inject_irq(vcpu);
> -		fallthrough;
> -	case INT_TI:
> -	case INT_IPI:
> -	case INT_SWI0:
> -	case INT_SWI1:
> +	irq = mask & KVM_ESTAT_INTI_MASK;
> +	if (irq) {
> +		if (irq & CPU_AVEC)
> +			dmsintc_inject_irq(vcpu);

[Severity: High]
This is a pre-existing issue, but does dmsintc_inject_irq() use an
uninitialized stack array?

Looking at dmsintc_inject_irq() in arch/loongarch/kvm/intc/dmsintc.c, the
array vector is declared without initialization:

arch/loongarch/kvm/intc/dmsintc.c:dmsintc_inject_irq() {
	unsigned int i;
	unsigned long vector[4], old;
	...
	for (i = 0; i < 4; i++) {
		old = atomic64_read(&(ds->vector_map[i]));
		if (old)
			vector[i] = atomic64_xchg(&(ds->vector_map[i]), 0);
	}

If old is zero, vector[i] is left uninitialized. Then later the
uninitialized memory is evaluated:

	if (vector[0]) {
		old = kvm_read_hw_gcsr(LOONGARCH_CSR_ISR0);
		kvm_write_hw_gcsr(LOONGARCH_CSR_ISR0, vector[0] | old);
	}
	...

Can this unpredictably corrupt the guest's interrupt state machine and
potentially cause guest kernel panics by writing garbage values to the
hardware CSR?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260611065815.1439154-1-maobibo@loongson.cn?part=4

  reply	other threads:[~2026-06-11  7:17 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-06-11  6:58 [PATCH v6 0/6] LoongArch: KVM: Small enhancement about interrupt injection Bibo Mao
2026-06-11  6:58 ` [PATCH v6 1/6] LoongArch: KVM: Check irq validility in kvm_vcpu_ioctl_interrupt() Bibo Mao
2026-06-11  7:08   ` sashiko-bot
2026-06-11  6:58 ` [PATCH v6 2/6] LoongArch: KVM: Check msgint feature in interrupt post Bibo Mao
2026-06-11  7:13   ` sashiko-bot
2026-06-11  6:58 ` [PATCH v6 3/6] LoongArch: KVM: Use existing macro about interrupt bit mask Bibo Mao
2026-06-11  6:58 ` [PATCH v6 4/6] LoongArch: KVM: Inject interrupt with batch method Bibo Mao
2026-06-11  7:17   ` sashiko-bot [this message]
2026-06-11  6:58 ` [PATCH v6 5/6] LoongArch: KVM: Add valid bit check when set ESTAT CSR register Bibo Mao
2026-06-11 10:10   ` Huacai Chen
2026-06-11 12:52     ` Bibo Mao
2026-06-11 13:01       ` Huacai Chen
2026-06-12  8:39         ` Huacai Chen
2026-06-11  6:58 ` [PATCH v6 6/6] LoongArch: KVM: Deliver interrupt after IN_GUEST_MODE is set Bibo Mao

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=20260611071722.41BF31F00893@smtp.kernel.org \
    --to=sashiko-bot@kernel.org \
    --cc=kvm@vger.kernel.org \
    --cc=maobibo@loongson.cn \
    --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