linux-kernel - Re: PCI: vmd: repeated kfree of vmdirq

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <20250717115220.jfvzVfuR@linutronix.de>
Date: Thu, 17 Jul 2025 13:52:20 +0200
From: Nam Cao <namcao@...utronix.de>
To: "Colin King (gmail)" <colin.i.king@...il.com>
Cc: Manivannan Sadhasivam <mani@...nel.org>,
	Thomas Gleixner <tglx@...utronix.de>,
	Bjorn Helgaas <bhelgaas@...gle.com>,
	Lorenzo Pieralisi <lpieralisi@...nel.org>,
	Krzysztof Wilczyński <kwilczynski@...nel.org>,
	linux-pci@...r.kernel.org,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: PCI: vmd: repeated kfree of vmdirq

On Thu, Jul 17, 2025 at 12:48:02PM +0100, Colin King (gmail) wrote:
> Hi,
> 
> Static analysis found an issue in the following commit in linux-next:
> 
> commit 2b96beffa42760513567919aa27eb72035f2db58
> Author: Nam Cao <namcao@...utronix.de>
> Date:   Thu Jun 26 16:48:06 2025 +0200
> 
>     PCI: vmd: Switch to msi_create_parent_irq_domain()
> 
> 
> The issue is as follows:
> 
> static void vmd_msi_free(struct irq_domain *domain, unsigned int virq,
> unsigned int nr_irqs)
> {
>         struct vmd_irq *vmdirq = irq_get_chip_data(virq);
> 
>         for (int i = 0; i < nr_irqs; ++i) {
>                 synchronize_srcu(&vmdirq->irq->srcu);
> 
>                 /* XXX: Potential optimization to rebalance */
>                 scoped_guard(raw_spinlock_irq, &list_lock)
>                         vmdirq->irq->count--;
> 
>                 kfree(vmdirq);
>         }
> }
> 
> The for-loop is repeatedly kfree'ing vmdirq which seems incorrect.

Thanks for the report. It has been resolved in PCI tree.

Best regards,
Nam