From: Thomas Gleixner <tglx@kernel.org>
To: Mark Brown <broonie@kernel.org>, linux-kernel@vger.kernel.org
Cc: Farhad Alemi <farhad.alemi@berkeley.edu>,
Mark Brown <broonie@kernel.org>
Subject: Re: [PATCH] regmap: irq: Free the irqdomain we create
Date: Fri, 04 Sep 2026 21:05:26 +0200 [thread overview]
Message-ID: <87jyp07s7d.ffs@fw13> (raw)
In-Reply-To: <87mrtw7sf3.ffs@fw13>
On Fri, Sep 04 2026 at 21:00, Thomas Gleixner wrote:
> On Tue, Sep 01 2026 at 22:55, Mark Brown wrote:
>> err_domain:
>> - /* Should really dispose of the domain but... */
>> + irq_domain_remove(d->domain);
>
> That won't work.
>
> The case which is affected is the one which allocates interrupts
> upfront via alloc_irq_descs().
>
> In that case the domain creation will associate allocated interrupts
> because info.virq_base is > 0.
>
> This wont trigger the WARN_ON() in irq_domain_remove() because it's a
> fixed sized linear domain, but irq_domain_remove() will leak the
> interrupt descriptors which still have a reference (pointer) to the irq
> chip and the domain. So the same UAF is still there :)
>
> What you need to do before removing the domain is
>
> if (irq_base > 0)
> irq_domain_free_irqs(irq_base, chip->num_irqs);
Hit send too fast. That stupidly works only when hierarchical domains
are enabled.
So you need:
if (irq_base > 0) {
for (unsigned int i = 0; i < chip->num_irqs; i++)
irq_dispose_mapping(irq_base + i);
}
irqdomain_remove_domain();
Thanks,
tglx
next prev parent reply other threads:[~2026-09-04 19:05 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-09-01 21:55 [PATCH] regmap: irq: Free the irqdomain we create Mark Brown
2026-09-01 22:49 ` Mark Brown
2026-09-04 19:00 ` Thomas Gleixner
2026-09-04 19:05 ` Thomas Gleixner [this message]
2026-09-04 19:21 ` Mark Brown
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=87jyp07s7d.ffs@fw13 \
--to=tglx@kernel.org \
--cc=broonie@kernel.org \
--cc=farhad.alemi@berkeley.edu \
--cc=linux-kernel@vger.kernel.org \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.