Linux GPIO subsystem development
 help / color / mirror / Atom feed
From: Stepan Ionichev <sozdayvek@gmail.com>
To: andy@kernel.org
Cc: mika.westerberg@linux.intel.com, linusw@kernel.org,
	linux-gpio@vger.kernel.org, linux-kernel@vger.kernel.org,
	sozdayvek@gmail.com
Subject: [PATCH] pinctrl: intel: zero-initialize capability_offset[] in probe
Date: Fri, 15 May 2026 20:00:49 +0500	[thread overview]
Message-ID: <20260515150049.33761-1-sozdayvek@gmail.com> (raw)

intel_pinctrl_probe() declares a per-community capability_offset[]
array on the stack and only writes the slots whose CAPLIST entries
the device actually advertises:

	unsigned short capability_offset[6];
	...
	do {
		value = readl(regs + offset);
		switch ((value & CAPLIST_ID_MASK) >> CAPLIST_ID_SHIFT) {
		case CAPLIST_ID_GPIO_HW_INFO:
			capability_offset[CAPLIST_ID_GPIO_HW_INFO] = offset;
			break;
		case CAPLIST_ID_PWM:
			capability_offset[CAPLIST_ID_PWM] = offset;
			break;
		...
		default:
			break;
		}
		...
	} while (offset);

	...
	ret = intel_pinctrl_probe_pwm(pctrl, community,
				      capability_offset[CAPLIST_ID_PWM]);

If a community does not advertise a PWM capability, the loop never
writes capability_offset[CAPLIST_ID_PWM] and the call to
intel_pinctrl_probe_pwm() reads an indeterminate stack value.

intel_pinctrl_probe_pwm() computes the PWM register base before it
checks PINCTRL_FEATURE_PWM:

	void __iomem *base = community->regs + capability_offset + 4;
	...
	if (!(community->features & PINCTRL_FEATURE_PWM))
		return 0;

so the base = community->regs + capability_offset + 4 expression is
evaluated using the uninitialized value on every probe for any
community that does not have CAPLIST_ID_PWM. The result is never
dereferenced when the PWM feature flag is clear, but reading an
indeterminate stack object is undefined behaviour by C, and KMSAN /
UBSAN report it at probe time.

Default the whole array to zero at declaration so unused slots are
defined (community->regs + 4) instead of indeterminate.

Fixes: 340bba73c545 ("pinctrl: intel: Improve capability support")
Signed-off-by: Stepan Ionichev <sozdayvek@gmail.com>
---
 drivers/pinctrl/intel/pinctrl-intel.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/pinctrl/intel/pinctrl-intel.c b/drivers/pinctrl/intel/pinctrl-intel.c
index 97bf5ec78..4fb4a3dab 100644
--- a/drivers/pinctrl/intel/pinctrl-intel.c
+++ b/drivers/pinctrl/intel/pinctrl-intel.c
@@ -1601,7 +1601,7 @@ int intel_pinctrl_probe(struct platform_device *pdev,
 
 	for (i = 0; i < pctrl->ncommunities; i++) {
 		struct intel_community *community = &pctrl->communities[i];
-		unsigned short capability_offset[6];
+		unsigned short capability_offset[6] = { };
 		void __iomem *regs;
 		u32 revision;
 		u32 offset;
-- 
2.43.0


                 reply	other threads:[~2026-05-15 15:01 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

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=20260515150049.33761-1-sozdayvek@gmail.com \
    --to=sozdayvek@gmail.com \
    --cc=andy@kernel.org \
    --cc=linusw@kernel.org \
    --cc=linux-gpio@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mika.westerberg@linux.intel.com \
    /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