linux-kernel - [PATCH 15/18] mfd: stpmic1: Fix swapped mask/unmask in irq chip

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

Message-Id: <20221112151835.39059-16-aidanmacdonald.0x0@gmail.com>
Date:   Sat, 12 Nov 2022 15:18:32 +0000
From:   Aidan MacDonald <aidanmacdonald.0x0@...il.com>
To:     lee@...nel.org
Cc:     mani@...nel.org, cristian.ciocaltea@...il.com, wens@...e.org,
        tharvey@...eworks.com, cw00.choi@...sung.com,
        krzysztof.kozlowski@...aro.org, brgl@...ev.pl,
        mazziesaccount@...il.com, orsonzhai@...il.com,
        baolin.wang@...ux.alibaba.com, zhang.lyra@...il.com,
        jernej.skrabec@...il.com, samuel@...lland.org,
        linux-kernel@...r.kernel.org, linux-actions@...ts.infradead.org,
        linux-arm-kernel@...ts.infradead.org, linux-sunxi@...ts.linux.dev
Subject: [PATCH 15/18] mfd: stpmic1: Fix swapped mask/unmask in irq chip

The usual behavior of mask registers is writing a '1' bit to
disable (mask) an interrupt; similarly, writing a '1' bit to
an unmask register enables (unmasks) an interrupt.

Due to a longstanding issue in regmap-irq, mask and unmask
registers were inverted when both kinds of registers were
present on the same chip, ie. regmap-irq actually wrote '1's
to the mask register to enable an IRQ and '1's to the unmask
register to disable an IRQ.

This was fixed by commit e8ffb12e7f06 ("regmap-irq: Fix
inverted handling of unmask registers") but the fix is opt-in
via mask_unmask_non_inverted = true because it requires manual
changes for each affected driver. The new behavior will become
the default once all drivers have been updated.

The STPMIC1 has a normal mask register with separate set and
clear registers. The driver intends to use the set & clear
registers with regmap-irq and has compensated for regmap-irq's
inverted behavior, and should currently be working properly.
Thus, swap mask_base and unmask_base, and opt in to the new
non-inverted behavior.

Signed-off-by: Aidan MacDonald <aidanmacdonald.0x0@...il.com>
---
 drivers/mfd/stpmic1.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/drivers/mfd/stpmic1.c b/drivers/mfd/stpmic1.c
index eb3da558c3fb..ee0469d5d435 100644
--- a/drivers/mfd/stpmic1.c
+++ b/drivers/mfd/stpmic1.c
@@ -108,8 +108,9 @@ static const struct regmap_irq stpmic1_irqs[] = {
 static const struct regmap_irq_chip stpmic1_regmap_irq_chip = {
 	.name = "pmic_irq",
 	.status_base = INT_PENDING_R1,
-	.mask_base = INT_CLEAR_MASK_R1,
-	.unmask_base = INT_SET_MASK_R1,
+	.mask_base = INT_SET_MASK_R1,
+	.unmask_base = INT_CLEAR_MASK_R1,
+	.mask_unmask_non_inverted = true,
 	.ack_base = INT_CLEAR_R1,
 	.num_regs = STPMIC1_PMIC_NUM_IRQ_REGS,
 	.irqs = stpmic1_irqs,
-- 
2.38.1