linux-kernel - [PATCH] iio: adc: stm32-adc: fix erroneous handling of spurious IRQs

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

Message-Id: <20210112152441.20665-1-a.fatoum@pengutronix.de>
Date:   Tue, 12 Jan 2021 16:24:42 +0100
From:   Ahmad Fatoum <a.fatoum@...gutronix.de>
To:     Jonathan Cameron <jic23@...nel.org>,
        Lars-Peter Clausen <lars@...afoo.de>,
        Peter Meerwald-Stadler <pmeerw@...erw.net>,
        Maxime Coquelin <mcoquelin.stm32@...il.com>,
        Alexandre Torgue <alexandre.torgue@...com>,
        Fabrice Gasnier <fabrice.gasnier@...com>,
        Olivier Moysan <olivier.moysan@...com>
Cc:     kernel@...gutronix.de, Thomas Gleixner <tglx@...utronix.de>,
        Lucas Stach <l.stach@...gutronix.de>,
        Holger Assmann <has@...gutronix.de>,
        Ahmad Fatoum <a.fatoum@...gutronix.de>,
        linux-iio@...r.kernel.org,
        linux-stm32@...md-mailman.stormreply.com,
        linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org
Subject: [PATCH] iio: adc: stm32-adc: fix erroneous handling of spurious IRQs

1c6c69525b40 ("genirq: Reject bogus threaded irq requests") makes sure
that threaded IRQs either
  - have IRQF_ONESHOT set
  - don't have the default just return IRQ_WAKE_THREAD primary handler

This is necessary because level-triggered interrupts need to be masked,
either at device or irqchip, to avoid an interrupt storm.

For spurious interrupts, the STM32 ADC driver still does this bogus
request though:
  - It doesn't set IRQF_ONESHOT
  - Its primary handler just returns IRQ_WAKE_THREAD if the interrupt
    is unexpected, i.e. !(status & enabled_mask)
  - stm32mp151.dtsi describes the ADC interrupt as level-triggered

Fix this by setting IRQF_ONESHOT to have the irqchip mask the IRQ
until the IRQ thread has finished.

IRQF_ONESHOT also has the effect that the primary handler is no longer
forced into a thread. This makes the issue with spurious interrupts
interrupts disappear when reading the ADC on a threadirqs=1 kernel.
This used to result in following kernel error message:

	iio iio:device1: Unexpected IRQ: IER=0x00000000, ISR=0x0000100e
or
	iio iio:device1: Unexpected IRQ: IER=0x00000004, ISR=0x0000100a

But with this patch applied (or threaded IRQs disabled), this no longer
occurs.

Cc: Lucas Stach <l.stach@...gutronix.de>
Reported-by: Holger Assmann <has@...gutronix.de>
Fixes: 695e2f5c289b ("iio: adc: stm32-adc: fix a regression when using dma and irq")
Signed-off-by: Ahmad Fatoum <a.fatoum@...gutronix.de>
---
 drivers/iio/adc/stm32-adc.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/iio/adc/stm32-adc.c b/drivers/iio/adc/stm32-adc.c
index c067c994dae2..7e0e21c79ac8 100644
--- a/drivers/iio/adc/stm32-adc.c
+++ b/drivers/iio/adc/stm32-adc.c
@@ -1910,7 +1910,7 @@ static int stm32_adc_probe(struct platform_device *pdev)
 
 	ret = devm_request_threaded_irq(&pdev->dev, adc->irq, stm32_adc_isr,
 					stm32_adc_threaded_isr,
-					0, pdev->name, indio_dev);
+					IRQF_ONESHOT, pdev->name, indio_dev);
 	if (ret) {
 		dev_err(&pdev->dev, "failed to request IRQ\n");
 		return ret;
-- 
2.30.0