linux-kernel - [PATCH] spi: davinci: Unset POWERDOWN bit when releasing resources

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: <20240624071745.17409-1-bastien.curutchet@bootlin.com>
Date: Mon, 24 Jun 2024 09:17:45 +0200
From: Bastien Curutchet <bastien.curutchet@...tlin.com>
To: Mark Brown <broonie@...nel.org>
Cc: linux-spi@...r.kernel.org,
	linux-kernel@...r.kernel.org,
	Thomas Petazzoni <thomas.petazzoni@...tlin.com>,
	Herve Codina <herve.codina@...tlin.com>,
	Christopher Cordahi <christophercordahi@...ometrics.ca>,
	Bastien Curutchet <bastien.curutchet@...tlin.com>
Subject: [PATCH] spi: davinci: Unset POWERDOWN bit when releasing resources

On the OMAPL138, the SPI reference clock is provided by the Power and
Sleep Controller (PSC). The PSC's datasheet says that 'some peripherals
have special programming requirements and additional recommended steps
you must take before you can invoke the PSC module state transition'. I
didn't find more details in documentation but it appears that PSC needs
the SPI to clear the POWERDOWN bit before disabling the clock. Indeed,
when this bit is set, the PSC gets stuck in transitions from enable to
disable state.

Clear the POWERDOWN bit when releasing driver's resources

Signed-off-by: Bastien Curutchet <bastien.curutchet@...tlin.com>
---
Hi,

I ran into this bug by enabling the 'cs-gpio' property. It causes the
probe to fail at first with -EPROBE_DEFER because the gpio provider is
not ready. So the clock gets disabled. In the clock controller's driver
(drivers/clk/davinci/psc.c) the clock_disable() calls a
regmap_read_poll_timeout() with an infinite timeout. This poll() polls
a transition bit status that never goes down so we end stuck in the
middle of the boot sequence.

 drivers/spi/spi-davinci.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/drivers/spi/spi-davinci.c b/drivers/spi/spi-davinci.c
index be3998104bfb..f7e8b5efa50e 100644
--- a/drivers/spi/spi-davinci.c
+++ b/drivers/spi/spi-davinci.c
@@ -984,6 +984,9 @@ static int davinci_spi_probe(struct platform_device *pdev)
 	return ret;

 free_dma:
+	/* This bit needs to be cleared to disable dpsi->clk */
+	clear_io_bits(dspi->base + SPIGCR1, SPIGCR1_POWERDOWN_MASK);
+
 	if (dspi->dma_rx) {
 		dma_release_channel(dspi->dma_rx);
 		dma_release_channel(dspi->dma_tx);
@@ -1013,6 +1016,9 @@ static void davinci_spi_remove(struct platform_device *pdev)

 	spi_bitbang_stop(&dspi->bitbang);

+	/* This bit needs to be cleared to disable dpsi->clk */
+	clear_io_bits(dspi->base + SPIGCR1, SPIGCR1_POWERDOWN_MASK);
+
 	if (dspi->dma_rx) {
 		dma_release_channel(dspi->dma_rx);
 		dma_release_channel(dspi->dma_tx);
-- 
2.45.0