linux-kernel - [RFC PATCH] dmaengine: xilinx_dma: device-wide directions cause ASoC cyclic DMA regression

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 for Android: free password hash cracker in your pocket

[<prev] [next>] [day] [month] [year] [list]

Message-ID: <20260211140051.8177-1-rahulnavale04@gmail.com>
Date: Wed, 11 Feb 2026 19:30:51 +0530
From: Rahul Navale <rahulnavale04@...il.com>
To: dmaengine@...r.kernel.org
Cc: vkoul@...nel.org,
	michal.simek@....com,
	linux-kernel@...r.kernel.org,
	linux-arm-kernel@...ts.infradead.org,
	thomas.gessler@...eckmann-gmbh.de,
	radhey.shyam.pandey@....com,
	Suraj.Gupta2@....com,
	marex@...x.de,
	manivannan.sadhasivam@...aro.org,
	harini.katakam@....com,
	marex@...ladev.com,
	Rahul Navale <rahul.navale@....com>
Subject: [RFC PATCH] dmaengine: xilinx_dma: device-wide directions cause ASoC cyclic DMA regression

From: Rahul Navale <rahul.navale@....com>

On ZynqMP platforms using AXI DMA for ASoC PCM playback, upstream commit
7e01511443c3 ("dmaengine: xilinx_dma: Set dma_device directions") causes
cyclic playback to fail after the first buffer period.

Background:
The upstream patch adds the following line in xilinx_dma_chan_probe():

    xdev->common.directions |= chan->direction;

Its purpose is to coalesce the directions of all enabled TX/RX channels into
the device-wide dma_device.directions mask so that dma_get_slave_caps()
works correctly. This is required by users such as IIO DMAEngine buffers
that rely on device-wide capability reporting.

Problem on ZynqMP ASoC audio (PCM):
On ZynqMP, Xilinx DMA provides fixed-direction channels:

    MM2S channels -> DMA_MEM_TO_DEV
    S2MM channels -> DMA_DEV_TO_MEM

ASoC dmaengine PCM relies on these fixed directions to select proper DMA
channels for cyclic playback and capture. Aggregating directions device-wide
can cause inconsistent capability reporting depending on channel probe order
or device tree layout.

This leads to the following behavior:
- The first DMA buffer plays correctly.
- Subsequent DMA periods repeat the first period indefinitely, breaking
  cyclic playback.

A temporary local patch that removes:

    xdev->common.directions |= chan->direction;

restores audio playback on ZynqMP platforms.

This RFC patch is a workaround to demonstrate a ZynqMP ASoC cyclic PCM
regression that appears in v6.12.36+ (backport of 7e01511443c3).

The workaround restores PCM playback but may regress users relying on
dma_get_slave_caps() (e.g. IIO DMAEngine buffers). The goal is to start
discussion on the correct upstream fix that supports both ASoC PCM and IIO.

Fixes: 7e01511443c3 ("dmaengine: xilinx_dma: Set dma_device directions")
Signed-off-by: Rahul Navale <rahul.navale@....com>
---
 drivers/dma/xilinx/xilinx_dma.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/drivers/dma/xilinx/xilinx_dma.c b/drivers/dma/xilinx/xilinx_dma.c
index aff046b03ef7..8da86e322c7a 100644
--- a/drivers/dma/xilinx/xilinx_dma.c
+++ b/drivers/dma/xilinx/xilinx_dma.c
@@ -2938,8 +2938,6 @@ static int xilinx_dma_chan_probe(struct xilinx_dma_device *xdev,
 		return -EINVAL;
 	}

-	xdev->common.directions |= chan->direction;
-
 	/* Request the interrupt */
 	chan->irq = of_irq_get(node, chan->tdest);
 	if (chan->irq < 0)
-- 
2.34.1