public inbox for dmaengine@vger.kernel.org
 help / color / mirror / Atom feed
From: "Tomasz Moń" <tomasz.mon@camlingroup.com>
To: dmaengine@vger.kernel.org
Cc: "Vinod Koul" <vkoul@kernel.org>,
	"Shawn Guo" <shawnguo@kernel.org>,
	"Sascha Hauer" <s.hauer@pengutronix.de>,
	"Pengutronix Kernel Team" <kernel@pengutronix.de>,
	"Fabio Estevam" <festevam@gmail.com>,
	"NXP Linux Team" <linux-imx@nxp.com>,
	k.drobinski@camlintechnologies.com,
	"Tomasz Moń" <tomasz.mon@camlingroup.com>
Subject: [PATCH 1/2] dmaengine: imx-sdma: restart cyclic channel if needed
Date: Mon, 17 Jan 2022 10:19:54 +0100	[thread overview]
Message-ID: <20220117091955.1038937-1-tomasz.mon@camlingroup.com> (raw)

Under heavy load resulting in high interrupt latencies, it is possible
for imx UART requests to completely fill DMA buffer. When DMA channel
is triggered and no SDMA owned buffer is available, SDMA stops. Thanks
to the autoRTS feature, there is no data loss due to the SDMA stop if
the UART is using hardware flow control.

According to DMA Engine API Guide, DMA cyclic operation is performed
until explicitly stopped. Restart the buffer after handling channel loop
if the channel was stopped by SDMA.

Signed-off-by: Tomasz Moń <tomasz.mon@camlingroup.com>
---
 drivers/dma/imx-sdma.c | 14 ++++++++++++++
 1 file changed, 14 insertions(+)

diff --git a/drivers/dma/imx-sdma.c b/drivers/dma/imx-sdma.c
index 75ec0754d4ad..330ff41cd614 100644
--- a/drivers/dma/imx-sdma.c
+++ b/drivers/dma/imx-sdma.c
@@ -701,6 +701,11 @@ static int sdma_config_ownership(struct sdma_channel *sdmac,
 	return 0;
 }
 
+static int is_sdma_channel_enabled(struct sdma_engine *sdma, int channel)
+{
+	return !!(readl(sdma->regs + SDMA_H_STATSTOP) & BIT(channel));
+}
+
 static void sdma_enable_channel(struct sdma_engine *sdma, int channel)
 {
 	writel(BIT(channel), sdma->regs + SDMA_H_START);
@@ -860,6 +865,15 @@ static void sdma_update_channel_loop(struct sdma_channel *sdmac)
 		if (error)
 			sdmac->status = old_status;
 	}
+
+	/*
+	 * SDMA stops cyclic channel when DMA request triggers a channel and no SDMA
+	 * owned buffer is available (i.e. BD_DONE was set too late).
+	 */
+	if (!is_sdma_channel_enabled(sdmac->sdma, sdmac->channel)) {
+		dev_warn(sdmac->sdma->dev, "restart cyclic channel %d\n", sdmac->channel);
+		sdma_enable_channel(sdmac->sdma, sdmac->channel);
+	}
 }
 
 static void mxc_sdma_handle_channel_normal(struct sdma_channel *data)
-- 
2.25.1


             reply	other threads:[~2022-01-17  9:27 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-01-17  9:19 Tomasz Moń [this message]
2022-01-17  9:19 ` [PATCH 2/2] dmaengine: imx-sdma: fix cyclic buffer race condition Tomasz Moń
2022-02-15  5:41 ` [PATCH 1/2] dmaengine: imx-sdma: restart cyclic channel if needed Vinod Koul

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=20220117091955.1038937-1-tomasz.mon@camlingroup.com \
    --to=tomasz.mon@camlingroup.com \
    --cc=dmaengine@vger.kernel.org \
    --cc=festevam@gmail.com \
    --cc=k.drobinski@camlintechnologies.com \
    --cc=kernel@pengutronix.de \
    --cc=linux-imx@nxp.com \
    --cc=s.hauer@pengutronix.de \
    --cc=shawnguo@kernel.org \
    --cc=vkoul@kernel.org \
    /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