dmaengine Archive on lore.kernel.org
 help / color / Atom feed
From: Philipp Puschmann <philipp.puschmann@emlix.com>
To: linux-kernel@vger.kernel.org
Cc: jlu@pengutronix.de, yibin.gong@nxp.com, fugang.duan@nxp.com,
	l.stach@pengutronix.de, dan.j.williams@intel.com,
	vkoul@kernel.org, shawnguo@kernel.org, s.hauer@pengutronix.de,
	kernel@pengutronix.de, festevam@gmail.com, linux-imx@nxp.com,
	dmaengine@vger.kernel.org, linux-arm-kernel@lists.infradead.org,
	Philipp Puschmann <philipp.puschmann@emlix.com>
Subject: [PATCH v5 2/3] dmaengine: imx-sdma: fix dma freezes
Date: Mon, 23 Sep 2019 15:58:07 +0200
Message-ID: <20190923135808.815-3-philipp.puschmann@emlix.com> (raw)
In-Reply-To: <20190923135808.815-1-philipp.puschmann@emlix.com>

For some years and since many kernel versions there are reports that the
RX UART SDMA channel stops working at some point. The workaround was to
disable DMA for RX. This commit fixes the problem itself. Cyclic DMA
transfers are used by uart and other drivers and these can fail in at
least two cases where we can run out of descriptors available to the
engine:
- Interrupts are disabled for too long and all buffers are filled with
  data, especially in a setup where many small dma transfers are being
  executed only using a tiny part of a single buffer
- DMA errors (such as generated by baud rate mismatch with imx-uart)
  use up all descriptors before we can react.

In this case, SDMA stops the channel and no further transfers are done
until the respective channel is disabled and re-enabled. We can check
if the channel has been stopped and re-enable it then. To distinguish
from the the case that the channel was stopped by upper-level driver
we introduce new flag IMX_DMA_ACTIVE.

As sdmac->desc is constant we can move desc out of the loop.

Fixes: 1ec1e82f2510 ("dmaengine: Add Freescale i.MX SDMA support")
Signed-off-by: Philipp Puschmann <philipp.puschmann@emlix.com>
Signed-off-by: Jan Luebbe <jlu@pengutronix.de>
Reviewed-by: Lucas Stach <l.stach@pengutronix.de>
---

Changelog v5:
 - join with patch version from Jan Luebbe
 - adapt comments and patch descriptions
 
Changelog v4:
 - fixed the fixes tag
 
Changelog v3:
 - use correct dma_wmb() instead of dma_wb()
 - add fixes tag
 
Changelog v2:
 - clarify comment and commit description

 drivers/dma/imx-sdma.c | 26 ++++++++++++++++++++++----
 1 file changed, 22 insertions(+), 4 deletions(-)

diff --git a/drivers/dma/imx-sdma.c b/drivers/dma/imx-sdma.c
index b42281604e54..0b1d6a62423d 100644
--- a/drivers/dma/imx-sdma.c
+++ b/drivers/dma/imx-sdma.c
@@ -383,6 +383,7 @@ struct sdma_channel {
 };
 
 #define IMX_DMA_SG_LOOP		BIT(0)
+#define IMX_DMA_ACTIVE		BIT(1)
 
 #define MAX_DMA_CHANNELS 32
 #define MXC_SDMA_DEFAULT_PRIORITY 1
@@ -658,6 +659,9 @@ static int sdma_config_ownership(struct sdma_channel *sdmac,
 
 static void sdma_enable_channel(struct sdma_engine *sdma, int channel)
 {
+	struct sdma_channel *sdmac = &sdma->channel[channel];
+
+	sdmac->flags |= IMX_DMA_ACTIVE;
 	writel(BIT(channel), sdma->regs + SDMA_H_START);
 }
 
@@ -774,16 +778,17 @@ static void sdma_start_desc(struct sdma_channel *sdmac)
 
 static void sdma_update_channel_loop(struct sdma_channel *sdmac)
 {
+	struct sdma_engine *sdma = sdmac->sdma;
 	struct sdma_buffer_descriptor *bd;
+	struct sdma_desc *desc = sdmac->desc;
 	int error = 0;
-	enum dma_status	old_status = sdmac->status;
+	enum dma_status old_status = sdmac->status;
 
 	/*
 	 * loop mode. Iterate over descriptors, re-setup them and
 	 * call callback function.
 	 */
-	while (sdmac->desc) {
-		struct sdma_desc *desc = sdmac->desc;
+	while (desc) {
 
 		bd = &desc->bd[desc->buf_tail];
 
@@ -822,6 +827,18 @@ static void sdma_update_channel_loop(struct sdma_channel *sdmac)
 		if (error)
 			sdmac->status = old_status;
 	}
+
+	/* In some situations it may happen that the sdma does not find any
+	 * usable descriptor in the ring to put data into. The channel is
+	 * stopped then and after having freed some buffers we have to restart
+	 * it manually.
+	 */
+	if ((sdmac->flags & IMX_DMA_ACTIVE) &&
+	    !(readl_relaxed(sdma->regs + SDMA_H_STATSTOP) & BIT(sdmac->channel))) {
+		dev_err_ratelimited(sdma->dev, "SDMA channel %d: cyclic transfer disabled by HW, reenabling\n",
+				    sdmac->channel);
+			writel(BIT(sdmac->channel), sdma->regs + SDMA_H_START);
+	};
 }
 
 static void mxc_sdma_handle_channel_normal(struct sdma_channel *data)
@@ -1051,7 +1068,8 @@ static int sdma_disable_channel(struct dma_chan *chan)
 	struct sdma_engine *sdma = sdmac->sdma;
 	int channel = sdmac->channel;
 
-	writel_relaxed(BIT(channel), sdma->regs + SDMA_H_STATSTOP);
+	sdmac->flags &= ~IMX_DMA_ACTIVE;
+	writel(BIT(channel), sdma->regs + SDMA_H_STATSTOP);
 	sdmac->status = DMA_ERROR;
 
 	return 0;
-- 
2.23.0


  parent reply index

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-09-23 13:58 [PATCH v5 0/3] Fix UART DMA freezes for i.MX SOCs Philipp Puschmann
2019-09-23 13:58 ` [PATCH v5 1/3] dmaengine: imx-sdma: fix buffer ownership Philipp Puschmann
2019-12-03  9:47   ` Lucas Stach
2019-12-04  9:19   ` Robin Gong
2019-12-10  9:44     ` Philipp Puschmann
2019-12-10 13:01       ` Robin Gong
2019-09-23 13:58 ` Philipp Puschmann [this message]
2019-09-24  3:14   ` [PATCH v5 2/3] dmaengine: imx-sdma: fix dma freezes Robin Gong
2019-09-23 13:58 ` [PATCH v5 3/3] dmaengine: imx-sdma: drop redundant variable Philipp Puschmann
2019-09-23 14:55 ` [PATCH v5 0/3] Fix UART DMA freezes for i.MX SOCs Adam Ford
2019-09-23 15:06   ` Philipp Puschmann
2019-09-23 15:09     ` Jan Lübbe

Reply instructions:

You may reply publically 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=20190923135808.815-3-philipp.puschmann@emlix.com \
    --to=philipp.puschmann@emlix.com \
    --cc=dan.j.williams@intel.com \
    --cc=dmaengine@vger.kernel.org \
    --cc=festevam@gmail.com \
    --cc=fugang.duan@nxp.com \
    --cc=jlu@pengutronix.de \
    --cc=kernel@pengutronix.de \
    --cc=l.stach@pengutronix.de \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-imx@nxp.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=s.hauer@pengutronix.de \
    --cc=shawnguo@kernel.org \
    --cc=vkoul@kernel.org \
    --cc=yibin.gong@nxp.com \
    /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

dmaengine Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/dmaengine/0 dmaengine/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 dmaengine dmaengine/ https://lore.kernel.org/dmaengine \
		dmaengine@vger.kernel.org
	public-inbox-index dmaengine

Example config snippet for mirrors

Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.kernel.vger.dmaengine


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git