dmaengine: imx-sdma: ack channel 0 IRQ in the interrupt handler

Commit Message

Lucas Stach July 7, 2016, 9:35 a.m. UTC

From: Michael Olbrich <[e-mail]>

Currently the handler ignores the channel 0 interrupt and thus doesn't ack
it properly. This is done in order to allow sdma_run_channel0() to poll
on the irq status bit, as this function may be called in atomic context,
but needs to know when the channel has finished.

This works mostly, as the polling happens under a spinlock, disabling IRQs
on the local CPU, leaving only a very slight race window for a spurious
IRQ to happen if the handler is executed on another CPU in an SMP system.
Still this is clearly suboptimal.
This behavior turns into a real problem on an RT system, where the spinlock
doesn't disable IRQs on the local CPU. Not acking the IRQ in the handler
in such a setup is very likely to drown the CPU in an IRQ storm, leaving
it unable to make any progress in the polling loop, leading to the IRQ
never being acked.

Fix this by properly acknowledging the channel 0 IRQ in the handler.
As the IRQ status bit can no longer be used to poll for the channel
completion, switch over to using the SDMA_H_STATSTOP register for this
purpose, where bit 0 is cleared by the hardware when the channel is done.

Signed-off-by: Michael Olbrich <[e-mail]>
Signed-off-by: Lucas Stach <[e-mail]>
---
v2 lst:
- use iopoll helper
- rephrase commit message
---
 drivers/dma/imx-sdma.c | 23 ++++++++---------------
 1 file changed, 8 insertions(+), 15 deletions(-)

Comments

Vinod Koul July 12, 2016, 4:48 a.m. UTC | #1

On Thu, Jul 07, 2016 at 11:35:51AM +0200, Lucas Stach wrote:
> From: Michael Olbrich <[e-mail]>
> 
> Currently the handler ignores the channel 0 interrupt and thus doesn't ack
> it properly. This is done in order to allow sdma_run_channel0() to poll
> on the irq status bit, as this function may be called in atomic context,
> but needs to know when the channel has finished.
> 
> This works mostly, as the polling happens under a spinlock, disabling IRQs
> on the local CPU, leaving only a very slight race window for a spurious
> IRQ to happen if the handler is executed on another CPU in an SMP system.
> Still this is clearly suboptimal.
> This behavior turns into a real problem on an RT system, where the spinlock
> doesn't disable IRQs on the local CPU. Not acking the IRQ in the handler
> in such a setup is very likely to drown the CPU in an IRQ storm, leaving
> it unable to make any progress in the polling loop, leading to the IRQ
> never being acked.
> 
> Fix this by properly acknowledging the channel 0 IRQ in the handler.
> As the IRQ status bit can no longer be used to poll for the channel
> completion, switch over to using the SDMA_H_STATSTOP register for this
> purpose, where bit 0 is cleared by the hardware when the channel is done.

Applied, thanks

Patch

diff --git a/drivers/dma/imx-sdma.c b/drivers/dma/imx-sdma.c
index 0f6fd42f55ca..ce865f68a8c7 100644
--- a/drivers/dma/imx-sdma.c
+++ b/drivers/dma/imx-sdma.c
@@ -18,6 +18,7 @@ 
  */
 
 #include <linux/init.h>
+#include <linux/iopoll.h>
 #include <linux/module.h>
 #include <linux/types.h>
 #include <linux/bitops.h>
@@ -571,28 +572,20 @@  static void sdma_enable_channel(struct sdma_engine *sdma, int channel)
 static int sdma_run_channel0(struct sdma_engine *sdma)
 {
 	int ret;
-	unsigned long timeout = 500;
+	u32 reg;
 
 	sdma_enable_channel(sdma, 0);
 
-	while (!(ret = readl_relaxed(sdma->regs + SDMA_H_INTR) & 1)) {
-		if (timeout-- <= 0)
-			break;
-		udelay(1);
-	}
-
-	if (ret) {
-		/* Clear the interrupt status */
-		writel_relaxed(ret, sdma->regs + SDMA_H_INTR);
-	} else {
+	ret = readl_relaxed_poll_timeout_atomic(sdma->regs + SDMA_H_STATSTOP,
+						reg, !(reg & 1), 1, 500);
+	if (ret)
 		dev_err(sdma->dev, "Timeout waiting for CH0 ready\n");
-	}
 
 	/* Set bits of CONFIG register with dynamic context switching */
 	if (readl(sdma->regs + SDMA_H_CONFIG) == 0)
 		writel_relaxed(SDMA_H_CONFIG_CSM, sdma->regs + SDMA_H_CONFIG);
 
-	return ret ? 0 : -ETIMEDOUT;
+	return ret;
 }
 
 static int sdma_load_script(struct sdma_engine *sdma, void *buf, int size,
@@ -727,9 +720,9 @@  static irqreturn_t sdma_int_handler(int irq, void *dev_id)
 	unsigned long stat;
 
 	stat = readl_relaxed(sdma->regs + SDMA_H_INTR);
-	/* not interested in channel 0 interrupts */
-	stat &= ~1;
 	writel_relaxed(stat, sdma->regs + SDMA_H_INTR);
+	/* channel 0 is special and not handled here, see run_channel0() */
+	stat &= ~1;
 
 	while (stat) {
 		int channel = fls(stat) - 1;