Re: [PATCH v3 8/8] DMA: Freescale: add suspend resume functions for DMA driver

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 




On 04/10/2014 08:05 PM, Andy Shevchenko wrote:
On Thu, 2014-04-10 at 15:10 +0800, hongbo.zhang@xxxxxxxxxxxxx wrote:
From: Hongbo Zhang <hongbo.zhang@xxxxxxxxxxxxx>

This patch adds suspend resume functions for Freescale DMA driver.
.prepare callback is used to stop further descriptors from being added into the
pending queue, and also issue pending queues into execution if there is any.
.suspend callback makes sure all the pending jobs are cleaned up and all the
channels are idle, and save the mode registers.
.resume callback re-initializes the channels by restore the mode registers.
Like we discussed with Vinod [1] the DMA controller drivers should go to
suspend after users and come back before them.

After you reconsider this point the patch logic might be modified a lot.

Looked through that discussions, I really had thought such problem for a while.

For the dma-controller and dma-user, we don't know which .suspend callback is executed firstly, so the idea would be: which ever is called earlier, the dma-controller driver suspend function should be as robust as possible.

It is better the dma-user .suspend callback is called earlier, some clean-ups should be done here such as stop dma request, but we cannot make sure every dma-user has a .suspend callback, some dma-users don't pay attention to or even don't care the suspend at all for some reason.

So even the suspend_late is used, we cannot assume every dma-user's activity is cleaned up neatly, dma-controller driver should be robust to handle this gracefully, that was my design target.

In the prepare() function, clean up the pending queue and stop receive new coming request, and in the suspend() function I do some register saving works. I don't think my code needs to be modified much, a possible change according to Vinod's idea would be:
use .suspend instead of my current .prepare
use . suspend_late instead of my current .suspend
e.g. postpone all my functions to be executed later, this method works and seems better for client/low-level module drivers.

The reason I didn't use the above functions was that I had read this Documentation/power/devices.txt, search the definitions of prepare, suspend, suspend_late and suspend_noirq, I think my usage complies with that definitions strictly, I can do the modification above if the maintainer like and if nobody says I break this law.

(Moreover, you abuse your own position to use only setters/getters to
access to the DMAc registers)

My shame, I will update it.
(reason is the setters/getters patch isn't merged into our internal tree, but this suspend patch has been done. I am trying to sync our internal kernel and the community now)

[1] http://www.spinics.net/lists/kernel/msg1650974.html


Signed-off-by: Hongbo Zhang <hongbo.zhang@xxxxxxxxxxxxx>
---
  drivers/dma/fsldma.c |  100 ++++++++++++++++++++++++++++++++++++++++++++++++++
  drivers/dma/fsldma.h |   16 ++++++++
  2 files changed, 116 insertions(+)

diff --git a/drivers/dma/fsldma.c b/drivers/dma/fsldma.c
index c9bf54a..d6da222 100644
--- a/drivers/dma/fsldma.c
+++ b/drivers/dma/fsldma.c
@@ -400,6 +400,14 @@ static dma_cookie_t fsl_dma_tx_submit(struct dma_async_tx_descriptor *tx)
spin_lock_bh(&chan->desc_lock); +#ifdef CONFIG_PM
+	if (unlikely(chan->pm_state != RUNNING)) {
+		chan_dbg(chan, "cannot submit due to suspend\n");
+		spin_unlock_bh(&chan->desc_lock);
+		return -1;
+	}
+#endif
+
  	/*
  	 * assign cookies to all of the software descriptors
  	 * that make up this transaction
@@ -1311,6 +1319,9 @@ static int fsl_dma_chan_probe(struct fsldma_device *fdev,
  	INIT_LIST_HEAD(&chan->ld_running);
  	INIT_LIST_HEAD(&chan->ld_completed);
  	chan->idle = true;
+#ifdef CONFIG_PM
+	chan->pm_state = RUNNING;
+#endif
chan->common.device = &fdev->common;
  	dma_cookie_init(&chan->common);
@@ -1450,6 +1461,92 @@ static int fsldma_of_remove(struct platform_device *op)
  	return 0;
  }
+#ifdef CONFIG_PM
+static int fsldma_prepare(struct device *dev)
+{
+	struct platform_device *pdev = to_platform_device(dev);
+	struct fsldma_device *fdev = platform_get_drvdata(pdev);
+	struct fsldma_chan *chan;
+	int i;
+
+	for (i = 0; i < FSL_DMA_MAX_CHANS_PER_DEVICE; i++) {
+		chan = fdev->chan[i];
+		if (!chan)
+			continue;
+
+		spin_lock_bh(&chan->desc_lock);
+		chan->pm_state = SUSPENDING;
+		if (!list_empty(&chan->ld_pending))
+			fsl_chan_xfer_ld_queue(chan);
+		spin_unlock_bh(&chan->desc_lock);
+	}
+
+	return 0;
+}
+
+static int fsldma_suspend(struct device *dev)
+{
+	struct platform_device *pdev = to_platform_device(dev);
+	struct fsldma_device *fdev = platform_get_drvdata(pdev);
+	struct fsldma_chan *chan;
+	int i;
+
+	for (i = 0; i < FSL_DMA_MAX_CHANS_PER_DEVICE; i++) {
+		chan = fdev->chan[i];
+		if (!chan)
+			continue;
+
+		spin_lock_bh(&chan->desc_lock);
+		if (!chan->idle)
+			goto out;
+		chan->regs_save.mr = DMA_IN(chan, &chan->regs->mr, 32);
+		chan->pm_state = SUSPENDED;
+		spin_unlock_bh(&chan->desc_lock);
+	}
+	return 0;
+
+out:
+	for (; i >= 0; i--) {
+		chan = fdev->chan[i];
+		if (!chan)
+			continue;
+		chan->pm_state = RUNNING;
+		spin_unlock_bh(&chan->desc_lock);
+	}
+	return -EBUSY;
+}
+
+static int fsldma_resume(struct device *dev)
+{
+	struct platform_device *pdev = to_platform_device(dev);
+	struct fsldma_device *fdev = platform_get_drvdata(pdev);
+	struct fsldma_chan *chan;
+	u32 mode;
+	int i;
+
+	for (i = 0; i < FSL_DMA_MAX_CHANS_PER_DEVICE; i++) {
+		chan = fdev->chan[i];
+		if (!chan)
+			continue;
+
+		spin_lock_bh(&chan->desc_lock);
+		mode = chan->regs_save.mr
+			& ~FSL_DMA_MR_CS & ~FSL_DMA_MR_CC & ~FSL_DMA_MR_CA;
+		DMA_OUT(chan, &chan->regs->mr, mode, 32);
+		chan->pm_state = RUNNING;
+		spin_unlock_bh(&chan->desc_lock);
+	}
+
+	return 0;
+}
+
+static const struct dev_pm_ops fsldma_pm_ops = {
+	.prepare	= fsldma_prepare,
+	.suspend	= fsldma_suspend,
+	.resume		= fsldma_resume,
+};
+#endif
+
  static const struct of_device_id fsldma_of_ids[] = {
  	{ .compatible = "fsl,elo3-dma", },
  	{ .compatible = "fsl,eloplus-dma", },
@@ -1462,6 +1559,9 @@ static struct platform_driver fsldma_of_driver = {
  		.name = "fsl-elo-dma",
  		.owner = THIS_MODULE,
  		.of_match_table = fsldma_of_ids,
+#ifdef CONFIG_PM
+		.pm = &fsldma_pm_ops,
+#endif
  	},
  	.probe = fsldma_of_probe,
  	.remove = fsldma_of_remove,
diff --git a/drivers/dma/fsldma.h b/drivers/dma/fsldma.h
index ec19517..eecaf9e 100644
--- a/drivers/dma/fsldma.h
+++ b/drivers/dma/fsldma.h
@@ -134,6 +134,18 @@ struct fsldma_device {
  #define FSL_DMA_CHAN_PAUSE_EXT	0x00001000
  #define FSL_DMA_CHAN_START_EXT	0x00002000
+#ifdef CONFIG_PM
+struct fsldma_chan_regs_save {
+	u32 mr;
+};
+
+enum fsldma_pm_state {
+	RUNNING = 0,
+	SUSPENDING,
+	SUSPENDED,
+};
+#endif
+
  struct fsldma_chan {
  	char name[8];			/* Channel name */
  	struct fsldma_chan_regs __iomem *regs;
@@ -161,6 +173,10 @@ struct fsldma_chan {
  	struct tasklet_struct tasklet;
  	u32 feature;
  	bool idle;			/* DMA controller is idle */
+#ifdef CONFIG_PM
+	struct fsldma_chan_regs_save regs_save;
+	enum fsldma_pm_state pm_state;
+#endif
void (*toggle_ext_pause)(struct fsldma_chan *fsl_chan, int enable);
  	void (*toggle_ext_start)(struct fsldma_chan *fsl_chan, int enable);




--
To unsubscribe from this list: send the line "unsubscribe dmaengine" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html




[Index of Archives]     [Linux Kernel]     [Linux ARM (vger)]     [Linux ARM MSM]     [Linux Omap]     [Linux Arm]     [Linux Tegra]     [Fedora ARM]     [Linux for Samsung SOC]     [eCos]     [Linux PCI]     [Linux Fastboot]     [Gcc Help]     [Git]     [DCCP]     [IETF Announce]     [Security]     [Linux MIPS]     [Yosemite Campsites]

  Powered by Linux