2015-07-15 09:37:54

by Yuan Yao

[permalink] [raw]
Subject:

This patch has been tested on Fresscale LS-1 SOCs.


2015-07-15 09:38:11

by Yuan Yao

[permalink] [raw]
Subject: [PATCH] dmaengine: fsl-edma: add PM suspend/resume support

This add power management suspend/resume support for the fsl-edma
driver.

eDMA acted as a basic function used by others. What it needs to do is
the two steps below to support power management.
In fsl_edma_suspend_late:
Check whether the DMA chan is idle and if it is not idle, stop PM
operation.
In fsl_edma_resume_early:
Enable the eDMA and wait for being used.

Signed-off-by: Yuan Yao <[email protected]>
---
drivers/dma/fsl-edma.c | 92 +++++++++++++++++++++++++++++++++++++++++++++++++-
1 file changed, 91 insertions(+), 1 deletion(-)

diff --git a/drivers/dma/fsl-edma.c b/drivers/dma/fsl-edma.c
index 915eec3..f987a4f 100644
--- a/drivers/dma/fsl-edma.c
+++ b/drivers/dma/fsl-edma.c
@@ -116,6 +116,12 @@
BIT(DMA_SLAVE_BUSWIDTH_2_BYTES) | \
BIT(DMA_SLAVE_BUSWIDTH_4_BYTES) | \
BIT(DMA_SLAVE_BUSWIDTH_8_BYTES)
+#ifdef CONFIG_PM
+enum fsl_edma_pm_state {
+ RUNNING = 0,
+ SUSPENDED,
+};
+#endif

struct fsl_edma_hw_tcd {
__le32 saddr;
@@ -151,6 +157,11 @@ struct fsl_edma_chan {
struct fsl_edma_desc *edesc;
struct fsl_edma_slave_config fsc;
struct dma_pool *tcd_pool;
+ bool idle;
+#ifdef CONFIG_PM
+ u32 slave_id;
+ enum fsl_edma_pm_state pm_state;
+#endif
};

struct fsl_edma_desc {
@@ -298,6 +309,7 @@ static int fsl_edma_terminate_all(struct dma_chan *chan)
spin_lock_irqsave(&fsl_chan->vchan.lock, flags);
fsl_edma_disable_request(fsl_chan);
fsl_chan->edesc = NULL;
+ fsl_chan->idle = true;
vchan_get_all_descriptors(&fsl_chan->vchan, &head);
spin_unlock_irqrestore(&fsl_chan->vchan.lock, flags);
vchan_dma_desc_free_list(&fsl_chan->vchan, &head);
@@ -313,6 +325,7 @@ static int fsl_edma_pause(struct dma_chan *chan)
if (fsl_chan->edesc) {
fsl_edma_disable_request(fsl_chan);
fsl_chan->status = DMA_PAUSED;
+ fsl_chan->idle = true;
}
spin_unlock_irqrestore(&fsl_chan->vchan.lock, flags);
return 0;
@@ -327,6 +340,7 @@ static int fsl_edma_resume(struct dma_chan *chan)
if (fsl_chan->edesc) {
fsl_edma_enable_request(fsl_chan);
fsl_chan->status = DMA_IN_PROGRESS;
+ fsl_chan->idle = false;
}
spin_unlock_irqrestore(&fsl_chan->vchan.lock, flags);
return 0;
@@ -648,6 +662,7 @@ static void fsl_edma_xfer_desc(struct fsl_edma_chan *fsl_chan)
fsl_edma_set_tcd_regs(fsl_chan, fsl_chan->edesc->tcd[0].vtcd);
fsl_edma_enable_request(fsl_chan);
fsl_chan->status = DMA_IN_PROGRESS;
+ fsl_chan->idle = false;
}

static irqreturn_t fsl_edma_tx_handler(int irq, void *dev_id)
@@ -676,6 +691,7 @@ static irqreturn_t fsl_edma_tx_handler(int irq, void *dev_id)
vchan_cookie_complete(&fsl_chan->edesc->vdesc);
fsl_chan->edesc = NULL;
fsl_chan->status = DMA_COMPLETE;
+ fsl_chan->idle = true;
} else {
vchan_cyclic_callback(&fsl_chan->edesc->vdesc);
}
@@ -704,6 +720,7 @@ static irqreturn_t fsl_edma_err_handler(int irq, void *dev_id)
edma_writeb(fsl_edma, EDMA_CERR_CERR(ch),
fsl_edma->membase + EDMA_CERR);
fsl_edma->chans[ch].status = DMA_ERROR;
+ fsl_edma->chans[ch].idle = true;
}
}
return IRQ_HANDLED;
@@ -724,6 +741,14 @@ static void fsl_edma_issue_pending(struct dma_chan *chan)

spin_lock_irqsave(&fsl_chan->vchan.lock, flags);

+#ifdef CONFIG_PM
+ if (unlikely(fsl_chan->pm_state != RUNNING)) {
+ spin_unlock_irqrestore(&fsl_chan->vchan.lock, flags);
+ /* cannot submit due to suspend */
+ return;
+ }
+#endif
+
if (vchan_issue_pending(&fsl_chan->vchan) && !fsl_chan->edesc)
fsl_edma_xfer_desc(fsl_chan);

@@ -750,6 +775,10 @@ static struct dma_chan *fsl_edma_xlate(struct of_phandle_args *dma_spec,
chan->device->privatecnt++;
fsl_edma_chan_mux(to_fsl_edma_chan(chan),
dma_spec->args[1], true);
+#ifdef CONFIG_PM
+ to_fsl_edma_chan(chan)->slave_id =
+ dma_spec->args[1];
+#endif
mutex_unlock(&fsl_edma->fsl_edma_mutex);
return chan;
}
@@ -888,7 +917,11 @@ static int fsl_edma_probe(struct platform_device *pdev)
struct fsl_edma_chan *fsl_chan = &fsl_edma->chans[i];

fsl_chan->edma = fsl_edma;
-
+#ifdef CONFIG_PM
+ fsl_chan->pm_state = RUNNING;
+ fsl_chan->slave_id = 0;
+#endif
+ fsl_chan->idle = true;
fsl_chan->vchan.desc_free = fsl_edma_free_desc;
vchan_init(&fsl_chan->vchan, &fsl_edma->dma_dev);

@@ -959,6 +992,60 @@ static int fsl_edma_remove(struct platform_device *pdev)
return 0;
}

+#ifdef CONFIG_PM
+static int fsl_edma_suspend_late(struct device *dev)
+{
+ struct fsl_edma_engine *fsl_edma = dev_get_drvdata(dev);
+ struct fsl_edma_chan *fsl_chan;
+ unsigned long flags;
+ int i;
+
+ for (i = 0; i < fsl_edma->n_chans; i++) {
+ fsl_chan = &fsl_edma->chans[i];
+ spin_lock_irqsave(&fsl_chan->vchan.lock, flags);
+ /* Make sure chan is idle or can not into suspend. */
+ if (unlikely(!fsl_chan->idle))
+ goto out;
+ fsl_chan->pm_state = SUSPENDED;
+ spin_unlock_irqrestore(&fsl_chan->vchan.lock, flags);
+ }
+ return 0;
+
+out:
+ for (; i >= 0; i--) {
+ fsl_chan = &fsl_edma->chans[i];
+ fsl_chan->pm_state = RUNNING;
+ spin_unlock_irqrestore(&fsl_chan->vchan.lock, flags);
+ }
+ return -EBUSY;
+}
+
+static int fsl_edma_resume_early(struct device *dev)
+{
+ struct fsl_edma_engine *fsl_edma = dev_get_drvdata(dev);
+ struct fsl_edma_chan *fsl_chan;
+ int i;
+
+ for (i = 0; i < fsl_edma->n_chans; i++) {
+ fsl_chan = &fsl_edma->chans[i];
+ fsl_chan->pm_state = RUNNING;
+ edma_writew(fsl_edma, 0x0, fsl_edma->membase + EDMA_TCD_CSR(i));
+ if (fsl_chan->slave_id != 0)
+ fsl_edma_chan_mux(fsl_chan, fsl_chan->slave_id, true);
+ }
+
+ edma_writel(fsl_edma, EDMA_CR_ERGA | EDMA_CR_ERCA,
+ fsl_edma->membase + EDMA_CR);
+
+ return 0;
+}
+
+static const struct dev_pm_ops fsl_edma_pm_ops = {
+ .suspend_late = fsl_edma_suspend_late,
+ .resume_early = fsl_edma_resume_early,
+};
+#endif
+
static const struct of_device_id fsl_edma_dt_ids[] = {
{ .compatible = "fsl,vf610-edma", },
{ /* sentinel */ }
@@ -969,6 +1056,9 @@ static struct platform_driver fsl_edma_driver = {
.driver = {
.name = "fsl-edma",
.of_match_table = fsl_edma_dt_ids,
+#ifdef CONFIG_PM
+ .pm = &fsl_edma_pm_ops,
+#endif
},
.probe = fsl_edma_probe,
.remove = fsl_edma_remove,
--
2.1.0.27.g96db324

2015-07-15 09:57:17

by Arnd Bergmann

[permalink] [raw]
Subject: Re: [PATCH] dmaengine: fsl-edma: add PM suspend/resume support

On Wednesday 15 July 2015 17:32:58 Yuan Yao wrote:
> +#ifdef CONFIG_PM
> +enum fsl_edma_pm_state {
> + RUNNING = 0,
> + SUSPENDED,
> +};
> +#endif
>
> struct fsl_edma_hw_tcd {
>

The #ifdefs here seem unnecessary, at least most of them, better
just do this all unconditionally.

Arnd

2015-07-15 10:29:58

by Yuan Yao

[permalink] [raw]
Subject: RE: [PATCH] dmaengine: fsl-edma: add PM suspend/resume support

Hi Arnd,

Thanks for your review.
And can you give me more information?
In my opinion, The fsl_edma_pm_state will just be used when CONFIG_PM support. So why not use the #ifdefs to remove the
unnecessary code? Since the PM will not be selected in many use cases.

Thanks.

Best Regards,
Yuan Yao

> -----Original Message-----
> From: Arnd Bergmann [mailto:[email protected]]
> Sent: Wednesday, July 15, 2015 5:57 PM
> To: [email protected]
> Cc: Yuan Yao-B46683; [email protected]; [email protected];
> [email protected]; [email protected]; linux-
> [email protected]
> Subject: Re: [PATCH] dmaengine: fsl-edma: add PM suspend/resume support
>
> On Wednesday 15 July 2015 17:32:58 Yuan Yao wrote:
> > +#ifdef CONFIG_PM
> > +enum fsl_edma_pm_state {
> > + RUNNING = 0,
> > + SUSPENDED,
> > +};
> > +#endif
> >
> > struct fsl_edma_hw_tcd {
> >
>
> The #ifdefs here seem unnecessary, at least most of them, better just do this
> all unconditionally.
>
> Arnd

2015-07-15 14:54:44

by Arnd Bergmann

[permalink] [raw]
Subject: Re: [PATCH] dmaengine: fsl-edma: add PM suspend/resume support

On Wednesday 15 July 2015 10:29:55 Yao Yuan wrote:
> Hi Arnd,
>
> Thanks for your review.
> And can you give me more information?
> In my opinion, The fsl_edma_pm_state will just be used when CONFIG_PM support. So why not use the #ifdefs to remove the
> unnecessary code? Since the PM will not be selected in many use cases.

I would consider code readability more important than saving a
few instructions, and the #ifdef interrupts the reading flow.
Another aspect is that the compiler does not produce warnings
for incorrect code in an #ifdef, so we try to use e.g. 'if (IS_ENABLED(CONFIG_FOO))'
instead of '#ifdef CONFIG_FOO' wherever possible.

Arnd

2015-07-16 08:17:42

by Yuan Yao

[permalink] [raw]
Subject: RE: [PATCH] dmaengine: fsl-edma: add PM suspend/resume support

Hi Arnd,
Ok, I will remove the #ifdef in the next version.
Thanks.

Hi All,
Is there any others comments?

Thanks for your review.

Best Regards,
Yuan Yao

On Wednesday, July 15, 2015 10:55 PM Arnd wrote:
> On Wednesday 15 July 2015 10:29:55 Yao Yuan wrote:
> > Hi Arnd,
> >
> > Thanks for your review.
> > And can you give me more information?
> > In my opinion, The fsl_edma_pm_state will just be used when CONFIG_PM
> > support. So why not use the #ifdefs to remove the unnecessary code? Since
> the PM will not be selected in many use cases.
>
> I would consider code readability more important than saving a few instructions,
> and the #ifdef interrupts the reading flow.
> Another aspect is that the compiler does not produce warnings for incorrect
> code in an #ifdef, so we try to use e.g. 'if (IS_ENABLED(CONFIG_FOO))'
> instead of '#ifdef CONFIG_FOO' wherever possible.
>
> Arnd

2015-07-16 13:22:31

by Vinod Koul

[permalink] [raw]
Subject: Re: [PATCH] dmaengine: fsl-edma: add PM suspend/resume support

On Wed, Jul 15, 2015 at 05:32:58PM +0800, Yuan Yao wrote:
> +
> +static const struct dev_pm_ops fsl_edma_pm_ops = {
> + .suspend_late = fsl_edma_suspend_late,
> + .resume_early = fsl_edma_resume_early,
any reason why you are using early/late handlers?

--
~Vinod

2015-07-17 05:56:24

by Yuan Yao

[permalink] [raw]
Subject: RE: [PATCH] dmaengine: fsl-edma: add PM suspend/resume support

Hi Vinod,

eDMA sullpies the data transmission service for other IPs, so it should be suspended later and resumed earlier.
For example:
Synchronous Audio Interface (SAI), SAI use DMA to transfer the data.
When suspend:
1, SAI using suspend handlers to stop the DMA transmission.
2, DMA using suspend_late handlers to check whether DMA chan is idle.
When resume:
1, DMA using resume_early to enable. And then can servicing for others.
2, SAI using resume and then can use DMA to data transmission.

We must sure the order is right, so I think DMA should using early/late handlers.

Thanks.

Best Regards,
Yuan Yao

On Thursday, July 16, 2015 9:24 PM, Vinod wrote:
> On Wed, Jul 15, 2015 at 05:32:58PM +0800, Yuan Yao wrote:
> > +
> > +static const struct dev_pm_ops fsl_edma_pm_ops = {
> > + .suspend_late = fsl_edma_suspend_late,
> > + .resume_early = fsl_edma_resume_early,
> any reason why you are using early/late handlers?
>
> --
> ~Vinod