[PATCH 1/2] dma: pxa_dma: fix the no-requestor case

[PATCH 1/2] dma: pxa_dma: fix the no-requestor case

[Robert Jarzmik]

A very small number of devices don't use the flow control offered by
requestor lines. In these specific cases, the pxa dma driver should be
aware of that and not try to use a requestor line.

Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr>
---
 drivers/dma/pxa_dma.c | 20 ++++++++++++++------
 1 file changed, 14 insertions(+), 6 deletions(-)

diff --git a/drivers/dma/pxa_dma.c b/drivers/dma/pxa_dma.c
index bb85cd7b5b70..5647c1579c44 100644
--- a/drivers/dma/pxa_dma.c
+++ b/drivers/dma/pxa_dma.c
@@ -474,8 +474,10 @@ static void pxad_free_phy(struct pxad_chan *chan)
 		return;
 
 	/* clear the channel mapping in DRCMR */
-	reg = pxad_drcmr(chan->drcmr);
-	writel_relaxed(0, chan->phy->base + reg);
+	if (chan->drcmr <= DRCMR_CHLNUM) {
+		reg = pxad_drcmr(chan->drcmr);
+		writel_relaxed(0, chan->phy->base + reg);
+	}
 
 	spin_lock_irqsave(&pdev->phy_lock, flags);
 	for (i = 0; i < 32; i++)
@@ -517,8 +519,10 @@ static void phy_enable(struct pxad_phy *phy, bool misaligned)
 		"%s(); phy=%p(%d) misaligned=%d\n", __func__,
 		phy, phy->idx, misaligned);
 
-	reg = pxad_drcmr(phy->vchan->drcmr);
-	writel_relaxed(DRCMR_MAPVLD | phy->idx, phy->base + reg);
+	if (phy->vchan->drcmr <= DRCMR_CHLNUM) {
+		reg = pxad_drcmr(phy->vchan->drcmr);
+		writel_relaxed(DRCMR_MAPVLD | phy->idx, phy->base + reg);
+	}
 
 	dalgn = phy_readl_relaxed(phy, DALGN);
 	if (misaligned)
@@ -911,14 +915,18 @@ static void pxad_get_config(struct pxad_chan *chan,
 		width = chan->cfg.src_addr_width;
 		dev_addr = chan->cfg.src_addr;
 		*dev_src = dev_addr;
-		*dcmd |= PXA_DCMD_INCTRGADDR | PXA_DCMD_FLOWSRC;
+		*dcmd |= PXA_DCMD_INCTRGADDR;
+		if (chan->drcmr <= DRCMR_CHLNUM)
+			*dcmd |= PXA_DCMD_FLOWSRC;
 	}
 	if (chan->cfg.direction == DMA_MEM_TO_DEV) {
 		maxburst = chan->cfg.dst_maxburst;
 		width = chan->cfg.dst_addr_width;
 		dev_addr = chan->cfg.dst_addr;
 		*dev_dst = dev_addr;
-		*dcmd |= PXA_DCMD_INCSRCADDR | PXA_DCMD_FLOWTRG;
+		*dcmd |= PXA_DCMD_INCSRCADDR;
+		if (chan->drcmr <= DRCMR_CHLNUM)
+			*dcmd |= PXA_DCMD_FLOWTRG;
 	}
 	if (chan->cfg.direction == DMA_MEM_TO_MEM)
 		*dcmd |= PXA_DCMD_BURST32 | PXA_DCMD_INCTRGADDR |
-- 
2.1.4

[PATCH 2/2] dma: pxa_dma: fix residue corner case

[Robert Jarzmik]

A very tiny temporal window exists in the residue calculation where :
 - upon entering residue calculation, the transfer is ongoing
 - when reading the current transfer pointer, it just changed to
   the "finisher/linker" descriptor

In this case, the he residue returned is thw whole transfer length
instead of 0. Fix it.

This appears almost in one extreme case, where the driver is used
by older clients which inquire for residue in interrupt context, such
as the xmsc91x ethernet driver, in a tight loop :
  interrupt_handler()
    dmaengine_submit()
    do {
      dmaengine_tx_status()
    } while (residue > 0 || status != DMA_ERROR)

Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr>
---
 drivers/dma/pxa_dma.c | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/drivers/dma/pxa_dma.c b/drivers/dma/pxa_dma.c
index 5647c1579c44..bf064277518e 100644
--- a/drivers/dma/pxa_dma.c
+++ b/drivers/dma/pxa_dma.c
@@ -1186,6 +1186,16 @@ static unsigned int pxad_residue(struct pxad_chan *chan,
 	else
 		curr = phy_readl_relaxed(chan->phy, DTADR);
 
+	/*
+	 * curr has to be actually read before checking descriptor
+	 * completion, so that a curr inside a status updater
+	 * descriptor implies the following test returns true, and
+	 * preventing reordering of curr load and the test.
+	 */
+	rmb();
+	if (is_desc_completed(vd))
+		goto out;
+
 	for (i = 0; i < sw_desc->nb_desc - 1; i++) {
 		hw_desc = sw_desc->hw_desc[i];
 		if (sw_desc->hw_desc[0]->dcmd & PXA_DCMD_INCSRCADDR)
-- 
2.1.4

[PATCH 2/2] dma: pxa_dma: fix residue corner case

[Vinod Koul]

On Thu, Sep 10, 2015 at 07:17:46PM +0200, Robert Jarzmik wrote:
> A very tiny temporal window exists in the residue calculation where :
>  - upon entering residue calculation, the transfer is ongoing
>  - when reading the current transfer pointer, it just changed to
>    the "finisher/linker" descriptor
> 
> In this case, the he residue returned is thw whole transfer length
> instead of 0. Fix it.
Two many typos here :(


-- 
~Vinod

[PATCH 1/2] dma: pxa_dma: fix the no-requestor case

[Vinod Koul]

On Thu, Sep 10, 2015 at 07:17:45PM +0200, Robert Jarzmik wrote:
> A very small number of devices don't use the flow control offered by
> requestor lines. In these specific cases, the pxa dma driver should be
> aware of that and not try to use a requestor line.

And these fail for me, what were these sent against?
Please rebase and resend

-- 
~Vinod

[PATCH 2/2] dma: pxa_dma: fix residue corner case

[Robert Jarzmik]

Vinod Koul <vinod.koul@intel.com> writes:

> On Thu, Sep 10, 2015 at 07:17:46PM +0200, Robert Jarzmik wrote:
>> A very tiny temporal window exists in the residue calculation where :
>>  - upon entering residue calculation, the transfer is ongoing
>>  - when reading the current transfer pointer, it just changed to
>>    the "finisher/linker" descriptor
>> 
>> In this case, the he residue returned is thw whole transfer length
>> instead of 0. Fix it.
> Two many typos here :(
Ah sorry for that, I'll resend a v2 today.

Cheers.

-- 
Robert

[PATCH 1/2] dma: pxa_dma: fix the no-requestor case

[Robert Jarzmik]

Vinod Koul <vinod.koul@intel.com> writes:

> On Thu, Sep 10, 2015 at 07:17:45PM +0200, Robert Jarzmik wrote:
>> A very small number of devices don't use the flow control offered by
>> requestor lines. In these specific cases, the pxa dma driver should be
>> aware of that and not try to use a requestor line.
>
> And these fail for me, what were these sent against?
v4.2. My fault, in v4.2-rc1 the patch 0e95fb9ceb82 ("dmaengine: pxa_dma: don't
use config direction parameter") impact this one.

> Please rebase and resend
Of course, I'm on it.

Cheers.

-- 
Robert