linux-toradex.git/drivers/dma/dw/core.c, branch v3.17.2

dmaengine: dw: don't perform DMA when dmaengine_submit is called

2014-07-15T16:44:30+00:00

Accordingly to discussion [1] and followed up documentation the DMA controller
driver shouldn't start any DMA operations when dmaengine_submit() is called.

This patch fixes the workflow in dw_dmac driver to follow the documentation.

[1] http://www.spinics.net/lists/arm-kernel/msg125987.html

Signed-off-by: Andy Shevchenko 
Signed-off-by: Vinod Koul

dmaengine: dw: add debug message to dwc_dostart_first_queued

2014-07-15T16:44:29+00:00

It would be useful to know when the first descriptor in the queue is started
along with its cookie.

Signed-off-by: Andy Shevchenko 
Signed-off-by: Vinod Koul

dmaengine: dw: introduce dwc_dostart_first_queued() helper

2014-07-15T16:44:28+00:00

We have a duplicate code which starts first descriptor in the queue. Let's make
this as a separate helper that can be used in future as well.

Signed-off-by: Andy Shevchenko 
Signed-off-by: Vinod Koul

dmaengine: dw: check return code of dma_async_device_register()

2014-05-22T10:07:25+00:00

dma_async_device_register() may return non-zero error code. In such case we
have to follow error path.

Signed-off-by: Andy Shevchenko 
Signed-off-by: Vinod Koul

dmaengine: dw: fix regression in dw_probe() function

2014-05-22T10:07:24+00:00

The commit dbde5c29 "dw_dmac: use devm_* functions to simplify code" turns
probe function to use devm_* helpers and simultaneously brings a regression.

We have to 1) call clk_disable_unprepare() on error path, and 2) check error
code of clk_enable_prepare(). First part was done in the original code, second
one is an update.

Signed-off-by: Andy Shevchenko 
Signed-off-by: Vinod Koul

dmaengine: dw: enable clock before access

2014-05-22T10:07:24+00:00

hclk signal is a bus clock. So, it means we have to have it enabled during
access to the DMA controller. This patch makes sure that we enable clock before
access to the device, though it currently works on Intel hardware.

Signed-off-by: Andy Shevchenko 
Signed-off-by: Vinod Koul

dmaengine: dw: went back to plain {request,free}_irq() calls

2014-05-07T10:04:45+00:00

The commit dbde5c29 "dw_dmac: use devm_* functions to simplify code" turns
probe function to use devm_* helpers and simultaneously brings a regression. We
need to ensure irq is disabled, followed by ensuring that don't schedule any
more tasklets and then its safe to use tasklet_kill().

The free_irq() will ensure that the irq is disabled and also wait till all
scheduled interrupts are executed by invoking synchronize_irq(). So we need to
only do tasklet_kill() after invoking free_irq().

Signed-off-by: Andy Shevchenko 
Cc: stable@vger.kernel.org # v3.11+
Signed-off-by: Vinod Koul

dma: dw: allocate memory in two stages in probe

2014-03-26T06:14:27+00:00

This makes the probe() function a little bit clearer.

Signed-off-by: Andy Shevchenko 
Signed-off-by: Vinod Koul

dma: dw: remove leftovers in the comment blocks

2014-02-17T08:46:54+00:00

There is couple of leftovers in the comment blocks. This patch modifies the
comments accordingly.

There is no functional change.

Signed-off-by: Andy Shevchenko 
Signed-off-by: Vinod Koul

dma: dw: use %pad instead of casting dma_addr_t

2014-01-20T08:26:37+00:00

Since we have nice helper to print dma_addr_t values by reference we may use it
instead of explicit casting to a longest type.

Signed-off-by: Andy Shevchenko 
Signed-off-by: Vinod Koul