[alsa-devel] [PATCH 00/12] dmaengine: remove users of device_control
The recent discussion [1] on the API have resulted in moving away from device_control ioctl method to proper channel APIs. There are still few users on the device_control which should use the wrappers existing rather than access device_control. This will aid us in deprecating and removing device_control, possibly after the merge window.
These can be merged thru respective subsystem tree or dmaengine tree. Either way please just let me know.
Feng's kbuild has tested these as well [2]
[1]: http://www.spinics.net/lists/dmaengine/msg02212.html [2]: http://git.infradead.org/users/vkoul/slave-dma.git/shortlog/refs/heads/topic...
Vinod Koul (12): pata_arasan_cf: use dmaengine_terminate_all() API dmaengine: coh901318: use dmaengine_terminate_all() API [media] V4L2: mx3_camer: use dmaengine_pause() API mtd: fsmc_nand: use dmaengine_terminate_all() API mtd: sh_flctl: use dmaengine_terminate_all() API net: ks8842: use dmaengine_terminate_all() API spi/atmel: use dmaengine_terminate_all() API spi/spi-dw-mid.c: use dmaengine_slave_config() API serial: sh-sci: use dmaengine_terminate_all() API usb: musb: ux500_dma: use dmaengine_xxx() APIs ASoC: txx9: use dmaengine_terminate_all() API video: mx3fb: use dmaengine_terminate_all() API
drivers/ata/pata_arasan_cf.c | 5 ++--- drivers/dma/coh901318.c | 2 +- drivers/media/platform/soc_camera/mx3_camera.c | 6 ++---- drivers/mtd/nand/fsmc_nand.c | 2 +- drivers/mtd/nand/sh_flctl.c | 2 +- drivers/net/ethernet/micrel/ks8842.c | 6 ++---- drivers/spi/spi-atmel.c | 6 ++---- drivers/spi/spi-dw-mid.c | 6 ++---- drivers/tty/serial/sh-sci.c | 2 +- drivers/usb/musb/ux500_dma.c | 7 ++----- drivers/video/fbdev/mx3fb.c | 3 +-- sound/soc/txx9/txx9aclc.c | 7 +++---- 12 files changed, 20 insertions(+), 34 deletions(-)
The drivers should use dmaengine_terminate_all() API instead of accessing the device_control which will be deprecated soon
Signed-off-by: Vinod Koul vinod.koul@intel.com --- sound/soc/txx9/txx9aclc.c | 7 +++---- 1 files changed, 3 insertions(+), 4 deletions(-)
diff --git a/sound/soc/txx9/txx9aclc.c b/sound/soc/txx9/txx9aclc.c index f0829de..3b122cd 100644 --- a/sound/soc/txx9/txx9aclc.c +++ b/sound/soc/txx9/txx9aclc.c @@ -160,7 +160,7 @@ static void txx9aclc_dma_tasklet(unsigned long data) void __iomem *base = drvdata->base;
spin_unlock_irqrestore(&dmadata->dma_lock, flags); - chan->device->device_control(chan, DMA_TERMINATE_ALL, 0); + dmaengine_terminate_all(chan); /* first time */ for (i = 0; i < NR_DMA_CHAIN; i++) { desc = txx9aclc_dma_submit(dmadata, @@ -266,7 +266,7 @@ static int txx9aclc_pcm_close(struct snd_pcm_substream *substream) struct dma_chan *chan = dmadata->dma_chan;
dmadata->frag_count = -1; - chan->device->device_control(chan, DMA_TERMINATE_ALL, 0); + dmaengine_terminate_all(chan); return 0; }
@@ -398,8 +398,7 @@ static int txx9aclc_pcm_remove(struct snd_soc_platform *platform) struct dma_chan *chan = dmadata->dma_chan; if (chan) { dmadata->frag_count = -1; - chan->device->device_control(chan, - DMA_TERMINATE_ALL, 0); + dmaengine_terminate_all(chan); dma_release_channel(chan); } dev->dmadata[i].dma_chan = NULL;
On 10/11/2014 05:40 PM, Vinod Koul wrote:
The drivers should use dmaengine_terminate_all() API instead of accessing the device_control which will be deprecated soon
I already send a similar patch a while ago. It is on it's way into 3.18.
Signed-off-by: Vinod Koul vinod.koul@intel.com
On Sat, Oct 11, 2014 at 06:21:47PM +0200, Lars-Peter Clausen wrote:
On 10/11/2014 05:40 PM, Vinod Koul wrote:
The drivers should use dmaengine_terminate_all() API instead of accessing the device_control which will be deprecated soon
I already send a similar patch a while ago. It is on it's way into 3.18.
I will drop this one then
I did grep the linux-next when I started on this and found no reductions from what I had in dmaengine-next
On Sat, Oct 11, 2014 at 09:24:11PM +0530, Vinod Koul wrote:
On Sat, Oct 11, 2014 at 06:21:47PM +0200, Lars-Peter Clausen wrote:
I already send a similar patch a while ago. It is on it's way into 3.18.
I will drop this one then
I did grep the linux-next when I started on this and found no reductions from what I had in dmaengine-next
It's ff495d3a8ea4d (ASoC: txx9: Don't opencode DMAengine API calls) from August which is in Linus' tree now.
On Sat, Oct 11, 2014 at 09:09:33PM +0530, Vinod Koul wrote:
The recent discussion [1] on the API have resulted in moving away from device_control ioctl method to proper channel APIs. There are still few users on the device_control which should use the wrappers existing rather than access device_control. This will aid us in deprecating and removing device_control, possibly after the merge window.
These can be merged thru respective subsystem tree or dmaengine tree. Either way please just let me know.
Applying to dmaengine-next with acks recived, dropping the ones which are applied by respective folks
participants (3)
-
Lars-Peter Clausen -
Mark Brown -
Vinod Koul