[PATCH v2 00/13] dmaengine: imx-dma: merge dma arch code into dmaengine.
From: javier Martin <hidden>
Date: 2012-03-09 12:04:38
Also in:
lkml
Hi Sascha, thank you for your time. On 9 March 2012 11:56, Sascha Hauer [off-list ref] wrote:
On Wed, Mar 07, 2012 at 03:33:32PM +0100, Javier Martin wrote:quoted
The following series have to be applied on top of the following patches: [PATCH v2 1/2] dmaengine: Add support for MEMCPY for imx-dma. [PATCH v3 2/2] dmaengine: Add support for multiple descriptors for imx-dma. As discussed with Vinod and Sascha, the file 'arch/arm/mach-imx/dma-v1.c' is deprecated and no new patches are accepted for it. This series merges this code into the imx-dma.c dmaengine driver. Changes since v1: ?- Merge patch 14 with 01. ?- Add a proper changelog to each patch of the series. [PATCH v2 01/13] dmaengine: imx-dma: merge old dma-v1.c with imx-dma.c [PATCH v2 02/13] dmaengine: imx-dma: remove data member from internal structure. [PATCH v2 03/13] dmaengine: imx-dma: remove dma_mode member of internal structure. [PATCH v2 04/13] dmaengine: imx-dma: remove 'imxdma_setup_mem2mem_hw' function. [PATCH v2 05/13] dmaengine: imx-dma: remove 'imxdma_config_channel_hw' function. [PATCH v2 06/13] dmaengine: imx-dma: remove 'imxdma_setup_sg_hw' function. [PATCH v2 07/13] dmaengine: imx-dma: remove sg member from internal structure. [PATCH v2 08/13] dmaengine: imx-dma: remove 'in_use' field of 'internal' structure. [PATCH v2 09/13] dmaengine: imx-dma: remove 'resbytes' field of 'internal' structure. [PATCH v2 10/13] dmaengine: imx-dma: remove internal structure. [PATCH v2 11/13] dmaengine: imx-dma: remove unused arg of imxdma_sg_next. [PATCH v2 12/13] dmaengine: imx-dma: remove 'imx_dmav1_baseaddr' and 'dma_clk'. [PATCH v2 13/13] dmaengine: imx-dma: use 'dev_dbg' and 'dev_warn' for messages.I didn't look at the individual patches but only on the end result and overall it looks good. There are several things left for further cleanup: - use resources for irq and base address - use platform device id rather than cpu_is_* to distinguish ?between the different device types. - request_irq for i.MX21/27 in alloc_chan_resources instead in probe - Use all available channels. In the time we had dmaengine users and ?legacy API users 8 channels were reserved for the old API and 8 ?channels for the dmaengine API. I think the first three points can be done in a later series, the last one should be trivial to fix now.
I agree. Let me fix the last issue you mention in a v3 of the involved patch.
Some comments follow to the individual patches. And, oh: Tested-by: Sascha Hauer <s.hauer@pengutronix.de
Cool. Did you test it for any other platform than i.MX27? Regards. -- Javier Martin Vista Silicon S.L. CDTUC - FASE C - Oficina S-345 Avda de los Castros s/n 39005- Santander. Cantabria. Spain +34 942 25 32 60 www.vista-silicon.com