I missed two comments first time around... On 06/02/15 13:31, Peter Ujfalusi wrote:
On 06/02/2015 11:58 AM, Jyri Sarha wrote:Find the configured DMA controller by asking for a DMA channel in the probe phase and releasing it right after. The controller device can be found via the dma_chan struct and the controller is recognized from the compatible property of its device node. The patch assumes EDMA if there is no device node. Signed-off-by: Jyri Sarha <jsarha@xxxxxx> --- sound/soc/davinci/davinci-mcasp.c | 61 ++++++++++++++++++++++++++++++++------- 1 file changed, 51 insertions(+), 10 deletions(-) diff --git a/sound/soc/davinci/davinci-mcasp.c b/sound/soc/davinci/davinci-mcasp.c index d793494..5bc2712 100644 --- a/sound/soc/davinci/davinci-mcasp.c +++ b/sound/soc/davinci/davinci-mcasp.c @@ -1565,6 +1565,46 @@ static int davinci_mcasp_init_ch_constraints(struct davinci_mcasp *mcasp) return ret; } +enum { + MCASP_EDMA = 1,Why start from 1? I'm not sure about the name. It somehow implies that McASP has something to do with the dma engine. MCASP_SERVICED_BY_E/SDMA PLATFORM_USES_E/SDMA I can not come up with a decent name :(+ MCASP_SDMA, +}; +static const char *sdma_prefix = "ti,omap"; + +static int davinci_mcasp_get_dma_type(struct davinci_mcasp *mcasp) +{ + struct dma_chan *chan; + const char *tmp; + int ret = MCASP_EDMA; + + tmp = mcasp->dma_data[SNDRV_PCM_STREAM_PLAYBACK].filter_data; + chan = dma_request_slave_channel_reason(mcasp->dev, tmp);this will return with error in case of !dev->of_node. The first check should be: if (!mcasp->dev->of_node) return MCASP_EDMA; since we can only boot in legacy mode with daVinci devices and they are using eDMA.
Yep, I'll add that back. I tought that I would not need it, but am not so familiar with the generic dmaengine pcm...
+ if (IS_ERR(chan)) { + if (PTR_ERR(chan) != -EPROBE_DEFER) + dev_err(mcasp->dev, + "Can't verify DMA configuration (%ld)\n", + PTR_ERR(chan)); + return PTR_ERR(chan); + } + BUG_ON(!chan->device || !chan->device->dev); + + if (chan->device->dev->of_node) + ret = of_property_read_string(chan->device->dev->of_node, + "compatible", &tmp); + else + dev_dbg(mcasp->dev, "DMA controller has no of-node\n");Now this is not likely to happen...+ + dma_release_channel(chan); + if (ret) + return ret; + + dev_dbg(mcasp->dev, "DMA controller compatible = \"%s\"\n", tmp); + if (!strncmp(tmp, sdma_prefix, strlen(sdma_prefix))) + return MCASP_SDMA; + + return MCASP_EDMA; +} + static int davinci_mcasp_probe(struct platform_device *pdev) { struct snd_dmaengine_dai_dma_data *dma_data; @@ -1763,27 +1803,28 @@ static int davinci_mcasp_probe(struct platform_device *pdev) if (ret != 0) goto err; - switch (mcasp->version) { + ret = -EINVAL;Why? What was the problem with setting the ret in the default case?
I would need to have that I three places now for the #else cases, when devicetree tries to configure pcm driver that has not been compiled.
+ switch (davinci_mcasp_get_dma_type(mcasp)) { + case MCASP_EDMA: #if IS_BUILTIN(CONFIG_SND_EDMA_SOC) || \ (IS_MODULE(CONFIG_SND_DAVINCI_SOC_MCASP) && \ IS_MODULE(CONFIG_SND_EDMA_SOC)) - case MCASP_VERSION_1: - case MCASP_VERSION_2: - case MCASP_VERSION_3: ret = edma_pcm_platform_register(&pdev->dev); - break; +#else + dev_err(&pdev->dev, "Missing SND_EDMA_SOC\n"); #endif + break; + case MCASP_SDMA: #if IS_BUILTIN(CONFIG_SND_OMAP_SOC) || \ (IS_MODULE(CONFIG_SND_DAVINCI_SOC_MCASP) && \ IS_MODULE(CONFIG_SND_OMAP_SOC)) - case MCASP_VERSION_4: ret = omap_pcm_platform_register(&pdev->dev); - break; +#else + dev_err(&pdev->dev, "Missing SND_SDMA_SOC\n"); #endif + break; default: - dev_err(&pdev->dev, "Invalid McASP version: %d\n", - mcasp->version); - ret = -EINVAL; + dev_err(&pdev->dev, "No DMA controller found\n"); break; }
-- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
