October 2012 - Alsa-devel - mailman.alsa-project.org

[alsa-devel] [PATCH 0/5] snd-ice1712: Add Philips PSC724 Ultimate Edge
by Ondrej Zary 19 Oct '12

19 Oct '12

Hello, this patch series adds full support for Philips PSC724 Ultimate Edge sound card to snd-ice1712 driver. Unlike other snd-ice1712 subdrivers, the codec code is splitted into separate files, reusable by other ice1712 subdrivers. Working: all analog outputs (front, rear, center+lfe, headphone) and inputs (front mic, rear mic, line, cd, aux), mixer, headphone jack detection, suspend/resume Untested: SPDIF -- Ondrej Zary

2 8

[alsa-devel] ESI drivers
by Neil Mcbennett 18 Oct '12

18 Oct '12

Hi, I am not a developer and probably shouldn't even be here, but I do have a question that is probably best answered from this forum. I've been looking at the ESI website and I notice that there is some ALSA support for a few of their cards... http://www.alsa-project.org/main/index.php/Matrix:Vendor-ESI The interface I'm most interested in isn't on this list (MaXiO 032e) I decided to call the technical support number in the UK and ask them what their position is regarding Linux. The guy I spoke to said that ESI are not going to provide an ALSA driver for these devices but they would be more than happy if someone from the ALSA community were to approach them. They would then assist in providing the necessary information. So, just wondering if anyone has already started on this or if there might be some interest in doing so? I was attracted to ESI by the fact that they have a number of PCIe interfaces which are still quite something of a rarity. Regards Neil

1 0

[alsa-devel] LAC 2013: the Linux Audio Conference - Call for Participation
by Robin Gareus 18 Oct '12

18 Oct '12

sorry for >< please >> << We are happy to announce the next issue of the Linux Audio Conference (LAC), May 9-12, 2013 @ IEM, the Institute of Electronic Music and Acoustics, in Graz, Austria. The Linux Audio Conference is an international conference that brings together musicians, sound artists, software developers and researchers, working with Linux as an open, stable, professional platform for audio and media research and music production. LAC includes paper sessions, workshops, and a diverse program of electronic music. *Call for Papers, Workshops, Music and Installations* We invite submissions of papers addressing all areas of audio processing and media creation based on Linux. Papers can focus on technical, artistic and scientific issues and should target developers or users. In our call for music, we are looking for works that have been produced or composed entirely/mostly using Linux. The online submission of papers, workshops, music and installations is now open at http://lac.iem.at/ The Deadline for all submissions is February 4th, 2013 (23:59 HAST) You are invited to register for participation on our conference website. There you will find up-to-date instructions, as well as important information about dates, travel, lodging, and so on. This year's conference is hosted by IEM, Graz, in cooperation with local artists and FLOSS enthusiasts. The Institute of Electronic Music and Acoustics (IEM) at the University of Music and Performing Arts Graz is considered Austria's leading institution in computer music, acoustics and audio engineering and has gained international reputation for its research on spatial audio and its artistic production and research. IEM has been embracing Linux audio as a production and research environment since the mid-1990s, and has contributed to FLOSS/Linux projects, amongst others by providing drivers for multichannel audio interfaces and hosting the Pure Data community portal and mailing lists. http://iem.at/ We look forward to seeing you in Graz in May! Sincerely, The LAC 2013 Organizing Team

1 0

[alsa-devel] [PATCH v2 0/2] Convert davinci ASoC to genalloc SRAM
by Matt Porter 18 Oct '12

18 Oct '12

Changes since v1: - Rebased against uio_pruss/genalloc series v4 patches, requiring another header to be included. This series converts davinci ASoC to use genalloc and enables that support on DA850. It applies to mainline on top of the uio_pruss/genalloc series [1] which allows DaVinci to provide a gen_pool via pdata for driver use. I've tested this on the AM180x EVM. Note that prior to this, the SRAM paths in the driver were completely unused. I've only enabled ping-pong buffering on the platform I can test as it's best to allow those with DM644x and similar platforms to set the playback/capture sram size to something that's known to work for them. [1] https://lkml.org/lkml/2012/10/5/425 Matt Porter (2): ASoC: davinci: replace private sram api with genalloc ARM: davinci: enable SRAM ping ping buffering on DA850 arch/arm/mach-davinci/board-da850-evm.c | 25 +++++++++++++++---------- include/linux/platform_data/davinci_asp.h | 3 +++ sound/soc/davinci/davinci-mcasp.c | 2 ++ sound/soc/davinci/davinci-pcm.c | 18 ++++++++++++------ sound/soc/davinci/davinci-pcm.h | 2 ++ 5 files changed, 34 insertions(+), 16 deletions(-) -- 1.7.9.5

4 9

Re: [alsa-devel] [PATCH] dma: add new DMA control commands
by Huang Shijie 18 Oct '12

18 Oct '12

On Thu, Oct 18, 2012 at 5:29 AM, Jassi Brar <jaswinder.singh(a)linaro.org> wrote: > On 18 October 2012 12:15, Huang Shijie <b32955(a)freescale.com> wrote: >> 于 2012年10月18日 14:18, Vinod Koul 写道: >> >>> Why cant you do start (prepare clock etc) when you submit the descriptor >>> to dmaengine. Can be done in tx_submit callback. >>> Similarly remove the clock when dma transaction gets completed. >> >> I ever thought this method too. >> >> But it will become low efficient in the following case: >> >> Assuming the gpmi-nand driver has to read out 1024 pages in one _SINGLE_ >> read operation. >> The gpmi-nand will submit the descriptor to dmaengine per page. So with your >> method, >> the system will repeat the enable/disable dma clock 1024 time. At every >> enable/disable dma clock, >> the system has to enable the clock chain and it's parents ... >> >> But with this patch, we only need to enable/disable dma clock one time, just >> at we select the nand chip. >> > If the clock is the dmac's property (not channels'), the toggling > seems too aggressive. > You could try using runtime_suspend/resume for clock > disabling/enabling. How about employing autosuspend with a few ms > delay? Yes, employing the autosuspend is workable method too.o But it's a little more complicated then this patch. You have to create a thread or workqueue to do the job. What's more, I think other DMA engine may also meets the same issue as the mxs-dma, such as the imx-sdma. It's somehow a common issue to disable the clocks when no one use the DMA engine. Do you also suggest employing the autosuspend? Why not use a simple way to handle this issue? When the dma device want to use a DMA engine, it just needs to issue a DMA_START command. thanks Huang Shijie Huang Shijie

1 0

Re: [alsa-devel] [PATCH v2 2/2] ARM: davinci: enable SRAM ping ping buffering on DA850
by Matt Porter 18 Oct '12

18 Oct '12

On Fri, Oct 12, 2012 at 07:11:06PM +0200, Sergei Shtylyov wrote: > Hello. > > On 11-10-2012 23:33, Matt Porter wrote: > > >Passes the DA850 shared SRAM gen_pool to the McASP driver > >and enables the ping-pong buffer DMA support. > > Here "ping-pong is correct but the subject have a typo. Thanks. Daniel: can you correct this when you squash these into your series? -Matt

1 0

[alsa-devel] [PATCH] dma: add new DMA control commands
by Huang Shijie 18 Oct '12

18 Oct '12

From: Huang Shijie <shijie8(a)gmail.com> [1] Why add these new DMA control commands? In mx6q, the gpmi-nand driver is the only user of the APBH-DMA. The dma clock is enabled when we have successfully requested a DMA channel. So even when the gpmi-nand driver does not work, the dma clock(apbh-dma) still runs in high speed (198MHz). To save some power, it is better to disable the dma clock when the dma device, such as gpmi-nand, is not working anymore. When the dma device becomes work again, enable the dma clock again. [2] add new DMA control commands: DMA_START/DMA_END DMA_START: do some preprations to start the DMA engine, such as enable the necessary clocks. DMA_END: do some works to end the DMA engine, such as disable the necessary clocks. [3] This patch does not change any logic in i2c-mxs driver and mxs-pcm driver. But for gpmi-nand driver, we will enable the the clock only when we select the nand chip, and want to do some real jobs with the nand chip. For mxs-mmc driver, disable the dma clock in the suspend; and enable the dma clock in the resume. Signed-off-by: Huang Shijie <b32955(a)freescale.com> --- drivers/dma/mxs-dma.c | 17 +++++++++++++++++ drivers/i2c/busses/i2c-mxs.c | 10 +++++++++- drivers/mmc/host/mxs-mmc.c | 19 ++++++++++++++++--- drivers/mtd/nand/gpmi-nand/gpmi-lib.c | 10 ++++++++++ drivers/mtd/nand/gpmi-nand/gpmi-nand.c | 10 ++++++---- include/linux/dmaengine.h | 6 ++++++ sound/soc/mxs/mxs-pcm.c | 12 ++++++++++++ 7 files changed, 76 insertions(+), 8 deletions(-) diff --git a/drivers/dma/mxs-dma.c b/drivers/dma/mxs-dma.c index 9f02e79..89286f4 100644 --- a/drivers/dma/mxs-dma.c +++ b/drivers/dma/mxs-dma.c @@ -384,6 +384,8 @@ static int mxs_dma_alloc_chan_resources(struct dma_chan *chan) /* the descriptor is ready */ async_tx_ack(&mxs_chan->desc); + clk_disable_unprepare(mxs_dma->clk); + return 0; err_clk: @@ -399,6 +401,14 @@ static void mxs_dma_free_chan_resources(struct dma_chan *chan) { struct mxs_dma_chan *mxs_chan = to_mxs_dma_chan(chan); struct mxs_dma_engine *mxs_dma = mxs_chan->mxs_dma; + int ret; + + ret = clk_prepare_enable(mxs_dma->clk); + if (ret) { + dev_err(mxs_dma->dma_device.dev, + "failed in enabling the dma clock\n"); + return; + } mxs_dma_disable_chan(mxs_chan); @@ -597,9 +607,13 @@ static int mxs_dma_control(struct dma_chan *chan, enum dma_ctrl_cmd cmd, unsigned long arg) { struct mxs_dma_chan *mxs_chan = to_mxs_dma_chan(chan); + struct mxs_dma_engine *mxs_dma = mxs_chan->mxs_dma; int ret = 0; switch (cmd) { + case DMA_START: + ret = clk_prepare_enable(mxs_dma->clk); + break; case DMA_TERMINATE_ALL: mxs_dma_reset_chan(mxs_chan); mxs_dma_disable_chan(mxs_chan); @@ -610,6 +624,9 @@ static int mxs_dma_control(struct dma_chan *chan, enum dma_ctrl_cmd cmd, case DMA_RESUME: mxs_dma_resume_chan(mxs_chan); break; + case DMA_END: + clk_disable_unprepare(mxs_dma->clk); + break; default: ret = -ENOSYS; } diff --git a/drivers/i2c/busses/i2c-mxs.c b/drivers/i2c/busses/i2c-mxs.c index 1f58197..da1e881 100644 --- a/drivers/i2c/busses/i2c-mxs.c +++ b/drivers/i2c/busses/i2c-mxs.c @@ -643,6 +643,12 @@ static int __devinit mxs_i2c_probe(struct platform_device *pdev) dev_err(dev, "Failed to request dma\n"); return -ENODEV; } + err = dmaengine_device_control(i2c->dmach, DMA_START, 0); + if (err) { + dma_release_channel(i2c->dmach); + dev_err(dev, "Failed to start dma\n"); + return err; + } } platform_set_drvdata(pdev, i2c); @@ -680,8 +686,10 @@ static int __devexit mxs_i2c_remove(struct platform_device *pdev) if (ret) return -EBUSY; - if (i2c->dmach) + if (i2c->dmach) { + dmaengine_device_control(i2c->dmach, DMA_END, 0); dma_release_channel(i2c->dmach); + } writel(MXS_I2C_CTRL0_SFTRST, i2c->regs + MXS_I2C_CTRL0_SET); diff --git a/drivers/mmc/host/mxs-mmc.c b/drivers/mmc/host/mxs-mmc.c index 80d1e6d..aa91830 100644 --- a/drivers/mmc/host/mxs-mmc.c +++ b/drivers/mmc/host/mxs-mmc.c @@ -676,6 +676,9 @@ static int mxs_mmc_probe(struct platform_device *pdev) "%s: failed to request dma\n", __func__); goto out_clk_put; } + ret = dmaengine_device_control(ssp->dmach, DMA_START, 0); + if (ret) + goto out_free_dma; /* set mmc core parameters */ mmc->ops = &mxs_mmc_ops; @@ -717,18 +720,20 @@ static int mxs_mmc_probe(struct platform_device *pdev) ret = devm_request_irq(&pdev->dev, irq_err, mxs_mmc_irq_handler, 0, DRIVER_NAME, host); if (ret) - goto out_free_dma; + goto out_end_free_dma; spin_lock_init(&host->lock); ret = mmc_add_host(mmc); if (ret) - goto out_free_dma; + goto out_end_free_dma; dev_info(mmc_dev(host->mmc), "initialized\n"); return 0; +out_end_free_dma: + dmaengine_device_control(ssp->dmach, DMA_END, 0); out_free_dma: if (ssp->dmach) dma_release_channel(ssp->dmach); @@ -750,8 +755,10 @@ static int mxs_mmc_remove(struct platform_device *pdev) platform_set_drvdata(pdev, NULL); - if (ssp->dmach) + if (ssp->dmach) { + dmaengine_device_control(ssp->dmach, DMA_END, 0); dma_release_channel(ssp->dmach); + } clk_disable_unprepare(ssp->clk); clk_put(ssp->clk); @@ -772,6 +779,7 @@ static int mxs_mmc_suspend(struct device *dev) ret = mmc_suspend_host(mmc); clk_disable_unprepare(ssp->clk); + dmaengine_device_control(ssp->dmach, DMA_END, 0); return ret; } @@ -784,8 +792,13 @@ static int mxs_mmc_resume(struct device *dev) int ret = 0; clk_prepare_enable(ssp->clk); + ret = dmaengine_device_control(ssp->dmach, DMA_START, 0); + if (ret) + return ret; ret = mmc_resume_host(mmc); + if (ret) + dmaengine_device_control(ssp->dmach, DMA_END, 0); return ret; } diff --git a/drivers/mtd/nand/gpmi-nand/gpmi-lib.c b/drivers/mtd/nand/gpmi-nand/gpmi-lib.c index 3502acc..20ed3f3 100644 --- a/drivers/mtd/nand/gpmi-nand/gpmi-lib.c +++ b/drivers/mtd/nand/gpmi-nand/gpmi-lib.c @@ -958,6 +958,7 @@ void gpmi_begin(struct gpmi_nand_data *this) uint32_t reg; unsigned int dll_wait_time_in_us; struct gpmi_nfc_hardware_timing hw; + struct dma_chan *channel = get_dma_chan(this); int ret; /* Enable the clock. */ @@ -967,6 +968,12 @@ void gpmi_begin(struct gpmi_nand_data *this) goto err_out; } + ret = dmaengine_device_control(channel, DMA_START, 0); + if (ret) { + gpmi_disable_clk(this); + goto err_out; + } + /* Only initialize the timing once */ if (this->flags & GPMI_TIMING_INIT_OK) return; @@ -1035,7 +1042,10 @@ err_out: void gpmi_end(struct gpmi_nand_data *this) { + struct dma_chan *channel = get_dma_chan(this); + gpmi_disable_clk(this); + dmaengine_device_control(channel, DMA_END, 0); } /* Clears a BCH interrupt. */ diff --git a/drivers/mtd/nand/gpmi-nand/gpmi-nand.c b/drivers/mtd/nand/gpmi-nand/gpmi-nand.c index e2c56fc..5694d03 100644 --- a/drivers/mtd/nand/gpmi-nand/gpmi-nand.c +++ b/drivers/mtd/nand/gpmi-nand/gpmi-nand.c @@ -815,12 +815,14 @@ static void gpmi_select_chip(struct mtd_info *mtd, int chipnr) struct nand_chip *chip = mtd->priv; struct gpmi_nand_data *this = chip->priv; - if ((this->current_chip < 0) && (chipnr >= 0)) + if ((this->current_chip < 0) && (chipnr >= 0)) { + /* set the current_chip before we call gpmi_begin(). */ + this->current_chip = chipnr; gpmi_begin(this); - else if ((this->current_chip >= 0) && (chipnr < 0)) + } else if ((this->current_chip >= 0) && (chipnr < 0)) { gpmi_end(this); - - this->current_chip = chipnr; + this->current_chip = chipnr; + } } static void gpmi_read_buf(struct mtd_info *mtd, uint8_t *buf, int len) diff --git a/include/linux/dmaengine.h b/include/linux/dmaengine.h index d3201e4..79f864a 100644 --- a/include/linux/dmaengine.h +++ b/include/linux/dmaengine.h @@ -199,6 +199,8 @@ enum dma_ctrl_flags { /** * enum dma_ctrl_cmd - DMA operations that can optionally be exercised * on a running channel. + * @DMA_START: do some preprations to start the DMA engine, such as enable the + * necessary clocks. * @DMA_TERMINATE_ALL: terminate all ongoing transfers * @DMA_PAUSE: pause ongoing transfers * @DMA_RESUME: resume paused transfer @@ -209,13 +211,17 @@ enum dma_ctrl_flags { * command. * @FSLDMA_EXTERNAL_START: this command will put the Freescale DMA controller * into external start mode. + * @DMA_END: do some works to end the DMA engine, such as disable the + * necessary clocks. */ enum dma_ctrl_cmd { + DMA_START, DMA_TERMINATE_ALL, DMA_PAUSE, DMA_RESUME, DMA_SLAVE_CONFIG, FSLDMA_EXTERNAL_START, + DMA_END, }; /** diff --git a/sound/soc/mxs/mxs-pcm.c b/sound/soc/mxs/mxs-pcm.c index f82d766..cfcc30f 100644 --- a/sound/soc/mxs/mxs-pcm.c +++ b/sound/soc/mxs/mxs-pcm.c @@ -92,6 +92,7 @@ static int snd_mxs_open(struct snd_pcm_substream *substream) { struct snd_soc_pcm_runtime *rtd = substream->private_data; struct mxs_pcm_dma_data *pcm_dma_data; + struct dma_chan *chan; int ret; pcm_dma_data = kzalloc(sizeof(*pcm_dma_data), GFP_KERNEL); @@ -107,6 +108,14 @@ static int snd_mxs_open(struct snd_pcm_substream *substream) return ret; } + chan = snd_dmaengine_pcm_get_chan(substream); + ret = dmaengine_device_control(chan, DMA_START, (unsigned long)0); + if (ret) { + snd_dmaengine_pcm_close(substream); + kfree(pcm_dma_data); + return ret; + } + snd_soc_set_runtime_hwparams(substream, &snd_mxs_hardware); snd_dmaengine_pcm_set_data(substream, pcm_dma_data); @@ -117,7 +126,10 @@ static int snd_mxs_open(struct snd_pcm_substream *substream) static int snd_mxs_close(struct snd_pcm_substream *substream) { struct mxs_pcm_dma_data *pcm_dma_data = snd_dmaengine_pcm_get_data(substream); + struct dma_chan *chan; + chan = snd_dmaengine_pcm_get_chan(substream); + dmaengine_device_control(chan, DMA_END, 0); snd_dmaengine_pcm_close(substream); kfree(pcm_dma_data); -- 1.7.0.4

5 13

[alsa-devel] Non-Interleaved access method?
by Prasant J 18 Oct '12

18 Oct '12

Hi, When the application set the access method to non-interleaved, that means that the ALSA application will deliver the non-interleaved buffers to alsa subsystem & the alsa subsystem will interleave the data before sending out the data to the hardware. Am I correct? Or non-interleaved means that the data will be sent out in non-interleaved manner itself? Regards, -Prasant

2 1

[alsa-devel] ASoC: mmap or read/write?
by Prasant J 18 Oct '12

18 Oct '12

Hi, I'm writing an alsa application for recording & playing 16 channel audio. The application is intended to run on an embedded platform & the resources are highly constrained. My application is not performing well with read/write methods. I'm considering using mmap but not able to get the right documentation or start, so I want to know whether mmap really offers any advantages or not. Does the mmap method has any advantages over read/write? Is there any thumb rule to decide when to use mmap or read/write? -Prasant

2 3

[alsa-devel] Interleaved or Non-interleaved?
by Prasant J 18 Oct '12

18 Oct '12

hi, My sound hardware supports only interleaved audio. I'm writing an alsa application. Which access method will be more performing : interleaved or non-interleaved? (Or it does not make a difference?) -Prasant

2 2