linux/drivers/dma, branch v3.16

Update imx-sdma cyclic handling to report residue

2014-07-01T06:53:42Z

I received a report this morning from one of the Novena developers that the behaviour of the iMX6 ASoC codec driver (using imx-pcm-dma.c) was sub-optimal under high system load. While there are issues relating to system load remaining, upon reviewing the ASoC imx-pcm-dma.c driver, it was noticed that it not using the residue support, because SDMA doesn't support it. This has the effect that SDMA has to make multiple calls into the ASoC and ALSA code, one for each period. Since ALSA's snd_pcm_elapsed() does not need to be called multiple times and it is entirely sufficient to call it once to update ALSA with the current buffer position via the pointer method, we can do better here. We can also avoid stopping the DMA entirely, just like real cyclic DMA implementations behave. While this means that we replay some old samples, this is a nicer behaviour than having audio stop and restart. The changes to achieve this are relatively minor - imx-sdma.c can track where the DMA is to the nearest descriptor boundary - it does this already when deciding how many callbacks to issue. In doing this, buf_tail always points at the descriptor which will complete next. The residue is defined by the bytes remaining to the end of the buffer, when the buffer is viewed as a single block of memory [start...end]. So, when we start out, there's a full buffer worth of residue, and this counts down as we approach the end of the buffer, eventually becoming zero at the end, before returning to the full buffer worth when we wrap back to the start. Moving the walking of the descriptors into the interrupt handler means that we can update the BD_DONE flag at interrupt time, thus avoiding a delayed tasklet stopping the cyclic DMA. This means that the residue can be calculated from (total descriptors - buf_tail) * descriptor size. This is what the change below does. We update imx-pcm-dma.c to remove the NO_RESIDUE flag since we now provide the residue. Signed-off-by: Russell King Tested-by: Shawn Guo Signed-off-by: Vinod Koul

dma: cppi41: handle 0-length packets

2014-07-01T06:45:48Z

When a 0-length packet is received on the bus, desc->pd0 yields 1, which confuses the driver's users. This information is clearly wrong and not in accordance to the datasheet, but it's been observed on an AM335x board, very reproducible. Fix this by looking at bit 19 in PD2 of the completed packet. This bit will tell us if a zero-length packet was received on a queue. If it's set, ignore the value in PD0 and report a total length of 0 instead. Signed-off-by: Daniel Mack Signed-off-by: Vinod Koul

Merge branch 'for-linus' of git://git.infradead.org/users/vkoul/slave-dma

2014-06-10T17:28:45Z

Pull slave-dmaengine updates from Vinod Koul: - new Xilixn VDMA driver from Srikanth - bunch of updates for edma driver by Thomas, Joel and Peter - fixes and updates on dw, ste_dma, freescale, mpc512x, sudmac etc * 'for-linus' of git://git.infradead.org/users/vkoul/slave-dma: (45 commits) dmaengine: sh: don't use dynamic static allocation dmaengine: sh: fix print specifier warnings dmaengine: sh: make shdma_prep_dma_cyclic static dmaengine: Kconfig: Update MXS_DMA help text to include MX6Q/MX6DL of: dma: Grammar s/requests/request/, s/used required/required/ dmaengine: shdma: Enable driver compilation with COMPILE_TEST dmaengine: rcar-hpbdma: Include linux/err.h dmaengine: sudmac: Include linux/err.h dmaengine: sudmac: Keep #include sorted alphabetically dmaengine: shdmac: Include linux/err.h dmaengine: shdmac: Keep #include sorted alphabetically dmaengine: s3c24xx-dma: Add cyclic transfer support dmaengine: s3c24xx-dma: Process whole SG chain dmaengine: imx: correct sdmac->status for cyclic dma tx dmaengine: pch: fix compilation for alpha target dmaengine: dw: check return code of dma_async_device_register() dmaengine: dw: fix regression in dw_probe() function dmaengine: dw: enable clock before access dma: pch_dma: Fix Kconfig dependencies dmaengine: mpc512x: add support for peripheral transfers ...

Merge branch 'topic/xilinx' into for-linus

2014-06-09T16:26:29Z

Merge branch 'topic/dw' into for-linus

2014-06-09T16:25:40Z

dmaengine: sh: don't use dynamic static allocation

2014-06-03T06:05:33Z

dynamic stack allocation in kernel is considered bad as kernel stack is low and we get warns on few archs as reported by kbuild test robot >> drivers/dma/sh/shdma-base.c:671:32: sparse: Variable length array is used. >> drivers/dma/sh/shdma-base.c:701:1: warning: 'shdma_prep_dma_cyclic' uses >> dynamic stack allocation [enabled by default] Fix this by making a static array of 32 which should be sufficient for shdma_prep_dma_cyclic which only user in kernel is audio and 32 periods for audio seems quite sufficient atm Reported-by: kbuild test robot Signed-off-by: Vinod Koul

dmaengine: sh: fix print specifier warnings

2014-06-03T06:05:24Z

As documented in Documentation/printk-formats.txt we should use %zu/%zx specifiers for size_t type variables for the code to compile on different architectures. This is uncovered as COMPILE_TEST has been enabled recently for this driver drivers/dma/sh/shdma-base.c: In function 'shdma_prep_dma_cyclic': >> drivers/dma/sh/shdma-base.c:683:4: warning: format '%d' expects argument of >> type 'int', but argument 4 has type 'size_t' [-Wformat=] __func__, buf_len, period_len, slave_id); >> drivers/dma/sh/shdma-base.c:683:4: warning: format '%d' expects argument of >> type 'int', but argument 5 has type 'size_t' [-Wformat=] Reported-by: kbuild test robot Acked-by: Laurent Pinchart Signed-off-by: Vinod Koul

dmaengine: sh: make shdma_prep_dma_cyclic static

2014-06-03T06:04:47Z

kbuild test robot reports that shdma_prep_dma_cyclic should be static, since symbol is not declared, quick check revails that is the case >> drivers/dma/sh/shdma-base.c:660:32: sparse: symbol 'shdma_prep_dma_cyclic' >> was not declared. Should it be static? Reported-by: kbuild test robot Acked-by: Laurent Pinchart Signed-off-by: Vinod Koul

Merge tag 'drivers-for-3.16' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc into next

2014-06-02T23:35:49Z

Pull ARM SoC driver changes from Olof Johansson: "SoC-near driver changes that we're merging through our tree. Mostly because they depend on other changes we have staged, but in some cases because the driver maintainers preferred that we did it this way. This contains a largeish cleanup series of the omap_l3_noc bus driver, cpuidle rework for Exynos, some reset driver conversions and a long branch of TI EDMA fixes and cleanups, with more to come next release. The TI EDMA cleanups is a shared branch with the dmaengine tree, with a handful of Davinci-specific fixes on top. After discussion at last year's KS (and some more on the mailing lists), we are here adding a drivers/soc directory. The purpose of this is to keep per-vendor shared code that's needed by different drivers but that doesn't fit into the MFD (nor drivers/platform) model. We expect to keep merging contents for this hierarchy through arm-soc so we can keep an eye on what the vendors keep adding here and not making it a free-for-all to shove in crazy stuff" * tag 'drivers-for-3.16' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (101 commits) cpufreq: exynos: Fix driver compilation with ARCH_MULTIPLATFORM tty: serial: msm: Remove direct access to GSBI power: reset: keystone-reset: introduce keystone reset driver Documentation: dt: add bindings for keystone pll control controller Documentation: dt: add bindings for keystone reset driver soc: qcom: fix of_device_id table ARM: EXYNOS: Fix kernel panic when unplugging CPU1 on exynos ARM: EXYNOS: Move the driver to drivers/cpuidle directory ARM: EXYNOS: Cleanup all unneeded headers from cpuidle.c ARM: EXYNOS: Pass the AFTR callback to the platform_data ARM: EXYNOS: Move S5P_CHECK_SLEEP into pm.c ARM: EXYNOS: Move the power sequence call in the cpu_pm notifier ARM: EXYNOS: Move the AFTR state function into pm.c ARM: EXYNOS: Encapsulate the AFTR code into a function ARM: EXYNOS: Disable cpuidle for exynos5440 ARM: EXYNOS: Encapsulate boot vector code into a function for cpuidle ARM: EXYNOS: Pass wakeup mask parameter to function for cpuidle ARM: EXYNOS: Remove ifdef for scu_enable in pm ARM: EXYNOS: Move scu_enable in the cpu_pm notifier ARM: EXYNOS: Use the cpu_pm notifier for pm ...

dmaengine: Kconfig: Update MXS_DMA help text to include MX6Q/MX6DL

2014-06-01T17:46:37Z

The APBX-DMA block is also found on MX6Q/MX6DL chips. Update the help text accordingly. Signed-off-by: Fabio Estevam Signed-off-by: Vinod Koul