The Renesas RZN1 DMA IP is based on a DW core, with eg. an additional
dmamux register located in the system control area which can take up to
32 requests (16 per DMA controller). Each DMA channel can be wired to
two different peripherals.
Signed-off-by: NMiquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: NGeert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: NRob Herring <robh@kernel.org>
Acked-by: NVinod Koul <vkoul@kernel.org>
Link: https://lore.kernel.org/r/20220427095653.91804-2-miquel.raynal@bootlin.comSigned-off-by: NVinod Koul <vkoul@kernel.org>

Calling synchronize_irq() right before free_irq() is quite useless. On one
hand the IRQ can easily fire again before free_irq() is entered, on the
other hand free_irq() itself calls synchronize_irq() internally (in a race
condition free way), before any state associated with the IRQ is freed.
Signed-off-by: NMinghao Chi <chi.minghao@zte.com.cn>
Link: https://lore.kernel.org/r/20220516115412.1651772-1-chi.minghao@zte.com.cnAcked-by: NDave Jiang <dave.jiang@intel.com>
Signed-off-by: NVinod Koul <vkoul@kernel.org>

When setting DMA_INTERRUPT capability, a callback function
dma->device_prep_dma_interrupt() is needed to support this capability.
Without setting the callback, dma_async_device_register() will fail dma
capability check.

Fixes: 4e5a4eb2 ("dmaengine: idxd: set DMA_INTERRUPT cap bit")
Signed-off-by: NDave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/165101232637.3951447.15765792791591763119.stgit@djiang5-desk3.ch.intel.comSigned-off-by: NVinod Koul <vkoul@kernel.org>

The generic property, used in most of the drivers and defined in generic
dma-common DT bindings, is 'dma-channels'.  Switch to new property while
keeping backward compatibility.
Signed-off-by: NKrzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Reviewed-by: NBaolin Wang <baolin.wang7@gmail.com>
Link: https://lore.kernel.org/r/20220503065147.51728-3-krzysztof.kozlowski@linaro.orgSigned-off-by: NVinod Koul <vkoul@kernel.org>

The generic property, used in most of the drivers and defined in generic
dma-common DT bindings, is 'dma-channels'.
Signed-off-by: NKrzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Reviewed-by: NRob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20220503065147.51728-2-krzysztof.kozlowski@linaro.orgSigned-off-by: NVinod Koul <vkoul@kernel.org>

Skip wq irq resources freeing when wq type is not kernel since the driver
skips the irq alloction during wq enable. Add check in wq type check in
idxd_wq_free_irq() to mirror idxd_wq_request_irq().

Fixes: 63c14ae6 ("dmaengine: idxd: refactor wq driver enable/disable operations")
Reported-by: NTony Zu <tony.zhu@intel.com>
Tested-by: NTony Zu <tony.zhu@intel.com>
Signed-off-by: NDave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/165176310726.2112428.7474366910758522079.stgit@djiang5-desk3.ch.intel.comSigned-off-by: NVinod Koul <vkoul@kernel.org>

Since idxd_register/unregister_dma_channel() are only called locally, make
them static.
Reported-by: Nkernel test robot <lkp@intel.com>
Signed-off-by: NDave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/165187583222.3287435.12882651040433040246.stgit@djiang5-desk3.ch.intel.comSigned-off-by: NVinod Koul <vkoul@kernel.org>

pm_runtime_resume_and_get() automatically handle dev->power.usage_count
decrement on errors, so prefer using it and also implement it's error
handling.
Signed-off-by: NRadhey Shyam Pandey <radhey.shyam.pandey@xilinx.com>
Link: https://lore.kernel.org/r/1652166762-18317-4-git-send-email-radhey.shyam.pandey@xilinx.comSigned-off-by: NVinod Koul <vkoul@kernel.org>

Add condition to check the return value of dma_async_device_register
and implement its error handling.

Addresses-Coverity: Event check_return.
Signed-off-by: NShravya Kumbham <shravya.kumbham@xilinx.com>
Signed-off-by: NHarini Katakam <harini.katakam@xilinx.com>
Signed-off-by: NRadhey Shyam Pandey <radhey.shyam.pandey@xilinx.com>
Link: https://lore.kernel.org/r/1652166762-18317-3-git-send-email-radhey.shyam.pandey@xilinx.comSigned-off-by: NVinod Koul <vkoul@kernel.org>

In zynqmp_dma_alloc/free_chan_resources functions there is a
potential overflow in the below expressions.

dma_alloc_coherent(chan->dev, (2 * chan->desc_size *
		   ZYNQMP_DMA_NUM_DESCS),
		   &chan->desc_pool_p, GFP_KERNEL);

dma_free_coherent(chan->dev,(2 * ZYNQMP_DMA_DESC_SIZE(chan) *
                 ZYNQMP_DMA_NUM_DESCS),
                chan->desc_pool_v, chan->desc_pool_p);

The arguments desc_size and ZYNQMP_DMA_NUM_DESCS were 32 bit. Though
this overflow condition is not observed but it is a potential problem
in the case of 32-bit multiplication. Hence fix it by changing the
desc_size data type to size_t.

In addition to coverity fix it also reuse ZYNQMP_DMA_DESC_SIZE macro in
dma_alloc_coherent API argument.

Addresses-Coverity: Event overflow_before_widen.
Signed-off-by: NRadhey Shyam Pandey <radhey.shyam.pandey@xilinx.com>
Link: https://lore.kernel.org/r/1652166762-18317-2-git-send-email-radhey.shyam.pandey@xilinx.comSigned-off-by: NVinod Koul <vkoul@kernel.org>

idxd_wq_device_reset_cleanup() already calls idxd_wq_disable_cleanup().
There is no need to call idxd_wq_disable_cleanup() again in
idxd_device_wqs_clear_state(). Remove redudant call from
idxd_wq_device_reset_cleanup().

Fixes: 0dcfe41e ("dmanegine: idxd: cleanup all device related bits after disabling device")
Signed-off-by: NDave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/165231365717.986350.2441351765955825964.stgit@djiang5-desk3.ch.intel.comSigned-off-by: NVinod Koul <vkoul@kernel.org>

Call idxd_wq_free_irq() in the drv_disable_wq() function before
idxd_wq_reset() is called. Otherwise the wq type is reset and the irq does
not get freed.

Fixes: 63c14ae6 ("dmaengine: idxd: refactor wq driver enable/disable operations")
Signed-off-by: NDave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/165231367316.986407.11001767338124941736.stgit@djiang5-desk3.ch.intel.comSigned-off-by: NVinod Koul <vkoul@kernel.org>

Jacob reported that with lockdep debug turned on, idxd_device_driver
removal causes kernel splat from lock assert warning for
idxd_device_wqs_clear_state(). Make sure
idxd_device_wqs_clear_state() holds the wq lock for each wq when
cleaning the wq state. Move the call outside of the device spinlock.
Reported-by: NJacob Pan <jacob.jun.pan@intel.com>
Signed-off-by: NDave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/165231364426.986304.9294302800482492780.stgit@djiang5-desk3.ch.intel.comSigned-off-by: NVinod Koul <vkoul@kernel.org>

The idxd driver always gated the pasid enabling under a single knob and
this assumption is incorrect. The pasid used for kernel operation can be
independently toggled and has no dependency on the user pasid (and vice
versa). Split the two so they are independent "enabled" flags.
Signed-off-by: NDave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/165231431746.986466.5666862038354800551.stgit@djiang5-desk3.ch.intel.comSigned-off-by: NVinod Koul <vkoul@kernel.org>

Despite the name, R-Car V3U is the first member of the R-Car Gen4
family.  Hence move its compatible value to the R-Car Gen4 section.
Signed-off-by: NGeert Uytterhoeven <geert+renesas@glider.be>
Acked-by: NKrzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/r/e6e4cf701f3a43b061b9c3f7f0adc4d6addd4722.1651497024.git.geert+renesas@glider.beSigned-off-by: NVinod Koul <vkoul@kernel.org>

If a call to alloc_chrdev_region() fails, the already allocated resources
are leaking.

Add the needed error handling path to fix the leak.

Fixes: 42d279f9 ("dmaengine: idxd: add char driver to expose submission portal to userland")
Signed-off-by: NChristophe JAILLET <christophe.jaillet@wanadoo.fr>
Acked-by: NDave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/1b5033dcc87b5f2a953c413f0306e883e6114542.1650521591.git.christophe.jaillet@wanadoo.frSigned-off-by: NVinod Koul <vkoul@kernel.org>

Use platform_irq_get() instead platform_get_resource() for IRQ resource
to fix the probe failure. platform_get_resource() fails to fetch the IRQ
resource as it might not be ready at that time.

platform_irq_get() is also the recommended way to get interrupt as it
directly gives the IRQ number and no conversion from resource is
required.

Fixes: ee170280 ("dmaengine: tegra: Add tegra gpcdma driver")
Reported-by: NJonathan Hunter <jonathanh@nvidia.com>
Signed-off-by: NAkhil R <akhilrajeev@nvidia.com>
Acked-by: NThierry Reding <treding@nvidia.com>
Reviewed-by: NJon Hunter <jonathanh@nvidia.com>
Link: https://lore.kernel.org/r/20220505091440.12981-1-akhilrajeev@nvidia.comSigned-off-by: NVinod Koul <vkoul@kernel.org>

It is unnecessary to call spin_lock_bh() for that you are already
in a tasklet.
Signed-off-by: NYunbo Yu <yuyunbo519@gmail.com>
Link: https://lore.kernel.org/r/20220420122754.148359-1-yuyunbo519@gmail.comSigned-off-by: NVinod Koul <vkoul@kernel.org>

Move the core driver operations from wq driver to the drv_enable_wq() and
drv_disable_wq() functions. The move should reduce the wq driver's
knowledge of the core driver operations and prevent code confusion for
future wq drivers.
Signed-off-by: NDave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/165047301643.3841827.11222723219862233060.stgit@djiang5-desk3.ch.intel.comSigned-off-by: NVinod Koul <vkoul@kernel.org>

Correct the RX PSIL thread for sa3ul.
Signed-off-by: NJayesh Choudhary <j-choudhary@ti.com>
Fixes: 5ac6bfb5 ("dmaengine: ti: k3-psil: Add AM62x PSIL and PDMA data")
Link: https://lore.kernel.org/r/20220421065323.16378-1-j-choudhary@ti.comSigned-off-by: NVinod Koul <vkoul@kernel.org>

LKP bot reports a new warning:
Warning:
drivers/dma/ptdma/ptdma-dmaengine.c:262:1: warning: no previous prototype for 'pt_tx_status' [-Wmissing-prototypes]

pt_tx_status() should be static, so declare as such.
Reported-by: Nkernel test robot <lkp@intel.com>
Fixes: d9650682 ("dmaengine: PTDMA: support polled mode")
Link: https://lore.kernel.org/r/20220421052407.745637-1-vkoul@kernel.orgSigned-off-by: NVinod Koul <vkoul@kernel.org>

The value parameter is a single byte, so duplicate it to the 8 byte
range that is used as the pattern.
Signed-off-by: NBen Walker <benjamin.walker@intel.com>
Cc: Sinan Kaya <okaya@kernel.org>
Link: https://lore.kernel.org/r/20220301182551.883474-5-benjamin.walker@intel.comSigned-off-by: NVinod Koul <vkoul@kernel.org>

The value passed in to .prep_dma_memset is to be treated as a single
byte repeating pattern.
Signed-off-by: NBen Walker <benjamin.walker@intel.com>
Cc: Ludovic Desroches <ludovic.desroches@microchip.com>
Cc: Tudor Ambarus <tudor.ambarus@microchip.com>
Link: https://lore.kernel.org/r/20220301182551.883474-4-benjamin.walker@intel.comSigned-off-by: NVinod Koul <vkoul@kernel.org>

The value passed in to .prep_dma_memset is to be treated as a single
byte repeating pattern.
Signed-off-by: NBen Walker <benjamin.walker@intel.com>
Cc: Ludovic Desroches <ludovic.desroches@microchip.com>
Cc: Tudor Ambarus <tudor.ambarus@microchip.com>
Link: https://lore.kernel.org/r/20220301182551.883474-3-benjamin.walker@intel.comSigned-off-by: NVinod Koul <vkoul@kernel.org>

Document this function to make clear the expected behavior of the
'value' parameter. It was intended to match the behavior of POSIX memset
as laid out here:

https://lore.kernel.org/dmaengine/YejrA5ZWZ3lTRO%2F1@matsya/Signed-off-by: NBen Walker <benjamin.walker@intel.com>
Link: https://lore.kernel.org/r/20220301182551.883474-2-benjamin.walker@intel.comSigned-off-by: NVinod Koul <vkoul@kernel.org>

Move the calling of request_irq() and other related irq setup code until
after the WQ is successfully enabled. This reduces the amount of
setup/teardown if the wq is not configured correctly and cannot be enabled.
Signed-off-by: NDave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/164642777730.179702.1880317757087484299.stgit@djiang5-desk3.ch.intel.comSigned-off-by: NVinod Koul <vkoul@kernel.org>

Eliminate the follow versioncheck warning:

./drivers/dma/tegra186-gpc-dma.c: 21 linux/version.h not needed.
Reported-by: NAbaci Robot <abaci@linux.alibaba.com>
Signed-off-by: NJiapeng Chong <jiapeng.chong@linux.alibaba.com>
Link: https://lore.kernel.org/r/20220413083842.69845-1-jiapeng.chong@linux.alibaba.comSigned-off-by: NVinod Koul <vkoul@kernel.org>

Looks like I forgot to add DMA_INTERRUPT cap setting to the idxd driver and
dmatest is still working regardless of this mistake. Add an explicit check
of DMA_INTERRUPT capability for dmatest to make sure the DMA device being used
actually supports interrupt before the test is launched and also that the
driver is programmed correctly.
Signed-off-by: NDave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/164978679251.2361020.5856734256126725993.stgit@djiang5-desk3.ch.intel.comSigned-off-by: NVinod Koul <vkoul@kernel.org>

This fixes the following sparse warnings:
drivers/dma/mediatek/mtk-hsdma.c:604:26: warning: Using plain integer
as NULL pointer
Signed-off-by: NHaowen Bai <baihaowen@meizu.com>
Link: https://lore.kernel.org/r/1649750340-30777-1-git-send-email-baihaowen@meizu.comSigned-off-by: NVinod Koul <vkoul@kernel.org>

Unneeded variable: "retval". Return "NULL" , so we have to make code clear.
better way, drop the function.
Signed-off-by: NHaowen Bai <baihaowen@meizu.com>
Link: https://lore.kernel.org/r/1649726180-13133-1-git-send-email-baihaowen@meizu.comSigned-off-by: NVinod Koul <vkoul@kernel.org>

Load the max_xfer_size and max_batch_size values from the values read from
registers to the shadow variables. This will allow the read-only device to
display the correct values for the sysfs attributes.
Signed-off-by: NDave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/164971507673.2201761.11244446608988838897.stgit@djiang5-desk3.ch.intel.comSigned-off-by: NVinod Koul <vkoul@kernel.org>

Even though idxd driver has always supported interrupt, it never actually
set the DMA_INTERRUPT cap bit. Rectify this mistake so the interrupt
capability is advertised.
Reported-by: NBen Walker <benjamin.walker@intel.com>
Signed-off-by: NDave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/164971497859.2201379.17925303210723708961.stgit@djiang5-desk3.ch.intel.comSigned-off-by: NVinod Koul <vkoul@kernel.org>

Set the maximum segment size, since the hardware can do transfers larger
than the default 64 KiB returned by dma_get_max_seg_size().

The maximum segment size is limited by the 24-bit transfer count field
in DMA descriptors. The number of bytes is equal to the transfer count
times the transfer size unit, which is selected by the driver based on
the DMA buffer address and length of the transfer. The size unit can be
as small as 1 byte, so set the maximum segment size to 2^24-1 bytes to
ensure the transfer count will not overflow regardless of the size unit
selected by the driver.
Signed-off-by: NAidan MacDonald <aidanmacdonald.0x0@gmail.com>
Link: https://lore.kernel.org/r/20220411153618.49876-1-aidanmacdonald.0x0@gmail.comSigned-off-by: NVinod Koul <vkoul@kernel.org>

If the DMA_PREP_INTERRUPT flag is not provided, run in polled mode,
which significantly improves IOPS: more than twice on chunks < 4K.
Signed-off-by: NIlya Novikov <i.m.novikov@yadro.com>
Link: https://lore.kernel.org/r/20220413113733.59041-1-i.m.novikov@yadro.comSigned-off-by: NVinod Koul <vkoul@kernel.org>

Document the compatible for GPI DMA controller on SC7280 SoC
Signed-off-by: NVinod Koul <vkoul@kernel.org>
Acked-by: NKrzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/r/20220414064216.1182177-1-vkoul@kernel.orgSigned-off-by: NVinod Koul <vkoul@kernel.org>

It is unnecessary to call spin_lock_bh() if you are already in a tasklet.
Signed-off-by: NYunbo Yu <yuyunbo519@gmail.com>
Reviewed-by: NLogan Gunthorpe <logang@deltatee.com>
Link: https://lore.kernel.org/r/20220418142021.1241558-1-yuyunbo519@gmail.comSigned-off-by: NVinod Koul <vkoul@kernel.org>

MCDMA IP provides up to 16 multiple channels of data movement each on
MM2S and S2MM paths. Inline with implementation, in the binding add
description for the channel ID start index and mention that it's fixed
irrespective of the MCDMA IP configuration(number of read/write channels).
Signed-off-by: NRadhey Shyam Pandey <radhey.shyam.pandey@xilinx.com>
Acked-by: NRob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/1649939061-6675-1-git-send-email-radhey.shyam.pandey@xilinx.comSigned-off-by: NVinod Koul <vkoul@kernel.org>

The Qualcomm SM8350 platform does, like the SM8450, provide a set of GPI
controllers with an ee-offset of 0x10000. Add this to the driver.
Signed-off-by: NBjorn Andersson <bjorn.andersson@linaro.org>
Link: https://lore.kernel.org/r/20220412212959.2385085-1-bjorn.andersson@linaro.orgSigned-off-by: NVinod Koul <vkoul@kernel.org>

Controller on newer SoCs like SM8450 have registers at at offset. Add
ee_offset to driver_data and add this compatible for the driver.
Signed-off-by: NVinod Koul <vkoul@kernel.org>
Reviewed-by: NBjorn Andersson <bjorn.andersson@linaro.org>
Tested-by: NBjorn Andersson <bjorn.andersson@linaro.org>
Link: https://lore.kernel.org/r/20220406132508.1029348-3-vkoul@kernel.orgSigned-off-by: NVinod Koul <vkoul@kernel.org>

Add the compatible for newer qcom socs with gpi dma i.e qcom sm8350 and
sm8450.
Signed-off-by: NVinod Koul <vkoul@kernel.org>
Reviewed-by: NBjorn Andersson <bjorn.andersson@linaro.org>
Acked-by: NKrzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/r/20220406132508.1029348-2-vkoul@kernel.orgSigned-off-by: NVinod Koul <vkoul@kernel.org>