linux.git/drivers/dma/ti, branch v6.7

dmaengine: ti: k3-psil-am62a: Fix SPI PDMA data

2023-11-24T08:44:00+00:00

AM62Ax has 3 SPI channels where each channel has 4x TX and 4x RX
threads. Also fix the thread numbers to match what the firmware expects
according to the PSI-L device description.

Link: http://downloads.ti.com/tisci/esd/latest/5_soc_doc/am62ax/psil_cfg.html [1]
Fixes: aac6db7e243a ("dmaengine: ti: k3-psil-am62a: Add AM62Ax PSIL and PDMA data")
Signed-off-by: Jai Luthra 
Link: https://lore.kernel.org/r/20231123-psil_fix-v1-1-6604d80819be@ti.com
Signed-off-by: Vinod Koul

dmaengine: ti: k3-psil-am62: Fix SPI PDMA data

2023-11-23T08:45:53+00:00

AM62x has 3 SPI channels where each channel has 4 TX and 4 RX threads.
This also fixes the thread numbers.

Signed-off-by: Ronald Wahl 
Fixes: 5ac6bfb58777 ("dmaengine: ti: k3-psil: Add AM62x PSIL and PDMA data")
Reviewed-by: Jai Luthra 
Acked-by: Peter Ujfalusi 
Link: https://lore.kernel.org/r/20231030190113.16782-1-rwahl@gmx.de
Signed-off-by: Vinod Koul

Merge tag 'dmaengine-6.7-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/vkoul/dmaengine

2023-11-04T04:56:51+00:00

Pull dmaengine updates from Vinod Koul:

 - Big pile of __counted_by attribute annotations to several structures
   for bounds checking of flexible arrays at run-time

 - Another big pile platform remove callback returning void changes

 - Device tree device_get_match_data() usage and dropping
   of_match_device() calls

 - Minor driver updates to pxa, idxd fsl, hisi etc drivers

* tag 'dmaengine-6.7-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/vkoul/dmaengine: (106 commits)
  dmaengine: stm32-mdma: correct desc prep when channel running
  dmaengine: dw-axi-dmac: Add support DMAX_NUM_CHANNELS > 16
  dmaengine: xilinx: xilinx_dma: Fix kernel doc about xilinx_dma_remove()
  dmaengine: mmp_tdma: drop unused variable 'of_id'
  MAINTAINERS: Add entries for NXP(Freescale) eDMA drivers
  dmaengine: xilinx: xdma: Support cyclic transfers
  dmaengine: xilinx: xdma: Prepare the introduction of cyclic transfers
  dmaengine: Drop unnecessary of_match_device() calls
  dmaengine: Use device_get_match_data()
  dmaengine: pxa_dma: Annotate struct pxad_desc_sw with __counted_by
  dmaengine: pxa_dma: Remove an erroneous BUG_ON() in pxad_free_desc()
  dmaengine: xilinx: xdma: Use resource_size() in xdma_probe()
  dmaengine: fsl-dpaa2-qdma: Remove redundant initialization owner in dpaa2_qdma_driver
  dmaengine: Remove unused declaration dma_chan_cleanup()
  dmaengine: mmp: fix Wvoid-pointer-to-enum-cast warning
  dmaengine: qcom: fix Wvoid-pointer-to-enum-cast warning
  dmaengine: fsl-edma: Remove redundant dev_err() for platform_get_irq()
  dmaengine: ep93xx_dma: Annotate struct ep93xx_dma_engine with __counted_by
  dmaengine: idxd: add wq driver name support for accel-config user tool
  dmaengine: fsl-edma: Annotate struct struct fsl_edma_engine with __counted_by
  ...

dmaengine: ti: k3-udma-glue: clean up k3_udma_glue_tx_get_irq() return

2023-10-04T09:29:55+00:00

The k3_udma_glue_tx_get_irq() function currently returns negative error
codes on error, zero on error and positive values for success.  This
complicates life for the callers who need to propagate the error code.
Also GCC will not warn about unsigned comparisons when you check:

	if (unsigned_irq <= 0)

All the callers have been fixed now but let's just make this easy going
forward.

Signed-off-by: Dan Carpenter 
Reviewed-by: Roger Quadros 
Acked-by: Vinod Koul 
Signed-off-by: David S. Miller

dmaengine: ti: edma: handle irq_of_parse_and_map() errors

2023-10-02T15:58:33+00:00

Zero is not a valid IRQ for in-kernel code and the irq_of_parse_and_map()
function returns zero on error.  So this check for valid IRQs should only
accept values > 0.

Fixes: 2b6b3b742019 ("ARM/dmaengine: edma: Merge the two drivers under drivers/dma/")
Signed-off-by: Dan Carpenter 
Acked-by: Peter Ujfalusi 
Link: https://lore.kernel.org/r/f15cb6a7-8449-4f79-98b6-34072f04edbc@moroto.mountain
Signed-off-by: Vinod Koul

dmaengine: ti: omap-dma: Annotate struct omap_desc with __counted_by

2023-09-28T11:12:15+00:00

Prepare for the coming implementation by GCC and Clang of the __counted_by
attribute. Flexible array members annotated with __counted_by can have
their accesses bounds-checked at run-time checking via CONFIG_UBSAN_BOUNDS
(for array indexing) and CONFIG_FORTIFY_SOURCE (for strcpy/memcpy-family
functions).

As found with Coccinelle[1], add __counted_by for struct omap_desc.
Additionally, since the element count member must be set before accessing
the annotated flexible array member, move its initialization earlier.

[1] https://github.com/kees/kernel-tools/blob/trunk/coccinelle/examples/counted_by.cocci

Cc: Vinod Koul 
Cc: Peter Ujfalusi 
Cc: dmaengine@vger.kernel.org
Signed-off-by: Kees Cook 
Reviewed-by: "Gustavo A. R. Silva" 
Link: https://lore.kernel.org/r/20230817235859.49846-18-keescook@chromium.org
Signed-off-by: Vinod Koul

dmaengine: ti: edma: Annotate struct edma_desc with __counted_by

2023-09-28T11:12:15+00:00

Prepare for the coming implementation by GCC and Clang of the __counted_by
attribute. Flexible array members annotated with __counted_by can have
their accesses bounds-checked at run-time checking via CONFIG_UBSAN_BOUNDS
(for array indexing) and CONFIG_FORTIFY_SOURCE (for strcpy/memcpy-family
functions).

As found with Coccinelle[1], add __counted_by for struct edma_desc.

[1] https://github.com/kees/kernel-tools/blob/trunk/coccinelle/examples/counted_by.cocci

Cc: Vinod Koul 
Cc: Peter Ujfalusi 
Cc: dmaengine@vger.kernel.org
Signed-off-by: Kees Cook 
Reviewed-by: "Gustavo A. R. Silva" 
Link: https://lore.kernel.org/r/20230817235859.49846-17-keescook@chromium.org
Signed-off-by: Vinod Koul

dmaengine: ti: omap-dma: Convert to platform remove callback returning void

2023-09-28T07:40:57+00:00

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.
To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new() which already returns void. Eventually after all drivers
are converted, .remove_new() is renamed to .remove().

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König 
Acked-by: Peter Ujfalusi 
Link: https://lore.kernel.org/r/20230919133207.1400430-53-u.kleine-koenig@pengutronix.de
Signed-off-by: Vinod Koul

dmaengine: ti: edma: Convert to platform remove callback returning void

2023-09-28T07:40:57+00:00

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.
To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new() which already returns void. Eventually after all drivers
are converted, .remove_new() is renamed to .remove().

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König 
Acked-by: Peter Ujfalusi 
Link: https://lore.kernel.org/r/20230919133207.1400430-52-u.kleine-koenig@pengutronix.de
Signed-off-by: Vinod Koul

dmaengine: ti: cppi41: Convert to platform remove callback returning void

2023-09-28T07:40:57+00:00

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.
To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new() which already returns void. Eventually after all drivers
are converted, .remove_new() is renamed to .remove().

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König 
Reviewed-by: Peter Ujfalusi 
Link: https://lore.kernel.org/r/20230919133207.1400430-51-u.kleine-koenig@pengutronix.de
Signed-off-by: Vinod Koul