linux-stable.git/drivers/dma, branch linux-6.13.y

dmaengine: fsl-edma: free irq correctly in remove path

2025-04-10T12:41:47+00:00

[ Upstream commit fa70c4c3c580c239a0f9e83a14770ab026e8d820 ]

Add fsl_edma->txirq/errirq check to avoid below warning because no
errirq at i.MX9 platform. Otherwise there will be kernel dump:
WARNING: CPU: 0 PID: 11 at kernel/irq/devres.c:144 devm_free_irq+0x74/0x80
Modules linked in:
CPU: 0 UID: 0 PID: 11 Comm: kworker/u8:0 Not tainted 6.12.0-rc7#18
Hardware name: NXP i.MX93 11X11 EVK board (DT)
Workqueue: events_unbound deferred_probe_work_func
pstate: 60400009 (nZCv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--)
pc : devm_free_irq+0x74/0x80
lr : devm_free_irq+0x48/0x80
Call trace:
 devm_free_irq+0x74/0x80 (P)
 devm_free_irq+0x48/0x80 (L)
 fsl_edma_remove+0xc4/0xc8
 platform_remove+0x28/0x44
 device_remove+0x4c/0x80

Fixes: 44eb827264de ("dmaengine: fsl-edma: request per-channel IRQ only when channel is allocated")
Reviewed-by: Frank Li 
Signed-off-by: Peng Fan 
Link: https://lore.kernel.org/r/20250228071720.3780479-2-peng.fan@oss.nxp.com
Signed-off-by: Vinod Koul 
Signed-off-by: Sasha Levin

dmaengine: fsl-edma: cleanup chan after dma_async_device_unregister

2025-04-10T12:41:47+00:00

[ Upstream commit c9c59da76ce9cb3f215b66eb3708cda1134a5206 ]

There is kernel dump when do module test:
sysfs: cannot create duplicate filename
/devices/platform/soc@0/44000000.bus/44000000.dma-controller/dma/dma0chan0
 __dma_async_device_channel_register+0x128/0x19c
 dma_async_device_register+0x150/0x454
 fsl_edma_probe+0x6cc/0x8a0
 platform_probe+0x68/0xc8

fsl_edma_cleanup_vchan will unlink vchan.chan.device_node, while
dma_async_device_unregister  needs the link to do
__dma_async_device_channel_unregister. So need move fsl_edma_cleanup_vchan
after dma_async_device_unregister to make sure channel could be freed.

So clean up chan after dma_async_device_unregister to address this.

Fixes: 6f93b93b2a1b ("dmaengine: fsl-edma: kill the tasklets upon exit")
Reviewed-by: Frank Li 
Signed-off-by: Peng Fan 
Link: https://lore.kernel.org/r/20250228071720.3780479-1-peng.fan@oss.nxp.com
Signed-off-by: Vinod Koul 
Signed-off-by: Sasha Levin

dmaengine: ti: edma: fix OF node reference leaks in edma_driver

2025-02-08T09:02:07+00:00

[ Upstream commit e883c64778e5a9905fce955681f8ee38c7197e0f ]

The .probe() of edma_driver calls of_parse_phandle_with_fixed_args() but
does not release the obtained OF nodes. Thus add a of_node_put() call.

This bug was found by an experimental verification tool that I am
developing.

Fixes: 1be5336bc7ba ("dmaengine: edma: New device tree binding")
Signed-off-by: Joe Hattori 
Reviewed-by: Dan Carpenter 
Link: https://lore.kernel.org/r/20241219020507.1983124-3-joe@pf.is.s.u-tokyo.ac.jp
Signed-off-by: Vinod Koul 
Signed-off-by: Sasha Levin

Merge tag 'dmaengine-fix-6.13' of git://git.kernel.org/pub/scm/linux/kernel/git/vkoul/dmaengine

2024-12-26T18:43:25+00:00

Pull dmaengine fixes from Vinod Koul:
 "Bunch of minor driver fixes for drivers in this cycle:

   - Kernel doc warning documentation fixes

   - apple driver fix for register access

   - amd driver dropping private dma_ops

   - freescale cleanup path fix

   - refcount fix for mv_xor driver

   - null pointer deref fix for at_xdmac driver

   - GENMASK to GENMASK_ULL fix for loongson2 apb driver

   - Tegra driver fix for correcting dma status"

* tag 'dmaengine-fix-6.13' of git://git.kernel.org/pub/scm/linux/kernel/git/vkoul/dmaengine:
  dmaengine: tegra: Return correct DMA status when paused
  dmaengine: mv_xor: fix child node refcount handling in early exit
  dmaengine: fsl-edma: implement the cleanup path of fsl_edma3_attach_pd()
  dmaengine: amd: qdma: Remove using the private get and set dma_ops APIs
  dmaengine: apple-admac: Avoid accessing registers in probe
  linux/dmaengine.h: fix a few kernel-doc warnings
  dmaengine: loongson2-apb: Change GENMASK to GENMASK_ULL
  dmaengine: dw: Select only supported masters for ACPI devices
  dmaengine: at_xdmac: avoid null_prt_deref in at_xdmac_prep_dma_memset

dmaengine: tegra: Return correct DMA status when paused

2024-12-24T10:19:30+00:00

Currently, the driver does not return the correct DMA status when a DMA
pause is issued by the client drivers. This causes GPCDMA users to
assume that DMA is still running, while in reality, the DMA is paused.

Return DMA_PAUSED for tx_status() if the channel is paused in the middle
of a transfer.

Fixes: ee17028009d4 ("dmaengine: tegra: Add tegra gpcdma driver")
Cc: stable@vger.kernel.org
Signed-off-by: Akhil R 
Signed-off-by: Kartik Rajput 
Link: https://lore.kernel.org/r/20241212124412.5650-1-kkartik@nvidia.com
Signed-off-by: Vinod Koul

dmaengine: mv_xor: fix child node refcount handling in early exit

2024-12-24T10:15:01+00:00

The for_each_child_of_node() loop requires explicit calls to
of_node_put() to decrement the child's refcount upon early exits (break,
goto, return).

Add the missing calls in the two early exits before the goto
instructions.

Cc: stable@vger.kernel.org
Fixes: f7d12ef53ddf ("dma: mv_xor: add Device Tree binding")
Signed-off-by: Javier Carrasco 
Link: https://lore.kernel.org/r/20241011-dma_mv_xor_of_node_put-v1-1-3c2de819f463@gmail.com
Signed-off-by: Vinod Koul

dmaengine: fsl-edma: implement the cleanup path of fsl_edma3_attach_pd()

2024-12-24T09:25:57+00:00

Current implementation of fsl_edma3_attach_pd() does not provide a
cleanup path, resulting in a memory leak. For example,
dev_pm_domain_detach() is not called after dev_pm_domain_attach_by_id(),
and the device link created with the DL_FLAG_STATELESS is not released
explicitly.

Therefore, provide a cleanup function fsl_edma3_detach_pd() and call it
upon failure. Also add a devm_add_action_or_reset() call with this
function after a successful fsl_edma3_attach_pd().

Fixes: 72f5801a4e2b ("dmaengine: fsl-edma: integrate v3 support")
Signed-off-by: Joe Hattori 
Link: https://lore.kernel.org/r/20241221075712.3297200-1-joe@pf.is.s.u-tokyo.ac.jp
Signed-off-by: Vinod Koul

dmaengine: amd: qdma: Remove using the private get and set dma_ops APIs

2024-12-04T12:57:32+00:00

The get_dma_ops and set_dma_ops APIs were never for driver to use. Remove
these calls from QDMA driver. Instead, pass the DMA device pointer from the
qdma_platdata structure.

Fixes: 73d5fc92a11c ("dmaengine: amd: qdma: Add AMD QDMA driver")
Signed-off-by: Lizhi Hou 
Reviewed-by: Christoph Hellwig 
Link: https://lore.kernel.org/r/20240918181022.2155715-1-lizhi.hou@amd.com
Signed-off-by: Vinod Koul

dmaengine: apple-admac: Avoid accessing registers in probe

2024-12-04T12:12:27+00:00

The ADMAC attached to the AOP has complex power sequencing, and is
power gated when the probe callback runs. Move the register reads
to other functions, where we can guarantee that the hardware is
switched on.

Fixes: 568aa6dd641f ("dmaengine: apple-admac: Allocate cache SRAM to channels")
Signed-off-by: Sasha Finkelstein 
Link: https://lore.kernel.org/r/20241124-admac-power-v1-1-58f2165a4d55@gmail.com
Signed-off-by: Vinod Koul

module: Convert default symbol namespace to string literal

2024-12-03T16:22:25+00:00

Commit cdd30ebb1b9f ("module: Convert symbol namespace to string
literal") only converted MODULE_IMPORT_NS() and EXPORT_SYMBOL_NS(),
leaving DEFAULT_SYMBOL_NAMESPACE as a macro expansion.

This commit converts DEFAULT_SYMBOL_NAMESPACE in the same way to avoid
annoyance for the default namespace as well.

Signed-off-by: Masahiro Yamada 
Reviewed-by: Uwe Kleine-König 
Signed-off-by: Linus Torvalds