linux.git/drivers/dma/idxd/cdev.c, branch v6.9

dmaengine: idxd: Convert spinlock to mutex to lock evl workqueue

2024-04-07T08:01:44+00:00

drain_workqueue() cannot be called safely in a spinlocked context due to
possible task rescheduling. In the multi-task scenario, calling
queue_work() while drain_workqueue() will lead to a Call Trace as
pushing a work on a draining workqueue is not permitted in spinlocked
context.
    Call Trace:
    
    ? __warn+0x7d/0x140
    ? __queue_work+0x2b2/0x440
    ? report_bug+0x1f8/0x200
    ? handle_bug+0x3c/0x70
    ? exc_invalid_op+0x18/0x70
    ? asm_exc_invalid_op+0x1a/0x20
    ? __queue_work+0x2b2/0x440
    queue_work_on+0x28/0x30
    idxd_misc_thread+0x303/0x5a0 [idxd]
    ? __schedule+0x369/0xb40
    ? __pfx_irq_thread_fn+0x10/0x10
    ? irq_thread+0xbc/0x1b0
    irq_thread_fn+0x21/0x70
    irq_thread+0x102/0x1b0
    ? preempt_count_add+0x74/0xa0
    ? __pfx_irq_thread_dtor+0x10/0x10
    ? __pfx_irq_thread+0x10/0x10
    kthread+0x103/0x140
    ? __pfx_kthread+0x10/0x10
    ret_from_fork+0x31/0x50
    ? __pfx_kthread+0x10/0x10
    ret_from_fork_asm+0x1b/0x30
    

The current implementation uses a spinlock to protect event log workqueue
and will lead to the Call Trace due to potential task rescheduling.

To address the locking issue, convert the spinlock to mutex, allowing
the drain_workqueue() to be called in a safe mutex-locked context.

This change ensures proper synchronization when accessing the event log
workqueue, preventing potential Call Trace and improving the overall
robustness of the code.

Fixes: c40bd7d9737b ("dmaengine: idxd: process user page faults for completion record")
Signed-off-by: Rex Zhang 
Reviewed-by: Dave Jiang 
Reviewed-by: Fenghua Yu 
Reviewed-by: Lijun Pan 
Link: https://lore.kernel.org/r/20240404223949.2885604-1-fenghua.yu@intel.com
Signed-off-by: Vinod Koul

Merge tag 'dmaengine-6.9-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/vkoul/dmaengine

2024-03-15T19:25:13+00:00

Pull dmaengine updates from Vinod Koul:
 "New hardware support:
   - Allwinner H616 dma support
   - Renesas r8a779h0 dma controller support
   - TI CSI2RX dma support

  Updates:
   - Freescale edma driver updates for TCD64csupport for i.MX95
   - constify of pointers and args
   - Yaml conversion for MediaTek High-Speed controller binding
   - TI k3 udma support for TX/RX DMA channels for thread IDs:

* tag 'dmaengine-6.9-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/vkoul/dmaengine: (25 commits)
  dmaengine: of: constify of_phandle_args in of_dma_find_controller()
  dmaengine: pl08x: constify pointer to char in filter function
  MAINTAINERS: change in AMD ptdma maintainer
  MAINTAINERS: adjust file entry in MEDIATEK DMA DRIVER
  dmaengine: idxd: constify the struct device_type usage
  dt-bindings: renesas,rcar-dmac: Add r8a779h0 support
  dt-bindings: dma: convert MediaTek High-Speed controller to the json-schema
  dmaengine: idxd: make dsa_bus_type const
  dmaengine: fsl-edma: integrate TCD64 support for i.MX95
  dt-bindings: fsl-dma: fsl-edma: add fsl,imx95-edma5 compatible string
  dmaengine: mcf-edma: utilize edma_write_tcdreg() macro for TCD Access
  dmaengine: fsl-edma: add address for channel mux register in fsl_edma_chan
  dmaengine: fsl-edma: fix spare build warning
  dmaengine: fsl-edma: involve help macro fsl_edma_set(get)_tcd()
  dt-bindings: mmp-dma: convert to YAML
  dmaengine: ti: k3-psil-j721s2: Add entry for CSI2RX
  dmaengine: ti: k3-udma-glue: Add function to request RX chan for thread ID
  dmaengine: ti: k3-udma-glue: Add function to request TX chan for thread ID
  dmaengine: ti: k3-udma-glue: Update name for remote RX channel device
  dmaengine: ti: k3-udma-glue: Add function to parse channel by ID
  ...

dmaengine: idxd: constify the struct device_type usage

2024-02-22T13:59:31+00:00

Since commit aed65af1cc2f ("drivers: make device_type const"), the driver
core can properly handle constant struct device_type. Move the
dsa_device_type, iax_device_type, idxd_wq_device_type, idxd_cdev_file_type,
idxd_cdev_device_type and idxd_group_device_type variables to be constant
structures as well, placing it into read-only memory which can not be
modified at runtime.

Cc: Greg Kroah-Hartman 
Signed-off-by: "Ricardo B. Marliere" 
Reviewed-by: Fenghua Yu 
Link: https://lore.kernel.org/r/20240219-device_cleanup-dmaengine-v1-1-9f72f3cf3587@marliere.net
Signed-off-by: Vinod Koul

dmaengine: idxd: Remove shadow Event Log head stored in idxd

2024-02-16T12:16:56+00:00

head is defined in idxd->evl as a shadow of head in the EVLSTATUS register.
There are two issues related to the shadow head:

1. Mismatch between the shadow head and the state of the EVLSTATUS
   register:
   If Event Log is supported, upon completion of the Enable Device command,
   the Event Log head in the variable idxd->evl->head should be cleared to
   match the state of the EVLSTATUS register. But the variable is not reset
   currently, leading mismatch between the variable and the register state.
   The mismatch causes incorrect processing of Event Log entries.

2. Unnecessary shadow head definition:
   The shadow head is unnecessary as head can be read directly from the
   EVLSTATUS register. Reading head from the register incurs no additional
   cost because event log head and tail are always read together and
   tail is already read directly from the register as required by hardware.

Remove the shadow Event Log head stored in idxd->evl to address the
mentioned issues.

Fixes: 244da66cda35 ("dmaengine: idxd: setup event log configuration")
Signed-off-by: Fenghua Yu 
Reviewed-by: Dave Jiang 
Link: https://lore.kernel.org/r/20240215024931.1739621-1-fenghua.yu@intel.com
Signed-off-by: Vinod Koul

Merge tag 'dmaengine-fix-6.8-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/vkoul/dmaengine

2024-01-20T23:03:25+00:00

Pull dmaengine updates from Vinod Koul:
 "New support:
   - Loongson LS2X APB DMA controller
   - sf-pdma: mpfs-pdma support
   - Qualcomm X1E80100 GPI dma controller support

  Updates:
   - Xilinx XDMA updates to support interleaved DMA transfers
   - TI PSIL threads for AM62P and J722S and cfg register regions
     description
   - axi-dmac Improving the cyclic DMA transfers
   - Tegra Support dma-channel-mask property
   - Remaining platform remove callback returning void conversions

 Driver fixes for:
   - Xilinx xdma driver operator precedence and initialization fix
   - Excess kernel-doc warning fix in imx-sdma xilinx xdma drivers
   - format-overflow warning fix for rz-dmac, sh usb dmac drivers
   - 'output may be truncated' fix for shdma, fsl-qdma and dw-edma
     drivers"

* tag 'dmaengine-fix-6.8-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/vkoul/dmaengine: (58 commits)
  dmaengine: dw-edma: increase size of 'name' in debugfs code
  dmaengine: fsl-qdma: increase size of 'irq_name'
  dmaengine: shdma: increase size of 'dev_id'
  dmaengine: xilinx: xdma: Fix kernel-doc warnings
  dmaengine: usb-dmac: Avoid format-overflow warning
  dmaengine: sh: rz-dmac: Avoid format-overflow warning
  dmaengine: imx-sdma: fix Excess kernel-doc warnings
  dmaengine: xilinx: xdma: Fix initialization location of desc in xdma_channel_isr()
  dmaengine: xilinx: xdma: Fix operator precedence in xdma_prep_interleaved_dma()
  dmaengine: xilinx: xdma: statify xdma_prep_interleaved_dma
  dmaengine: xilinx: xdma: Workaround truncation compilation error
  dmaengine: pl330: issue_pending waits until WFP state
  dmaengine: xilinx: xdma: Implement interleaved DMA transfers
  dmaengine: xilinx: xdma: Prepare the introduction of interleaved DMA transfers
  dmaengine: xilinx: xdma: Add transfer error reporting
  dmaengine: xilinx: xdma: Add error checking in xdma_channel_isr()
  dmaengine: xilinx: xdma: Rework xdma_terminate_all()
  dmaengine: xilinx: xdma: Ease dma_pool alignment requirements
  dmaengine: xilinx: xdma: Add necessary macro definitions
  dmaengine: xilinx: xdma: Get rid of unused code
  ...

dmaengine: idxd: Remove usage of the deprecated ida_simple_xx() API

2023-12-21T15:59:21+00:00

ida_alloc() and ida_free() should be preferred to the deprecated
ida_simple_get() and ida_simple_remove().

This is less verbose.

Note that the upper limit of ida_simple_get() is exclusive, but the one of
ida_alloc_range() is inclusive. Sothis change allows one more device.

MINORMASK is ((1U << MINORBITS) - 1), so allowing MINORMASK as a maximum value
makes sense. It is also consistent with other "ida_.*MINORMASK" and
"ida_*MINOR()" usages.

Signed-off-by: Christophe JAILLET 
Reviewed-by: Fenghua Yu 
Acked-by: Lijun Pan 
Link: https://lore.kernel.org/r/ac991f5f42112fa782a881d391d447529cbc4a23.1702967302.git.christophe.jaillet@wanadoo.fr
Signed-off-by: Vinod Koul

dmaengine: idxd: Rename drv_enable/disable_wq to idxd_drv_enable/disable_wq, and export

2023-12-15T09:52:53+00:00

Rename drv_enable_wq and drv_disable_wq to idxd_drv_enable_wq and
idxd_drv_disable_wq respectively, so that they're no longer too
generic to be exported.  This also matches existing naming within the
idxd driver.

And to allow idxd sub-drivers to enable and disable wqs, export them.

Signed-off-by: Tom Zanussi 
Reviewed-by: Dave Jiang 
Reviewed-by: Fenghua Yu 
Acked-by: Vinod Koul 
Signed-off-by: Herbert Xu

dmaengine: idxd: add wq driver name support for accel-config user tool

2023-10-04T07:30:34+00:00

With the possibility of multiple wq drivers that can be bound to the wq,
the user config tool accel-config needs a way to know which wq driver to
bind to the wq. Introduce per wq driver_name sysfs attribute where the user
can indicate the driver to be bound to the wq. This allows accel-config to
just bind to the driver using wq->driver_name.

Signed-off-by: Dave Jiang 
Signed-off-by: Tom Zanussi 
Reviewed-by: Fenghua Yu 
Acked-by: Vinod Koul 
Link: https://lore.kernel.org/r/20230908201045.4115614-1-fenghua.yu@intel.com
Signed-off-by: Vinod Koul

dmaengine: idxd: Fix passing freed memory in idxd_cdev_open()

2023-05-17T06:45:09+00:00

Smatch warns:
	drivers/dma/idxd/cdev.c:327:
		idxd_cdev_open() warn: 'sva' was already freed.

When idxd_wq_set_pasid() fails, the current code unbinds sva and then
goes to 'failed_set_pasid' where iommu_sva_unbind_device is called
again causing the above warning.
[ device_user_pasid_enabled(idxd) is still true when calling
failed_set_pasid ]

Fix this by removing additional unbind when idxd_wq_set_pasid() fails

Fixes: b022f59725f0 ("dmaengine: idxd: add idxd_copy_cr() to copy user completion record during page fault handling")
Signed-off-by: Harshit Mogalapalli 
Acked-by: Fenghua Yu 
Acked-by: Dave Jiang 
Link: https://lore.kernel.org/r/20230509060716.2830630-1-harshit.m.mogalapalli@oracle.com
Signed-off-by: Vinod Koul

dmaengine: idxd: add pid to exported sysfs attribute for opened file

2023-04-12T17:48:46+00:00

Provide the pid of the application for the opened file. This allows the
monitor daemon to easily correlate which app opened the file and easily
kill the app by pid if that is desired action.

Tested-by: Tony Zhu 
Signed-off-by: Dave Jiang 
Co-developed-by: Fenghua Yu 
Signed-off-by: Fenghua Yu 
Link: https://lore.kernel.org/r/20230407203143.2189681-16-fenghua.yu@intel.com
Signed-off-by: Vinod Koul