linux-stable.git/drivers/firmware/xilinx, branch v6.3

firmware: xilinx: don't make a sleepable memory allocation from an atomic context

2023-03-09T17:00:31+00:00

The following issue was discovered using lockdep:
[    6.691371] BUG: sleeping function called from invalid context at include/linux/sched/mm.h:209
[    6.694602] in_atomic(): 1, irqs_disabled(): 128, non_block: 0, pid: 1, name: swapper/0
[    6.702431] 2 locks held by swapper/0/1:
[    6.706300]  #0: ffffff8800f6f188 (&dev->mutex){....}-{3:3}, at: __device_driver_lock+0x4c/0x90
[    6.714900]  #1: ffffffc009a2abb8 (enable_lock){....}-{2:2}, at: clk_enable_lock+0x4c/0x140
[    6.723156] irq event stamp: 304030
[    6.726596] hardirqs last  enabled at (304029): [] _raw_spin_unlock_irqrestore+0xc0/0xd0
[    6.736142] hardirqs last disabled at (304030): [] clk_enable_lock+0xfc/0x140
[    6.744742] softirqs last  enabled at (303958): [] _stext+0x4f0/0x894
[    6.752655] softirqs last disabled at (303951): [] irq_exit+0x238/0x280
[    6.760744] CPU: 1 PID: 1 Comm: swapper/0 Tainted: G     U            5.15.36 #2
[    6.768048] Hardware name: xlnx,zynqmp (DT)
[    6.772179] Call trace:
[    6.774584]  dump_backtrace+0x0/0x300
[    6.778197]  show_stack+0x18/0x30
[    6.781465]  dump_stack_lvl+0xb8/0xec
[    6.785077]  dump_stack+0x1c/0x38
[    6.788345]  ___might_sleep+0x1a8/0x2a0
[    6.792129]  __might_sleep+0x6c/0xd0
[    6.795655]  kmem_cache_alloc_trace+0x270/0x3d0
[    6.800127]  do_feature_check_call+0x100/0x220
[    6.804513]  zynqmp_pm_invoke_fn+0x8c/0xb0
[    6.808555]  zynqmp_pm_clock_getstate+0x90/0xe0
[    6.813027]  zynqmp_pll_is_enabled+0x8c/0x120
[    6.817327]  zynqmp_pll_enable+0x38/0xc0
[    6.821197]  clk_core_enable+0x144/0x400
[    6.825067]  clk_core_enable+0xd4/0x400
[    6.828851]  clk_core_enable+0xd4/0x400
[    6.832635]  clk_core_enable+0xd4/0x400
[    6.836419]  clk_core_enable+0xd4/0x400
[    6.840203]  clk_core_enable+0xd4/0x400
[    6.843987]  clk_core_enable+0xd4/0x400
[    6.847771]  clk_core_enable+0xd4/0x400
[    6.851555]  clk_core_enable_lock+0x24/0x50
[    6.855683]  clk_enable+0x24/0x40
[    6.858952]  fclk_probe+0x84/0xf0
[    6.862220]  platform_probe+0x8c/0x110
[    6.865918]  really_probe+0x110/0x5f0
[    6.869530]  __driver_probe_device+0xcc/0x210
[    6.873830]  driver_probe_device+0x64/0x140
[    6.877958]  __driver_attach+0x114/0x1f0
[    6.881828]  bus_for_each_dev+0xe8/0x160
[    6.885698]  driver_attach+0x34/0x50
[    6.889224]  bus_add_driver+0x228/0x300
[    6.893008]  driver_register+0xc0/0x1e0
[    6.896792]  __platform_driver_register+0x44/0x60
[    6.901436]  fclk_driver_init+0x1c/0x28
[    6.905220]  do_one_initcall+0x104/0x590
[    6.909091]  kernel_init_freeable+0x254/0x2bc
[    6.913390]  kernel_init+0x24/0x130
[    6.916831]  ret_from_fork+0x10/0x20

Fix it by passing the GFP_ATOMIC gfp flag for the corresponding
memory allocation.

Fixes: acfdd18591ea ("firmware: xilinx: Use hash-table for api feature check")
Cc: stable 
Signed-off-by: Roman Gushchin 
Cc: Amit Sunil Dhamne 
Cc: Michal Simek 
Cc: linux-arm-kernel@lists.infradead.org
Cc: linux-kernel@vger.kernel.org
Link: https://lore.kernel.org/r/20230308222602.123866-1-roman.gushchin@linux.dev
Signed-off-by: Greg Kroah-Hartman

firmware: xilinx: Clear IOCTL_SET_SD_TAPDELAY using PM_MMIO_WRITE

2023-01-05T09:18:07+00:00

In case the tap delay required by Arasan SDHCI is set to 0, the current
embeddedsw firmware unconditionally writes IOU_SLCR SD_ITAPDLY to 0x100
(SD0_ITAPDLYENA=1, SD0_ITAPDLYSEL=0). Previous behavior was to keep the
IOU_SLCR SD_ITAPDLY set to 0x0. There is some sort of difference in the
behavior between SD0_ITAPDLYENA=1/0 with the same SD0_ITAPDLYSEL=0, even
though the behavior should be identical -- zero delay added to rxclk_in
line. The former breaks HS200 training in low temperature conditions.

Write IOU_SLCR SD_ITAPDLY register to 0 using PM_MMIO_WRITE which seem
to allow unrestricted WRITE access (and PM_MMIO_READ which allows read
access) to the entire address space. This way, it is possible to work
around the defect in IOCTL_SET_SD_TAPDELAY design which does not permit
clearing SDx_ITAPDLYENA bit.

Note that the embeddedsw firmware does not permit clearing the SD_ITAPDLY
SD0_ITAPDLYENA bit, this bit can only ever be set by the firmware and it
is often impossible to update the possibly broken firmware.

Signed-off-by: Marek Vasut 
Link: https://lore.kernel.org/r/20221215152023.8387-1-marex@denx.de
Signed-off-by: Michal Simek

Merge tag 'rproc-v6.2' of git://git.kernel.org/pub/scm/linux/kernel/git/remoteproc/linux

2022-12-21T17:37:14+00:00

Pull remoteproc updates from Bjorn Andersson:
 "rproc-virtio device names are now auto generated, to avoid conflicts
  between remoteproc instances.

  The imx_rproc driver is extended with support for communicating with
  and attaching to a running M4 on i.MX8QXP, as well as support for
  attaching to the M4 after self-recovering from a crash. Support is
  added for i.MX8QM and mailbox channels are reconnected during the
  recovery process, in order to avoid data corruption.

  The Xilinx Zynqmp firmware interface is extended and support for the
  Xilinx R5 RPU is introduced.

  Various resources leaks, primarily in error paths, throughout the
  Qualcomm drivers are corrected.

  Lastly a fix to ensure that pm_relax is invoked even if the remoteproc
  instance is stopped between a crash is being reported and the recovery
  handler is scheduled"

* tag 'rproc-v6.2' of git://git.kernel.org/pub/scm/linux/kernel/git/remoteproc/linux: (25 commits)
  remoteproc: core: Do pm_relax when in RPROC_OFFLINE state
  remoteproc: qcom: q6v5: Fix missing clk_disable_unprepare() in q6v5_wcss_qcs404_power_on()
  remoteproc: qcom_q6v5_pas: Fix missing of_node_put() in adsp_alloc_memory_region()
  remoteproc: qcom_q6v5_pas: detach power domains on remove
  remoteproc: qcom_q6v5_pas: disable wakeup on probe fail or remove
  remoteproc: qcom: q6v5: Fix potential null-ptr-deref in q6v5_wcss_init_mmio()
  remoteproc: sysmon: fix memory leak in qcom_add_sysmon_subdev()
  remoteproc: sysmon: Make QMI message rules const
  drivers: remoteproc: Add Xilinx r5 remoteproc driver
  firmware: xilinx: Add RPU configuration APIs
  firmware: xilinx: Add shutdown/wakeup APIs
  firmware: xilinx: Add ZynqMP firmware ioctl enums for RPU configuration.
  arm64: dts: xilinx: zynqmp: Add RPU subsystem device node
  dt-bindings: remoteproc: Add Xilinx RPU subsystem bindings
  remoteproc: core: Use device_match_of_node()
  remoteproc: imx_rproc: Correct i.MX93 DRAM mapping
  remoteproc: imx_rproc: Enable attach recovery for i.MX8QM/QXP
  remoteproc: imx_rproc: Request mbox channel later
  remoteproc: imx_rproc: Support i.MX8QM
  remoteproc: imx_rproc: Support kicking Mcore from Linux for i.MX8QXP
  ...

firmware: xilinx: Add RPU configuration APIs

2022-11-25T16:11:47+00:00

This patch adds APIs to access to configure RPU and its
processor-specific memory.

That is query the run-time mode of RPU as either split or lockstep as well
as API to set this mode. In addition add APIs to access configuration of
the RPUs' tightly coupled memory (TCM).

Signed-off-by: Ben Levinsky 
Signed-off-by: Tanmay Shah 
Acked-by: Michal Simek 
Link: https://lore.kernel.org/r/20221114233940.2096237-6-tanmay.shah@amd.com
Signed-off-by: Mathieu Poirier

firmware: xilinx: Add shutdown/wakeup APIs

2022-11-25T16:10:54+00:00

Add shutdown/wakeup a resource eemi operations to shutdown
or bringup a resource.

Note alignment of args matches convention of other fn's in this file.
The reason being that the long fn name results in aligned args that
otherwise go over 80 chars so shift right to avoid this

Signed-off-by: Ben Levinsky 
Signed-off-by: Tanmay Shah 
Acked-by: Michal Simek 
Link: https://lore.kernel.org/r/20221114233940.2096237-5-tanmay.shah@amd.com
Signed-off-by: Mathieu Poirier

firmware: xilinx: Add qspi firmware interface

2022-10-19T13:34:37+00:00

Add support for QSPI ioctl functions and enums.

Signed-off-by: Rajan Vaja 
Signed-off-by: Michal Simek 
Signed-off-by: Amit Kumar Mahapatra 
Link: https://lore.kernel.org/r/20221011062040.12116-5-amit.kumar-mahapatra@amd.com
Signed-off-by: Mark Brown

firmware: xilinx: add support for sd/gem config

2022-09-20T15:33:04+00:00

Add new APIs in firmware to configure SD/GEM registers. Internally
it calls PM IOCTL for below SD/GEM register configuration:
- SD/EMMC select
- SD slot type
- SD base clock
- SD 8 bit support
- SD fixed config
- GEM SGMII Mode
- GEM fixed config

Signed-off-by: Ronak Jain 
Signed-off-by: Radhey Shyam Pandey 
Reviewed-by: Claudiu Beznea 
Signed-off-by: Jakub Kicinski

firmware: xilinx: Add TF_A_PM_REGISTER_SGI SMC call

2022-06-29T12:46:22+00:00

SGI interrupt register and reset is performed by EEMI ioctl
IOCTL_REGISTER_SGI. However, this is not correct use of EEMI call.
SGI registration functionality does not qualify as energy management
activity and so shouldn't be mapped to EEMI call.

This new call will replace IOCTL_REGISTER_SGI and will  be handled by TF-A
specific handler in TF-A. To maintain backward compatibility for a while
firmware driver will still use IOCTL_REGISTER_SGI as fallback strategy if
new call fails or is not supported by TF-A.

This new design also helps to make TF-A as pass through layer for EEMI
calls. So we don't have to maintain PM_IOCTL as EEMI API ID in TF-A.

Signed-off-by: Tanmay Shah 
Acked-by: Michal Simek 
Link: https://lore.kernel.org/r/20220607224253.54919-1-tanmay.shah@xilinx.com
Signed-off-by: Michal Simek

firmware: xilinx: enable feature check for ZynqMP

2022-04-22T15:20:03+00:00

Enable the feature check if the PM_FEATURE_CHECK API returns success
with the supported version for the ZynqMP. Currently, it is enabled
for Versal only.

Move get_set_conduit_method() at the beginning as the Linux is
requesting to TF-A for the PM_FEATURE_CHECK API version for which the
interface should be enabled with TF-A.

Signed-off-by: Ronak Jain 
Link: https://lore.kernel.org/r/1649242526-17493-5-git-send-email-ronak.jain@xilinx.com
Signed-off-by: Greg Kroah-Hartman

firmware: xilinx: always check API version for IOCTL/QUERY

2022-04-22T15:20:03+00:00

Currently, we are not checking feature check version for PM APIs as
the support may or may not there in the firmware. To check the whether
the feature check API is supported or not in the firmware, allow
checking for its own version.

Signed-off-by: Ronak Jain 
Link: https://lore.kernel.org/r/1649242526-17493-4-git-send-email-ronak.jain@xilinx.com
Signed-off-by: Greg Kroah-Hartman