[PATCH 0/3] firmware: qcom: scm: pass struct qcom_scm through the API

[Albert Esteve <aesteve@redhat.com>]

The SCM driver still funnels every call through a process-wide
__scm singleton.

This series threads struct qcom_scm through the driver and the
exported API, then drops the global.

Series summary
--------------
Patch 1: Internal helpers take the instance. Callbacks that only
have a struct device or reset_controller_dev recover it
with drvdata / container_of. Exported signatures stay
unchanged so this patch is bisectable on its own.

Patch 2: Exported calls take struct qcom_scm * as the first
argument. The type stays opaque in qcom_scm.h.

    Children of the SCM device (qseecom, tzmem, qcomtee)
    use dev_get_drvdata() on the parent. Other consumers
    call qcom_scm_get(), which finds the bound platform
    device and returns NULL until probe has marked the
    instance ready. Callers that need SCM return
    -EPROBE_DEFER; callers that only use it for an optional
    path (rmtfs VMIDs, fastrpc vmids, PAS region assign,
    Venus CP, HDMI HDCP, Adreno fuse poke) skip the lookup
    or tolerate NULL. qcom_pas_* is unchanged: PAS has a
    TEE backend, so those ops keep struct device * and the
    SCM backend uses drvdata.

Patch 3: Drop __scm. Readiness is scm->available with the same
release/acquire pairing as today. Module-parameter
stores reuse qcom_scm_get(); shutdown uses
platform_get_drvdata().

No intended change in SCM calling convention or wait-queue
behaviour.

Testing
-------

Booted to a shell on Qualcomm SA8775P Ride4 with current series.
qcom_scm bound as firmware:qcom_scm; probe logged SMC ARM 64 and
the tz-ffi reserved-memory assignment. Consumers were not exercised.

Suggested-by: Maxime Ripard <mripard@kernel.org>
Signed-off-by: Albert Esteve <aesteve@redhat.com>
---
Albert Esteve (3):
      firmware: qcom: scm: pass qcom_scm to internal helpers
      firmware: qcom: scm: pass qcom_scm through the exported API
      firmware: qcom: scm: drop the __scm global

 arch/arm/mach-qcom/platsmp.c                     |   3 +-
 drivers/cpuidle/cpuidle-qcom-spm.c               |  28 +-
 drivers/firmware/qcom/qcom_qseecom.c             |   3 +-
 drivers/firmware/qcom/qcom_scm-smc.c             |   2 +-
 drivers/firmware/qcom/qcom_scm.c                 | 518 +++++++++++++----------
 drivers/firmware/qcom/qcom_scm.h                 |   2 +-
 drivers/firmware/qcom/qcom_tzmem.c               |   6 +-
 drivers/gpu/drm/msm/adreno/a6xx_gmu.c            |   6 +-
 drivers/gpu/drm/msm/adreno/adreno_gpu.c          |   8 +-
 drivers/gpu/drm/msm/hdmi/hdmi_hdcp.c             |   6 +-
 drivers/iommu/arm/arm-smmu/arm-smmu-qcom-debug.c |   9 +-
 drivers/iommu/arm/arm-smmu/arm-smmu-qcom.c       |   8 +-
 drivers/iommu/arm/arm-smmu/arm-smmu-qcom.h       |   3 +
 drivers/iommu/arm/arm-smmu/qcom_iommu.c          |  18 +-
 drivers/irqchip/qcom-pdc.c                       |   6 +-
 drivers/media/platform/qcom/venus/firmware.c     |  10 +-
 drivers/misc/fastrpc.c                           |  31 +-
 drivers/net/wireless/ath/ath10k/qmi.c            |  12 +-
 drivers/net/wireless/ath/ath10k/qmi.h            |   3 +
 drivers/nvmem/sec-qfprom.c                       |   6 +-
 drivers/pinctrl/qcom/pinctrl-msm.c               |  11 +-
 drivers/remoteproc/qcom_q6v5_mss.c               |   6 +-
 drivers/remoteproc/qcom_q6v5_pas.c               |  12 +-
 drivers/soc/qcom/ice.c                           |  23 +-
 drivers/soc/qcom/ocmem.c                         |  22 +-
 drivers/soc/qcom/rmtfs_mem.c                     |   8 +-
 drivers/tee/qcomtee/call.c                       |   7 +
 drivers/tee/qcomtee/core.c                       |   6 +-
 drivers/tee/qcomtee/qcomtee.h                    |   2 +
 drivers/thermal/qcom/lmh.c                       |  22 +-
 include/linux/firmware/qcom/qcom_qseecom.h       |   5 +-
 include/linux/firmware/qcom/qcom_scm.h           | 117 ++---
 32 files changed, 566 insertions(+), 363 deletions(-)
---
base-commit: bc35965f6940a9bf834d54187b6088b8eb09206d
change-id: 20260908-scm-device-api-634ac7a7f377

Best regards,
-- 
Albert Esteve <aesteve@redhat.com>