[PATCH v8 0/2] spi: add new_device/delete_device sysfs interface

[Vishwaroop A <va@nvidia.com>]

Add I2C-style new_device/delete_device sysfs attributes to SPI host
controllers, allowing userspace to instantiate and remove SPI devices
at runtime without device-tree changes.

Patch 1 adds the new_device/delete_device attributes and the supporting
infrastructure (userspace_clients list, manual sysfs group registration).
Patch 2 adds the ABI and user-facing documentation.

Changes since v7:
  - Remove the redundant CONFIG_SPI_DYNAMIC guard around
    spi_controller_groups[].  (Jon Hunter)
  - Restore the original add_lock release point; the userspace sysfs
    group is removed and drained before the lock is taken.  (Jon Hunter)
  - Remove the redundant dead flag and retain the base-tree
    device_is_registered() teardown check.  (Jon Hunter)
  - Simplify parsing, diagnostics and comments.  (Jon Hunter)

Changes since v6:
  - new_device_store(): reject empty device names.  (Mark Brown)
  - Move sysfs userspace group registration out of the automatic
    ctlr->dev.groups path; create it manually at the end of
    spi_register_controller() and remove it at the start of
    spi_unregister_controller().  (Mark Brown)
  - Drop userspace_clients_lock; reuse the existing add_lock.
    new_device_store() calls __spi_add_device() directly under
    add_lock; delete_device_store() also takes add_lock.
    new_device is DEVICE_ATTR_IGNORE_LOCKDEP alongside delete_device
    to silence the class-level splat.  (Jon Hunter)
  - kernel-doc / inline comment updates for the new struct fields.

Changes since v5:
  - Rebased on next-20260727 (v7.2-rc5); applies cleanly.
  - kernel-doc entries for the new struct spi_controller /
    struct spi_device fields.
  - new_device_store(): reject chip_select > U8_MAX; parse with %u.
  - new_device_store(): reject mode bits outside SPI_MODE_USER_MASK.
  - new_device_store(): drop the unused 'n' / '%n' from sscanf().

Changes since v4:
  - Removed spi_unregister_device() call from new_device_store()'s
    controller-teardown error path.

Changes since v3:
  - Replaced holding add_lock across __spi_add_device() + list
    insertion with the design that eventually landed as v8: sysfs
    group removed before add_lock is taken, so kernfs_drain()
    serialises us with unregister.

Changes since v2:
  - Gated sysfs attributes and locking on CONFIG_SPI_DYNAMIC.

Changes since v1:
  - Added locking to prevent races between new_device_store() and
    concurrent spi_unregister_controller().

Link: https://lore.kernel.org/linux-spi/20260728030541.2279518-1-va@nvidia.com/  # v7
Link: https://lore.kernel.org/linux-spi/cover.1784000000.git.va@nvidia.com/  # v6
Link: https://lore.kernel.org/linux-spi/20260517201602.498135-1-va@nvidia.com/  # v5
Link: https://lore.kernel.org/linux-tegra/909f0c92-d110-4253-903e-5c81e21e12c9@nvidia.com/

Vishwaroop A (2):
  spi: add new_device/delete_device sysfs interface
  docs: spi: add documentation for userspace device instantiation

 .../ABI/testing/sysfs-class-spi-master        |  34 +++
 Documentation/spi/index.rst                   |   1 +
 Documentation/spi/instantiating-devices.rst   |  88 +++++++
 drivers/spi/spi.c                             | 241 ++++++++++++++++++
 include/linux/spi/spi.h                       |  18 ++
 5 files changed, 382 insertions(+)
 create mode 100644 Documentation/ABI/testing/sysfs-class-spi-master
 create mode 100644 Documentation/spi/instantiating-devices.rst


base-commit: 0d33d21e47d9dc66f91e44da3fc9220c74d93df7
-- 
2.17.1