Claude review: bus: mhi: Move sahara protocol driver under drivers/bus/mhi

[Claude Code Review Bot <claude-review@example.com>]

Patch Review

Moves the driver, creates Kconfig/Makefile, and converts to `module_init`/`module_exit`.

**Issues:**
- **Copyright change is suspicious.** The original file had:
  ```c
  // SPDX-License-Identifier: GPL-2.0-only
  /* Copyright (c) 2024 Qualcomm Innovation Center, Inc. All rights reserved. */
  ```
  The new file has:
  ```c
  // SPDX-License-Identifier: GPL-2.0
  /* Copyright (c) 2018-2020, The Linux Foundation. All rights reserved. */
  ```
  Changing `GPL-2.0-only` to `GPL-2.0` (which in SPDX means "GPL-2.0-only" anyway but is ambiguous) and replacing the copyright holder with a different entity/date range is a red flag. This needs justification.

- **QAIC init/exit error path change looks buggy.** The patch modifies `qaic_init()` but the goto labels appear swapped:
  ```c
  ret = pci_register_driver(&qaic_pci_driver);
  if (ret) {
      pr_debug("qaic: pci_register_driver failed %d\n", ret);
  -   return ret;
  +   goto free_pci;
  }
  ret = mhi_driver_register(&qaic_mhi_driver);
  if (ret) {
      pr_debug("qaic: mhi_driver_register failed %d\n", ret);
  -   goto free_pci;
  +   goto free_mhi;
  }
  ```
  After the `sahara_register()` call is removed, if `mhi_driver_register` fails, the code now jumps to `free_mhi` which would unregister MHI (which just failed), instead of `free_pci` which should unregister the PCI driver. I'd need to see the full error path, but this looks inverted — the original code had `free_pci` for MHI failure (correct: unwind PCI), and now after removing sahara, the labels may have shifted incorrectly.

- The Kconfig `select FW_LOADER_COMPRESS` and `select FW_LOADER_COMPRESS_XZ` / `FW_LOADER_COMPRESS_ZSTD` is heavy-handed. Not all firmware necessarily needs compressed loading. This should be documented or justified.

---
Generated by Claude Code Patch Reviewer