[PATCH v2 0/3] usb: dwc3: xilinx: error path and teardown fixes

[Radhey Shyam Pandey <radhey.shyam.pandey@amd.com>]

The ZynqMP path in dwc3-xilinx deasserts three resets and initialises the
USB3 PHY, but nothing unwinds that state when a later probe step fails,
and nothing unwinds it on remove either.  System suspend and resume have
the same problem: suspend calls phy_exit() without powering the PHY off
and ignores the result, and resume can leave clocks enabled if PHY
reinitialisation fails.

These three patches make each of those paths release what it acquired.

v1 [1] sent them together with two platform-data cleanups.  Per review
feedback they are now split out so the fixes come first and can be
backported on their own; the cleanups follow in a separate series against
usb-next.

Patch 3 was reworked for that split.  In v1 it registered the teardown as
plat->exit in struct dwc3_xlnx_platdata, which one of the cleanup patches
introduces, so it could not have been backported.  It now adds a
pltfm_exit pointer alongside the existing pltfm_init in struct dwc3_xlnx
and is self-contained.  The follow-up series folds both pointers into the
platform data struct.

All three are tagged for stable.

Link to v1:
https://lore.kernel.org/all/20260810174713.2325292-1-radhey.shyam.pandey@amd.com

Radhey Shyam Pandey (3):
  usb: dwc3: xilinx: fix system suspend and resume PHY handling
  usb: dwc3: xilinx: re-assert resets on ZynqMP init error paths
  usb: dwc3: xilinx: unwind ZynqMP platform init on probe failure and
    remove

 drivers/usb/dwc3/dwc3-xilinx.c | 101 ++++++++++++++++++++++++---------
 1 file changed, 77 insertions(+), 24 deletions(-)


base-commit: abc36cbda29d8f19cf3a580cd86ca9e865186a41
-- 
2.43.0