Two-patch series addressing Andy's review of the leak-fix on v1.

Patch 1 keeps the same single-line leak fix as v1, but with:
- the correct "serial: 8250_dw:" prefix (underscore),
- a Fixes: tag pointing at the original clk_notifier introduction,
- Cc: stable@ so the fix gets picked up by stable branches that
  still carry the notifier code.

Patch 2 drops the clock-notifier infrastructure entirely from
mainline, as suggested by Andy. The notifier was introduced for the
Baikal-T1 SoC (shared baudclk between UART ports) and has no other
in-tree user; Baikal-T1 support has been removed from the kernel.

If a future platform needs the cross-device baudclk-rate notification
pattern again, it can be reintroduced in a more general form.

Stepan Ionichev (2):
  serial: 8250_dw: unregister 8250 port if clk_notifier_register() fails
  serial: 8250_dw: remove clock-notifier infrastructure

 drivers/tty/serial/8250/8250_dw.c | 79 -------------------------------
 1 file changed, 79 deletions(-)

-- 
2.43.0

On Thu, May 14, 2026 at 07:37:44PM +0500, Stepan Ionichev wrote:
> Two-patch series addressing Andy's review of the leak-fix on v1.
> 
> Patch 1 keeps the same single-line leak fix as v1, but with:
> - the correct "serial: 8250_dw:" prefix (underscore),
> - a Fixes: tag pointing at the original clk_notifier introduction,
> - Cc: stable@ so the fix gets picked up by stable branches that
>   still carry the notifier code.
> 
> Patch 2 drops the clock-notifier infrastructure entirely from
> mainline, as suggested by Andy. The notifier was introduced for the
> Baikal-T1 SoC (shared baudclk between UART ports) and has no other
> in-tree user; Baikal-T1 support has been removed from the kernel.
> 
> If a future platform needs the cross-device baudclk-rate notification
> pattern again, it can be reintroduced in a more general form.

Seems legit, especially the second patch.
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>


-- 
With Best Regards,
Andy Shevchenko