Re: [PATCH] Input: apm-power - remove driver

[Nicolas Schier <nsc@kernel.org>]

On Mon, Nov 09, 2026 at 11:44:25AM -0500, Nathan Chancellor wrote:
> On Mon, Nov 09, 2026 at 12:09:44AM -0800, Dmitry Torokhov wrote:
> > The apm-power input handler solely listens for EV_PWR KEY_SUSPEND
> > events to queue APM_USER_SUSPEND via legacy APM emulation. The only
> > in-tree platforms generating those events are legacy Sharp Zaurus ARM
> > machines (spitz/borzoi/akita and collie), none of which were converted
> > to Device Tree. Both mach-pxa board files and mach-sa1100 were marked
> > deprecated in v7.3 by commit cc7852314abd ("ARM: PXA: mark remaining
> > board files as deprecated") and commit 7d92e50ceacd ("ARM: mark
> > mach-sa1100 as deprecated"), and are queued for removal in v7.5. None
> > of the remaining ARM platforms or other APM_EMULATION targets
> > (MIPS_ALCHEMY, PMAC_APM_EMU, and SH_HP6XX) emit EV_PWR input events or
> > enable CONFIG_INPUT_APMPOWER, and arch/arm/configs/pxa_defconfig
> > already dropped CONFIG_PXA_SHARPSL at the same time, leaving its
> > CONFIG_INPUT_APMPOWER=m entry inert.
> > 
> > Normalizing dashes and underscores in scripts/modules-check.sh for v7.5
> > exposes a module name collision (apm_power) between
> > drivers/input/apm-power.ko and drivers/power/supply/apm_power.ko.
> > Remove this obsolete input bridge to resolve the collision rather than
> > renaming a module whose last target platforms are going away in the
> > same cycle.
> > 
> > Reported-by: Sasha Levin <sashal@kernel.org>
> > Closes: https://lore.kernel.org/r/20261108225808.1809266-1-sashal@kernel.org
> > Link: https://lore.kernel.org/r/20261109-apm-power-input-rename-v1-1-0609e7b91825@kernel.org
> > Link: https://lore.kernel.org/r/e4da2f9d-d3b9-4267-9b64-b0f2cb43402a@app.fastmail.com
> > Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
> 
> Thanks a lot for helping with this.
> 
> Reviewed-by: Nathan Chancellor <nathan@kernel.org>
> 
> Nicolas, do you prefer taking this via kbuild-next directly or having
> Dmitry take it via the input tree and giving us an immutable branch? I
> assume the chance of conflicts or follow-ups are low, so we can probably
> just take it directly but I don't want to assume your preference.

Thanks!  I'd like to take it via kbuild tree.  Please complain, if I
shall drop it again.

-- 
Nicolas