Re: [PATCH v3] pinctrl: sunxi: fix gpiochip_lock_as_irq() failure when pinmux is unknown

[Michal Piekos]

On Thu, Mar 19, 2026 at 03:16:33PM +0100, Andre Przywara wrote:
> Hi,
> 
> On 3/19/26 15:06, Chen-Yu Tsai wrote:
> > On Thu, Mar 19, 2026 at 9:46 PM Linus Walleij <linusw@xxxxxxxxxx> wrote:
> > > 
> > > On Sat, Mar 14, 2026 at 9:10 AM Michal Piekos
> > > <michal.piekos@xxxxxxxxxxxxx> wrote:
> > > 
> > > > Fixes kernel hang during boot due to inability to set up IRQ on AXP313a.
> > > > 
> > > > The issue is caused by gpiochip_lock_as_irq() which is failing when gpio
> > > > is in unitialized state.
> > > > 
> > > > Solution is to set pinmux to GPIO INPUT in
> > > > sunxi_pinctrl_irq_request_resources() if it wasn't initialized
> > > > earlier.
> > > > 
> > > > Tested on Orange Pi Zero 3.
> > > > 
> > > > Signed-off-by: Michal Piekos <michal.piekos@xxxxxxxxxxxxx>
> > > > Suggested-by: Chen-Yu Tsai <wens@xxxxxxxxxx>
> > > 
> > > Should this have a Fixes: tag and should it be applied as an urgent fix?
> > 
> > Fixes: 01e10d0272b9 ("pinctrl: sunxi: Implement gpiochip::get_direction()")
> > 
> > > Tag for stable?
> > 
> > The original patch wasn't tagged for stable. And it doesn't seem like it
> > was auto picked.
> > 
> > > Also: is there consensus with Andre that this is the way to proceed?
> > 
> > As Andre mentioned, the mux value for "disabled" is different between
> > generations. So we likely need to make the value part of
> > |struct sunxi_pinctrl_desc|.
> 
> There is already the SUNXI_PINCTRL_NEW_REG_LAYOUT flag, which describes
> exactly that: the new layout increased the bit field to 4 bits, moving the
> HiZ value to 0xf.
> But it's just a flag passed to sunxi_pinctrl_init_with_flags(), not stored
> in desc, and not available during runtime. Which is what I tried to fix for
> my A733 series:
> https://lore.kernel.org/linux-arm-kernel/20250821004232.8134-3-andre.przywara@xxxxxxx/
> 
> So maybe we can cherry-pick just this patch, then check for that flag?

I'll prepare v4 with your suggestion.

> 
> Cheers,
> Andre
> 
> > 
> > Michal, can you respin a version so that we can get this fixed in the
> > same release?
> > 
> > 
> > ChenYu
> > 
>