linux-kernel - Re: [PATCH 2/2] pinctrl: spacemit: support I/O power domain configuration

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <47781E30EC71C60F+aUzitZG78aGBxHOG@kernel.org>
Date: Thu, 25 Dec 2025 15:07:33 +0800
From: Troy Mitchell <troy.mitchell@...ux.spacemit.com>
To: Yao Zi <me@...ao.cc>, Troy Mitchell <troy.mitchell@...ux.spacemit.com>,
	Inochi Amaoto <inochiama@...il.com>, Rob Herring <robh@...nel.org>,
	Krzysztof Kozlowski <krzk+dt@...nel.org>,
	Conor Dooley <conor+dt@...nel.org>, Yixun Lan <dlan@...too.org>,
	Paul Walmsley <pjw@...nel.org>, Palmer Dabbelt <palmer@...belt.com>,
	Albert Ou <aou@...s.berkeley.edu>, Alexandre Ghiti <alex@...ti.fr>,
	Linus Walleij <linusw@...nel.org>
Cc: devicetree@...r.kernel.org, linux-riscv@...ts.infradead.org,
	spacemit@...ts.linux.dev, linux-kernel@...r.kernel.org,
	linux-gpio@...r.kernel.org
Subject: Re: [PATCH 2/2] pinctrl: spacemit: support I/O power domain
 configuration

On Tue, Dec 23, 2025 at 01:03:27PM +0000, Yao Zi wrote:
> On Tue, Dec 23, 2025 at 05:50:08PM +0800, Troy Mitchell wrote:
> > On Tue, Dec 23, 2025 at 05:42:26PM +0800, Inochi Amaoto wrote:
> > > On Tue, Dec 23, 2025 at 05:11:12PM +0800, Troy Mitchell wrote:
> > > > IO domain power control registers are used to configure the operating
> > > > voltage of dual-voltage GPIO banks. By default, these registers are
> > > > configured for 3.3V operation. As a result, even when a GPIO bank is
> > > > externally supplied with 1.8V, the internal logic continues to
> > > > operate in the 3.3V domain, which may lead to functional failures.
> > > > 
> > > > This patch adds support for programming the IO domain power control
> > > > registers, allowing dual-voltage GPIO banks to be explicitly configured
> > > > for 1.8V operation when required.
> > > > 
> > > > Care must be taken when configuring these registers. If a GPIO bank is
> > > > externally supplied with 3.3V while the corresponding IO power domain
> > > > is configured for 1.8V, external current injection (back-powering)
> > > > may occur, potentially causing damage to the GPIO pin.
> > > > 
> > > > Due to these hardware constraints and safety considerations, the IO
> > > > domain power control registers are implemented as secure registers.
> > > > Access to these registers requires unlocking via the AIB Secure Access
> > > > Register (ASAR) in the APBC block before a single read or write
> > > > operation can be performed.
> > > > 
> > > > Signed-off-by: Troy Mitchell <troy.mitchell@...ux.spacemit.com>
> > > > ---
> > > >  arch/riscv/boot/dts/spacemit/k1.dtsi  |   4 +-
> > > >  drivers/pinctrl/spacemit/pinctrl-k1.c | 131 +++++++++++++++++++++++++++++++++-
> > > >  2 files changed, 131 insertions(+), 4 deletions(-)
> > > > 
> > > 
> > > > diff --git a/arch/riscv/boot/dts/spacemit/k1.dtsi b/arch/riscv/boot/dts/spacemit/k1.dtsi
> > > > index 7818ca4979b6a7755722919a5958512aa11950ab..23ecb19624f227f3c39de35bf3078379f7a2490e 100644
> > > > --- a/arch/riscv/boot/dts/spacemit/k1.dtsi
> > > > +++ b/arch/riscv/boot/dts/spacemit/k1.dtsi
> > > > @@ -565,10 +565,12 @@ i2c8: i2c@...1d800 {
> > > >  
> > > >  		pinctrl: pinctrl@...1e000 {
> > > >  			compatible = "spacemit,k1-pinctrl";
> > > > -			reg = <0x0 0xd401e000 0x0 0x400>;
> > > > +			reg = <0x0 0xd401e000 0x0 0x400>,
> > > > +			      <0x0 0xd401e800 0x0 0x34>;
> > > >  			clocks = <&syscon_apbc CLK_AIB>,
> > > >  				 <&syscon_apbc CLK_AIB_BUS>;
> > > >  			clock-names = "func", "bus";
> > > > +			spacemit,apbc = <&syscon_apbc 0x50>;
> > > >  		};
> > > 
> > > If you insist on a new reg field, you should change the binding as well.
> > Yes, I forgot to modify the binding.
> 
> This will also break ABI compatibility with older devicetrees, I
> strongly suggest against a new item in reg property.
> 
> Furthermore, it's unreasonable to describe d401_e000 - d401_e400 and
> d401_e800 - d401_e834 as separate memory regions. TRM claims the region
> starting from 0xd401_e000 with length 0xc00 is "Pad Configuration". So I
> think this separation neither simplifies anything nor matches the
> hardware.
> 
> > > This change breaks binding, can we use something like <0x0 0xd401e000 0x0 0x1000>?
> 
> If the TRM is correct, we probably can and should.
I have double checked. Inochi is correct.
I'll fix it in the next version.
Thanks!

                              - Troy