linux-kernel - Re: [PATCH 5/6] i2c: spacemit: ensure SDA is released after bus reset

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

Message-ID: <88C40EE66EF5519F+aNHukycHdAifY9du@LT-Guozexi>
Date: Tue, 23 Sep 2025 08:49:23 +0800
From: Troy Mitchell <troy.mitchell@...ux.spacemit.com>
To: Troy Mitchell <troy.mitchell@...ux.spacemit.com>,
	Andi Shyti <andi.shyti@...nel.org>, Yixun Lan <dlan@...too.org>,
	Alex Elder <elder@...cstar.com>,
	Troy Mitchell <troymitchell988@...il.com>,
	linux-i2c@...r.kernel.org, linux-kernel@...r.kernel.org,
	linux-riscv@...ts.infradead.org, spacemit@...ts.linux.dev
Subject: Re: [PATCH 5/6] i2c: spacemit: ensure SDA is released after bus reset

On Mon, Sep 22, 2025 at 10:56:10PM +0200, Aurelien Jarno wrote:
> On 2025-08-27 15:39, Troy Mitchell wrote:
> > After performing a conditional bus reset, the controller must ensure
> > that the SDA line is actually released.
> > 
> > Previously, the reset routine only performed a single check,
> > which could leave the bus in a locked state in some situations.
> > 
> > This patch introduces a loop that toggles the reset cycle and issues
> > a reset request up to SPACEMIT_BUS_RESET_CLK_CNT_MAX times, checking
> > SDA after each attempt. If SDA is released before the maximum count,
> > the function returns early. Otherwise, a warning is emitted.
> > 
> > This change improves bus recovery reliability.
> > 
> > Fixes: 5ea558473fa31 ("i2c: spacemit: add support for SpacemiT K1 SoC")
> > Signed-off-by: Troy Mitchell <troy.mitchell@...ux.spacemit.com>
> > ---
> >  drivers/i2c/busses/i2c-k1.c | 23 ++++++++++++++++++++++-
> >  1 file changed, 22 insertions(+), 1 deletion(-)
> > 
> > diff --git a/drivers/i2c/busses/i2c-k1.c b/drivers/i2c/busses/i2c-k1.c
> > index 4d78ee7b6929ee43771e500d4f85d9e55e68b221..d2c0d20d19ba73baa8b2e9a6acb02b2cc3b7243f 100644
> > --- a/drivers/i2c/busses/i2c-k1.c
> > +++ b/drivers/i2c/busses/i2c-k1.c
> > @@ -3,6 +3,7 @@
> >   * Copyright (C) 2024-2025 Troy Mitchell <troymitchell988@...il.com>
> >   */
> >  
> > +#include <linux/bitfield.h>
> >   #include <linux/clk.h>
> >   #include <linux/i2c.h>
> >   #include <linux/iopoll.h>
> > @@ -26,7 +27,8 @@
> >  #define SPACEMIT_CR_MODE_FAST    BIT(8)		/* bus mode (master operation) */
> >  /* Bit 9 is reserved */
> >  #define SPACEMIT_CR_UR           BIT(10)	/* unit reset */
> > -/* Bits 11-12 are reserved */
> > +#define SPACEMIT_CR_RSTREQ	 BIT(11)	/* i2c bus reset request */
> > +/* Bit 12 is reserved */
> >  #define SPACEMIT_CR_SCLE         BIT(13)	/* master clock enable */
> >  #define SPACEMIT_CR_IUE          BIT(14)	/* unit enable */
> >  /* Bits 15-17 are reserved */
> > @@ -78,6 +80,7 @@
> >  					SPACEMIT_SR_ALD)
> >  
> >  #define SPACEMIT_RCR_SDA_GLITCH_NOFIX		BIT(7)		/* bypass the SDA glitch fix */
> > +#define SPACEMIT_RCR_FIELD_RST_CYC		GENMASK(3, 0)	/* bypass the SDA glitch fix */
> 
> The comment here seems wrong, datasheet says "The cycles of SCL during 
> bus reset."
Yes, that's my mistake.
I'll fix it.

> 
> >  /* SPACEMIT_IBMR register fields */
> >  #define SPACEMIT_BMR_SDA         BIT(0)		/* SDA line level */
> > @@ -91,6 +94,8 @@
> >  
> >  #define SPACEMIT_SR_ERR	(SPACEMIT_SR_BED | SPACEMIT_SR_RXOV | SPACEMIT_SR_ALD)
> >  
> > +#define SPACEMIT_BUS_RESET_CLK_CNT_MAX		9
> > +
> >  enum spacemit_i2c_state {
> >  	SPACEMIT_STATE_IDLE,
> >  	SPACEMIT_STATE_START,
> > @@ -163,6 +168,7 @@ static int spacemit_i2c_handle_err(struct spacemit_i2c_dev *i2c)
> >  static void spacemit_i2c_conditionally_reset_bus(struct spacemit_i2c_dev *i2c)
> >  {
> >  	u32 status;
> > +	u8 clk_cnt;
> >  
> >  	/* if bus is locked, reset unit. 0: locked */
> >  	status = readl(i2c->base + SPACEMIT_IBMR);
> > @@ -172,6 +178,21 @@ static void spacemit_i2c_conditionally_reset_bus(struct spacemit_i2c_dev *i2c)
> >  	spacemit_i2c_reset(i2c);
> >  	usleep_range(10, 20);
> >  
> > +	for (clk_cnt = 0; clk_cnt < SPACEMIT_BUS_RESET_CLK_CNT_MAX; clk_cnt++) {
> > +		status = readl(i2c->base + SPACEMIT_IBMR);
> > +		if (status & SPACEMIT_BMR_SDA)
> > +			break;
> 
> What about just adding the return here instead of checking clk_cnt 
> below?
> 
See below.
> > +
> > +		/* There's nothing left to save here, we are about to exit */
> > +		writel(FIELD_PREP(SPACEMIT_RCR_FIELD_RST_CYC, 1),
> > +		       i2c->base + SPACEMIT_IRCR);
> > +		writel(SPACEMIT_CR_RSTREQ, i2c->base + SPACEMIT_ICR);
> > +		usleep_range(20, 30);
> > +	}
> > +
> > +	if (clk_cnt < SPACEMIT_BUS_RESET_CLK_CNT_MAX)
> > +		return;
> > +
> >  	/* check sda again here */
> >  	status = readl(i2c->base + SPACEMIT_IBMR);
> >  	if (!(status & SPACEMIT_BMR_SDA))
> 
> Once we have exited the loop, I am not sure we should check SDA once 
> more, maybe just display the error message directly.
I'll double check it.
If it's unnecessary, I'll just return directly.

                - Troy
> 
> -- 
> Aurelien Jarno                          GPG: 4096R/1DDD8C9B
> aurelien@...el32.net                     http://aurel32.net
>