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: <20210301100239.orbt7km6lgjwvzii@maple.lan>
Date:   Mon, 1 Mar 2021 10:02:39 +0000
From:   Daniel Thompson <daniel.thompson@...aro.org>
To:     Kiran Gunda <kgunda@...eaurora.org>
Cc:     bjorn.andersson@...aro.org, jingoohan1@...il.com,
        lee.jones@...aro.org, b.zolnierkie@...sung.com,
        dri-devel@...ts.freedesktop.org, jacek.anaszewski@...il.com,
        pavel@....cz, robh+dt@...nel.org, mark.rutland@....com,
        linux-leds@...r.kernel.org, devicetree@...r.kernel.org,
        linux-kernel@...r.kernel.org, Andy Gross <agross@...nel.org>,
        linux-arm-msm@...r.kernel.org, linux-fbdev@...r.kernel.org,
        phone-devel@...r.kernel.org
Subject: Re: [PATCH V3 2/2] backlight: qcom-wled: Correct the sync_toggle
 sequence

On Mon, Mar 01, 2021 at 02:58:36PM +0530, Kiran Gunda wrote:
> As per the current implementation, after FSC (Full Scale Current)
> and brightness update the sync bits are transitioned from set-then-clear.

This does not makes sense since there are too many verbs. Set and clear
are both verbs so in this context: "the code will set the bit and then
the code will clear the bit".

Either:

s/transitioned from set-then-clear/set-then-cleared/.

Or:

s/transitioned from set-then-clear/using a set-then-clear approach/.

> But, the FSC and brightness sync takes place during a clear-then-set
> transition of the sync bits.

Likewise this no longer makes sense and had also become misleading.
Two changes of state, clear and then set, do not usually result in a
single transition.

Either:

s/clear-then-set/0 to 1/

Alternatively, if you want to stick exclusively to the set/clear
terminology then replace the whole quoted section with:

  But, the FSC and brightness sync takes place when the sync bits are
  set (e.g. on a rising edge).


> So the hardware team recommends a
> clear-then-set approach in order to guarantee such a transition
> regardless of the previous register state.
> 
> Signed-off-by: Kiran Gunda <kgunda@...eaurora.org>

With one of each of the changes proposed above:
Reviewed-by: Daniel Thompson <daniel.thompson@...aro.org>


Daniel.


> ---
>  drivers/video/backlight/qcom-wled.c | 12 ++++++------
>  1 file changed, 6 insertions(+), 6 deletions(-)
> 
> diff --git a/drivers/video/backlight/qcom-wled.c b/drivers/video/backlight/qcom-wled.c
> index aef52b9..19f83ac 100644
> --- a/drivers/video/backlight/qcom-wled.c
> +++ b/drivers/video/backlight/qcom-wled.c
> @@ -337,13 +337,13 @@ static int wled3_sync_toggle(struct wled *wled)
>  
>  	rc = regmap_update_bits(wled->regmap,
>  				wled->ctrl_addr + WLED3_SINK_REG_SYNC,
> -				mask, mask);
> +				mask, WLED3_SINK_REG_SYNC_CLEAR);
>  	if (rc < 0)
>  		return rc;
>  
>  	rc = regmap_update_bits(wled->regmap,
>  				wled->ctrl_addr + WLED3_SINK_REG_SYNC,
> -				mask, WLED3_SINK_REG_SYNC_CLEAR);
> +				mask, mask);
>  
>  	return rc;
>  }
> @@ -353,17 +353,17 @@ static int wled5_mod_sync_toggle(struct wled *wled)
>  	int rc;
>  	u8 val;
>  
> -	val = (wled->cfg.mod_sel == MOD_A) ? WLED5_SINK_REG_SYNC_MOD_A_BIT :
> -					     WLED5_SINK_REG_SYNC_MOD_B_BIT;
>  	rc = regmap_update_bits(wled->regmap,
>  				wled->sink_addr + WLED5_SINK_REG_MOD_SYNC_BIT,
> -				WLED5_SINK_REG_SYNC_MASK, val);
> +				WLED5_SINK_REG_SYNC_MASK, 0);
>  	if (rc < 0)
>  		return rc;
>  
> +	val = (wled->cfg.mod_sel == MOD_A) ? WLED5_SINK_REG_SYNC_MOD_A_BIT :
> +					     WLED5_SINK_REG_SYNC_MOD_B_BIT;
>  	return regmap_update_bits(wled->regmap,
>  				  wled->sink_addr + WLED5_SINK_REG_MOD_SYNC_BIT,
> -				  WLED5_SINK_REG_SYNC_MASK, 0);
> +				  WLED5_SINK_REG_SYNC_MASK, val);
>  }
>  
>  static int wled_ovp_fault_status(struct wled *wled, bool *fault_set)
> -- 
> The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
>  a Linux Foundation Collaborative Project
> 

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ