linux-kernel - Re: [RFT] [PATCH] regulator: tps65219: Fix .bypass_val

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]

Date:   Tue, 13 Sep 2022 12:00:30 +0200
From:   jerome Neanne <jneanne@...libre.com>
To:     Axel Lin <axel.lin@...ics.com>
Cc:     Mark Brown <broonie@...nel.org>,
        Liam Girdwood <lgirdwood@...il.com>,
        Tony Lindgren <tony@...mide.com>, linux-kernel@...r.kernel.org
Subject: Re: [RFT] [PATCH] regulator: tps65219: Fix .bypass_val_on setting

Got it! And validate it.

On 13/09/2022 01:48, Axel Lin wrote:
> jerome Neanne <jneanne@...libre.com> 於 2022年9月12日 週一 下午5:49寫道：
>> Hi Axel,
>>
>> On 28/08/2022 14:01, Axel Lin wrote:
>>> The .bypass_val_on setting does not match the .bypass_mask setting, so the
>>> .bypass_mask bit will never get set.  Fix it by removing .bypass_val_on
>>> setting, the regulator_set_bypass_regmap and regulator_get_bypass_regmap
>>> helpers will use rdev->desc->bypass_mask as val_on if the val_on is 0.
>> I think this will result in exact same behavior. val would be assigned
>> to 1 when enable is set and 0 otherwise. Anyway you are right this line
>> is useless.
> Setting .bypass_val_on=1 won't set TPS65219_LDOS_BYP_CONFIG_MASK bit.
> The TPS65219_LDOS_BYP_CONFIG_MASK is BIT(6), so you need to set BIT(6)
> instead of 1 for .bypass_val_on.
> Remove .bypass_val_on setting then it will use .bypass_mask as
> .bypass_val_on setting.

Got your point. The issue is on the get path not on the set path...

I misinterpreted val_on usage.

I was able to confirm on board that cat 
/sys/class/regulator/regulator.10/bypass
always return disabled before your fix is applied.

After your fix is applied, it accurately return enabled/disabled 
depending how LDO is set (forced at probe).

side note:

- I added virtual regulator and userspace consumer to validate (that's 
why VDDSHV_SD_IO_PMIC is regulator.10)

- This bug was missed because I did not re-validated the bypass after I 
changed from custom to standard helpers (I apologize).

Thanks for raising this

Regards,

Jerome.