linux-kernel - Re: [PATCH] regulator: bd718x7, bd71828, Fix dvs voltage levels

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:   Fri, 15 Jan 2021 14:10:34 +0000
From:   "Vaittinen, Matti" <Matti.Vaittinen@...rohmeurope.com>
To:     "lee.jones@...aro.org" <lee.jones@...aro.org>
CC:     "lgirdwood@...il.com" <lgirdwood@...il.com>,
        linux-power <linux-power@...rohmeurope.com>,
        "broonie@...nel.org" <broonie@...nel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] regulator: bd718x7, bd71828, Fix dvs voltage levels


On Fri, 2021-01-15 at 13:47 +0000, Lee Jones wrote:
> On Fri, 15 Jan 2021, Matti Vaittinen wrote:
> 
> > The ROHM BD718x7 and BD71828 drivers support setting HW state
> > specific voltages from device-tree. This is used also by various
> > in-tree DTS files.
> > 
> > These drivers do incorrectly try to compose bit-map using enum
> > values. By a chance this works for first two valid levels having
> > values 1 and 2 - but setting values for the rest of the levels
> > do indicate capbility of setting values for first levels as
> > well. Luckily the regulators which support settin values for
> > SUSPEND/LPSR do usually also support setting values for RUN
> > and IDLE too - thus this has not been such a fatal issue.
> > 
> > Fix this by defining the old enum values as bits and using
> > new enum in parsing code. This allows keeping existing IC
> > specific drivers intact and only adding the defines and
> > slightly changing the rohm-regulator.c
> > 
> > Fixes: 21b72156ede8b ("regulator: bd718x7: Split driver to common
> > and bd718x7 specific parts")
> > Signed-off-by: Matti Vaittinen <matti.vaittinen@...rohmeurope.com>
> > ---
> >  drivers/regulator/rohm-regulator.c |  8 ++++----
> >  include/linux/mfd/rohm-generic.h   | 22 ++++++++++++++++------
> >  2 files changed, 20 insertions(+), 10 deletions(-)
> > 
> > diff --git a/drivers/regulator/rohm-regulator.c
> > b/drivers/regulator/rohm-regulator.c
> > index 399002383b28..96caae7dbef4 100644
> > --- a/drivers/regulator/rohm-regulator.c
> > +++ b/drivers/regulator/rohm-regulator.c
> > @@ -55,25 +55,25 @@ int rohm_regulator_set_dvs_levels(const struct
> > rohm_dvs_config *dvs,
> >  	for (i = 0; i < ROHM_DVS_LEVEL_MAX && !ret; i++) {
> >  		if (dvs->level_map & (1 << i)) {
> >  			switch (i + 1) {
> > -			case ROHM_DVS_LEVEL_RUN:
> > +			case _ROHM_DVS_LEVEL_RUN:
> >  				prop = "rohm,dvs-run-voltage";
> >  				reg = dvs->run_reg;
> >  				mask = dvs->run_mask;
> >  				omask = dvs->run_on_mask;
> >  				break;
> > -			case ROHM_DVS_LEVEL_IDLE:
> > +			case _ROHM_DVS_LEVEL_IDLE:
> >  				prop = "rohm,dvs-idle-voltage";
> >  				reg = dvs->idle_reg;
> >  				mask = dvs->idle_mask;
> >  				omask = dvs->idle_on_mask;
> >  				break;
> > -			case ROHM_DVS_LEVEL_SUSPEND:
> > +			case _ROHM_DVS_LEVEL_SUSPEND:
> >  				prop = "rohm,dvs-suspend-voltage";
> >  				reg = dvs->suspend_reg;
> >  				mask = dvs->suspend_mask;
> >  				omask = dvs->suspend_on_mask;
> >  				break;
> > -			case ROHM_DVS_LEVEL_LPSR:
> > +			case _ROHM_DVS_LEVEL_LPSR:
> >  				prop = "rohm,dvs-lpsr-voltage";
> >  				reg = dvs->lpsr_reg;
> >  				mask = dvs->lpsr_mask;
> > diff --git a/include/linux/mfd/rohm-generic.h
> > b/include/linux/mfd/rohm-generic.h
> > index 4283b5b33e04..a557988831d7 100644
> > --- a/include/linux/mfd/rohm-generic.h
> > +++ b/include/linux/mfd/rohm-generic.h
> > @@ -20,15 +20,25 @@ struct rohm_regmap_dev {
> >  	struct regmap *regmap;
> >  };
> >  
> > +/*
> > + * Do not use these in IC specific driver - the bit map should be
> > created using
> > + * defines without the underscore. These should be used only in
> > rohm-regulator.c
> > + */
> >  enum {
> > -	ROHM_DVS_LEVEL_UNKNOWN,
> > -	ROHM_DVS_LEVEL_RUN,
> > -	ROHM_DVS_LEVEL_IDLE,
> > -	ROHM_DVS_LEVEL_SUSPEND,
> > -	ROHM_DVS_LEVEL_LPSR,
> > -	ROHM_DVS_LEVEL_MAX = ROHM_DVS_LEVEL_LPSR,
> > +	_ROHM_DVS_LEVEL_UNKNOWN,
> > +	_ROHM_DVS_LEVEL_RUN,
> > +	_ROHM_DVS_LEVEL_IDLE,
> > +	_ROHM_DVS_LEVEL_SUSPEND,
> > +	_ROHM_DVS_LEVEL_LPSR,
> > +	ROHM_DVS_LEVEL_MAX = _ROHM_DVS_LEVEL_LPSR,
> 
> I don't understand how this is still not broken.
> 
> I think you either need to change the for() loop that consumes this
> to
> use "<=" or push the MAX enum to the last line (on its own).
> 
> The latter would be my personal preference.

Bah. Occasionally getting it right is just hard. The for loop condition
is allright now as it does as it was originally intended and scans the
amount of valid values. Starting at 0, using condition i <
ROHM_DVS_LEVEL_MAX guarantees we process all valid values _knowing_
that value 0 is 'invalid'.

But now I made a new error in following if-condition. I didn't take
into account the fact that left-sifting the 1 by _ROHM_DVS_LEVEL_RUN
for first valid value (ROHM_DVS_LEVEL_RUN) makes the first valid
bitmask to be 2 not 1. So the if-condition

if (dvs->level_map & (1 << i))

following the for loop is now broken. It would probably work if it was
2 << i. Anyways - we're now wasting one bit out of 32 for invalid mask.

I think you're correct though. It needs a change. This is quite good
indication that this loop/check is overly complex :) So please forget
this patch - I will do better at next week when I get to the office and
can also test this with a break-out board. 

Sorry for wasting the time and thanks for the review again.

Best Regards
	Matti Vaittinen