linux-kernel - Re: [PATCH 13/17] pwm: cros-ec: Remove cros_ec_cmd_xfer

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]

Date:   Mon, 3 Feb 2020 10:39:39 -0800
From:   Prashant Malani <pmalani@...omium.org>
To:     Enric Balletbo i Serra <enric.balletbo@...labora.com>
Cc:     Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        Thierry Reding <thierry.reding@...il.com>,
        Uwe Kleine-König 
        <u.kleine-koenig@...gutronix.de>,
        Benson Leung <bleung@...omium.org>,
        Guenter Roeck <groeck@...omium.org>,
        "open list:PWM SUBSYSTEM" <linux-pwm@...r.kernel.org>
Subject: Re: [PATCH 13/17] pwm: cros-ec: Remove cros_ec_cmd_xfer_status()

Hi Enric,

On Mon, Feb 3, 2020 at 10:26 AM Prashant Malani <pmalani@...omium.org> wrote:
>
> On Mon, Feb 3, 2020 at 7:33 AM Enric Balletbo i Serra
> <enric.balletbo@...labora.com> wrote:
> >
> >
> > On 30/1/20 21:31, Prashant Malani wrote:
> > > Convert one existing usage of cros_ec_cmd_xfer_status() to
> > > cros_ec_send_cmd_msg(), which accomplishes the same thing but also does
> > > the EC message struct setup,and is defined in platform/chrome and is
> > > accessible by other modules.
> > >
> > > For the other usage, switch it to using cros_ec_cmd_xfer(), since the
> > > calling functions rely on the result field of the struct cros_ec_command
> > > struct that is used.
> > >
> > > Signed-off-by: Prashant Malani <pmalani@...omium.org>
> > > ---
> > >  drivers/pwm/pwm-cros-ec.c | 27 +++++++++------------------
> > >  1 file changed, 9 insertions(+), 18 deletions(-)
> > >
> > > diff --git a/drivers/pwm/pwm-cros-ec.c b/drivers/pwm/pwm-cros-ec.c
> > > index 89497448d21775..8bf610a6529e7e 100644
> > > --- a/drivers/pwm/pwm-cros-ec.c
> > > +++ b/drivers/pwm/pwm-cros-ec.c
> > > @@ -32,25 +32,14 @@ static inline struct cros_ec_pwm_device *pwm_to_cros_ec_pwm(struct pwm_chip *c)
> > >
> > >  static int cros_ec_pwm_set_duty(struct cros_ec_device *ec, u8 index, u16 duty)
> > >  {
> > > -     struct {
> > > -             struct cros_ec_command msg;
> > > -             struct ec_params_pwm_set_duty params;
> > > -     } __packed buf;
> > > -     struct ec_params_pwm_set_duty *params = &buf.params;
> > > -     struct cros_ec_command *msg = &buf.msg;
> > > -
> > > -     memset(&buf, 0, sizeof(buf));
> > > +     struct ec_params_pwm_set_duty params = {0};
> > >
> > > -     msg->version = 0;
> > > -     msg->command = EC_CMD_PWM_SET_DUTY;
> > > -     msg->insize = 0;
> > > -     msg->outsize = sizeof(*params);
> > > -
> > > -     params->duty = duty;
> > > -     params->pwm_type = EC_PWM_TYPE_GENERIC;
> > > -     params->index = index;
> > > +     params.duty = duty;
> > > +     params.pwm_type = EC_PWM_TYPE_GENERIC;
> > > +     params.index = index;
> > >
> > > -     return cros_ec_cmd_xfer_status(ec, msg);
> > > +     return cros_ec_send_cmd_msg(ec, 0, EC_CMD_PWM_SET_DUTY, &params,
> > > +                                 sizeof(params), NULL, 0);
> > >  }
> > >
> > >  static int __cros_ec_pwm_get_duty(struct cros_ec_device *ec, u8 index,
> > > @@ -78,11 +67,13 @@ static int __cros_ec_pwm_get_duty(struct cros_ec_device *ec, u8 index,
> > >       params->pwm_type = EC_PWM_TYPE_GENERIC;
> > >       params->index = index;
> > >
> > > -     ret = cros_ec_cmd_xfer_status(ec, msg);
> > > +     ret = cros_ec_cmd_xfer(ec, msg);
> >
> > Why? There is a good reason we introduced the cros_ec_cmd_xfer_status.
> >
> > IMO the purpose of introduce the new wrapper only makes sense if we can cover
> > _all_ the cases, so we can remove cros_ec_cmd_xfer_status and make
> > cros_ec_cmd_xfer private to cros_ec_proto.
I'm hoping for that too, but as we saw below (and some in some other
drivers), some callers of cros_ec_cmd_xfer() actually use the
msg->result.
Should we change the new wrapper to return the message via a pointer
(if not NULL), so something like this ? :

int cros_ec_send_cmd_msg(struct cros_ec_device *ec, unsigned int
version, uint32_t command, void *outdata, unsigned int outsize,
void *indata, unsigned int insize, uint32_t *result) ?

> >
> > Is not possible to use the new wrapper here?
> >
> > >       if (result)
> > >               *result = msg->result;
> >
> > Hmm, I see, that's the problem ...
> >
> > This driver will need a bit of rework but I think could be possible to use the
> > wrapper.
> Yeah, I looked around, and it seems to use msg->result.
> Perhaps we should work on reworking this driver before doing the large
> patch series? I would be happy to work on it, unless you feel there is
> someone else who'd be better suited. Kindly let me know.
> >
> > >       if (ret < 0)
> > >               return ret;
> > > +     else if (msg->result != EC_RES_SUCCESS)
> > > +             return -EPROTO;
> > >
> > >       return resp->duty;
> > >  }
> > >