linux-kernel - RE: [PATCH V2 2/2] firmware: arm

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, 14 Feb 2020 01:04:49 +0000
From:   Peng Fan <peng.fan@....com>
To:     Sudeep Holla <sudeep.holla@....com>
CC:     "robh+dt@...nel.org" <robh+dt@...nel.org>,
        "mark.rutland@....com" <mark.rutland@....com>,
        "viresh.kumar@...aro.org" <viresh.kumar@...aro.org>,
        "f.fainelli@...il.com" <f.fainelli@...il.com>,
        dl-linux-imx <linux-imx@....com>,
        "linux-arm-kernel@...ts.infradead.org" 
        <linux-arm-kernel@...ts.infradead.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "devicetree@...r.kernel.org" <devicetree@...r.kernel.org>,
        "andre.przywara@....com" <andre.przywara@....com>
Subject: RE: [PATCH V2 2/2] firmware: arm_scmi: add smc/hvc transports

Hi Sudeep,

> Subject: Re: [PATCH V2 2/2] firmware: arm_scmi: add smc/hvc transports
> 
> On Thu, Feb 13, 2020 at 11:58:50AM +0800, peng.fan@....com wrote:
> > From: Peng Fan <peng.fan@....com>
> >
> > Add SCMI smc/hvc transports support.
> >
> > Take smc-id as the 2nd arg, and protocol id as the 2nd arg when
> > invokding SMC/HVC. Since we need protocol id, so add this parrameter
> > to chan_setup, then smc transport driver could directly use it.
> > There is no Rx, only Tx because of smc/hvc not support Rx.
> >
> > Signed-off-by: Peng Fan <peng.fan@....com>
> > ---
> >  drivers/firmware/arm_scmi/Makefile  |   2 +-
> >  drivers/firmware/arm_scmi/common.h  |   4 +-
> >  drivers/firmware/arm_scmi/driver.c  |  11 +-
> >  drivers/firmware/arm_scmi/mailbox.c |   2 +-
> >  drivers/firmware/arm_scmi/smc.c     | 222
> ++++++++++++++++++++++++++++++++++++
> >  5 files changed, 234 insertions(+), 7 deletions(-)  create mode
> > 100644 drivers/firmware/arm_scmi/smc.c
> 
> [...]
> 
> >
> > diff --git a/drivers/firmware/arm_scmi/driver.c
> > b/drivers/firmware/arm_scmi/driver.c
> > index dbec767222e9..65c56328e6d8 100644
> > --- a/drivers/firmware/arm_scmi/driver.c
> > +++ b/drivers/firmware/arm_scmi/driver.c
> > @@ -595,7 +595,7 @@ static int scmi_chan_setup(struct scmi_info *info,
> > struct device *dev,
> >
> >  	cinfo->dev = dev;
> >
> > -	ret = info->desc->ops->chan_setup(cinfo, info->dev, tx);
> > +	ret = info->desc->ops->chan_setup(cinfo, info->dev, prot_id, tx);
> 
> Why do you need this ?

For smc tranports, all protocols share same smd-id, but if protocols
not share the same shmem, we need let firmare know which protocol
issues the smc call. So I take prot_id as an arguments of smc call.

> 
> >  	if (ret)
> >  		return ret;
> >
> > @@ -826,7 +829,7 @@ ATTRIBUTE_GROUPS(versions);
> >
> >  /* Each compatible listed below must have descriptor associated with
> > it */  static const struct of_device_id scmi_of_match[] = {
> > -	{ .compatible = "arm,scmi", .data = &scmi_mailbox_desc },
> > +	{ .compatible = "arm,scmi",  },
> 
> Don't do this, get "arm,scmi-smc"

You mean code as below?
/* Each compatible listed below must have descriptor associated with it */
static const struct of_device_id scmi_of_match[] = {
         { .compatible = "arm,scmi", .data = &scmi_mailbox_desc },
         { .compatible = "arm,scmi-smc", .data = &scmi_smc_desc },
         { /* Sentinel */ },
};

But since we could use mboxes and smc-id to know the tranports type,
do we really need arm,scmi-smc?

> 
> >  	{ /* Sentinel */ },
> >  };
> >
> [...]
> 
> 
> > +static unsigned long
> > +__invoke_scmi_fn_hvc(unsigned long function_id, unsigned long arg0,
> > +		     unsigned long arg1, unsigned long arg2,
> > +		     unsigned long arg3, unsigned long arg4,
> > +		     unsigned long arg5, unsigned long arg6) {
> > +	struct arm_smccc_res res;
> > +
> > +	arm_smccc_hvc(function_id, arg0, arg1, arg2, arg3, arg4, arg5,
> > +		      arg6, &res);
> > +
> > +	return res.a0;
> > +}
> > +
> > +static unsigned long
> > +__invoke_scmi_fn_smc(unsigned long function_id, unsigned long arg0,
> > +		     unsigned long arg1, unsigned long arg2,
> > +		     unsigned long arg3, unsigned long arg4,
> > +		     unsigned long arg5, unsigned long arg6) {
> > +	struct arm_smccc_res res;
> > +
> > +	arm_smccc_smc(function_id, arg0, arg1, arg2, arg3, arg4, arg5,
> > +		      arg6, &res);
> > +
> > +	return res.a0;
> > +}
> > +
> > +static int scmi_smc_conduit_method(struct device_node *np) {
> > +	const char *method;
> > +
> > +	if (invoke_scmi_smc_fn)
> > +		return 0;
> > +
> > +	if (of_property_read_string(np, "method", &method))
> > +		return -ENXIO;
> > +
> > +	if (!strcmp("hvc", method)) {
> > +		invoke_scmi_smc_fn = __invoke_scmi_fn_hvc;
> > +	} else if (!strcmp("smc", method)) {
> > +		invoke_scmi_smc_fn = __invoke_scmi_fn_smc;
> > +	} else {
> > +		pr_warn("invalid \"method\" property: %s\n", method);
> > +		return -EINVAL;
> > +	}
> > +
> > +	return 0;
> > +}
> > +
> 
> You don't the above functions

ok

> 
> [...]
> 
> > +
> > +	np = of_find_node_by_path("/psci");
> > +	if (!np) {
> > +		dev_err(dev, "Not able to find /psci node\n");
> > +		return -ENODEV;
> > +	}
> 
> No need for this as mentioned below.

ok

> 
> > +
> > +	ret = scmi_smc_conduit_method(np);
> 
> Just use arm_smccc_1_1_get_conduit if you want to get the conduit which I
> don't think you need. You can just use arm_smccc_1_1_invoke() directly.

Fix in v3.

I'll post v3 after we have an agreement on whether we need a new compatible
string arm,scmi-smc and the prot_id introduced in chan_setup.

Thanks,
Peng.

> 
> --
> Regards,
> Sudeep