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  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, 25 Sep 2018 11:31:50 +0300
From:   Andy Shevchenko <andy.shevchenko@...il.com>
To:     Rajat Jain <rajatja@...gle.com>
Cc:     subrata.banik@...el.com,
        Mika Westerberg <mika.westerberg@...ux.intel.com>,
        Linus Walleij <linus.walleij@...aro.org>,
        "open list:GPIO SUBSYSTEM" <linux-gpio@...r.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        Rajat Jain <rajatxjain@...il.com>, aamir.bohra@...el.com
Subject: Re: [PATCH] pinctrl: icelake: Fix the resource number for community-4/5

On Tue, Sep 25, 2018 at 12:32 AM Rajat Jain <rajatja@...gle.com> wrote:
> On Mon, Sep 24, 2018 at 2:09 PM Andy Shevchenko
> <andy.shevchenko@...il.com> wrote:
> > On Mon, Sep 24, 2018 at 8:04 PM Rajat Jain <rajatja@...gle.com> wrote:
> > > On Mon, Sep 24, 2018 at 7:54 AM Banik, Subrata <subrata.banik@...el.com> wrote:

> > > > First of all, this is pre-production chip, so, I don't think there is a bug in the driver (yet) discovered.
> > > >
> > > > Looking to the above ASL code I may conclude that is definitely is *not* from our reference BIOS.
> > > > I have checked two versions of it and found that in both we have the following mapping:
> > > > for LP variant: there are only 4 communities are exported for H variant: there are only 5 communities are exported
> > > >
> > > > So, I guess the problem is in ASL code you provided. It simple should not export that community at all.
> > > >
> > > > In case you need to do so, there are ways:
> > > >  - contact Intel and ask for a change in reference BIOS
> > > >  - acquire another ACPI ID for the case, or, perhaps use special constants like
> > > >    _HRV for that purpose (also need to contact Intel while doing that)

> > > Andy, can you please let us know in what order are the resources laid
> > > out in your reference BIOS for the case when it exports 5 communities
> > > (i.e. community-0-2, 4-5)?
> >
> > We have no hardware with such PCH to answer to this question. LP
> > version exports only 4 communities and order of them as per existing
> > driver.
>
> Got it, thanks. From your response earlier:
>
> > Looking to the above ASL code I may conclude that is definitely is *not* from our reference BIOS.
> > I have checked two versions of it and found that in both we have the following mapping:
> > for LP variant: there are only 4 communities are exported
> > for H variant: there are only 5 communities are exported
>
> 1) Do you know or recall, the order of communities in ACPI in the H
> variant? Of course, this is a request for help for getting the info.

I can't say anything about hardware I didn't see.
Even if BIOS code has something, it's not fully guaranteed that in
production it will be same. Better to ask Intel BIOS team for the
details.

> 2) Trying to understand how would the kernel support both LP and H
> variant. Is it the assumption that the H variant will have a different
> ACPI ID than the LP variant  (i.e. not "INT3455")? Because it will
> also have the same problem that we are seeing I think.

Yes, definitely they have different IDs. You may consult with
pinctrl-cannonlake.c driver (better from latest possible kernel
version like v4.19-rc5) to see how it's implemented.
Moreover, LP and H variants have different pin lists inside
communities, so, they can't be substituted and one may consider them
as different IPs.

-- 
With Best Regards,
Andy Shevchenko

Powered by blists - more mailing lists