linux-kernel - Re: [PATCH v6 1/2] of: Documentation: Specify local APIC ID in "reg"

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:   Tue, 20 Mar 2018 11:10:18 -0700
From:   Ivan Gorinov <ivan.gorinov@...el.com>
To:     Rob Herring <robh+dt@...nel.org>
Cc:     Thomas Gleixner <tglx@...utronix.de>,
        Frank Rowand <frowand.list@...il.com>,
        Andy Shevchenko <andy.shevchenko@...il.com>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        Ingo Molnar <mingo@...hat.com>,
        Mark Rutland <mark.rutland@....com>
Subject: Re: [PATCH v6 1/2] of: Documentation: Specify local APIC ID in "reg"

On Mon, Mar 19, 2018 at 07:39:52PM -0500, Rob Herring wrote:
> On Tue, Mar 13, 2018 at 5:05 PM, Ivan Gorinov <ivan.gorinov@...el.com> wrote:
> > Use the "reg" property to specify the processor's local APIC ID.
> > Local APIC ID is assigned by hardware and may differ from CPU number.
> 
> Is "CPU number" a s/w visible h/w number or has it just been an index
> for DT? In the latter case, I'm okay with this change. In the former
> case, you should stick to the existing numbering. For example on ARM,
> the number here corresponds to a core ID number in a register called
> MPIDR.

The latter case. Apparently, "CPU number" was just an index in the list.
Local APIC ID is the s/w visible h/w assigned number.
Some processor models allow local APIC ID to be changed by software, but
CPUID instruction executed with %eax = 0x0b always returns the initial ID
assigned by hardware in %edx.

APIC ID does not match index in the list in many systems.

> >
> > Signed-off-by: Ivan Gorinov <ivan.gorinov@...el.com>
> > ---
> >  Documentation/devicetree/bindings/x86/ce4100.txt | 37 ++++++++++++++++++------
> >  1 file changed, 28 insertions(+), 9 deletions(-)
> >
> > diff --git a/Documentation/devicetree/bindings/x86/ce4100.txt b/Documentation/devicetree/bindings/x86/ce4100.txt
> > index b49ae59..1c41cbd 100644
> > --- a/Documentation/devicetree/bindings/x86/ce4100.txt
> > +++ b/Documentation/devicetree/bindings/x86/ce4100.txt
> > @@ -7,17 +7,36 @@ Many of the "generic" devices like HPET or IO APIC have the ce4100
> >  name in their compatible property because they first appeared in this
> >  SoC.
> >
> > -The CPU node
> > -------------
> > -       cpu@0 {
> > -               device_type = "cpu";
> > -               compatible = "intel,ce4100";
> > -               reg = <0>;
> > -               lapic = <&lapic0>;
> > +The CPU nodes
> > +-------------
> > +
> > +       cpus {
> > +               #address-cells = <1>;
> > +               #size-cells = <0>;
> > +
> > +               cpu@...0 {
> 
> Drop the '0x' and leading 0s.
> 
> > +                       device_type = "cpu";
> > +                       compatible = "intel,ce4100";
> > +                       reg = <0x00>;
> > +               };
> > +
> > +               cpu@...2 {
> > +                       device_type = "cpu";
> > +                       compatible = "intel,ce4100";
> > +                       reg = <0x02>;
> > +               };
> >         };
> >
> > -The reg property describes the CPU number. The lapic property points to
> > -the local APIC timer.
> > +A "cpu" node describes one logical processor (hardware thread).
> > +
> > +Required properties:
> > +
> > +- device_type
> > +       Device type, must be "cpu".
> > +
> > +- reg
> > +       Local APIC ID, the unique number assigned to each processor by
> > +       system hardware.
> >
> >  The SoC node
> >  ------------
> > --
> > 2.7.4
> >