Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Thu, 28 Jul 2022 00:22:05 +0200
From:   Petr Vorel <pvorel@suse.cz>
To:     Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc:     linux-kernel@vger.kernel.org,
        "Rafael J. Wysocki" <rafael@kernel.org>,
        David Sterba <dsterba@suse.com>
Subject: Re: [PATCH 1/1] drivers/base/cpu: Print kernel arch
Message-ID: <YuG6jfeT9p47Ekn1@pevik>
Reply-To: Petr Vorel <pvorel@suse.cz>
References: <20220727161135.24531-1-pvorel@suse.cz>
 <YuFqh6OrWEQsZV04@kroah.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <YuFqh6OrWEQsZV04@kroah.com>
Precedence: bulk

Hi Greg,

thanks for your review!

> On Wed, Jul 27, 2022 at 06:11:35PM +0200, Petr Vorel wrote:
> > Print kernel architecture in /sys/devices/system/cpu/arch
> > using UTS_MACHINE, i.e. member of struct uts_namespace.machine.

> > This helps people who debug kernel with initramfs with minimal
> > environment (i.e. without coreutils or even busybox) or allow to open
> > sysfs file instead of run uname -m in high level languages.

> > Signed-off-by: Petr Vorel <pvorel@suse.cz>
> > ---
> >  drivers/base/cpu.c | 9 +++++++++
> >  1 file changed, 9 insertions(+)

> You can't add a new sysfs file without a Documentation/ABI/ update as
> well.  Please fix that up.
I'm sorry. Yes, I realized it later on (once I got offline).
Sure, I'll fix this in v2. But the main question is whether this feature is
acceptable and what is the best place for the file.


> > diff --git a/drivers/base/cpu.c b/drivers/base/cpu.c
> > index 4c98849577d4..7c8032e3ff10 100644
> > --- a/drivers/base/cpu.c
> > +++ b/drivers/base/cpu.c
> > @@ -3,6 +3,7 @@
> >   * CPU subsystem support
> >   */

> > +#include <generated/compile.h>
> >  #include <linux/kernel.h>
> >  #include <linux/module.h>
> >  #include <linux/init.h>
> > @@ -232,6 +233,13 @@ static ssize_t print_cpus_kernel_max(struct device *dev,
> >  }
> >  static DEVICE_ATTR(kernel_max, 0444, print_cpus_kernel_max, NULL);

> > +static ssize_t print_cpus_arch(struct device *dev,
> > +				     struct device_attribute *attr, char *buf)
> > +{
> > +	return sysfs_emit(buf, "%s\n", UTS_MACHINE);
> > +}
> > +static DEVICE_ATTR(arch, 0444, print_cpus_arch, NULL);

> why just UTS_MACHINE?  Doesn't 'uname' show this already?  And I thought
> this was in /proc/cpuinfo but odd, it isn't...
Sure, this info is in uname(). But for certain cases is really easier to read
file from /proc or /sys than run create custom C bindings or a binary which
calls uname(2) libc wrapper or run uname binary via execve(2).

Yes, I also expected /proc/cpuinfo would have it but it does not have it. I
don't think it's a good idea to add 'arch  : foo' line to it, but I can do if
there is consensus that it's the best place.

> Also what about the other things in compile.h?
UTS_VERSION is in /proc/version.
LINUX_COMPILE_BY (e.g. "user"), LINUX_COMPILE_HOST (e.g. "host") and
LINUX_COMPILER (e.g. "aarch64-alpine-linux-musl-gcc (Alpine 11.2.1_git20220219)
...") are IMHO useless, but I can add it if you wish.

Well, there is hostname in /proc/sys/kernel/hostname, there are also ostype and
osrelease.. Thinking about it twice /proc/sys/kernel/ looks to me a better place
for arch file than current /sys/devices/system/cpu/. WDYT?

Kind regards,
Petr

> thanks,

> greg k-h