[PATCH 1/1] ARM: thumb: Have the machine name indicate operation in thumb mode.

From: dave.martin@linaro.org (Dave Martin)
To: linux-arm-kernel@lists.infradead.org
Subject: [PATCH 1/1] ARM: thumb: Have the machine name indicate operation in thumb mode.
Date: Thu, 26 May 2011 14:56:04 +0100	[thread overview]
Message-ID: <20110526135604.GA2119@arm.com> (raw)
In-Reply-To: <20110526130104.GA5762@n2100.arm.linux.org.uk>

On Thu, May 26, 2011 at 02:01:04PM +0100, Russell King - ARM Linux wrote:
> On Wed, May 25, 2011 at 03:08:05PM -0700, Vadim Bendebury (??) wrote:
> > On Wed, May 18, 2011 at 3:57 PM, Russell King - ARM Linux <
> > linux at arm.linux.org.uk> wrote:
> > > In which case I think we'll have to talk to kbuild people to get
> > > mkcompile_h modified.  Also maybe asking what other architectures
> > > do for this kind of problem may be a good idea - especially before
> > > we start abusing and hacking stuff like the hardware name.
> > 
> > So, what would be an equivalent in other architectures to ask about - it
> > seems that the ability to run in Thumb/ARM mode is fairly unique for ARM.
> 
> What about 32-bit vs 64-bit x86 ?

Thumb seems different because is is irrelevant to the kernel/user ABI
(or, if it's not irrelevant, we have a bug somewhere...)

Documenting what ISA the kernel was built with is therefore similar to
documenting what optimisation options were used -- i.e., it's interesting
and occasionally useful for humans, but not something software should
normally be querying or care about in any way.

> > Consulting /boot/config file is not reliable: in my own setup I can netboot
> > either thumb or ARM kernel without changing the root file system contents. I
> > agree though that changing the hardware name is not the best way to do this
> > either.
> > 
> > I could look into adding 'thumb' to any element of the uname output, or as a
> > watered down version - just add a line to dmesg output?
> 
> The thing that I worry about is that userspace will start using this
> information inappropriately.  Whether the kernel is built as T2 or ARM
> is really only a matter for the kernel itself (and people debugging the
> kernel), rather than random userspace programs.

Agreed

> 
> Maybe a way forward would be to produce a patch for mkcompile_h, post it
> to here, kbuild people and LKML and see what response is forthcoming.
> 
> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel at lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel