All of lore.kernel.org
 help / color / mirror / Atom feed
* [Buildroot] BR2_ARCH_IS_64 NEON support
@ 2021-11-15 21:46 Tim Harvey
  2021-11-15 22:36 ` Thomas Petazzoni
  0 siblings, 1 reply; 2+ messages in thread
From: Tim Harvey @ 2021-11-15 21:46 UTC (permalink / raw)
  To: buildroot

Greetings,

I'm curious why NEON is not available for BR2_ARCH_IS_64?

I'm working with boards with IMX8M SoC's which are arm64 and do have
NEON support (per the NXP literature and I've run binaries using NEON
on IMX8M). Is this a simple oversight or is there something wrong with
my understanding of this?

Best regards,

Tim
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: [Buildroot] BR2_ARCH_IS_64 NEON support
  2021-11-15 21:46 [Buildroot] BR2_ARCH_IS_64 NEON support Tim Harvey
@ 2021-11-15 22:36 ` Thomas Petazzoni
  0 siblings, 0 replies; 2+ messages in thread
From: Thomas Petazzoni @ 2021-11-15 22:36 UTC (permalink / raw)
  To: Tim Harvey; +Cc: buildroot

Hello Tim,

On Mon, 15 Nov 2021 13:46:40 -0800
Tim Harvey <tharvey@gateworks.com> wrote:

> I'm curious why NEON is not available for BR2_ARCH_IS_64?
> 
> I'm working with boards with IMX8M SoC's which are arm64 and do have
> NEON support (per the NXP literature and I've run binaries using NEON
> on IMX8M). Is this a simple oversight or is there something wrong with
> my understanding of this?

This is explained in commit 0976cd6cd64a128a2ac921e4f35f0e7cbe306692,
which states:

commit 0976cd6cd64a128a2ac921e4f35f0e7cbe306692
Author: Peter Korsgaard <peter@korsgaard.com>
Date:   Wed Dec 7 10:25:21 2016 +0100

    arch/Config.in.arm: only enable BR2_ARM_CPU_HAS_NEON for ARMv8 in 32bit mode
    
    A number of packages use BR2_ARM_CPU_HAS_NEON to know if the target handles
    aarch32 neon instructions, which is only true for ARMv8 cores when they are
    running in 32bit mode.
    
    Notice: These cores do support neon-like instructions using a different
    encoding in 64bit mode (it is a required part of ARMv8, similar to the FPU).
    
    Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
    Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

However, it does mean that we need to improve the situation, to also
properly support NEON for ARM 64-bit platforms, exactly as you say.

I'm not sure if this means:

 * Fixing packages to use BR2_ARM_CPU_HAS_NEON && !BR2_ARCH_IS_64 when
   they use aarch32 neon instructions

 * Separating BR2_ARM_CPU_HAS_NEON into BR2_ARM_CPU_HAS_NEON32 and
   BR2_ARM_CPU_HAS_NEON64 ?

 * Something else?

Thomas
-- 
Thomas Petazzoni, co-owner and CEO, Bootlin
Embedded Linux and Kernel engineering and training
https://bootlin.com
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2021-11-15 22:37 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2021-11-15 21:46 [Buildroot] BR2_ARCH_IS_64 NEON support Tim Harvey
2021-11-15 22:36 ` Thomas Petazzoni

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.