All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Andreas Färber" <afaerber@suse.de>
To: u-boot@lists.denx.de
Subject: [U-Boot] [PATCH] ARM: bcm2835: config.mk isn't needed
Date: Thu, 30 Jan 2014 08:03:54 +0100	[thread overview]
Message-ID: <52E9F95A.4080004@suse.de> (raw)
In-Reply-To: <52E95ABA.2070502@ti.com>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Am 29.01.2014 20:47, schrieb Tom Rini:
> On 01/29/2014 02:41 PM, Andreas F?rber wrote:
>> Am 29.01.2014 19:38, schrieb Stephen Warren:
>>> On 01/29/2014 02:22 AM, Andreas F?rber wrote:
>>>> Am 29.01.2014 06:42, schrieb Stephen Warren:
>>>>> The entries in config.mk were needed so that U-Boot could
>>>>> be built with an old version of the Raspberry Pi
>>>>> Foundation's toolchain. Without them, the build would error
>>>>> out with:
>>>>> 
>>>>> ...-ld: error: .../libgcc.a(_bswapsi2.o) uses VFP register 
>>>>> arguments, u-boot does not
>>>>> 
>>>>> However, none of the 3 toolchains in the latest version of 
>>>>> their tools.git, nor the Ubuntu/Linaro ARM compilers in at 
>>>>> least Ubuntu Quantal or Saucy, need these options set in 
>>>>> order to compile a working U-Boot. Hence, remove the
>>>>> options for simplicity.
>>>>> 
>>>>> Reported-by: Tom Rini <trini@ti.com> Signed-off-by:
>>>>> Stephen Warren <swarren@wwwdotorg.org> --- 
>>>>> arch/arm/cpu/arm1176/bcm2835/config.mk | 19 
>>>>> ------------------- 1 file changed, 19 deletions(-) delete 
>>>>> mode 100644 arch/arm/cpu/arm1176/bcm2835/config.mk
>>>> 
>>>> No objection against the patch, but why was
>>>> arch/arm/config.mk using -msoft-float in the first place?
>>> 
>>> It wasn't. It was filtering -msoft-float out of the compiler 
>>> flags, which then left the ABI up to the toolchain default,
>>> which was hardfp for one of the Raspberry Pi Foundation's
>>> toolchains - the same one that caused the error messages quoted
>>> above unless -msoft-float was removed from the compiler
>>> options.
> 
>> You seem to have misread my question, it was not about 
>> bcm2835/config.mk but about arm/config.mk, which bcm2835 was 
>> filtering.
> 
> No, I know what you said.

Tom, I didn't dispute that. But Stephen's statement above was clearly
wrong if you look at v2013.10:

http://git.denx.de/?p=u-boot.git;a=blob;f=arch/arm/config.mk;h=bdabcf407e2c7cf783e4e73b05acec2031d2189c;hb=183acb700378a8cfc5d50a01a65de93fb2c24586#l19

It was in fact using -msoft-float, as you confirm below, and bcm2835
was the one filtering it out.

In v2014.01 -msoft-float gets compile-tested before usage and
therefore the bcm2835-specific filtering here can be dropped.

>> I didn't quite get Tom's response about U-Boot being soft-float.
>> It is what I compile it to be, no? If I compile it with an
>> hardfp toolchain such as openSUSE's, without -msoft-float it is
>> going to be hardfp IIUC. softfp is for compatibility with non-VFP
>> hardware I thought, but U-Boot is clearly built with a per-board
>> config, so it's not portable anyway.
> 
> This isn't quite correct.  U-Boot is like the kernel.  We say how 
> we're built and like the kernel, we are soft float (and unless the 
> board says we need it, we aren't thumb/thumb2 either).  That's why 
> arch/arm/config.mk (like arch/arm/Makefile in the kernel) enforces 
> -msoft-float.

OK, thanks for explaining. Our v2013.10 package was carrying a patch
dropping the hardcoded -msoft-float for our armv7hl toolchain and it
worked okay as far as I could tell. I've dropped that patch now for
v2014.01 and things at least still build for me. Question solved and FWIW

Reviewed-by: Andreas F?rber <afaerber@suse.de>

Cheers,
Andreas

- -- 
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 N?rnberg, Germany
GF: Jeff Hawn, Jennifer Guild, Felix Imend?rffer; HRB 16746 AG N?rnberg
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.22 (GNU/Linux)
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=Bgtl
-----END PGP SIGNATURE-----

  reply	other threads:[~2014-01-30  7:03 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-01-29  5:42 [U-Boot] [PATCH] ARM: bcm2835: config.mk isn't needed Stephen Warren
2014-01-29  9:22 ` Andreas Färber
2014-01-29 17:32   ` Tom Rini
2014-01-29 18:38   ` Stephen Warren
2014-01-29 19:41     ` Andreas Färber
2014-01-29 19:47       ` Tom Rini
2014-01-30  7:03         ` Andreas Färber [this message]
2014-02-05  2:19 ` Stephen Warren
2014-02-13 14:07 ` Albert ARIBAUD

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=52E9F95A.4080004@suse.de \
    --to=afaerber@suse.de \
    --cc=u-boot@lists.denx.de \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
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.