Re: [PATCH] Upstreaming a large amount of patches

Christian Seiler <[email protected]> Tue, 29 Mar 2016 19:43:31 +0200
Newsgroups gmane.linux.lib.dietlibc
Message-ID <[email protected]>
Hi Felix,

On 03/29/2016 06:06 PM, Felix von Leitner wrote:
> I just imported a ton of your patches. Thanks for your work!

Unfortunately, you made a mistake that broke ARMv7 builds:

arm/__longjmp.S:
-# if __ARM_ARCH__ == 7
+# if __ARM_ARCH__ == 6

If you look at my patch, __longjmp.S explicitly replaced the check
for __ARM_ARCH__ == 7 with a check against >= 6, but you now just
replaced it by == 6, which is wrong.

Funnily enough, in setjump.S, the check was already >= 7, so changing
that to a 6 wasn't a problem there, just in __longjmp.S.

(My patch, for comparison:)
https://gist.githubusercontent.com/chris-se/e35381c75583a4c4bce2/raw/8bf327a048f396220c48fb4ffbd504d12b042007/porting-07-ARMv6-VFPU.patch

(I did like how you greatly simplified the -fno-strict-aliasing
and the _newselect stuff, that's _much_ more elegant than what the
original patches did.)

Thanks again for applying this all.

Regards,
Christian