Implement more aeabi float functions

[andrew]

Summary

Implement the aeabi floating point functions that return their result
using the PSR Z and C flags.

[andrew]

Fix a few typos and add __aeabi_c[df]cmpeq to Symbol.map

[andrew]

This should fix the build of print/tex-luatex

[andrew]

The *_soft functions don't exist, call the correct soft float function

[sbruno]

starting test build with head + this patch.

[andrew]

The float{32,64}_lt symbols have macros to rename them

[andrew]

Rework the *eq functions to check if we are passed NaNs first

[andrew]

lib/libc/arm/aeabi/aeabi_asm_double.S

106

This should be __aeabi_cdcmpeq_helper.

[rpaulo]

I've reviewed this, but I think it needs to be reviewed by someone else more experienced in VFP.

lib/libc/arm/aeabi/aeabi_asm_double.S

31

It would be great if you could move these to a common header.

[andrew]

Fix the name of the double helper

[sbruno]

This is looking very good at the moment on my tests. http://chips.ysv.freebsd.org/build.html?mastername=11armv6-default&build=2014-10-05_21h42m23s

Is this committable yet?

[sbruno]

I guess the only way this is getting reviewed is by pushing it into base. :-)

[andrew]

Yes, I've asked on two lists now with no feedback. It's likely I'll commit it tomorrow if I get nothing by then.

[andrew]

Committed as r273088