Clean up register constraints.
authorTimothy B. Terriberry <tterribe@xiph.org>
Mon, 20 May 2013 22:29:04 +0000 (15:29 -0700)
committerTimothy B. Terriberry <tterribe@xiph.org>
Tue, 21 May 2013 06:15:15 +0000 (23:15 -0700)
commitb518b56fe11bf53f88fe30d57ea9d668337983a9
tree1cf7d43f65b1122942d5175df40b01c78ac31bc9
parent9880c4cdebf7e4db5616546e801749d36fdd7202
Clean up register constraints.

http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.dui0068b/CIHBJEHG.html
 says that "Rd cannot be the same as Rm."
http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.dui0068b/CIHBJEHG.html
 says that "RdLo, RdHi, and Rm must all be different registers."
This means that some of the early clobbers I removed really should
 have been there (to prevent aliasing Rd, RdLo, or RdHi with Rm).
It also means that we should reverse some of the operands in the
 FFT's complex multiplies.
This should only affect the ARMv4 optimizations.

Thanks to Nils Wallménius for the report.

While we're here, audit the commutative pair flags again, since I
 screwed up at least one of them, and eliminate some dead code.
celt/_kiss_fft_guts.h
celt/fixed_armv4.h
celt/fixed_armv5e.h
silk/SigProc_FIX_armv4.h
silk/SigProc_FIX_armv5e.h
silk/macros_armv4.h
silk/macros_armv5e.h