- User Since
- Jan 7 2014, 10:13 PM (176 w, 6 d)
Sun, May 28
Sat, May 27
Fri, May 26
Thu, May 25
Wed, May 24
This patch looks fine to me, but it would be nice to add some #ifdef __FreeBSD_version parts so it can be upstreamed.
It would be nice to have some tests, even if they are rudimentary... :)
Mon, May 22
Sun, May 21
Using -target mips-freebsd, the assembly produced for e.g. bswapsi2.c is:
__bswapsi2: # @__bswapsi2 .frame $fp,8,$ra .mask 0x40000000,-4 .fmask 0x00000000,0 .set noreorder .set nomacro .set noat # BB#0: # %entry addiu $sp, $sp, -8 sw $fp, 4($sp) # 4-byte Folded Spill move $fp, $sp wsbh $1, $4 rotr $2, $1, 16 move $sp, $fp lw $fp, 4($sp) # 4-byte Folded Reload jr $ra addiu $sp, $sp, 8 .set at .set macro .set reorder .end __bswapsi2 $func_end0: .size __bswapsi2, ($func_end0)-__bswapsi2
I'm not a mips expert, but it looks pretty OK to me... :)
LGTM, we should upstream the contrib part, though. Unfortunately llvm.org is down at the moment...
Let's use D10840 instead.
No, not like this, please. The experimental sources are unstable, and there is no guarantee whatsoever that anything in there will stay as it is. This is also why I never added them to our libc++.so, to discourage people from actually using them. I have actively tried to avoid adding the experimental features for now, as I don't think these have a place in our base system. But if there really is some use for them, I will not stand in the way, of course.