ack/mach/mips/libem/c_uf_i.s

#
.sect .text
.sect .rom
.sect .data
.sect .bss

.sect .text
.define .c_uf_i
.c_uf_i:
	/* Input: f0
	 * Output: r2
	 * Only at and f30/f31 may be used.
	 */

	lui at, ha16[.ff_80000000]
	lwc1 f30, lo16[.ff_80000000] (at)
	c.le.s 0, f30, f0
	bc1t toobig
	nop

	trunc.w.s f0, f0
	mfc1 r2, f0
	jr ra
	nop

toobig:
	sub.s f0, f0, f30
	trunc.w.s f0, f0
	mfc1 r2, f0
	lui at, 0x8000			! load 0x80000000
	addu r2, r2, at
	jr ra
	nop

.sect .rom
.ff_80000000:
	.dataf4 2147483648.0
MIPS appears to hate converting unsigneds to floats and vice versa. 2018-09-05 21:53:38 +00:00			`#`
			`.sect .text`
			`.sect .rom`
			`.sect .data`
			`.sect .bss`

Added hi[] and lo[] functions to the assembler; fix things enough that we now get to the point of failing to link programs. 2018-09-08 20:42:06 +00:00			`.sect .text`
Rework floating point conversion. Bash enough of the table into workingness that the libraries build now. 2018-09-08 20:06:38 +00:00			`.define .c_uf_i`
			`.c_uf_i:`
MIPS appears to hate converting unsigneds to floats and vice versa. 2018-09-05 21:53:38 +00:00			`/* Input: f0`
			`* Output: r2`
Yikes! Turns out that FPU registers are only 32 bits wide, and doubles are stored in pairs, just like on the PowerPC! 2018-09-09 16:25:00 +00:00			`* Only at and f30/f31 may be used.`
MIPS appears to hate converting unsigneds to floats and vice versa. 2018-09-05 21:53:38 +00:00			`*/`
Added hi[] and lo[] functions to the assembler; fix things enough that we now get to the point of failing to link programs. 2018-09-08 20:42:06 +00:00
Make sure li works properly when loading fixed up values which end up as 0. Use lui rather than li to load high values. 2018-09-12 20:28:40 +00:00			`lui at, ha16[.ff_80000000]`
Yikes! Turns out that FPU registers are only 32 bits wide, and doubles are stored in pairs, just like on the PowerPC! 2018-09-09 16:25:00 +00:00			`lwc1 f30, lo16[.ff_80000000] (at)`
			`c.le.s 0, f30, f0`
MIPS appears to hate converting unsigneds to floats and vice versa. 2018-09-05 21:53:38 +00:00			`bc1t toobig`
			`nop`

Added hi[] and lo[] functions to the assembler; fix things enough that we now get to the point of failing to link programs. 2018-09-08 20:42:06 +00:00			`trunc.w.s f0, f0`
MIPS appears to hate converting unsigneds to floats and vice versa. 2018-09-05 21:53:38 +00:00			`mfc1 r2, f0`
			`jr ra`
			`nop`

			`toobig:`
Yikes! Turns out that FPU registers are only 32 bits wide, and doubles are stored in pairs, just like on the PowerPC! 2018-09-09 16:25:00 +00:00			`sub.s f0, f0, f30`
Added hi[] and lo[] functions to the assembler; fix things enough that we now get to the point of failing to link programs. 2018-09-08 20:42:06 +00:00			`trunc.w.s f0, f0`
MIPS appears to hate converting unsigneds to floats and vice versa. 2018-09-05 21:53:38 +00:00			`mfc1 r2, f0`
Converting floats and doubles to integers now works, as much as these ever do on MIPS; turns out that it can't (or at least can't in qemu) reliably turn INT_MIN from a double to an int. 2018-09-12 21:58:48 +00:00			`lui at, 0x8000 ! load 0x80000000`
			`addu r2, r2, at`
MIPS appears to hate converting unsigneds to floats and vice versa. 2018-09-05 21:53:38 +00:00			`jr ra`
			`nop`

Added some more incredibly untested helper functions. 2018-09-09 22:21:16 +00:00			`.sect .rom`
			`.ff_80000000:`
Add the .dataf4 and .dataf8 directives to the assembler --- manually converting floats and doubles to bytes is not fun. It might even work! 2018-09-10 20:56:18 +00:00			`.dataf4 2147483648.0`
Added some more incredibly untested helper functions. 2018-09-09 22:21:16 +00:00