Re: [PATCH] powerpc: emulate power5 popcntb instruction
From: Will Schmidt <hidden>
Date: 2006-08-18 18:11:53
Subsystem:
linux for powerpc (32-bit and 64-bit), the rest · Maintainers:
Madhavan Srinivasan, Michael Ellerman, Linus Torvalds
In an attempt to make it easier for a power5 optimized app to run on a power4 or a 970 or random earlier machine, this provides emulation of the popcntb instruction. Rewritten to use a slicker algorithm as suggested by Segher. I left a 'tmp' variable in play, as it seemed cleaner to use tmp than referring to regs->gpr[rs] and [ra] multiple times within the magic algorithm. Also tested on power4 with both 32 and 64 userspace this time. *blush*. :-) Signed-off-by: Will Schmidt <redacted> ---
diff --git a/arch/powerpc/kernel/traps.c b/arch/powerpc/kernel/traps.c
index 2105767..a0e80dd 100644
--- a/arch/powerpc/kernel/traps.c
+++ b/arch/powerpc/kernel/traps.c@@ -588,6 +588,8 @@ #define INST_LSWX 0x7c00042a #define INST_STSWI 0x7c0005aa #define INST_STSWX 0x7c00052a +#define INST_POPCNTB 0x7c0000f4 + static int emulate_string_inst(struct pt_regs *regs, u32 instword) { u8 rT = (instword >> 21) & 0x1f;
@@ -656,6 +658,23 @@ static int emulate_string_inst(struct pt return 0; } +static int emulate_popcntb_inst(struct pt_regs *regs, u32 instword) +{ + u32 ra,rs; + unsigned long tmp; + + ra = (instword >> 16) & 0x1f; + rs = (instword >> 21) & 0x1f; + + tmp = regs->gpr[rs]; + tmp = tmp - ((tmp >> 1) & 0x5555555555555555); + tmp = (tmp & 0x3333333333333333) + ((tmp >> 2) & 0x3333333333333333); + tmp = (tmp + (tmp >> 4)) & 0x0f0f0f0f0f0f0f0f; + regs->gpr[ra] = tmp; + + return 0; +} + static int emulate_instruction(struct pt_regs *regs) { u32 instword;
@@ -693,6 +712,11 @@ static int emulate_instruction(struct pt if ((instword & INST_STRING_GEN_MASK) == INST_STRING) return emulate_string_inst(regs, instword); + /* Emulate the popcntb (Population Count Bytes) instruction. */ + if ((instword & INST_POPCNTB) == INST_POPCNTB) { + return emulate_popcntb_inst(regs, instword); + } + return -EINVAL; }