Message ID | 1347206679-428-9-git-send-email-jcmvbkbc@gmail.com |
---|---|
State | New |
Headers | show |
Hi Max, Max Filippov wrote: > On Sun, Sep 9, 2012 at 8:16 PM, Peter Maydell > <peter.maydell@linaro.org> wrote: > > On 9 September 2012 17:04, Max Filippov <jcmvbkbc@gmail.com> wrote: > >> These are FP to integer and integer to FP conversion opcodes. > >> See ISA, 4.3.10 for more details. > >> > >> Note that utrunc.s implementation follows ISS behaviour, not ISA. > > > > ISS here means "instruction set simulator", right? Do you > > have any actual silicon you can check behaviour against? > > Basically there are three votes here (documentation, simulator > > and silicon) and QEMU should follow the majority opinion in > > the absence of any more official word. > > I have no silicon core with FP and I doubt that I can easily > access one. > IIUC Tensilica ISS core-specific code is autogenerated from > the hardware > description, without human intervention. > Looks like it's either documentation error or silicon error, probably > there's an erratum issued. Marc, can you please comment? If it's the Tensilica ISS, behavior indeed should match hardware, from a common description (am simplifying, but I expect here this to hold). It also depends on which release of hardware/ISS (do you know that exactly?). I've forwarded the question to someone at Tensilica who may know the answer, hopefully will have some answer during weekdays. Thanks, -Marc
On Sun, Sep 9, 2012 at 9:24 PM, Marc Gauthier <marc@tensilica.com> wrote: > Hi Max, > > Max Filippov wrote: >> On Sun, Sep 9, 2012 at 8:16 PM, Peter Maydell >> <peter.maydell@linaro.org> wrote: >> > On 9 September 2012 17:04, Max Filippov <jcmvbkbc@gmail.com> wrote: >> >> These are FP to integer and integer to FP conversion opcodes. >> >> See ISA, 4.3.10 for more details. >> >> >> >> Note that utrunc.s implementation follows ISS behaviour, not ISA. >> > >> > ISS here means "instruction set simulator", right? Do you >> > have any actual silicon you can check behaviour against? >> > Basically there are three votes here (documentation, simulator >> > and silicon) and QEMU should follow the majority opinion in >> > the absence of any more official word. >> >> I have no silicon core with FP and I doubt that I can easily >> access one. >> IIUC Tensilica ISS core-specific code is autogenerated from >> the hardware >> description, without human intervention. >> Looks like it's either documentation error or silicon error, probably >> there's an erratum issued. Marc, can you please comment? > > If it's the Tensilica ISS, behavior indeed should match hardware, > from a common description (am simplifying, but I expect here this > to hold). It also depends on which release of hardware/ISS > (do you know that exactly?). ISS is from the RC-2010.2-linux package, xt-run --help shows the following: Xtensa 8.0.2 Instruction Set Simulator core-isa.h for the core on which I run tests has the following: #define XCHAL_SW_VERSION 800002 /* sw version of this header */ #define XCHAL_CORE_ID "dsp3400_RC2" /* alphanum core name (CoreID) set in the Xtensa Processor Generator */ #define XCHAL_BUILD_UNIQUE_ID 0x0002DC22 /* 22-bit sw build ID */ /* * These definitions describe the hardware targeted by this software. */ #define XCHAL_HW_CONFIGID0 0xC3F3DBFE /* ConfigID hi 32 bits*/ #define XCHAL_HW_CONFIGID1 0x1082C3B0 /* ConfigID lo 32 bits*/ #define XCHAL_HW_VERSION_NAME "LX3.0.1" /* full version name */ #define XCHAL_HW_VERSION_MAJOR 2300 /* major ver# of targeted hw */ #define XCHAL_HW_VERSION_MINOR 1 /* minor ver# of targeted hw */ #define XCHAL_HW_VERSION 230001 /* major*100+minor */
On Sun, Sep 9, 2012 at 8:38 PM, Max Filippov <jcmvbkbc@gmail.com> wrote: > On Sun, Sep 9, 2012 at 8:16 PM, Peter Maydell <peter.maydell@linaro.org> wrote: >> On 9 September 2012 17:04, Max Filippov <jcmvbkbc@gmail.com> wrote: >>> These are FP to integer and integer to FP conversion opcodes. >>> See ISA, 4.3.10 for more details. >>> >>> Note that utrunc.s implementation follows ISS behaviour, not ISA. >> >> ISS here means "instruction set simulator", right? Do you >> have any actual silicon you can check behaviour against? >> Basically there are three votes here (documentation, simulator >> and silicon) and QEMU should follow the majority opinion in >> the absence of any more official word. Short summary for the answer that I've got from Tensilica is the following: - hardware and ISS behaviour must match, - documentation (ISA) is wrong and will be fixed. > I have no silicon core with FP and I doubt that I can easily access one. > IIUC Tensilica ISS core-specific code is autogenerated from the hardware > description, without human intervention. > Looks like it's either documentation error or silicon error, probably > there's an erratum issued. Marc, can you please comment?
diff --git a/target-xtensa/helper.h b/target-xtensa/helper.h index 4e6e417..9557347 100644 --- a/target-xtensa/helper.h +++ b/target-xtensa/helper.h @@ -44,5 +44,9 @@ DEF_HELPER_3(sub_s, f32, env, f32, f32) DEF_HELPER_3(mul_s, f32, env, f32, f32) DEF_HELPER_4(madd_s, f32, env, f32, f32, f32) DEF_HELPER_4(msub_s, f32, env, f32, f32, f32) +DEF_HELPER_FLAGS_3(ftoi, TCG_CALL_CONST | TCG_CALL_PURE, i32, f32, i32, i32) +DEF_HELPER_FLAGS_3(ftoui, TCG_CALL_CONST | TCG_CALL_PURE, i32, f32, i32, i32) +DEF_HELPER_3(itof, f32, env, i32, i32) +DEF_HELPER_3(uitof, f32, env, i32, i32) #include "def-helper.h" diff --git a/target-xtensa/op_helper.c b/target-xtensa/op_helper.c index ba935a8..5cf9c02 100644 --- a/target-xtensa/op_helper.c +++ b/target-xtensa/op_helper.c @@ -821,3 +821,40 @@ float32 HELPER(msub_s)(CPUXtensaState *env, float32 a, float32 b, float32 c) return float32_muladd(b, c, a, float_muladd_negate_product, &env->fp_status); } + +uint32_t HELPER(ftoi)(float32 v, uint32_t rounding_mode, uint32_t scale) +{ + float_status fp_status = {0}; + + set_float_rounding_mode(rounding_mode, &fp_status); + return float32_to_int32( + float32_scalbn(v, scale, &fp_status), &fp_status); +} + +uint32_t HELPER(ftoui)(float32 v, uint32_t rounding_mode, uint32_t scale) +{ + float_status fp_status = {0}; + float32 res; + + set_float_rounding_mode(rounding_mode, &fp_status); + + res = float32_scalbn(v, scale, &fp_status); + + if (float32_is_neg(v) && !float32_is_any_nan(v)) { + return float32_to_int32(res, &fp_status); + } else { + return float32_to_uint32(res, &fp_status); + } +} + +float32 HELPER(itof)(CPUXtensaState *env, uint32_t v, uint32_t scale) +{ + return float32_scalbn(int32_to_float32(v, &env->fp_status), + (int32_t)scale, &env->fp_status); +} + +float32 HELPER(uitof)(CPUXtensaState *env, uint32_t v, uint32_t scale) +{ + return float32_scalbn(uint32_to_float32(v, &env->fp_status), + (int32_t)scale, &env->fp_status); +} diff --git a/target-xtensa/translate.c b/target-xtensa/translate.c index a2ce286..fabde4f 100644 --- a/target-xtensa/translate.c +++ b/target-xtensa/translate.c @@ -1915,6 +1915,54 @@ static void disas_xtensa_insn(DisasContext *dc) cpu_FR[RRR_R], cpu_FR[RRR_S], cpu_FR[RRR_T]); break; + case 8: /*ROUND.Sf*/ + case 9: /*TRUNC.Sf*/ + case 10: /*FLOOR.Sf*/ + case 11: /*CEIL.Sf*/ + case 14: /*UTRUNC.Sf*/ + gen_window_check1(dc, RRR_R); + { + static const unsigned rounding_mode_const[] = { + float_round_nearest_even, + float_round_to_zero, + float_round_down, + float_round_up, + [6] = float_round_to_zero, + }; + TCGv_i32 rounding_mode = tcg_const_i32( + rounding_mode_const[OP2 & 7]); + TCGv_i32 scale = tcg_const_i32(RRR_T); + + if (OP2 == 14) { + gen_helper_ftoui(cpu_R[RRR_R], cpu_FR[RRR_S], + rounding_mode, scale); + } else { + gen_helper_ftoi(cpu_R[RRR_R], cpu_FR[RRR_S], + rounding_mode, scale); + } + + tcg_temp_free(rounding_mode); + tcg_temp_free(scale); + } + break; + + case 12: /*FLOAT.Sf*/ + case 13: /*UFLOAT.Sf*/ + gen_window_check1(dc, RRR_S); + { + TCGv_i32 scale = tcg_const_i32(-RRR_T); + + if (OP2 == 13) { + gen_helper_uitof(cpu_FR[RRR_R], cpu_env, + cpu_R[RRR_S], scale); + } else { + gen_helper_itof(cpu_FR[RRR_R], cpu_env, + cpu_R[RRR_S], scale); + } + tcg_temp_free(scale); + } + break; + case 15: /*FP1OP*/ switch (RRR_T) { case 0: /*MOV.Sf*/
These are FP to integer and integer to FP conversion opcodes. See ISA, 4.3.10 for more details. Note that utrunc.s implementation follows ISS behaviour, not ISA. Signed-off-by: Max Filippov <jcmvbkbc@gmail.com> --- Changes v1->v2: - replace float32_mul/div in FP-to-integer convertors with float32_scalbn - reimplement comparison with zero in FP-to-integer convertors target-xtensa/helper.h | 4 +++ target-xtensa/op_helper.c | 37 ++++++++++++++++++++++++++++++++++ target-xtensa/translate.c | 48 +++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 89 insertions(+), 0 deletions(-)