diff options
| author | Petar Jovanovic <petar.jovanovic@imgtec.com> | 2013-04-15 13:54:07 +0200 |
|---|---|---|
| committer | Aurelien Jarno <aurelien@aurel32.net> | 2013-04-15 16:07:57 +0200 |
| commit | b1ca31d7ce648a10e3513bb4b5e7f89a8702aec7 (patch) | |
| tree | 6afd4ed29078d01fc1253ac0b78a3f81de016527 /target-mips | |
| parent | e3d142d073d02f0a3a4aad79eb838c15b6f99c01 (diff) | |
| download | qemu-b1ca31d7ce648a10e3513bb4b5e7f89a8702aec7.zip qemu-b1ca31d7ce648a10e3513bb4b5e7f89a8702aec7.tar.gz qemu-b1ca31d7ce648a10e3513bb4b5e7f89a8702aec7.tar.bz2 | |
target-mips: fix mipsdsp_mul_q15_q15 and tests for MAQ_SA_W_PHL/PHR
The operands for MAQ_SA_W.PHL/MAQ_SA_W.PHR must in specified format.
Otherwise, the results are unpredictable. Once the operands were corrected
in the tests (part of this change), a bug in mipsdsp_mul_q15_q15 became
visible.
This change corrects the tests for MAQ_SA_W.PHL/MAQ_SA_W.PHR and fixes
sign-related issue in mipsdsp_mul_q15_q15. It also removes unnecessary
comment.
Signed-off-by: Petar Jovanovic <petar.jovanovic@imgtec.com>
Diffstat (limited to 'target-mips')
| -rw-r--r-- | target-mips/dsp_helper.c | 14 |
1 files changed, 1 insertions, 13 deletions
diff --git a/target-mips/dsp_helper.c b/target-mips/dsp_helper.c index c7df595..f975da0 100644 --- a/target-mips/dsp_helper.c +++ b/target-mips/dsp_helper.c @@ -269,18 +269,6 @@ static inline int32_t mipsdsp_sat32_acc_q31(int32_t acc, int32_t a, temp31 = (temp_sum >> 31) & 0x01; result = temp_sum & 0xFFFFFFFF; - /* FIXME - This sat function may wrong, because user manual wrote: - temp127..0 ← temp + ( (signA) || a31..0 - if ( temp32 ≠ temp31 ) then - if ( temp32 = 0 ) then - temp31..0 ← 0x80000000 - else - temp31..0 ← 0x7FFFFFFF - endif - DSPControlouflag:16+acc ← 1 - endif - */ if (temp32 != temp31) { if (temp32 == 0) { result = 0x7FFFFFFF; @@ -578,7 +566,7 @@ static inline int32_t mipsdsp_mul_q15_q15(int32_t ac, uint16_t a, uint16_t b, temp = 0x7FFFFFFF; set_DSPControl_overflow_flag(1, 16 + ac, env); } else { - temp = ((uint32_t)a * (uint32_t)b) << 1; + temp = ((int16_t)a * (int16_t)b) << 1; } return temp; |