From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:58953) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1WLhoZ-0001IX-SS for qemu-devel@nongnu.org; Thu, 06 Mar 2014 18:35:38 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1WLhnz-0003Pw-4d for qemu-devel@nongnu.org; Thu, 06 Mar 2014 18:34:59 -0500 From: Alexander Graf Date: Fri, 7 Mar 2014 00:33:55 +0100 Message-Id: <1394148857-19607-109-git-send-email-agraf@suse.de> In-Reply-To: <1394148857-19607-1-git-send-email-agraf@suse.de> References: <1394148857-19607-1-git-send-email-agraf@suse.de> Subject: [Qemu-devel] [PULL 108/130] target-ppc: Altivec 2.07: Change Bit Masks to Support 64-bit Rotates and Shifts List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: qemu-devel@nongnu.org Cc: Peter Maydell , Tom Musta , blauwirbel@gmail.com, qemu-ppc@nongnu.org, aliguori@amazon.com, aurelien@aurel32.net From: Tom Musta Existing code in the VROTATE, VSL and VSR macros for the Altivec rotate and shift helpers uses a formula to compute a bit mask used to extract the rotate/shift amount from the VRB register. What is desired is: mask = (1 << (3 + log2(sizeof(element)))) - 1 but what is implemented is: mask = (1 << (3 + (sizeof(element)/2))) - 1 This produces correct answers when "element" is uint8_t, uint16_t or uint_32t. But it breaks down when element is uint64_t. This patch corrects the situation. Since the mask is known at compile time, the macros are changed to simply accept the mask as an argument. Subsequent patches in this series will add double-word variants of rotates and shifts and thus take advantage of this fix. Signed-off-by: Tom Musta Signed-off-by: Alexander Graf --- target-ppc/int_helper.c | 40 +++++++++++++++------------------------- 1 file changed, 15 insertions(+), 25 deletions(-) diff --git a/target-ppc/int_helper.c b/target-ppc/int_helper.c index 56e8d9a..59b5a1f 100644 --- a/target-ppc/int_helper.c +++ b/target-ppc/int_helper.c @@ -1128,23 +1128,20 @@ VRFI(p, float_round_up) VRFI(z, float_round_to_zero) #undef VRFI -#define VROTATE(suffix, element) \ +#define VROTATE(suffix, element, mask) \ void helper_vrl##suffix(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b) \ { \ int i; \ \ for (i = 0; i < ARRAY_SIZE(r->element); i++) { \ - unsigned int mask = ((1 << \ - (3 + (sizeof(a->element[0]) >> 1))) \ - - 1); \ unsigned int shift = b->element[i] & mask; \ r->element[i] = (a->element[i] << shift) | \ (a->element[i] >> (sizeof(a->element[0]) * 8 - shift)); \ } \ } -VROTATE(b, u8) -VROTATE(h, u16) -VROTATE(w, u32) +VROTATE(b, u8, 0x7) +VROTATE(h, u16, 0xF) +VROTATE(w, u32, 0x1F) #undef VROTATE void helper_vrsqrtefp(CPUPPCState *env, ppc_avr_t *r, ppc_avr_t *b) @@ -1225,23 +1222,20 @@ VSHIFT(r, RIGHT) #undef LEFT #undef RIGHT -#define VSL(suffix, element) \ +#define VSL(suffix, element, mask) \ void helper_vsl##suffix(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b) \ { \ int i; \ \ for (i = 0; i < ARRAY_SIZE(r->element); i++) { \ - unsigned int mask = ((1 << \ - (3 + (sizeof(a->element[0]) >> 1))) \ - - 1); \ unsigned int shift = b->element[i] & mask; \ \ r->element[i] = a->element[i] << shift; \ } \ } -VSL(b, u8) -VSL(h, u16) -VSL(w, u32) +VSL(b, u8, 0x7) +VSL(h, u16, 0x0F) +VSL(w, u32, 0x1F) #undef VSL void helper_vsldoi(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t shift) @@ -1325,26 +1319,22 @@ VSPLTI(h, s16, int16_t) VSPLTI(w, s32, int32_t) #undef VSPLTI -#define VSR(suffix, element) \ +#define VSR(suffix, element, mask) \ void helper_vsr##suffix(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b) \ { \ int i; \ \ for (i = 0; i < ARRAY_SIZE(r->element); i++) { \ - unsigned int mask = ((1 << \ - (3 + (sizeof(a->element[0]) >> 1))) \ - - 1); \ unsigned int shift = b->element[i] & mask; \ - \ r->element[i] = a->element[i] >> shift; \ } \ } -VSR(ab, s8) -VSR(ah, s16) -VSR(aw, s32) -VSR(b, u8) -VSR(h, u16) -VSR(w, u32) +VSR(ab, s8, 0x7) +VSR(ah, s16, 0xF) +VSR(aw, s32, 0x1F) +VSR(b, u8, 0x7) +VSR(h, u16, 0xF) +VSR(w, u32, 0x1F) #undef VSR void helper_vsro(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b) -- 1.8.1.4