yading@10: /* yading@10: * Copyright (C) 2003 David S. Miller yading@10: * yading@10: * This file is part of FFmpeg. yading@10: * yading@10: * FFmpeg is free software; you can redistribute it and/or yading@10: * modify it under the terms of the GNU Lesser General Public yading@10: * License as published by the Free Software Foundation; either yading@10: * version 2.1 of the License, or (at your option) any later version. yading@10: * yading@10: * FFmpeg is distributed in the hope that it will be useful, yading@10: * but WITHOUT ANY WARRANTY; without even the implied warranty of yading@10: * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU yading@10: * Lesser General Public License for more details. yading@10: * yading@10: * You should have received a copy of the GNU Lesser General Public yading@10: * License along with FFmpeg; if not, write to the Free Software yading@10: * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA yading@10: */ yading@10: yading@10: /* You may be asking why I hard-code the instruction opcodes and don't yading@10: * use the normal VIS assembler mnenomics for the VIS instructions. yading@10: * yading@10: * The reason is that Sun, in their infinite wisdom, decided that a binary yading@10: * using a VIS instruction will cause it to be marked (in the ELF headers) yading@10: * as doing so, and this prevents the OS from loading such binaries if the yading@10: * current cpu doesn't have VIS. There is no way to easily override this yading@10: * behavior of the assembler that I am aware of. yading@10: * yading@10: * This totally defeats what libmpeg2 is trying to do which is allow a yading@10: * single binary to be created, and then detect the availability of VIS yading@10: * at runtime. yading@10: * yading@10: * I'm not saying that tainting the binary by default is bad, rather I'm yading@10: * saying that not providing a way to override this easily unnecessarily yading@10: * ties people's hands. yading@10: * yading@10: * Thus, we do the opcode encoding by hand and output 32-bit words in yading@10: * the assembler to keep the binary from becoming tainted. yading@10: */ yading@10: yading@10: #ifndef AVCODEC_SPARC_VIS_H yading@10: #define AVCODEC_SPARC_VIS_H yading@10: yading@10: #define ACCEL_SPARC_VIS 1 yading@10: #define ACCEL_SPARC_VIS2 2 yading@10: yading@10: static inline int vis_level(void) yading@10: { yading@10: int accel = 0; yading@10: accel |= ACCEL_SPARC_VIS; yading@10: accel |= ACCEL_SPARC_VIS2; yading@10: return accel; yading@10: } yading@10: yading@10: #define vis_opc_base ((0x1 << 31) | (0x36 << 19)) yading@10: #define vis_opf(X) ((X) << 5) yading@10: #define vis_sreg(X) (X) yading@10: #define vis_dreg(X) (((X)&0x1f)|((X)>>5)) yading@10: #define vis_rs1_s(X) (vis_sreg(X) << 14) yading@10: #define vis_rs1_d(X) (vis_dreg(X) << 14) yading@10: #define vis_rs2_s(X) (vis_sreg(X) << 0) yading@10: #define vis_rs2_d(X) (vis_dreg(X) << 0) yading@10: #define vis_rd_s(X) (vis_sreg(X) << 25) yading@10: #define vis_rd_d(X) (vis_dreg(X) << 25) yading@10: yading@10: #define vis_ss2s(opf,rs1,rs2,rd) \ yading@10: __asm__ volatile (".word %0" \ yading@10: : : "i" (vis_opc_base | vis_opf(opf) | \ yading@10: vis_rs1_s(rs1) | \ yading@10: vis_rs2_s(rs2) | \ yading@10: vis_rd_s(rd))) yading@10: yading@10: #define vis_dd2d(opf,rs1,rs2,rd) \ yading@10: __asm__ volatile (".word %0" \ yading@10: : : "i" (vis_opc_base | vis_opf(opf) | \ yading@10: vis_rs1_d(rs1) | \ yading@10: vis_rs2_d(rs2) | \ yading@10: vis_rd_d(rd))) yading@10: yading@10: #define vis_ss2d(opf,rs1,rs2,rd) \ yading@10: __asm__ volatile (".word %0" \ yading@10: : : "i" (vis_opc_base | vis_opf(opf) | \ yading@10: vis_rs1_s(rs1) | \ yading@10: vis_rs2_s(rs2) | \ yading@10: vis_rd_d(rd))) yading@10: yading@10: #define vis_sd2d(opf,rs1,rs2,rd) \ yading@10: __asm__ volatile (".word %0" \ yading@10: : : "i" (vis_opc_base | vis_opf(opf) | \ yading@10: vis_rs1_s(rs1) | \ yading@10: vis_rs2_d(rs2) | \ yading@10: vis_rd_d(rd))) yading@10: yading@10: #define vis_d2s(opf,rs2,rd) \ yading@10: __asm__ volatile (".word %0" \ yading@10: : : "i" (vis_opc_base | vis_opf(opf) | \ yading@10: vis_rs2_d(rs2) | \ yading@10: vis_rd_s(rd))) yading@10: yading@10: #define vis_s2d(opf,rs2,rd) \ yading@10: __asm__ volatile (".word %0" \ yading@10: : : "i" (vis_opc_base | vis_opf(opf) | \ yading@10: vis_rs2_s(rs2) | \ yading@10: vis_rd_d(rd))) yading@10: yading@10: #define vis_d12d(opf,rs1,rd) \ yading@10: __asm__ volatile (".word %0" \ yading@10: : : "i" (vis_opc_base | vis_opf(opf) | \ yading@10: vis_rs1_d(rs1) | \ yading@10: vis_rd_d(rd))) yading@10: yading@10: #define vis_d22d(opf,rs2,rd) \ yading@10: __asm__ volatile (".word %0" \ yading@10: : : "i" (vis_opc_base | vis_opf(opf) | \ yading@10: vis_rs2_d(rs2) | \ yading@10: vis_rd_d(rd))) yading@10: yading@10: #define vis_s12s(opf,rs1,rd) \ yading@10: __asm__ volatile (".word %0" \ yading@10: : : "i" (vis_opc_base | vis_opf(opf) | \ yading@10: vis_rs1_s(rs1) | \ yading@10: vis_rd_s(rd))) yading@10: yading@10: #define vis_s22s(opf,rs2,rd) \ yading@10: __asm__ volatile (".word %0" \ yading@10: : : "i" (vis_opc_base | vis_opf(opf) | \ yading@10: vis_rs2_s(rs2) | \ yading@10: vis_rd_s(rd))) yading@10: yading@10: #define vis_s(opf,rd) \ yading@10: __asm__ volatile (".word %0" \ yading@10: : : "i" (vis_opc_base | vis_opf(opf) | \ yading@10: vis_rd_s(rd))) yading@10: yading@10: #define vis_d(opf,rd) \ yading@10: __asm__ volatile (".word %0" \ yading@10: : : "i" (vis_opc_base | vis_opf(opf) | \ yading@10: vis_rd_d(rd))) yading@10: yading@10: #define vis_r2m(op,rd,mem) \ yading@10: __asm__ volatile (#op "\t%%f" #rd ", [%0]" : : "r" (&(mem)) ) yading@10: yading@10: #define vis_r2m_2(op,rd,mem1,mem2) \ yading@10: __asm__ volatile (#op "\t%%f" #rd ", [%0 + %1]" : : "r" (mem1), "r" (mem2) ) yading@10: yading@10: #define vis_m2r(op,mem,rd) \ yading@10: __asm__ volatile (#op "\t[%0], %%f" #rd : : "r" (&(mem)) ) yading@10: yading@10: #define vis_m2r_2(op,mem1,mem2,rd) \ yading@10: __asm__ volatile (#op "\t[%0 + %1], %%f" #rd : : "r" (mem1), "r" (mem2) ) yading@10: yading@10: static inline void vis_set_gsr(unsigned int _val) yading@10: { yading@10: register unsigned int val __asm__("g1"); yading@10: yading@10: val = _val; yading@10: __asm__ volatile(".word 0xa7804000" yading@10: : : "r" (val)); yading@10: } yading@10: yading@10: #define VIS_GSR_ALIGNADDR_MASK 0x0000007 yading@10: #define VIS_GSR_ALIGNADDR_SHIFT 0 yading@10: #define VIS_GSR_SCALEFACT_MASK 0x0000078 yading@10: #define VIS_GSR_SCALEFACT_SHIFT 3 yading@10: yading@10: #define vis_ld32(mem,rs1) vis_m2r(ld, mem, rs1) yading@10: #define vis_ld32_2(mem1,mem2,rs1) vis_m2r_2(ld, mem1, mem2, rs1) yading@10: #define vis_st32(rs1,mem) vis_r2m(st, rs1, mem) yading@10: #define vis_st32_2(rs1,mem1,mem2) vis_r2m_2(st, rs1, mem1, mem2) yading@10: #define vis_ld64(mem,rs1) vis_m2r(ldd, mem, rs1) yading@10: #define vis_ld64_2(mem1,mem2,rs1) vis_m2r_2(ldd, mem1, mem2, rs1) yading@10: #define vis_st64(rs1,mem) vis_r2m(std, rs1, mem) yading@10: #define vis_st64_2(rs1,mem1,mem2) vis_r2m_2(std, rs1, mem1, mem2) yading@10: yading@10: #define vis_ldblk(mem, rd) \ yading@10: do { register void *__mem __asm__("g1"); \ yading@10: __mem = &(mem); \ yading@10: __asm__ volatile(".word 0xc1985e00 | %1" \ yading@10: : \ yading@10: : "r" (__mem), \ yading@10: "i" (vis_rd_d(rd)) \ yading@10: : "memory"); \ yading@10: } while (0) yading@10: yading@10: #define vis_stblk(rd, mem) \ yading@10: do { register void *__mem __asm__("g1"); \ yading@10: __mem = &(mem); \ yading@10: __asm__ volatile(".word 0xc1b85e00 | %1" \ yading@10: : \ yading@10: : "r" (__mem), \ yading@10: "i" (vis_rd_d(rd)) \ yading@10: : "memory"); \ yading@10: } while (0) yading@10: yading@10: #define vis_membar_storestore() \ yading@10: __asm__ volatile(".word 0x8143e008" : : : "memory") yading@10: yading@10: #define vis_membar_sync() \ yading@10: __asm__ volatile(".word 0x8143e040" : : : "memory") yading@10: yading@10: /* 16 and 32 bit partitioned addition and subtraction. The normal yading@10: * versions perform 4 16-bit or 2 32-bit additions or subtractions. yading@10: * The 's' versions perform 2 16-bit or 1 32-bit additions or yading@10: * subtractions. yading@10: */ yading@10: yading@10: #define vis_padd16(rs1,rs2,rd) vis_dd2d(0x50, rs1, rs2, rd) yading@10: #define vis_padd16s(rs1,rs2,rd) vis_ss2s(0x51, rs1, rs2, rd) yading@10: #define vis_padd32(rs1,rs2,rd) vis_dd2d(0x52, rs1, rs2, rd) yading@10: #define vis_padd32s(rs1,rs2,rd) vis_ss2s(0x53, rs1, rs2, rd) yading@10: #define vis_psub16(rs1,rs2,rd) vis_dd2d(0x54, rs1, rs2, rd) yading@10: #define vis_psub16s(rs1,rs2,rd) vis_ss2s(0x55, rs1, rs2, rd) yading@10: #define vis_psub32(rs1,rs2,rd) vis_dd2d(0x56, rs1, rs2, rd) yading@10: #define vis_psub32s(rs1,rs2,rd) vis_ss2s(0x57, rs1, rs2, rd) yading@10: yading@10: /* Pixel formatting instructions. */ yading@10: yading@10: #define vis_pack16(rs2,rd) vis_d2s( 0x3b, rs2, rd) yading@10: #define vis_pack32(rs1,rs2,rd) vis_dd2d(0x3a, rs1, rs2, rd) yading@10: #define vis_packfix(rs2,rd) vis_d2s( 0x3d, rs2, rd) yading@10: #define vis_expand(rs2,rd) vis_s2d( 0x4d, rs2, rd) yading@10: #define vis_pmerge(rs1,rs2,rd) vis_ss2d(0x4b, rs1, rs2, rd) yading@10: yading@10: /* Partitioned multiply instructions. */ yading@10: yading@10: #define vis_mul8x16(rs1,rs2,rd) vis_sd2d(0x31, rs1, rs2, rd) yading@10: #define vis_mul8x16au(rs1,rs2,rd) vis_ss2d(0x33, rs1, rs2, rd) yading@10: #define vis_mul8x16al(rs1,rs2,rd) vis_ss2d(0x35, rs1, rs2, rd) yading@10: #define vis_mul8sux16(rs1,rs2,rd) vis_dd2d(0x36, rs1, rs2, rd) yading@10: #define vis_mul8ulx16(rs1,rs2,rd) vis_dd2d(0x37, rs1, rs2, rd) yading@10: #define vis_muld8sux16(rs1,rs2,rd) vis_ss2d(0x38, rs1, rs2, rd) yading@10: #define vis_muld8ulx16(rs1,rs2,rd) vis_ss2d(0x39, rs1, rs2, rd) yading@10: yading@10: /* Alignment instructions. */ yading@10: yading@10: static inline const void *vis_alignaddr(const void *_ptr) yading@10: { yading@10: register const void *ptr __asm__("g1"); yading@10: yading@10: ptr = _ptr; yading@10: yading@10: __asm__ volatile(".word %2" yading@10: : "=&r" (ptr) yading@10: : "0" (ptr), yading@10: "i" (vis_opc_base | vis_opf(0x18) | yading@10: vis_rs1_s(1) | yading@10: vis_rs2_s(0) | yading@10: vis_rd_s(1))); yading@10: yading@10: return ptr; yading@10: } yading@10: yading@10: static inline void vis_alignaddr_g0(void *_ptr) yading@10: { yading@10: register void *ptr __asm__("g1"); yading@10: yading@10: ptr = _ptr; yading@10: yading@10: __asm__ volatile(".word %2" yading@10: : "=&r" (ptr) yading@10: : "0" (ptr), yading@10: "i" (vis_opc_base | vis_opf(0x18) | yading@10: vis_rs1_s(1) | yading@10: vis_rs2_s(0) | yading@10: vis_rd_s(0))); yading@10: } yading@10: yading@10: static inline void *vis_alignaddrl(void *_ptr) yading@10: { yading@10: register void *ptr __asm__("g1"); yading@10: yading@10: ptr = _ptr; yading@10: yading@10: __asm__ volatile(".word %2" yading@10: : "=&r" (ptr) yading@10: : "0" (ptr), yading@10: "i" (vis_opc_base | vis_opf(0x19) | yading@10: vis_rs1_s(1) | yading@10: vis_rs2_s(0) | yading@10: vis_rd_s(1))); yading@10: yading@10: return ptr; yading@10: } yading@10: yading@10: static inline void vis_alignaddrl_g0(void *_ptr) yading@10: { yading@10: register void *ptr __asm__("g1"); yading@10: yading@10: ptr = _ptr; yading@10: yading@10: __asm__ volatile(".word %2" yading@10: : "=&r" (ptr) yading@10: : "0" (ptr), yading@10: "i" (vis_opc_base | vis_opf(0x19) | yading@10: vis_rs1_s(1) | yading@10: vis_rs2_s(0) | yading@10: vis_rd_s(0))); yading@10: } yading@10: yading@10: #define vis_faligndata(rs1,rs2,rd) vis_dd2d(0x48, rs1, rs2, rd) yading@10: yading@10: /* Logical operate instructions. */ yading@10: yading@10: #define vis_fzero(rd) vis_d( 0x60, rd) yading@10: #define vis_fzeros(rd) vis_s( 0x61, rd) yading@10: #define vis_fone(rd) vis_d( 0x7e, rd) yading@10: #define vis_fones(rd) vis_s( 0x7f, rd) yading@10: #define vis_src1(rs1,rd) vis_d12d(0x74, rs1, rd) yading@10: #define vis_src1s(rs1,rd) vis_s12s(0x75, rs1, rd) yading@10: #define vis_src2(rs2,rd) vis_d22d(0x78, rs2, rd) yading@10: #define vis_src2s(rs2,rd) vis_s22s(0x79, rs2, rd) yading@10: #define vis_not1(rs1,rd) vis_d12d(0x6a, rs1, rd) yading@10: #define vis_not1s(rs1,rd) vis_s12s(0x6b, rs1, rd) yading@10: #define vis_not2(rs2,rd) vis_d22d(0x66, rs2, rd) yading@10: #define vis_not2s(rs2,rd) vis_s22s(0x67, rs2, rd) yading@10: #define vis_or(rs1,rs2,rd) vis_dd2d(0x7c, rs1, rs2, rd) yading@10: #define vis_ors(rs1,rs2,rd) vis_ss2s(0x7d, rs1, rs2, rd) yading@10: #define vis_nor(rs1,rs2,rd) vis_dd2d(0x62, rs1, rs2, rd) yading@10: #define vis_nors(rs1,rs2,rd) vis_ss2s(0x63, rs1, rs2, rd) yading@10: #define vis_and(rs1,rs2,rd) vis_dd2d(0x70, rs1, rs2, rd) yading@10: #define vis_ands(rs1,rs2,rd) vis_ss2s(0x71, rs1, rs2, rd) yading@10: #define vis_nand(rs1,rs2,rd) vis_dd2d(0x6e, rs1, rs2, rd) yading@10: #define vis_nands(rs1,rs2,rd) vis_ss2s(0x6f, rs1, rs2, rd) yading@10: #define vis_xor(rs1,rs2,rd) vis_dd2d(0x6c, rs1, rs2, rd) yading@10: #define vis_xors(rs1,rs2,rd) vis_ss2s(0x6d, rs1, rs2, rd) yading@10: #define vis_xnor(rs1,rs2,rd) vis_dd2d(0x72, rs1, rs2, rd) yading@10: #define vis_xnors(rs1,rs2,rd) vis_ss2s(0x73, rs1, rs2, rd) yading@10: #define vis_ornot1(rs1,rs2,rd) vis_dd2d(0x7a, rs1, rs2, rd) yading@10: #define vis_ornot1s(rs1,rs2,rd) vis_ss2s(0x7b, rs1, rs2, rd) yading@10: #define vis_ornot2(rs1,rs2,rd) vis_dd2d(0x76, rs1, rs2, rd) yading@10: #define vis_ornot2s(rs1,rs2,rd) vis_ss2s(0x77, rs1, rs2, rd) yading@10: #define vis_andnot1(rs1,rs2,rd) vis_dd2d(0x68, rs1, rs2, rd) yading@10: #define vis_andnot1s(rs1,rs2,rd) vis_ss2s(0x69, rs1, rs2, rd) yading@10: #define vis_andnot2(rs1,rs2,rd) vis_dd2d(0x64, rs1, rs2, rd) yading@10: #define vis_andnot2s(rs1,rs2,rd) vis_ss2s(0x65, rs1, rs2, rd) yading@10: yading@10: /* Pixel component distance. */ yading@10: yading@10: #define vis_pdist(rs1,rs2,rd) vis_dd2d(0x3e, rs1, rs2, rd) yading@10: yading@10: #endif /* AVCODEC_SPARC_VIS_H */