diff options
author | Peter Maydell <peter.maydell@linaro.org> | 2018-08-24 23:10:15 +0100 |
---|---|---|
committer | Peter Maydell <peter.maydell@linaro.org> | 2018-08-24 23:10:15 +0100 |
commit | e2e6fa67931fdba493e10cc55abcc99a65c92c7b (patch) | |
tree | a429e9f81e874002e3840e89e50b1fb3403cb4b3 | |
parent | 746b7907feeba7eced022b96b8effa079bd27a2e (diff) | |
parent | d45942d908edee175a90f915ab92ac302eedf33a (diff) |
Merge remote-tracking branch 'remotes/amarkovic/tags/mips-queue-aug-2018' into staging
MIPS queue August 2018 v6
# gpg: Signature made Fri 24 Aug 2018 16:52:27 BST
# gpg: using RSA key D4972A8967F75A65
# gpg: Good signature from "Aleksandar Markovic <amarkovic@wavecomp.com>"
# gpg: WARNING: This key is not certified with a trusted signature!
# gpg: There is no indication that the signature belongs to the owner.
# Primary key fingerprint: 8526 FBF1 5DA3 811F 4A01 DD75 D497 2A89 67F7 5A65
* remotes/amarkovic/tags/mips-queue-aug-2018: (45 commits)
target/mips: Add definition of nanoMIPS I7200 CPU
mips_malta: Fix semihosting argument passing for nanoMIPS bare metal
mips_malta: Add setting up GT64120 BARs to the nanoMIPS bootloader
mips_malta: Add basic nanoMIPS boot code for Malta board
elf: Don't check FCR31_NAN2008 bit for nanoMIPS
elf: On elf loading, treat both EM_MIPS and EM_NANOMIPS as legal for MIPS
elf: Relax MIPS' elf_check_arch() to accept EM_NANOMIPS too
elf: Add EM_NANOMIPS value as a valid one for e_machine field
target/mips: Fix ERET/ERETNC behavior related to ADEL exception
target/mips: Add updating BadInstr and BadInstrX for nanoMIPS
target/mips: Add availability control via bit NMS
target/mips: Add emulation of DSP ASE for nanoMIPS - part 6
target/mips: Add emulation of DSP ASE for nanoMIPS - part 5
target/mips: Add emulation of DSP ASE for nanoMIPS - part 4
target/mips: Add emulation of DSP ASE for nanoMIPS - part 3
target/mips: Add emulation of DSP ASE for nanoMIPS - part 2
target/mips: Add emulation of DSP ASE for nanoMIPS - part 1
target/mips: Implement MT ASE support for nanoMIPS
target/mips: Fix pre-nanoMIPS MT ASE instructions availability control
target/mips: Add emulation of nanoMIPS 32-bit branch instructions
...
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
-rw-r--r-- | hw/mips/mips_malta.c | 212 | ||||
-rw-r--r-- | include/elf.h | 2 | ||||
-rw-r--r-- | include/hw/elf_ops.h | 8 | ||||
-rw-r--r-- | linux-user/elfload.c | 2 | ||||
-rw-r--r-- | linux-user/mips/cpu_loop.c | 28 | ||||
-rw-r--r-- | target/mips/cpu.h | 2 | ||||
-rw-r--r-- | target/mips/helper.c | 16 | ||||
-rw-r--r-- | target/mips/helper.h | 2 | ||||
-rw-r--r-- | target/mips/mips-defs.h | 4 | ||||
-rw-r--r-- | target/mips/op_helper.c | 98 | ||||
-rw-r--r-- | target/mips/translate.c | 4899 | ||||
-rw-r--r-- | target/mips/translate_init.inc.c | 39 |
12 files changed, 5275 insertions, 37 deletions
diff --git a/hw/mips/mips_malta.c b/hw/mips/mips_malta.c index 3467451482..40041d5ec0 100644 --- a/hw/mips/mips_malta.c +++ b/hw/mips/mips_malta.c @@ -599,6 +599,208 @@ static void network_init(PCIBus *pci_bus) } } +static void write_bootloader_nanomips(uint8_t *base, int64_t run_addr, + int64_t kernel_entry) +{ + uint16_t *p; + + /* Small bootloader */ + p = (uint16_t *)base; + +#define NM_HI1(VAL) (((VAL) >> 16) & 0x1f) +#define NM_HI2(VAL) \ + (((VAL) & 0xf000) | (((VAL) >> 19) & 0xffc) | (((VAL) >> 31) & 0x1)) +#define NM_LO(VAL) ((VAL) & 0xfff) + + stw_p(p++, 0x2800); stw_p(p++, 0x001c); + /* bc to_here */ + stw_p(p++, 0x8000); stw_p(p++, 0xc000); + /* nop */ + stw_p(p++, 0x8000); stw_p(p++, 0xc000); + /* nop */ + stw_p(p++, 0x8000); stw_p(p++, 0xc000); + /* nop */ + stw_p(p++, 0x8000); stw_p(p++, 0xc000); + /* nop */ + stw_p(p++, 0x8000); stw_p(p++, 0xc000); + /* nop */ + stw_p(p++, 0x8000); stw_p(p++, 0xc000); + /* nop */ + stw_p(p++, 0x8000); stw_p(p++, 0xc000); + /* nop */ + + /* to_here: */ + if (semihosting_get_argc()) { + /* Preserve a0 content as arguments have been passed */ + stw_p(p++, 0x8000); stw_p(p++, 0xc000); + /* nop */ + } else { + stw_p(p++, 0x0080); stw_p(p++, 0x0002); + /* li a0,2 */ + } + + stw_p(p++, 0xe3a0 | NM_HI1(ENVP_ADDR - 64)); + + stw_p(p++, NM_HI2(ENVP_ADDR - 64)); + /* lui sp,%hi(ENVP_ADDR - 64) */ + + stw_p(p++, 0x83bd); stw_p(p++, NM_LO(ENVP_ADDR - 64)); + /* ori sp,sp,%lo(ENVP_ADDR - 64) */ + + stw_p(p++, 0xe0a0 | NM_HI1(ENVP_ADDR)); + + stw_p(p++, NM_HI2(ENVP_ADDR)); + /* lui a1,%hi(ENVP_ADDR) */ + + stw_p(p++, 0x80a5); stw_p(p++, NM_LO(ENVP_ADDR)); + /* ori a1,a1,%lo(ENVP_ADDR) */ + + stw_p(p++, 0xe0c0 | NM_HI1(ENVP_ADDR + 8)); + + stw_p(p++, NM_HI2(ENVP_ADDR + 8)); + /* lui a2,%hi(ENVP_ADDR + 8) */ + + stw_p(p++, 0x80c6); stw_p(p++, NM_LO(ENVP_ADDR + 8)); + /* ori a2,a2,%lo(ENVP_ADDR + 8) */ + + stw_p(p++, 0xe0e0 | NM_HI1(loaderparams.ram_low_size)); + + stw_p(p++, NM_HI2(loaderparams.ram_low_size)); + /* lui a3,%hi(loaderparams.ram_low_size) */ + + stw_p(p++, 0x80e7); stw_p(p++, NM_LO(loaderparams.ram_low_size)); + /* ori a3,a3,%lo(loaderparams.ram_low_size) */ + + /* + * Load BAR registers as done by YAMON: + * + * - set up PCI0 I/O BARs from 0x18000000 to 0x181fffff + * - set up PCI0 MEM0 at 0x10000000, size 0x8000000 + * - set up PCI0 MEM1 at 0x18200000, size 0xbe00000 + * + */ + stw_p(p++, 0xe040); stw_p(p++, 0x0681); + /* lui t1, %hi(0xb4000000) */ + +#ifdef TARGET_WORDS_BIGENDIAN + + stw_p(p++, 0xe020); stw_p(p++, 0x0be1); + /* lui t0, %hi(0xdf000000) */ + + /* 0x68 corresponds to GT_ISD (from hw/mips/gt64xxx_pci.c) */ + stw_p(p++, 0x8422); stw_p(p++, 0x9068); + /* sw t0, 0x68(t1) */ + + stw_p(p++, 0xe040); stw_p(p++, 0x077d); + /* lui t1, %hi(0xbbe00000) */ + + stw_p(p++, 0xe020); stw_p(p++, 0x0801); + /* lui t0, %hi(0xc0000000) */ + + /* 0x48 corresponds to GT_PCI0IOLD */ + stw_p(p++, 0x8422); stw_p(p++, 0x9048); + /* sw t0, 0x48(t1) */ + + stw_p(p++, 0xe020); stw_p(p++, 0x0800); + /* lui t0, %hi(0x40000000) */ + + /* 0x50 corresponds to GT_PCI0IOHD */ + stw_p(p++, 0x8422); stw_p(p++, 0x9050); + /* sw t0, 0x50(t1) */ + + stw_p(p++, 0xe020); stw_p(p++, 0x0001); + /* lui t0, %hi(0x80000000) */ + + /* 0x58 corresponds to GT_PCI0M0LD */ + stw_p(p++, 0x8422); stw_p(p++, 0x9058); + /* sw t0, 0x58(t1) */ + + stw_p(p++, 0xe020); stw_p(p++, 0x07e0); + /* lui t0, %hi(0x3f000000) */ + + /* 0x60 corresponds to GT_PCI0M0HD */ + stw_p(p++, 0x8422); stw_p(p++, 0x9060); + /* sw t0, 0x60(t1) */ + + stw_p(p++, 0xe020); stw_p(p++, 0x0821); + /* lui t0, %hi(0xc1000000) */ + + /* 0x80 corresponds to GT_PCI0M1LD */ + stw_p(p++, 0x8422); stw_p(p++, 0x9080); + /* sw t0, 0x80(t1) */ + + stw_p(p++, 0xe020); stw_p(p++, 0x0bc0); + /* lui t0, %hi(0x5e000000) */ + +#else + + stw_p(p++, 0x0020); stw_p(p++, 0x00df); + /* addiu[32] t0, $0, 0xdf */ + + /* 0x68 corresponds to GT_ISD */ + stw_p(p++, 0x8422); stw_p(p++, 0x9068); + /* sw t0, 0x68(t1) */ + + /* Use kseg2 remapped address 0x1be00000 */ + stw_p(p++, 0xe040); stw_p(p++, 0x077d); + /* lui t1, %hi(0xbbe00000) */ + + stw_p(p++, 0x0020); stw_p(p++, 0x00c0); + /* addiu[32] t0, $0, 0xc0 */ + + /* 0x48 corresponds to GT_PCI0IOLD */ + stw_p(p++, 0x8422); stw_p(p++, 0x9048); + /* sw t0, 0x48(t1) */ + + stw_p(p++, 0x0020); stw_p(p++, 0x0040); + /* addiu[32] t0, $0, 0x40 */ + + /* 0x50 corresponds to GT_PCI0IOHD */ + stw_p(p++, 0x8422); stw_p(p++, 0x9050); + /* sw t0, 0x50(t1) */ + + stw_p(p++, 0x0020); stw_p(p++, 0x0080); + /* addiu[32] t0, $0, 0x80 */ + + /* 0x58 corresponds to GT_PCI0M0LD */ + stw_p(p++, 0x8422); stw_p(p++, 0x9058); + /* sw t0, 0x58(t1) */ + + stw_p(p++, 0x0020); stw_p(p++, 0x003f); + /* addiu[32] t0, $0, 0x3f */ + + /* 0x60 corresponds to GT_PCI0M0HD */ + stw_p(p++, 0x8422); stw_p(p++, 0x9060); + /* sw t0, 0x60(t1) */ + + stw_p(p++, 0x0020); stw_p(p++, 0x00c1); + /* addiu[32] t0, $0, 0xc1 */ + + /* 0x80 corresponds to GT_PCI0M1LD */ + stw_p(p++, 0x8422); stw_p(p++, 0x9080); + /* sw t0, 0x80(t1) */ + + stw_p(p++, 0x0020); stw_p(p++, 0x005e); + /* addiu[32] t0, $0, 0x5e */ + +#endif + + /* 0x88 corresponds to GT_PCI0M1HD */ + stw_p(p++, 0x8422); stw_p(p++, 0x9088); + /* sw t0, 0x88(t1) */ + + stw_p(p++, 0xe320 | NM_HI1(kernel_entry)); + + stw_p(p++, NM_HI2(kernel_entry)); + /* lui t9,%hi(kernel_entry) */ + + stw_p(p++, 0x8339); stw_p(p++, NM_LO(kernel_entry)); + /* ori t9,t9,%lo(kernel_entry) */ + + stw_p(p++, 0x4bf9); stw_p(p++, 0x0000); + /* jalrc t8 */ +} + /* ROM and pseudo bootloader The following code implements a very very simple bootloader. It first @@ -620,7 +822,6 @@ static void network_init(PCIBus *pci_bus) a2 - 32-bit address of the environment variables table a3 - RAM size in bytes */ - static void write_bootloader(uint8_t *base, int64_t run_addr, int64_t kernel_entry) { @@ -1096,8 +1297,13 @@ void mips_malta_init(MachineState *machine) loaderparams.initrd_filename = initrd_filename; kernel_entry = load_kernel(); - write_bootloader(memory_region_get_ram_ptr(bios), - bootloader_run_addr, kernel_entry); + if (!cpu_supports_isa(machine->cpu_type, ISA_NANOMIPS32)) { + write_bootloader(memory_region_get_ram_ptr(bios), + bootloader_run_addr, kernel_entry); + } else { + write_bootloader_nanomips(memory_region_get_ram_ptr(bios), + bootloader_run_addr, kernel_entry); + } if (kvm_enabled()) { /* Write the bootloader code @ the end of RAM, 1MB reserved */ write_bootloader(memory_region_get_ram_ptr(ram_low_preio) + diff --git a/include/elf.h b/include/elf.h index 28a5a638e0..312f68af81 100644 --- a/include/elf.h +++ b/include/elf.h @@ -143,6 +143,8 @@ typedef int64_t Elf64_Sxword; #define EM_RISCV 243 /* RISC-V */ +#define EM_NANOMIPS 249 /* Wave Computing nanoMIPS */ + /* * This is an interim value that we will use until the committee comes * up with a final number. diff --git a/include/hw/elf_ops.h b/include/hw/elf_ops.h index b6e19e35d0..81cecaf27e 100644 --- a/include/hw/elf_ops.h +++ b/include/hw/elf_ops.h @@ -327,6 +327,14 @@ static int glue(load_elf, SZ)(const char *name, int fd, } } break; + case EM_MIPS: + case EM_NANOMIPS: + if ((ehdr.e_machine != EM_MIPS) && + (ehdr.e_machine != EM_NANOMIPS)) { + ret = ELF_LOAD_WRONG_ARCH; + goto fail; + } + break; default: if (elf_machine != ehdr.e_machine) { ret = ELF_LOAD_WRONG_ARCH; diff --git a/linux-user/elfload.c b/linux-user/elfload.c index df07055361..8638612aec 100644 --- a/linux-user/elfload.c +++ b/linux-user/elfload.c @@ -853,6 +853,8 @@ static void elf_core_copy_regs(target_elf_gregset_t *regs, const CPUPPCState *en #endif #define ELF_ARCH EM_MIPS +#define elf_check_arch(x) ((x) == EM_MIPS || (x) == EM_NANOMIPS) + static inline void init_thread(struct target_pt_regs *regs, struct image_info *infop) { diff --git a/linux-user/mips/cpu_loop.c b/linux-user/mips/cpu_loop.c index 084ad6a041..c9c20cf8b7 100644 --- a/linux-user/mips/cpu_loop.c +++ b/linux-user/mips/cpu_loop.c @@ -397,10 +397,13 @@ static int do_store_exclusive(CPUMIPSState *env) target_ulong addr; target_ulong page_addr; target_ulong val; + uint32_t val_wp = 0; + uint32_t llnewval_wp = 0; int flags; int segv = 0; int reg; int d; + int wp; addr = env->lladdr; page_addr = addr & TARGET_PAGE_MASK; @@ -412,19 +415,31 @@ static int do_store_exclusive(CPUMIPSState *env) } else { reg = env->llreg & 0x1f; d = (env->llreg & 0x20) != 0; - if (d) { - segv = get_user_s64(val, addr); + wp = (env->llreg & 0x40) != 0; + if (!wp) { + if (d) { + segv = get_user_s64(val, addr); + } else { + segv = get_user_s32(val, addr); + } } else { segv = get_user_s32(val, addr); + segv |= get_user_s32(val_wp, addr); + llnewval_wp = env->llnewval_wp; } if (!segv) { - if (val != env->llval) { + if (val != env->llval && val_wp == llnewval_wp) { env->active_tc.gpr[reg] = 0; } else { - if (d) { - segv = put_user_u64(env->llnewval, addr); + if (!wp) { + if (d) { + segv = put_user_u64(env->llnewval, addr); + } else { + segv = put_user_u32(env->llnewval, addr); + } } else { segv = put_user_u32(env->llnewval, addr); + segv |= put_user_u32(env->llnewval_wp, addr + 4); } if (!segv) { env->active_tc.gpr[reg] = 1; @@ -732,6 +747,9 @@ void target_cpu_copy_regs(CPUArchState *env, struct target_pt_regs *regs) if (regs->cp0_epc & 1) { env->hflags |= MIPS_HFLAG_M16; } + if (env->insn_flags & ISA_NANOMIPS32) { + return; + } if (((info->elf_flags & EF_MIPS_NAN2008) != 0) != ((env->active_fpu.fcr31 & (1 << FCR31_NAN2008)) != 0)) { if ((env->active_fpu.fcr31_rw_bitmask & diff --git a/target/mips/cpu.h b/target/mips/cpu.h index 009202cf64..28af4d191c 100644 --- a/target/mips/cpu.h +++ b/target/mips/cpu.h @@ -506,6 +506,8 @@ struct CPUMIPSState { uint64_t lladdr; target_ulong llval; target_ulong llnewval; + uint64_t llval_wp; + uint32_t llnewval_wp; target_ulong llreg; uint64_t CP0_LLAddr_rw_bitmask; int CP0_LLAddr_shift; diff --git a/target/mips/helper.c b/target/mips/helper.c index e215af9a41..f0c268b83c 100644 --- a/target/mips/helper.c +++ b/target/mips/helper.c @@ -682,6 +682,22 @@ static void set_hflags_for_handler (CPUMIPSState *env) static inline void set_badinstr_registers(CPUMIPSState *env) { + if (env->insn_flags & ISA_NANOMIPS32) { + if (env->CP0_Config3 & (1 << CP0C3_BI)) { + uint32_t instr = (cpu_lduw_code(env, env->active_tc.PC)) << 16; + if ((instr & 0x10000000) == 0) { + instr |= cpu_lduw_code(env, env->active_tc.PC + 2); + } + env->CP0_BadInstr = instr; + + if ((instr & 0xFC000000) == 0x60000000) { + instr = cpu_lduw_code(env, env->active_tc.PC + 4) << 16; + env->CP0_BadInstrX = instr; + } + } + return; + } + if (env->hflags & MIPS_HFLAG_M16) { /* TODO: add BadInstr support for microMIPS */ return; diff --git a/target/mips/helper.h b/target/mips/helper.h index 5f492348dd..b2a780a6f2 100644 --- a/target/mips/helper.h +++ b/target/mips/helper.h @@ -40,6 +40,8 @@ DEF_HELPER_FLAGS_1(bitswap, TCG_CALL_NO_RWG_SE, tl, tl) DEF_HELPER_FLAGS_1(dbitswap, TCG_CALL_NO_RWG_SE, tl, tl) #endif +DEF_HELPER_FLAGS_4(rotx, TCG_CALL_NO_RWG_SE, tl, tl, i32, i32, i32) + #ifndef CONFIG_USER_ONLY /* CP0 helpers */ DEF_HELPER_1(mfc0_mvpcontrol, tl, env) diff --git a/target/mips/mips-defs.h b/target/mips/mips-defs.h index d239069975..c8e99791ad 100644 --- a/target/mips/mips-defs.h +++ b/target/mips/mips-defs.h @@ -39,6 +39,7 @@ #define ISA_MIPS64R5 0x00001000 #define ISA_MIPS32R6 0x00002000 #define ISA_MIPS64R6 0x00004000 +#define ISA_NANOMIPS32 0x00008000 /* MIPS ASEs. */ #define ASE_MIPS16 0x00010000 @@ -87,6 +88,9 @@ #define CPU_MIPS32R6 (CPU_MIPS32R5 | ISA_MIPS32R6) #define CPU_MIPS64R6 (CPU_MIPS64R5 | CPU_MIPS32R6 | ISA_MIPS64R6) +/* Wave Computing: "nanoMIPS" */ +#define CPU_NANOMIPS32 (CPU_MIPS32R6 | ISA_NANOMIPS32) + /* Strictly follow the architecture standard: - Disallow "special" instruction handling for PMON/SPIM. Note that we still maintain Count/Compare to match the host clock. */ diff --git a/target/mips/op_helper.c b/target/mips/op_helper.c index 0b2663b73a..c148b310cd 100644 --- a/target/mips/op_helper.c +++ b/target/mips/op_helper.c @@ -249,6 +249,100 @@ target_ulong helper_bitswap(target_ulong rt) return (int32_t)bitswap(rt); } +target_ulong helper_rotx(target_ulong rs, uint32_t shift, uint32_t shiftx, + uint32_t stripe) +{ + int i; + uint64_t tmp0 = ((uint64_t)rs) << 32 | ((uint64_t)rs & 0xffffffff); + uint64_t tmp1 = tmp0; + for (i = 0; i <= 46; i++) { + int s; + if (i & 0x8) { + s = shift; + } else { + s = shiftx; + } + + if (stripe != 0 && !(i & 0x4)) { + s = ~s; + } + if (s & 0x10) { + if (tmp0 & (1LL << (i + 16))) { + tmp1 |= 1LL << i; + } else { + tmp1 &= ~(1LL << i); + } + } + } + + uint64_t tmp2 = tmp1; + for (i = 0; i <= 38; i++) { + int s; + if (i & 0x4) { + s = shift; + } else { + s = shiftx; + } + + if (s & 0x8) { + if (tmp1 & (1LL << (i + 8))) { + tmp2 |= 1LL << i; + } else { + tmp2 &= ~(1LL << i); + } + } + } + + uint64_t tmp3 = tmp2; + for (i = 0; i <= 34; i++) { + int s; + if (i & 0x2) { + s = shift; + } else { + s = shiftx; + } + if (s & 0x4) { + if (tmp2 & (1LL << (i + 4))) { + tmp3 |= 1LL << i; + } else { + tmp3 &= ~(1LL << i); + } + } + } + + uint64_t tmp4 = tmp3; + for (i = 0; i <= 32; i++) { + int s; + if (i & 0x1) { + s = shift; + } else { + s = shiftx; + } + if (s & 0x2) { + if (tmp3 & (1LL << (i + 2))) { + tmp4 |= 1LL << i; + } else { + tmp4 &= ~(1LL << i); + } + } + } + + uint64_t tmp5 = tmp4; + for (i = 0; i <= 31; i++) { + int s; + s = shift; + if (s & 0x1) { + if (tmp4 & (1LL << (i + 1))) { + tmp5 |= 1LL << i; + } else { + tmp5 &= ~(1LL << i); + } + } + } + + return (int64_t)(int32_t)(uint32_t)tmp5; +} + #ifndef CONFIG_USER_ONLY static inline hwaddr do_translate_address(CPUMIPSState *env, @@ -2333,10 +2427,12 @@ void helper_eretnc(CPUMIPSState *env) void helper_deret(CPUMIPSState *env) { debug_pre_eret(env); - set_pc(env, env->CP0_DEPC); env->hflags &= ~MIPS_HFLAG_DM; compute_hflags(env); + + set_pc(env, env->CP0_DEPC); + debug_post_eret(env); } #endif /* !CONFIG_USER_ONLY */ diff --git a/target/mips/translate.c b/target/mips/translate.c index bdd880bb77..ab16cdb911 100644 --- a/target/mips/translate.c +++ b/target/mips/translate.c @@ -1449,6 +1449,8 @@ typedef struct DisasContext { uint32_t opcode; int insn_flags; int32_t CP0_Config1; + int32_t CP0_Config3; + int32_t CP0_Config5; /* Routine used to access memory */ int mem_idx; TCGMemOp default_tcg_memop_mask; @@ -1746,6 +1748,18 @@ static inline void gen_op_addr_add (DisasContext *ctx, TCGv ret, TCGv arg0, TCGv #endif } +static inline void gen_op_addr_addi(DisasContext *ctx, TCGv ret, TCGv base, + target_long ofs) +{ + tcg_gen_addi_tl(ret, base, ofs); + +#if defined(TARGET_MIPS64) + if (ctx->hflags & MIPS_HFLAG_AWRAP) { + tcg_gen_ext32s_i64(ret, ret); + } +#endif +} + /* Addresses computation (translation time) */ static target_long addr_add(DisasContext *ctx, target_long base, target_long offset) @@ -1902,6 +1916,59 @@ static inline void check_mvh(DisasContext *ctx) } #endif +/* + * This code generates a "reserved instruction" exception if the + * Config5 XNP bit is set. + */ +static inline void check_xnp(DisasContext *ctx) +{ + if (unlikely(ctx->CP0_Config5 & (1 << CP0C5_XNP))) { + generate_exception_end(ctx, EXCP_RI); + } +} + +/* + * This code generates a "reserved instruction" exception if the + * Config3 MT bit is NOT set. + */ +static inline void check_mt(DisasContext *ctx) +{ + if (unlikely(!(ctx->CP0_Config3 & (1 << CP0C3_MT)))) { + generate_exception_end(ctx, EXCP_RI); + } +} + +#ifndef CONFIG_USER_ONLY +/* + * This code generates a "coprocessor unusable" exception if CP0 is not + * available, and, if that is not the case, generates a "reserved instruction" + * exception if the Config5 MT bit is NOT set. This is needed for availability + * control of some of MT ASE instructions. + */ +static inline void check_cp0_mt(DisasContext *ctx) +{ + if (unlikely(!(ctx->hflags & MIPS_HFLAG_CP0))) { + generate_exception_err(ctx, EXCP_CpU, 0); + } else { + if (unlikely(!(ctx->CP0_Config3 & (1 << CP0C3_MT)))) { + generate_exception_err(ctx, EXCP_RI, 0); + } + } +} +#endif + +/* + * This code generates a "reserved instruction" exception if the + * Config5 NMS bit is set. + */ +static inline void check_nms(DisasContext *ctx) +{ + if (unlikely(ctx->CP0_Config5 & (1 << CP0C5_NMS))) { + generate_exception_end(ctx, EXCP_RI); + } +} + + /* Define small wrappers for gen_load_fpr* so that we have a uniform calling interface for 32 and 64-bit FPRs. No sense in changing all callers for gen_load_fpr32 when we need the CTX parameter for @@ -2348,6 +2415,31 @@ static void gen_ld(DisasContext *ctx, uint32_t opc, tcg_temp_free(t0); } +static void gen_llwp(DisasContext *ctx, uint32_t base, int16_t offset, + uint32_t reg1, uint32_t reg2) +{ + TCGv taddr = tcg_temp_new(); + TCGv_i64 tval = tcg_temp_new_i64(); + TCGv tmp1 = tcg_temp_new(); + TCGv tmp2 = tcg_temp_new(); + + gen_base_offset_addr(ctx, taddr, base, offset); + tcg_gen_qemu_ld64(tval, taddr, ctx->mem_idx); +#ifdef TARGET_WORDS_BIGENDIAN + tcg_gen_extr_i64_tl(tmp2, tmp1, tval); +#else + tcg_gen_extr_i64_tl(tmp1, tmp2, tval); +#endif + gen_store_gpr(tmp1, reg1); + tcg_temp_free(tmp1); + gen_store_gpr(tmp2, reg2); + tcg_temp_free(tmp2); + tcg_gen_st_i64(tval, cpu_env, offsetof(CPUMIPSState, llval_wp)); + tcg_temp_free_i64(tval); + tcg_gen_st_tl(taddr, cpu_env, offsetof(CPUMIPSState, lladdr)); + tcg_temp_free(taddr); +} + /* Store */ static void gen_st (DisasContext *ctx, uint32_t opc, int rt, int base, int offset) @@ -2444,6 +2536,51 @@ static void gen_st_cond (DisasContext *ctx, uint32_t opc, int rt, tcg_temp_free(t0); } +static void gen_scwp(DisasContext *ctx, uint32_t base, int16_t offset, + uint32_t reg1, uint32_t reg2) +{ + TCGv taddr = tcg_temp_local_new(); + TCGv lladdr = tcg_temp_local_new(); + TCGv_i64 tval = tcg_temp_new_i64(); + TCGv_i64 llval = tcg_temp_new_i64(); + TCGv_i64 val = tcg_temp_new_i64(); + TCGv tmp1 = tcg_temp_new(); + TCGv tmp2 = tcg_temp_new(); + TCGLabel *lab_fail = gen_new_label(); + TCGLabel *lab_done = gen_new_label(); + + gen_base_offset_addr(ctx, taddr, base, offset); + + tcg_gen_ld_tl(lladdr, cpu_env, offsetof(CPUMIPSState, lladdr)); + tcg_gen_brcond_tl(TCG_COND_NE, taddr, lladdr, lab_fail); + + gen_load_gpr(tmp1, reg1); + gen_load_gpr(tmp2, reg2); + +#ifdef TARGET_WORDS_BIGENDIAN + tcg_gen_concat_tl_i64(tval, tmp2, tmp1); +#else + tcg_gen_concat_tl_i64(tval, tmp1, tmp2); +#endif + + tcg_gen_ld_i64(llval, cpu_env, offsetof(CPUMIPSState, llval_wp)); + tcg_gen_atomic_cmpxchg_i64(val, taddr, llval, tval, + ctx->mem_idx, MO_64); + if (reg1 != 0) { + tcg_gen_movi_tl(cpu_gpr[reg1], 1); + } + tcg_gen_brcond_i64(TCG_COND_EQ, val, llval, lab_done); + + gen_set_label(lab_fail); + + if (reg1 != 0) { + tcg_gen_movi_tl(cpu_gpr[reg1], 0); + } + gen_set_label(lab_done); + tcg_gen_movi_tl(lladdr, -1); + tcg_gen_st_tl(lladdr, cpu_env, offsetof(CPUMIPSState, lladdr)); +} + /* Load and store */ static void gen_flt_ldst (DisasContext *ctx, uint32_t opc, int ft, TCGv t0) @@ -4564,6 +4701,128 @@ static void gen_compute_branch (DisasContext *ctx, uint32_t opc, tcg_temp_free(t1); } + +/* nanoMIPS Branches */ +static void gen_compute_branch_nm(DisasContext *ctx, uint32_t opc, + int insn_bytes, + int rs, int rt, int32_t offset) +{ + target_ulong btgt = -1; + int bcond_compute = 0; + TCGv t0 = tcg_temp_new(); + TCGv t1 = tcg_temp_new(); + + /* Load needed operands */ + switch (opc) { + case OPC_BEQ: + case OPC_BNE: + /* Compare two registers */ + if (rs != rt) { + gen_load_gpr(t0, rs); + gen_load_gpr(t1, rt); + bcond_compute = 1; + } + btgt = ctx->base.pc_next + insn_bytes + offset; + break; + case OPC_BGEZAL: + /* Compare to zero */ + if (rs != 0) { + gen_load_gpr(t0, rs); + bcond_compute = 1; + } + btgt = ctx->base.pc_next + insn_bytes + offset; + break; + case OPC_BPOSGE32: + tcg_gen_andi_tl(t0, cpu_dspctrl, 0x3F); + bcond_compute = 1; + btgt = ctx->base.pc_next + insn_bytes + offset; + break; + case OPC_JR: + case OPC_JALR: + /* Jump to register */ + if (offset != 0 && offset != 16) { + /* Hint = 0 is JR/JALR, hint 16 is JR.HB/JALR.HB, the + others are reserved. */ + MIPS_INVAL("jump hint"); + generate_exception_end(ctx, EXCP_RI); + goto out; + } + gen_load_gpr(btarget, rs); + break; + default: + MIPS_INVAL("branch/jump"); + generate_exception_end(ctx, EXCP_RI); + goto out; + } + if (bcond_compute == 0) { + /* No condition to be computed */ + switch (opc) { + case OPC_BEQ: /* rx == rx */ + /* Always take */ + ctx->hflags |= MIPS_HFLAG_B; + break; + case OPC_BGEZAL: /* 0 >= 0 */ + /* Always take and link */ + tcg_gen_movi_tl(cpu_gpr[31], + ctx->base.pc_next + insn_bytes); + ctx->hflags |= MIPS_HFLAG_B; + break; + case OPC_BNE: /* rx != rx */ + tcg_gen_movi_tl(cpu_gpr[31], ctx->base.pc_next + 8); + /* Skip the instruction in the delay slot */ + ctx->base.pc_next += 4; + goto out; + case OPC_JR: + ctx->hflags |= MIPS_HFLAG_BR; + break; + case OPC_JALR: + if (rt > 0) { + tcg_gen_movi_tl(cpu_gpr[rt], + ctx->base.pc_next + insn_bytes); + } + ctx->hflags |= MIPS_HFLAG_BR; + break; + default: + MIPS_INVAL("branch/jump"); + generate_exception_end(ctx, EXCP_RI); + goto out; + } + } else { + switch (opc) { + case OPC_BEQ: + tcg_gen_setcond_tl(TCG_COND_EQ, bcond, t0, t1); + goto not_likely; + case OPC_BNE: + tcg_gen_setcond_tl(TCG_COND_NE, bcond, t0, t1); + goto not_likely; + case OPC_BGEZAL: + tcg_gen_setcondi_tl(TCG_COND_GE, bcond, t0, 0); + tcg_gen_movi_tl(cpu_gpr[31], + ctx->base.pc_next + insn_bytes); + goto not_likely; + case OPC_BPOSGE32: + tcg_gen_setcondi_tl(TCG_COND_GE, bcond, t0, 32); + not_likely: + ctx->hflags |= MIPS_HFLAG_BC; + break; + default: + MIPS_INVAL("conditional branch/jump"); + generate_exception_end(ctx, EXCP_RI); + goto out; + } + } + + ctx->btarget = btgt; + + out: + if (insn_bytes == 2) { + ctx->hflags |= MIPS_HFLAG_B16; + } + tcg_temp_free(t0); + tcg_temp_free(t1); +} + + /* special3 bitfield operations */ static void gen_bitops (DisasContext *ctx, uint32_t opc, int rt, int rs, int lsb, int msb) @@ -4736,8 +4995,8 @@ static void gen_lsa(DisasContext *ctx, int opc, int rd, int rs, int rt, return; } -static void gen_align(DisasContext *ctx, int opc, int rd, int rs, int rt, - int bp) +static void gen_align_bits(DisasContext *ctx, int wordsz, int rd, int rs, + int rt, int bits) { TCGv t0; if (rd == 0) { @@ -4745,35 +5004,40 @@ static void gen_align(DisasContext *ctx, int opc, int rd, int rs, int rt, return; } t0 = tcg_temp_new(); - gen_load_gpr(t0, rt); - if (bp == 0) { - switch (opc) { - case OPC_ALIGN: + if (bits == 0 || bits == wordsz) { + if (bits == 0) { + gen_load_gpr(t0, rt); + } else { + gen_load_gpr(t0, rs); + } + switch (wordsz) { + case 32: tcg_gen_ext32s_tl(cpu_gpr[rd], t0); break; #if defined(TARGET_MIPS64) - case OPC_DALIGN: + case 64: tcg_gen_mov_tl(cpu_gpr[rd], t0); break; #endif } } else { TCGv t1 = tcg_temp_new(); + gen_load_gpr(t0, rt); gen_load_gpr(t1, rs); - switch (opc) { - case OPC_ALIGN: + switch (wordsz) { + case 32: { TCGv_i64 t2 = tcg_temp_new_i64(); tcg_gen_concat_tl_i64(t2, t1, t0); - tcg_gen_shri_i64(t2, t2, 8 * (4 - bp)); + tcg_gen_shri_i64(t2, t2, 32 - bits); gen_move_low32(cpu_gpr[rd], t2); tcg_temp_free_i64(t2); } break; #if defined(TARGET_MIPS64) - case OPC_DALIGN: - tcg_gen_shli_tl(t0, t0, 8 * bp); - tcg_gen_shri_tl(t1, t1, 8 * (8 - bp)); + case 64: + tcg_gen_shli_tl(t0, t0, bits); + tcg_gen_shri_tl(t1, t1, 64 - bits); tcg_gen_or_tl(cpu_gpr[rd], t1, t0); break; #endif @@ -4784,6 +5048,18 @@ static void gen_align(DisasContext *ctx, int opc, int rd, int rs, int rt, tcg_temp_free(t0); } +static void gen_align(DisasContext *ctx, int wordsz, int rd, int rs, int rt, + int bp) +{ + gen_align_bits(ctx, wordsz, rd, rs, rt, bp * 8); +} + +static void gen_ext(DisasContext *ctx, int wordsz, int rd, int rs, int rt, + int shift) +{ + gen_align_bits(ctx, wordsz, rd, rs, rt, wordsz - shift); +} + static void gen_bitswap(DisasContext *ctx, int opc, int rd, int rt) { TCGv t0; @@ -8360,7 +8636,7 @@ static void gen_cp0 (CPUMIPSState *env, DisasContext *ctx, uint32_t opc, int rt, opn = "mthc0"; break; case OPC_MFTR: - check_insn(ctx, ASE_MT); + check_cp0_enabled(ctx); if (rd == 0) { /* Treat as NOP. */ return; @@ -8370,7 +8646,7 @@ static void gen_cp0 (CPUMIPSState *env, DisasContext *ctx, uint32_t opc, int rt, opn = "mftr"; break; case OPC_MTTR: - check_insn(ctx, ASE_MT); + check_cp0_enabled(ctx); gen_mttr(env, ctx, rd, rt, (ctx->opcode >> 5) & 1, ctx->opcode & 0x7, (ctx->opcode >> 4) & 1); opn = "mttr"; @@ -14276,8 +14552,7 @@ static void decode_micromips32_opc(CPUMIPSState *env, DisasContext *ctx) break; case ALIGN: check_insn(ctx, ISA_MIPS32R6); - gen_align(ctx, OPC_ALIGN, rd, rs, rt, - extract32(ctx->opcode, 9, 2)); + gen_align(ctx, 32, rd, rs, rt, extract32(ctx->opcode, 9, 2)); break; case EXT: gen_bitops(ctx, OPC_EXT, rt, rs, rr, rd); @@ -15701,6 +15976,4569 @@ static int decode_micromips_opc (CPUMIPSState *env, DisasContext *ctx) return 2; } +/* + * + * nanoMIPS opcodes + * + */ + +/* MAJOR, P16, and P32 pools opcodes */ +enum { + NM_P_ADDIU = 0x00, + NM_ADDIUPC = 0x01, + NM_MOVE_BALC = 0x02, + NM_P16_MV = 0x04, + NM_LW16 = 0x05, + NM_BC16 = 0x06, + NM_P16_SR = 0x07, + + NM_POOL32A = 0x08, + NM_P_BAL = 0x0a, + NM_P16_SHIFT = 0x0c, + NM_LWSP16 = 0x0d, + NM_BALC16 = 0x0e, + NM_P16_4X4 = 0x0f, + + NM_P_GP_W = 0x10, + NM_P_GP_BH = 0x11, + NM_P_J = 0x12, + NM_P16C = 0x14, + NM_LWGP16 = 0x15, + NM_P16_LB = 0x17, + + NM_P48I = 0x18, + NM_P16_A1 = 0x1c, + NM_LW4X4 = 0x1d, + NM_P16_LH = 0x1f, + + NM_P_U12 = 0x20, + NM_P_LS_U12 = 0x21, + NM_P_BR1 = 0x22, + NM_P16_A2 = 0x24, + NM_SW16 = 0x25, + NM_BEQZC16 = 0x26, + + NM_POOL32F = 0x28, + NM_P_LS_S9 = 0x29, + NM_P_BR2 = 0x2a, + + NM_P16_ADDU = 0x2c, + NM_SWSP16 = 0x2d, + NM_BNEZC16 = 0x2e, + NM_MOVEP = 0x2f, + + NM_POOL32S = 0x30, + NM_P_BRI = 0x32, + NM_LI16 = 0x34, + NM_SWGP16 = 0x35, + NM_P16_BR = 0x36, + + NM_P_LUI = 0x38, + NM_ANDI16 = 0x3c, + NM_SW4X4 = 0x3d, + NM_MOVEPREV = 0x3f, +}; + +/* POOL32A instruction pool */ +enum { + NM_POOL32A0 = 0x00, + NM_SPECIAL2 = 0x01, + NM_COP2_1 = 0x02, + NM_UDI = 0x03, + NM_POOL32A5 = 0x05, + NM_POOL32A7 = 0x07, +}; + +/* P.GP.W instruction pool */ +enum { + NM_ADDIUGP_W = 0x00, + NM_LWGP = 0x02, + NM_SWGP = 0x03, +}; + +/* P48I instruction pool */ +enum { + NM_LI48 = 0x00, + NM_ADDIU48 = 0x01, + NM_ADDIUGP48 = 0x02, + NM_ADDIUPC48 = 0x03, + NM_LWPC48 = 0x0b, + NM_SWPC48 = 0x0f, +}; + +/* P.U12 instruction pool */ +enum { + NM_ORI = 0x00, + NM_XORI = 0x01, + NM_ANDI = 0x02, + NM_P_SR = 0x03, + NM_SLTI = 0x04, + NM_SLTIU = 0x05, + NM_SEQI = 0x06, + NM_ADDIUNEG = 0x08, + NM_P_SHIFT = 0x0c, + NM_P_ROTX = 0x0d, + NM_P_INS = 0x0e, + NM_P_EXT = 0x0f, +}; + +/* POOL32F instruction pool */ +enum { + NM_POOL32F_0 = 0x00, + NM_POOL32F_3 = 0x03, + NM_POOL32F_5 = 0x05, +}; + +/* POOL32S instruction pool */ +enum { + NM_POOL32S_0 = 0x00, + NM_POOL32S_4 = 0x04, +}; + +/* P.LUI instruction pool */ +enum { + NM_LUI = 0x00, + NM_ALUIPC = 0x01, +}; + +/* P.GP.BH instruction pool */ +enum { + NM_LBGP = 0x00, + NM_SBGP = 0x01, + NM_LBUGP = 0x02, + NM_ADDIUGP_B = 0x03, + NM_P_GP_LH = 0x04, + NM_P_GP_SH = 0x05, + NM_P_GP_CP1 = 0x06, +}; + +/* P.LS.U12 instruction pool */ +enum { + NM_LB = 0x00, + NM_SB = 0x01, + NM_LBU = 0x02, + NM_P_PREFU12 = 0x03, + NM_LH = 0x04, + NM_SH = 0x05, + NM_LHU = 0x06, + NM_LWU = 0x07, + NM_LW = 0x08, + NM_SW = 0x09, + NM_LWC1 = 0x0a, + NM_SWC1 = 0x0b, + NM_LDC1 = 0x0e, + NM_SDC1 = 0x0f, +}; + +/* P.LS.S9 instruction pool */ +enum { + NM_P_LS_S0 = 0x00, + NM_P_LS_S1 = 0x01, + NM_P_LS_E0 = 0x02, + NM_P_LS_WM = 0x04, + NM_P_LS_UAWM = 0x05, +}; + +/* P.BAL instruction pool */ +enum { + NM_BC = 0x00, + NM_BALC = 0x01, +}; + +/* P.J instruction pool */ +enum { + NM_JALRC = 0x00, + NM_JALRC_HB = 0x01, + NM_P_BALRSC = 0x08, +}; + +/* P.BR1 instruction pool */ +enum { + NM_BEQC = 0x00, + NM_P_BR3A = 0x01, + NM_BGEC = 0x02, + NM_BGEUC = 0x03, +}; + +/* P.BR2 instruction pool */ +enum { + NM_BNEC = 0x00, + NM_BLTC = 0x02, + NM_BLTUC = 0x03, +}; + +/* P.BRI instruction pool */ +enum { + NM_BEQIC = 0x00, + NM_BBEQZC = 0x01, + NM_BGEIC = 0x02, + NM_BGEIUC = 0x03, + NM_BNEIC = 0x04, + NM_BBNEZC = 0x05, + NM_BLTIC = 0x06, + NM_BLTIUC = 0x07, +}; + +/* P16.SHIFT instruction pool */ +enum { + NM_SLL16 = 0x00, + NM_SRL16 = 0x01, +}; + +/* POOL16C instruction pool */ +enum { + NM_POOL16C_0 = 0x00, + NM_LWXS16 = 0x01, +}; + +/* P16.A1 instruction pool */ +enum { + NM_ADDIUR1SP = 0x01, +}; + +/* P16.A2 instruction pool */ +enum { + NM_ADDIUR2 = 0x00, + NM_P_ADDIURS5 = 0x01, +}; + +/* P16.ADDU instruction pool */ +enum { + NM_ADDU16 = 0x00, + NM_SUBU16 = 0x01, +}; + +/* P16.SR instruction pool */ +enum { + NM_SAVE16 = 0x00, + NM_RESTORE_JRC16 = 0x01, +}; + +/* P16.4X4 instruction pool */ +enum { + NM_ADDU4X4 = 0x00, + NM_MUL4X4 = 0x01, +}; + +/* P16.LB instruction pool */ +enum { + NM_LB16 = 0x00, + NM_SB16 = 0x01, + NM_LBU16 = 0x02, +}; + +/* P16.LH instruction pool */ +enum { + NM_LH16 = 0x00, + NM_SH16 = 0x01, + NM_LHU16 = 0x02, +}; + +/* P.RI instruction pool */ +enum { + NM_SIGRIE = 0x00, + NM_P_SYSCALL = 0x01, + NM_BREAK = 0x02, + NM_SDBBP = 0x03, +}; + +/* POOL32A0 instruction pool */ +enum { + NM_P_TRAP = 0x00, + NM_SEB = 0x01, + NM_SLLV = 0x02, + NM_MUL = 0x03, + NM_MFC0 = 0x06, + NM_MFHC0 = 0x07, + NM_SEH = 0x09, + NM_SRLV = 0x0a, + NM_MUH = 0x0b, + NM_MTC0 = 0x0e, + NM_MTHC0 = 0x0f, + NM_SRAV = 0x12, + NM_MULU = 0x13, + NM_ROTRV = 0x1a, + NM_MUHU = 0x1b, + NM_ADD = 0x22, + NM_DIV = 0x23, + NM_ADDU = 0x2a, + NM_MOD = 0x2b, + NM_SUB = 0x32, + NM_DIVU = 0x33, + NM_RDHWR = 0x38, + NM_SUBU = 0x3a, + NM_MODU = 0x3b, + NM_P_CMOVE = 0x42, + NM_FORK = 0x45, + NM_MFTR = 0x46, + NM_MFHTR = 0x47, + NM_AND = 0x4a, + NM_YIELD = 0x4d, + NM_MTTR = 0x4e, + NM_MTHTR = 0x4f, + NM_OR = 0x52, + NM_D_E_MT_VPE = 0x56, + NM_NOR = 0x5a, + NM_XOR = 0x62, + NM_SLT = 0x6a, + NM_P_SLTU = 0x72, + NM_SOV = 0x7a, +}; + +/* POOL32A5 instruction pool */ +enum { + NM_CMP_EQ_PH = 0x00, + NM_CMP_LT_PH = 0x08, + NM_CMP_LE_PH = 0x10, + NM_CMPGU_EQ_QB = 0x18, + NM_CMPGU_LT_QB = 0x20, + NM_CMPGU_LE_QB = 0x28, + NM_CMPGDU_EQ_QB = 0x30, + NM_CMPGDU_LT_QB = 0x38, + NM_CMPGDU_LE_QB = 0x40, + NM_CMPU_EQ_QB = 0x48, + NM_CMPU_LT_QB = 0x50, + NM_CMPU_LE_QB = 0x58, + NM_ADDQ_S_W = 0x60, + NM_SUBQ_S_W = 0x68, + NM_ADDSC = 0x70, + NM_ADDWC = 0x78, + + NM_ADDQ_S_PH = 0x01, + NM_ADDQH_R_PH = 0x09, + NM_ADDQH_R_W = 0x11, + NM_ADDU_S_QB = 0x19, + NM_ADDU_S_PH = 0x21, + NM_ADDUH_R_QB = 0x29, + NM_SHRAV_R_PH = 0x31, + NM_SHRAV_R_QB = 0x39, + NM_SUBQ_S_PH = 0x41, + NM_SUBQH_R_PH = 0x49, + NM_SUBQH_R_W = 0x51, + NM_SUBU_S_QB = 0x59, + NM_SUBU_S_PH = 0x61, + NM_SUBUH_R_QB = 0x69, + NM_SHLLV_S_PH = 0x71, + NM_PRECR_SRA_R_PH_W = 0x79, + + NM_MULEU_S_PH_QBL = 0x12, + NM_MULEU_S_PH_QBR = 0x1a, + NM_MULQ_RS_PH = 0x22, + NM_MULQ_S_PH = 0x2a, + NM_MULQ_RS_W = 0x32, + NM_MULQ_S_W = 0x3a, + NM_APPEND = 0x42, + NM_MODSUB = 0x52, + NM_SHRAV_R_W = 0x5a, + NM_SHRLV_PH = 0x62, + NM_SHRLV_QB = 0x6a, + NM_SHLLV_QB = 0x72, + NM_SHLLV_S_W = 0x7a, + + NM_SHILO = 0x03, + + NM_MULEQ_S_W_PHL = 0x04, + NM_MULEQ_S_W_PHR = 0x0c, + + NM_MUL_S_PH = 0x05, + NM_PRECR_QB_PH = 0x0d, + NM_PRECRQ_QB_PH = 0x15, + NM_PRECRQ_PH_W = 0x1d, + NM_PRECRQ_RS_PH_W = 0x25, + NM_PRECRQU_S_QB_PH = 0x2d, + NM_PACKRL_PH = 0x35, + NM_PICK_QB = 0x3d, + NM_PICK_PH = 0x45, + + NM_SHRA_R_W = 0x5e, + NM_SHRA_R_PH = 0x66, + NM_SHLL_S_PH = 0x76, + NM_SHLL_S_W = 0x7e, + + NM_REPL_PH = 0x07 +}; + +/* POOL32A7 instruction pool */ +enum { + NM_P_LSX = 0x00, + NM_LSA = 0x01, + NM_EXTW = 0x03, + NM_POOL32AXF = 0x07, +}; + +/* P.SR instruction pool */ +enum { + NM_PP_SR = 0x00, + NM_P_SR_F = 0x01, +}; + +/* P.SHIFT instruction pool */ +enum { + NM_P_SLL = 0x00, + NM_SRL = 0x02, + NM_SRA = 0x04, + NM_ROTR = 0x06, +}; + +/* P.ROTX instruction pool */ +enum { + NM_ROTX = 0x00, +}; + +/* P.INS instruction pool */ +enum { + NM_INS = 0x00, +}; + +/* P.EXT instruction pool */ +enum { + NM_EXT = 0x00, +}; + +/* POOL32F_0 (fmt) instruction pool */ +enum { + NM_RINT_S = 0x04, + NM_RINT_D = 0x44, + NM_ADD_S = 0x06, + NM_SELEQZ_S = 0x07, + NM_SELEQZ_D = 0x47, + NM_CLASS_S = 0x0c, + NM_CLASS_D = 0x4c, + NM_SUB_S = 0x0e, + NM_SELNEZ_S = 0x0f, + NM_SELNEZ_D = 0x4f, + NM_MUL_S = 0x16, + NM_SEL_S = 0x17, + NM_SEL_D = 0x57, + NM_DIV_S = 0x1e, + NM_ADD_D = 0x26, + NM_SUB_D = 0x2e, + NM_MUL_D = 0x36, + NM_MADDF_S = 0x37, + NM_MADDF_D = 0x77, + NM_DIV_D = 0x3e, + NM_MSUBF_S = 0x3f, + NM_MSUBF_D = 0x7f, +}; + +/* POOL32F_3 instruction pool */ +enum { + NM_MIN_FMT = 0x00, + NM_MAX_FMT = 0x01, + NM_MINA_FMT = 0x04, + NM_MAXA_FMT = 0x05, + NM_POOL32FXF = 0x07, +}; + +/* POOL32F_5 instruction pool */ +enum { + NM_CMP_CONDN_S = 0x00, + NM_CMP_CONDN_D = 0x02, +}; + +/* P.GP.LH instruction pool */ +enum { + NM_LHGP = 0x00, + NM_LHUGP = 0x01, +}; + +/* P.GP.SH instruction pool */ +enum { + NM_SHGP = 0x00, +}; + +/* P.GP.CP1 instruction pool */ +enum { + NM_LWC1GP = 0x00, + NM_SWC1GP = 0x01, + NM_LDC1GP = 0x02, + NM_SDC1GP = 0x03, +}; + +/* P.LS.S0 instruction pool */ +enum { + NM_LBS9 = 0x00, + NM_LHS9 = 0x04, + NM_LWS9 = 0x08, + NM_LDS9 = 0x0c, + + NM_SBS9 = 0x01, + NM_SHS9 = 0x05, + NM_SWS9 = 0x09, + NM_SDS9 = 0x0d, + + NM_LBUS9 = 0x02, + NM_LHUS9 = 0x06, + NM_LWC1S9 = 0x0a, + NM_LDC1S9 = 0x0e, + + NM_P_PREFS9 = 0x03, + NM_LWUS9 = 0x07, + NM_SWC1S9 = 0x0b, + NM_SDC1S9 = 0x0f, +}; + +/* P.LS.S1 instruction pool */ +enum { + NM_ASET_ACLR = 0x02, + NM_UALH = 0x04, + NM_UASH = 0x05, + NM_CACHE = 0x07, + NM_P_LL = 0x0a, + NM_P_SC = 0x0b, +}; + +/* P.LS.WM instruction pool */ +enum { + NM_LWM = 0x00, + NM_SWM = 0x01, +}; + +/* P.LS.UAWM instruction pool */ +enum { + NM_UALWM = 0x00, + NM_UASWM = 0x01, +}; + +/* P.BR3A instruction pool */ +enum { + NM_BC1EQZC = 0x00, + NM_BC1NEZC = 0x01, + NM_BC2EQZC = 0x02, + NM_BC2NEZC = 0x03, + NM_BPOSGE32C = 0x04, +}; + +/* P16.RI instruction pool */ +enum { + NM_P16_SYSCALL = 0x01, + NM_BREAK16 = 0x02, + NM_SDBBP16 = 0x03, +}; + +/* POOL16C_0 instruction pool */ +enum { + NM_POOL16C_00 = 0x00, +}; + +/* P16.JRC instruction pool */ +enum { + NM_JRC = 0x00, + NM_JALRC16 = 0x01, +}; + +/* P.SYSCALL instruction pool */ +enum { + NM_SYSCALL = 0x00, + NM_HYPCALL = 0x01, +}; + +/* P.TRAP instruction pool */ +enum { + NM_TEQ = 0x00, + NM_TNE = 0x01, +}; + +/* P.CMOVE instruction pool */ +enum { + NM_MOVZ = 0x00, + NM_MOVN = 0x01, +}; + +/* POOL32Axf instruction pool */ +enum { + NM_POOL32AXF_1 = 0x01, + NM_POOL32AXF_2 = 0x02, + NM_POOL32AXF_4 = 0x04, + NM_POOL32AXF_5 = 0x05, + NM_POOL32AXF_7 = 0x07, +}; + +/* POOL32Axf_1 instruction pool */ +enum { + NM_POOL32AXF_1_0 = 0x00, + NM_POOL32AXF_1_1 = 0x01, + NM_POOL32AXF_1_3 = 0x03, + NM_POOL32AXF_1_4 = 0x04, + NM_POOL32AXF_1_5 = 0x05, + NM_POOL32AXF_1_7 = 0x07, +}; + +/* POOL32Axf_2 instruction pool */ +enum { + NM_POOL32AXF_2_0_7 = 0x00, + NM_POOL32AXF_2_8_15 = 0x01, + NM_POOL32AXF_2_16_23 = 0x02, + NM_POOL32AXF_2_24_31 = 0x03, +}; + +/* POOL32Axf_7 instruction pool */ +enum { + NM_SHRA_R_QB = 0x0, + NM_SHRL_PH = 0x1, + NM_REPL_QB = 0x2, +}; + +/* POOL32Axf_1_0 instruction pool */ +enum { + NM_MFHI = 0x0, + NM_MFLO = 0x1, + NM_MTHI = 0x2, + NM_MTLO = 0x3, +}; + +/* POOL32Axf_1_1 instruction pool */ +enum { + NM_MTHLIP = 0x0, + NM_SHILOV = 0x1, +}; + +/* POOL32Axf_1_3 instruction pool */ +enum { + NM_RDDSP = 0x0, + NM_WRDSP = 0x1, + NM_EXTP = 0x2, + NM_EXTPDP = 0x3, +}; + +/* POOL32Axf_1_4 instruction pool */ +enum { + NM_SHLL_QB = 0x0, + NM_SHRL_QB = 0x1, +}; + +/* POOL32Axf_1_5 instruction pool */ +enum { + NM_MAQ_S_W_PHR = 0x0, + NM_MAQ_S_W_PHL = 0x1, + NM_MAQ_SA_W_PHR = 0x2, + NM_MAQ_SA_W_PHL = 0x3, +}; + +/* POOL32Axf_1_7 instruction pool */ +enum { + NM_EXTR_W = 0x0, + NM_EXTR_R_W = 0x1, + NM_EXTR_RS_W = 0x2, + NM_EXTR_S_H = 0x3, +}; + +/* POOL32Axf_2_0_7 instruction pool */ +enum { + NM_DPA_W_PH = 0x0, + NM_DPAQ_S_W_PH = 0x1, + NM_DPS_W_PH = 0x2, + NM_DPSQ_S_W_PH = 0x3, + NM_BALIGN = 0x4, + NM_MADD = 0x5, + NM_MULT = 0x6, + NM_EXTRV_W = 0x7, +}; + +/* POOL32Axf_2_8_15 instruction pool */ +enum { + NM_DPAX_W_PH = 0x0, + NM_DPAQ_SA_L_W = 0x1, + NM_DPSX_W_PH = 0x2, + NM_DPSQ_SA_L_W = 0x3, + NM_MADDU = 0x5, + NM_MULTU = 0x6, + NM_EXTRV_R_W = 0x7, +}; + +/* POOL32Axf_2_16_23 instruction pool */ +enum { + NM_DPAU_H_QBL = 0x0, + NM_DPAQX_S_W_PH = 0x1, + NM_DPSU_H_QBL = 0x2, + NM_DPSQX_S_W_PH = 0x3, + NM_EXTPV = 0x4, + NM_MSUB = 0x5, + NM_MULSA_W_PH = 0x6, + NM_EXTRV_RS_W = 0x7, +}; + +/* POOL32Axf_2_24_31 instruction pool */ +enum { + NM_DPAU_H_QBR = 0x0, + NM_DPAQX_SA_W_PH = 0x1, + NM_DPSU_H_QBR = 0x2, + NM_DPSQX_SA_W_PH = 0x3, + NM_EXTPDPV = 0x4, + NM_MSUBU = 0x5, + NM_MULSAQ_S_W_PH = 0x6, + NM_EXTRV_S_H = 0x7, +}; + +/* POOL32Axf_{4, 5} instruction pool */ +enum { + NM_CLO = 0x25, + NM_CLZ = 0x2d, + + NM_TLBP = 0x01, + NM_TLBR = 0x09, + NM_TLBWI = 0x11, + NM_TLBWR = 0x19, + NM_TLBINV = 0x03, + NM_TLBINVF = 0x0b, + NM_DI = 0x23, + NM_EI = 0x2b, + NM_RDPGPR = 0x70, + NM_WRPGPR = 0x78, + NM_WAIT = 0x61, + NM_DERET = 0x71, + NM_ERETX = 0x79, + + /* nanoMIPS DSP instructions */ + NM_ABSQ_S_QB = 0x00, + NM_ABSQ_S_PH = 0x08, + NM_ABSQ_S_W = 0x10, + NM_PRECEQ_W_PHL = 0x28, + NM_PRECEQ_W_PHR = 0x30, + NM_PRECEQU_PH_QBL = 0x38, + NM_PRECEQU_PH_QBR = 0x48, + NM_PRECEU_PH_QBL = 0x58, + NM_PRECEU_PH_QBR = 0x68, + NM_PRECEQU_PH_QBLA = 0x39, + NM_PRECEQU_PH_QBRA = 0x49, + NM_PRECEU_PH_QBLA = 0x59, + NM_PRECEU_PH_QBRA = 0x69, + NM_REPLV_PH = 0x01, + NM_REPLV_QB = 0x09, + NM_BITREV = 0x18, + NM_INSV = 0x20, + NM_RADDU_W_QB = 0x78, + + NM_BITSWAP = 0x05, + NM_WSBH = 0x3d, +}; + +/* PP.SR instruction pool */ +enum { + NM_SAVE = 0x00, + NM_RESTORE = 0x02, + NM_RESTORE_JRC = 0x03, +}; + +/* P.SR.F instruction pool */ +enum { + NM_SAVEF = 0x00, + NM_RESTOREF = 0x01, +}; + +/* P16.SYSCALL instruction pool */ +enum { + NM_SYSCALL16 = 0x00, + NM_HYPCALL16 = 0x01, +}; + +/* POOL16C_00 instruction pool */ +enum { + NM_NOT16 = 0x00, + NM_XOR16 = 0x01, + NM_AND16 = 0x02, + NM_OR16 = 0x03, +}; + +/* PP.LSX and PP.LSXS instruction pool */ +enum { + NM_LBX = 0x00, + NM_LHX = 0x04, + NM_LWX = 0x08, + NM_LDX = 0x0c, + + NM_SBX = 0x01, + NM_SHX = 0x05, + NM_SWX = 0x09, + NM_SDX = 0x0d, + + NM_LBUX = 0x02, + NM_LHUX = 0x06, + NM_LWC1X = 0x0a, + NM_LDC1X = 0x0e, + + NM_LWUX = 0x07, + NM_SWC1X = 0x0b, + NM_SDC1X = 0x0f, + + NM_LHXS = 0x04, + NM_LWXS = 0x08, + NM_LDXS = 0x0c, + + NM_SHXS = 0x05, + NM_SWXS = 0x09, + NM_SDXS = 0x0d, + + NM_LHUXS = 0x06, + NM_LWC1XS = 0x0a, + NM_LDC1XS = 0x0e, + + NM_LWUXS = 0x07, + NM_SWC1XS = 0x0b, + NM_SDC1XS = 0x0f, +}; + +/* ERETx instruction pool */ +enum { + NM_ERET = 0x00, + NM_ERETNC = 0x01, +}; + +/* POOL32FxF_{0, 1} insturction pool */ +enum { + NM_CFC1 = 0x40, + NM_CTC1 = 0x60, + NM_MFC1 = 0x80, + NM_MTC1 = 0xa0, + NM_MFHC1 = 0xc0, + NM_MTHC1 = 0xe0, + + NM_CVT_S_PL = 0x84, + NM_CVT_S_PU = 0xa4, + + NM_CVT_L_S = 0x004, + NM_CVT_L_D = 0x104, + NM_CVT_W_S = 0x024, + NM_CVT_W_D = 0x124, + + NM_RSQRT_S = 0x008, + NM_RSQRT_D = 0x108, + + NM_SQRT_S = 0x028, + NM_SQRT_D = 0x128, + + NM_RECIP_S = 0x048, + NM_RECIP_D = 0x148, + + NM_FLOOR_L_S = 0x00c, + NM_FLOOR_L_D = 0x10c, + + NM_FLOOR_W_S = 0x02c, + NM_FLOOR_W_D = 0x12c, + + NM_CEIL_L_S = 0x04c, + NM_CEIL_L_D = 0x14c, + NM_CEIL_W_S = 0x06c, + NM_CEIL_W_D = 0x16c, + NM_TRUNC_L_S = 0x08c, + NM_TRUNC_L_D = 0x18c, + NM_TRUNC_W_S = 0x0ac, + NM_TRUNC_W_D = 0x1ac, + NM_ROUND_L_S = 0x0cc, + NM_ROUND_L_D = 0x1cc, + NM_ROUND_W_S = 0x0ec, + NM_ROUND_W_D = 0x1ec, + + NM_MOV_S = 0x01, + NM_MOV_D = 0x81, + NM_ABS_S = 0x0d, + NM_ABS_D = 0x8d, + NM_NEG_S = 0x2d, + NM_NEG_D = 0xad, + NM_CVT_D_S = 0x04d, + NM_CVT_D_W = 0x0cd, + NM_CVT_D_L = 0x14d, + NM_CVT_S_D = 0x06d, + NM_CVT_S_W = 0x0ed, + NM_CVT_S_L = 0x16d, +}; + +/* P.LL instruction pool */ +enum { + NM_LL = 0x00, + NM_LLWP = 0x01, +}; + +/* P.SC instruction pool */ +enum { + NM_SC = 0x00, + NM_SCWP = 0x01, +}; + +/* P.DVP instruction pool */ +enum { + NM_DVP = 0x00, + NM_EVP = 0x01, +}; + + +/* + * + * nanoMIPS decoding engine + * + */ + + +/* extraction utilities */ + +#define NANOMIPS_EXTRACT_RD(op) ((op >> 7) & 0x7) +#define NANOMIPS_EXTRACT_RS(op) ((op >> 4) & 0x7) +#define NANOMIPS_EXTRACT_RS2(op) uMIPS_RS(op) +#define NANOMIPS_EXTRACT_RS1(op) ((op >> 1) & 0x7) +#define NANOMIPS_EXTRACT_RD5(op) ((op >> 5) & 0x1f) +#define NANOMIPS_EXTRACT_RS5(op) (op & 0x1f) + +/* Implement nanoMIPS pseudocode decode_gpr(encoded_gpr, 'gpr3'). */ +static inline int decode_gpr_gpr3(int r) +{ + static const int map[] = { 16, 17, 18, 19, 4, 5, 6, 7 }; + + return map[r & 0x7]; +} + +/* Implement nanoMIPS pseudocode decode_gpr(encoded_gpr, 'gpr3.src.store'). */ +static inline int decode_gpr_gpr3_src_store(int r) +{ + static const int map[] = { 0, 17, 18, 19, 4, 5, 6, 7 }; + + return map[r & 0x7]; +} + +/* Implement nanoMIPS pseudocode decode_gpr(encoded_gpr, 'gpr4'). */ +static inline int decode_gpr_gpr4(int r) +{ + static const int map[] = { 8, 9, 10, 11, 4, 5, 6, 7, + 16, 17, 18, 19, 20, 21, 22, 23 }; + + return map[r & 0xf]; +} + +/* Implement nanoMIPS pseudocode decode_gpr(encoded_gpr, 'gpr4.zero'). */ +static inline int decode_gpr_gpr4_zero(int r) +{ + static const int map[] = { 8, 9, 10, 0, 4, 5, 6, 7, + 16, 17, 18, 19, 20, 21, 22, 23 }; + + return map[r & 0xf]; +} + + +/* extraction utilities */ + +#define NANOMIPS_EXTRACT_RD(op) ((op >> 7) & 0x7) +#define NANOMIPS_EXTRACT_RS(op) ((op >> 4) & 0x7) +#define NANOMIPS_EXTRACT_RS2(op) uMIPS_RS(op) +#define NANOMIPS_EXTRACT_RS1(op) ((op >> 1) & 0x7) +#define NANOMIPS_EXTRACT_RD5(op) ((op >> 5) & 0x1f) +#define NANOMIPS_EXTRACT_RS5(op) (op & 0x1f) + + +static void gen_adjust_sp(DisasContext *ctx, int u) +{ + gen_op_addr_addi(ctx, cpu_gpr[29], cpu_gpr[29], u); +} + +static void gen_save(DisasContext *ctx, uint8_t rt, uint8_t count, + uint8_t gp, uint16_t u) +{ + int counter = 0; + TCGv va = tcg_temp_new(); + TCGv t0 = tcg_temp_new(); + + while (counter != count) { + bool use_gp = gp && (counter == count - 1); + int this_rt = use_gp ? 28 : (rt & 0x10) | ((rt + counter) & 0x1f); + int this_offset = -((counter + 1) << 2); + gen_base_offset_addr(ctx, va, 29, this_offset); + gen_load_gpr(t0, this_rt); + tcg_gen_qemu_st_tl(t0, va, ctx->mem_idx, + (MO_TEUL | ctx->default_tcg_memop_mask)); + counter++; + } + + /* adjust stack pointer */ + gen_adjust_sp(ctx, -u); + + tcg_temp_free(t0); + tcg_temp_free(va); +} + +static void gen_restore(DisasContext *ctx, uint8_t rt, uint8_t count, + uint8_t gp, uint16_t u) +{ + int counter = 0; + TCGv va = tcg_temp_new(); + TCGv t0 = tcg_temp_new(); + + while (counter != count) { + bool use_gp = gp && (counter == count - 1); + int this_rt = use_gp ? 28 : (rt & 0x10) | ((rt + counter) & 0x1f); + int this_offset = u - ((counter + 1) << 2); + gen_base_offset_addr(ctx, va, 29, this_offset); + tcg_gen_qemu_ld_tl(t0, va, ctx->mem_idx, MO_TESL | + ctx->default_tcg_memop_mask); + tcg_gen_ext32s_tl(t0, t0); + gen_store_gpr(t0, this_rt); + counter++; + } + + /* adjust stack pointer */ + gen_adjust_sp(ctx, u); + + tcg_temp_free(t0); + tcg_temp_free(va); +} + +static void gen_pool16c_nanomips_insn(DisasContext *ctx) +{ + int rt = decode_gpr_gpr3(NANOMIPS_EXTRACT_RD(ctx->opcode)); + int rs = decode_gpr_gpr3(NANOMIPS_EXTRACT_RS(ctx->opcode)); + + switch (extract32(ctx->opcode, 2, 2)) { + case NM_NOT16: + gen_logic(ctx, OPC_NOR, rt, rs, 0); + break; + case NM_AND16: + gen_logic(ctx, OPC_AND, rt, rt, rs); + break; + case NM_XOR16: + gen_logic(ctx, OPC_XOR, rt, rt, rs); + break; + case NM_OR16: + gen_logic(ctx, OPC_OR, rt, rt, rs); + break; + } +} + +static void gen_pool32a0_nanomips_insn(CPUMIPSState *env, DisasContext *ctx) +{ + int rt = extract32(ctx->opcode, 21, 5); + int rs = extract32(ctx->opcode, 16, 5); + int rd = extract32(ctx->opcode, 11, 5); + + switch (extract32(ctx->opcode, 3, 7)) { + case NM_P_TRAP: + switch (extract32(ctx->opcode, 10, 1)) { + case NM_TEQ: + check_nms(ctx); + gen_trap(ctx, OPC_TEQ, rs, rt, -1); + break; + case NM_TNE: + check_nms(ctx); + gen_trap(ctx, OPC_TNE, rs, rt, -1); + break; + } + break; + case NM_RDHWR: + check_nms(ctx); + gen_rdhwr(ctx, rt, rs, extract32(ctx->opcode, 11, 3)); + break; + case NM_SEB: + check_nms(ctx); + gen_bshfl(ctx, OPC_SEB, rs, rt); + break; + case NM_SEH: + gen_bshfl(ctx, OPC_SEH, rs, rt); + break; + case NM_SLLV: + gen_shift(ctx, OPC_SLLV, rd, rt, rs); + break; + case NM_SRLV: + gen_shift(ctx, OPC_SRLV, rd, rt, rs); + break; + case NM_SRAV: + gen_shift(ctx, OPC_SRAV, rd, rt, rs); + break; + case NM_ROTRV: + gen_shift(ctx, OPC_ROTRV, rd, rt, rs); + break; + case NM_ADD: + gen_arith(ctx, OPC_ADD, rd, rs, rt); + break; + case NM_ADDU: + gen_arith(ctx, OPC_ADDU, rd, rs, rt); + break; + case NM_SUB: + check_nms(ctx); + gen_arith(ctx, OPC_SUB, rd, rs, rt); + break; + case NM_SUBU: + gen_arith(ctx, OPC_SUBU, rd, rs, rt); + break; + case NM_P_CMOVE: + switch (extract32(ctx->opcode, 10, 1)) { + case NM_MOVZ: + gen_cond_move(ctx, OPC_MOVZ, rd, rs, rt); + break; + case NM_MOVN: + gen_cond_move(ctx, OPC_MOVN, rd, rs, rt); + break; + } + break; + case NM_AND: + gen_logic(ctx, OPC_AND, rd, rs, rt); + break; + case NM_OR: + gen_logic(ctx, OPC_OR, rd, rs, rt); + break; + case NM_NOR: + gen_logic(ctx, OPC_NOR, rd, rs, rt); + break; + case NM_XOR: + gen_logic(ctx, OPC_XOR, rd, rs, rt); + break; + case NM_SLT: + gen_slt(ctx, OPC_SLT, rd, rs, rt); + break; + case NM_P_SLTU: + if (rd == 0) { + /* P_DVP */ +#ifndef CONFIG_USER_ONLY + TCGv t0 = tcg_temp_new(); + switch (extract32(ctx->opcode, 10, 1)) { + case NM_DVP: + if (ctx->vp) { + check_cp0_enabled(ctx); + gen_helper_dvp(t0, cpu_env); + gen_store_gpr(t0, rt); + } + break; + case NM_EVP: + if (ctx->vp) { + check_cp0_enabled(ctx); + gen_helper_evp(t0, cpu_env); + gen_store_gpr(t0, rt); + } + break; + } + tcg_temp_free(t0); +#endif + } else { + gen_slt(ctx, OPC_SLTU, rd, rs, rt); + } + break; + case NM_SOV: + { + TCGv t0 = tcg_temp_new(); + TCGv t1 = tcg_temp_new(); + TCGv t2 = tcg_temp_new(); + + gen_load_gpr(t1, rs); + gen_load_gpr(t2, rt); + tcg_gen_add_tl(t0, t1, t2); + tcg_gen_ext32s_tl(t0, t0); + tcg_gen_xor_tl(t1, t1, t2); + tcg_gen_xor_tl(t2, t0, t2); + tcg_gen_andc_tl(t1, t2, t1); + + /* operands of same sign, result different sign */ + tcg_gen_setcondi_tl(TCG_COND_LT, t0, t1, 0); + gen_store_gpr(t0, rd); + + tcg_temp_free(t0); + tcg_temp_free(t1); + tcg_temp_free(t2); + } + break; + case NM_MUL: + gen_r6_muldiv(ctx, R6_OPC_MUL, rd, rs, rt); + break; + case NM_MUH: + gen_r6_muldiv(ctx, R6_OPC_MUH, rd, rs, rt); + break; + case NM_MULU: + gen_r6_muldiv(ctx, R6_OPC_MULU, rd, rs, rt); + break; + case NM_MUHU: + gen_r6_muldiv(ctx, R6_OPC_MUHU, rd, rs, rt); + break; + case NM_DIV: + gen_r6_muldiv(ctx, R6_OPC_DIV, rd, rs, rt); + break; + case NM_MOD: + gen_r6_muldiv(ctx, R6_OPC_MOD, rd, rs, rt); + break; + case NM_DIVU: + gen_r6_muldiv(ctx, R6_OPC_DIVU, rd, rs, rt); + break; + case NM_MODU: + gen_r6_muldiv(ctx, R6_OPC_MODU, rd, rs, rt); + break; +#ifndef CONFIG_USER_ONLY + case NM_MFC0: + check_cp0_enabled(ctx); + if (rt == 0) { + /* Treat as NOP. */ + break; + } + gen_mfc0(ctx, cpu_gpr[rt], rs, extract32(ctx->opcode, 11, 3)); + break; + case NM_MTC0: + check_cp0_enabled(ctx); + { + TCGv t0 = tcg_temp_new(); + + gen_load_gpr(t0, rt); + gen_mtc0(ctx, t0, rs, extract32(ctx->opcode, 11, 3)); + tcg_temp_free(t0); + } + break; + case NM_D_E_MT_VPE: + { + uint8_t sc = extract32(ctx->opcode, 10, 1); + TCGv t0 = tcg_temp_new(); + + switch (sc) { + case 0: + if (rs == 1) { + /* DMT */ + check_cp0_mt(ctx); + gen_helper_dmt(t0); + gen_store_gpr(t0, rt); + } else if (rs == 0) { + /* DVPE */ + check_cp0_mt(ctx); + gen_helper_dvpe(t0, cpu_env); + gen_store_gpr(t0, rt); + } else { + generate_exception_end(ctx, EXCP_RI); + } + break; + case 1: + if (rs == 1) { + /* EMT */ + check_cp0_mt(ctx); + gen_helper_emt(t0); + gen_store_gpr(t0, rt); + } else if (rs == 0) { + /* EVPE */ + check_cp0_mt(ctx); + gen_helper_evpe(t0, cpu_env); + gen_store_gpr(t0, rt); + } else { + generate_exception_end(ctx, EXCP_RI); + } + break; + } + + tcg_temp_free(t0); + } + break; + case NM_FORK: + check_mt(ctx); + { + TCGv t0 = tcg_temp_new(); + TCGv t1 = tcg_temp_new(); + + gen_load_gpr(t0, rt); + gen_load_gpr(t1, rs); + gen_helper_fork(t0, t1); + tcg_temp_free(t0); + tcg_temp_free(t1); + } + break; + case NM_MFTR: + case NM_MFHTR: + check_cp0_enabled(ctx); + if (rd == 0) { + /* Treat as NOP. */ + return; + } + gen_mftr(env, ctx, rs, rt, extract32(ctx->opcode, 10, 1), + extract32(ctx->opcode, 11, 5), extract32(ctx->opcode, 3, 1)); + break; + case NM_MTTR: + case NM_MTHTR: + check_cp0_enabled(ctx); + gen_mttr(env, ctx, rs, rt, extract32(ctx->opcode, 10, 1), + extract32(ctx->opcode, 11, 5), extract32(ctx->opcode, 3, 1)); + break; + case NM_YIELD: + check_mt(ctx); + { + TCGv t0 = tcg_temp_new(); + + gen_load_gpr(t0, rs); + gen_helper_yield(t0, cpu_env, t0); + gen_store_gpr(t0, rt); + tcg_temp_free(t0); + } + break; +#endif + default: + generate_exception_end(ctx, EXCP_RI); + break; + } +} + +/* dsp */ +static void gen_pool32axf_1_5_nanomips_insn(DisasContext *ctx, uint32_t opc, + int ret, int v1, int v2) +{ + TCGv_i32 t0; + TCGv v0_t; + TCGv v1_t; + + t0 = tcg_temp_new_i32(); + + v0_t = tcg_temp_new(); + v1_t = tcg_temp_new(); + + tcg_gen_movi_i32(t0, v2 >> 3); + + gen_load_gpr(v0_t, ret); + gen_load_gpr(v1_t, v1); + + switch (opc) { + case NM_MAQ_S_W_PHR: + check_dsp(ctx); + gen_helper_maq_s_w_phr(t0, v1_t, v0_t, cpu_env); + break; + case NM_MAQ_S_W_PHL: + check_dsp(ctx); + gen_helper_maq_s_w_phl(t0, v1_t, v0_t, cpu_env); + break; + case NM_MAQ_SA_W_PHR: + check_dsp(ctx); + gen_helper_maq_sa_w_phr(t0, v1_t, v0_t, cpu_env); + break; + case NM_MAQ_SA_W_PHL: + check_dsp(ctx); + gen_helper_maq_sa_w_phl(t0, v1_t, v0_t, cpu_env); + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + + tcg_temp_free_i32(t0); + + tcg_temp_free(v0_t); + tcg_temp_free(v1_t); +} + + +static void gen_pool32axf_1_nanomips_insn(DisasContext *ctx, uint32_t opc, + int ret, int v1, int v2) +{ + int16_t imm; + TCGv t0 = tcg_temp_new(); + TCGv t1 = tcg_temp_new(); + TCGv v0_t = tcg_temp_new(); + + gen_load_gpr(v0_t, v1); + + switch (opc) { + case NM_POOL32AXF_1_0: + check_dsp(ctx); + switch (extract32(ctx->opcode, 12, 2)) { + case NM_MFHI: + gen_HILO(ctx, OPC_MFHI, v2 >> 3, ret); + break; + case NM_MFLO: + gen_HILO(ctx, OPC_MFLO, v2 >> 3, ret); + break; + case NM_MTHI: + gen_HILO(ctx, OPC_MTHI, v2 >> 3, v1); + break; + case NM_MTLO: + gen_HILO(ctx, OPC_MTLO, v2 >> 3, v1); + break; + } + break; + case NM_POOL32AXF_1_1: + check_dsp(ctx); + switch (extract32(ctx->opcode, 12, 2)) { + case NM_MTHLIP: + tcg_gen_movi_tl(t0, v2); + gen_helper_mthlip(t0, v0_t, cpu_env); + break; + case NM_SHILOV: + tcg_gen_movi_tl(t0, v2 >> 3); + gen_helper_shilo(t0, v0_t, cpu_env); + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + break; + case NM_POOL32AXF_1_3: + check_dsp(ctx); + imm = extract32(ctx->opcode, 14, 7); + switch (extract32(ctx->opcode, 12, 2)) { + case NM_RDDSP: + tcg_gen_movi_tl(t0, imm); + gen_helper_rddsp(t0, t0, cpu_env); + gen_store_gpr(t0, ret); + break; + case NM_WRDSP: + gen_load_gpr(t0, ret); + tcg_gen_movi_tl(t1, imm); + gen_helper_wrdsp(t0, t1, cpu_env); + break; + case NM_EXTP: + tcg_gen_movi_tl(t0, v2 >> 3); + tcg_gen_movi_tl(t1, v1); + gen_helper_extp(t0, t0, t1, cpu_env); + gen_store_gpr(t0, ret); + break; + case NM_EXTPDP: + tcg_gen_movi_tl(t0, v2 >> 3); + tcg_gen_movi_tl(t1, v1); + gen_helper_extpdp(t0, t0, t1, cpu_env); + gen_store_gpr(t0, ret); + break; + } + break; + case NM_POOL32AXF_1_4: + check_dsp(ctx); + tcg_gen_movi_tl(t0, v2 >> 2); + switch (extract32(ctx->opcode, 12, 1)) { + case NM_SHLL_QB: + gen_helper_shll_qb(t0, t0, v0_t, cpu_env); + gen_store_gpr(t0, ret); + break; + case NM_SHRL_QB: + gen_helper_shrl_qb(t0, t0, v0_t); + gen_store_gpr(t0, ret); + break; + } + break; + case NM_POOL32AXF_1_5: + opc = extract32(ctx->opcode, 12, 2); + gen_pool32axf_1_5_nanomips_insn(ctx, opc, ret, v1, v2); + break; + case NM_POOL32AXF_1_7: + check_dsp(ctx); + tcg_gen_movi_tl(t0, v2 >> 3); + tcg_gen_movi_tl(t1, v1); + switch (extract32(ctx->opcode, 12, 2)) { + case NM_EXTR_W: + gen_helper_extr_w(t0, t0, t1, cpu_env); + gen_store_gpr(t0, ret); + break; + case NM_EXTR_R_W: + gen_helper_extr_r_w(t0, t0, t1, cpu_env); + gen_store_gpr(t0, ret); + break; + case NM_EXTR_RS_W: + gen_helper_extr_rs_w(t0, t0, t1, cpu_env); + gen_store_gpr(t0, ret); + break; + case NM_EXTR_S_H: + gen_helper_extr_s_h(t0, t0, t1, cpu_env); + gen_store_gpr(t0, ret); + break; + } + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + + tcg_temp_free(t0); + tcg_temp_free(t1); + tcg_temp_free(v0_t); +} + +static void gen_pool32axf_2_multiply(DisasContext *ctx, uint32_t opc, + TCGv v0, TCGv v1, int rd) +{ + TCGv_i32 t0; + + t0 = tcg_temp_new_i32(); + + tcg_gen_movi_i32(t0, rd >> 3); + + switch (opc) { + case NM_POOL32AXF_2_0_7: + switch (extract32(ctx->opcode, 9, 3)) { + case NM_DPA_W_PH: + check_dspr2(ctx); + gen_helper_dpa_w_ph(t0, v1, v0, cpu_env); + break; + case NM_DPAQ_S_W_PH: + check_dsp(ctx); + gen_helper_dpaq_s_w_ph(t0, v1, v0, cpu_env); + break; + case NM_DPS_W_PH: + check_dspr2(ctx); + gen_helper_dps_w_ph(t0, v1, v0, cpu_env); + break; + case NM_DPSQ_S_W_PH: + check_dsp(ctx); + gen_helper_dpsq_s_w_ph(t0, v1, v0, cpu_env); + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + break; + case NM_POOL32AXF_2_8_15: + switch (extract32(ctx->opcode, 9, 3)) { + case NM_DPAX_W_PH: + check_dspr2(ctx); + gen_helper_dpax_w_ph(t0, v0, v1, cpu_env); + break; + case NM_DPAQ_SA_L_W: + check_dsp(ctx); + gen_helper_dpaq_sa_l_w(t0, v0, v1, cpu_env); + break; + case NM_DPSX_W_PH: + check_dspr2(ctx); + gen_helper_dpsx_w_ph(t0, v0, v1, cpu_env); + break; + case NM_DPSQ_SA_L_W: + check_dsp(ctx); + gen_helper_dpsq_sa_l_w(t0, v0, v1, cpu_env); + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + break; + case NM_POOL32AXF_2_16_23: + switch (extract32(ctx->opcode, 9, 3)) { + case NM_DPAU_H_QBL: + check_dsp(ctx); + gen_helper_dpau_h_qbl(t0, v0, v1, cpu_env); + break; + case NM_DPAQX_S_W_PH: + check_dspr2(ctx); + gen_helper_dpaqx_s_w_ph(t0, v0, v1, cpu_env); + break; + case NM_DPSU_H_QBL: + check_dsp(ctx); + gen_helper_dpsu_h_qbl(t0, v0, v1, cpu_env); + break; + case NM_DPSQX_S_W_PH: + check_dspr2(ctx); + gen_helper_dpsqx_s_w_ph(t0, v0, v1, cpu_env); + break; + case NM_MULSA_W_PH: + check_dspr2(ctx); + gen_helper_mulsa_w_ph(t0, v0, v1, cpu_env); + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + break; + case NM_POOL32AXF_2_24_31: + switch (extract32(ctx->opcode, 9, 3)) { + case NM_DPAU_H_QBR: + check_dsp(ctx); + gen_helper_dpau_h_qbr(t0, v1, v0, cpu_env); + break; + case NM_DPAQX_SA_W_PH: + check_dspr2(ctx); + gen_helper_dpaqx_sa_w_ph(t0, v1, v0, cpu_env); + break; + case NM_DPSU_H_QBR: + check_dsp(ctx); + gen_helper_dpsu_h_qbr(t0, v1, v0, cpu_env); + break; + case NM_DPSQX_SA_W_PH: + check_dspr2(ctx); + gen_helper_dpsqx_sa_w_ph(t0, v1, v0, cpu_env); + break; + case NM_MULSAQ_S_W_PH: + check_dsp(ctx); + gen_helper_mulsaq_s_w_ph(t0, v1, v0, cpu_env); + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + + tcg_temp_free_i32(t0); +} + +static void gen_pool32axf_2_nanomips_insn(DisasContext *ctx, uint32_t opc, + int rt, int rs, int rd) +{ + int ret = rt; + TCGv t0 = tcg_temp_new(); + TCGv t1 = tcg_temp_new(); + TCGv v0_t = tcg_temp_new(); + TCGv v1_t = tcg_temp_new(); + + gen_load_gpr(v0_t, rt); + gen_load_gpr(v1_t, rs); + + switch (opc) { + case NM_POOL32AXF_2_0_7: + switch (extract32(ctx->opcode, 9, 3)) { + case NM_DPA_W_PH: + case NM_DPAQ_S_W_PH: + case NM_DPS_W_PH: + case NM_DPSQ_S_W_PH: + gen_pool32axf_2_multiply(ctx, opc, v0_t, v1_t, rd); + break; + case NM_BALIGN: + check_dspr2(ctx); + if (rt != 0) { + gen_load_gpr(t0, rs); + rd &= 3; + if (rd != 0 && rd != 2) { + tcg_gen_shli_tl(cpu_gpr[ret], cpu_gpr[ret], 8 * rd); + tcg_gen_ext32u_tl(t0, t0); + tcg_gen_shri_tl(t0, t0, 8 * (4 - rd)); + tcg_gen_or_tl(cpu_gpr[ret], cpu_gpr[ret], t0); + } + tcg_gen_ext32s_tl(cpu_gpr[ret], cpu_gpr[ret]); + } + break; + case NM_MADD: + check_dsp(ctx); + { + int acc = extract32(ctx->opcode, 14, 2); + TCGv_i64 t2 = tcg_temp_new_i64(); + TCGv_i64 t3 = tcg_temp_new_i64(); + + gen_load_gpr(t0, rt); + gen_load_gpr(t1, rs); + tcg_gen_ext_tl_i64(t2, t0); + tcg_gen_ext_tl_i64(t3, t1); + tcg_gen_mul_i64(t2, t2, t3); + tcg_gen_concat_tl_i64(t3, cpu_LO[acc], cpu_HI[acc]); + tcg_gen_add_i64(t2, t2, t3); + tcg_temp_free_i64(t3); + gen_move_low32(cpu_LO[acc], t2); + gen_move_high32(cpu_HI[acc], t2); + tcg_temp_free_i64(t2); + } + break; + case NM_MULT: + check_dsp(ctx); + { + int acc = extract32(ctx->opcode, 14, 2); + TCGv_i32 t2 = tcg_temp_new_i32(); + TCGv_i32 t3 = tcg_temp_new_i32(); + + gen_load_gpr(t0, rs); + gen_load_gpr(t1, rt); + tcg_gen_trunc_tl_i32(t2, t0); + tcg_gen_trunc_tl_i32(t3, t1); + tcg_gen_muls2_i32(t2, t3, t2, t3); + tcg_gen_ext_i32_tl(cpu_LO[acc], t2); + tcg_gen_ext_i32_tl(cpu_HI[acc], t3); + tcg_temp_free_i32(t2); + tcg_temp_free_i32(t3); + } + break; + case NM_EXTRV_W: + check_dsp(ctx); + gen_load_gpr(v1_t, rs); + tcg_gen_movi_tl(t0, rd >> 3); + gen_helper_extr_w(t0, t0, v1_t, cpu_env); + gen_store_gpr(t0, ret); + break; + } + break; + case NM_POOL32AXF_2_8_15: + switch (extract32(ctx->opcode, 9, 3)) { + case NM_DPAX_W_PH: + case NM_DPAQ_SA_L_W: + case NM_DPSX_W_PH: + case NM_DPSQ_SA_L_W: + gen_pool32axf_2_multiply(ctx, opc, v0_t, v1_t, rd); + break; + case NM_MADDU: + check_dsp(ctx); + { + int acc = extract32(ctx->opcode, 14, 2); + TCGv_i64 t2 = tcg_temp_new_i64(); + TCGv_i64 t3 = tcg_temp_new_i64(); + + gen_load_gpr(t0, rs); + gen_load_gpr(t1, rt); + tcg_gen_ext32u_tl(t0, t0); + tcg_gen_ext32u_tl(t1, t1); + tcg_gen_extu_tl_i64(t2, t0); + tcg_gen_extu_tl_i64(t3, t1); + tcg_gen_mul_i64(t2, t2, t3); + tcg_gen_concat_tl_i64(t3, cpu_LO[acc], cpu_HI[acc]); + tcg_gen_add_i64(t2, t2, t3); + tcg_temp_free_i64(t3); + gen_move_low32(cpu_LO[acc], t2); + gen_move_high32(cpu_HI[acc], t2); + tcg_temp_free_i64(t2); + } + break; + case NM_MULTU: + check_dsp(ctx); + { + int acc = extract32(ctx->opcode, 14, 2); + TCGv_i32 t2 = tcg_temp_new_i32(); + TCGv_i32 t3 = tcg_temp_new_i32(); + + gen_load_gpr(t0, rs); + gen_load_gpr(t1, rt); + tcg_gen_trunc_tl_i32(t2, t0); + tcg_gen_trunc_tl_i32(t3, t1); + tcg_gen_mulu2_i32(t2, t3, t2, t3); + tcg_gen_ext_i32_tl(cpu_LO[acc], t2); + tcg_gen_ext_i32_tl(cpu_HI[acc], t3); + tcg_temp_free_i32(t2); + tcg_temp_free_i32(t3); + } + break; + case NM_EXTRV_R_W: + check_dsp(ctx); + tcg_gen_movi_tl(t0, rd >> 3); + gen_helper_extr_r_w(t0, t0, v1_t, cpu_env); + gen_store_gpr(t0, ret); + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + break; + case NM_POOL32AXF_2_16_23: + switch (extract32(ctx->opcode, 9, 3)) { + case NM_DPAU_H_QBL: + case NM_DPAQX_S_W_PH: + case NM_DPSU_H_QBL: + case NM_DPSQX_S_W_PH: + case NM_MULSA_W_PH: + gen_pool32axf_2_multiply(ctx, opc, v0_t, v1_t, rd); + break; + case NM_EXTPV: + check_dsp(ctx); + tcg_gen_movi_tl(t0, rd >> 3); + gen_helper_extp(t0, t0, v1_t, cpu_env); + gen_store_gpr(t0, ret); + break; + case NM_MSUB: + check_dsp(ctx); + { + int acc = extract32(ctx->opcode, 14, 2); + TCGv_i64 t2 = tcg_temp_new_i64(); + TCGv_i64 t3 = tcg_temp_new_i64(); + + gen_load_gpr(t0, rs); + gen_load_gpr(t1, rt); + tcg_gen_ext_tl_i64(t2, t0); + tcg_gen_ext_tl_i64(t3, t1); + tcg_gen_mul_i64(t2, t2, t3); + tcg_gen_concat_tl_i64(t3, cpu_LO[acc], cpu_HI[acc]); + tcg_gen_sub_i64(t2, t3, t2); + tcg_temp_free_i64(t3); + gen_move_low32(cpu_LO[acc], t2); + gen_move_high32(cpu_HI[acc], t2); + tcg_temp_free_i64(t2); + } + break; + case NM_EXTRV_RS_W: + check_dsp(ctx); + tcg_gen_movi_tl(t0, rd >> 3); + gen_helper_extr_rs_w(t0, t0, v1_t, cpu_env); + gen_store_gpr(t0, ret); + break; + } + break; + case NM_POOL32AXF_2_24_31: + switch (extract32(ctx->opcode, 9, 3)) { + case NM_DPAU_H_QBR: + case NM_DPAQX_SA_W_PH: + case NM_DPSU_H_QBR: + case NM_DPSQX_SA_W_PH: + case NM_MULSAQ_S_W_PH: + gen_pool32axf_2_multiply(ctx, opc, v0_t, v1_t, rd); + break; + case NM_EXTPDPV: + check_dsp(ctx); + tcg_gen_movi_tl(t0, rd >> 3); + gen_helper_extpdp(t0, t0, v1_t, cpu_env); + gen_store_gpr(t0, ret); + break; + case NM_MSUBU: + check_dsp(ctx); + { + int acc = extract32(ctx->opcode, 14, 2); + TCGv_i64 t2 = tcg_temp_new_i64(); + TCGv_i64 t3 = tcg_temp_new_i64(); + + gen_load_gpr(t0, rs); + gen_load_gpr(t1, rt); + tcg_gen_ext32u_tl(t0, t0); + tcg_gen_ext32u_tl(t1, t1); + tcg_gen_extu_tl_i64(t2, t0); + tcg_gen_extu_tl_i64(t3, t1); + tcg_gen_mul_i64(t2, t2, t3); + tcg_gen_concat_tl_i64(t3, cpu_LO[acc], cpu_HI[acc]); + tcg_gen_sub_i64(t2, t3, t2); + tcg_temp_free_i64(t3); + gen_move_low32(cpu_LO[acc], t2); + gen_move_high32(cpu_HI[acc], t2); + tcg_temp_free_i64(t2); + } + break; + case NM_EXTRV_S_H: + check_dsp(ctx); + tcg_gen_movi_tl(t0, rd >> 3); + gen_helper_extr_s_h(t0, t0, v0_t, cpu_env); + gen_store_gpr(t0, ret); + break; + } + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + + tcg_temp_free(t0); + tcg_temp_free(t1); + + tcg_temp_free(v0_t); + tcg_temp_free(v1_t); +} + +static void gen_pool32axf_4_nanomips_insn(DisasContext *ctx, uint32_t opc, + int rt, int rs) +{ + int ret = rt; + TCGv t0 = tcg_temp_new(); + TCGv v0_t = tcg_temp_new(); + + gen_load_gpr(v0_t, rs); + + switch (opc) { + case NM_ABSQ_S_QB: + check_dspr2(ctx); + gen_helper_absq_s_qb(v0_t, v0_t, cpu_env); + gen_store_gpr(v0_t, ret); + break; + case NM_ABSQ_S_PH: + check_dsp(ctx); + gen_helper_absq_s_ph(v0_t, v0_t, cpu_env); + gen_store_gpr(v0_t, ret); + break; + case NM_ABSQ_S_W: + check_dsp(ctx); + gen_helper_absq_s_w(v0_t, v0_t, cpu_env); + gen_store_gpr(v0_t, ret); + break; + case NM_PRECEQ_W_PHL: + check_dsp(ctx); + tcg_gen_andi_tl(v0_t, v0_t, 0xFFFF0000); + tcg_gen_ext32s_tl(v0_t, v0_t); + gen_store_gpr(v0_t, ret); + break; + case NM_PRECEQ_W_PHR: + check_dsp(ctx); + tcg_gen_andi_tl(v0_t, v0_t, 0x0000FFFF); + tcg_gen_shli_tl(v0_t, v0_t, 16); + tcg_gen_ext32s_tl(v0_t, v0_t); + gen_store_gpr(v0_t, ret); + break; + case NM_PRECEQU_PH_QBL: + check_dsp(ctx); + gen_helper_precequ_ph_qbl(v0_t, v0_t); + gen_store_gpr(v0_t, ret); + break; + case NM_PRECEQU_PH_QBR: + check_dsp(ctx); + gen_helper_precequ_ph_qbr(v0_t, v0_t); + gen_store_gpr(v0_t, ret); + break; + case NM_PRECEQU_PH_QBLA: + check_dsp(ctx); + gen_helper_precequ_ph_qbla(v0_t, v0_t); + gen_store_gpr(v0_t, ret); + break; + case NM_PRECEQU_PH_QBRA: + check_dsp(ctx); + gen_helper_precequ_ph_qbra(v0_t, v0_t); + gen_store_gpr(v0_t, ret); + break; + case NM_PRECEU_PH_QBL: + check_dsp(ctx); + gen_helper_preceu_ph_qbl(v0_t, v0_t); + gen_store_gpr(v0_t, ret); + break; + case NM_PRECEU_PH_QBR: + check_dsp(ctx); + gen_helper_preceu_ph_qbr(v0_t, v0_t); + gen_store_gpr(v0_t, ret); + break; + case NM_PRECEU_PH_QBLA: + check_dsp(ctx); + gen_helper_preceu_ph_qbla(v0_t, v0_t); + gen_store_gpr(v0_t, ret); + break; + case NM_PRECEU_PH_QBRA: + check_dsp(ctx); + gen_helper_preceu_ph_qbra(v0_t, v0_t); + gen_store_gpr(v0_t, ret); + break; + case NM_REPLV_PH: + check_dsp(ctx); + tcg_gen_ext16u_tl(v0_t, v0_t); + tcg_gen_shli_tl(t0, v0_t, 16); + tcg_gen_or_tl(v0_t, v0_t, t0); + tcg_gen_ext32s_tl(v0_t, v0_t); + gen_store_gpr(v0_t, ret); + break; + case NM_REPLV_QB: + check_dsp(ctx); + tcg_gen_ext8u_tl(v0_t, v0_t); + tcg_gen_shli_tl(t0, v0_t, 8); + tcg_gen_or_tl(v0_t, v0_t, t0); + tcg_gen_shli_tl(t0, v0_t, 16); + tcg_gen_or_tl(v0_t, v0_t, t0); + tcg_gen_ext32s_tl(v0_t, v0_t); + gen_store_gpr(v0_t, ret); + break; + case NM_BITREV: + check_dsp(ctx); + gen_helper_bitrev(v0_t, v0_t); + gen_store_gpr(v0_t, ret); + break; + case NM_INSV: + check_dsp(ctx); + { + TCGv tv0 = tcg_temp_new(); + + gen_load_gpr(tv0, rt); + gen_helper_insv(v0_t, cpu_env, v0_t, tv0); + gen_store_gpr(v0_t, ret); + tcg_temp_free(tv0); + } + break; + case NM_RADDU_W_QB: + check_dsp(ctx); + gen_helper_raddu_w_qb(v0_t, v0_t); + gen_store_gpr(v0_t, ret); + break; + case NM_BITSWAP: + gen_bitswap(ctx, OPC_BITSWAP, ret, rs); + break; + case NM_CLO: + check_nms(ctx); + gen_cl(ctx, OPC_CLO, ret, rs); + break; + case NM_CLZ: + check_nms(ctx); + gen_cl(ctx, OPC_CLZ, ret, rs); + break; + case NM_WSBH: + gen_bshfl(ctx, OPC_WSBH, ret, rs); + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + + tcg_temp_free(v0_t); + tcg_temp_free(t0); +} + +static void gen_pool32axf_7_nanomips_insn(DisasContext *ctx, uint32_t opc, + int rt, int rs, int rd) +{ + TCGv t0 = tcg_temp_new(); + TCGv rs_t = tcg_temp_new(); + + gen_load_gpr(rs_t, rs); + + switch (opc) { + case NM_SHRA_R_QB: + check_dspr2(ctx); + tcg_gen_movi_tl(t0, rd >> 2); + switch (extract32(ctx->opcode, 12, 1)) { + case 0: + /* NM_SHRA_QB */ + gen_helper_shra_qb(t0, t0, rs_t); + gen_store_gpr(t0, rt); + break; + case 1: + /* NM_SHRA_R_QB */ + gen_helper_shra_r_qb(t0, t0, rs_t); + gen_store_gpr(t0, rt); + break; + } + break; + case NM_SHRL_PH: + check_dspr2(ctx); + tcg_gen_movi_tl(t0, rd >> 1); + gen_helper_shrl_ph(t0, t0, rs_t); + gen_store_gpr(t0, rt); + break; + case NM_REPL_QB: + check_dsp(ctx); + { + int16_t imm; + target_long result; + imm = extract32(ctx->opcode, 13, 8); + result = (uint32_t)imm << 24 | + (uint32_t)imm << 16 | + (uint32_t)imm << 8 | + (uint32_t)imm; + result = (int32_t)result; + tcg_gen_movi_tl(t0, result); + gen_store_gpr(t0, rt); + } + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + tcg_temp_free(t0); + tcg_temp_free(rs_t); +} + + +static void gen_pool32axf_nanomips_insn(CPUMIPSState *env, DisasContext *ctx) +{ + int rt = extract32(ctx->opcode, 21, 5); + int rs = extract32(ctx->opcode, 16, 5); + int rd = extract32(ctx->opcode, 11, 5); + + switch (extract32(ctx->opcode, 6, 3)) { + case NM_POOL32AXF_1: + { + int32_t op1 = extract32(ctx->opcode, 9, 3); + gen_pool32axf_1_nanomips_insn(ctx, op1, rt, rs, rd); + } + break; + case NM_POOL32AXF_2: + { + int32_t op1 = extract32(ctx->opcode, 12, 2); + gen_pool32axf_2_nanomips_insn(ctx, op1, rt, rs, rd); + } + break; + case NM_POOL32AXF_4: + { + int32_t op1 = extract32(ctx->opcode, 9, 7); + gen_pool32axf_4_nanomips_insn(ctx, op1, rt, rs); + } + break; + case NM_POOL32AXF_5: + switch (extract32(ctx->opcode, 9, 7)) { +#ifndef CONFIG_USER_ONLY + case NM_TLBP: + gen_cp0(env, ctx, OPC_TLBP, 0, 0); + break; + case NM_TLBR: + gen_cp0(env, ctx, OPC_TLBR, 0, 0); + break; + case NM_TLBWI: + gen_cp0(env, ctx, OPC_TLBWI, 0, 0); + break; + case NM_TLBWR: + gen_cp0(env, ctx, OPC_TLBWR, 0, 0); + break; + case NM_TLBINV: + gen_cp0(env, ctx, OPC_TLBINV, 0, 0); + break; + case NM_TLBINVF: + gen_cp0(env, ctx, OPC_TLBINVF, 0, 0); + break; + case NM_DI: + check_cp0_enabled(ctx); + { + TCGv t0 = tcg_temp_new(); + + save_cpu_state(ctx, 1); + gen_helper_di(t0, cpu_env); + gen_store_gpr(t0, rt); + /* Stop translation as we may have switched the execution mode */ + ctx->base.is_jmp = DISAS_STOP; + tcg_temp_free(t0); + } + break; + case NM_EI: + check_cp0_enabled(ctx); + { + TCGv t0 = tcg_temp_new(); + + save_cpu_state(ctx, 1); + gen_helper_ei(t0, cpu_env); + gen_store_gpr(t0, rt); + /* Stop translation as we may have switched the execution mode */ + ctx->base.is_jmp = DISAS_STOP; + tcg_temp_free(t0); + } + break; + case NM_RDPGPR: + gen_load_srsgpr(rs, rt); + break; + case NM_WRPGPR: + gen_store_srsgpr(rs, rt); + break; + case NM_WAIT: + gen_cp0(env, ctx, OPC_WAIT, 0, 0); + break; + case NM_DERET: + gen_cp0(env, ctx, OPC_DERET, 0, 0); + break; + case NM_ERETX: + gen_cp0(env, ctx, OPC_ERET, 0, 0); + break; +#endif + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + break; + case NM_POOL32AXF_7: + { + int32_t op1 = extract32(ctx->opcode, 9, 3); + gen_pool32axf_7_nanomips_insn(ctx, op1, rt, rs, rd); + } + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } +} + +/* Immediate Value Compact Branches */ +static void gen_compute_imm_branch(DisasContext *ctx, uint32_t opc, + int rt, int32_t imm, int32_t offset) +{ + TCGCond cond; + int bcond_compute = 0; + TCGv t0 = tcg_temp_new(); + TCGv t1 = tcg_temp_new(); + + gen_load_gpr(t0, rt); + tcg_gen_movi_tl(t1, imm); + ctx->btarget = addr_add(ctx, ctx->base.pc_next + 4, offset); + + /* Load needed operands and calculate btarget */ + switch (opc) { + case NM_BEQIC: + if (rt == 0 && imm == 0) { + /* Unconditional branch */ + } else if (rt == 0 && imm != 0) { + /* Treat as NOP */ + goto out; + } else { + bcond_compute = 1; + cond = TCG_COND_EQ; + } + break; + case NM_BBEQZC: + case NM_BBNEZC: + check_nms(ctx); + if (imm >= 32 && !(ctx->hflags & MIPS_HFLAG_64)) { + generate_exception_end(ctx, EXCP_RI); + goto out; + } else if (rt == 0 && opc == NM_BBEQZC) { + /* Unconditional branch */ + } else if (rt == 0 && opc == NM_BBNEZC) { + /* Treat as NOP */ + goto out; + } else { + tcg_gen_shri_tl(t0, t0, imm); + tcg_gen_andi_tl(t0, t0, 1); + tcg_gen_movi_tl(t1, 0); + bcond_compute = 1; + if (opc == NM_BBEQZC) { + cond = TCG_COND_EQ; + } else { + cond = TCG_COND_NE; + } + } + break; + case NM_BNEIC: + if (rt == 0 && imm == 0) { + /* Treat as NOP */ + goto out; + } else if (rt == 0 && imm != 0) { + /* Unconditional branch */ + } else { + bcond_compute = 1; + cond = TCG_COND_NE; + } + break; + case NM_BGEIC: + if (rt == 0 && imm == 0) { + /* Unconditional branch */ + } else { + bcond_compute = 1; + cond = TCG_COND_GE; + } + break; + case NM_BLTIC: + bcond_compute = 1; + cond = TCG_COND_LT; + break; + case NM_BGEIUC: + if (rt == 0 && imm == 0) { + /* Unconditional branch */ + } else { + bcond_compute = 1; + cond = TCG_COND_GEU; + } + break; + case NM_BLTIUC: + bcond_compute = 1; + cond = TCG_COND_LTU; + break; + default: + MIPS_INVAL("Immediate Value Compact branch"); + generate_exception_end(ctx, EXCP_RI); + goto out; + } + + if (bcond_compute == 0) { + /* Uncoditional compact branch */ + gen_goto_tb(ctx, 0, ctx->btarget); + } else { + /* Conditional compact branch */ + TCGLabel *fs = gen_new_label(); + + tcg_gen_brcond_tl(tcg_invert_cond(cond), t0, t1, fs); + + gen_goto_tb(ctx, 1, ctx->btarget); + gen_set_label(fs); + + gen_goto_tb(ctx, 0, ctx->base.pc_next + 4); + } + +out: + tcg_temp_free(t0); + tcg_temp_free(t1); +} + +/* P.BALRSC type nanoMIPS R6 branches: BALRSC and BRSC */ +static void gen_compute_nanomips_pbalrsc_branch(DisasContext *ctx, int rs, + int rt) +{ + TCGv t0 = tcg_temp_new(); + TCGv t1 = tcg_temp_new(); + + /* load rs */ + gen_load_gpr(t0, rs); + + /* link */ + if (rt != 0) { + tcg_gen_movi_tl(cpu_gpr[rt], ctx->base.pc_next + 4); + } + + /* calculate btarget */ + tcg_gen_shli_tl(t0, t0, 1); + tcg_gen_movi_tl(t1, ctx->base.pc_next + 4); + gen_op_addr_add(ctx, btarget, t1, t0); + + /* unconditional branch to register */ + tcg_gen_mov_tl(cpu_PC, btarget); + tcg_gen_lookup_and_goto_ptr(); + + tcg_temp_free(t0); + tcg_temp_free(t1); +} + +/* nanoMIPS Branches */ +static void gen_compute_compact_branch_nm(DisasContext *ctx, uint32_t opc, + int rs, int rt, int32_t offset) +{ + int bcond_compute = 0; + TCGv t0 = tcg_temp_new(); + TCGv t1 = tcg_temp_new(); + + /* Load needed operands and calculate btarget */ + switch (opc) { + /* compact branch */ + case OPC_BGEC: + case OPC_BLTC: + gen_load_gpr(t0, rs); + gen_load_gpr(t1, rt); + bcond_compute = 1; + ctx->btarget = addr_add(ctx, ctx->base.pc_next + 4, offset); + break; + case OPC_BGEUC: + case OPC_BLTUC: + if (rs == 0 || rs == rt) { + /* OPC_BLEZALC, OPC_BGEZALC */ + /* OPC_BGTZALC, OPC_BLTZALC */ + tcg_gen_movi_tl(cpu_gpr[31], ctx->base.pc_next + 4); + } + gen_load_gpr(t0, rs); + gen_load_gpr(t1, rt); + bcond_compute = 1; + ctx->btarget = addr_add(ctx, ctx->base.pc_next + 4, offset); + break; + case OPC_BC: + ctx->btarget = addr_add(ctx, ctx->base.pc_next + 4, offset); + break; + case OPC_BEQZC: + if (rs != 0) { + /* OPC_BEQZC, OPC_BNEZC */ + gen_load_gpr(t0, rs); + bcond_compute = 1; + ctx->btarget = addr_add(ctx, ctx->base.pc_next + 4, offset); + } else { + /* OPC_JIC, OPC_JIALC */ + TCGv tbase = tcg_temp_new(); + TCGv toffset = tcg_temp_new(); + + gen_load_gpr(tbase, rt); + tcg_gen_movi_tl(toffset, offset); + gen_op_addr_add(ctx, btarget, tbase, toffset); + tcg_temp_free(tbase); + tcg_temp_free(toffset); + } + break; + default: + MIPS_INVAL("Compact branch/jump"); + generate_exception_end(ctx, EXCP_RI); + goto out; + } + + if (bcond_compute == 0) { + /* Uncoditional compact branch */ + switch (opc) { + case OPC_BC: + gen_goto_tb(ctx, 0, ctx->btarget); + break; + default: + MIPS_INVAL("Compact branch/jump"); + generate_exception_end(ctx, EXCP_RI); + goto out; + } + } else { + /* Conditional compact branch */ + TCGLabel *fs = gen_new_label(); + + switch (opc) { + case OPC_BGEUC: + if (rs == 0 && rt != 0) { + /* OPC_BLEZALC */ + tcg_gen_brcondi_tl(tcg_invert_cond(TCG_COND_LE), t1, 0, fs); + } else if (rs != 0 && rt != 0 && rs == rt) { + /* OPC_BGEZALC */ + tcg_gen_brcondi_tl(tcg_invert_cond(TCG_COND_GE), t1, 0, fs); + } else { + /* OPC_BGEUC */ + tcg_gen_brcond_tl(tcg_invert_cond(TCG_COND_GEU), t0, t1, fs); + } + break; + case OPC_BLTUC: + if (rs == 0 && rt != 0) { + /* OPC_BGTZALC */ + tcg_gen_brcondi_tl(tcg_invert_cond(TCG_COND_GT), t1, 0, fs); + } else if (rs != 0 && rt != 0 && rs == rt) { + /* OPC_BLTZALC */ + tcg_gen_brcondi_tl(tcg_invert_cond(TCG_COND_LT), t1, 0, fs); + } else { + /* OPC_BLTUC */ + tcg_gen_brcond_tl(tcg_invert_cond(TCG_COND_LTU), t0, t1, fs); + } + break; + case OPC_BGEC: + if (rs == 0 && rt != 0) { + /* OPC_BLEZC */ + tcg_gen_brcondi_tl(tcg_invert_cond(TCG_COND_LE), t1, 0, fs); + } else if (rs != 0 && rt != 0 && rs == rt) { + /* OPC_BGEZC */ + tcg_gen_brcondi_tl(tcg_invert_cond(TCG_COND_GE), t1, 0, fs); + } else { + /* OPC_BGEC */ + tcg_gen_brcond_tl(tcg_invert_cond(TCG_COND_GE), t0, t1, fs); + } + break; + case OPC_BLTC: + if (rs == 0 && rt != 0) { + /* OPC_BGTZC */ + tcg_gen_brcondi_tl(tcg_invert_cond(TCG_COND_GT), t1, 0, fs); + } else if (rs != 0 && rt != 0 && rs == rt) { + /* OPC_BLTZC */ + tcg_gen_brcondi_tl(tcg_invert_cond(TCG_COND_LT), t1, 0, fs); + } else { + /* OPC_BLTC */ + tcg_gen_brcond_tl(tcg_invert_cond(TCG_COND_LT), t0, t1, fs); + } + break; + case OPC_BEQZC: + tcg_gen_brcondi_tl(tcg_invert_cond(TCG_COND_EQ), t0, 0, fs); + break; + default: + MIPS_INVAL("Compact conditional branch/jump"); + generate_exception_end(ctx, EXCP_RI); + goto out; + } + + /* Generating branch here as compact branches don't have delay slot */ + gen_goto_tb(ctx, 1, ctx->btarget); + gen_set_label(fs); + + gen_goto_tb(ctx, 0, ctx->base.pc_next + 4); + } + +out: + tcg_temp_free(t0); + tcg_temp_free(t1); +} + + +/* nanoMIPS CP1 Branches */ +static void gen_compute_branch_cp1_nm(DisasContext *ctx, uint32_t op, + int32_t ft, int32_t offset) +{ + target_ulong btarget; + TCGv_i64 t0 = tcg_temp_new_i64(); + + gen_load_fpr64(ctx, t0, ft); + tcg_gen_andi_i64(t0, t0, 1); + + btarget = addr_add(ctx, ctx->base.pc_next + 4, offset); + + switch (op) { + case NM_BC1EQZC: + tcg_gen_xori_i64(t0, t0, 1); + ctx->hflags |= MIPS_HFLAG_BC; + break; + case NM_BC1NEZC: + /* t0 already set */ + ctx->hflags |= MIPS_HFLAG_BC; + break; + default: + MIPS_INVAL("cp1 cond branch"); + generate_exception_end(ctx, EXCP_RI); + goto out; + } + + tcg_gen_trunc_i64_tl(bcond, t0); + + ctx->btarget = btarget; + +out: + tcg_temp_free_i64(t0); +} + + +static void gen_p_lsx(DisasContext *ctx, int rd, int rs, int rt) +{ + TCGv t0, t1; + t0 = tcg_temp_new(); + t1 = tcg_temp_new(); + + gen_load_gpr(t0, rs); + gen_load_gpr(t1, rt); + + if ((extract32(ctx->opcode, 6, 1)) == 1) { + /* PP.LSXS instructions require shifting */ + switch (extract32(ctx->opcode, 7, 4)) { + case NM_SHXS: + check_nms(ctx); + case NM_LHXS: + case NM_LHUXS: + tcg_gen_shli_tl(t0, t0, 1); + break; + case NM_SWXS: + check_nms(ctx); + case NM_LWXS: + case NM_LWC1XS: + case NM_SWC1XS: + tcg_gen_shli_tl(t0, t0, 2); + break; + case NM_LDC1XS: + case NM_SDC1XS: + tcg_gen_shli_tl(t0, t0, 3); + break; + } + } + gen_op_addr_add(ctx, t0, t0, t1); + + switch (extract32(ctx->opcode, 7, 4)) { + case NM_LBX: + tcg_gen_qemu_ld_tl(t0, t0, ctx->mem_idx, + MO_SB); + gen_store_gpr(t0, rd); + break; + case NM_LHX: + /*case NM_LHXS:*/ + tcg_gen_qemu_ld_tl(t0, t0, ctx->mem_idx, + MO_TESW); + gen_store_gpr(t0, rd); + break; + case NM_LWX: + /*case NM_LWXS:*/ + tcg_gen_qemu_ld_tl(t0, t0, ctx->mem_idx, + MO_TESL); + gen_store_gpr(t0, rd); + break; + case NM_LBUX: + tcg_gen_qemu_ld_tl(t0, t0, ctx->mem_idx, + MO_UB); + gen_store_gpr(t0, rd); + break; + case NM_LHUX: + /*case NM_LHUXS:*/ + tcg_gen_qemu_ld_tl(t0, t0, ctx->mem_idx, + MO_TEUW); + gen_store_gpr(t0, rd); + break; + case NM_SBX: + check_nms(ctx); + gen_load_gpr(t1, rd); + tcg_gen_qemu_st_tl(t1, t0, ctx->mem_idx, + MO_8); + break; + case NM_SHX: + /*case NM_SHXS:*/ + check_nms(ctx); + gen_load_gpr(t1, rd); + tcg_gen_qemu_st_tl(t1, t0, ctx->mem_idx, + MO_TEUW); + break; + case NM_SWX: + /*case NM_SWXS:*/ + check_nms(ctx); + gen_load_gpr(t1, rd); + tcg_gen_qemu_st_tl(t1, t0, ctx->mem_idx, + MO_TEUL); + break; + case NM_LWC1X: + /*case NM_LWC1XS:*/ + case NM_LDC1X: + /*case NM_LDC1XS:*/ + case NM_SWC1X: + /*case NM_SWC1XS:*/ + case NM_SDC1X: + /*case NM_SDC1XS:*/ + if (ctx->CP0_Config1 & (1 << CP0C1_FP)) { + check_cp1_enabled(ctx); + switch (extract32(ctx->opcode, 7, 4)) { + case NM_LWC1X: + /*case NM_LWC1XS:*/ + gen_flt_ldst(ctx, OPC_LWC1, rd, t0); + break; + case NM_LDC1X: + /*case NM_LDC1XS:*/ + gen_flt_ldst(ctx, OPC_LDC1, rd, t0); + break; + case NM_SWC1X: + /*case NM_SWC1XS:*/ + gen_flt_ldst(ctx, OPC_SWC1, rd, t0); + break; + case NM_SDC1X: + /*case NM_SDC1XS:*/ + gen_flt_ldst(ctx, OPC_SDC1, rd, t0); + break; + } + } else { + generate_exception_err(ctx, EXCP_CpU, 1); + } + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + + tcg_temp_free(t0); + tcg_temp_free(t1); +} + +static void gen_pool32f_nanomips_insn(DisasContext *ctx) +{ + int rt, rs, rd; + + rt = extract32(ctx->opcode, 21, 5); + rs = extract32(ctx->opcode, 16, 5); + rd = extract32(ctx->opcode, 11, 5); + + if (!(ctx->CP0_Config1 & (1 << CP0C1_FP))) { + generate_exception_end(ctx, EXCP_RI); + return; + } + check_cp1_enabled(ctx); + switch (extract32(ctx->opcode, 0, 3)) { + case NM_POOL32F_0: + switch (extract32(ctx->opcode, 3, 7)) { + case NM_RINT_S: + gen_farith(ctx, OPC_RINT_S, 0, rt, rs, 0); + break; + case NM_RINT_D: + gen_farith(ctx, OPC_RINT_D, 0, rt, rs, 0); + break; + case NM_CLASS_S: + gen_farith(ctx, OPC_CLASS_S, 0, rt, rs, 0); + break; + case NM_CLASS_D: + gen_farith(ctx, OPC_CLASS_D, 0, rt, rs, 0); + break; + case NM_ADD_S: + gen_farith(ctx, OPC_ADD_S, rt, rs, rd, 0); + break; + case NM_ADD_D: + gen_farith(ctx, OPC_ADD_D, rt, rs, rd, 0); + break; + case NM_SUB_S: + gen_farith(ctx, OPC_SUB_S, rt, rs, rd, 0); + break; + case NM_SUB_D: + gen_farith(ctx, OPC_SUB_D, rt, rs, rd, 0); + break; + case NM_MUL_S: + gen_farith(ctx, OPC_MUL_S, rt, rs, rd, 0); + break; + case NM_MUL_D: + gen_farith(ctx, OPC_MUL_D, rt, rs, rd, 0); + break; + case NM_DIV_S: + gen_farith(ctx, OPC_DIV_S, rt, rs, rd, 0); + break; + case NM_DIV_D: + gen_farith(ctx, OPC_DIV_D, rt, rs, rd, 0); + break; + case NM_SELEQZ_S: + gen_sel_s(ctx, OPC_SELEQZ_S, rd, rt, rs); + break; + case NM_SELEQZ_D: + gen_sel_d(ctx, OPC_SELEQZ_D, rd, rt, rs); + break; + case NM_SELNEZ_S: + gen_sel_s(ctx, OPC_SELNEZ_S, rd, rt, rs); + break; + case NM_SELNEZ_D: + gen_sel_d(ctx, OPC_SELNEZ_D, rd, rt, rs); + break; + case NM_SEL_S: + gen_sel_s(ctx, OPC_SEL_S, rd, rt, rs); + break; + case NM_SEL_D: + gen_sel_d(ctx, OPC_SEL_D, rd, rt, rs); + break; + case NM_MADDF_S: + gen_farith(ctx, OPC_MADDF_S, rt, rs, rd, 0); + break; + case NM_MADDF_D: + gen_farith(ctx, OPC_MADDF_D, rt, rs, rd, 0); + break; + case NM_MSUBF_S: + gen_farith(ctx, OPC_MSUBF_S, rt, rs, rd, 0); + break; + case NM_MSUBF_D: + gen_farith(ctx, OPC_MSUBF_D, rt, rs, rd, 0); + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + break; + case NM_POOL32F_3: + switch (extract32(ctx->opcode, 3, 3)) { + case NM_MIN_FMT: + switch (extract32(ctx->opcode, 9, 1)) { + case FMT_SDPS_S: + gen_farith(ctx, OPC_MIN_S, rt, rs, rd, 0); + break; + case FMT_SDPS_D: + gen_farith(ctx, OPC_MIN_D, rt, rs, rd, 0); + break; + } + break; + case NM_MAX_FMT: + switch (extract32(ctx->opcode, 9, 1)) { + case FMT_SDPS_S: + gen_farith(ctx, OPC_MAX_S, rt, rs, rd, 0); + break; + case FMT_SDPS_D: + gen_farith(ctx, OPC_MAX_D, rt, rs, rd, 0); + break; + } + break; + case NM_MINA_FMT: + switch (extract32(ctx->opcode, 9, 1)) { + case FMT_SDPS_S: + gen_farith(ctx, OPC_MINA_S, rt, rs, rd, 0); + break; + case FMT_SDPS_D: + gen_farith(ctx, OPC_MINA_D, rt, rs, rd, 0); + break; + } + break; + case NM_MAXA_FMT: + switch (extract32(ctx->opcode, 9, 1)) { + case FMT_SDPS_S: + gen_farith(ctx, OPC_MAXA_S, rt, rs, rd, 0); + break; + case FMT_SDPS_D: + gen_farith(ctx, OPC_MAXA_D, rt, rs, rd, 0); + break; + } + break; + case NM_POOL32FXF: + switch (extract32(ctx->opcode, 6, 8)) { + case NM_CFC1: + gen_cp1(ctx, OPC_CFC1, rt, rs); + break; + case NM_CTC1: + gen_cp1(ctx, OPC_CTC1, rt, rs); + break; + case NM_MFC1: + gen_cp1(ctx, OPC_MFC1, rt, rs); + break; + case NM_MTC1: + gen_cp1(ctx, OPC_MTC1, rt, rs); + break; + case NM_MFHC1: + gen_cp1(ctx, OPC_MFHC1, rt, rs); + break; + case NM_MTHC1: + gen_cp1(ctx, OPC_MTHC1, rt, rs); + break; + case NM_CVT_S_PL: + gen_farith(ctx, OPC_CVT_S_PL, -1, rs, rt, 0); + break; + case NM_CVT_S_PU: + gen_farith(ctx, OPC_CVT_S_PU, -1, rs, rt, 0); + break; + default: + switch (extract32(ctx->opcode, 6, 9)) { + case NM_CVT_L_S: + gen_farith(ctx, OPC_CVT_L_S, -1, rs, rt, 0); + break; + case NM_CVT_L_D: + gen_farith(ctx, OPC_CVT_L_D, -1, rs, rt, 0); + break; + case NM_CVT_W_S: + gen_farith(ctx, OPC_CVT_W_S, -1, rs, rt, 0); + break; + case NM_CVT_W_D: + gen_farith(ctx, OPC_CVT_W_D, -1, rs, rt, 0); + break; + case NM_RSQRT_S: + gen_farith(ctx, OPC_RSQRT_S, -1, rs, rt, 0); + break; + case NM_RSQRT_D: + gen_farith(ctx, OPC_RSQRT_D, -1, rs, rt, 0); + break; + case NM_SQRT_S: + gen_farith(ctx, OPC_SQRT_S, -1, rs, rt, 0); + break; + case NM_SQRT_D: + gen_farith(ctx, OPC_SQRT_D, -1, rs, rt, 0); + break; + case NM_RECIP_S: + gen_farith(ctx, OPC_RECIP_S, -1, rs, rt, 0); + break; + case NM_RECIP_D: + gen_farith(ctx, OPC_RECIP_D, -1, rs, rt, 0); + break; + case NM_FLOOR_L_S: + gen_farith(ctx, OPC_FLOOR_L_S, -1, rs, rt, 0); + break; + case NM_FLOOR_L_D: + gen_farith(ctx, OPC_FLOOR_L_D, -1, rs, rt, 0); + break; + case NM_FLOOR_W_S: + gen_farith(ctx, OPC_FLOOR_W_S, -1, rs, rt, 0); + break; + case NM_FLOOR_W_D: + gen_farith(ctx, OPC_FLOOR_W_D, -1, rs, rt, 0); + break; + case NM_CEIL_L_S: + gen_farith(ctx, OPC_CEIL_L_S, -1, rs, rt, 0); + break; + case NM_CEIL_L_D: + gen_farith(ctx, OPC_CEIL_L_D, -1, rs, rt, 0); + break; + case NM_CEIL_W_S: + gen_farith(ctx, OPC_CEIL_W_S, -1, rs, rt, 0); + break; + case NM_CEIL_W_D: + gen_farith(ctx, OPC_CEIL_W_D, -1, rs, rt, 0); + break; + case NM_TRUNC_L_S: + gen_farith(ctx, OPC_TRUNC_L_S, -1, rs, rt, 0); + break; + case NM_TRUNC_L_D: + gen_farith(ctx, OPC_TRUNC_L_D, -1, rs, rt, 0); + break; + case NM_TRUNC_W_S: + gen_farith(ctx, OPC_TRUNC_W_S, -1, rs, rt, 0); + break; + case NM_TRUNC_W_D: + gen_farith(ctx, OPC_TRUNC_W_D, -1, rs, rt, 0); + break; + case NM_ROUND_L_S: + gen_farith(ctx, OPC_ROUND_L_S, -1, rs, rt, 0); + break; + case NM_ROUND_L_D: + gen_farith(ctx, OPC_ROUND_L_D, -1, rs, rt, 0); + break; + case NM_ROUND_W_S: + gen_farith(ctx, OPC_ROUND_W_S, -1, rs, rt, 0); + break; + case NM_ROUND_W_D: + gen_farith(ctx, OPC_ROUND_W_D, -1, rs, rt, 0); + break; + case NM_MOV_S: + gen_farith(ctx, OPC_MOV_S, -1, rs, rt, 0); + break; + case NM_MOV_D: + gen_farith(ctx, OPC_MOV_D, -1, rs, rt, 0); + break; + case NM_ABS_S: + gen_farith(ctx, OPC_ABS_S, -1, rs, rt, 0); + break; + case NM_ABS_D: + gen_farith(ctx, OPC_ABS_D, -1, rs, rt, 0); + break; + case NM_NEG_S: + gen_farith(ctx, OPC_NEG_S, -1, rs, rt, 0); + break; + case NM_NEG_D: + gen_farith(ctx, OPC_NEG_D, -1, rs, rt, 0); + break; + case NM_CVT_D_S: + gen_farith(ctx, OPC_CVT_D_S, -1, rs, rt, 0); + break; + case NM_CVT_D_W: + gen_farith(ctx, OPC_CVT_D_W, -1, rs, rt, 0); + break; + case NM_CVT_D_L: + gen_farith(ctx, OPC_CVT_D_L, -1, rs, rt, 0); + break; + case NM_CVT_S_D: + gen_farith(ctx, OPC_CVT_S_D, -1, rs, rt, 0); + break; + case NM_CVT_S_W: + gen_farith(ctx, OPC_CVT_S_W, -1, rs, rt, 0); + break; + case NM_CVT_S_L: + gen_farith(ctx, OPC_CVT_S_L, -1, rs, rt, 0); + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + break; + } + break; + } + break; + case NM_POOL32F_5: + switch (extract32(ctx->opcode, 3, 3)) { + case NM_CMP_CONDN_S: + gen_r6_cmp_s(ctx, extract32(ctx->opcode, 6, 5), rt, rs, rd); + break; + case NM_CMP_CONDN_D: + gen_r6_cmp_d(ctx, extract32(ctx->opcode, 6, 5), rt, rs, rd); + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } +} + +static void gen_pool32a5_nanomips_insn(DisasContext *ctx, int opc, + int rd, int rs, int rt) +{ + int ret = rd; + TCGv t0 = tcg_temp_new(); + TCGv v1_t = tcg_temp_new(); + TCGv v2_t = tcg_temp_new(); + + gen_load_gpr(v1_t, rs); + gen_load_gpr(v2_t, rt); + + switch (opc) { + case NM_CMP_EQ_PH: + check_dsp(ctx); + gen_helper_cmp_eq_ph(v1_t, v2_t, cpu_env); + break; + case NM_CMP_LT_PH: + check_dsp(ctx); + gen_helper_cmp_lt_ph(v1_t, v2_t, cpu_env); + break; + case NM_CMP_LE_PH: + check_dsp(ctx); + gen_helper_cmp_le_ph(v1_t, v2_t, cpu_env); + break; + case NM_CMPU_EQ_QB: + check_dsp(ctx); + gen_helper_cmpu_eq_qb(v1_t, v2_t, cpu_env); + break; + case NM_CMPU_LT_QB: + check_dsp(ctx); + gen_helper_cmpu_lt_qb(v1_t, v2_t, cpu_env); + break; + case NM_CMPU_LE_QB: + check_dsp(ctx); + gen_helper_cmpu_le_qb(v1_t, v2_t, cpu_env); + break; + case NM_CMPGU_EQ_QB: + check_dsp(ctx); + gen_helper_cmpgu_eq_qb(v1_t, v1_t, v2_t); + gen_store_gpr(v1_t, ret); + break; + case NM_CMPGU_LT_QB: + check_dsp(ctx); + gen_helper_cmpgu_lt_qb(v1_t, v1_t, v2_t); + gen_store_gpr(v1_t, ret); + break; + case NM_CMPGU_LE_QB: + check_dsp(ctx); + gen_helper_cmpgu_le_qb(v1_t, v1_t, v2_t); + gen_store_gpr(v1_t, ret); + break; + case NM_CMPGDU_EQ_QB: + check_dspr2(ctx); + gen_helper_cmpgu_eq_qb(v1_t, v1_t, v2_t); + tcg_gen_deposit_tl(cpu_dspctrl, cpu_dspctrl, v1_t, 24, 4); + gen_store_gpr(v1_t, ret); + break; + case NM_CMPGDU_LT_QB: + check_dspr2(ctx); + gen_helper_cmpgu_lt_qb(v1_t, v1_t, v2_t); + tcg_gen_deposit_tl(cpu_dspctrl, cpu_dspctrl, v1_t, 24, 4); + gen_store_gpr(v1_t, ret); + break; + case NM_CMPGDU_LE_QB: + check_dspr2(ctx); + gen_helper_cmpgu_le_qb(v1_t, v1_t, v2_t); + tcg_gen_deposit_tl(cpu_dspctrl, cpu_dspctrl, v1_t, 24, 4); + gen_store_gpr(v1_t, ret); + break; + case NM_PACKRL_PH: + check_dsp(ctx); + gen_helper_packrl_ph(v1_t, v1_t, v2_t); + gen_store_gpr(v1_t, ret); + break; + case NM_PICK_QB: + check_dsp(ctx); + gen_helper_pick_qb(v1_t, v1_t, v2_t, cpu_env); + gen_store_gpr(v1_t, ret); + break; + case NM_PICK_PH: + check_dsp(ctx); + gen_helper_pick_ph(v1_t, v1_t, v2_t, cpu_env); + gen_store_gpr(v1_t, ret); + break; + case NM_ADDQ_S_W: + check_dsp(ctx); + gen_helper_addq_s_w(v1_t, v1_t, v2_t, cpu_env); + gen_store_gpr(v1_t, ret); + break; + case NM_SUBQ_S_W: + check_dsp(ctx); + gen_helper_subq_s_w(v1_t, v1_t, v2_t, cpu_env); + gen_store_gpr(v1_t, ret); + break; + case NM_ADDSC: + check_dsp(ctx); + gen_helper_addsc(v1_t, v1_t, v2_t, cpu_env); + gen_store_gpr(v1_t, ret); + break; + case NM_ADDWC: + check_dsp(ctx); + gen_helper_addwc(v1_t, v1_t, v2_t, cpu_env); + gen_store_gpr(v1_t, ret); + break; + case NM_ADDQ_S_PH: + check_dsp(ctx); + switch (extract32(ctx->opcode, 10, 1)) { + case 0: + /* ADDQ_PH */ + gen_helper_addq_ph(v1_t, v1_t, v2_t, cpu_env); + gen_store_gpr(v1_t, ret); + break; + case 1: + /* ADDQ_S_PH */ + gen_helper_addq_s_ph(v1_t, v1_t, v2_t, cpu_env); + gen_store_gpr(v1_t, ret); + break; + } + break; + case NM_ADDQH_R_PH: + check_dspr2(ctx); + switch (extract32(ctx->opcode, 10, 1)) { + case 0: + /* ADDQH_PH */ + gen_helper_addqh_ph(v1_t, v1_t, v2_t); + gen_store_gpr(v1_t, ret); + break; + case 1: + /* ADDQH_R_PH */ + gen_helper_addqh_r_ph(v1_t, v1_t, v2_t); + gen_store_gpr(v1_t, ret); + break; + } + break; + case NM_ADDQH_R_W: + check_dspr2(ctx); + switch (extract32(ctx->opcode, 10, 1)) { + case 0: + /* ADDQH_W */ + gen_helper_addqh_w(v1_t, v1_t, v2_t); + gen_store_gpr(v1_t, ret); + break; + case 1: + /* ADDQH_R_W */ + gen_helper_addqh_r_w(v1_t, v1_t, v2_t); + gen_store_gpr(v1_t, ret); + break; + } + break; + case NM_ADDU_S_QB: + check_dsp(ctx); + switch (extract32(ctx->opcode, 10, 1)) { + case 0: + /* ADDU_QB */ + gen_helper_addu_qb(v1_t, v1_t, v2_t, cpu_env); + gen_store_gpr(v1_t, ret); + break; + case 1: + /* ADDU_S_QB */ + gen_helper_addu_s_qb(v1_t, v1_t, v2_t, cpu_env); + gen_store_gpr(v1_t, ret); + break; + } + break; + case NM_ADDU_S_PH: + check_dspr2(ctx); + switch (extract32(ctx->opcode, 10, 1)) { + case 0: + /* ADDU_PH */ + gen_helper_addu_ph(v1_t, v1_t, v2_t, cpu_env); + gen_store_gpr(v1_t, ret); + break; + case 1: + /* ADDU_S_PH */ + gen_helper_addu_s_ph(v1_t, v1_t, v2_t, cpu_env); + gen_store_gpr(v1_t, ret); + break; + } + break; + case NM_ADDUH_R_QB: + check_dspr2(ctx); + switch (extract32(ctx->opcode, 10, 1)) { + case 0: + /* ADDUH_QB */ + gen_helper_adduh_qb(v1_t, v1_t, v2_t); + gen_store_gpr(v1_t, ret); + break; + case 1: + /* ADDUH_R_QB */ + gen_helper_adduh_r_qb(v1_t, v1_t, v2_t); + gen_store_gpr(v1_t, ret); + break; + } + break; + case NM_SHRAV_R_PH: + check_dsp(ctx); + switch (extract32(ctx->opcode, 10, 1)) { + case 0: + /* SHRAV_PH */ + gen_helper_shra_ph(v1_t, v1_t, v2_t); + gen_store_gpr(v1_t, ret); + break; + case 1: + /* SHRAV_R_PH */ + gen_helper_shra_r_ph(v1_t, v1_t, v2_t); + gen_store_gpr(v1_t, ret); + break; + } + break; + case NM_SHRAV_R_QB: + check_dspr2(ctx); + switch (extract32(ctx->opcode, 10, 1)) { + case 0: + /* SHRAV_QB */ + gen_helper_shra_qb(v1_t, v1_t, v2_t); + gen_store_gpr(v1_t, ret); + break; + case 1: + /* SHRAV_R_QB */ + gen_helper_shra_r_qb(v1_t, v1_t, v2_t); + gen_store_gpr(v1_t, ret); + break; + } + break; + case NM_SUBQ_S_PH: + check_dsp(ctx); + switch (extract32(ctx->opcode, 10, 1)) { + case 0: + /* SUBQ_PH */ + gen_helper_subq_ph(v1_t, v1_t, v2_t, cpu_env); + gen_store_gpr(v1_t, ret); + break; + case 1: + /* SUBQ_S_PH */ + gen_helper_subq_s_ph(v1_t, v1_t, v2_t, cpu_env); + gen_store_gpr(v1_t, ret); + break; + } + break; + case NM_SUBQH_R_PH: + check_dspr2(ctx); + switch (extract32(ctx->opcode, 10, 1)) { + case 0: + /* SUBQH_PH */ + gen_helper_subqh_ph(v1_t, v1_t, v2_t); + gen_store_gpr(v1_t, ret); + break; + case 1: + /* SUBQH_R_PH */ + gen_helper_subqh_r_ph(v1_t, v1_t, v2_t); + gen_store_gpr(v1_t, ret); + break; + } + break; + case NM_SUBQH_R_W: + check_dspr2(ctx); + switch (extract32(ctx->opcode, 10, 1)) { + case 0: + /* SUBQH_W */ + gen_helper_subqh_w(v1_t, v1_t, v2_t); + gen_store_gpr(v1_t, ret); + break; + case 1: + /* SUBQH_R_W */ + gen_helper_subqh_r_w(v1_t, v1_t, v2_t); + gen_store_gpr(v1_t, ret); + break; + } + break; + case NM_SUBU_S_QB: + check_dsp(ctx); + switch (extract32(ctx->opcode, 10, 1)) { + case 0: + /* SUBU_QB */ + gen_helper_subu_qb(v1_t, v1_t, v2_t, cpu_env); + gen_store_gpr(v1_t, ret); + break; + case 1: + /* SUBU_S_QB */ + gen_helper_subu_s_qb(v1_t, v1_t, v2_t, cpu_env); + gen_store_gpr(v1_t, ret); + break; + } + break; + case NM_SUBU_S_PH: + check_dspr2(ctx); + switch (extract32(ctx->opcode, 10, 1)) { + case 0: + /* SUBU_PH */ + gen_helper_subu_ph(v1_t, v1_t, v2_t, cpu_env); + gen_store_gpr(v1_t, ret); + break; + case 1: + /* SUBU_S_PH */ + gen_helper_subu_s_ph(v1_t, v1_t, v2_t, cpu_env); + gen_store_gpr(v1_t, ret); + break; + } + break; + case NM_SUBUH_R_QB: + check_dspr2(ctx); + switch (extract32(ctx->opcode, 10, 1)) { + case 0: + /* SUBUH_QB */ + gen_helper_subuh_qb(v1_t, v1_t, v2_t); + gen_store_gpr(v1_t, ret); + break; + case 1: + /* SUBUH_R_QB */ + gen_helper_subuh_r_qb(v1_t, v1_t, v2_t); + gen_store_gpr(v1_t, ret); + break; + } + break; + case NM_SHLLV_S_PH: + check_dsp(ctx); + switch (extract32(ctx->opcode, 10, 1)) { + case 0: + /* SHLLV_PH */ + gen_helper_shll_ph(v1_t, v1_t, v2_t, cpu_env); + gen_store_gpr(v1_t, ret); + break; + case 1: + /* SHLLV_S_PH */ + gen_helper_shll_s_ph(v1_t, v1_t, v2_t, cpu_env); + gen_store_gpr(v1_t, ret); + break; + } + break; + case NM_PRECR_SRA_R_PH_W: + check_dspr2(ctx); + switch (extract32(ctx->opcode, 10, 1)) { + case 0: + /* PRECR_SRA_PH_W */ + { + TCGv_i32 sa_t = tcg_const_i32(rd); + gen_helper_precr_sra_ph_w(v1_t, sa_t, v1_t, + cpu_gpr[rt]); + gen_store_gpr(v1_t, rt); + tcg_temp_free_i32(sa_t); + } + break; + case 1: + /* PRECR_SRA_R_PH_W */ + { + TCGv_i32 sa_t = tcg_const_i32(rd); + gen_helper_precr_sra_r_ph_w(v1_t, sa_t, v1_t, + cpu_gpr[rt]); + gen_store_gpr(v1_t, rt); + tcg_temp_free_i32(sa_t); + } + break; + } + break; + case NM_MULEU_S_PH_QBL: + check_dsp(ctx); + gen_helper_muleu_s_ph_qbl(v1_t, v1_t, v2_t, cpu_env); + gen_store_gpr(v1_t, ret); + break; + case NM_MULEU_S_PH_QBR: + check_dsp(ctx); + gen_helper_muleu_s_ph_qbr(v1_t, v1_t, v2_t, cpu_env); + gen_store_gpr(v1_t, ret); + break; + case NM_MULQ_RS_PH: + check_dsp(ctx); + gen_helper_mulq_rs_ph(v1_t, v1_t, v2_t, cpu_env); + gen_store_gpr(v1_t, ret); + break; + case NM_MULQ_S_PH: + check_dspr2(ctx); + gen_helper_mulq_s_ph(v1_t, v1_t, v2_t, cpu_env); + gen_store_gpr(v1_t, ret); + break; + case NM_MULQ_RS_W: + check_dspr2(ctx); + gen_helper_mulq_rs_w(v1_t, v1_t, v2_t, cpu_env); + gen_store_gpr(v1_t, ret); + break; + case NM_MULQ_S_W: + check_dspr2(ctx); + gen_helper_mulq_s_w(v1_t, v1_t, v2_t, cpu_env); + gen_store_gpr(v1_t, ret); + break; + case NM_APPEND: + check_dspr2(ctx); + gen_load_gpr(t0, rs); + if (rd != 0) { + tcg_gen_deposit_tl(cpu_gpr[rt], t0, cpu_gpr[rt], rd, 32 - rd); + } + tcg_gen_ext32s_tl(cpu_gpr[rt], cpu_gpr[rt]); + break; + case NM_MODSUB: + check_dsp(ctx); + gen_helper_modsub(v1_t, v1_t, v2_t); + gen_store_gpr(v1_t, ret); + break; + case NM_SHRAV_R_W: + check_dsp(ctx); + gen_helper_shra_r_w(v1_t, v1_t, v2_t); + gen_store_gpr(v1_t, ret); + break; + case NM_SHRLV_PH: + check_dspr2(ctx); + gen_helper_shrl_ph(v1_t, v1_t, v2_t); + gen_store_gpr(v1_t, ret); + break; + case NM_SHRLV_QB: + check_dsp(ctx); + gen_helper_shrl_qb(v1_t, v1_t, v2_t); + gen_store_gpr(v1_t, ret); + break; + case NM_SHLLV_QB: + check_dsp(ctx); + gen_helper_shll_qb(v1_t, v1_t, v2_t, cpu_env); + gen_store_gpr(v1_t, ret); + break; + case NM_SHLLV_S_W: + check_dsp(ctx); + gen_helper_shll_s_w(v1_t, v1_t, v2_t, cpu_env); + gen_store_gpr(v1_t, ret); + break; + case NM_SHILO: + check_dsp(ctx); + { + TCGv tv0 = tcg_temp_new(); + TCGv tv1 = tcg_temp_new(); + int16_t imm = extract32(ctx->opcode, 16, 7); + + tcg_gen_movi_tl(tv0, rd >> 3); + tcg_gen_movi_tl(tv1, imm); + gen_helper_shilo(tv0, tv1, cpu_env); + } + break; + case NM_MULEQ_S_W_PHL: + check_dsp(ctx); + gen_helper_muleq_s_w_phl(v1_t, v1_t, v2_t, cpu_env); + gen_store_gpr(v1_t, ret); + break; + case NM_MULEQ_S_W_PHR: + check_dsp(ctx); + gen_helper_muleq_s_w_phr(v1_t, v1_t, v2_t, cpu_env); + gen_store_gpr(v1_t, ret); + break; + case NM_MUL_S_PH: + check_dspr2(ctx); + switch (extract32(ctx->opcode, 10, 1)) { + case 0: + /* MUL_PH */ + gen_helper_mul_ph(v1_t, v1_t, v2_t, cpu_env); + gen_store_gpr(v1_t, ret); + break; + case 1: + /* MUL_S_PH */ + gen_helper_mul_s_ph(v1_t, v1_t, v2_t, cpu_env); + gen_store_gpr(v1_t, ret); + break; + } + break; + case NM_PRECR_QB_PH: + check_dspr2(ctx); + gen_helper_precr_qb_ph(v1_t, v1_t, v2_t); + gen_store_gpr(v1_t, ret); + break; + case NM_PRECRQ_QB_PH: + check_dsp(ctx); + gen_helper_precrq_qb_ph(v1_t, v1_t, v2_t); + gen_store_gpr(v1_t, ret); + break; + case NM_PRECRQ_PH_W: + check_dsp(ctx); + gen_helper_precrq_ph_w(v1_t, v1_t, v2_t); + gen_store_gpr(v1_t, ret); + break; + case NM_PRECRQ_RS_PH_W: + check_dsp(ctx); + gen_helper_precrq_rs_ph_w(v1_t, v1_t, v2_t, cpu_env); + gen_store_gpr(v1_t, ret); + break; + case NM_PRECRQU_S_QB_PH: + check_dsp(ctx); + gen_helper_precrqu_s_qb_ph(v1_t, v1_t, v2_t, cpu_env); + gen_store_gpr(v1_t, ret); + break; + case NM_SHRA_R_W: + check_dsp(ctx); + tcg_gen_movi_tl(t0, rd); + gen_helper_shra_r_w(v1_t, t0, v1_t); + gen_store_gpr(v1_t, rt); + break; + case NM_SHRA_R_PH: + check_dsp(ctx); + tcg_gen_movi_tl(t0, rd >> 1); + switch (extract32(ctx->opcode, 10, 1)) { + case 0: + /* SHRA_PH */ + gen_helper_shra_ph(v1_t, t0, v1_t); + break; + gen_store_gpr(v1_t, rt); + case 1: + /* SHRA_R_PH */ + gen_helper_shra_r_ph(v1_t, t0, v1_t); + gen_store_gpr(v1_t, rt); + break; + } + break; + case NM_SHLL_S_PH: + check_dsp(ctx); + tcg_gen_movi_tl(t0, rd >> 1); + switch (extract32(ctx->opcode, 10, 2)) { + case 0: + /* SHLL_PH */ + gen_helper_shll_ph(v1_t, t0, v1_t, cpu_env); + gen_store_gpr(v1_t, rt); + break; + case 2: + /* SHLL_S_PH */ + gen_helper_shll_s_ph(v1_t, t0, v1_t, cpu_env); + gen_store_gpr(v1_t, rt); + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + break; + case NM_SHLL_S_W: + check_dsp(ctx); + tcg_gen_movi_tl(t0, rd); + gen_helper_shll_s_w(v1_t, t0, v1_t, cpu_env); + gen_store_gpr(v1_t, rt); + break; + case NM_REPL_PH: + check_dsp(ctx); + { + int16_t imm; + imm = sextract32(ctx->opcode, 11, 11); + imm = (int16_t)(imm << 6) >> 6; + if (rt != 0) { + tcg_gen_movi_tl(cpu_gpr[rt], dup_const(MO_16, imm)); + } + } + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } +} + +static int decode_nanomips_32_48_opc(CPUMIPSState *env, DisasContext *ctx) +{ + uint16_t insn; + uint32_t op; + int rt, rs, rd; + int offset; + int imm; + + insn = cpu_lduw_code(env, ctx->base.pc_next + 2); + ctx->opcode = (ctx->opcode << 16) | insn; + + rt = extract32(ctx->opcode, 21, 5); + rs = extract32(ctx->opcode, 16, 5); + rd = extract32(ctx->opcode, 11, 5); + + op = extract32(ctx->opcode, 26, 6); + switch (op) { + case NM_P_ADDIU: + if (rt == 0) { + /* P.RI */ + switch (extract32(ctx->opcode, 19, 2)) { + case NM_SIGRIE: + default: + generate_exception_end(ctx, EXCP_RI); + break; + case NM_P_SYSCALL: + if ((extract32(ctx->opcode, 18, 1)) == NM_SYSCALL) { + generate_exception_end(ctx, EXCP_SYSCALL); + } else { + generate_exception_end(ctx, EXCP_RI); + } + break; + case NM_BREAK: + generate_exception_end(ctx, EXCP_BREAK); + break; + case NM_SDBBP: + if (is_uhi(extract32(ctx->opcode, 0, 19))) { + gen_helper_do_semihosting(cpu_env); + } else { + if (ctx->hflags & MIPS_HFLAG_SBRI) { + generate_exception_end(ctx, EXCP_RI); + } else { + generate_exception_end(ctx, EXCP_DBp); + } + } + break; + } + } else { + /* NM_ADDIU */ + imm = extract32(ctx->opcode, 0, 16); + if (rs != 0) { + tcg_gen_addi_tl(cpu_gpr[rt], cpu_gpr[rs], imm); + } else { + tcg_gen_movi_tl(cpu_gpr[rt], imm); + } + tcg_gen_ext32s_tl(cpu_gpr[rt], cpu_gpr[rt]); + } + break; + case NM_ADDIUPC: + if (rt != 0) { + offset = sextract32(ctx->opcode, 0, 1) << 21 | + extract32(ctx->opcode, 1, 20) << 1; + target_long addr = addr_add(ctx, ctx->base.pc_next + 4, offset); + tcg_gen_movi_tl(cpu_gpr[rt], addr); + } + break; + case NM_POOL32A: + switch (ctx->opcode & 0x07) { + case NM_POOL32A0: + gen_pool32a0_nanomips_insn(env, ctx); + break; + case NM_POOL32A5: + { + int32_t op1 = extract32(ctx->opcode, 3, 7); + gen_pool32a5_nanomips_insn(ctx, op1, rd, rs, rt); + } + break; + case NM_POOL32A7: + switch (extract32(ctx->opcode, 3, 3)) { + case NM_P_LSX: + gen_p_lsx(ctx, rd, rs, rt); + break; + case NM_LSA: + /* In nanoMIPS, the shift field directly encodes the shift + * amount, meaning that the supported shift values are in + * the range 0 to 3 (instead of 1 to 4 in MIPSR6). */ + gen_lsa(ctx, OPC_LSA, rd, rs, rt, + extract32(ctx->opcode, 9, 2) - 1); + break; + case NM_EXTW: + gen_ext(ctx, 32, rd, rs, rt, extract32(ctx->opcode, 6, 5)); + break; + case NM_POOL32AXF: + gen_pool32axf_nanomips_insn(env, ctx); + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + break; + case NM_P_GP_W: + switch (ctx->opcode & 0x03) { + case NM_ADDIUGP_W: + if (rt != 0) { + offset = extract32(ctx->opcode, 0, 21); + gen_op_addr_addi(ctx, cpu_gpr[rt], cpu_gpr[28], offset); + } + break; + case NM_LWGP: + gen_ld(ctx, OPC_LW, rt, 28, extract32(ctx->opcode, 2, 19) << 2); + break; + case NM_SWGP: + gen_st(ctx, OPC_SW, rt, 28, extract32(ctx->opcode, 2, 19) << 2); + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + break; + case NM_P48I: + { + insn = cpu_lduw_code(env, ctx->base.pc_next + 4); + target_long addr_off = extract32(ctx->opcode, 0, 16) | insn << 16; + switch (extract32(ctx->opcode, 16, 5)) { + case NM_LI48: + check_nms(ctx); + if (rt != 0) { + tcg_gen_movi_tl(cpu_gpr[rt], addr_off); + } + break; + case NM_ADDIU48: + check_nms(ctx); + if (rt != 0) { + tcg_gen_addi_tl(cpu_gpr[rt], cpu_gpr[rt], addr_off); + tcg_gen_ext32s_tl(cpu_gpr[rt], cpu_gpr[rt]); + } + break; + case NM_ADDIUGP48: + check_nms(ctx); + if (rt != 0) { + gen_op_addr_addi(ctx, cpu_gpr[rt], cpu_gpr[28], addr_off); + } + break; + case NM_ADDIUPC48: + check_nms(ctx); + if (rt != 0) { + target_long addr = addr_add(ctx, ctx->base.pc_next + 6, + addr_off); + + tcg_gen_movi_tl(cpu_gpr[rt], addr); + } + break; + case NM_LWPC48: + check_nms(ctx); + if (rt != 0) { + TCGv t0; + t0 = tcg_temp_new(); + + target_long addr = addr_add(ctx, ctx->base.pc_next + 6, + addr_off); + + tcg_gen_movi_tl(t0, addr); + tcg_gen_qemu_ld_tl(cpu_gpr[rt], t0, ctx->mem_idx, MO_TESL); + tcg_temp_free(t0); + } + break; + case NM_SWPC48: + check_nms(ctx); + { + TCGv t0, t1; + t0 = tcg_temp_new(); + t1 = tcg_temp_new(); + + target_long addr = addr_add(ctx, ctx->base.pc_next + 6, + addr_off); + + tcg_gen_movi_tl(t0, addr); + gen_load_gpr(t1, rt); + + tcg_gen_qemu_st_tl(t1, t0, ctx->mem_idx, MO_TEUL); + + tcg_temp_free(t0); + tcg_temp_free(t1); + } + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + return 6; + } + case NM_P_U12: + switch (extract32(ctx->opcode, 12, 4)) { + case NM_ORI: + gen_logic_imm(ctx, OPC_ORI, rt, rs, extract32(ctx->opcode, 0, 12)); + break; + case NM_XORI: + gen_logic_imm(ctx, OPC_XORI, rt, rs, extract32(ctx->opcode, 0, 12)); + break; + case NM_ANDI: + gen_logic_imm(ctx, OPC_ANDI, rt, rs, extract32(ctx->opcode, 0, 12)); + break; + case NM_P_SR: + switch (extract32(ctx->opcode, 20, 1)) { + case NM_PP_SR: + switch (ctx->opcode & 3) { + case NM_SAVE: + gen_save(ctx, rt, extract32(ctx->opcode, 16, 4), + extract32(ctx->opcode, 2, 1), + extract32(ctx->opcode, 3, 9) << 3); + break; + case NM_RESTORE: + case NM_RESTORE_JRC: + gen_restore(ctx, rt, extract32(ctx->opcode, 16, 4), + extract32(ctx->opcode, 2, 1), + extract32(ctx->opcode, 3, 9) << 3); + if ((ctx->opcode & 3) == NM_RESTORE_JRC) { + gen_compute_branch_nm(ctx, OPC_JR, 2, 31, 0, 0); + } + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + break; + case NM_P_SR_F: + generate_exception_end(ctx, EXCP_RI); + break; + } + break; + case NM_SLTI: + gen_slt_imm(ctx, OPC_SLTI, rt, rs, extract32(ctx->opcode, 0, 12)); + break; + case NM_SLTIU: + gen_slt_imm(ctx, OPC_SLTIU, rt, rs, extract32(ctx->opcode, 0, 12)); + break; + case NM_SEQI: + { + TCGv t0 = tcg_temp_new(); + + imm = extract32(ctx->opcode, 0, 12); + gen_load_gpr(t0, rs); + tcg_gen_setcondi_tl(TCG_COND_EQ, t0, t0, imm); + gen_store_gpr(t0, rt); + + tcg_temp_free(t0); + } + break; + case NM_ADDIUNEG: + imm = (int16_t) extract32(ctx->opcode, 0, 12); + gen_arith_imm(ctx, OPC_ADDIU, rt, rs, -imm); + break; + case NM_P_SHIFT: + { + int shift = extract32(ctx->opcode, 0, 5); + switch (extract32(ctx->opcode, 5, 4)) { + case NM_P_SLL: + if (rt == 0 && shift == 0) { + /* NOP */ + } else if (rt == 0 && shift == 3) { + /* EHB - treat as NOP */ + } else if (rt == 0 && shift == 5) { + /* PAUSE - treat as NOP */ + } else if (rt == 0 && shift == 6) { + /* SYNC */ + gen_sync(extract32(ctx->opcode, 16, 5)); + } else { + /* SLL */ + gen_shift_imm(ctx, OPC_SLL, rt, rs, + extract32(ctx->opcode, 0, 5)); + } + break; + case NM_SRL: + gen_shift_imm(ctx, OPC_SRL, rt, rs, + extract32(ctx->opcode, 0, 5)); + break; + case NM_SRA: + gen_shift_imm(ctx, OPC_SRA, rt, rs, + extract32(ctx->opcode, 0, 5)); + break; + case NM_ROTR: + gen_shift_imm(ctx, OPC_ROTR, rt, rs, + extract32(ctx->opcode, 0, 5)); + break; + } + } + break; + case NM_P_ROTX: + check_nms(ctx); + if (rt != 0) { + TCGv t0 = tcg_temp_new(); + TCGv_i32 shift = tcg_const_i32(extract32(ctx->opcode, 0, 5)); + TCGv_i32 shiftx = tcg_const_i32(extract32(ctx->opcode, 7, 4) + << 1); + TCGv_i32 stripe = tcg_const_i32(extract32(ctx->opcode, 6, 1)); + + gen_load_gpr(t0, rs); + gen_helper_rotx(cpu_gpr[rt], t0, shift, shiftx, stripe); + tcg_temp_free(t0); + + tcg_temp_free_i32(shift); + tcg_temp_free_i32(shiftx); + tcg_temp_free_i32(stripe); + } + break; + case NM_P_INS: + switch (((ctx->opcode >> 10) & 2) | + (extract32(ctx->opcode, 5, 1))) { + case NM_INS: + check_nms(ctx); + gen_bitops(ctx, OPC_INS, rt, rs, extract32(ctx->opcode, 0, 5), + extract32(ctx->opcode, 6, 5)); + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + break; + case NM_P_EXT: + switch (((ctx->opcode >> 10) & 2) | + (extract32(ctx->opcode, 5, 1))) { + case NM_EXT: + check_nms(ctx); + gen_bitops(ctx, OPC_EXT, rt, rs, extract32(ctx->opcode, 0, 5), + extract32(ctx->opcode, 6, 5)); + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + break; + case NM_POOL32F: + gen_pool32f_nanomips_insn(ctx); + break; + case NM_POOL32S: + break; + case NM_P_LUI: + switch (extract32(ctx->opcode, 1, 1)) { + case NM_LUI: + if (rt != 0) { + tcg_gen_movi_tl(cpu_gpr[rt], + sextract32(ctx->opcode, 0, 1) << 31 | + extract32(ctx->opcode, 2, 10) << 21 | + extract32(ctx->opcode, 12, 9) << 12); + } + break; + case NM_ALUIPC: + if (rt != 0) { + offset = sextract32(ctx->opcode, 0, 1) << 31 | + extract32(ctx->opcode, 2, 10) << 21 | + extract32(ctx->opcode, 12, 9) << 12; + target_long addr; + addr = ~0xFFF & addr_add(ctx, ctx->base.pc_next + 4, offset); + tcg_gen_movi_tl(cpu_gpr[rt], addr); + } + break; + } + break; + case NM_P_GP_BH: + { + uint32_t u = extract32(ctx->opcode, 0, 18); + + switch (extract32(ctx->opcode, 18, 3)) { + case NM_LBGP: + gen_ld(ctx, OPC_LB, rt, 28, u); + break; + case NM_SBGP: + gen_st(ctx, OPC_SB, rt, 28, u); + break; + case NM_LBUGP: + gen_ld(ctx, OPC_LBU, rt, 28, u); + break; + case NM_ADDIUGP_B: + if (rt != 0) { + gen_op_addr_addi(ctx, cpu_gpr[rt], cpu_gpr[28], u); + } + break; + case NM_P_GP_LH: + u &= ~1; + switch (ctx->opcode & 1) { + case NM_LHGP: + gen_ld(ctx, OPC_LH, rt, 28, u); + break; + case NM_LHUGP: + gen_ld(ctx, OPC_LHU, rt, 28, u); + break; + } + break; + case NM_P_GP_SH: + u &= ~1; + switch (ctx->opcode & 1) { + case NM_SHGP: + gen_st(ctx, OPC_SH, rt, 28, u); + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + break; + case NM_P_GP_CP1: + u &= ~0x3; + switch (ctx->opcode & 0x3) { + case NM_LWC1GP: + gen_cop1_ldst(ctx, OPC_LWC1, rt, 28, u); + break; + case NM_LDC1GP: + gen_cop1_ldst(ctx, OPC_LDC1, rt, 28, u); + break; + case NM_SWC1GP: + gen_cop1_ldst(ctx, OPC_SWC1, rt, 28, u); + break; + case NM_SDC1GP: + gen_cop1_ldst(ctx, OPC_SDC1, rt, 28, u); + break; + } + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + } + break; + case NM_P_LS_U12: + { + uint32_t u = extract32(ctx->opcode, 0, 12); + + switch (extract32(ctx->opcode, 12, 4)) { + case NM_P_PREFU12: + if (rt == 31) { + /* SYNCI */ + /* Break the TB to be able to sync copied instructions + immediately */ + ctx->base.is_jmp = DISAS_STOP; + } else { + /* PREF */ + /* Treat as NOP. */ + } + break; + case NM_LB: + gen_ld(ctx, OPC_LB, rt, rs, u); + break; + case NM_LH: + gen_ld(ctx, OPC_LH, rt, rs, u); + break; + case NM_LW: + gen_ld(ctx, OPC_LW, rt, rs, u); + break; + case NM_LBU: + gen_ld(ctx, OPC_LBU, rt, rs, u); + break; + case NM_LHU: + gen_ld(ctx, OPC_LHU, rt, rs, u); + break; + case NM_SB: + gen_st(ctx, OPC_SB, rt, rs, u); + break; + case NM_SH: + gen_st(ctx, OPC_SH, rt, rs, u); + break; + case NM_SW: + gen_st(ctx, OPC_SW, rt, rs, u); + break; + case NM_LWC1: + gen_cop1_ldst(ctx, OPC_LWC1, rt, rs, u); + break; + case NM_LDC1: + gen_cop1_ldst(ctx, OPC_LDC1, rt, rs, u); + break; + case NM_SWC1: + gen_cop1_ldst(ctx, OPC_SWC1, rt, rs, u); + break; + case NM_SDC1: + gen_cop1_ldst(ctx, OPC_SDC1, rt, rs, u); + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + } + break; + case NM_P_LS_S9: + { + int32_t s = (sextract32(ctx->opcode, 15, 1) << 8) | + extract32(ctx->opcode, 0, 8); + + switch (extract32(ctx->opcode, 8, 3)) { + case NM_P_LS_S0: + switch (extract32(ctx->opcode, 11, 4)) { + case NM_LBS9: + gen_ld(ctx, OPC_LB, rt, rs, s); + break; + case NM_LHS9: + gen_ld(ctx, OPC_LH, rt, rs, s); + break; + case NM_LWS9: + gen_ld(ctx, OPC_LW, rt, rs, s); + break; + case NM_LBUS9: + gen_ld(ctx, OPC_LBU, rt, rs, s); + break; + case NM_LHUS9: + gen_ld(ctx, OPC_LHU, rt, rs, s); + break; + case NM_SBS9: + gen_st(ctx, OPC_SB, rt, rs, s); + break; + case NM_SHS9: + gen_st(ctx, OPC_SH, rt, rs, s); + break; + case NM_SWS9: + gen_st(ctx, OPC_SW, rt, rs, s); + break; + case NM_LWC1S9: + gen_cop1_ldst(ctx, OPC_LWC1, rt, rs, s); + break; + case NM_LDC1S9: + gen_cop1_ldst(ctx, OPC_LDC1, rt, rs, s); + break; + case NM_SWC1S9: + gen_cop1_ldst(ctx, OPC_SWC1, rt, rs, s); + break; + case NM_SDC1S9: + gen_cop1_ldst(ctx, OPC_SDC1, rt, rs, s); + break; + case NM_P_PREFS9: + if (rt == 31) { + /* SYNCI */ + /* Break the TB to be able to sync copied instructions + immediately */ + ctx->base.is_jmp = DISAS_STOP; + } else { + /* PREF */ + /* Treat as NOP. */ + } + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + break; + case NM_P_LS_S1: + switch (extract32(ctx->opcode, 11, 4)) { + case NM_UALH: + case NM_UASH: + check_nms(ctx); + { + TCGv t0 = tcg_temp_new(); + TCGv t1 = tcg_temp_new(); + + gen_base_offset_addr(ctx, t0, rs, s); + + switch (extract32(ctx->opcode, 11, 4)) { + case NM_UALH: + tcg_gen_qemu_ld_tl(t0, t0, ctx->mem_idx, MO_TESW | + MO_UNALN); + gen_store_gpr(t0, rt); + break; + case NM_UASH: + gen_load_gpr(t1, rt); + tcg_gen_qemu_st_tl(t1, t0, ctx->mem_idx, MO_TEUW | + MO_UNALN); + break; + } + tcg_temp_free(t0); + tcg_temp_free(t1); + } + break; + case NM_P_LL: + switch (ctx->opcode & 0x03) { + case NM_LL: + gen_ld(ctx, OPC_LL, rt, rs, s); + break; + case NM_LLWP: + check_xnp(ctx); + gen_llwp(ctx, rs, 0, rt, extract32(ctx->opcode, 3, 5)); + break; + } + break; + case NM_P_SC: + switch (ctx->opcode & 0x03) { + case NM_SC: + gen_st_cond(ctx, OPC_SC, rt, rs, s); + break; + case NM_SCWP: + check_xnp(ctx); + gen_scwp(ctx, rs, 0, rt, extract32(ctx->opcode, 3, 5)); + break; + } + break; + case NM_CACHE: + check_cp0_enabled(ctx); + if (ctx->hflags & MIPS_HFLAG_ITC_CACHE) { + gen_cache_operation(ctx, rt, rs, s); + } + break; + } + break; + case NM_P_LS_WM: + case NM_P_LS_UAWM: + check_nms(ctx); + { + int count = extract32(ctx->opcode, 12, 3); + int counter = 0; + + offset = sextract32(ctx->opcode, 15, 1) << 8 | + extract32(ctx->opcode, 0, 8); + TCGv va = tcg_temp_new(); + TCGv t1 = tcg_temp_new(); + TCGMemOp memop = (extract32(ctx->opcode, 8, 3)) == + NM_P_LS_UAWM ? MO_UNALN : 0; + + count = (count == 0) ? 8 : count; + while (counter != count) { + int this_rt = ((rt + counter) & 0x1f) | (rt & 0x10); + int this_offset = offset + (counter << 2); + + gen_base_offset_addr(ctx, va, rs, this_offset); + + switch (extract32(ctx->opcode, 11, 1)) { + case NM_LWM: + tcg_gen_qemu_ld_tl(t1, va, ctx->mem_idx, + memop | MO_TESL); + gen_store_gpr(t1, this_rt); + if ((this_rt == rs) && + (counter != (count - 1))) { + /* UNPREDICTABLE */ + } + break; + case NM_SWM: + this_rt = (rt == 0) ? 0 : this_rt; + gen_load_gpr(t1, this_rt); + tcg_gen_qemu_st_tl(t1, va, ctx->mem_idx, + memop | MO_TEUL); + break; + } + counter++; + } + tcg_temp_free(va); + tcg_temp_free(t1); + } + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + } + break; + case NM_MOVE_BALC: + check_nms(ctx); + { + TCGv t0 = tcg_temp_new(); + int32_t s = sextract32(ctx->opcode, 0, 1) << 21 | + extract32(ctx->opcode, 1, 20) << 1; + rd = (extract32(ctx->opcode, 24, 1)) == 0 ? 4 : 5; + rt = decode_gpr_gpr4_zero(extract32(ctx->opcode, 25, 1) << 3 | + extract32(ctx->opcode, 21, 3)); + gen_load_gpr(t0, rt); + tcg_gen_mov_tl(cpu_gpr[rd], t0); + gen_compute_branch_nm(ctx, OPC_BGEZAL, 4, 0, 0, s); + tcg_temp_free(t0); + } + break; + case NM_P_BAL: + { + int32_t s = sextract32(ctx->opcode, 0, 1) << 25 | + extract32(ctx->opcode, 1, 24) << 1; + + if ((extract32(ctx->opcode, 25, 1)) == 0) { + /* BC */ + gen_compute_branch_nm(ctx, OPC_BEQ, 4, 0, 0, s); + } else { + /* BALC */ + gen_compute_branch_nm(ctx, OPC_BGEZAL, 4, 0, 0, s); + } + } + break; + case NM_P_J: + switch (extract32(ctx->opcode, 12, 4)) { + case NM_JALRC: + case NM_JALRC_HB: + gen_compute_branch_nm(ctx, OPC_JALR, 4, rs, rt, 0); + break; + case NM_P_BALRSC: + gen_compute_nanomips_pbalrsc_branch(ctx, rs, rt); + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + break; + case NM_P_BR1: + { + int32_t s = sextract32(ctx->opcode, 0, 1) << 14 | + extract32(ctx->opcode, 1, 13) << 1; + switch (extract32(ctx->opcode, 14, 2)) { + case NM_BEQC: + check_nms(ctx); + gen_compute_branch_nm(ctx, OPC_BEQ, 4, rs, rt, s); + break; + case NM_P_BR3A: + s = sextract32(ctx->opcode, 0, 1) << 14 | + extract32(ctx->opcode, 1, 13) << 1; + check_cp1_enabled(ctx); + switch (extract32(ctx->opcode, 16, 5)) { + case NM_BC1EQZC: + gen_compute_branch_cp1_nm(ctx, OPC_BC1EQZ, rt, s); + break; + case NM_BC1NEZC: + gen_compute_branch_cp1_nm(ctx, OPC_BC1NEZ, rt, s); + break; + case NM_BPOSGE32C: + check_dspr2(ctx); + { + int32_t imm = extract32(ctx->opcode, 1, 13) | + extract32(ctx->opcode, 0, 1) << 13; + + gen_compute_branch_nm(ctx, OPC_BPOSGE32, 4, -1, -2, + imm); + } + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + break; + case NM_BGEC: + if (rs == rt) { + gen_compute_compact_branch_nm(ctx, OPC_BC, rs, rt, s); + } else { + gen_compute_compact_branch_nm(ctx, OPC_BGEC, rs, rt, s); + } + break; + case NM_BGEUC: + if (rs == rt || rt == 0) { + gen_compute_compact_branch_nm(ctx, OPC_BC, 0, 0, s); + } else if (rs == 0) { + gen_compute_compact_branch_nm(ctx, OPC_BEQZC, rt, 0, s); + } else { + gen_compute_compact_branch_nm(ctx, OPC_BGEUC, rs, rt, s); + } + break; + } + } + break; + case NM_P_BR2: + { + int32_t s = sextract32(ctx->opcode, 0, 1) << 14 | + extract32(ctx->opcode, 1, 13) << 1; + switch (extract32(ctx->opcode, 14, 2)) { + case NM_BNEC: + check_nms(ctx); + gen_compute_branch_nm(ctx, OPC_BNE, 4, rs, rt, s); + break; + case NM_BLTC: + if (rs != 0 && rt != 0 && rs == rt) { + /* NOP */ + ctx->hflags |= MIPS_HFLAG_FBNSLOT; + } else { + gen_compute_compact_branch_nm(ctx, OPC_BLTC, rs, rt, s); + } + break; + case NM_BLTUC: + if (rs == 0 || rs == rt) { + /* NOP */ + ctx->hflags |= MIPS_HFLAG_FBNSLOT; + } else { + gen_compute_compact_branch_nm(ctx, OPC_BLTUC, rs, rt, s); + } + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + } + break; + case NM_P_BRI: + { + int32_t s = sextract32(ctx->opcode, 0, 1) << 11 | + extract32(ctx->opcode, 1, 10) << 1; + uint32_t u = extract32(ctx->opcode, 11, 7); + + gen_compute_imm_branch(ctx, extract32(ctx->opcode, 18, 3), + rt, u, s); + } + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + return 4; +} + +static int decode_nanomips_opc(CPUMIPSState *env, DisasContext *ctx) +{ + uint32_t op; + int rt = decode_gpr_gpr3(NANOMIPS_EXTRACT_RD(ctx->opcode)); + int rs = decode_gpr_gpr3(NANOMIPS_EXTRACT_RS(ctx->opcode)); + int rd = decode_gpr_gpr3(NANOMIPS_EXTRACT_RS1(ctx->opcode)); + int offset; + int imm; + + /* make sure instructions are on a halfword boundary */ + if (ctx->base.pc_next & 0x1) { + TCGv tmp = tcg_const_tl(ctx->base.pc_next); + tcg_gen_st_tl(tmp, cpu_env, offsetof(CPUMIPSState, CP0_BadVAddr)); + tcg_temp_free(tmp); + generate_exception_end(ctx, EXCP_AdEL); + return 2; + } + + op = extract32(ctx->opcode, 10, 6); + switch (op) { + case NM_P16_MV: + rt = NANOMIPS_EXTRACT_RD5(ctx->opcode); + if (rt != 0) { + /* MOVE */ + rs = NANOMIPS_EXTRACT_RS5(ctx->opcode); + gen_arith(ctx, OPC_ADDU, rt, rs, 0); + } else { + /* P16.RI */ + switch (extract32(ctx->opcode, 3, 2)) { + case NM_P16_SYSCALL: + if (extract32(ctx->opcode, 2, 1) == 0) { + generate_exception_end(ctx, EXCP_SYSCALL); + } else { + generate_exception_end(ctx, EXCP_RI); + } + break; + case NM_BREAK16: + generate_exception_end(ctx, EXCP_BREAK); + break; + case NM_SDBBP16: + if (is_uhi(extract32(ctx->opcode, 0, 3))) { + gen_helper_do_semihosting(cpu_env); + } else { + if (ctx->hflags & MIPS_HFLAG_SBRI) { + generate_exception_end(ctx, EXCP_RI); + } else { + generate_exception_end(ctx, EXCP_DBp); + } + } + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + } + break; + case NM_P16_SHIFT: + { + int shift = extract32(ctx->opcode, 0, 3); + uint32_t opc = 0; + shift = (shift == 0) ? 8 : shift; + + switch (extract32(ctx->opcode, 3, 1)) { + case NM_SLL16: + opc = OPC_SLL; + break; + case NM_SRL16: + opc = OPC_SRL; + break; + } + gen_shift_imm(ctx, opc, rt, rs, shift); + } + break; + case NM_P16C: + switch (ctx->opcode & 1) { + case NM_POOL16C_0: + gen_pool16c_nanomips_insn(ctx); + break; + case NM_LWXS16: + gen_ldxs(ctx, rt, rs, rd); + break; + } + break; + case NM_P16_A1: + switch (extract32(ctx->opcode, 6, 1)) { + case NM_ADDIUR1SP: + imm = extract32(ctx->opcode, 0, 6) << 2; + gen_arith_imm(ctx, OPC_ADDIU, rt, 29, imm); + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + break; + case NM_P16_A2: + switch (extract32(ctx->opcode, 3, 1)) { + case NM_ADDIUR2: + imm = extract32(ctx->opcode, 0, 3) << 2; + gen_arith_imm(ctx, OPC_ADDIU, rt, rs, imm); + break; + case NM_P_ADDIURS5: + rt = extract32(ctx->opcode, 5, 5); + if (rt != 0) { + /* imm = sign_extend(s[3] . s[2:0] , from_nbits = 4) */ + imm = (sextract32(ctx->opcode, 4, 1) << 3) | + (extract32(ctx->opcode, 0, 3)); + gen_arith_imm(ctx, OPC_ADDIU, rt, rt, imm); + } + break; + } + break; + case NM_P16_ADDU: + switch (ctx->opcode & 0x1) { + case NM_ADDU16: + gen_arith(ctx, OPC_ADDU, rd, rs, rt); + break; + case NM_SUBU16: + gen_arith(ctx, OPC_SUBU, rd, rs, rt); + break; + } + break; + case NM_P16_4X4: + rt = (extract32(ctx->opcode, 9, 1) << 3) | + extract32(ctx->opcode, 5, 3); + rs = (extract32(ctx->opcode, 4, 1) << 3) | + extract32(ctx->opcode, 0, 3); + rt = decode_gpr_gpr4(rt); + rs = decode_gpr_gpr4(rs); + switch ((extract32(ctx->opcode, 7, 2) & 0x2) | + (extract32(ctx->opcode, 3, 1))) { + case NM_ADDU4X4: + check_nms(ctx); + gen_arith(ctx, OPC_ADDU, rt, rs, rt); + break; + case NM_MUL4X4: + check_nms(ctx); + gen_r6_muldiv(ctx, R6_OPC_MUL, rt, rs, rt); + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + break; + case NM_LI16: + { + int imm = extract32(ctx->opcode, 0, 7); + imm = (imm == 0x7f ? -1 : imm); + if (rt != 0) { + tcg_gen_movi_tl(cpu_gpr[rt], imm); + } + } + break; + case NM_ANDI16: + { + uint32_t u = extract32(ctx->opcode, 0, 4); + u = (u == 12) ? 0xff : + (u == 13) ? 0xffff : u; + gen_logic_imm(ctx, OPC_ANDI, rt, rs, u); + } + break; + case NM_P16_LB: + offset = extract32(ctx->opcode, 0, 2); + switch (extract32(ctx->opcode, 2, 2)) { + case NM_LB16: + gen_ld(ctx, OPC_LB, rt, rs, offset); + break; + case NM_SB16: + rt = decode_gpr_gpr3_src_store( + NANOMIPS_EXTRACT_RD(ctx->opcode)); + gen_st(ctx, OPC_SB, rt, rs, offset); + break; + case NM_LBU16: + gen_ld(ctx, OPC_LBU, rt, rs, offset); + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + break; + case NM_P16_LH: + offset = extract32(ctx->opcode, 1, 2) << 1; + switch ((extract32(ctx->opcode, 3, 1) << 1) | (ctx->opcode & 1)) { + case NM_LH16: + gen_ld(ctx, OPC_LH, rt, rs, offset); + break; + case NM_SH16: + rt = decode_gpr_gpr3_src_store( + NANOMIPS_EXTRACT_RD(ctx->opcode)); + gen_st(ctx, OPC_SH, rt, rs, offset); + break; + case NM_LHU16: + gen_ld(ctx, OPC_LHU, rt, rs, offset); + break; + default: + generate_exception_end(ctx, EXCP_RI); + break; + } + break; + case NM_LW16: + offset = extract32(ctx->opcode, 0, 4) << 2; + gen_ld(ctx, OPC_LW, rt, rs, offset); + break; + case NM_LWSP16: + rt = NANOMIPS_EXTRACT_RD5(ctx->opcode); + offset = extract32(ctx->opcode, 0, 5) << 2; + gen_ld(ctx, OPC_LW, rt, 29, offset); + break; + case NM_LW4X4: + check_nms(ctx); + rt = (extract32(ctx->opcode, 9, 1) << 3) | + extract32(ctx->opcode, 5, 3); + rs = (extract32(ctx->opcode, 4, 1) << 3) | + extract32(ctx->opcode, 0, 3); + offset = (extract32(ctx->opcode, 3, 1) << 3) | + (extract32(ctx->opcode, 8, 1) << 2); + rt = decode_gpr_gpr4(rt); + rs = decode_gpr_gpr4(rs); + gen_ld(ctx, OPC_LW, rt, rs, offset); + break; + case NM_SW4X4: + check_nms(ctx); + rt = (extract32(ctx->opcode, 9, 1) << 3) | + extract32(ctx->opcode, 5, 3); + rs = (extract32(ctx->opcode, 4, 1) << 3) | + extract32(ctx->opcode, 0, 3); + offset = (extract32(ctx->opcode, 3, 1) << 3) | + (extract32(ctx->opcode, 8, 1) << 2); + rt = decode_gpr_gpr4_zero(rt); + rs = decode_gpr_gpr4(rs); + gen_st(ctx, OPC_SW, rt, rs, offset); + break; + case NM_LWGP16: + offset = extract32(ctx->opcode, 0, 7) << 2; + gen_ld(ctx, OPC_LW, rt, 28, offset); + break; + case NM_SWSP16: + rt = NANOMIPS_EXTRACT_RD5(ctx->opcode); + offset = extract32(ctx->opcode, 0, 5) << 2; + gen_st(ctx, OPC_SW, rt, 29, offset); + break; + case NM_SW16: + rt = decode_gpr_gpr3_src_store( + NANOMIPS_EXTRACT_RD(ctx->opcode)); + rs = decode_gpr_gpr3(NANOMIPS_EXTRACT_RS(ctx->opcode)); + offset = extract32(ctx->opcode, 0, 4) << 2; + gen_st(ctx, OPC_SW, rt, rs, offset); + break; + case NM_SWGP16: + rt = decode_gpr_gpr3_src_store( + NANOMIPS_EXTRACT_RD(ctx->opcode)); + offset = extract32(ctx->opcode, 0, 7) << 2; + gen_st(ctx, OPC_SW, rt, 28, offset); + break; + case NM_BC16: + gen_compute_branch_nm(ctx, OPC_BEQ, 2, 0, 0, + (sextract32(ctx->opcode, 0, 1) << 10) | + (extract32(ctx->opcode, 1, 9) << 1)); + break; + case NM_BALC16: + gen_compute_branch_nm(ctx, OPC_BGEZAL, 2, 0, 0, + (sextract32(ctx->opcode, 0, 1) << 10) | + (extract32(ctx->opcode, 1, 9) << 1)); + break; + case NM_BEQZC16: + gen_compute_branch_nm(ctx, OPC_BEQ, 2, rt, 0, + (sextract32(ctx->opcode, 0, 1) << 7) | + (extract32(ctx->opcode, 1, 6) << 1)); + break; + case NM_BNEZC16: + gen_compute_branch_nm(ctx, OPC_BNE, 2, rt, 0, + (sextract32(ctx->opcode, 0, 1) << 7) | + (extract32(ctx->opcode, 1, 6) << 1)); + break; + case NM_P16_BR: + switch (ctx->opcode & 0xf) { + case 0: + /* P16.JRC */ + switch (extract32(ctx->opcode, 4, 1)) { + case NM_JRC: + gen_compute_branch_nm(ctx, OPC_JR, 2, + extract32(ctx->opcode, 5, 5), 0, 0); + break; + case NM_JALRC16: + gen_compute_branch_nm(ctx, OPC_JALR, 2, + extract32(ctx->opcode, 5, 5), 31, 0); + break; + } + break; + default: + { + /* P16.BRI */ + uint32_t opc = extract32(ctx->opcode, 4, 3) < + extract32(ctx->opcode, 7, 3) ? OPC_BEQ : OPC_BNE; + gen_compute_branch_nm(ctx, opc, 2, rs, rt, + extract32(ctx->opcode, 0, 4) << 1); + } + break; + } + break; + case NM_P16_SR: + { + int count = extract32(ctx->opcode, 0, 4); + int u = extract32(ctx->opcode, 4, 4) << 4; + + rt = 30 + extract32(ctx->opcode, 9, 1); + switch (extract32(ctx->opcode, 8, 1)) { + case NM_SAVE16: + gen_save(ctx, rt, count, 0, u); + break; + case NM_RESTORE_JRC16: + gen_restore(ctx, rt, count, 0, u); + gen_compute_branch_nm(ctx, OPC_JR, 2, 31, 0, 0); + break; + } + } + break; + case NM_MOVEP: + case NM_MOVEPREV: + check_nms(ctx); + { + static const int gpr2reg1[] = {4, 5, 6, 7}; + static const int gpr2reg2[] = {5, 6, 7, 8}; + int re; + int rd2 = extract32(ctx->opcode, 3, 1) << 1 | + extract32(ctx->opcode, 8, 1); + int r1 = gpr2reg1[rd2]; + int r2 = gpr2reg2[rd2]; + int r3 = extract32(ctx->opcode, 4, 1) << 3 | + extract32(ctx->opcode, 0, 3); + int r4 = extract32(ctx->opcode, 9, 1) << 3 | + extract32(ctx->opcode, 5, 3); + TCGv t0 = tcg_temp_new(); + TCGv t1 = tcg_temp_new(); + if (op == NM_MOVEP) { + rd = r1; + re = r2; + rs = decode_gpr_gpr4_zero(r3); + rt = decode_gpr_gpr4_zero(r4); + } else { + rd = decode_gpr_gpr4(r3); + re = decode_gpr_gpr4(r4); + rs = r1; + rt = r2; + } + gen_load_gpr(t0, rs); + gen_load_gpr(t1, rt); + tcg_gen_mov_tl(cpu_gpr[rd], t0); + tcg_gen_mov_tl(cpu_gpr[re], t1); + tcg_temp_free(t0); + tcg_temp_free(t1); + } + break; + default: + return decode_nanomips_32_48_opc(env, ctx); + } + + return 2; +} + + /* SmartMIPS extension to MIPS32 */ #if defined(TARGET_MIPS64) @@ -17901,7 +22739,7 @@ static void decode_opc_special3_r6(CPUMIPSState *env, DisasContext *ctx) switch (op2) { case OPC_ALIGN: case OPC_ALIGN_END: - gen_align(ctx, OPC_ALIGN, rd, rs, rt, sa & 3); + gen_align(ctx, 32, rd, rs, rt, sa & 3); break; case OPC_BITSWAP: gen_bitswap(ctx, op2, rd, rt); @@ -17927,7 +22765,7 @@ static void decode_opc_special3_r6(CPUMIPSState *env, DisasContext *ctx) switch (op2) { case OPC_DALIGN: case OPC_DALIGN_END: - gen_align(ctx, OPC_DALIGN, rd, rs, rt, sa & 7); + gen_align(ctx, 64, rd, rs, rt, sa & 7); break; case OPC_DBITSWAP: gen_bitswap(ctx, op2, rd, rt); @@ -18586,7 +23424,7 @@ static void decode_opc_special3(CPUMIPSState *env, DisasContext *ctx) gen_rdhwr(ctx, rt, rd, extract32(ctx->opcode, 6, 3)); break; case OPC_FORK: - check_insn(ctx, ASE_MT); + check_mt(ctx); { TCGv t0 = tcg_temp_new(); TCGv t1 = tcg_temp_new(); @@ -18599,7 +23437,7 @@ static void decode_opc_special3(CPUMIPSState *env, DisasContext *ctx) } break; case OPC_YIELD: - check_insn(ctx, ASE_MT); + check_mt(ctx); { TCGv t0 = tcg_temp_new(); @@ -19896,22 +24734,22 @@ static void decode_opc(CPUMIPSState *env, DisasContext *ctx) op2 = MASK_MFMC0(ctx->opcode); switch (op2) { case OPC_DMT: - check_insn(ctx, ASE_MT); + check_cp0_mt(ctx); gen_helper_dmt(t0); gen_store_gpr(t0, rt); break; case OPC_EMT: - check_insn(ctx, ASE_MT); + check_cp0_mt(ctx); gen_helper_emt(t0); gen_store_gpr(t0, rt); break; case OPC_DVPE: - check_insn(ctx, ASE_MT); + check_cp0_mt(ctx); gen_helper_dvpe(t0, cpu_env); gen_store_gpr(t0, rt); break; case OPC_EVPE: - check_insn(ctx, ASE_MT); + check_cp0_mt(ctx); gen_helper_evpe(t0, cpu_env); gen_store_gpr(t0, rt); break; @@ -20447,6 +25285,8 @@ static void mips_tr_init_disas_context(DisasContextBase *dcbase, CPUState *cs) ctx->saved_pc = -1; ctx->insn_flags = env->insn_flags; ctx->CP0_Config1 = env->CP0_Config1; + ctx->CP0_Config3 = env->CP0_Config3; + ctx->CP0_Config5 = env->CP0_Config5; ctx->btarget = 0; ctx->kscrexist = (env->CP0_Config4 >> CP0C4_KScrExist) & 0xff; ctx->rxi = (env->CP0_Config3 >> CP0C3_RXI) & 1; @@ -20517,7 +25357,10 @@ static void mips_tr_translate_insn(DisasContextBase *dcbase, CPUState *cs) int is_slot; is_slot = ctx->hflags & MIPS_HFLAG_BMASK; - if (!(ctx->hflags & MIPS_HFLAG_M16)) { + if (ctx->insn_flags & ISA_NANOMIPS32) { + ctx->opcode = cpu_lduw_code(env, ctx->base.pc_next); + insn_bytes = decode_nanomips_opc(env, ctx); + } else if (!(ctx->hflags & MIPS_HFLAG_M16)) { ctx->opcode = cpu_ldl_code(env, ctx->base.pc_next); insn_bytes = 4; decode_opc(env, ctx); @@ -20956,8 +25799,8 @@ void cpu_state_reset(CPUMIPSState *env) env->CP0_Status |= (1 << CP0St_FR); } - if (env->CP0_Config3 & (1 << CP0C3_ISA)) { - /* microMIPS on reset when Config3.ISA == {1, 3} */ + if (env->CP0_Config3 & (1 << CP0C3_ISA) & (1 << (CP0C3_ISA + 1))) { + /* microMIPS on reset when Config3.ISA is 3 */ env->hflags |= MIPS_HFLAG_M16; } diff --git a/target/mips/translate_init.inc.c b/target/mips/translate_init.inc.c index c7ba6ee5f9..b3320b9dc7 100644 --- a/target/mips/translate_init.inc.c +++ b/target/mips/translate_init.inc.c @@ -449,6 +449,45 @@ const mips_def_t mips_defs[] = .insn_flags = CPU_MIPS32R6 | ASE_MICROMIPS, .mmu_type = MMU_TYPE_R4000, }, + { + .name = "I7200", + .CP0_PRid = 0x00010000, + .CP0_Config0 = MIPS_CONFIG0 | (1 << CP0C0_MM) | (0x2 << CP0C0_AR) | + (MMU_TYPE_R4000 << CP0C0_MT), + .CP0_Config1 = (1U << CP0C1_M) | (15 << CP0C1_MMU) | (2 << CP0C1_IS) | + (4 << CP0C1_IL) | (3 << CP0C1_IA) | (2 << CP0C1_DS) | + (4 << CP0C1_DL) | (3 << CP0C1_DA) | (1 << CP0C1_PC) | + (1 << CP0C1_EP), + .CP0_Config2 = MIPS_CONFIG2, + .CP0_Config3 = MIPS_CONFIG3 | (1U << CP0C3_M) | (1 << CP0C3_CMGCR) | + (1 << CP0C3_BI) | (1 << CP0C3_SC) | (3 << CP0C3_MMAR) | + (1 << CP0C3_ISA_ON_EXC) | (1 << CP0C3_ISA) | + (1 << CP0C3_ULRI) | (1 << CP0C3_RXI) | + (1 << CP0C3_DSP2P) | (1 << CP0C3_DSPP) | + (1 << CP0C3_CTXTC) | (1 << CP0C3_VInt) | + (1 << CP0C3_CDMM) | (1 << CP0C3_MT) | (1 << CP0C3_TL), + .CP0_Config4 = MIPS_CONFIG4 | (0xfc << CP0C4_KScrExist) | + (2 << CP0C4_IE) | (1U << CP0C4_M), + .CP0_Config5 = MIPS_CONFIG5 | (1 << CP0C5_MVH) | (1 << CP0C5_LLB), + .CP0_Config5_rw_bitmask = (1 << CP0C5_SBRI) | (1 << CP0C5_FRE) | + (1 << CP0C5_UFE), + .CP0_LLAddr_rw_bitmask = 0, + .CP0_LLAddr_shift = 0, + .SYNCI_Step = 32, + .CCRes = 2, + .CP0_Status_rw_bitmask = 0x3158FF1F, + .CP0_PageGrain = (1 << CP0PG_IEC) | (1 << CP0PG_XIE) | + (1U << CP0PG_RIE), + .CP0_PageGrain_rw_bitmask = 0, + .CP1_fcr0 = (1 << FCR0_FREP) | (1 << FCR0_HAS2008) | (1 << FCR0_F64) | + (1 << FCR0_L) | (1 << FCR0_W) | (1 << FCR0_D) | + (1 << FCR0_S) | (0x02 << FCR0_PRID) | (0x0 << FCR0_REV), + .CP1_fcr31 = (1 << FCR31_ABS2008) | (1 << FCR31_NAN2008), + .SEGBITS = 32, + .PABITS = 32, + .insn_flags = CPU_NANOMIPS32 | ASE_DSP | ASE_DSPR2 | ASE_MT, + .mmu_type = MMU_TYPE_R4000, + }, #if defined(TARGET_MIPS64) { .name = "R4000", |