Files
upkr/asm_unpackers/unpack_riscv.S

143 lines
2.0 KiB
ArmAsm

.section .text
#define FRAME_SIZE (256+32*4+4)
// x8 prob array ptr
// x9 prev was literal
// x10 out ptr
// x11 in ptr
// x12 offset
// x13 state
.global upkr_unpack
.type upkr_unpack, %function
upkr_unpack:
mv t4, ra
mv x17, x8
mv t6, x9
li x13, FRAME_SIZE
li x9, 128
1:
addi sp, sp, -1
sb x9, 0(sp)
addi x13, x13, -1
bnez x13, 1b
.Lmainloop:
li x14, 0
jal upkr_decode_bit
beqz x15, .Lliteral
slli x14, x14, 8
beqz x9, .Lread_offset_inc_x14
jal upkr_decode_bit
bnez x15, .Lread_offset
.Lfinished_offset:
addi x14, x14, 64
jal t3, upkr_decode_number
1:
add x14, x10, t0
lbu x14, (x14)
.Lstore_byte:
sb x14, (x10)
addi x10, x10, 1
addi x9, x9, 1
blt x9, x0, 1b
j .Lmainloop
.Lliteral:
jal upkr_decode_bit
addi x14, x14, -1
slli x14, x14, 1
add x14, x14, x15
srli x9, x14, 8
beqz x9, .Lliteral
j .Lstore_byte
.Lread_offset_inc_x14:
addi x14, x14, 1
.Lread_offset:
jal t3, upkr_decode_number
addi t0, x9, 1
bnez t0, .Lfinished_offset
.Ldone:
addi sp, sp, FRAME_SIZE
mv x8, x17
mv x9, t6
jr t4
// x14 context index
// return: x9 negtive decoded number
upkr_decode_number:
mv t5, x14
li x9, 0
li x8, -1
1:
jal upkr_decode_bit
beqz x15, 1f
jal upkr_decode_bit
beqz x15, 2f
add x9, x9, x8
2:
slli x8, x8, 1
j 1b
1:
add x9, x9, x8
mv x14, t5
jr t3
upkr_load_byte:
lbu x15, 0(x11)
addi x11, x11, 1
slli x13, x13, 8
add x13, x13, x15
// x8 prob array ptr
// x11 in ptr
// x13 state
// x14 context index
// return:
// x14 context index + 1
// x15 decoded bit
upkr_decode_bit:
srli x15, x13, 12
beqz x15, upkr_load_byte
mv t1, x14
mv t2, x10
add x14, x14, sp
lbu x12, 0(x14)
andi x10, x13, 255
sltu x15, x10, x12
srli x13, x13, 8
beqz x15, .Lelse
mul x13, x13, x12
add x13, x13, x10
li x10, 256 + 8
sub x10, x10, x12
srli x10, x10, 4
add x12, x12, x10
j .Lendif
.Lelse:
li x16, 256
sub x16, x16, x12
mul x13, x13, x16
add x13, x13, x10
sub x13, x13, x12
addi x10, x12, 8
srli x10, x10, 4
sub x12, x12, x10
.Lendif:
sb x12, 0(x14)
addi x14, t1, 1
mv x10, t2
ret