* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
-#include "asm.S"
+#include "libavutil/arm/asm.S"
#define W1 22725 /* cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5 */
#define W2 21407 /* cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5 */
#define W26 (W2 | (W6 << 16))
#define W57 (W5 | (W7 << 16))
- .text
- .align
-w13: .long W13
-w26: .long W26
-w57: .long W57
-
function idct_row_armv5te
str lr, [sp, #-4]!
- ldrd v1, [a1, #8]
- ldrd a3, [a1] /* a3 = row[1:0], a4 = row[3:2] */
+ ldrd v1, v2, [a1, #8]
+ ldrd a3, a4, [a1] /* a3 = row[1:0], a4 = row[3:2] */
orrs v1, v1, v2
+ itt eq
cmpeq v1, a4
cmpeq v1, a3, lsr #16
beq row_dc_only
mov ip, #16384
sub ip, ip, #1 /* ip = W4 */
smlabb v1, ip, a3, v1 /* v1 = W4*row[0]+(1<<(RS-1)) */
- ldr ip, w26 /* ip = W2 | (W6 << 16) */
+ ldr ip, =W26 /* ip = W2 | (W6 << 16) */
smultb a2, ip, a4
smulbb lr, ip, a4
add v2, v1, a2
sub v4, v1, lr
add v1, v1, lr
- ldr ip, w13 /* ip = W1 | (W3 << 16) */
- ldr lr, w57 /* lr = W5 | (W7 << 16) */
+ ldr ip, =W13 /* ip = W1 | (W3 << 16) */
+ ldr lr, =W57 /* lr = W5 | (W7 << 16) */
smulbt v5, ip, a3
smultt v6, lr, a4
smlatt v5, ip, a4, v5
smultt fp, lr, a3
sub v7, v7, a2
smulbt a2, lr, a4
- ldrd a3, [a1, #8] /* a3=row[5:4] a4=row[7:6] */
+ ldrd a3, a4, [a1, #8] /* a3=row[5:4] a4=row[7:6] */
sub fp, fp, a2
orrs a2, a3, a4
smlatt v7, ip, a4, v7
sub fp, fp, a2
- ldr ip, w26 /* ip = W2 | (W6 << 16) */
+ ldr ip, =W26 /* ip = W2 | (W6 << 16) */
mov a2, #16384
sub a2, a2, #1 /* a2 = W4 */
smulbb a2, a2, a3 /* a2 = W4*row[4] */
add a2, v4, fp
mov a2, a2, lsr #11
add a4, a4, a2, lsl #16
- strd a3, [a1]
+ strd a3, a4, [a1]
sub a2, v4, fp
mov a3, a2, lsr #11
sub a2, v1, v5
mov a2, a2, lsr #11
add a4, a4, a2, lsl #16
- strd a3, [a1, #8]
+ strd a3, a4, [a1, #8]
ldr pc, [sp], #4
bic a3, a3, #0xe000
mov a3, a3, lsl #3
mov a4, a3
- strd a3, [a1]
- strd a3, [a1, #8]
+ strd a3, a4, [a1]
+ strd a3, a4, [a1, #8]
ldr pc, [sp], #4
endfunc
sub v4, v2, a3
sub v6, v2, a3
add fp, v2, a3
- ldr ip, w26
+ ldr ip, =W26
ldr a4, [a1, #(16*2)]
add v2, v2, a3
stmfd sp!, {v1, v2, v3, v4, v5, v6, v7, fp}
- ldr ip, w13
+ ldr ip, =W13
ldr a4, [a1, #(16*1)]
- ldr lr, w57
+ ldr lr, =W57
smulbb v1, ip, a4
smultb v3, ip, a4
smulbb v5, lr, a4
ldmfd sp!, {a3, a4}
adds a2, a3, v1
mov a2, a2, lsr #20
+ it mi
orrmi a2, a2, #0xf000
add ip, a4, v2
mov ip, ip, asr #20
str a2, [a1]
subs a3, a3, v1
mov a2, a3, lsr #20
+ it mi
orrmi a2, a2, #0xf000
sub a4, a4, v2
mov a4, a4, asr #20
subs a2, a3, v3
mov a2, a2, lsr #20
+ it mi
orrmi a2, a2, #0xf000
sub ip, a4, v4
mov ip, ip, asr #20
str a2, [a1, #(16*1)]
adds a3, a3, v3
mov a2, a3, lsr #20
+ it mi
orrmi a2, a2, #0xf000
add a4, a4, v4
mov a4, a4, asr #20
adds a2, a3, v5
mov a2, a2, lsr #20
+ it mi
orrmi a2, a2, #0xf000
add ip, a4, v6
mov ip, ip, asr #20
str a2, [a1, #(16*2)]
subs a3, a3, v5
mov a2, a3, lsr #20
+ it mi
orrmi a2, a2, #0xf000
sub a4, a4, v6
mov a4, a4, asr #20
adds a2, a3, v7
mov a2, a2, lsr #20
+ it mi
orrmi a2, a2, #0xf000
add ip, a4, fp
mov ip, ip, asr #20
str a2, [a1, #(16*3)]
subs a3, a3, v7
mov a2, a3, lsr #20
+ it mi
orrmi a2, a2, #0xf000
sub a4, a4, fp
mov a4, a4, asr #20
ldr pc, [sp], #4
endfunc
+.macro clip dst, src:vararg
+ movs \dst, \src
+ it mi
+ movmi \dst, #0
+ cmp \dst, #255
+ it gt
+ movgt \dst, #255
+.endm
+
+.macro aclip dst, src:vararg
+ adds \dst, \src
+ it mi
+ movmi \dst, #0
+ cmp \dst, #255
+ it gt
+ movgt \dst, #255
+.endm
+
function idct_col_put_armv5te
str lr, [sp, #-4]!
ldmfd sp!, {a3, a4}
ldr lr, [sp, #32]
add a2, a3, v1
- movs a2, a2, asr #20
- movmi a2, #0
- cmp a2, #255
- movgt a2, #255
+ clip a2, a2, asr #20
add ip, a4, v2
- movs ip, ip, asr #20
- movmi ip, #0
- cmp ip, #255
- movgt ip, #255
+ clip ip, ip, asr #20
orr a2, a2, ip, lsl #8
sub a3, a3, v1
- movs a3, a3, asr #20
- movmi a3, #0
- cmp a3, #255
- movgt a3, #255
+ clip a3, a3, asr #20
sub a4, a4, v2
- movs a4, a4, asr #20
- movmi a4, #0
- cmp a4, #255
+ clip a4, a4, asr #20
ldr v1, [sp, #28]
- movgt a4, #255
strh a2, [v1]
add a2, v1, #2
str a2, [sp, #28]
orr a2, a3, a4, lsl #8
rsb v2, lr, lr, lsl #3
ldmfd sp!, {a3, a4}
- strh a2, [v2, v1]!
+ strh_pre a2, v2, v1
sub a2, a3, v3
- movs a2, a2, asr #20
- movmi a2, #0
- cmp a2, #255
- movgt a2, #255
+ clip a2, a2, asr #20
sub ip, a4, v4
- movs ip, ip, asr #20
- movmi ip, #0
- cmp ip, #255
- movgt ip, #255
+ clip ip, ip, asr #20
orr a2, a2, ip, lsl #8
- strh a2, [v1, lr]!
+ strh_pre a2, v1, lr
add a3, a3, v3
- movs a2, a3, asr #20
- movmi a2, #0
- cmp a2, #255
- movgt a2, #255
+ clip a2, a3, asr #20
add a4, a4, v4
- movs a4, a4, asr #20
- movmi a4, #0
- cmp a4, #255
- movgt a4, #255
+ clip a4, a4, asr #20
orr a2, a2, a4, lsl #8
ldmfd sp!, {a3, a4}
- strh a2, [v2, -lr]!
+ strh_dpre a2, v2, lr
add a2, a3, v5
- movs a2, a2, asr #20
- movmi a2, #0
- cmp a2, #255
- movgt a2, #255
+ clip a2, a2, asr #20
add ip, a4, v6
- movs ip, ip, asr #20
- movmi ip, #0
- cmp ip, #255
- movgt ip, #255
+ clip ip, ip, asr #20
orr a2, a2, ip, lsl #8
- strh a2, [v1, lr]!
+ strh_pre a2, v1, lr
sub a3, a3, v5
- movs a2, a3, asr #20
- movmi a2, #0
- cmp a2, #255
- movgt a2, #255
+ clip a2, a3, asr #20
sub a4, a4, v6
- movs a4, a4, asr #20
- movmi a4, #0
- cmp a4, #255
- movgt a4, #255
+ clip a4, a4, asr #20
orr a2, a2, a4, lsl #8
ldmfd sp!, {a3, a4}
- strh a2, [v2, -lr]!
+ strh_dpre a2, v2, lr
add a2, a3, v7
- movs a2, a2, asr #20
- movmi a2, #0
- cmp a2, #255
- movgt a2, #255
+ clip a2, a2, asr #20
add ip, a4, fp
- movs ip, ip, asr #20
- movmi ip, #0
- cmp ip, #255
- movgt ip, #255
+ clip ip, ip, asr #20
orr a2, a2, ip, lsl #8
strh a2, [v1, lr]
sub a3, a3, v7
- movs a2, a3, asr #20
- movmi a2, #0
- cmp a2, #255
- movgt a2, #255
+ clip a2, a3, asr #20
sub a4, a4, fp
- movs a4, a4, asr #20
- movmi a4, #0
- cmp a4, #255
- movgt a4, #255
+ clip a4, a4, asr #20
orr a2, a2, a4, lsl #8
- strh a2, [v2, -lr]
+ strh_dpre a2, v2, lr
ldr pc, [sp], #4
endfunc
ldmfd sp!, {a3, a4}
ldrh ip, [lr]
add a2, a3, v1
- mov a2, a2, asr #20
sub a3, a3, v1
and v1, ip, #255
- adds a2, a2, v1
- movmi a2, #0
- cmp a2, #255
- movgt a2, #255
+ aclip a2, v1, a2, asr #20
add v1, a4, v2
mov v1, v1, asr #20
- adds v1, v1, ip, lsr #8
- movmi v1, #0
- cmp v1, #255
- movgt v1, #255
+ aclip v1, v1, ip, lsr #8
orr a2, a2, v1, lsl #8
ldr v1, [sp, #32]
sub a4, a4, v2
rsb v2, v1, v1, lsl #3
- ldrh ip, [v2, lr]!
+ ldrh_pre ip, v2, lr
strh a2, [lr]
- mov a3, a3, asr #20
and a2, ip, #255
- adds a3, a3, a2
- movmi a3, #0
- cmp a3, #255
- movgt a3, #255
+ aclip a3, a2, a3, asr #20
mov a4, a4, asr #20
- adds a4, a4, ip, lsr #8
- movmi a4, #0
- cmp a4, #255
- movgt a4, #255
+ aclip a4, a4, ip, lsr #8
add a2, lr, #2
str a2, [sp, #28]
orr a2, a3, a4, lsl #8
strh a2, [v2]
ldmfd sp!, {a3, a4}
- ldrh ip, [lr, v1]!
+ ldrh_pre ip, lr, v1
sub a2, a3, v3
- mov a2, a2, asr #20
add a3, a3, v3
and v3, ip, #255
- adds a2, a2, v3
- movmi a2, #0
- cmp a2, #255
- movgt a2, #255
+ aclip a2, v3, a2, asr #20
sub v3, a4, v4
mov v3, v3, asr #20
- adds v3, v3, ip, lsr #8
- movmi v3, #0
- cmp v3, #255
- movgt v3, #255
+ aclip v3, v3, ip, lsr #8
orr a2, a2, v3, lsl #8
add a4, a4, v4
- ldrh ip, [v2, -v1]!
+ ldrh_dpre ip, v2, v1
strh a2, [lr]
- mov a3, a3, asr #20
and a2, ip, #255
- adds a3, a3, a2
- movmi a3, #0
- cmp a3, #255
- movgt a3, #255
+ aclip a3, a2, a3, asr #20
mov a4, a4, asr #20
- adds a4, a4, ip, lsr #8
- movmi a4, #0
- cmp a4, #255
- movgt a4, #255
+ aclip a4, a4, ip, lsr #8
orr a2, a3, a4, lsl #8
strh a2, [v2]
ldmfd sp!, {a3, a4}
- ldrh ip, [lr, v1]!
+ ldrh_pre ip, lr, v1
add a2, a3, v5
- mov a2, a2, asr #20
sub a3, a3, v5
and v3, ip, #255
- adds a2, a2, v3
- movmi a2, #0
- cmp a2, #255
- movgt a2, #255
+ aclip a2, v3, a2, asr #20
add v3, a4, v6
mov v3, v3, asr #20
- adds v3, v3, ip, lsr #8
- movmi v3, #0
- cmp v3, #255
- movgt v3, #255
+ aclip v3, v3, ip, lsr #8
orr a2, a2, v3, lsl #8
sub a4, a4, v6
- ldrh ip, [v2, -v1]!
+ ldrh_dpre ip, v2, v1
strh a2, [lr]
- mov a3, a3, asr #20
and a2, ip, #255
- adds a3, a3, a2
- movmi a3, #0
- cmp a3, #255
- movgt a3, #255
+ aclip a3, a2, a3, asr #20
mov a4, a4, asr #20
- adds a4, a4, ip, lsr #8
- movmi a4, #0
- cmp a4, #255
- movgt a4, #255
+ aclip a4, a4, ip, lsr #8
orr a2, a3, a4, lsl #8
strh a2, [v2]
ldmfd sp!, {a3, a4}
- ldrh ip, [lr, v1]!
+ ldrh_pre ip, lr, v1
add a2, a3, v7
- mov a2, a2, asr #20
sub a3, a3, v7
and v3, ip, #255
- adds a2, a2, v3
- movmi a2, #0
- cmp a2, #255
- movgt a2, #255
+ aclip a2, v3, a2, asr #20
add v3, a4, fp
mov v3, v3, asr #20
- adds v3, v3, ip, lsr #8
- movmi v3, #0
- cmp v3, #255
- movgt v3, #255
+ aclip v3, v3, ip, lsr #8
orr a2, a2, v3, lsl #8
sub a4, a4, fp
- ldrh ip, [v2, -v1]!
+ ldrh_dpre ip, v2, v1
strh a2, [lr]
- mov a3, a3, asr #20
and a2, ip, #255
- adds a3, a3, a2
- movmi a3, #0
- cmp a3, #255
- movgt a3, #255
+ aclip a3, a2, a3, asr #20
mov a4, a4, asr #20
- adds a4, a4, ip, lsr #8
- movmi a4, #0
- cmp a4, #255
- movgt a4, #255
+ aclip a4, a4, ip, lsr #8
orr a2, a3, a4, lsl #8
strh a2, [v2]
projects / ffmpeg / blobdiff