[PATCH v2] ARM: fix Thumb-2 bug in AES assembler code
Ard Biesheuvel
ard.biesheuvel at linaro.org
Tue Sep 17 13:16:35 EDT 2013
Patch 638591c enabled building the AES assembler code in Thumb2 mode.
However, this code used arithmetic involving PC rather than adr{l}
instructions to generate PC-relative references to the lookup tables,
and this break in Thumb mode due to the different PC offset.
Use adr instructions instead.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel at linaro.org>
---
v2:
- use 'adr' instruction rather than explicit PC arithmetic
- move the tables around so they are in range for 'adr' in each mode
arch/arm/crypto/aes-armv4.S | 388 ++++++++++++++++++++++----------------------
1 file changed, 193 insertions(+), 195 deletions(-)
diff --git a/arch/arm/crypto/aes-armv4.S b/arch/arm/crypto/aes-armv4.S
index 19d6cd6..7c10292 100644
--- a/arch/arm/crypto/aes-armv4.S
+++ b/arch/arm/crypto/aes-armv4.S
@@ -38,6 +38,102 @@
.text
+@ void AES_encrypt(const unsigned char *in, unsigned char *out,
+@ const AES_KEY *key) {
+.align 5
+ENTRY(AES_encrypt)
+ stmdb sp!,{r1,r4-r12,lr}
+ mov r12,r0 @ inp
+ mov r11,r2
+ adr r10, AES_Te @ Te
+#if __ARM_ARCH__<7
+ ldrb r0,[r12,#3] @ load input data in endian-neutral
+ ldrb r4,[r12,#2] @ manner...
+ ldrb r5,[r12,#1]
+ ldrb r6,[r12,#0]
+ orr r0,r0,r4,lsl#8
+ ldrb r1,[r12,#7]
+ orr r0,r0,r5,lsl#16
+ ldrb r4,[r12,#6]
+ orr r0,r0,r6,lsl#24
+ ldrb r5,[r12,#5]
+ ldrb r6,[r12,#4]
+ orr r1,r1,r4,lsl#8
+ ldrb r2,[r12,#11]
+ orr r1,r1,r5,lsl#16
+ ldrb r4,[r12,#10]
+ orr r1,r1,r6,lsl#24
+ ldrb r5,[r12,#9]
+ ldrb r6,[r12,#8]
+ orr r2,r2,r4,lsl#8
+ ldrb r3,[r12,#15]
+ orr r2,r2,r5,lsl#16
+ ldrb r4,[r12,#14]
+ orr r2,r2,r6,lsl#24
+ ldrb r5,[r12,#13]
+ ldrb r6,[r12,#12]
+ orr r3,r3,r4,lsl#8
+ orr r3,r3,r5,lsl#16
+ orr r3,r3,r6,lsl#24
+#else
+ ldr r0,[r12,#0]
+ ldr r1,[r12,#4]
+ ldr r2,[r12,#8]
+ ldr r3,[r12,#12]
+#ifdef __ARMEL__
+ rev r0,r0
+ rev r1,r1
+ rev r2,r2
+ rev r3,r3
+#endif
+#endif
+ bl _armv4_AES_encrypt
+
+ ldr r12,[sp],#4 @ pop out
+#if __ARM_ARCH__>=7
+#ifdef __ARMEL__
+ rev r0,r0
+ rev r1,r1
+ rev r2,r2
+ rev r3,r3
+#endif
+ str r0,[r12,#0]
+ str r1,[r12,#4]
+ str r2,[r12,#8]
+ str r3,[r12,#12]
+#else
+ mov r4,r0,lsr#24 @ write output in endian-neutral
+ mov r5,r0,lsr#16 @ manner...
+ mov r6,r0,lsr#8
+ strb r4,[r12,#0]
+ strb r5,[r12,#1]
+ mov r4,r1,lsr#24
+ strb r6,[r12,#2]
+ mov r5,r1,lsr#16
+ strb r0,[r12,#3]
+ mov r6,r1,lsr#8
+ strb r4,[r12,#4]
+ strb r5,[r12,#5]
+ mov r4,r2,lsr#24
+ strb r6,[r12,#6]
+ mov r5,r2,lsr#16
+ strb r1,[r12,#7]
+ mov r6,r2,lsr#8
+ strb r4,[r12,#8]
+ strb r5,[r12,#9]
+ mov r4,r3,lsr#24
+ strb r6,[r12,#10]
+ mov r5,r3,lsr#16
+ strb r2,[r12,#11]
+ mov r6,r3,lsr#8
+ strb r4,[r12,#12]
+ strb r5,[r12,#13]
+ strb r6,[r12,#14]
+ strb r3,[r12,#15]
+#endif
+ ldmia sp!,{r4-r12,pc}
+ENDPROC(AES_encrypt)
+
.type AES_Te,%object
.align 5
AES_Te:
@@ -144,103 +240,6 @@ AES_Te:
.word 0x1B000000, 0x36000000, 0, 0, 0, 0, 0, 0
.size AES_Te,.-AES_Te
-@ void AES_encrypt(const unsigned char *in, unsigned char *out,
-@ const AES_KEY *key) {
-.align 5
-ENTRY(AES_encrypt)
- sub r3,pc,#8 @ AES_encrypt
- stmdb sp!,{r1,r4-r12,lr}
- mov r12,r0 @ inp
- mov r11,r2
- sub r10,r3,#AES_encrypt-AES_Te @ Te
-#if __ARM_ARCH__<7
- ldrb r0,[r12,#3] @ load input data in endian-neutral
- ldrb r4,[r12,#2] @ manner...
- ldrb r5,[r12,#1]
- ldrb r6,[r12,#0]
- orr r0,r0,r4,lsl#8
- ldrb r1,[r12,#7]
- orr r0,r0,r5,lsl#16
- ldrb r4,[r12,#6]
- orr r0,r0,r6,lsl#24
- ldrb r5,[r12,#5]
- ldrb r6,[r12,#4]
- orr r1,r1,r4,lsl#8
- ldrb r2,[r12,#11]
- orr r1,r1,r5,lsl#16
- ldrb r4,[r12,#10]
- orr r1,r1,r6,lsl#24
- ldrb r5,[r12,#9]
- ldrb r6,[r12,#8]
- orr r2,r2,r4,lsl#8
- ldrb r3,[r12,#15]
- orr r2,r2,r5,lsl#16
- ldrb r4,[r12,#14]
- orr r2,r2,r6,lsl#24
- ldrb r5,[r12,#13]
- ldrb r6,[r12,#12]
- orr r3,r3,r4,lsl#8
- orr r3,r3,r5,lsl#16
- orr r3,r3,r6,lsl#24
-#else
- ldr r0,[r12,#0]
- ldr r1,[r12,#4]
- ldr r2,[r12,#8]
- ldr r3,[r12,#12]
-#ifdef __ARMEL__
- rev r0,r0
- rev r1,r1
- rev r2,r2
- rev r3,r3
-#endif
-#endif
- bl _armv4_AES_encrypt
-
- ldr r12,[sp],#4 @ pop out
-#if __ARM_ARCH__>=7
-#ifdef __ARMEL__
- rev r0,r0
- rev r1,r1
- rev r2,r2
- rev r3,r3
-#endif
- str r0,[r12,#0]
- str r1,[r12,#4]
- str r2,[r12,#8]
- str r3,[r12,#12]
-#else
- mov r4,r0,lsr#24 @ write output in endian-neutral
- mov r5,r0,lsr#16 @ manner...
- mov r6,r0,lsr#8
- strb r4,[r12,#0]
- strb r5,[r12,#1]
- mov r4,r1,lsr#24
- strb r6,[r12,#2]
- mov r5,r1,lsr#16
- strb r0,[r12,#3]
- mov r6,r1,lsr#8
- strb r4,[r12,#4]
- strb r5,[r12,#5]
- mov r4,r2,lsr#24
- strb r6,[r12,#6]
- mov r5,r2,lsr#16
- strb r1,[r12,#7]
- mov r6,r2,lsr#8
- strb r4,[r12,#8]
- strb r5,[r12,#9]
- mov r4,r3,lsr#24
- strb r6,[r12,#10]
- mov r5,r3,lsr#16
- strb r2,[r12,#11]
- mov r6,r3,lsr#8
- strb r4,[r12,#12]
- strb r5,[r12,#13]
- strb r6,[r12,#14]
- strb r3,[r12,#15]
-#endif
- ldmia sp!,{r4-r12,pc}
-ENDPROC(AES_encrypt)
-
.type _armv4_AES_encrypt,%function
.align 2
_armv4_AES_encrypt:
@@ -381,7 +380,6 @@ _armv4_AES_encrypt:
.align 5
ENTRY(private_AES_set_encrypt_key)
_armv4_AES_set_encrypt_key:
- sub r3,pc,#8 @ AES_set_encrypt_key
teq r0,#0
moveq r0,#-1
beq .Labrt
@@ -398,7 +396,7 @@ _armv4_AES_set_encrypt_key:
bne .Labrt
.Lok: stmdb sp!,{r4-r12,lr}
- sub r10,r3,#_armv4_AES_set_encrypt_key-AES_Te-1024 @ Te4
+ adr r10, (AES_Te + 1024) @ Te4
mov r12,r0 @ inp
mov lr,r1 @ bits
@@ -737,6 +735,102 @@ ENTRY(private_AES_set_decrypt_key)
ldmia sp!,{r4-r12,pc}
ENDPROC(private_AES_set_decrypt_key)
+@ void AES_decrypt(const unsigned char *in, unsigned char *out,
+@ const AES_KEY *key) {
+.align 5
+ENTRY(AES_decrypt)
+ stmdb sp!,{r1,r4-r12,lr}
+ mov r12,r0 @ inp
+ mov r11,r2
+ adr r10, AES_Td @ Td
+#if __ARM_ARCH__<7
+ ldrb r0,[r12,#3] @ load input data in endian-neutral
+ ldrb r4,[r12,#2] @ manner...
+ ldrb r5,[r12,#1]
+ ldrb r6,[r12,#0]
+ orr r0,r0,r4,lsl#8
+ ldrb r1,[r12,#7]
+ orr r0,r0,r5,lsl#16
+ ldrb r4,[r12,#6]
+ orr r0,r0,r6,lsl#24
+ ldrb r5,[r12,#5]
+ ldrb r6,[r12,#4]
+ orr r1,r1,r4,lsl#8
+ ldrb r2,[r12,#11]
+ orr r1,r1,r5,lsl#16
+ ldrb r4,[r12,#10]
+ orr r1,r1,r6,lsl#24
+ ldrb r5,[r12,#9]
+ ldrb r6,[r12,#8]
+ orr r2,r2,r4,lsl#8
+ ldrb r3,[r12,#15]
+ orr r2,r2,r5,lsl#16
+ ldrb r4,[r12,#14]
+ orr r2,r2,r6,lsl#24
+ ldrb r5,[r12,#13]
+ ldrb r6,[r12,#12]
+ orr r3,r3,r4,lsl#8
+ orr r3,r3,r5,lsl#16
+ orr r3,r3,r6,lsl#24
+#else
+ ldr r0,[r12,#0]
+ ldr r1,[r12,#4]
+ ldr r2,[r12,#8]
+ ldr r3,[r12,#12]
+#ifdef __ARMEL__
+ rev r0,r0
+ rev r1,r1
+ rev r2,r2
+ rev r3,r3
+#endif
+#endif
+ bl _armv4_AES_decrypt
+
+ ldr r12,[sp],#4 @ pop out
+#if __ARM_ARCH__>=7
+#ifdef __ARMEL__
+ rev r0,r0
+ rev r1,r1
+ rev r2,r2
+ rev r3,r3
+#endif
+ str r0,[r12,#0]
+ str r1,[r12,#4]
+ str r2,[r12,#8]
+ str r3,[r12,#12]
+#else
+ mov r4,r0,lsr#24 @ write output in endian-neutral
+ mov r5,r0,lsr#16 @ manner...
+ mov r6,r0,lsr#8
+ strb r4,[r12,#0]
+ strb r5,[r12,#1]
+ mov r4,r1,lsr#24
+ strb r6,[r12,#2]
+ mov r5,r1,lsr#16
+ strb r0,[r12,#3]
+ mov r6,r1,lsr#8
+ strb r4,[r12,#4]
+ strb r5,[r12,#5]
+ mov r4,r2,lsr#24
+ strb r6,[r12,#6]
+ mov r5,r2,lsr#16
+ strb r1,[r12,#7]
+ mov r6,r2,lsr#8
+ strb r4,[r12,#8]
+ strb r5,[r12,#9]
+ mov r4,r3,lsr#24
+ strb r6,[r12,#10]
+ mov r5,r3,lsr#16
+ strb r2,[r12,#11]
+ mov r6,r3,lsr#8
+ strb r4,[r12,#12]
+ strb r5,[r12,#13]
+ strb r6,[r12,#14]
+ strb r3,[r12,#15]
+#endif
+ ldmia sp!,{r4-r12,pc}
+ENDPROC(AES_decrypt)
+
.type AES_Td,%object
.align 5
AES_Td:
@@ -839,102 +933,6 @@ AES_Td:
.byte 0xe1, 0x69, 0x14, 0x63, 0x55, 0x21, 0x0c, 0x7d
.size AES_Td,.-AES_Td
-@ void AES_decrypt(const unsigned char *in, unsigned char *out,
-@ const AES_KEY *key) {
-.align 5
-ENTRY(AES_decrypt)
- sub r3,pc,#8 @ AES_decrypt
- stmdb sp!,{r1,r4-r12,lr}
- mov r12,r0 @ inp
- mov r11,r2
- sub r10,r3,#AES_decrypt-AES_Td @ Td
-#if __ARM_ARCH__<7
- ldrb r0,[r12,#3] @ load input data in endian-neutral
- ldrb r4,[r12,#2] @ manner...
- ldrb r5,[r12,#1]
- ldrb r6,[r12,#0]
- orr r0,r0,r4,lsl#8
- ldrb r1,[r12,#7]
- orr r0,r0,r5,lsl#16
- ldrb r4,[r12,#6]
- orr r0,r0,r6,lsl#24
- ldrb r5,[r12,#5]
- ldrb r6,[r12,#4]
- orr r1,r1,r4,lsl#8
- ldrb r2,[r12,#11]
- orr r1,r1,r5,lsl#16
- ldrb r4,[r12,#10]
- orr r1,r1,r6,lsl#24
- ldrb r5,[r12,#9]
- ldrb r6,[r12,#8]
- orr r2,r2,r4,lsl#8
- ldrb r3,[r12,#15]
- orr r2,r2,r5,lsl#16
- ldrb r4,[r12,#14]
- orr r2,r2,r6,lsl#24
- ldrb r5,[r12,#13]
- ldrb r6,[r12,#12]
- orr r3,r3,r4,lsl#8
- orr r3,r3,r5,lsl#16
- orr r3,r3,r6,lsl#24
-#else
- ldr r0,[r12,#0]
- ldr r1,[r12,#4]
- ldr r2,[r12,#8]
- ldr r3,[r12,#12]
-#ifdef __ARMEL__
- rev r0,r0
- rev r1,r1
- rev r2,r2
- rev r3,r3
-#endif
-#endif
- bl _armv4_AES_decrypt
-
- ldr r12,[sp],#4 @ pop out
-#if __ARM_ARCH__>=7
-#ifdef __ARMEL__
- rev r0,r0
- rev r1,r1
- rev r2,r2
- rev r3,r3
-#endif
- str r0,[r12,#0]
- str r1,[r12,#4]
- str r2,[r12,#8]
- str r3,[r12,#12]
-#else
- mov r4,r0,lsr#24 @ write output in endian-neutral
- mov r5,r0,lsr#16 @ manner...
- mov r6,r0,lsr#8
- strb r4,[r12,#0]
- strb r5,[r12,#1]
- mov r4,r1,lsr#24
- strb r6,[r12,#2]
- mov r5,r1,lsr#16
- strb r0,[r12,#3]
- mov r6,r1,lsr#8
- strb r4,[r12,#4]
- strb r5,[r12,#5]
- mov r4,r2,lsr#24
- strb r6,[r12,#6]
- mov r5,r2,lsr#16
- strb r1,[r12,#7]
- mov r6,r2,lsr#8
- strb r4,[r12,#8]
- strb r5,[r12,#9]
- mov r4,r3,lsr#24
- strb r6,[r12,#10]
- mov r5,r3,lsr#16
- strb r2,[r12,#11]
- mov r6,r3,lsr#8
- strb r4,[r12,#12]
- strb r5,[r12,#13]
- strb r6,[r12,#14]
- strb r3,[r12,#15]
-#endif
- ldmia sp!,{r4-r12,pc}
-ENDPROC(AES_decrypt)
.type _armv4_AES_decrypt,%function
.align 2
--
1.8.1.2
More information about the linux-arm-kernel
mailing list