Using ARM Unified Assembler Language

Change-Id: Iae32a8ba8eff6087330e458fafc912a12fee4509

23 files changed, 139 insertions, 127 deletions

diff --git a/apps/plugins/mpegplayer/libmpeg2/idct_arm.S b/apps/plugins/mpegplayer/libmpeg2/idct_arm.S
index 97a87a8b59..90eb5031c7 100644
--- a/apps/plugins/mpegplayer/libmpeg2/idct_arm.S
+++ b/apps/plugins/mpegplayer/libmpeg2/idct_arm.S
@@ -43,8 +43,8 @@
     ldrsh  r7, [r0, #12]          /* d2 */
     ldrsh  r8, [r0, #14]          /* d3 */
     orrs   r9, r2, r3
-    orreqs r9, r4, r5
-    orreqs r9, r6, r7
+    orrseq r9, r4, r5
+    orrseq r9, r6, r7
     cmpeq  r8, #0
     bne    2f
     mov    r1, r1, asl #15
@@ -320,7 +320,7 @@ mpeg2_idct_copy:
 mpeg2_idct_add:
     cmp    r0, #129
     mov    r0, r1
-    ldreqsh r1, [r0, #0]
+    ldrsheq r1, [r0, #0]
     bne    1f
     and    r1, r1, #0x70
     cmp    r1, #0x40
diff --git a/apps/plugins/mpegplayer/libmpeg2/idct_armv6.S b/apps/plugins/mpegplayer/libmpeg2/idct_armv6.S
index dc53cbd7bd..a259721410 100644
--- a/apps/plugins/mpegplayer/libmpeg2/idct_armv6.S
+++ b/apps/plugins/mpegplayer/libmpeg2/idct_armv6.S
@@ -19,6 +19,7 @@
  *
  ****************************************************************************/
 
+#include "config.h"
 
     .global     mpeg2_idct_copy
     .type       mpeg2_idct_copy, %function
@@ -228,7 +229,7 @@ mpeg2_idct_copy:
 mpeg2_idct_add:
     cmp    r0, #129
     mov    r0, r1
-    ldreqsh r1, [r0, #0]
+    ldrsheq r1, [r0, #0]
     bne    1f
     and    r1, r1, #0x70
     cmp    r1, #0x40
@@ -260,7 +261,7 @@ mpeg2_idct_add:
     strd   r4, [r1]             @ r4, r5
     add    r1, r1, r2
     cmp    r0, r3
-    ldrlod r8, [r1]             @ r8, r9
+    ldrdlo r8, [r1]             @ r8, r9
     blo    2b
 
     ldmfd  sp!, {r4-r11, pc}
@@ -291,7 +292,7 @@ mpeg2_idct_add:
     strd   r0, [r2]             @ r0, r1
     add    r2, r2, r3
     cmp    r2, r12
-    ldrlod r0, [r2]             @ r0, r1
+    ldrdlo r0, [r2]             @ r0, r1
     blo    4b
 
     ldmfd  sp!, {r4, pc}
diff --git a/apps/plugins/mpegplayer/mpeg_misc.h b/apps/plugins/mpegplayer/mpeg_misc.h
index e04db0e19d..68ee8cac3c 100644
--- a/apps/plugins/mpegplayer/mpeg_misc.h
+++ b/apps/plugins/mpegplayer/mpeg_misc.h
@@ -53,12 +53,13 @@ enum state_enum
 #define CMP_3_CONST(_a, _b) \
     ({  int _x;                             \
         asm volatile (                      \
+            ".syntax unified          \n"   \
             "ldrb   %[x], [%[a], #0]  \n"   \
             "eors   %[x], %[x], %[b0] \n"   \
-            "ldreqb %[x], [%[a], #1]  \n"   \
-            "eoreqs %[x], %[x], %[b1] \n"   \
-            "ldreqb %[x], [%[a], #2]  \n"   \
-            "eoreqs %[x], %[x], %[b2] \n"   \
+            "ldrbeq %[x], [%[a], #1]  \n"   \
+            "eorseq %[x], %[x], %[b1] \n"   \
+            "ldrbeq %[x], [%[a], #2]  \n"   \
+            "eorseq %[x], %[x], %[b2] \n"   \
             : [x]"=&r"(_x)                  \
             : [a]"r"(_a),                   \
               [b0]"i"(((_b) >> 24) & 0xff), \
@@ -70,14 +71,15 @@ enum state_enum
 #define CMP_4_CONST(_a, _b) \
     ({  int _x;                             \
         asm volatile (                      \
+            ".syntax unified          \n"   \
             "ldrb   %[x], [%[a], #0]  \n"   \
             "eors   %[x], %[x], %[b0] \n"   \
-            "ldreqb %[x], [%[a], #1]  \n"   \
-            "eoreqs %[x], %[x], %[b1] \n"   \
-            "ldreqb %[x], [%[a], #2]  \n"   \
-            "eoreqs %[x], %[x], %[b2] \n"   \
-            "ldreqb %[x], [%[a], #3]  \n"   \
-            "eoreqs %[x], %[x], %[b3] \n"   \
+            "ldrbeq %[x], [%[a], #1]  \n"   \
+            "eorseq %[x], %[x], %[b1] \n"   \
+            "ldrbeq %[x], [%[a], #2]  \n"   \
+            "eorseq %[x], %[x], %[b2] \n"   \
+            "ldrbeq %[x], [%[a], #3]  \n"   \
+            "eorseq %[x], %[x], %[b3] \n"   \
             : [x]"=&r"(_x)                  \
             : [a]"r"(_a),                   \
               [b0]"i"(((_b) >> 24) & 0xff), \
diff --git a/apps/recorder/jpeg_idct_arm.S b/apps/recorder/jpeg_idct_arm.S
index e7eb4b87f1..1f2603da1b 100644
--- a/apps/recorder/jpeg_idct_arm.S
+++ b/apps/recorder/jpeg_idct_arm.S
@@ -410,7 +410,7 @@ jpeg_idct8v:
 #if ARM_ARCH < 5
     mov    r8,  r4,  lsl #16
     orrs   r9,  r6,  r7
-    orreqs r9,  r5,  r4, lsr #16
+    orrseq r9,  r5,  r4, lsr #16
     bne    2f
     mov    r8,  r8,  asr #14
     strh   r8,  [r2]
@@ -505,7 +505,7 @@ jpeg_idct8v:
 #else /* ARMv5+ */
     mov    r12, r4,  lsl #16
     orrs   r9,  r6,  r7
-    orreqs r9,  r5,  r4,  lsr #16
+    orrseq r9,  r5,  r4,  lsr #16
     bne    2f
     mov    r12, r12, asr #14
     strh   r12, [r2]
@@ -615,7 +615,7 @@ jpeg_idct8h:
 #if ARM_ARCH < 5
     add    r8,  r14, r4,  lsl #16
     orrs   r9,  r6,  r7
-    orreqs r9,  r5,  r4, lsr #16
+    orrseq r9,  r5,  r4, lsr #16
     bne    2f
     mov    r8,  r8,  asr #21
     cmp    r8,  #255
@@ -727,7 +727,7 @@ jpeg_idct8h:
 #else /* ARMv5+ */
     add    r12, r14, r4,  lsl #16
     orrs   r9,  r6,  r7
-    orreqs r9,  r5,  r4,  lsr #16
+    orrseq r9,  r5,  r4,  lsr #16
     bne    2f
     mov    r12, r12, asr #21
     cmp    r12, #255
@@ -835,7 +835,7 @@ jpeg_idct8v:
 1:
     ldmia  r0!, { r4-r7 }
     orrs   r9,  r6,  r7
-    orreqs r9,  r5,  r4,  lsr #16
+    orrseq r9,  r5,  r4,  lsr #16
     bne    2f
     mov    r4,  r4,  lsl #2
     strh   r4,  [r2]
@@ -939,7 +939,7 @@ jpeg_idct8h:
     ldmia  r0!, { r4-r7 }
     sadd16 r4,  r4,  r14
     orrs   r9,  r6,  r7
-    orreqs r9,  r5,  r4,  lsr #16
+    orrseq r9,  r5,  r4,  lsr #16
     bne    2f
     sxth   r4,  r4
     usat   r4,  #8,  r4,  asr #5
diff --git a/firmware/asm/arm/corelock.c b/firmware/asm/arm/corelock.c
index b36a40b45b..07ec77a60e 100644
--- a/firmware/asm/arm/corelock.c
+++ b/firmware/asm/arm/corelock.c
@@ -61,6 +61,7 @@ int corelock_try_lock(struct corelock *cl)
 
     /* Relies on the fact that core IDs are complementary bitmasks (0x55,0xaa) */
     asm volatile (
+        ".syntax unified             \n"
         "mov    r1, %[id]            \n" /* r1 = PROCESSOR_ID */
         "ldrb   r1, [r1]             \n"
         "strb   r1, [%[cl], r1, lsr #7] \n" /* cl->myl[core] = core */
@@ -71,7 +72,7 @@ int corelock_try_lock(struct corelock *cl)
         "bne    1f                   \n" /* yes? lock acquired */
         "ldrb   %[rv], [%[cl], #2]   \n" /* || cl->turn == core? */
         "ands   %[rv], %[rv], r1     \n"
-        "streqb %[rv], [%[cl], r1, lsr #7] \n" /* if not, cl->myl[core] = 0 */
+        "strbeq %[rv], [%[cl], r1, lsr #7] \n" /* if not, cl->myl[core] = 0 */
     "1:                              \n" /* Done */
         : [rv] "=r"(rval)
         : [id] "i" (&PROCESSOR_ID), [cl] "r" (cl)
diff --git a/firmware/asm/arm/lcd-as-memframe.S b/firmware/asm/arm/lcd-as-memframe.S
index 52ab0447c2..d42b2a920d 100644
--- a/firmware/asm/arm/lcd-as-memframe.S
+++ b/firmware/asm/arm/lcd-as-memframe.S
@@ -91,9 +91,9 @@ lcd_copy_buffer_rect:                   @
     stmia   r0!, { r6-r12, r14 }        @
     bgt     30b @ octword loop          @
 40: @ finish line                       @
-    ldreqh  r6, [r1], #2                @ finish last halfword if eq ...
+    ldrheq  r6, [r1], #2                @ finish last halfword if eq ...
     add     r1, r1, r4, lsl #1          @
-    streqh  r6, [r0], #2                @ ...
+    strheq  r6, [r0], #2                @ ...
     add     r0, r0, r4, lsl #1          @
     subs    r3, r3, #1                  @ next line
     bgt     10b @ copy line             @
diff --git a/firmware/asm/arm/memcpy.S b/firmware/asm/arm/memcpy.S
index 83d43293e6..86fc6b7930 100644
--- a/firmware/asm/arm/memcpy.S
+++ b/firmware/asm/arm/memcpy.S
@@ -99,22 +99,22 @@ memcpy:
 7:      ldmfd   sp!, {r5 - r8}
 
 8:      movs    r2, r2, lsl #31
-        ldrneb  r3, [r1], #1
-        ldrcsb  r4, [r1], #1
-        ldrcsb  ip, [r1]
-        strneb  r3, [r0], #1
-        strcsb  r4, [r0], #1
-        strcsb  ip, [r0]
+        ldrbne  r3, [r1], #1
+        ldrbcs  r4, [r1], #1
+        ldrbcs  ip, [r1]
+        strbne  r3, [r0], #1
+        strbcs  r4, [r0], #1
+        strbcs  ip, [r0]
 
         ldmpc   regs="r0, r4"
 
 9:      rsb ip, ip, #4
         cmp ip, #2
-        ldrgtb  r3, [r1], #1
-        ldrgeb  r4, [r1], #1
+        ldrbgt  r3, [r1], #1
+        ldrbge  r4, [r1], #1
         ldrb    lr, [r1], #1
-        strgtb  r3, [r0], #1
-        strgeb  r4, [r0], #1
+        strbgt  r3, [r0], #1
+        strbge  r4, [r0], #1
         subs    r2, r2, ip
         strb    lr, [r0], #1
         blt 8b
diff --git a/firmware/asm/arm/memmove.S b/firmware/asm/arm/memmove.S
index d8cab048be..e5c9b42928 100644
--- a/firmware/asm/arm/memmove.S
+++ b/firmware/asm/arm/memmove.S
@@ -106,20 +106,20 @@ memmove:
 7:      ldmfd   sp!, {r5 - r8}
 
 8:      movs    r2, r2, lsl #31
-        ldrneb  r3, [r1, #-1]!
-        ldrcsb  r4, [r1, #-1]!
-        ldrcsb  ip, [r1, #-1]
-        strneb  r3, [r0, #-1]!
-        strcsb  r4, [r0, #-1]!
-        strcsb  ip, [r0, #-1]
+        ldrbne  r3, [r1, #-1]!
+        ldrbcs  r4, [r1, #-1]!
+        ldrbcs  ip, [r1, #-1]
+        strbne  r3, [r0, #-1]!
+        strbcs  r4, [r0, #-1]!
+        strbcs  ip, [r0, #-1]
         ldmpc   regs="r0, r4"
 
 9:      cmp ip, #2
-        ldrgtb  r3, [r1, #-1]!
-        ldrgeb  r4, [r1, #-1]!
+        ldrbgt  r3, [r1, #-1]!
+        ldrbge  r4, [r1, #-1]!
         ldrb    lr, [r1, #-1]!
-        strgtb  r3, [r0, #-1]!
-        strgeb  r4, [r0, #-1]!
+        strbgt  r3, [r0, #-1]!
+        strbge  r4, [r0, #-1]!
         subs    r2, r2, ip
         strb    lr, [r0, #-1]!
         blt 8b
diff --git a/firmware/asm/arm/memset.S b/firmware/asm/arm/memset.S
index 64cd95cc9e..d727f2a5ec 100644
--- a/firmware/asm/arm/memset.S
+++ b/firmware/asm/arm/memset.S
@@ -34,8 +34,8 @@
 1:      cmp     r2, #4                  @ 1 do we have enough
         blt     5f                      @ 1 bytes to align with?
         cmp     r3, #2                  @ 1
-        strgtb  r1, [r0, #-1]!          @ 1
-        strgeb  r1, [r0, #-1]!          @ 1
+        strbgt  r1, [r0, #-1]!          @ 1
+        strbge  r1, [r0, #-1]!          @ 1
         strb    r1, [r0, #-1]!          @ 1
         sub     r2, r2, r3              @ 1 r2 = r2 - r3
         b 2f
@@ -65,24 +65,24 @@ memset:
         mov     lr, r1
 
 3:      subs    r2, r2, #64
-        stmgedb r0!, {r1, r3, ip, lr}   @ 64 bytes at a time.
-        stmgedb r0!, {r1, r3, ip, lr}
-        stmgedb r0!, {r1, r3, ip, lr}
-        stmgedb r0!, {r1, r3, ip, lr}
+        stmdbge r0!, {r1, r3, ip, lr}   @ 64 bytes at a time.
+        stmdbge r0!, {r1, r3, ip, lr}
+        stmdbge r0!, {r1, r3, ip, lr}
+        stmdbge r0!, {r1, r3, ip, lr}
         bgt     3b
         ldrpc   cond=eq                 @ Now <64 bytes to go.
 /*
  * No need to correct the count; we're only testing bits from now on
  */
         tst     r2, #32
-        stmnedb r0!, {r1, r3, ip, lr}
-        stmnedb r0!, {r1, r3, ip, lr}
+        stmdbne r0!, {r1, r3, ip, lr}
+        stmdbne r0!, {r1, r3, ip, lr}
         tst     r2, #16
-        stmnedb r0!, {r1, r3, ip, lr}
+        stmdbne r0!, {r1, r3, ip, lr}
         ldr     lr, [sp], #4
 
 5:      tst     r2, #8
-        stmnedb r0!, {r1, r3}
+        stmdbne r0!, {r1, r3}
         tst     r2, #4
         strne   r1, [r0, #-4]!
 /*
@@ -90,10 +90,10 @@ memset:
  * may have an unaligned pointer as well.
  */
 6:      tst     r2, #2
-        strneb  r1, [r0, #-1]!
-        strneb  r1, [r0, #-1]!
+        strbne  r1, [r0, #-1]!
+        strbne  r1, [r0, #-1]!
         tst     r2, #1
-        strneb  r1, [r0, #-1]!
+        strbne  r1, [r0, #-1]!
         bx      lr
 .end:
         .size   memset,.end-memset
diff --git a/firmware/asm/arm/memset16.S b/firmware/asm/arm/memset16.S
index 5c787b1bed..226eac39e1 100644
--- a/firmware/asm/arm/memset16.S
+++ b/firmware/asm/arm/memset16.S
@@ -35,7 +35,7 @@
 memset16:
         tst     r0, #2                  @ unaligned?
         cmpne   r2, #0
-        strneh  r1, [r0], #2            @ store one halfword to align
+        strhne  r1, [r0], #2            @ store one halfword to align
         subne   r2, r2, #1
 
 /*
@@ -54,29 +54,29 @@ memset16:
         mov     lr, r1
 
 2:      subs    r2, r2, #32
-        stmgeia r0!, {r1, r3, ip, lr}   @ 64 bytes at a time.
-        stmgeia r0!, {r1, r3, ip, lr}
-        stmgeia r0!, {r1, r3, ip, lr}
-        stmgeia r0!, {r1, r3, ip, lr}
+        stmiage r0!, {r1, r3, ip, lr}   @ 64 bytes at a time.
+        stmiage r0!, {r1, r3, ip, lr}
+        stmiage r0!, {r1, r3, ip, lr}
+        stmiage r0!, {r1, r3, ip, lr}
         bgt     2b
         ldrpc   cond=eq                 @ Now <64 bytes to go.
 /*
  * No need to correct the count; we're only testing bits from now on
  */
         tst     r2, #16
-        stmneia r0!, {r1, r3, ip, lr}
-        stmneia r0!, {r1, r3, ip, lr}
+        stmiane r0!, {r1, r3, ip, lr}
+        stmiane r0!, {r1, r3, ip, lr}
         tst     r2, #8
-        stmneia r0!, {r1, r3, ip, lr}
+        stmiane r0!, {r1, r3, ip, lr}
         ldr     lr, [sp], #4
 
 4:      tst     r2, #4
-        stmneia r0!, {r1, r3}
+        stmiane r0!, {r1, r3}
         tst     r2, #2
         strne   r1, [r0], #4
 
         tst     r2, #1
-        strneh  r1, [r0], #2
+        strhne  r1, [r0], #2
         bx      lr
 .end:
         .size   memset16,.end-memset16
diff --git a/firmware/asm/arm/thread.c b/firmware/asm/arm/thread.c
index cf685526e3..bd9f950616 100644
--- a/firmware/asm/arm/thread.c
+++ b/firmware/asm/arm/thread.c
@@ -73,15 +73,16 @@ static inline void store_context(void* addr)
 static inline void load_context(const void* addr)
 {
     asm volatile(
+        ".syntax unified                \n"
         "ldr     r0, [%0, #40]          \n" /* Load start pointer */
         "cmp     r0, #0                 \n" /* Check for NULL */
 
         /* If not already running, jump to start */
 #if ARM_ARCH == 4 && defined(USE_THUMB)
-        "ldmneia %0, { r0, r12 }        \n"
+        "ldmiane %0, { r0, r12 }        \n"
         "bxne    r12                    \n"
 #else
-        "ldmneia %0, { r0, pc }         \n"
+        "ldmiane %0, { r0, pc }         \n"
 #endif
 
         "ldmia   %0, { r4-r11, sp, lr } \n" /* Load regs r4 to r14 from context */
diff --git a/firmware/export/config.h b/firmware/export/config.h
index 0882cad61c..19ee03b4c7 100644
--- a/firmware/export/config.h
+++ b/firmware/export/config.h
@@ -1006,13 +1006,14 @@ Lyre prototype 1 */
 #endif
 
 #if defined(CPU_ARM) && defined(__ASSEMBLER__)
+.syntax unified
 /* ARMv4T doesn't switch the T bit when popping pc directly, we must use BX */
 .macro ldmpc cond="", order="ia", regs
 #if ARM_ARCH == 4 && defined(USE_THUMB)
-    ldm\cond\order sp!, { \regs, lr }
+    ldm\order\cond sp!, { \regs, lr }
     bx\cond lr
 #else
-    ldm\cond\order sp!, { \regs, pc }
+    ldm\order\cond sp!, { \regs, pc }
 #endif
 .endm
 .macro ldrpc cond=""
diff --git a/firmware/target/arm/ata-as-arm.S b/firmware/target/arm/ata-as-arm.S
index 101bc4dcc1..16c2928bf1 100644
--- a/firmware/target/arm/ata-as-arm.S
+++ b/firmware/target/arm/ata-as-arm.S
@@ -139,9 +139,9 @@ copy_read_sectors:
 .r_end2_u:
 
     tst     r1, #1              /* one halfword left? */
-    ldrneh  r4, [r2]
+    ldrhne  r4, [r2]
     orrne   r3, r3, r4, lsl #8
-    strneh  r3, [r0], #2
+    strhne  r3, [r0], #2
     movne   r3, r4, lsr #8
 
     strb    r3, [r0], #1        /* store final byte */
@@ -151,8 +151,8 @@ copy_read_sectors:
     /* 16-bit aligned */
 .r_aligned:
     tst     r0, #2              /* 32 bit aligned? */
-    ldrneh  r3, [r2]            /* no: read first halfword */
-    strneh  r3, [r0], #2        /*     store */
+    ldrhne  r3, [r2]            /* no: read first halfword */
+    strhne  r3, [r0], #2        /*     store */
     subne   r1, r1, #1          /*     one halfword taken */
 
     sub     r1, r1, #8          /* adjust for zero-check and doing 8 halfwords/loop */
@@ -186,14 +186,14 @@ copy_read_sectors:
 .r_end4_a:
 
     tst     r1, #2              /* 2 or more halfwords left? */
-    ldrneh  r3, [r2]
-    ldrneh  r4, [r2]
+    ldrhne  r3, [r2]
+    ldrhne  r4, [r2]
     orrne   r3, r3, r4, lsl #16
     strne   r3, [r0], #4
 
     tst     r1, #1              /* one halfword left? */
-    ldrneh  r3, [r2]
-    strneh  r3, [r0], #2
+    ldrhne  r3, [r2]
+    strhne  r3, [r0], #2
 
     ldmpc   regs=r4-r5
 
@@ -291,9 +291,9 @@ copy_write_sectors:
 .w_end2_u:
 
     tst     r1, #1              /* one halfword left? */
-    ldrneh  r4, [r0], #2
+    ldrhne  r4, [r0], #2
     orrne   r3, r3, r4, lsl #8
-    strneh  r3, [r2]
+    strhne  r3, [r2]
     movne   r3, r3, lsr #16
     
     ldrb    r4, [r0], #1        /* load final byte */
@@ -305,8 +305,8 @@ copy_write_sectors:
     /* 16-bit aligned */
 .w_aligned:
     tst     r0, #2              /* 32 bit aligned? */
-    ldrneh  r3, [r0], #2        /* no: load first halfword */
-    strneh  r3, [r2]            /*     write */
+    ldrhne  r3, [r0], #2        /* no: load first halfword */
+    strhne  r3, [r2]            /*     write */
     subne   r1, r1, #1          /*     one halfword taken */
 
     sub     r1, r1, #8          /* adjust for zero-check and doing 8 halfwords/loop */
@@ -341,13 +341,13 @@ copy_write_sectors:
 
     tst     r1, #2              /* 2 or more halfwords left? */
     ldrne   r3, [r0], #4
-    strneh  r3, [r2]
+    strhne  r3, [r2]
     movne   r3, r3, lsr #16
-    strneh  r3, [r2]
+    strhne  r3, [r2]
 
     tst     r1, #1              /* one halfword left? */
-    ldrneh  r3, [r0], #2
-    strneh  r3, [r2]
+    ldrhne  r3, [r0], #2
+    strhne  r3, [r2]
 
     ldmpc   regs=r4-r5
 
diff --git a/firmware/target/arm/ipod/video/lcd-as-video.S b/firmware/target/arm/ipod/video/lcd-as-video.S
index 47155b8c75..7d6caef448 100644
--- a/firmware/target/arm/ipod/video/lcd-as-video.S
+++ b/firmware/target/arm/ipod/video/lcd-as-video.S
@@ -40,24 +40,24 @@ lcd_write_data:                       /* r1 = pixel count, must be even */
 
     subs    r1, r1, #16
 .loop16:
-    ldmgeia r0!, {r2-r3}
-    stmgeia lr,  {r2-r3}
-    ldmgeia r0!, {r2-r3}
-    stmgeia lr,  {r2-r3}
-    ldmgeia r0!, {r2-r3}
-    stmgeia lr,  {r2-r3}
-    ldmgeia r0!, {r2-r3}
-    stmgeia lr,  {r2-r3}
-    subges  r1, r1, #16
+    ldmiage r0!, {r2-r3}
+    stmiage lr,  {r2-r3}
+    ldmiage r0!, {r2-r3}
+    stmiage lr,  {r2-r3}
+    ldmiage r0!, {r2-r3}
+    stmiage lr,  {r2-r3}
+    ldmiage r0!, {r2-r3}
+    stmiage lr,  {r2-r3}
+    subsge  r1, r1, #16
     bge     .loop16
 
     /* no need to correct the count, we're just checking bits from now */
     tst     r1, #8
-    ldmneia r0!, {r2-r4, r12}
-    stmneia lr,  {r2-r4, r12}
+    ldmiane r0!, {r2-r4, r12}
+    stmiane lr,  {r2-r4, r12}
     tst     r1, #4
-    ldmneia r0!, {r2-r3}
-    stmneia lr,  {r2-r3}
+    ldmiane r0!, {r2-r3}
+    stmiane lr,  {r2-r3}
     tst     r1, #2
     ldrne   r3, [r0], #4
     strne   r3, [lr]
diff --git a/firmware/target/arm/pcm-telechips.c b/firmware/target/arm/pcm-telechips.c
index 336b5626ca..747765d8fb 100644
--- a/firmware/target/arm/pcm-telechips.c
+++ b/firmware/target/arm/pcm-telechips.c
@@ -218,6 +218,7 @@ void fiq_handler(void)
      * r0-r3 and r12 is a working register.
      */
     asm volatile (
+        ".syntax unified             \n"
         "sub     lr, lr, #4          \n"
         "stmfd   sp!, { r0-r3, lr }  \n" /* stack scratch regs and lr */
         "mov     r14, #0             \n" /* Was the callback called? */
@@ -251,7 +252,7 @@ void fiq_handler(void)
         "stmia   r11, { r8-r9 }      \n" /* save p and size */
 
         "cmp     r14, #0             \n" /* Callback called? */
-        "ldmeqfd sp!, { r0-r3, pc }^ \n" /* no? -> exit */
+        "ldmfdeq sp!, { r0-r3, pc }^ \n" /* no? -> exit */
 
         "ldr     r1, =pcm_play_status_callback \n"
         "ldr     r1, [r1]            \n"
@@ -268,7 +269,7 @@ void fiq_handler(void)
         "mov     lr, pc              \n"
         "ldr     pc, =pcm_play_dma_complete_callback \n"
         "cmp     r0, #0              \n" /* any more to play? */
-        "ldmneia r11, { r8-r9 }      \n" /* load new p and size */
+        "ldmiane r11, { r8-r9 }      \n" /* load new p and size */
         "cmpne   r9, #0x0f           \n" /* did we actually get enough data? */
         "bhi     .fill_fifo          \n" /* not stop and enough? refill */
         "ldmfd   sp!, { r0-r3, pc }^ \n" /* exit */
diff --git a/firmware/target/arm/pp/pcm-pp.c b/firmware/target/arm/pp/pcm-pp.c
index 0d61eb44ff..fd798f0506 100644
--- a/firmware/target/arm/pp/pcm-pp.c
+++ b/firmware/target/arm/pp/pcm-pp.c
@@ -327,6 +327,7 @@ void fiq_playback(void)
      */
     asm volatile (
     /* No external calls */
+        ".syntax unified              \n"
         "sub     lr, lr, #4           \n" /* Prepare return address */
         "stmfd   sp!, { lr }          \n" /* stack lr so we can use it */
         "ldr     r12, =0xcf001040     \n" /* Some magic from iPodLinux ... */
@@ -349,8 +350,8 @@ void fiq_playback(void)
         "bhi     0b                   \n" /* ... yes, continue */
 
         "cmp     r9, #0               \n" /* either FIFO full or size empty? */
-        "stmneia r11, { r8-r9 }       \n" /* save p and size, if not empty */
-        "ldmnefd sp!, { pc }^         \n" /* RFE if not empty */
+        "stmiane r11, { r8-r9 }       \n" /* save p and size, if not empty */
+        "ldmfdne sp!, { pc }^         \n" /* RFE if not empty */
 
     /* Making external calls */
     "1:                               \n"
@@ -363,7 +364,7 @@ void fiq_playback(void)
         "mov     lr, pc               \n" /* long call (not in same section) */
         "bx      r3                   \n"
         "cmp     r0, #0               \n" /* more data? */
-        "ldmeqfd sp!, { r0-r3, pc }^  \n" /* no? -> exit */
+        "ldmfdeq sp!, { r0-r3, pc }^  \n" /* no? -> exit */
 
         "ldr     r14, [r10, #0x1c]    \n" /* read IISFIFO_CFG to check FIFO status */
         "ands    r14, r14, #(0xe<<23) \n" /* r14 = (IIS_TX_FREE_COUNT & ~1) << 23 */
diff --git a/lib/arm_support/support-arm.S b/lib/arm_support/support-arm.S
index f99d086b0b..3af0c70975 100644
--- a/lib/arm_support/support-arm.S
+++ b/lib/arm_support/support-arm.S
@@ -227,7 +227,7 @@
     /* Test whether divisor is 2^N */
     cmp     \inv, #1<<31
     /* Load approximate reciprocal */
-    ldrhib  \inv, [\neg, #.L_udiv_est_table-.-64]
+    ldrbhi  \inv, [\neg, #.L_udiv_est_table-.-64]
     bls     20f
     subs    \bits, \bits, #7
     rsb     \neg, \divisor, #0
diff --git a/lib/rbcodec/codecs/demac/libdemac/udiv32_arm.S b/lib/rbcodec/codecs/demac/libdemac/udiv32_arm.S
index 7b851659bd..1d19160a91 100644
--- a/lib/rbcodec/codecs/demac/libdemac/udiv32_arm.S
+++ b/lib/rbcodec/codecs/demac/libdemac/udiv32_arm.S
@@ -225,7 +225,7 @@ udiv32_arm:
     mov     \inv, \divisor, lsl \bits
     add     \neg, pc, \inv, lsr #25
     cmp     \inv, #1<<31
-    ldrhib  \inv, [\neg, #.L_udiv_est_table-.-64]
+    ldrbhi  \inv, [\neg, #.L_udiv_est_table-.-64]
     bls     20f
     subs    \bits, \bits, #7
     rsb     \neg, \divisor, #0
diff --git a/lib/rbcodec/codecs/demac/libdemac/vector_math16_armv6.h b/lib/rbcodec/codecs/demac/libdemac/vector_math16_armv6.h
index 8d27331b62..1da090efbb 100644
--- a/lib/rbcodec/codecs/demac/libdemac/vector_math16_armv6.h
+++ b/lib/rbcodec/codecs/demac/libdemac/vector_math16_armv6.h
@@ -45,6 +45,7 @@ static inline int32_t vector_sp_add(int16_t* v1, int16_t* f2, int16_t* s2)
 #endif
 
     asm volatile (
+        ".syntax unified                 \n"
 #if ORDER > 32
         "mov     %[res], #0              \n"
 #endif
@@ -117,7 +118,7 @@ static inline int32_t vector_sp_add(int16_t* v1, int16_t* f2, int16_t* s2)
         "smladx  %[res], r1, r2, %[res]  \n"
 #if ORDER > 32
         "subs    %[cnt], %[cnt], #1      \n"
-        "ldmneia %[f2]!, {r2,r4}         \n"
+        "ldmiane %[f2]!, {r2,r4}         \n"
         "sadd16  r0, r0, r7              \n"
         "sadd16  r1, r1, r5              \n"
         "strd    r0, [%[v1]], #8         \n"
@@ -172,8 +173,8 @@ static inline int32_t vector_sp_add(int16_t* v1, int16_t* f2, int16_t* s2)
         "smlad   %[res], r3, r5, %[res]  \n"
 #if ORDER > 32
         "subs    %[cnt], %[cnt], #1      \n"
-        "ldrned  r4, [%[f2]], #8         \n"
-        "ldrned  r0, [%[v1], #8]         \n"
+        "ldrdne  r4, [%[f2]], #8         \n"
+        "ldrdne  r0, [%[v1], #8]         \n"
         "sadd16  r2, r2, r6              \n"
         "sadd16  r3, r3, r7              \n"
         "strd    r2, [%[v1]], #8         \n"
@@ -214,6 +215,7 @@ static inline int32_t vector_sp_sub(int16_t* v1, int16_t* f2, int16_t* s2)
 #endif
 
     asm volatile (
+        ".syntax unified                 \n"
 #if ORDER > 32
         "mov     %[res], #0              \n"
 #endif
@@ -286,7 +288,7 @@ static inline int32_t vector_sp_sub(int16_t* v1, int16_t* f2, int16_t* s2)
         "smladx  %[res], r1, r2, %[res]  \n"
 #if ORDER > 32
         "subs    %[cnt], %[cnt], #1      \n"
-        "ldmneia %[f2]!, {r2,r4}         \n"
+        "ldmiane %[f2]!, {r2,r4}         \n"
         "ssub16  r0, r0, r7              \n"
         "ssub16  r1, r1, r5              \n"
         "strd    r0, [%[v1]], #8         \n"
@@ -341,8 +343,8 @@ static inline int32_t vector_sp_sub(int16_t* v1, int16_t* f2, int16_t* s2)
         "smlad   %[res], r3, r5, %[res]  \n"
 #if ORDER > 32
         "subs    %[cnt], %[cnt], #1      \n"
-        "ldrned  r4, [%[f2]], #8         \n"
-        "ldrned  r0, [%[v1], #8]         \n"
+        "ldrdne  r4, [%[f2]], #8         \n"
+        "ldrdne  r0, [%[v1], #8]         \n"
         "ssub16  r2, r2, r6              \n"
         "ssub16  r3, r3, r7              \n"
         "strd    r2, [%[v1]], #8         \n"
@@ -381,6 +383,7 @@ static inline int32_t scalarproduct(int16_t* v1, int16_t* v2)
 #endif
 
     asm volatile (
+        ".syntax unified                 \n"
 #if ORDER > 32
         "mov     %[res], #0              \n"
 #endif
@@ -421,10 +424,10 @@ static inline int32_t scalarproduct(int16_t* v1, int16_t* v2)
         "pkhtb   r1, r7, r4              \n"
 #if ORDER > 32
         "subs    %[cnt], %[cnt], #1      \n"
-        "ldrned  r6, [%[v2]], #8         \n"
+        "ldrdne  r6, [%[v2]], #8         \n"
         "smladx  %[res], r2, r1, %[res]  \n"
         "pkhtb   r2, r4, r5              \n"
-        "ldrned  r0, [%[v1]], #8         \n"
+        "ldrdne  r0, [%[v1]], #8         \n"
         "smladx  %[res], r3, r2, %[res]  \n"
         "bne     1b                      \n"
 #else
@@ -461,9 +464,9 @@ static inline int32_t scalarproduct(int16_t* v1, int16_t* v2)
         "ldrd    r4, [%[v2]], #8         \n"
         "smlad   %[res], r1, r6, %[res]  \n"
         "subs    %[cnt], %[cnt], #1      \n"
-        "ldrned  r0, [%[v1]], #8         \n"
+        "ldrdne  r0, [%[v1]], #8         \n"
         "smlad   %[res], r2, r7, %[res]  \n"
-        "ldrned  r6, [%[v2]], #8         \n"
+        "ldrdne  r6, [%[v2]], #8         \n"
         "smlad   %[res], r3, r4, %[res]  \n"
         "bne     1b                      \n"
 #else
diff --git a/lib/rbcodec/codecs/libtta/filter_arm.S b/lib/rbcodec/codecs/libtta/filter_arm.S
index f3959b83ca..10f1491796 100644
--- a/lib/rbcodec/codecs/libtta/filter_arm.S
+++ b/lib/rbcodec/codecs/libtta/filter_arm.S
@@ -165,8 +165,8 @@ hybrid_filter:
     sub      r10, r11, r10
 
     @ set to the memory: *pA, *(pA-1), *(pA-2), *(pA-3), *pM, *(pM-1), *(pM-2), *(pM-3)
-    stmneda  r2,  {r10, r11, r12, lr}
-    stmneda  r3,  {r5,  r6,  r7,  r8}
+    stmdane  r2,  {r10, r11, r12, lr}
+    stmdane  r3,  {r5,  r6,  r7,  r8}
     ldmpc    cond=ne regs=r4-r12     @ hybrid_filter end (when fs->index != 0)
 
 .hf_memshl:
diff --git a/lib/rbcodec/dsp/dsp_arm.S b/lib/rbcodec/dsp/dsp_arm.S
index 16394b8690..864abee4b6 100644
--- a/lib/rbcodec/dsp/dsp_arm.S
+++ b/lib/rbcodec/dsp/dsp_arm.S
@@ -323,7 +323,7 @@ resample_hermite:
     add     r6, r6, r0, lsl #2          @ r6 = &s[pos]
 
     cmp     r0, #3                      @ pos >= 3? history not needed
-    ldmgedb r6, { r1-r3 }               @ x3..x1 = s[pos-3]..s[pos-1]
+    ldmdbge r6, { r1-r3 }               @ x3..x1 = s[pos-3]..s[pos-1]
     bge     .hrs_loadhist_done          @
     add     r10, r0, r0, lsl #1         @ branch pc + pos*12
     add     pc, pc, r10, lsl #2         @
@@ -496,7 +496,7 @@ resample_hermite:
     ldmfd   sp!, { r10, r12 }           @ recover ch, h
     subs    r10, r10, #1                @ --ch
     stmia   r12!, { r1-r3 }             @ h[0..2] = x3..x1
-    ldmgtia sp, { r0-r2 }               @ load data, src, dst
+    ldmiagt sp, { r0-r2 }               @ load data, src, dst
     bgt     .hrs_channel_loop
 
     ldmfd   sp!, { r1-r3 }              @ pop data, src, dst
@@ -614,7 +614,7 @@ filter_process:
     ldr     r0, [sp]                @ r0 = history[channels-ch-1]
     subs    r3, r3, #1              @ all channels processed?
     stmia   r0!, { r9-r12 }         @ save back history, history++
-    ldmhsib sp, { r1-r2 }           @ r1 = buf, r2 = count
+    ldmibhs sp, { r1-r2 }           @ r1 = buf, r2 = count
     strhs   r3, [sp, #12]           @ store ch
     strhs   r0, [sp]                @ store history[channels-ch-1]
     bhs     .fp_channelloop
diff --git a/lib/rbcodec/dsp/dsp_arm_v6.S b/lib/rbcodec/dsp/dsp_arm_v6.S
index a36760f744..aa27ec90f6 100644
--- a/lib/rbcodec/dsp/dsp_arm_v6.S
+++ b/lib/rbcodec/dsp/dsp_arm_v6.S
@@ -18,6 +18,7 @@
  * KIND, either express or implied.
  *
  ****************************************************************************/
+ #include "rbcodecconfig.h"
 
 /****************************************************************************
  *  void sample_output_mono(struct sample_io_data *this,
@@ -56,7 +57,7 @@ sample_output_mono:
     stmia   r3!, { r12, r14 }          @ store So0, So1
     bgt     1b                         @
                                        @
-    ldmltfd sp!, { r4, pc }            @ if count was even, we're done
+    ldmfdlt sp!, { r4, pc }            @ if count was even, we're done
                                        @
 2:                                     @
     ldr     r12, [r2]                  @ round, scale, saturate
@@ -113,7 +114,7 @@ sample_output_stereo:
     stmia   r3!, { r6, r7 }            @ store So0, So1
     bgt     1b                         @
                                        @
-    ldmltfd sp!, { r4-r7, pc }         @ if count was even, we're done
+    ldmfdlt sp!, { r4-r7, pc }         @ if count was even, we're done
                                        @
 2:                                     @
     ldr     r6, [r2]                   @ r6 = Li
diff --git a/lib/unwarminder/safe_read.S b/lib/unwarminder/safe_read.S
index 1969c7e0c6..ce4913870c 100644
--- a/lib/unwarminder/safe_read.S
+++ b/lib/unwarminder/safe_read.S
@@ -45,7 +45,7 @@ safe_read8_faulty_addr:
     @ if(value != NULL)
     cmp     r1, #0
     @   *value = r0
-    strneb  r0, [r1]
+    strbne  r0, [r1]
     @ return true;
     mov     r0, #1
     bx      lr
@@ -72,7 +72,7 @@ safe_read16_faulty_addr:
     @ if(value != NULL)
     cmp     r1, #0
     @   *value = r0
-    strneh  r0, [r1]
+    strhne  r0, [r1]
     @ return true;
     mov     r0, #1
     bx      lr