MdePkg/BaseMemoryLibOptDxe: replace deprecated uses of IT blocks

The ARM architecture version 8 deprecates all uses of the IT instruction
except cases where it is followed by a single narrow instruction. So
replace any occurrences with equivalent sequences that adhere to the
new rules.

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Liming Gao <liming.gao@intel.com>

MdePkg/Library/BaseMemoryLibOptDxe/Arm/CompareGuid.S

@@ -42,7 +42,7 @@ ASM_PFX(InternalMemCompareGuid):
     ldr     lr, [r1, #4]
     cmp     r2, ip
     it      eq
-    cmpeq   r3, lr
+    cmpeq.n r3, lr
     beq     0f
     movs    r0, #0
     pop     {r4, pc}
@@ -51,7 +51,7 @@ ASM_PFX(InternalMemCompareGuid):
     ldr     r3, [r1, #12]
     cmp     r4, r2
     it      eq
-    cmpeq   r0, r3
+    cmpeq.n r0, r3
     bne     2f
     movs    r0, #1
     pop     {r4, pc}
@@ -61,5 +61,5 @@ ASM_PFX(InternalMemCompareGuid):
     movs    r0, #1
     orrs    r2, r2, r4
 2:  it      ne
-    movne   r0, #0
+    movne.n r0, #0
     pop     {r4, pc}

MdePkg/Library/BaseMemoryLibOptDxe/Arm/CompareMem.S

@@ -132,7 +132,7 @@ ASM_PFX(InternalMemCompareMem):
     ldrb    data2, [src2], #1
     subs    limit, limit, #1
     it      cs
-    cmpcs   data1, data2
+    cmpcs.n data1, data2
     beq     1b
     sub     result, data1, data2
     pop     {r4-r8, pc}

MdePkg/Library/BaseMemoryLibOptDxe/Arm/CopyMem.S

@@ -65,16 +65,18 @@ memcopy_check_optim_default:
     // Check if we can use an optimized path ((length >= 32) && destination word-aligned && source word-aligned) for the memcopy (optimized path if r0 == 1)
     tst     r0, #0xF
     it      ne
-    movne   r0, #0
+    movne.n r0, #0
     bne     memcopy_default
     tst     r1, #0xF
-    ite     ne
-    movne   r3, #0
-    moveq   r3, #1
+    it      ne
+    movne.n r3, #0
+    it      eq
+    moveq.n r3, #1
     cmp     r2, #31
-    ite     ls
-    movls   r0, #0
-    andhi   r0, r3, #1
+    it      ls
+    movls.n r0, #0
+    bls     memcopy_default
+    and     r0, r3, #1
     b       memcopy_default
 
 memcopy_check_optim_overlap:
@@ -84,15 +86,16 @@ memcopy_check_optim_overlap:
 
     // Are we in the optimized case ((length >= 32) && dest_end word-aligned && source_end word-aligned)
     cmp     r2, #31
-    ite     ls
-    movls   r0, #0
-    movhi   r0, #1
+    it      ls
+    movls.n r0, #0
+    it      hi
+    movhi.n r0, #1
     tst     r10, #0xF
     it      ne
-    movne   r0, #0
+    movne.n r0, #0
     tst     r14, #0xF
     it      ne
-    movne   r0, #0
+    movne.n r0, #0
     b       memcopy_overlapped
 
 memcopy_overlapped_non_optim:
@@ -123,7 +126,7 @@ memcopy_overlapped:
 
     // If length is less than 32 then disable optim
     it      ls
-    movls   r0, #0
+    movls.n r0, #0
 
     cmp     r12, #0
 
@@ -157,7 +160,7 @@ memcopy_default_loop:
 
     // If length is less than 32 then disable optim
     it      ls
-    movls   r0, #0
+    movls.n r0, #0
 
     cmp     r12, #0
 

MdePkg/Library/BaseMemoryLibOptDxe/Arm/ScanMem.S

@@ -123,11 +123,12 @@ ASM_PFX(InternalMemScanMem8):
 60: // We're here because the fast path found a hit - now we have to track down exactly which word it was
     // r0 points to the start of the double word after the one that was tested
     // r5 has the 00/ff pattern for the first word, r6 has the chained value
+    subs    r0, r0, #3
     cmp     r5, #0
-    itte    eq
-    moveq   r5, r6        // the end is in the 2nd word
-    subeq   r0, r0, #3    // Points to 2nd byte of 2nd word
-    subne   r0, r0, #7    // or 2nd byte of 1st word
+    it      eq
+    moveq.n r5, r6        // the end is in the 2nd word
+    it      ne
+    subne.n r0, r0, #4    // or 2nd byte of 1st word
 
     // r0 currently points to the 3rd byte of the word containing the hit
     tst     r5, #CHARTSTMASK(0)     // 1st character

MdePkg/Library/BaseMemoryLibOptDxe/Arm/SetMem.S

@@ -73,15 +73,17 @@ ASM_PFX(InternalMemZeroMem):
     cmp     r4, #4                  // between 4 and 15 bytes?
     blt     4f
     cmp     r4, #8                  // between 8 and 15 bytes?
-    str     r2, [lr, #-16]          // overlapping store of 4 + (4 + 4) + 4 bytes
-    itt     gt
-    strgt   r3, [lr, #-12]
-    strgt   r2, [r1]
+    sub     r4, lr, #16
+    str     r2, [r4]                // overlapping store of 4 + (4 + 4) + 4 bytes
+    it      gt
+    strgt.n r3, [r4, #4]
+    it      gt
+    strgt.n r2, [r1]
     str     r3, [r1, #4]
     pop     {r4, pc}
 
 4:  cmp     r4, #2                  // 2 or 3 bytes?
     strb    r2, [lr, #-16]          // store 1 byte
     it      ge
-    strhge  r2, [r1, #6]            // store 2 bytes
+    strhge.n r2, [r1, #6]           // store 2 bytes
     pop     {r4, pc}