Merge branch 'drm-patches' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied...

[pandora-kernel.git] / arch / mips / lib / csum_partial.S

diff --git a/arch/mips/lib/csum_partial.S b/arch/mips/lib/csum_partial.S
index 3bffdbb..9db3572 100644 (file)
--- a/arch/mips/lib/csum_partial.S
+++ b/arch/mips/lib/csum_partial.S
@@ -29,30 +29,49 @@
 #define t5     $13
 #define t6     $14
 #define t7     $15
+
+#define USE_DOUBLE
 #endif
 
+#ifdef USE_DOUBLE
+
+#define LOAD   ld
+#define ADD    daddu
+#define NBYTES 8
+
+#else
+
+#define LOAD   lw
+#define ADD    addu
+#define NBYTES 4
+
+#endif /* USE_DOUBLE */
+
+#define UNIT(unit)  ((unit)*NBYTES)
+
 #define ADDC(sum,reg)                                          \
-       addu    sum, reg;                                       \
+       ADD     sum, reg;                                       \
        sltu    v1, sum, reg;                                   \
-       addu    sum, v1
+       ADD     sum, v1
 
-#define CSUM_BIGCHUNK(src, offset, sum, _t0, _t1, _t2, _t3)    \
-       lw      _t0, (offset + 0x00)(src);                      \
-       lw      _t1, (offset + 0x04)(src);                      \
-       lw      _t2, (offset + 0x08)(src);                      \
-       lw      _t3, (offset + 0x0c)(src);                      \
+#define CSUM_BIGCHUNK1(src, offset, sum, _t0, _t1, _t2, _t3)   \
+       LOAD    _t0, (offset + UNIT(0))(src);                   \
+       LOAD    _t1, (offset + UNIT(1))(src);                   \
+       LOAD    _t2, (offset + UNIT(2))(src);                   \
+       LOAD    _t3, (offset + UNIT(3))(src);                   \
        ADDC(sum, _t0);                                         \
        ADDC(sum, _t1);                                         \
        ADDC(sum, _t2);                                         \
-       ADDC(sum, _t3);                                         \
-       lw      _t0, (offset + 0x10)(src);                      \
-       lw      _t1, (offset + 0x14)(src);                      \
-       lw      _t2, (offset + 0x18)(src);                      \
-       lw      _t3, (offset + 0x1c)(src);                      \
-       ADDC(sum, _t0);                                         \
-       ADDC(sum, _t1);                                         \
-       ADDC(sum, _t2);                                         \
-       ADDC(sum, _t3);                                         \
+       ADDC(sum, _t3)
+
+#ifdef USE_DOUBLE
+#define CSUM_BIGCHUNK(src, offset, sum, _t0, _t1, _t2, _t3)    \
+       CSUM_BIGCHUNK1(src, offset, sum, _t0, _t1, _t2, _t3)
+#else
+#define CSUM_BIGCHUNK(src, offset, sum, _t0, _t1, _t2, _t3)    \
+       CSUM_BIGCHUNK1(src, offset, sum, _t0, _t1, _t2, _t3);   \
+       CSUM_BIGCHUNK1(src, offset + 0x10, sum, _t0, _t1, _t2, _t3)
+#endif
 
 /*
  * a0: source address
@@ -65,64 +84,6 @@
 
        .text
        .set    noreorder
-
-/* unknown src alignment and < 8 bytes to go  */
-small_csumcpy:
-       move    a1, t2
-
-       andi    t0, a1, 4
-       beqz    t0, 1f
-        andi   t0, a1, 2
-
-       /* Still a full word to go  */
-       ulw     t1, (src)
-       PTR_ADDIU       src, 4
-       ADDC(sum, t1)
-
-1:     move    t1, zero
-       beqz    t0, 1f
-        andi   t0, a1, 1
-
-       /* Still a halfword to go  */
-       ulhu    t1, (src)
-       PTR_ADDIU       src, 2
-
-1:     beqz    t0, 1f
-        sll    t1, t1, 16
-
-       lbu     t2, (src)
-        nop
-
-#ifdef __MIPSEB__
-       sll     t2, t2, 8
-#endif
-       or      t1, t2
-
-1:     ADDC(sum, t1)
-
-       /* fold checksum */
-       sll     v1, sum, 16
-       addu    sum, v1
-       sltu    v1, sum, v1
-       srl     sum, sum, 16
-       addu    sum, v1
-
-       /* odd buffer alignment? */
-       beqz    t7, 1f
-        nop
-       sll     v1, sum, 8
-       srl     sum, sum, 8
-       or      sum, v1
-       andi    sum, 0xffff
-1:
-       .set    reorder
-       /* Add the passed partial csum.  */
-       ADDC(sum, a2)
-       jr      ra
-       .set    noreorder
-
-/* ------------------------------------------------------------------------- */
-
        .align  5
 LEAF(csum_partial)
        move    sum, zero
@@ -132,8 +93,7 @@ LEAF(csum_partial)
        bnez    t8, small_csumcpy               /* < 8 bytes to copy */
         move   t2, a1
 
-       beqz    a1, out
-        andi   t7, src, 0x1                    /* odd buffer? */
+       andi    t7, src, 0x1                    /* odd buffer? */
 
 hword_align:
        beqz    t7, word_align
@@ -176,11 +136,17 @@ qword_align:
        beqz    t8, oword_align
         andi   t8, src, 0x10
 
+#ifdef USE_DOUBLE
+       ld      t0, 0x00(src)
+       LONG_SUBU       a1, a1, 0x8
+       ADDC(sum, t0)
+#else
        lw      t0, 0x00(src)
        lw      t1, 0x04(src)
        LONG_SUBU       a1, a1, 0x8
        ADDC(sum, t0)
        ADDC(sum, t1)
+#endif
        PTR_ADDU        src, src, 0x8
        andi    t8, src, 0x10
 
@@ -188,14 +154,14 @@ oword_align:
        beqz    t8, begin_movement
         LONG_SRL       t8, a1, 0x7
 
-       lw      t3, 0x08(src)
-       lw      t4, 0x0c(src)
-       lw      t0, 0x00(src)
-       lw      t1, 0x04(src)
-       ADDC(sum, t3)
-       ADDC(sum, t4)
+#ifdef USE_DOUBLE
+       ld      t0, 0x00(src)
+       ld      t1, 0x08(src)
        ADDC(sum, t0)
        ADDC(sum, t1)
+#else
+       CSUM_BIGCHUNK1(src, 0x00, sum, t0, t1, t3, t4)
+#endif
        LONG_SUBU       a1, a1, 0x10
        PTR_ADDU        src, src, 0x10
        LONG_SRL        t8, a1, 0x7
@@ -232,8 +198,9 @@ move_32bytes:
        PTR_ADDU        src, src, 0x20
 
 do_end_words:
-       beqz    t8, maybe_end_cruft
-        LONG_SRL       t8, t8, 0x2
+       beqz    t8, small_csumcpy
+        andi   t2, a1, 0x3
+       LONG_SRL        t8, t8, 0x2
 
 end_words:
        lw      t0, (src)
@@ -242,21 +209,65 @@ end_words:
        bnez    t8, end_words
         PTR_ADDU       src, src, 0x4
 
-maybe_end_cruft:
-       andi    t2, a1, 0x3
+/* unknown src alignment and < 8 bytes to go  */
+small_csumcpy:
+       move    a1, t2
 
-small_memcpy:
- j small_csumcpy; move a1, t2          /* XXX ??? */
-       beqz    t2, out
-        move   a1, t2
+       andi    t0, a1, 4
+       beqz    t0, 1f
+        andi   t0, a1, 2
 
-end_bytes:
-       lb      t0, (src)
-       LONG_SUBU       a1, a1, 0x1
-       bnez    a2, end_bytes
-        PTR_ADDU       src, src, 0x1
+       /* Still a full word to go  */
+       ulw     t1, (src)
+       PTR_ADDIU       src, 4
+       ADDC(sum, t1)
+
+1:     move    t1, zero
+       beqz    t0, 1f
+        andi   t0, a1, 1
+
+       /* Still a halfword to go  */
+       ulhu    t1, (src)
+       PTR_ADDIU       src, 2
+
+1:     beqz    t0, 1f
+        sll    t1, t1, 16
+
+       lbu     t2, (src)
+        nop
+
+#ifdef __MIPSEB__
+       sll     t2, t2, 8
+#endif
+       or      t1, t2
 
-out:
+1:     ADDC(sum, t1)
+
+       /* fold checksum */
+#ifdef USE_DOUBLE
+       dsll32  v1, sum, 0
+       daddu   sum, v1
+       sltu    v1, sum, v1
+       dsra32  sum, sum, 0
+       addu    sum, v1
+#endif
+       sll     v1, sum, 16
+       addu    sum, v1
+       sltu    v1, sum, v1
+       srl     sum, sum, 16
+       addu    sum, v1
+
+       /* odd buffer alignment? */
+       beqz    t7, 1f
+        nop
+       sll     v1, sum, 8
+       srl     sum, sum, 8
+       or      sum, v1
+       andi    sum, 0xffff
+1:
+       .set    reorder
+       /* Add the passed partial csum.  */
+       ADDC(sum, a2)
        jr      ra
-        move   v0, sum
+       .set    noreorder
        END(csum_partial)