2013-01-17 Marcus Shawcroft <marcus.shawcroft@linaro.org>

	    * libc/machine/aarch64/strncmp.S: Correct arithmetic for
	    argument N values close to the maximum representable
	    value in an unsigned 64 bit value.

diff --git a/newlib/ChangeLog b/newlib/ChangeLog
index 15123c0fc0bd3bcadd139e2fc3f7b953186bc239..d65758498f3d30de3964ad5d7f31435f30339960 100644 (file)
--- a/newlib/ChangeLog
+++ b/newlib/ChangeLog
@@ -1,3 +1,9 @@
+2013-01-17  Marcus Shawcroft  <marcus.shawcroft@linaro.org>
+
+       * libc/machine/aarch64/strncmp.S: Correct arithmetic for
+       argument N values close to the maximum representable
+       value in an unsigned 64 bit value.
+
 2013-01-17  Marcus Shawcroft  <marcus.shawcroft@linaro.org>
 
        * libc/machine/aarch64/strnlen.S: Correct arithmetic for

diff --git a/newlib/libc/machine/aarch64/strncmp.S b/newlib/libc/machine/aarch64/strncmp.S
index 6c2246dfb8c0e7b4e17ce597b22e4b3481eac08a..0b90dd8a6c65dae604b6e0aaaeaae0d588a07078 100644 (file)
--- a/newlib/libc/machine/aarch64/strncmp.S
+++ b/newlib/libc/machine/aarch64/strncmp.S
@@ -81,8 +81,10 @@ def_fn strncmp
        b.ne    .Lmisaligned8
        ands    tmp1, src1, #7
        b.ne    .Lmutual_align
-       add     limit_wd, limit, #7
-       lsr     limit_wd, limit_wd, #3
+       /* Calculate the number of full and partial words -1.  */
+       sub     limit_wd, limit, #1     /* limit != 0, so no underflow.  */
+       lsr     limit_wd, limit_wd, #3  /* Convert to Dwords.  */
+
        /* NUL detection works on the principle that (X - 1) & (~X) & 0x80
           (=> (X - 1) & ~(X | 0x7f)) is non-zero iff a byte is zero, and
           can be done in parallel across the entire word.  */
@@ -95,14 +97,14 @@ def_fn strncmp
        sub     tmp1, data1, zeroones
        orr     tmp2, data1, #REP8_7f
        eor     diff, data1, data2      /* Non-zero if differences found.  */
-       csinv   endloop, diff, xzr, ne  /* Last Dword or differences.  */
+       csinv   endloop, diff, xzr, pl  /* Last Dword or differences.  */
        bics    has_nul, tmp1, tmp2     /* Non-zero if NUL terminator.  */
        ccmp    endloop, #0, #0, eq
        b.eq    .Lloop_aligned
        /* End of performance-critical section  -- one 64B cache line.  */
 
        /* Not reached the limit, must have found the end or a diff.  */
-       cbnz    limit_wd, .Lnot_limit
+       tbz     limit_wd, #63, .Lnot_limit
 
        /* Limit % 8 == 0 => all bytes significant.  */
        ands    limit, limit, #7
@@ -177,26 +179,31 @@ def_fn strncmp
 .Lmutual_align:
        /* Sources are mutually aligned, but are not currently at an
           alignment boundary.  Round down the addresses and then mask off
-          the bytes that precede the start point.  */
+          the bytes that precede the start point.
+          We also need to adjust the limit calculations, but without
+          overflowing if the limit is near ULONG_MAX.  */
        bic     src1, src1, #7
        bic     src2, src2, #7
-       add     limit, limit, tmp1      /* Adjust the limit for the extra.  */
-       lsl     tmp1, tmp1, #3          /* Bytes beyond alignment -> bits.  */
        ldr     data1, [src1], #8
-       neg     tmp1, tmp1              /* Bits to alignment -64.  */
+       neg     tmp3, tmp1, lsl #3      /* 64 - bits(bytes beyond align). */
        ldr     data2, [src2], #8
        mov     tmp2, #~0
+       sub     limit_wd, limit, #1     /* limit != 0, so no underflow.  */
 #ifdef __AARCH64EB__
        /* Big-endian.  Early bytes are at MSB.  */
-       lsl     tmp2, tmp2, tmp1        /* Shift (tmp1 & 63).  */
+       lsl     tmp2, tmp2, tmp3        /* Shift (tmp1 & 63).  */
 #else
        /* Little-endian.  Early bytes are at LSB.  */
-       lsr     tmp2, tmp2, tmp1        /* Shift (tmp1 & 63).  */
+       lsr     tmp2, tmp2, tmp3        /* Shift (tmp1 & 63).  */
 #endif
-       add     limit_wd, limit, #7
+       and     tmp3, limit_wd, #7
+       lsr     limit_wd, limit_wd, #3
+       /* Adjust the limit. Only low 3 bits used, so overflow irrelevant.  */
+       add     limit, limit, tmp1
+       add     tmp3, tmp3, tmp1
        orr     data1, data1, tmp2
        orr     data2, data2, tmp2
-       lsr     limit_wd, limit_wd, #3
+       add     limit_wd, limit_wd, tmp3, lsr #3
        b       .Lstart_realigned
 
 .Lret0: