summaryrefslogtreecommitdiff
path: root/arch/arm/lib/csumpartial.S
diff options
context:
space:
mode:
authorLinus Torvalds <torvalds@ppc970.osdl.org>2005-04-16 15:20:36 -0700
committerLinus Torvalds <torvalds@ppc970.osdl.org>2005-04-16 15:20:36 -0700
commit1da177e4c3f41524e886b7f1b8a0c1fc7321cac2 (patch)
tree0bba044c4ce775e45a88a51686b5d9f90697ea9d /arch/arm/lib/csumpartial.S
Linux-2.6.12-rc2
Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
Diffstat (limited to 'arch/arm/lib/csumpartial.S')
-rw-r--r--arch/arm/lib/csumpartial.S137
1 files changed, 137 insertions, 0 deletions
diff --git a/arch/arm/lib/csumpartial.S b/arch/arm/lib/csumpartial.S
new file mode 100644
index 000000000000..cb5e3708f118
--- /dev/null
+++ b/arch/arm/lib/csumpartial.S
@@ -0,0 +1,137 @@
+/*
+ * linux/arch/arm/lib/csumpartial.S
+ *
+ * Copyright (C) 1995-1998 Russell King
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#include <linux/linkage.h>
+#include <asm/assembler.h>
+
+ .text
+
+/*
+ * Function: __u32 csum_partial(const char *src, int len, __u32 sum)
+ * Params : r0 = buffer, r1 = len, r2 = checksum
+ * Returns : r0 = new checksum
+ */
+
+buf .req r0
+len .req r1
+sum .req r2
+td0 .req r3
+td1 .req r4 @ save before use
+td2 .req r5 @ save before use
+td3 .req lr
+
+.zero: mov r0, sum
+ add sp, sp, #4
+ ldr pc, [sp], #4
+
+ /*
+ * Handle 0 to 7 bytes, with any alignment of source and
+ * destination pointers. Note that when we get here, C = 0
+ */
+.less8: teq len, #0 @ check for zero count
+ beq .zero
+
+ /* we must have at least one byte. */
+ tst buf, #1 @ odd address?
+ ldrneb td0, [buf], #1
+ subne len, len, #1
+ adcnes sum, sum, td0, put_byte_1
+
+.less4: tst len, #6
+ beq .less8_byte
+
+ /* we are now half-word aligned */
+
+.less8_wordlp:
+#if __LINUX_ARM_ARCH__ >= 4
+ ldrh td0, [buf], #2
+ sub len, len, #2
+#else
+ ldrb td0, [buf], #1
+ ldrb td3, [buf], #1
+ sub len, len, #2
+#ifndef __ARMEB__
+ orr td0, td0, td3, lsl #8
+#else
+ orr td0, td3, td0, lsl #8
+#endif
+#endif
+ adcs sum, sum, td0
+ tst len, #6
+ bne .less8_wordlp
+
+.less8_byte: tst len, #1 @ odd number of bytes
+ ldrneb td0, [buf], #1 @ include last byte
+ adcnes sum, sum, td0, put_byte_0 @ update checksum
+
+.done: adc r0, sum, #0 @ collect up the last carry
+ ldr td0, [sp], #4
+ tst td0, #1 @ check buffer alignment
+ movne r0, r0, ror #8 @ rotate checksum by 8 bits
+ ldr pc, [sp], #4 @ return
+
+.not_aligned: tst buf, #1 @ odd address
+ ldrneb td0, [buf], #1 @ make even
+ subne len, len, #1
+ adcnes sum, sum, td0, put_byte_1 @ update checksum
+
+ tst buf, #2 @ 32-bit aligned?
+#if __LINUX_ARM_ARCH__ >= 4
+ ldrneh td0, [buf], #2 @ make 32-bit aligned
+ subne len, len, #2
+#else
+ ldrneb td0, [buf], #1
+ ldrneb ip, [buf], #1
+ subne len, len, #2
+#ifndef __ARMEB__
+ orrne td0, td0, ip, lsl #8
+#else
+ orrne td0, ip, td0, lsl #8
+#endif
+#endif
+ adcnes sum, sum, td0 @ update checksum
+ mov pc, lr
+
+ENTRY(csum_partial)
+ stmfd sp!, {buf, lr}
+ cmp len, #8 @ Ensure that we have at least
+ blo .less8 @ 8 bytes to copy.
+
+ adds sum, sum, #0 @ C = 0
+ tst buf, #3 @ Test destination alignment
+ blne .not_aligned @ aligh destination, return here
+
+1: bics ip, len, #31
+ beq 3f
+
+ stmfd sp!, {r4 - r5}
+2: ldmia buf!, {td0, td1, td2, td3}
+ adcs sum, sum, td0
+ adcs sum, sum, td1
+ adcs sum, sum, td2
+ adcs sum, sum, td3
+ ldmia buf!, {td0, td1, td2, td3}
+ adcs sum, sum, td0
+ adcs sum, sum, td1
+ adcs sum, sum, td2
+ adcs sum, sum, td3
+ sub ip, ip, #32
+ teq ip, #0
+ bne 2b
+ ldmfd sp!, {r4 - r5}
+
+3: tst len, #0x1c @ should not change C
+ beq .less4
+
+4: ldr td0, [buf], #4
+ sub len, len, #4
+ adcs sum, sum, td0
+ tst len, #0x1c
+ bne 4b
+ b .less4