Lines Matching full:bytes

64 # Fold reg1, reg2 into the next 32 data bytes, storing the result back into
102 	# For sizes less than 256 bytes, we can't fold 128 bytes at a time.
106 	# Load the first 128 data bytes.  Byte swapping is necessary to make the
133 	# Subtract 128 for the 128 data bytes just consumed.  Subtract another
137 	# While >= 128 data bytes remain (not counting xmm0-7), fold the 128
138 	# bytes xmm0-7 into them, storing the result back into xmm0-7.
148 	# Now fold the 112 bytes in xmm0-xmm6 into the 16 bytes in xmm7.
150 	# Fold across 64 bytes.
156 	# Fold across 32 bytes.
160 	# Fold across 16 bytes.
164 	# Add 128 to get the correct number of data bytes remaining in 0...127
170 	# While >= 16 data bytes remain (not counting xmm7), fold the 16 bytes
186 	# Add 16 to get the correct number of data bytes remaining in 0...15
192 	# Reduce the last '16 + len' bytes where 1 <= len <= 15 and the first 16
193 	# bytes are in xmm7 and the rest are the remaining data in 'buf'.  To do
195 	# the bytes into a first chunk of 'len' bytes and a second chunk of 16
196 	# bytes, then fold the first chunk into the second.
200 	# xmm1 = last 16 original data bytes
204 	# xmm2 = high order part of second chunk: xmm7 left-shifted by 'len' bytes.
210 	# xmm7 = first chunk: xmm7 right-shifted by '16-len' bytes.
214 	# xmm1 = second chunk: 'len' bytes from xmm1 (low-order bytes),
215 	# then '16-len' bytes from xmm2 (high-order bytes).
264 	# Checksumming a buffer of length 16...255 bytes
266 	# Load the first 16 data bytes.
327 # is the index vector to shift left by 'len' bytes, and is also {0x80, ...,
328 # 0x80} XOR the index vector to shift right by '16 - len' bytes.