Loading...
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 | /* * arch/alpha/lib/stxcpy.S * Contributed by Richard Henderson (rth@tamu.edu) * * Copy a null-terminated string from SRC to DST. * * This is an internal routine used by strcpy, stpcpy, and strcat. * As such, it uses special linkage conventions to make implementation * of these public functions more efficient. * * On input: * t9 = return address * a0 = DST * a1 = SRC * * On output: * t12 = bitmask (with one bit set) indicating the last byte written * a0 = unaligned address of the last *word* written * * Furthermore, v0, a3-a5, t11, and t12 are untouched. */ #include <asm/regdef.h> .set noat .set noreorder .text /* There is a problem with either gdb (as of 4.16) or gas (as of 2.7) that doesn't like putting the entry point for a procedure somewhere in the middle of the procedure descriptor. Work around this by putting the aligned copy in its own procedure descriptor */ .ent stxcpy_aligned .align 3 stxcpy_aligned: .frame sp, 0, t9 .prologue 0 /* On entry to this basic block: t0 == the first destination word for masking back in t1 == the first source word. */ /* Create the 1st output word and detect 0's in the 1st input word. */ lda t2, -1 # e1 : build a mask against false zero mskqh t2, a1, t2 # e0 : detection in the src word mskqh t1, a1, t3 # e0 : ornot t1, t2, t2 # .. e1 : mskql t0, a1, t0 # e0 : assemble the first output word cmpbge zero, t2, t8 # .. e1 : bits set iff null found or t0, t3, t1 # e0 : bne t8, $a_eos # .. e1 : /* On entry to this basic block: t0 == the first destination word for masking back in t1 == a source word not containing a null. */ $a_loop: stq_u t1, 0(a0) # e0 : addq a0, 8, a0 # .. e1 : ldq_u t1, 0(a1) # e0 : addq a1, 8, a1 # .. e1 : cmpbge zero, t1, t8 # e0 (stall) beq t8, $a_loop # .. e1 (zdb) /* Take care of the final (partial) word store. On entry to this basic block we have: t1 == the source word containing the null t8 == the cmpbge mask that found it. */ $a_eos: negq t8, t6 # e0 : find low bit set and t8, t6, t12 # e1 (stall) /* For the sake of the cache, don't read a destination word if we're not going to need it. */ and t12, 0x80, t6 # e0 : bne t6, 1f # .. e1 (zdb) /* We're doing a partial word store and so need to combine our source and original destination words. */ ldq_u t0, 0(a0) # e0 : subq t12, 1, t6 # .. e1 : zapnot t1, t6, t1 # e0 : clear src bytes >= null or t12, t6, t8 # .. e1 : zap t0, t8, t0 # e0 : clear dst bytes <= null or t0, t1, t1 # e1 : 1: stq_u t1, 0(a0) # e0 : ret (t9) # .. e1 : .end stxcpy_aligned .align 3 .ent __stxcpy .globl __stxcpy __stxcpy: .frame sp, 0, t9 .prologue 0 /* Are source and destination co-aligned? */ xor a0, a1, t0 # e0 : unop # : and t0, 7, t0 # e0 : bne t0, $unaligned # .. e1 : /* We are co-aligned; take care of a partial first word. */ ldq_u t1, 0(a1) # e0 : load first src word and a0, 7, t0 # .. e1 : take care not to load a word ... addq a1, 8, a1 # e0 : beq t0, stxcpy_aligned # .. e1 : ... if we wont need it ldq_u t0, 0(a0) # e0 : br stxcpy_aligned # .. e1 : /* The source and destination are not co-aligned. Align the destination and cope. We have to be very careful about not reading too much and causing a SEGV. */ .align 3 $u_head: /* We know just enough now to be able to assemble the first full source word. We can still find a zero at the end of it that prevents us from outputting the whole thing. On entry to this basic block: t0 == the first dest word, for masking back in, if needed else 0 t1 == the low bits of the first source word t6 == bytemask that is -1 in dest word bytes */ ldq_u t2, 8(a1) # e0 : addq a1, 8, a1 # .. e1 : extql t1, a1, t1 # e0 : extqh t2, a1, t4 # e0 : mskql t0, a0, t0 # e0 : or t1, t4, t1 # .. e1 : mskqh t1, a0, t1 # e0 : or t0, t1, t1 # e1 : or t1, t6, t6 # e0 : cmpbge zero, t6, t8 # .. e1 : lda t6, -1 # e0 : for masking just below bne t8, $u_final # .. e1 : mskql t6, a1, t6 # e0 : mask out the bits we have or t6, t2, t2 # e1 : already extracted before cmpbge zero, t2, t8 # e0 : testing eos bne t8, $u_late_head_exit # .. e1 (zdb) /* Finally, we've got all the stupid leading edge cases taken care of and we can set up to enter the main loop. */ stq_u t1, 0(a0) # e0 : store first output word addq a0, 8, a0 # .. e1 : extql t2, a1, t0 # e0 : position ho-bits of lo word ldq_u t2, 8(a1) # .. e1 : read next high-order source word addq a1, 8, a1 # e0 : cmpbge zero, t2, t8 # .. e1 : nop # e0 : bne t8, $u_eos # .. e1 : /* Unaligned copy main loop. In order to avoid reading too much, the loop is structured to detect zeros in aligned source words. This has, unfortunately, effectively pulled half of a loop iteration out into the head and half into the tail, but it does prevent nastiness from accumulating in the very thing we want to run as fast as possible. On entry to this basic block: t0 == the shifted high-order bits from the previous source word t2 == the unshifted current source word We further know that t2 does not contain a null terminator. */ .align 3 $u_loop: extqh t2, a1, t1 # e0 : extract high bits for current word addq a1, 8, a1 # .. e1 : extql t2, a1, t3 # e0 : extract low bits for next time addq a0, 8, a0 # .. e1 : or t0, t1, t1 # e0 : current dst word now complete ldq_u t2, 0(a1) # .. e1 : load high word for next time stq_u t1, -8(a0) # e0 : save the current word mov t3, t0 # .. e1 : cmpbge zero, t2, t8 # e0 : test new word for eos beq t8, $u_loop # .. e1 : /* We've found a zero somewhere in the source word we just read. If it resides in the lower half, we have one (probably partial) word to write out, and if it resides in the upper half, we have one full and one partial word left to write out. On entry to this basic block: t0 == the shifted high-order bits from the previous source word t2 == the unshifted current source word. */ $u_eos: extqh t2, a1, t1 # e0 : or t0, t1, t1 # e1 : first (partial) source word complete cmpbge zero, t1, t8 # e0 : is the null in this first bit? bne t8, $u_final # .. e1 (zdb) $u_late_head_exit: stq_u t1, 0(a0) # e0 : the null was in the high-order bits addq a0, 8, a0 # .. e1 : extql t2, a1, t1 # e0 : cmpbge zero, t1, t8 # .. e1 : /* Take care of a final (probably partial) result word. On entry to this basic block: t1 == assembled source word t8 == cmpbge mask that found the null. */ $u_final: negq t8, t6 # e0 : isolate low bit set and t6, t8, t12 # e1 : and t12, 0x80, t6 # e0 : avoid dest word load if we can bne t6, 1f # .. e1 (zdb) ldq_u t0, 0(a0) # e0 : subq t12, 1, t6 # .. e1 : or t6, t12, t8 # e0 : zapnot t1, t6, t1 # .. e1 : kill source bytes >= null zap t0, t8, t0 # e0 : kill dest bytes <= null or t0, t1, t1 # e1 : 1: stq_u t1, 0(a0) # e0 : ret (t9) # .. e1 : /* Unaligned copy entry point. */ .align 3 $unaligned: ldq_u t1, 0(a1) # e0 : load first source word and a0, 7, t4 # .. e1 : find dest misalignment and a1, 7, t5 # e0 : find src misalignment /* Conditionally load the first destination word and a bytemask with 0xff indicating that the destination byte is sacrosanct. */ mov zero, t0 # .. e1 : mov zero, t6 # e0 : beq t4, 1f # .. e1 : ldq_u t0, 0(a0) # e0 : lda t6, -1 # .. e1 : mskql t6, a0, t6 # e0 : 1: subq a1, t4, a1 # .. e1 : sub dest misalignment from src addr /* If source misalignment is larger than dest misalignment, we need extra startup checks to avoid SEGV. */ cmplt t4, t5, t12 # e0 : beq t12, $u_head # .. e1 (zdb) lda t2, -1 # e1 : mask out leading garbage in source mskqh t2, t5, t2 # e0 : nop # e0 : ornot t1, t2, t3 # .. e1 : cmpbge zero, t3, t8 # e0 : is there a zero? beq t8, $u_head # .. e1 (zdb) /* At this point we've found a zero in the first partial word of the source. We need to isolate the valid source data and mask it into the original destination data. (Incidentally, we know that we'll need at least one byte of that original dest word.) */ ldq_u t0, 0(a0) # e0 : negq t8, t6 # .. e1 : build bitmask of bytes <= zero and t6, t8, t12 # e0 : and a1, 7, t5 # .. e1 : subq t12, 1, t6 # e0 : or t6, t12, t8 # e1 : srl t12, t5, t12 # e0 : adjust final null return value zapnot t2, t8, t2 # .. e1 : prepare source word; mirror changes and t1, t2, t1 # e1 : to source validity mask extql t2, a1, t2 # .. e0 : extql t1, a1, t1 # e0 : andnot t0, t2, t0 # .. e1 : zero place for source to reside or t0, t1, t1 # e1 : and put it there stq_u t1, 0(a0) # .. e0 : ret (t9) # e1 : .end __stxcpy |