diff options
Diffstat (limited to 'REORG.TODO/sysdeps/alpha/stxncpy.S')
| -rw-r--r-- | REORG.TODO/sysdeps/alpha/stxncpy.S | 352 |
1 files changed, 352 insertions, 0 deletions
diff --git a/REORG.TODO/sysdeps/alpha/stxncpy.S b/REORG.TODO/sysdeps/alpha/stxncpy.S new file mode 100644 index 0000000000..718a37ad0a --- /dev/null +++ b/REORG.TODO/sysdeps/alpha/stxncpy.S @@ -0,0 +1,352 @@ +/* Copyright (C) 1996-2017 Free Software Foundation, Inc. + Contributed by Richard Henderson (rth@tamu.edu) + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library. If not, see + <http://www.gnu.org/licenses/>. */ + +/* Copy no more than COUNT bytes of the null-terminated string from + SRC to DST. + + This is an internal routine used by strncpy, stpncpy, and strncat. + 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 + a2 = COUNT + + Furthermore, COUNT may not be zero. + + On output: + t0 = last word written + t8 = bitmask (with one bit set) indicating the last byte written + t10 = bitmask (with one bit set) indicating the byte position of + the end of the range specified by COUNT + a0 = unaligned address of the last *word* written + a2 = the number of full words left in COUNT + + Furthermore, v0, a3-a5, t11, and t12 are untouched. +*/ + + +/* This is generally scheduled for the EV5, but should still be pretty + good for the EV4 too. */ + +#include <sysdep.h> + + .set noat + .set noreorder + + .text + .type __stxncpy, @function + .globl __stxncpy + .usepv __stxncpy, no + + cfi_startproc + cfi_return_column (t9) + + /* On entry to this basic block: + t0 == the first destination word for masking back in + t1 == the first source word. */ + .align 3 +stxncpy_aligned: + /* 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, t7 # .. e1 : bits set iff null found + or t0, t3, t0 # e0 : + beq a2, $a_eoc # .. e1 : + bne t7, $a_eos # .. e1 : + + /* On entry to this basic block: + t0 == a source word not containing a null. */ +$a_loop: + stq_u t0, 0(a0) # e0 : + addq a0, 8, a0 # .. e1 : + ldq_u t0, 0(a1) # e0 : + addq a1, 8, a1 # .. e1 : + subq a2, 1, a2 # e0 : + cmpbge zero, t0, t7 # .. e1 (stall) + beq a2, $a_eoc # e1 : + beq t7, $a_loop # e1 : + + /* Take care of the final (partial) word store. At this point + the end-of-count bit is set in t7 iff it applies. + + On entry to this basic block we have: + t0 == the source word containing the null + t7 == the cmpbge mask that found it. */ +$a_eos: + negq t7, t8 # e0 : find low bit set + and t7, t8, t8 # e1 (stall) + + /* For the sake of the cache, don't read a destination word + if we're not going to need it. */ + and t8, 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 t1, 0(a0) # e0 : + subq t8, 1, t6 # .. e1 : + or t8, t6, t7 # e0 : + unop # + zapnot t0, t7, t0 # e0 : clear src bytes > null + zap t1, t7, t1 # .. e1 : clear dst bytes <= null + or t0, t1, t0 # e1 : + +1: stq_u t0, 0(a0) # e0 : + ret (t9) # e1 : + + /* Add the end-of-count bit to the eos detection bitmask. */ +$a_eoc: + or t10, t7, t7 + br $a_eos + + .align 3 +__stxncpy: + /* Are source and destination co-aligned? */ + lda t2, -1 + xor a0, a1, t1 + srl t2, 1, t2 + and a0, 7, t0 # find dest misalignment + cmovlt a2, t2, a2 # bound neg count to LONG_MAX + and t1, 7, t1 + addq a2, t0, a2 # bias count by dest misalignment + subq a2, 1, a2 + and a2, 7, t2 + srl a2, 3, a2 # a2 = loop counter = (count - 1)/8 + addq zero, 1, t10 + sll t10, t2, t10 # t10 = bitmask of last count byte + bne t1, $unaligned + + /* We are co-aligned; take care of a partial first word. */ + + ldq_u t1, 0(a1) # e0 : load first src word + addq a1, 8, a1 # .. e1 : + + beq t0, stxncpy_aligned # avoid loading dest word if not needed + ldq_u t0, 0(a0) # e0 : + br stxncpy_aligned # .. e1 : + + +/* The source and destination are not co-aligned. Align the destination + and cope. We have to be very careful about not reading too much and + causing a SEGV. */ + + .align 3 +$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, unmasked + t1 == the shifted low bits of the first source word + t6 == bytemask that is -1 in dest word bytes */ + + ldq_u t2, 8(a1) # e0 : load second src word + addq a1, 8, a1 # .. e1 : + mskql t0, a0, t0 # e0 : mask trailing garbage in dst + extqh t2, a1, t4 # e0 : + or t1, t4, t1 # e1 : first aligned src word complete + mskqh t1, a0, t1 # e0 : mask leading garbage in src + or t0, t1, t0 # e0 : first output word complete + or t0, t6, t6 # e1 : mask original data for zero test + cmpbge zero, t6, t7 # e0 : + beq a2, $u_eocfin # .. e1 : + lda t6, -1 # e0 : + bne t7, $u_final # .. e1 : + + mskql t6, a1, t6 # e0 : mask out bits already seen + nop # .. e1 : + stq_u t0, 0(a0) # e0 : store first output word + or t6, t2, t2 # .. e1 : + cmpbge zero, t2, t7 # e0 : find nulls in second partial + addq a0, 8, a0 # .. e1 : + subq a2, 1, a2 # e0 : + bne t7, $u_late_head_exit # .. e1 : + + /* Finally, we've got all the stupid leading edge cases taken care + of and we can set up to enter the main loop. */ + + extql t2, a1, t1 # e0 : position hi-bits of lo word + beq a2, $u_eoc # .. e1 : + ldq_u t2, 8(a1) # e0 : read next high-order source word + addq a1, 8, a1 # .. e1 : + extqh t2, a1, t0 # e0 : position lo-bits of hi word + cmpbge zero, t2, t7 # .. e1 : test new word for eos + nop # e0 : + bne t7, $u_eos # .. e1 : + + /* Unaligned copy main loop. In order to avoid reading too much, + the loop is structured to detect zeros in aligned source words. + This has, unfortunately, effectively pulled half of a loop + iteration out into the head and half into the tail, but it does + prevent nastiness from accumulating in the very thing we want + to run as fast as possible. + + On entry to this basic block: + t0 == the shifted low-order bits from the current source word + t1 == the shifted high-order bits from the previous source word + t2 == the unshifted current source word + + We further know that t2 does not contain a null terminator. */ + + .align 3 +$u_loop: + or t0, t1, t0 # e0 : current dst word now complete + subq a2, 1, a2 # .. e1 : decrement word count + stq_u t0, 0(a0) # e0 : save the current word + addq a0, 8, a0 # .. e1 : + extql t2, a1, t1 # e0 : extract high bits for next time + beq a2, $u_eoc # .. e1 : + ldq_u t2, 8(a1) # e0 : load high word for next time + addq a1, 8, a1 # .. e1 : + nop # e0 : + cmpbge zero, t2, t7 # .. e1 : test new word for eos + extqh t2, a1, t0 # e0 : extract low bits for current word + beq t7, $u_loop # .. e1 : + + /* We've found a zero somewhere in the source word we just read. + If it resides in the lower half, we have one (probably partial) + word to write out, and if it resides in the upper half, we + have one full and one partial word left to write out. + + On entry to this basic block: + t0 == the shifted low-order bits from the current source word + t1 == the shifted high-order bits from the previous source word + t2 == the unshifted current source word. */ +$u_eos: + or t0, t1, t0 # e0 : first (partial) source word complete + cmpbge zero, t0, t7 # e0 : is the null in this first bit? + bne t7, $u_final # .. e1 (zdb) + + stq_u t0, 0(a0) # e0 : the null was in the high-order bits + addq a0, 8, a0 # .. e1 : + subq a2, 1, a2 # e0 : + +$u_late_head_exit: + extql t2, a1, t0 # e0 : + cmpbge zero, t0, t7 # e0 : + or t7, t10, t6 # e1 : + cmoveq a2, t6, t7 # e0 : + + /* Take care of a final (probably partial) result word. + On entry to this basic block: + t0 == assembled source word + t7 == cmpbge mask that found the null. */ +$u_final: + negq t7, t6 # e0 : isolate low bit set + and t6, t7, t8 # e1 : + + and t8, 0x80, t6 # e0 : avoid dest word load if we can + bne t6, 1f # .. e1 (zdb) + + ldq_u t1, 0(a0) # e0 : + subq t8, 1, t6 # .. e1 : + or t6, t8, t7 # e0 : + zapnot t0, t7, t0 # .. e1 : kill source bytes > null + zap t1, t7, t1 # e0 : kill dest bytes <= null + or t0, t1, t0 # e1 : + +1: stq_u t0, 0(a0) # e0 : + ret (t9) # .. e1 : + + /* Got to end-of-count before end of string. + On entry to this basic block: + t1 == the shifted high-order bits from the previous source word */ +$u_eoc: + and a1, 7, t6 # e1 : + sll t10, t6, t6 # e0 : + and t6, 0xff, t6 # e0 : + bne t6, 1f # e1 : avoid src word load if we can + + ldq_u t2, 8(a1) # e0 : load final src word + nop # .. e1 : + extqh t2, a1, t0 # e0 : extract high bits for last word + or t1, t0, t1 # e1 : + +1: cmpbge zero, t1, t7 + mov t1, t0 + +$u_eocfin: # end-of-count, final word + or t10, t7, t7 + br $u_final + + /* Unaligned copy entry point. */ + .align 3 +$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, t8 # e1 : + extql t1, a1, t1 # .. e0 : shift src into place + lda t2, -1 # e0 : for creating masks later + beq t8, $u_head # e1 : + + mskqh t2, t5, t2 # e0 : begin src byte validity mask + cmpbge zero, t1, t7 # .. e1 : is there a zero? + extql t2, a1, t2 # e0 : + or t7, t10, t5 # .. e1 : test for end-of-count too + cmpbge zero, t2, t3 # e0 : + cmoveq a2, t5, t7 # .. e1 : + andnot t7, t3, t7 # e0 : + beq t7, $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 t7, t6 # .. e1 : build bitmask of bytes <= zero + mskqh t1, t4, t1 # e0 : + and t6, t7, t8 # .. e1 : + subq t8, 1, t6 # e0 : + or t6, t8, t7 # e1 : + + zapnot t2, t7, t2 # e0 : prepare source word; mirror changes + zapnot t1, t7, t1 # .. e1 : to source validity mask + + andnot t0, t2, t0 # e0 : zero place for source to reside + or t0, t1, t0 # e1 : and put it there + stq_u t0, 0(a0) # e0 : + ret (t9) # .. e1 : + + cfi_endproc |