/* SPDX-License-Identifier: GPL-2.0 */
/*
 * arch/alpha/lib/ev6-stxcpy.S
 * 21264 version contributed by Rick Gorton <rick.gorton@alpha-processor.com>
 *
 * 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.
 *
 * Much of the information about 21264 scheduling/coding comes from:
 *      Compiler Writer's Guide for the Alpha 21264
 *      abbreviated as 'CWG' in other comments here
 *      ftp.digital.com/pub/Digital/info/semiconductor/literature/dsc-library.html
 * Scheduling notation:
 *      E       - either cluster
 *      U       - upper subcluster; U0 - subcluster U0; U1 - subcluster U1
 *      L       - lower subcluster; L0 - subcluster L0; L1 - subcluster L1
 * Try not to change the actual algorithm if possible for consistency.
 */

#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 4
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          # E : build a mask against false zero
        mskqh   t2, a1, t2      # U :   detection in the src word (stall)
        mskqh   t1, a1, t3      # U :
        ornot   t1, t2, t2      # E : (stall)

        mskql   t0, a1, t0      # U : assemble the first output word
        cmpbge  zero, t2, t8    # E : bits set iff null found
        or      t0, t3, t1      # E : (stall)
        bne     t8, $a_eos      # U : (stall)

        /* On entry to this basic block:
           t0 == the first destination word for masking back in
           t1 == a source word not containing a null.  */
        /* Nops here to separate store quads from load quads */

$a_loop:
        stq_u   t1, 0(a0)       # L :
        addq    a0, 8, a0       # E :
        nop
        nop

        ldq_u   t1, 0(a1)       # L : Latency=3
        addq    a1, 8, a1       # E :
        cmpbge  zero, t1, t8    # E : (3 cycle stall)
        beq     t8, $a_loop     # U : (stall for t8)

        /* 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          # E : find low bit set
        and     t8, t6, t12     # E : (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   # E : (stall)
        bne     t6, 1f          # U : (stall)

        /* We're doing a partial word store and so need to combine
           our source and original destination words.  */
        ldq_u   t0, 0(a0)       # L : Latency=3
        subq    t12, 1, t6      # E :
        zapnot  t1, t6, t1      # U : clear src bytes >= null (stall)
        or      t12, t6, t8     # E : (stall)

        zap     t0, t8, t0      # E : clear dst bytes <= null
        or      t0, t1, t1      # E : (stall)
        nop
        nop

1:      stq_u   t1, 0(a0)       # L :
        ret     (t9)            # L0 : Latency=3
        nop
        nop

        .end stxcpy_aligned

        .align 4
        .ent __stxcpy
        .globl __stxcpy
__stxcpy:
        .frame sp, 0, t9
        .prologue 0

        /* Are source and destination co-aligned?  */
        xor     a0, a1, t0      # E :
        unop                    # E :
        and     t0, 7, t0       # E : (stall)
        bne     t0, $unaligned  # U : (stall)

        /* We are co-aligned; take care of a partial first word.  */
        ldq_u   t1, 0(a1)               # L : load first src word
        and     a0, 7, t0               # E : take care not to load a word ...
        addq    a1, 8, a1               # E :
        beq     t0, stxcpy_aligned      # U : ... if we wont need it (stall)

        ldq_u   t0, 0(a0)       # L :
        br      stxcpy_aligned  # L0 : Latency=3
        nop
        nop


/* 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 4
$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)       # L :
        addq    a1, 8, a1       # E :
        extql   t1, a1, t1      # U : (stall on a1)
        extqh   t2, a1, t4      # U : (stall on a1)

        mskql   t0, a0, t0      # U :
        or      t1, t4, t1      # E :
        mskqh   t1, a0, t1      # U : (stall on t1)
        or      t0, t1, t1      # E : (stall on t1)

        or      t1, t6, t6      # E :
        cmpbge  zero, t6, t8    # E : (stall)
        lda     t6, -1          # E : for masking just below
        bne     t8, $u_final    # U : (stall)

        mskql   t6, a1, t6              # U : mask out the bits we have
        or      t6, t2, t2              # E :   already extracted before (stall)
        cmpbge  zero, t2, t8            # E :   testing eos (stall)
        bne     t8, $u_late_head_exit   # U : (stall)

        /* 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)       # L : store first output word
        addq    a0, 8, a0       # E :
        extql   t2, a1, t0      # U : position ho-bits of lo word
        ldq_u   t2, 8(a1)       # U : read next high-order source word

        addq    a1, 8, a1       # E :
        cmpbge  zero, t2, t8    # E : (stall for t2)
        nop                     # E :
        bne     t8, $u_eos      # U : (stall)

        /* 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      # U : extract high bits for current word
        addq    a1, 8, a1       # E : (stall)
        extql   t2, a1, t3      # U : extract low bits for next time (stall)
        addq    a0, 8, a0       # E :

        or      t0, t1, t1      # E : current dst word now complete
        ldq_u   t2, 0(a1)       # L : Latency=3 load high word for next time
        stq_u   t1, -8(a0)      # L : save the current word (stall)
        mov     t3, t0          # E :

        cmpbge  zero, t2, t8    # E : test new word for eos
        beq     t8, $u_loop     # U : (stall)
        nop
        nop

        /* 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      # U :
        or      t0, t1, t1      # E : first (partial) source word complete (stall)
        cmpbge  zero, t1, t8    # E : is the null in this first bit? (stall)
        bne     t8, $u_final    # U : (stall)

$u_late_head_exit:
        stq_u   t1, 0(a0)       # L : the null was in the high-order bits
        addq    a0, 8, a0       # E :
        extql   t2, a1, t1      # U :
        cmpbge  zero, t1, t8    # E : (stall)

        /* 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          # E : isolate low bit set
        and     t6, t8, t12     # E : (stall)
        and     t12, 0x80, t6   # E : avoid dest word load if we can (stall)
        bne     t6, 1f          # U : (stall)

        ldq_u   t0, 0(a0)       # E :
        subq    t12, 1, t6      # E :
        or      t6, t12, t8     # E : (stall)
        zapnot  t1, t6, t1      # U : kill source bytes >= null (stall)

        zap     t0, t8, t0      # U : kill dest bytes <= null (2 cycle data stall)
        or      t0, t1, t1      # E : (stall)
        nop
        nop

1:      stq_u   t1, 0(a0)       # L :
        ret     (t9)            # L0 : Latency=3
        nop
        nop

        /* Unaligned copy entry point.  */
        .align 4
$unaligned:

        ldq_u   t1, 0(a1)       # L : load first source word
        and     a0, 7, t4       # E : find dest misalignment
        and     a1, 7, t5       # E : find src misalignment
        /* Conditionally load the first destination word and a bytemask
           with 0xff indicating that the destination byte is sacrosanct.  */
        mov     zero, t0        # E :

        mov     zero, t6        # E :
        beq     t4, 1f          # U :
        ldq_u   t0, 0(a0)       # L :
        lda     t6, -1          # E :

        mskql   t6, a0, t6      # U :
        nop
        nop
        nop
1:
        subq    a1, t4, a1      # E : 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     # E :
        beq     t12, $u_head    # U :
        lda     t2, -1          # E : mask out leading garbage in source

        mskqh   t2, t5, t2      # U :
        ornot   t1, t2, t3      # E : (stall)
        cmpbge  zero, t3, t8    # E : is there a zero? (stall)
        beq     t8, $u_head     # U : (stall)

        /* 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)       # L :
        negq    t8, t6          # E : build bitmask of bytes <= zero
        and     t6, t8, t12     # E : (stall)
        and     a1, 7, t5       # E :

        subq    t12, 1, t6      # E :
        or      t6, t12, t8     # E : (stall)
        srl     t12, t5, t12    # U : adjust final null return value
        zapnot  t2, t8, t2      # U : prepare source word; mirror changes (stall)

        and     t1, t2, t1      # E : to source validity mask
        extql   t2, a1, t2      # U :
        extql   t1, a1, t1      # U : (stall)
        andnot  t0, t2, t0      # .. e1 : zero place for source to reside (stall)

        or      t0, t1, t1      # e1    : and put it there
        stq_u   t1, 0(a0)       # .. e0 : (stall)
        ret     (t9)            # e1    :
        nop

        .end __stxcpy