root/sys/arch/hppa/hppa/support.S
/*      $NetBSD: support.S,v 1.9 2025/09/06 02:53:22 riastradh Exp $    */

/*      $OpenBSD: locore.S,v 1.46 2001/09/20 18:33:03 mickey Exp $      */

/*
 * Copyright (c) 1998-2004 Michael Shalayeff
 * All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 *
 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
 * IN NO EVENT SHALL THE AUTHOR OR HIS RELATIVES BE LIABLE FOR ANY DIRECT,
 * INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
 * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
 * SERVICES; LOSS OF MIND, USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
 * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
 * IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF
 * THE POSSIBILITY OF SUCH DAMAGE.
 *
 * Portitions of this file are derived from other sources, see
 * the copyrights and acknowledgements below.
 */
/*
 * Copyright (c) 1990,1991,1992,1994 The University of Utah and
 * the Computer Systems Laboratory (CSL).  All rights reserved.
 *
 * THE UNIVERSITY OF UTAH AND CSL PROVIDE THIS SOFTWARE IN ITS "AS IS"
 * CONDITION, AND DISCLAIM ANY LIABILITY OF ANY KIND FOR ANY DAMAGES
 * WHATSOEVER RESULTING FROM ITS USE.
 *
 * CSL requests users of this software to return to csl-dist@cs.utah.edu any
 * improvements that they make and grant CSL redistribution rights.
 *
 *      Utah $Hdr: locore.s 1.62 94/12/15$
 */
/*
 *  (c) Copyright 1988 HEWLETT-PACKARD COMPANY
 *
 *  To anyone who acknowledges that this file is provided "AS IS"
 *  without any express or implied warranty:
 *      permission to use, copy, modify, and distribute this file
 *  for any purpose is hereby granted without fee, provided that
 *  the above copyright notice and this notice appears in all
 *  copies, and that the name of Hewlett-Packard Company not be
 *  used in advertising or publicity pertaining to distribution
 *  of the software without specific, written prior permission.
 *  Hewlett-Packard Company makes no representations about the
 *  suitability of this software for any purpose.
 */

/*
 * NOTICE: This is not a standalone file.  To use it, #include it in
 * your port's locore.S, like so:
 *
 *      #include <hppa/hppa/support.S>
 */

/*
 * Miscellaneous support routines common to all hppa ports.
 */

/*
 * void fdcache(pa_space_t sp, vaddr_t va, vsize_t size);
 */
        .import dcache_stride, data
LEAF_ENTRY(fdcache)
        ldil    L%dcache_stride,%t1
        ldw     R%dcache_stride(%t1), %arg3

        comb,=,n %arg2, %r0, fdc_none   /* no bytes, no flush */

        mtsp    %arg0, %sr1             /* move the space register to sr1 */
        add     %arg1, %arg2, %arg0     /* get the last byte to flush in arg0 */

        zdep    %arg3, 27, 28, %t1      /* get size of a 16X loop in t1 */
        comb,<  %arg2, %t1, fdc_short   /* check for count < 16 * stride */
        addi    -1, %t1, %t1            /* compute size of large loop - 1 */

        andcm   %arg2, %t1, %t1         /* L = count - (count mod lenbigloop) */
        add     %arg1, %t1, %t1         /* ub for big loop is lb + L */

        fdc,m   %arg3(%sr1, %arg1)      /* Start flushing first cache line. */
fdc_long:
        fdc,m   %arg3(%sr1, %arg1)
        fdc,m   %arg3(%sr1, %arg1)
        fdc,m   %arg3(%sr1, %arg1)
        fdc,m   %arg3(%sr1, %arg1)
        fdc,m   %arg3(%sr1, %arg1)
        fdc,m   %arg3(%sr1, %arg1)
        fdc,m   %arg3(%sr1, %arg1)
        fdc,m   %arg3(%sr1, %arg1)
        fdc,m   %arg3(%sr1, %arg1)
        fdc,m   %arg3(%sr1, %arg1)
        fdc,m   %arg3(%sr1, %arg1)
        fdc,m   %arg3(%sr1, %arg1)
        fdc,m   %arg3(%sr1, %arg1)
        fdc,m   %arg3(%sr1, %arg1)
        fdc,m   %arg3(%sr1, %arg1)
        comb,<<,n %arg1, %t1, fdc_long
        fdc,m   %arg3(%sr1, %arg1)
fdc_short:                              /* flush one line at a time */
        comb,<<,n %arg1, %arg0, fdc_short
        fdc,m   %arg3(%sr1, %arg1)

        addi    -1, %arg0, %arg1
        fdc     %r0(%sr1, %arg1)

fdc_none:
        sync
        syncdma
        bv      %r0(%rp)
        sync
EXIT(fdcache)

/*
 * void pdcache(pa_space_t sp, vaddr_t va, vsize_t size);
 */
        .import dcache_stride, data
LEAF_ENTRY(pdcache)
        ldil    L%dcache_stride,%t1
        ldw     R%dcache_stride(%t1), %arg3

        comb,=  %arg2, %r0, pdc_none    /* no bytes, no purge */

        mtsp    %arg0, %sr1             /* move the space register to sr1 */
        add     %arg1, %arg2, %arg0     /* get the last byte to flush in arg0 */

        zdep    %arg3, 27, 28, %t1      /* get size of a 16X loop in t1 */
        comb,<  %arg2, %t1, pdc_short   /* check for count < 16 * stride */
        addi    -1, %t1, %t1            /* compute size of large loop - 1 */

        andcm   %arg2, %t1, %t1         /* L = count - (count mod lenbigloop) */
        add     %arg1, %t1, %t1         /* ub for big loop is lb + L */

        pdc,m   %arg3(%sr1, %arg1)      /* Start flushing first cache line. */
pdc_long:
        pdc,m   %arg3(%sr1, %arg1)
        pdc,m   %arg3(%sr1, %arg1)
        pdc,m   %arg3(%sr1, %arg1)
        pdc,m   %arg3(%sr1, %arg1)
        pdc,m   %arg3(%sr1, %arg1)
        pdc,m   %arg3(%sr1, %arg1)
        pdc,m   %arg3(%sr1, %arg1)
        pdc,m   %arg3(%sr1, %arg1)
        pdc,m   %arg3(%sr1, %arg1)
        pdc,m   %arg3(%sr1, %arg1)
        pdc,m   %arg3(%sr1, %arg1)
        pdc,m   %arg3(%sr1, %arg1)
        pdc,m   %arg3(%sr1, %arg1)
        pdc,m   %arg3(%sr1, %arg1)
        pdc,m   %arg3(%sr1, %arg1)
        comb,<<,n %arg1, %t1, pdc_long
        pdc,m   %arg3(%sr1, %arg1)
pdc_short:                              /* flush one line at a time */
        comb,<<,n %arg1, %arg0, pdc_short
        pdc,m   %arg3(%sr1, %arg1)

        addi    -1, %arg0, %arg1
        pdc     %r0(%sr1, %arg1)

pdc_none:
        sync
        syncdma
        bv      %r0(%rp)
        sync
EXIT(pdcache)

/*
 * void ficache(pa_space_t sp, vaddr_t va, vsize_t size);
 */
        .import icache_stride, data
LEAF_ENTRY(ficache)
        ldil    L%icache_stride,%t1
        ldw     R%icache_stride(%t1), %arg3

        comb,=  %arg2, %r0, fic_none    /* no bytes, no flush */

        mtsp    %arg0, %sr1             /* move the space register to sr1 */
        add     %arg1, %arg2, %arg0     /* get the last byte to flush in arg0 */

        zdep    %arg3, 27, 28, %t1      /* get size of a 16X loop in t1 */
        comb,<  %arg2, %t1, fic_short   /* check for count < 16 * stride */
        addi    -1, %t1, %t1            /* compute size of large loop - 1 */

        andcm   %arg2, %t1, %t1         /* L = count - (count mod lenbigloop) */
        add     %arg1, %t1, %t1         /* ub for big loop is lb + L */

        fic,m   %arg3(%sr1, %arg1)      /* Start flushing first cache line. */
fic_long:
        fic,m   %arg3(%sr1, %arg1)
        fic,m   %arg3(%sr1, %arg1)
        fic,m   %arg3(%sr1, %arg1)
        fic,m   %arg3(%sr1, %arg1)
        fic,m   %arg3(%sr1, %arg1)
        fic,m   %arg3(%sr1, %arg1)
        fic,m   %arg3(%sr1, %arg1)
        fic,m   %arg3(%sr1, %arg1)
        fic,m   %arg3(%sr1, %arg1)
        fic,m   %arg3(%sr1, %arg1)
        fic,m   %arg3(%sr1, %arg1)
        fic,m   %arg3(%sr1, %arg1)
        fic,m   %arg3(%sr1, %arg1)
        fic,m   %arg3(%sr1, %arg1)
        fic,m   %arg3(%sr1, %arg1)
        comb,<<,n %arg1, %t1, fic_long
        fic,m   %arg3(%sr1, %arg1)
fic_short:                              /* flush one line at a time */
        comb,<<,n %arg1, %arg0, fic_short
        fic,m   %arg3(%sr1, %arg1)

        addi    -1, %arg0, %arg1
        fic     %r0(%sr1, %arg1)

fic_none:
        sync
        syncdma
        bv      %r0(%rp)
        sync
EXIT(ficache)

#ifdef HP7300LC_CPU
        .section .bss
eaio_l2_mask:
        .block 4
        .text
LEAF_ENTRY(eaio_l2)
        ldil    L%eaio_l2_mask, %t2
        ldw     R%eaio_l2_mask(%t2), %t1
        or      %t1, %arg0, %t1
        MTCPU_C(22, DR0_PCXL2_ACCEL_IO)
        nop
        nop
        bv      0(%rp)
        stw     %t1, R%eaio_l2_mask(%t2)
EXIT(eaio_l2)
#endif /* HP7300LC_CPU */

LEAF_ENTRY(setjmp)
/*
 * Save the other general registers whose contents are expected to remain
 * across function calls.  According to the "HP9000 Series 800 Assembly
 * Language Reference Manual", procedures can use general registers 19-26,
 * 28, 29, 1, and 31 without restoring them.  Hence, we do not save these.
 */
        stwm    %r3,4(%arg0)
        stwm    %r4,4(%arg0)
        stwm    %r5,4(%arg0)
        stwm    %r6,4(%arg0)
        stwm    %r7,4(%arg0)
        stwm    %r8,4(%arg0)
        stwm    %r9,4(%arg0)
        stwm    %r10,4(%arg0)
        stwm    %r11,4(%arg0)
        stwm    %r12,4(%arg0)
        stwm    %r13,4(%arg0)
        stwm    %r14,4(%arg0)
        stwm    %r15,4(%arg0)
        stwm    %r16,4(%arg0)
        stwm    %r17,4(%arg0)
        stwm    %r18,4(%arg0)
        stwm    %r27,4(%arg0)   /* Good idea to save the data pointer (dp) */
        stwm    %rp,4(%arg0)    /* Save the return pointer */
        stwm    %sp,4(%arg0)    /* Save the original stack pointer */

        bv      %r0(%rp)
        copy    %r0, %ret0
EXIT(setjmp)

LEAF_ENTRY(longjmp)
/*
 * Restore general registers.
 */
        ldwm    4(%arg0),%r3
        ldwm    4(%arg0),%r4
        ldwm    4(%arg0),%r5
        ldwm    4(%arg0),%r6
        ldwm    4(%arg0),%r7
        ldwm    4(%arg0),%r8
        ldwm    4(%arg0),%r9
        ldwm    4(%arg0),%r10
        ldwm    4(%arg0),%r11
        ldwm    4(%arg0),%r12
        ldwm    4(%arg0),%r13
        ldwm    4(%arg0),%r14
        ldwm    4(%arg0),%r15
        ldwm    4(%arg0),%r16
        ldwm    4(%arg0),%r17
        ldwm    4(%arg0),%r18
        ldwm    4(%arg0),%r27
        ldwm    4(%arg0),%rp    /* Restore return address pointer, */
        ldwm    4(%arg0),%sp    /* stack pointer, */

        bv      %r0(%rp)
        ldi     1, %ret0
EXIT(longjmp)

LEAF_ENTRY(paravirt_membar_sync)
        /*
         * Store-before-load ordering with respect to matching logic
         * on the hypervisor side.
         *
         * This is the same as membar_sync, but guaranteed never to be
         * conditionalized or hotpatched away even on uniprocessor
         * builds and boots -- because under virtualization, we still
         * have to coordinate with a `device' backed by a hypervisor
         * that is potentially on another physical CPU even if we
         * observe only one virtual CPU as the guest.
         */
        bv      %r0(%rp)
         sync
EXIT(paravirt_membar_sync)