1 /* SPDX-License-Identifier: GPL-2.0-only */
2 /*
3  * Cache flushing routines.
4  *
5  * Copyright (C) 1999-2001, 2005 Hewlett-Packard Co
6  *	David Mosberger-Tang <davidm@hpl.hp.com>
7  *
8  * 05/28/05 Zoltan Menyhart	Dynamic stride size
9  */
10 
11 #include <asm/asmmacro.h>
12 #include <asm/export.h>
13 
14 
15 	/*
16 	 * flush_icache_range(start,end)
17 	 *
18 	 *	Make i-cache(s) coherent with d-caches.
19 	 *
20 	 *	Must deal with range from start to end-1 but nothing else (need to
21 	 *	be careful not to touch addresses that may be unmapped).
22 	 *
23 	 *	Note: "in0" and "in1" are preserved for debugging purposes.
24 	 */
25 	.section .kprobes.text,"ax"
26 GLOBAL_ENTRY(flush_icache_range)
27 
28 	.prologue
29 	alloc	r2=ar.pfs,2,0,0,0
30 	movl	r3=ia64_i_cache_stride_shift
31  	mov	r21=1
32 	;;
33 	ld8	r20=[r3]		// r20: stride shift
34 	sub	r22=in1,r0,1		// last byte address
35 	;;
36 	shr.u	r23=in0,r20		// start / (stride size)
37 	shr.u	r22=r22,r20		// (last byte address) / (stride size)
38 	shl	r21=r21,r20		// r21: stride size of the i-cache(s)
39 	;;
40 	sub	r8=r22,r23		// number of strides - 1
41 	shl	r24=r23,r20		// r24: addresses for "fc.i" =
42 					//	"start" rounded down to stride boundary
43 	.save	ar.lc,r3
44 	mov	r3=ar.lc		// save ar.lc
45 	;;
46 
47 	.body
48 	mov	ar.lc=r8
49 	;;
50 	/*
51 	 * 32 byte aligned loop, even number of (actually 2) bundles
52 	 */
53 .Loop:	fc.i	r24			// issuable on M0 only
54 	add	r24=r21,r24		// we flush "stride size" bytes per iteration
55 	nop.i	0
56 	br.cloop.sptk.few .Loop
57 	;;
58 	sync.i
59 	;;
60 	srlz.i
61 	;;
62 	mov	ar.lc=r3		// restore ar.lc
63 	br.ret.sptk.many rp
64 END(flush_icache_range)
65 EXPORT_SYMBOL_GPL(flush_icache_range)
66 
67 	/*
68 	 * clflush_cache_range(start,size)
69 	 *
70 	 *	Flush cache lines from start to start+size-1.
71 	 *
72 	 *	Must deal with range from start to start+size-1 but nothing else
73 	 *	(need to be careful not to touch addresses that may be
74 	 *	unmapped).
75 	 *
76 	 *	Note: "in0" and "in1" are preserved for debugging purposes.
77 	 */
78 	.section .kprobes.text,"ax"
79 GLOBAL_ENTRY(clflush_cache_range)
80 
81 	.prologue
82 	alloc	r2=ar.pfs,2,0,0,0
83 	movl	r3=ia64_cache_stride_shift
84 	mov	r21=1
85 	add     r22=in1,in0
86 	;;
87 	ld8	r20=[r3]		// r20: stride shift
88 	sub	r22=r22,r0,1		// last byte address
89 	;;
90 	shr.u	r23=in0,r20		// start / (stride size)
91 	shr.u	r22=r22,r20		// (last byte address) / (stride size)
92 	shl	r21=r21,r20		// r21: stride size of the i-cache(s)
93 	;;
94 	sub	r8=r22,r23		// number of strides - 1
95 	shl	r24=r23,r20		// r24: addresses for "fc" =
96 					//	"start" rounded down to stride
97 					//	boundary
98 	.save	ar.lc,r3
99 	mov	r3=ar.lc		// save ar.lc
100 	;;
101 
102 	.body
103 	mov	ar.lc=r8
104 	;;
105 	/*
106 	 * 32 byte aligned loop, even number of (actually 2) bundles
107 	 */
108 .Loop_fc:
109 	fc	r24		// issuable on M0 only
110 	add	r24=r21,r24	// we flush "stride size" bytes per iteration
111 	nop.i	0
112 	br.cloop.sptk.few .Loop_fc
113 	;;
114 	sync.i
115 	;;
116 	srlz.i
117 	;;
118 	mov	ar.lc=r3		// restore ar.lc
119 	br.ret.sptk.many rp
120 END(clflush_cache_range)
121