Mark Brown | b77e995 | 2021-12-10 18:41:02 +0000 | [diff] [blame] | 1 | // SPDX-License-Identifier: GPL-2.0-only |
| 2 | // Copyright (C) 2021 ARM Limited. |
| 3 | // |
| 4 | // Assembly portion of the syscall ABI test |
| 5 | |
| 6 | // |
| 7 | // Load values from memory into registers, invoke a syscall and save the |
| 8 | // register values back to memory for later checking. The syscall to be |
| 9 | // invoked is configured in x8 of the input GPR data. |
| 10 | // |
| 11 | // x0: SVE VL, 0 for FP only |
| 12 | // |
| 13 | // GPRs: gpr_in, gpr_out |
| 14 | // FPRs: fpr_in, fpr_out |
| 15 | // Zn: z_in, z_out |
| 16 | // Pn: p_in, p_out |
| 17 | // FFR: ffr_in, ffr_out |
| 18 | |
| 19 | .arch_extension sve |
| 20 | |
| 21 | .globl do_syscall |
| 22 | do_syscall: |
| 23 | // Store callee saved registers x19-x29 (80 bytes) plus x0 and x1 |
| 24 | stp x29, x30, [sp, #-112]! |
| 25 | mov x29, sp |
| 26 | stp x0, x1, [sp, #16] |
| 27 | stp x19, x20, [sp, #32] |
| 28 | stp x21, x22, [sp, #48] |
| 29 | stp x23, x24, [sp, #64] |
| 30 | stp x25, x26, [sp, #80] |
| 31 | stp x27, x28, [sp, #96] |
| 32 | |
| 33 | // Load GPRs x8-x28, and save our SP/FP for later comparison |
| 34 | ldr x2, =gpr_in |
| 35 | add x2, x2, #64 |
| 36 | ldp x8, x9, [x2], #16 |
| 37 | ldp x10, x11, [x2], #16 |
| 38 | ldp x12, x13, [x2], #16 |
| 39 | ldp x14, x15, [x2], #16 |
| 40 | ldp x16, x17, [x2], #16 |
| 41 | ldp x18, x19, [x2], #16 |
| 42 | ldp x20, x21, [x2], #16 |
| 43 | ldp x22, x23, [x2], #16 |
| 44 | ldp x24, x25, [x2], #16 |
| 45 | ldp x26, x27, [x2], #16 |
| 46 | ldr x28, [x2], #8 |
| 47 | str x29, [x2], #8 // FP |
| 48 | str x30, [x2], #8 // LR |
| 49 | |
| 50 | // Load FPRs if we're not doing SVE |
| 51 | cbnz x0, 1f |
| 52 | ldr x2, =fpr_in |
| 53 | ldp q0, q1, [x2] |
| 54 | ldp q2, q3, [x2, #16 * 2] |
| 55 | ldp q4, q5, [x2, #16 * 4] |
| 56 | ldp q6, q7, [x2, #16 * 6] |
| 57 | ldp q8, q9, [x2, #16 * 8] |
| 58 | ldp q10, q11, [x2, #16 * 10] |
| 59 | ldp q12, q13, [x2, #16 * 12] |
| 60 | ldp q14, q15, [x2, #16 * 14] |
| 61 | ldp q16, q17, [x2, #16 * 16] |
| 62 | ldp q18, q19, [x2, #16 * 18] |
| 63 | ldp q20, q21, [x2, #16 * 20] |
| 64 | ldp q22, q23, [x2, #16 * 22] |
| 65 | ldp q24, q25, [x2, #16 * 24] |
| 66 | ldp q26, q27, [x2, #16 * 26] |
| 67 | ldp q28, q29, [x2, #16 * 28] |
| 68 | ldp q30, q31, [x2, #16 * 30] |
| 69 | 1: |
| 70 | |
| 71 | // Load the SVE registers if we're doing SVE |
| 72 | cbz x0, 1f |
| 73 | |
| 74 | ldr x2, =z_in |
| 75 | ldr z0, [x2, #0, MUL VL] |
| 76 | ldr z1, [x2, #1, MUL VL] |
| 77 | ldr z2, [x2, #2, MUL VL] |
| 78 | ldr z3, [x2, #3, MUL VL] |
| 79 | ldr z4, [x2, #4, MUL VL] |
| 80 | ldr z5, [x2, #5, MUL VL] |
| 81 | ldr z6, [x2, #6, MUL VL] |
| 82 | ldr z7, [x2, #7, MUL VL] |
| 83 | ldr z8, [x2, #8, MUL VL] |
| 84 | ldr z9, [x2, #9, MUL VL] |
| 85 | ldr z10, [x2, #10, MUL VL] |
| 86 | ldr z11, [x2, #11, MUL VL] |
| 87 | ldr z12, [x2, #12, MUL VL] |
| 88 | ldr z13, [x2, #13, MUL VL] |
| 89 | ldr z14, [x2, #14, MUL VL] |
| 90 | ldr z15, [x2, #15, MUL VL] |
| 91 | ldr z16, [x2, #16, MUL VL] |
| 92 | ldr z17, [x2, #17, MUL VL] |
| 93 | ldr z18, [x2, #18, MUL VL] |
| 94 | ldr z19, [x2, #19, MUL VL] |
| 95 | ldr z20, [x2, #20, MUL VL] |
| 96 | ldr z21, [x2, #21, MUL VL] |
| 97 | ldr z22, [x2, #22, MUL VL] |
| 98 | ldr z23, [x2, #23, MUL VL] |
| 99 | ldr z24, [x2, #24, MUL VL] |
| 100 | ldr z25, [x2, #25, MUL VL] |
| 101 | ldr z26, [x2, #26, MUL VL] |
| 102 | ldr z27, [x2, #27, MUL VL] |
| 103 | ldr z28, [x2, #28, MUL VL] |
| 104 | ldr z29, [x2, #29, MUL VL] |
| 105 | ldr z30, [x2, #30, MUL VL] |
| 106 | ldr z31, [x2, #31, MUL VL] |
| 107 | |
| 108 | ldr x2, =ffr_in |
| 109 | ldr p0, [x2, #0] |
| 110 | wrffr p0.b |
| 111 | |
| 112 | ldr x2, =p_in |
| 113 | ldr p0, [x2, #0, MUL VL] |
| 114 | ldr p1, [x2, #1, MUL VL] |
| 115 | ldr p2, [x2, #2, MUL VL] |
| 116 | ldr p3, [x2, #3, MUL VL] |
| 117 | ldr p4, [x2, #4, MUL VL] |
| 118 | ldr p5, [x2, #5, MUL VL] |
| 119 | ldr p6, [x2, #6, MUL VL] |
| 120 | ldr p7, [x2, #7, MUL VL] |
| 121 | ldr p8, [x2, #8, MUL VL] |
| 122 | ldr p9, [x2, #9, MUL VL] |
| 123 | ldr p10, [x2, #10, MUL VL] |
| 124 | ldr p11, [x2, #11, MUL VL] |
| 125 | ldr p12, [x2, #12, MUL VL] |
| 126 | ldr p13, [x2, #13, MUL VL] |
| 127 | ldr p14, [x2, #14, MUL VL] |
| 128 | ldr p15, [x2, #15, MUL VL] |
| 129 | 1: |
| 130 | |
| 131 | // Do the syscall |
| 132 | svc #0 |
| 133 | |
| 134 | // Save GPRs x8-x30 |
| 135 | ldr x2, =gpr_out |
| 136 | add x2, x2, #64 |
| 137 | stp x8, x9, [x2], #16 |
| 138 | stp x10, x11, [x2], #16 |
| 139 | stp x12, x13, [x2], #16 |
| 140 | stp x14, x15, [x2], #16 |
| 141 | stp x16, x17, [x2], #16 |
| 142 | stp x18, x19, [x2], #16 |
| 143 | stp x20, x21, [x2], #16 |
| 144 | stp x22, x23, [x2], #16 |
| 145 | stp x24, x25, [x2], #16 |
| 146 | stp x26, x27, [x2], #16 |
| 147 | stp x28, x29, [x2], #16 |
| 148 | str x30, [x2] |
| 149 | |
| 150 | // Restore x0 and x1 for feature checks |
| 151 | ldp x0, x1, [sp, #16] |
| 152 | |
| 153 | // Save FPSIMD state |
| 154 | ldr x2, =fpr_out |
| 155 | stp q0, q1, [x2] |
| 156 | stp q2, q3, [x2, #16 * 2] |
| 157 | stp q4, q5, [x2, #16 * 4] |
| 158 | stp q6, q7, [x2, #16 * 6] |
| 159 | stp q8, q9, [x2, #16 * 8] |
| 160 | stp q10, q11, [x2, #16 * 10] |
| 161 | stp q12, q13, [x2, #16 * 12] |
| 162 | stp q14, q15, [x2, #16 * 14] |
| 163 | stp q16, q17, [x2, #16 * 16] |
| 164 | stp q18, q19, [x2, #16 * 18] |
| 165 | stp q20, q21, [x2, #16 * 20] |
| 166 | stp q22, q23, [x2, #16 * 22] |
| 167 | stp q24, q25, [x2, #16 * 24] |
| 168 | stp q26, q27, [x2, #16 * 26] |
| 169 | stp q28, q29, [x2, #16 * 28] |
| 170 | stp q30, q31, [x2, #16 * 30] |
| 171 | |
| 172 | // Save the SVE state if we have some |
| 173 | cbz x0, 1f |
| 174 | |
| 175 | ldr x2, =z_out |
| 176 | str z0, [x2, #0, MUL VL] |
| 177 | str z1, [x2, #1, MUL VL] |
| 178 | str z2, [x2, #2, MUL VL] |
| 179 | str z3, [x2, #3, MUL VL] |
| 180 | str z4, [x2, #4, MUL VL] |
| 181 | str z5, [x2, #5, MUL VL] |
| 182 | str z6, [x2, #6, MUL VL] |
| 183 | str z7, [x2, #7, MUL VL] |
| 184 | str z8, [x2, #8, MUL VL] |
| 185 | str z9, [x2, #9, MUL VL] |
| 186 | str z10, [x2, #10, MUL VL] |
| 187 | str z11, [x2, #11, MUL VL] |
| 188 | str z12, [x2, #12, MUL VL] |
| 189 | str z13, [x2, #13, MUL VL] |
| 190 | str z14, [x2, #14, MUL VL] |
| 191 | str z15, [x2, #15, MUL VL] |
| 192 | str z16, [x2, #16, MUL VL] |
| 193 | str z17, [x2, #17, MUL VL] |
| 194 | str z18, [x2, #18, MUL VL] |
| 195 | str z19, [x2, #19, MUL VL] |
| 196 | str z20, [x2, #20, MUL VL] |
| 197 | str z21, [x2, #21, MUL VL] |
| 198 | str z22, [x2, #22, MUL VL] |
| 199 | str z23, [x2, #23, MUL VL] |
| 200 | str z24, [x2, #24, MUL VL] |
| 201 | str z25, [x2, #25, MUL VL] |
| 202 | str z26, [x2, #26, MUL VL] |
| 203 | str z27, [x2, #27, MUL VL] |
| 204 | str z28, [x2, #28, MUL VL] |
| 205 | str z29, [x2, #29, MUL VL] |
| 206 | str z30, [x2, #30, MUL VL] |
| 207 | str z31, [x2, #31, MUL VL] |
| 208 | |
| 209 | ldr x2, =p_out |
| 210 | str p0, [x2, #0, MUL VL] |
| 211 | str p1, [x2, #1, MUL VL] |
| 212 | str p2, [x2, #2, MUL VL] |
| 213 | str p3, [x2, #3, MUL VL] |
| 214 | str p4, [x2, #4, MUL VL] |
| 215 | str p5, [x2, #5, MUL VL] |
| 216 | str p6, [x2, #6, MUL VL] |
| 217 | str p7, [x2, #7, MUL VL] |
| 218 | str p8, [x2, #8, MUL VL] |
| 219 | str p9, [x2, #9, MUL VL] |
| 220 | str p10, [x2, #10, MUL VL] |
| 221 | str p11, [x2, #11, MUL VL] |
| 222 | str p12, [x2, #12, MUL VL] |
| 223 | str p13, [x2, #13, MUL VL] |
| 224 | str p14, [x2, #14, MUL VL] |
| 225 | str p15, [x2, #15, MUL VL] |
| 226 | |
| 227 | ldr x2, =ffr_out |
| 228 | rdffr p0.b |
| 229 | str p0, [x2, #0] |
| 230 | 1: |
| 231 | |
| 232 | // Restore callee saved registers x19-x30 |
| 233 | ldp x19, x20, [sp, #32] |
| 234 | ldp x21, x22, [sp, #48] |
| 235 | ldp x23, x24, [sp, #64] |
| 236 | ldp x25, x26, [sp, #80] |
| 237 | ldp x27, x28, [sp, #96] |
| 238 | ldp x29, x30, [sp], #112 |
| 239 | |
| 240 | ret |