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// x1:	SME VL
13//
14//	GPRs:	gpr_in, gpr_out
15//	FPRs:	fpr_in, fpr_out
16//	Zn:	z_in, z_out
17//	Pn:	p_in, p_out
18//	FFR:	ffr_in, ffr_out
19//	ZA:	za_in, za_out
20//	SVCR:	svcr_in, svcr_out
21
22#include "syscall-abi.h"
23
24.arch_extension sve
25
26#define ID_AA64SMFR0_EL1_SMEver_SHIFT           56
27#define ID_AA64SMFR0_EL1_SMEver_WIDTH           4
28
29/*
30 * LDR (vector to ZA array):
31 *	LDR ZA[\nw, #\offset], [X\nxbase, #\offset, MUL VL]
32 */
33.macro _ldr_za nw, nxbase, offset=0
34	.inst	0xe1000000			\
35		| (((\nw) & 3) << 13)		\
36		| ((\nxbase) << 5)		\
37		| ((\offset) & 7)
38.endm
39
40/*
41 * STR (vector from ZA array):
42 *	STR ZA[\nw, #\offset], [X\nxbase, #\offset, MUL VL]
43 */
44.macro _str_za nw, nxbase, offset=0
45	.inst	0xe1200000			\
46		| (((\nw) & 3) << 13)		\
47		| ((\nxbase) << 5)		\
48		| ((\offset) & 7)
49.endm
50
51/*
52 * LDR (ZT0)
53 *
54 *	LDR ZT0, nx
55 */
56.macro _ldr_zt nx
57	.inst	0xe11f8000			\
58		| (((\nx) & 0x1f) << 5)
59.endm
60
61/*
62 * STR (ZT0)
63 *
64 *	STR ZT0, nx
65 */
66.macro _str_zt nx
67	.inst	0xe13f8000			\
68		| (((\nx) & 0x1f) << 5)
69.endm
70
71.globl do_syscall
72do_syscall:
73	// Store callee saved registers x19-x29 (80 bytes) plus x0 and x1
74	stp	x29, x30, [sp, #-112]!
75	mov	x29, sp
76	stp	x0, x1, [sp, #16]
77	stp	x19, x20, [sp, #32]
78	stp	x21, x22, [sp, #48]
79	stp	x23, x24, [sp, #64]
80	stp	x25, x26, [sp, #80]
81	stp	x27, x28, [sp, #96]
82
83	// Set SVCR if we're doing SME
84	cbz	x1, 1f
85	adrp	x2, svcr_in
86	ldr	x2, [x2, :lo12:svcr_in]
87	msr	S3_3_C4_C2_2, x2
881:
89
90	// Load ZA and ZT0 if enabled - uses x12 as scratch due to SME LDR
91	tbz	x2, #SVCR_ZA_SHIFT, 1f
92	mov	w12, #0
93	ldr	x2, =za_in
942:	_ldr_za 12, 2
95	add	x2, x2, x1
96	add	x12, x12, #1
97	cmp	x1, x12
98	bne	2b
99
100	// ZT0
101	mrs	x2, S3_0_C0_C4_5	// ID_AA64SMFR0_EL1
102	ubfx	x2, x2, #ID_AA64SMFR0_EL1_SMEver_SHIFT, \
103			 #ID_AA64SMFR0_EL1_SMEver_WIDTH
104	cbz	x2, 1f
105	adrp	x2, zt_in
106	add	x2, x2, :lo12:zt_in
107	_ldr_zt 2
1081:
109
110	// Load GPRs x8-x28, and save our SP/FP for later comparison
111	ldr	x2, =gpr_in
112	add	x2, x2, #64
113	ldp	x8, x9, [x2], #16
114	ldp	x10, x11, [x2], #16
115	ldp	x12, x13, [x2], #16
116	ldp	x14, x15, [x2], #16
117	ldp	x16, x17, [x2], #16
118	ldp	x18, x19, [x2], #16
119	ldp	x20, x21, [x2], #16
120	ldp	x22, x23, [x2], #16
121	ldp	x24, x25, [x2], #16
122	ldp	x26, x27, [x2], #16
123	ldr	x28, [x2], #8
124	str	x29, [x2], #8		// FP
125	str	x30, [x2], #8		// LR
126
127	// Load FPRs if we're not doing neither SVE nor streaming SVE
128	cbnz	x0, 1f
129	ldr	x2, =svcr_in
130	tbnz	x2, #SVCR_SM_SHIFT, 1f
131
132	ldr	x2, =fpr_in
133	ldp	q0, q1, [x2]
134	ldp	q2, q3, [x2, #16 * 2]
135	ldp	q4, q5, [x2, #16 * 4]
136	ldp	q6, q7, [x2, #16 * 6]
137	ldp	q8, q9, [x2, #16 * 8]
138	ldp	q10, q11, [x2, #16 * 10]
139	ldp	q12, q13, [x2, #16 * 12]
140	ldp	q14, q15, [x2, #16 * 14]
141	ldp	q16, q17, [x2, #16 * 16]
142	ldp	q18, q19, [x2, #16 * 18]
143	ldp	q20, q21, [x2, #16 * 20]
144	ldp	q22, q23, [x2, #16 * 22]
145	ldp	q24, q25, [x2, #16 * 24]
146	ldp	q26, q27, [x2, #16 * 26]
147	ldp	q28, q29, [x2, #16 * 28]
148	ldp	q30, q31, [x2, #16 * 30]
149
150	b	2f
1511:
152
153	// Load the SVE registers if we're doing SVE/SME
154
155	ldr	x2, =z_in
156	ldr	z0, [x2, #0, MUL VL]
157	ldr	z1, [x2, #1, MUL VL]
158	ldr	z2, [x2, #2, MUL VL]
159	ldr	z3, [x2, #3, MUL VL]
160	ldr	z4, [x2, #4, MUL VL]
161	ldr	z5, [x2, #5, MUL VL]
162	ldr	z6, [x2, #6, MUL VL]
163	ldr	z7, [x2, #7, MUL VL]
164	ldr	z8, [x2, #8, MUL VL]
165	ldr	z9, [x2, #9, MUL VL]
166	ldr	z10, [x2, #10, MUL VL]
167	ldr	z11, [x2, #11, MUL VL]
168	ldr	z12, [x2, #12, MUL VL]
169	ldr	z13, [x2, #13, MUL VL]
170	ldr	z14, [x2, #14, MUL VL]
171	ldr	z15, [x2, #15, MUL VL]
172	ldr	z16, [x2, #16, MUL VL]
173	ldr	z17, [x2, #17, MUL VL]
174	ldr	z18, [x2, #18, MUL VL]
175	ldr	z19, [x2, #19, MUL VL]
176	ldr	z20, [x2, #20, MUL VL]
177	ldr	z21, [x2, #21, MUL VL]
178	ldr	z22, [x2, #22, MUL VL]
179	ldr	z23, [x2, #23, MUL VL]
180	ldr	z24, [x2, #24, MUL VL]
181	ldr	z25, [x2, #25, MUL VL]
182	ldr	z26, [x2, #26, MUL VL]
183	ldr	z27, [x2, #27, MUL VL]
184	ldr	z28, [x2, #28, MUL VL]
185	ldr	z29, [x2, #29, MUL VL]
186	ldr	z30, [x2, #30, MUL VL]
187	ldr	z31, [x2, #31, MUL VL]
188
189	// Only set a non-zero FFR, test patterns must be zero since the
190	// syscall should clear it - this lets us handle FA64.
191	ldr	x2, =ffr_in
192	ldr	p0, [x2]
193	ldr	x2, [x2, #0]
194	cbz	x2, 1f
195	wrffr	p0.b
1961:
197
198	ldr	x2, =p_in
199	ldr	p0, [x2, #0, MUL VL]
200	ldr	p1, [x2, #1, MUL VL]
201	ldr	p2, [x2, #2, MUL VL]
202	ldr	p3, [x2, #3, MUL VL]
203	ldr	p4, [x2, #4, MUL VL]
204	ldr	p5, [x2, #5, MUL VL]
205	ldr	p6, [x2, #6, MUL VL]
206	ldr	p7, [x2, #7, MUL VL]
207	ldr	p8, [x2, #8, MUL VL]
208	ldr	p9, [x2, #9, MUL VL]
209	ldr	p10, [x2, #10, MUL VL]
210	ldr	p11, [x2, #11, MUL VL]
211	ldr	p12, [x2, #12, MUL VL]
212	ldr	p13, [x2, #13, MUL VL]
213	ldr	p14, [x2, #14, MUL VL]
214	ldr	p15, [x2, #15, MUL VL]
2152:
216
217	// Do the syscall
218	svc	#0
219
220	// Save GPRs x8-x30
221	ldr	x2, =gpr_out
222	add	x2, x2, #64
223	stp	x8, x9, [x2], #16
224	stp	x10, x11, [x2], #16
225	stp	x12, x13, [x2], #16
226	stp	x14, x15, [x2], #16
227	stp	x16, x17, [x2], #16
228	stp	x18, x19, [x2], #16
229	stp	x20, x21, [x2], #16
230	stp	x22, x23, [x2], #16
231	stp	x24, x25, [x2], #16
232	stp	x26, x27, [x2], #16
233	stp	x28, x29, [x2], #16
234	str	x30, [x2]
235
236	// Restore x0 and x1 for feature checks
237	ldp	x0, x1, [sp, #16]
238
239	// Save FPSIMD state
240	ldr	x2, =fpr_out
241	stp	q0, q1, [x2]
242	stp	q2, q3, [x2, #16 * 2]
243	stp	q4, q5, [x2, #16 * 4]
244	stp	q6, q7, [x2, #16 * 6]
245	stp	q8, q9, [x2, #16 * 8]
246	stp	q10, q11, [x2, #16 * 10]
247	stp	q12, q13, [x2, #16 * 12]
248	stp	q14, q15, [x2, #16 * 14]
249	stp	q16, q17, [x2, #16 * 16]
250	stp	q18, q19, [x2, #16 * 18]
251	stp	q20, q21, [x2, #16 * 20]
252	stp	q22, q23, [x2, #16 * 22]
253	stp	q24, q25, [x2, #16 * 24]
254	stp	q26, q27, [x2, #16 * 26]
255	stp	q28, q29, [x2, #16 * 28]
256	stp	q30, q31, [x2, #16 * 30]
257
258	// Save SVCR if we're doing SME
259	cbz	x1, 1f
260	mrs	x2, S3_3_C4_C2_2
261	adrp	x3, svcr_out
262	str	x2, [x3, :lo12:svcr_out]
2631:
264
265	// Save ZA if it's enabled - uses x12 as scratch due to SME STR
266	tbz	x2, #SVCR_ZA_SHIFT, 1f
267	mov	w12, #0
268	ldr	x2, =za_out
2692:	_str_za 12, 2
270	add	x2, x2, x1
271	add	x12, x12, #1
272	cmp	x1, x12
273	bne	2b
274
275	// ZT0
276	mrs	x2, S3_0_C0_C4_5	// ID_AA64SMFR0_EL1
277	ubfx	x2, x2, #ID_AA64SMFR0_EL1_SMEver_SHIFT, \
278			#ID_AA64SMFR0_EL1_SMEver_WIDTH
279	cbz	x2, 1f
280	adrp	x2, zt_out
281	add	x2, x2, :lo12:zt_out
282	_str_zt 2
2831:
284
285	// Save the SVE state if we have some
286	cbz	x0, 1f
287
288	ldr	x2, =z_out
289	str	z0, [x2, #0, MUL VL]
290	str	z1, [x2, #1, MUL VL]
291	str	z2, [x2, #2, MUL VL]
292	str	z3, [x2, #3, MUL VL]
293	str	z4, [x2, #4, MUL VL]
294	str	z5, [x2, #5, MUL VL]
295	str	z6, [x2, #6, MUL VL]
296	str	z7, [x2, #7, MUL VL]
297	str	z8, [x2, #8, MUL VL]
298	str	z9, [x2, #9, MUL VL]
299	str	z10, [x2, #10, MUL VL]
300	str	z11, [x2, #11, MUL VL]
301	str	z12, [x2, #12, MUL VL]
302	str	z13, [x2, #13, MUL VL]
303	str	z14, [x2, #14, MUL VL]
304	str	z15, [x2, #15, MUL VL]
305	str	z16, [x2, #16, MUL VL]
306	str	z17, [x2, #17, MUL VL]
307	str	z18, [x2, #18, MUL VL]
308	str	z19, [x2, #19, MUL VL]
309	str	z20, [x2, #20, MUL VL]
310	str	z21, [x2, #21, MUL VL]
311	str	z22, [x2, #22, MUL VL]
312	str	z23, [x2, #23, MUL VL]
313	str	z24, [x2, #24, MUL VL]
314	str	z25, [x2, #25, MUL VL]
315	str	z26, [x2, #26, MUL VL]
316	str	z27, [x2, #27, MUL VL]
317	str	z28, [x2, #28, MUL VL]
318	str	z29, [x2, #29, MUL VL]
319	str	z30, [x2, #30, MUL VL]
320	str	z31, [x2, #31, MUL VL]
321
322	ldr	x2, =p_out
323	str	p0, [x2, #0, MUL VL]
324	str	p1, [x2, #1, MUL VL]
325	str	p2, [x2, #2, MUL VL]
326	str	p3, [x2, #3, MUL VL]
327	str	p4, [x2, #4, MUL VL]
328	str	p5, [x2, #5, MUL VL]
329	str	p6, [x2, #6, MUL VL]
330	str	p7, [x2, #7, MUL VL]
331	str	p8, [x2, #8, MUL VL]
332	str	p9, [x2, #9, MUL VL]
333	str	p10, [x2, #10, MUL VL]
334	str	p11, [x2, #11, MUL VL]
335	str	p12, [x2, #12, MUL VL]
336	str	p13, [x2, #13, MUL VL]
337	str	p14, [x2, #14, MUL VL]
338	str	p15, [x2, #15, MUL VL]
339
340	// Only save FFR if we wrote a value for SME
341	ldr	x2, =ffr_in
342	ldr	x2, [x2, #0]
343	cbz	x2, 1f
344	ldr	x2, =ffr_out
345	rdffr	p0.b
346	str	p0, [x2]
3471:
348
349	// Restore callee saved registers x19-x30
350	ldp	x19, x20, [sp, #32]
351	ldp	x21, x22, [sp, #48]
352	ldp	x23, x24, [sp, #64]
353	ldp	x25, x26, [sp, #80]
354	ldp	x27, x28, [sp, #96]
355	ldp	x29, x30, [sp], #112
356
357	// Clear SVCR if we were doing SME so future tests don't have ZA
358	cbz	x1, 1f
359	msr	S3_3_C4_C2_2, xzr
3601:
361
362	ret
363