xref: /openbmc/qemu/target/hexagon/gen_tcg.h (revision 2f0a771d)
1 /*
2  *  Copyright(c) 2019-2024 Qualcomm Innovation Center, Inc. All Rights Reserved.
3  *
4  *  This program is free software; you can redistribute it and/or modify
5  *  it under the terms of the GNU General Public License as published by
6  *  the Free Software Foundation; either version 2 of the License, or
7  *  (at your option) any later version.
8  *
9  *  This program is distributed in the hope that it will be useful,
10  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
11  *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
12  *  GNU General Public License for more details.
13  *
14  *  You should have received a copy of the GNU General Public License
15  *  along with this program; if not, see <http://www.gnu.org/licenses/>.
16  */
17 
18 #ifndef HEXAGON_GEN_TCG_H
19 #define HEXAGON_GEN_TCG_H
20 
21 /*
22  * Here is a primer to understand the tag names for load/store instructions
23  *
24  * Data types
25  *      b        signed byte                       r0 = memb(r2+#0)
26  *     ub        unsigned byte                     r0 = memub(r2+#0)
27  *      h        signed half word (16 bits)        r0 = memh(r2+#0)
28  *     uh        unsigned half word                r0 = memuh(r2+#0)
29  *      i        integer (32 bits)                 r0 = memw(r2+#0)
30  *      d        double word (64 bits)             r1:0 = memd(r2+#0)
31  *
32  * Addressing modes
33  *     _io       indirect with offset              r0 = memw(r1+#4)
34  *     _ur       absolute with register offset     r0 = memw(r1<<#4+##variable)
35  *     _rr       indirect with register offset     r0 = memw(r1+r4<<#2)
36  *     gp        global pointer relative           r0 = memw(gp+#200)
37  *     _sp       stack pointer relative            r0 = memw(r29+#12)
38  *     _ap       absolute set                      r0 = memw(r1=##variable)
39  *     _pr       post increment register           r0 = memw(r1++m1)
40  *     _pbr      post increment bit reverse        r0 = memw(r1++m1:brev)
41  *     _pi       post increment immediate          r0 = memb(r1++#1)
42  *     _pci      post increment circular immediate r0 = memw(r1++#4:circ(m0))
43  *     _pcr      post increment circular register  r0 = memw(r1++I:circ(m0))
44  */
45 
46 /* Macros for complex addressing modes */
47 #define GET_EA_ap \
48     do { \
49         fEA_IMM(UiV); \
50         tcg_gen_movi_tl(ReV, UiV); \
51     } while (0)
52 #define GET_EA_pr \
53     do { \
54         fEA_REG(RxV); \
55         fPM_M(RxV, MuV); \
56     } while (0)
57 #define GET_EA_pbr \
58     do { \
59         gen_helper_fbrev(EA, RxV); \
60         tcg_gen_add_tl(RxV, RxV, MuV); \
61     } while (0)
62 #define GET_EA_pi \
63     do { \
64         fEA_REG(RxV); \
65         fPM_I(RxV, siV); \
66     } while (0)
67 #define GET_EA_pci \
68     do { \
69         TCGv tcgv_siV = tcg_constant_tl(siV); \
70         tcg_gen_mov_tl(EA, RxV); \
71         gen_helper_fcircadd(RxV, RxV, tcgv_siV, MuV, CS); \
72     } while (0)
73 #define GET_EA_pcr(SHIFT) \
74     do { \
75         TCGv ireg = tcg_temp_new(); \
76         tcg_gen_mov_tl(EA, RxV); \
77         gen_read_ireg(ireg, MuV, (SHIFT)); \
78         gen_helper_fcircadd(RxV, RxV, ireg, MuV, CS); \
79     } while (0)
80 
81 /* Instructions with multiple definitions */
82 #define fGEN_TCG_LOAD_AP(RES, SIZE, SIGN) \
83     do { \
84         fMUST_IMMEXT(UiV); \
85         fEA_IMM(UiV); \
86         fLOAD(1, SIZE, SIGN, EA, RES); \
87         tcg_gen_movi_tl(ReV, UiV); \
88     } while (0)
89 
90 #define fGEN_TCG_L4_loadrub_ap(SHORTCODE) \
91     fGEN_TCG_LOAD_AP(RdV, 1, u)
92 #define fGEN_TCG_L4_loadrb_ap(SHORTCODE) \
93     fGEN_TCG_LOAD_AP(RdV, 1, s)
94 #define fGEN_TCG_L4_loadruh_ap(SHORTCODE) \
95     fGEN_TCG_LOAD_AP(RdV, 2, u)
96 #define fGEN_TCG_L4_loadrh_ap(SHORTCODE) \
97     fGEN_TCG_LOAD_AP(RdV, 2, s)
98 #define fGEN_TCG_L4_loadri_ap(SHORTCODE) \
99     fGEN_TCG_LOAD_AP(RdV, 4, u)
100 #define fGEN_TCG_L4_loadrd_ap(SHORTCODE) \
101     fGEN_TCG_LOAD_AP(RddV, 8, u)
102 
103 #define fGEN_TCG_L2_loadrub_pci(SHORTCODE)    SHORTCODE
104 #define fGEN_TCG_L2_loadrb_pci(SHORTCODE)     SHORTCODE
105 #define fGEN_TCG_L2_loadruh_pci(SHORTCODE)    SHORTCODE
106 #define fGEN_TCG_L2_loadrh_pci(SHORTCODE)     SHORTCODE
107 #define fGEN_TCG_L2_loadri_pci(SHORTCODE)     SHORTCODE
108 #define fGEN_TCG_L2_loadrd_pci(SHORTCODE)     SHORTCODE
109 
110 #define fGEN_TCG_LOAD_pcr(SHIFT, LOAD) \
111     do { \
112         TCGv ireg = tcg_temp_new(); \
113         tcg_gen_mov_tl(EA, RxV); \
114         gen_read_ireg(ireg, MuV, SHIFT); \
115         gen_helper_fcircadd(RxV, RxV, ireg, MuV, CS); \
116         LOAD; \
117     } while (0)
118 
119 #define fGEN_TCG_L2_loadrub_pcr(SHORTCODE) \
120       fGEN_TCG_LOAD_pcr(0, fLOAD(1, 1, u, EA, RdV))
121 #define fGEN_TCG_L2_loadrb_pcr(SHORTCODE) \
122       fGEN_TCG_LOAD_pcr(0, fLOAD(1, 1, s, EA, RdV))
123 #define fGEN_TCG_L2_loadruh_pcr(SHORTCODE) \
124       fGEN_TCG_LOAD_pcr(1, fLOAD(1, 2, u, EA, RdV))
125 #define fGEN_TCG_L2_loadrh_pcr(SHORTCODE) \
126       fGEN_TCG_LOAD_pcr(1, fLOAD(1, 2, s, EA, RdV))
127 #define fGEN_TCG_L2_loadri_pcr(SHORTCODE) \
128       fGEN_TCG_LOAD_pcr(2, fLOAD(1, 4, u, EA, RdV))
129 #define fGEN_TCG_L2_loadrd_pcr(SHORTCODE) \
130       fGEN_TCG_LOAD_pcr(3, fLOAD(1, 8, u, EA, RddV))
131 
132 #define fGEN_TCG_L2_loadrub_pr(SHORTCODE)      SHORTCODE
133 #define fGEN_TCG_L2_loadrub_pbr(SHORTCODE)     SHORTCODE
134 #define fGEN_TCG_L2_loadrub_pi(SHORTCODE)      SHORTCODE
135 #define fGEN_TCG_L2_loadrb_pr(SHORTCODE)       SHORTCODE
136 #define fGEN_TCG_L2_loadrb_pbr(SHORTCODE)      SHORTCODE
137 #define fGEN_TCG_L2_loadrb_pi(SHORTCODE)       SHORTCODE
138 #define fGEN_TCG_L2_loadruh_pr(SHORTCODE)      SHORTCODE
139 #define fGEN_TCG_L2_loadruh_pbr(SHORTCODE)     SHORTCODE
140 #define fGEN_TCG_L2_loadruh_pi(SHORTCODE)      SHORTCODE
141 #define fGEN_TCG_L2_loadrh_pr(SHORTCODE)       SHORTCODE
142 #define fGEN_TCG_L2_loadrh_pbr(SHORTCODE)      SHORTCODE
143 #define fGEN_TCG_L2_loadrh_pi(SHORTCODE)       SHORTCODE
144 #define fGEN_TCG_L2_loadri_pr(SHORTCODE)       SHORTCODE
145 #define fGEN_TCG_L2_loadri_pbr(SHORTCODE)      SHORTCODE
146 #define fGEN_TCG_L2_loadri_pi(SHORTCODE)       SHORTCODE
147 #define fGEN_TCG_L2_loadrd_pr(SHORTCODE)       SHORTCODE
148 #define fGEN_TCG_L2_loadrd_pbr(SHORTCODE)      SHORTCODE
149 #define fGEN_TCG_L2_loadrd_pi(SHORTCODE)       SHORTCODE
150 
151 /*
152  * These instructions load 2 bytes and places them in
153  * two halves of the destination register.
154  * The GET_EA macro determines the addressing mode.
155  * The SIGN argument determines whether to zero-extend or
156  * sign-extend.
157  */
158 #define fGEN_TCG_loadbXw2(GET_EA, SIGN) \
159     do { \
160         TCGv tmp = tcg_temp_new(); \
161         TCGv byte = tcg_temp_new(); \
162         GET_EA; \
163         fLOAD(1, 2, u, EA, tmp); \
164         tcg_gen_movi_tl(RdV, 0); \
165         for (int i = 0; i < 2; i++) { \
166             gen_set_half(i, RdV, gen_get_byte(byte, i, tmp, (SIGN))); \
167         } \
168     } while (0)
169 
170 #define fGEN_TCG_L2_loadbzw2_io(SHORTCODE) \
171     fGEN_TCG_loadbXw2(fEA_RI(RsV, siV), false)
172 #define fGEN_TCG_L4_loadbzw2_ur(SHORTCODE) \
173     fGEN_TCG_loadbXw2(fEA_IRs(UiV, RtV, uiV), false)
174 #define fGEN_TCG_L2_loadbsw2_io(SHORTCODE) \
175     fGEN_TCG_loadbXw2(fEA_RI(RsV, siV), true)
176 #define fGEN_TCG_L4_loadbsw2_ur(SHORTCODE) \
177     fGEN_TCG_loadbXw2(fEA_IRs(UiV, RtV, uiV), true)
178 #define fGEN_TCG_L4_loadbzw2_ap(SHORTCODE) \
179     fGEN_TCG_loadbXw2(GET_EA_ap, false)
180 #define fGEN_TCG_L2_loadbzw2_pr(SHORTCODE) \
181     fGEN_TCG_loadbXw2(GET_EA_pr, false)
182 #define fGEN_TCG_L2_loadbzw2_pbr(SHORTCODE) \
183     fGEN_TCG_loadbXw2(GET_EA_pbr, false)
184 #define fGEN_TCG_L2_loadbzw2_pi(SHORTCODE) \
185     fGEN_TCG_loadbXw2(GET_EA_pi, false)
186 #define fGEN_TCG_L4_loadbsw2_ap(SHORTCODE) \
187     fGEN_TCG_loadbXw2(GET_EA_ap, true)
188 #define fGEN_TCG_L2_loadbsw2_pr(SHORTCODE) \
189     fGEN_TCG_loadbXw2(GET_EA_pr, true)
190 #define fGEN_TCG_L2_loadbsw2_pbr(SHORTCODE) \
191     fGEN_TCG_loadbXw2(GET_EA_pbr, true)
192 #define fGEN_TCG_L2_loadbsw2_pi(SHORTCODE) \
193     fGEN_TCG_loadbXw2(GET_EA_pi, true)
194 #define fGEN_TCG_L2_loadbzw2_pci(SHORTCODE) \
195     fGEN_TCG_loadbXw2(GET_EA_pci, false)
196 #define fGEN_TCG_L2_loadbsw2_pci(SHORTCODE) \
197     fGEN_TCG_loadbXw2(GET_EA_pci, true)
198 #define fGEN_TCG_L2_loadbzw2_pcr(SHORTCODE) \
199     fGEN_TCG_loadbXw2(GET_EA_pcr(1), false)
200 #define fGEN_TCG_L2_loadbsw2_pcr(SHORTCODE) \
201     fGEN_TCG_loadbXw2(GET_EA_pcr(1), true)
202 
203 /*
204  * These instructions load 4 bytes and places them in
205  * four halves of the destination register pair.
206  * The GET_EA macro determines the addressing mode.
207  * The SIGN argument determines whether to zero-extend or
208  * sign-extend.
209  */
210 #define fGEN_TCG_loadbXw4(GET_EA, SIGN) \
211     do { \
212         TCGv tmp = tcg_temp_new(); \
213         TCGv byte = tcg_temp_new(); \
214         GET_EA; \
215         fLOAD(1, 4, u, EA, tmp);  \
216         tcg_gen_movi_i64(RddV, 0); \
217         for (int i = 0; i < 4; i++) { \
218             gen_set_half_i64(i, RddV, gen_get_byte(byte, i, tmp, (SIGN)));  \
219         }  \
220     } while (0)
221 
222 #define fGEN_TCG_L2_loadbzw4_io(SHORTCODE) \
223     fGEN_TCG_loadbXw4(fEA_RI(RsV, siV), false)
224 #define fGEN_TCG_L4_loadbzw4_ur(SHORTCODE) \
225     fGEN_TCG_loadbXw4(fEA_IRs(UiV, RtV, uiV), false)
226 #define fGEN_TCG_L2_loadbsw4_io(SHORTCODE) \
227     fGEN_TCG_loadbXw4(fEA_RI(RsV, siV), true)
228 #define fGEN_TCG_L4_loadbsw4_ur(SHORTCODE) \
229     fGEN_TCG_loadbXw4(fEA_IRs(UiV, RtV, uiV), true)
230 #define fGEN_TCG_L2_loadbzw4_pci(SHORTCODE) \
231     fGEN_TCG_loadbXw4(GET_EA_pci, false)
232 #define fGEN_TCG_L2_loadbsw4_pci(SHORTCODE) \
233     fGEN_TCG_loadbXw4(GET_EA_pci, true)
234 #define fGEN_TCG_L2_loadbzw4_pcr(SHORTCODE) \
235     fGEN_TCG_loadbXw4(GET_EA_pcr(2), false)
236 #define fGEN_TCG_L2_loadbsw4_pcr(SHORTCODE) \
237     fGEN_TCG_loadbXw4(GET_EA_pcr(2), true)
238 #define fGEN_TCG_L4_loadbzw4_ap(SHORTCODE) \
239     fGEN_TCG_loadbXw4(GET_EA_ap, false)
240 #define fGEN_TCG_L2_loadbzw4_pr(SHORTCODE) \
241     fGEN_TCG_loadbXw4(GET_EA_pr, false)
242 #define fGEN_TCG_L2_loadbzw4_pbr(SHORTCODE) \
243     fGEN_TCG_loadbXw4(GET_EA_pbr, false)
244 #define fGEN_TCG_L2_loadbzw4_pi(SHORTCODE) \
245     fGEN_TCG_loadbXw4(GET_EA_pi, false)
246 #define fGEN_TCG_L4_loadbsw4_ap(SHORTCODE) \
247     fGEN_TCG_loadbXw4(GET_EA_ap, true)
248 #define fGEN_TCG_L2_loadbsw4_pr(SHORTCODE) \
249     fGEN_TCG_loadbXw4(GET_EA_pr, true)
250 #define fGEN_TCG_L2_loadbsw4_pbr(SHORTCODE) \
251     fGEN_TCG_loadbXw4(GET_EA_pbr, true)
252 #define fGEN_TCG_L2_loadbsw4_pi(SHORTCODE) \
253     fGEN_TCG_loadbXw4(GET_EA_pi, true)
254 
255 /*
256  * These instructions load a half word, shift the destination right by 16 bits
257  * and place the loaded value in the high half word of the destination pair.
258  * The GET_EA macro determines the addressing mode.
259  */
260 #define fGEN_TCG_loadalignh(GET_EA) \
261     do { \
262         TCGv tmp = tcg_temp_new(); \
263         TCGv_i64 tmp_i64 = tcg_temp_new_i64(); \
264         GET_EA;  \
265         fLOAD(1, 2, u, EA, tmp);  \
266         tcg_gen_extu_i32_i64(tmp_i64, tmp); \
267         tcg_gen_shri_i64(RyyV, RyyV, 16); \
268         tcg_gen_deposit_i64(RyyV, RyyV, tmp_i64, 48, 16); \
269     } while (0)
270 
271 #define fGEN_TCG_L4_loadalignh_ur(SHORTCODE) \
272     fGEN_TCG_loadalignh(fEA_IRs(UiV, RtV, uiV))
273 #define fGEN_TCG_L2_loadalignh_io(SHORTCODE) \
274     fGEN_TCG_loadalignh(fEA_RI(RsV, siV))
275 #define fGEN_TCG_L2_loadalignh_pci(SHORTCODE) \
276     fGEN_TCG_loadalignh(GET_EA_pci)
277 #define fGEN_TCG_L2_loadalignh_pcr(SHORTCODE) \
278     fGEN_TCG_loadalignh(GET_EA_pcr(1))
279 #define fGEN_TCG_L4_loadalignh_ap(SHORTCODE) \
280     fGEN_TCG_loadalignh(GET_EA_ap)
281 #define fGEN_TCG_L2_loadalignh_pr(SHORTCODE) \
282     fGEN_TCG_loadalignh(GET_EA_pr)
283 #define fGEN_TCG_L2_loadalignh_pbr(SHORTCODE) \
284     fGEN_TCG_loadalignh(GET_EA_pbr)
285 #define fGEN_TCG_L2_loadalignh_pi(SHORTCODE) \
286     fGEN_TCG_loadalignh(GET_EA_pi)
287 
288 /* Same as above, but loads a byte instead of half word */
289 #define fGEN_TCG_loadalignb(GET_EA) \
290     do { \
291         TCGv tmp = tcg_temp_new(); \
292         TCGv_i64 tmp_i64 = tcg_temp_new_i64(); \
293         GET_EA;  \
294         fLOAD(1, 1, u, EA, tmp);  \
295         tcg_gen_extu_i32_i64(tmp_i64, tmp); \
296         tcg_gen_shri_i64(RyyV, RyyV, 8); \
297         tcg_gen_deposit_i64(RyyV, RyyV, tmp_i64, 56, 8); \
298     } while (0)
299 
300 #define fGEN_TCG_L2_loadalignb_io(SHORTCODE) \
301     fGEN_TCG_loadalignb(fEA_RI(RsV, siV))
302 #define fGEN_TCG_L4_loadalignb_ur(SHORTCODE) \
303     fGEN_TCG_loadalignb(fEA_IRs(UiV, RtV, uiV))
304 #define fGEN_TCG_L2_loadalignb_pci(SHORTCODE) \
305     fGEN_TCG_loadalignb(GET_EA_pci)
306 #define fGEN_TCG_L2_loadalignb_pcr(SHORTCODE) \
307     fGEN_TCG_loadalignb(GET_EA_pcr(0))
308 #define fGEN_TCG_L4_loadalignb_ap(SHORTCODE) \
309     fGEN_TCG_loadalignb(GET_EA_ap)
310 #define fGEN_TCG_L2_loadalignb_pr(SHORTCODE) \
311     fGEN_TCG_loadalignb(GET_EA_pr)
312 #define fGEN_TCG_L2_loadalignb_pbr(SHORTCODE) \
313     fGEN_TCG_loadalignb(GET_EA_pbr)
314 #define fGEN_TCG_L2_loadalignb_pi(SHORTCODE) \
315     fGEN_TCG_loadalignb(GET_EA_pi)
316 
317 /*
318  * Predicated loads
319  * Here is a primer to understand the tag names
320  *
321  * Predicate used
322  *      t        true "old" value                  if (p0) r0 = memb(r2+#0)
323  *      f        false "old" value                 if (!p0) r0 = memb(r2+#0)
324  *      tnew     true "new" value                  if (p0.new) r0 = memb(r2+#0)
325  *      fnew     false "new" value                 if (!p0.new) r0 = memb(r2+#0)
326  */
327 #define fGEN_TCG_PRED_LOAD(GET_EA, PRED, SIZE, SIGN) \
328     do { \
329         TCGv LSB = tcg_temp_new(); \
330         TCGLabel *label = gen_new_label(); \
331         tcg_gen_movi_tl(EA, 0); \
332         PRED;  \
333         CHECK_NOSHUF_PRED(GET_EA, SIZE, LSB); \
334         tcg_gen_brcondi_tl(TCG_COND_EQ, LSB, 0, label); \
335         fLOAD(1, SIZE, SIGN, EA, RdV); \
336         gen_set_label(label); \
337     } while (0)
338 
339 #define fGEN_TCG_L2_ploadrubt_pi(SHORTCODE) \
340     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBOLD(PtV), 1, u)
341 #define fGEN_TCG_L2_ploadrubf_pi(SHORTCODE) \
342     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBOLDNOT(PtV), 1, u)
343 #define fGEN_TCG_L2_ploadrubtnew_pi(SHORTCODE) \
344     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBNEW(PtN), 1, u)
345 #define fGEN_TCG_L2_ploadrubfnew_pi(SHORTCODE) \
346     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBNEWNOT(PtN), 1, u)
347 #define fGEN_TCG_L2_ploadrbt_pi(SHORTCODE) \
348     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBOLD(PtV), 1, s)
349 #define fGEN_TCG_L2_ploadrbf_pi(SHORTCODE) \
350     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBOLDNOT(PtV), 1, s)
351 #define fGEN_TCG_L2_ploadrbtnew_pi(SHORTCODE) \
352     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBNEW(PtN), 1, s)
353 #define fGEN_TCG_L2_ploadrbfnew_pi(SHORTCODE) \
354     fGEN_TCG_PRED_LOAD({ fEA_REG(RxV); fPM_I(RxV, siV); }, \
355                        fLSBNEWNOT(PtN), 1, s)
356 
357 #define fGEN_TCG_L2_ploadruht_pi(SHORTCODE) \
358     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBOLD(PtV), 2, u)
359 #define fGEN_TCG_L2_ploadruhf_pi(SHORTCODE) \
360     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBOLDNOT(PtV), 2, u)
361 #define fGEN_TCG_L2_ploadruhtnew_pi(SHORTCODE) \
362     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBNEW(PtN), 2, u)
363 #define fGEN_TCG_L2_ploadruhfnew_pi(SHORTCODE) \
364     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBNEWNOT(PtN), 2, u)
365 #define fGEN_TCG_L2_ploadrht_pi(SHORTCODE) \
366     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBOLD(PtV), 2, s)
367 #define fGEN_TCG_L2_ploadrhf_pi(SHORTCODE) \
368     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBOLDNOT(PtV), 2, s)
369 #define fGEN_TCG_L2_ploadrhtnew_pi(SHORTCODE) \
370     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBNEW(PtN), 2, s)
371 #define fGEN_TCG_L2_ploadrhfnew_pi(SHORTCODE) \
372     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBNEWNOT(PtN), 2, s)
373 
374 #define fGEN_TCG_L2_ploadrit_pi(SHORTCODE) \
375     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBOLD(PtV), 4, u)
376 #define fGEN_TCG_L2_ploadrif_pi(SHORTCODE) \
377     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBOLDNOT(PtV), 4, u)
378 #define fGEN_TCG_L2_ploadritnew_pi(SHORTCODE) \
379     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBNEW(PtN), 4, u)
380 #define fGEN_TCG_L2_ploadrifnew_pi(SHORTCODE) \
381     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBNEWNOT(PtN), 4, u)
382 
383 /* Predicated loads into a register pair */
384 #define fGEN_TCG_PRED_LOAD_PAIR(GET_EA, PRED) \
385     do { \
386         TCGv LSB = tcg_temp_new(); \
387         TCGLabel *label = gen_new_label(); \
388         tcg_gen_movi_tl(EA, 0); \
389         PRED;  \
390         CHECK_NOSHUF_PRED(GET_EA, 8, LSB); \
391         tcg_gen_brcondi_tl(TCG_COND_EQ, LSB, 0, label); \
392         fLOAD(1, 8, u, EA, RddV); \
393         gen_set_label(label); \
394     } while (0)
395 
396 #define fGEN_TCG_L2_ploadrdt_pi(SHORTCODE) \
397     fGEN_TCG_PRED_LOAD_PAIR(GET_EA_pi, fLSBOLD(PtV))
398 #define fGEN_TCG_L2_ploadrdf_pi(SHORTCODE) \
399     fGEN_TCG_PRED_LOAD_PAIR(GET_EA_pi, fLSBOLDNOT(PtV))
400 #define fGEN_TCG_L2_ploadrdtnew_pi(SHORTCODE) \
401     fGEN_TCG_PRED_LOAD_PAIR(GET_EA_pi, fLSBNEW(PtN))
402 #define fGEN_TCG_L2_ploadrdfnew_pi(SHORTCODE) \
403     fGEN_TCG_PRED_LOAD_PAIR(GET_EA_pi, fLSBNEWNOT(PtN))
404 
405 /* load-locked and store-locked */
406 #define fGEN_TCG_L2_loadw_locked(SHORTCODE) \
407     SHORTCODE
408 #define fGEN_TCG_L4_loadd_locked(SHORTCODE) \
409     SHORTCODE
410 #define fGEN_TCG_S2_storew_locked(SHORTCODE) \
411     SHORTCODE
412 #define fGEN_TCG_S4_stored_locked(SHORTCODE) \
413     SHORTCODE
414 
415 #define fGEN_TCG_STORE(SHORTCODE) \
416     do { \
417         TCGv HALF G_GNUC_UNUSED = tcg_temp_new(); \
418         TCGv BYTE G_GNUC_UNUSED = tcg_temp_new(); \
419         SHORTCODE; \
420     } while (0)
421 
422 #define fGEN_TCG_STORE_pcr(SHIFT, STORE) \
423     do { \
424         TCGv ireg = tcg_temp_new(); \
425         TCGv HALF G_GNUC_UNUSED = tcg_temp_new(); \
426         TCGv BYTE G_GNUC_UNUSED = tcg_temp_new(); \
427         tcg_gen_mov_tl(EA, RxV); \
428         gen_read_ireg(ireg, MuV, SHIFT); \
429         gen_helper_fcircadd(RxV, RxV, ireg, MuV, CS); \
430         STORE; \
431     } while (0)
432 
433 #define fGEN_TCG_S2_storerb_pbr(SHORTCODE) \
434     fGEN_TCG_STORE(SHORTCODE)
435 #define fGEN_TCG_S2_storerb_pci(SHORTCODE) \
436     fGEN_TCG_STORE(SHORTCODE)
437 #define fGEN_TCG_S2_storerb_pcr(SHORTCODE) \
438     fGEN_TCG_STORE_pcr(0, fSTORE(1, 1, EA, fGETBYTE(0, RtV)))
439 
440 #define fGEN_TCG_S2_storerh_pbr(SHORTCODE) \
441     fGEN_TCG_STORE(SHORTCODE)
442 #define fGEN_TCG_S2_storerh_pci(SHORTCODE) \
443     fGEN_TCG_STORE(SHORTCODE)
444 #define fGEN_TCG_S2_storerh_pcr(SHORTCODE) \
445     fGEN_TCG_STORE_pcr(1, fSTORE(1, 2, EA, fGETHALF(0, RtV)))
446 
447 #define fGEN_TCG_S2_storerf_pbr(SHORTCODE) \
448     fGEN_TCG_STORE(SHORTCODE)
449 #define fGEN_TCG_S2_storerf_pci(SHORTCODE) \
450     fGEN_TCG_STORE(SHORTCODE)
451 #define fGEN_TCG_S2_storerf_pcr(SHORTCODE) \
452     fGEN_TCG_STORE_pcr(1, fSTORE(1, 2, EA, fGETHALF(1, RtV)))
453 
454 #define fGEN_TCG_S2_storeri_pbr(SHORTCODE) \
455     fGEN_TCG_STORE(SHORTCODE)
456 #define fGEN_TCG_S2_storeri_pci(SHORTCODE) \
457     fGEN_TCG_STORE(SHORTCODE)
458 #define fGEN_TCG_S2_storeri_pcr(SHORTCODE) \
459     fGEN_TCG_STORE_pcr(2, fSTORE(1, 4, EA, RtV))
460 
461 #define fGEN_TCG_S2_storerd_pbr(SHORTCODE) \
462     fGEN_TCG_STORE(SHORTCODE)
463 #define fGEN_TCG_S2_storerd_pci(SHORTCODE) \
464     fGEN_TCG_STORE(SHORTCODE)
465 #define fGEN_TCG_S2_storerd_pcr(SHORTCODE) \
466     fGEN_TCG_STORE_pcr(3, fSTORE(1, 8, EA, RttV))
467 
468 #define fGEN_TCG_S2_storerbnew_pbr(SHORTCODE) \
469     fGEN_TCG_STORE(SHORTCODE)
470 #define fGEN_TCG_S2_storerbnew_pci(SHORTCODE) \
471     fGEN_TCG_STORE(SHORTCODE)
472 #define fGEN_TCG_S2_storerbnew_pcr(SHORTCODE) \
473     fGEN_TCG_STORE_pcr(0, fSTORE(1, 1, EA, fGETBYTE(0, NtN)))
474 
475 #define fGEN_TCG_S2_storerhnew_pbr(SHORTCODE) \
476     fGEN_TCG_STORE(SHORTCODE)
477 #define fGEN_TCG_S2_storerhnew_pci(SHORTCODE) \
478     fGEN_TCG_STORE(SHORTCODE)
479 #define fGEN_TCG_S2_storerhnew_pcr(SHORTCODE) \
480     fGEN_TCG_STORE_pcr(1, fSTORE(1, 2, EA, fGETHALF(0, NtN)))
481 
482 #define fGEN_TCG_S2_storerinew_pbr(SHORTCODE) \
483     fGEN_TCG_STORE(SHORTCODE)
484 #define fGEN_TCG_S2_storerinew_pci(SHORTCODE) \
485     fGEN_TCG_STORE(SHORTCODE)
486 #define fGEN_TCG_S2_storerinew_pcr(SHORTCODE) \
487     fGEN_TCG_STORE_pcr(2, fSTORE(1, 4, EA, NtN))
488 
489 /* dczeroa clears the 32 byte cache line at the address given */
490 #define fGEN_TCG_Y2_dczeroa(SHORTCODE) SHORTCODE
491 
492 /* In linux-user mode, these are not modelled, suppress compiler warning */
493 #define fGEN_TCG_Y2_dcinva(SHORTCODE) \
494     do { RsV = RsV; } while (0)
495 #define fGEN_TCG_Y2_dccleaninva(SHORTCODE) \
496     do { RsV = RsV; } while (0)
497 #define fGEN_TCG_Y2_dccleana(SHORTCODE) \
498     do { RsV = RsV; } while (0)
499 #define fGEN_TCG_Y2_icinva(SHORTCODE) \
500     do { RsV = RsV; } while (0)
501 
502 /*
503  * allocframe(#uiV)
504  *     RxV == r29
505  */
506 #define fGEN_TCG_S2_allocframe(SHORTCODE) \
507     gen_allocframe(ctx, RxV, uiV)
508 
509 /* sub-instruction version (no RxV, so handle it manually) */
510 #define fGEN_TCG_SS2_allocframe(SHORTCODE) \
511     do { \
512         TCGv r29 = tcg_temp_new(); \
513         tcg_gen_mov_tl(r29, hex_gpr[HEX_REG_SP]); \
514         gen_allocframe(ctx, r29, uiV); \
515         gen_log_reg_write(ctx, HEX_REG_SP, r29); \
516     } while (0)
517 
518 /*
519  * Rdd32 = deallocframe(Rs32):raw
520  *     RddV == r31:30
521  *     RsV  == r30
522  */
523 #define fGEN_TCG_L2_deallocframe(SHORTCODE) \
524     gen_deallocframe(ctx, RddV, RsV)
525 
526 /* sub-instruction version (no RddV/RsV, so handle it manually) */
527 #define fGEN_TCG_SL2_deallocframe(SHORTCODE) \
528     do { \
529         TCGv_i64 r31_30 = tcg_temp_new_i64(); \
530         gen_deallocframe(ctx, r31_30, hex_gpr[HEX_REG_FP]); \
531         gen_log_reg_write_pair(ctx, HEX_REG_FP, r31_30); \
532     } while (0)
533 
534 /*
535  * dealloc_return
536  * Assembler mapped to
537  * r31:30 = dealloc_return(r30):raw
538  */
539 #define fGEN_TCG_L4_return(SHORTCODE) \
540     gen_return(ctx, RddV, RsV)
541 
542 /*
543  * sub-instruction version (no RddV, so handle it manually)
544  */
545 #define fGEN_TCG_SL2_return(SHORTCODE) \
546     do { \
547         TCGv_i64 RddV = get_result_gpr_pair(ctx, HEX_REG_FP); \
548         gen_return(ctx, RddV, hex_gpr[HEX_REG_FP]); \
549         gen_log_reg_write_pair(ctx, HEX_REG_FP, RddV); \
550     } while (0)
551 
552 /*
553  * Conditional returns follow this naming convention
554  *     _t                 predicate true
555  *     _f                 predicate false
556  *     _tnew_pt           predicate.new true predict taken
557  *     _fnew_pt           predicate.new false predict taken
558  *     _tnew_pnt          predicate.new true predict not taken
559  *     _fnew_pnt          predicate.new false predict not taken
560  * Predictions are not modelled in QEMU
561  *
562  * Example:
563  *     if (p1) r31:30 = dealloc_return(r30):raw
564  */
565 #define fGEN_TCG_L4_return_t(SHORTCODE) \
566     gen_cond_return(ctx, RddV, RsV, PvV, TCG_COND_EQ);
567 #define fGEN_TCG_L4_return_f(SHORTCODE) \
568     gen_cond_return(ctx, RddV, RsV, PvV, TCG_COND_NE)
569 #define fGEN_TCG_L4_return_tnew_pt(SHORTCODE) \
570     gen_cond_return(ctx, RddV, RsV, PvN, TCG_COND_EQ)
571 #define fGEN_TCG_L4_return_fnew_pt(SHORTCODE) \
572     gen_cond_return(ctx, RddV, RsV, PvN, TCG_COND_NE)
573 #define fGEN_TCG_L4_return_tnew_pnt(SHORTCODE) \
574     gen_cond_return(ctx, RddV, RsV, PvN, TCG_COND_EQ)
575 #define fGEN_TCG_L4_return_fnew_pnt(SHORTCODE) \
576     gen_cond_return(ctx, RddV, RsV, PvN, TCG_COND_NE)
577 
578 #define fGEN_TCG_SL2_return_t(SHORTCODE) \
579     gen_cond_return_subinsn(ctx, TCG_COND_EQ, hex_pred[0])
580 #define fGEN_TCG_SL2_return_f(SHORTCODE) \
581     gen_cond_return_subinsn(ctx, TCG_COND_NE, hex_pred[0])
582 #define fGEN_TCG_SL2_return_tnew(SHORTCODE) \
583     gen_cond_return_subinsn(ctx, TCG_COND_EQ, ctx->new_pred_value[0])
584 #define fGEN_TCG_SL2_return_fnew(SHORTCODE) \
585     gen_cond_return_subinsn(ctx, TCG_COND_NE, ctx->new_pred_value[0])
586 
587 /*
588  * Mathematical operations with more than one definition require
589  * special handling
590  */
591 #define fGEN_TCG_A5_ACS(SHORTCODE) \
592     do { \
593         gen_helper_vacsh_pred(PeV, tcg_env, RxxV, RssV, RttV); \
594         gen_helper_vacsh_val(RxxV, tcg_env, RxxV, RssV, RttV, \
595                              tcg_constant_tl(ctx->need_commit)); \
596     } while (0)
597 
598 #define fGEN_TCG_S2_cabacdecbin(SHORTCODE) \
599     do { \
600         TCGv p0 = tcg_temp_new(); \
601         gen_helper_cabacdecbin_pred(p0, RssV, RttV); \
602         gen_helper_cabacdecbin_val(RddV, RssV, RttV); \
603         gen_log_pred_write(ctx, 0, p0); \
604     } while (0)
605 
606 /*
607  * Approximate reciprocal
608  * r3,p1 = sfrecipa(r0, r1)
609  *
610  * The helper packs the 2 32-bit results into a 64-bit value,
611  * so unpack them into the proper results.
612  */
613 #define fGEN_TCG_F2_sfrecipa(SHORTCODE) \
614     do { \
615         TCGv_i64 tmp = tcg_temp_new_i64(); \
616         gen_helper_sfrecipa(tmp, tcg_env, RsV, RtV);  \
617         tcg_gen_extrh_i64_i32(RdV, tmp); \
618         tcg_gen_extrl_i64_i32(PeV, tmp); \
619     } while (0)
620 
621 /*
622  * Approximation of the reciprocal square root
623  * r1,p0 = sfinvsqrta(r0)
624  *
625  * The helper packs the 2 32-bit results into a 64-bit value,
626  * so unpack them into the proper results.
627  */
628 #define fGEN_TCG_F2_sfinvsqrta(SHORTCODE) \
629     do { \
630         TCGv_i64 tmp = tcg_temp_new_i64(); \
631         gen_helper_sfinvsqrta(tmp, tcg_env, RsV); \
632         tcg_gen_extrh_i64_i32(RdV, tmp); \
633         tcg_gen_extrl_i64_i32(PeV, tmp); \
634     } while (0)
635 
636 /*
637  * Add or subtract with carry.
638  * Predicate register is used as an extra input and output.
639  * r5:4 = add(r1:0, r3:2, p1):carry
640  */
641 #define fGEN_TCG_A4_addp_c(SHORTCODE) \
642     do { \
643         TCGv_i64 carry = tcg_temp_new_i64(); \
644         TCGv_i64 zero = tcg_constant_i64(0); \
645         tcg_gen_extu_i32_i64(carry, PxV); \
646         tcg_gen_andi_i64(carry, carry, 1); \
647         tcg_gen_add2_i64(RddV, carry, RssV, zero, carry, zero); \
648         tcg_gen_add2_i64(RddV, carry, RddV, carry, RttV, zero); \
649         tcg_gen_extrl_i64_i32(PxV, carry); \
650         gen_8bitsof(PxV, PxV); \
651     } while (0)
652 
653 /* r5:4 = sub(r1:0, r3:2, p1):carry */
654 #define fGEN_TCG_A4_subp_c(SHORTCODE) \
655     do { \
656         TCGv_i64 carry = tcg_temp_new_i64(); \
657         TCGv_i64 zero = tcg_constant_i64(0); \
658         TCGv_i64 not_RttV = tcg_temp_new_i64(); \
659         tcg_gen_extu_i32_i64(carry, PxV); \
660         tcg_gen_andi_i64(carry, carry, 1); \
661         tcg_gen_not_i64(not_RttV, RttV); \
662         tcg_gen_add2_i64(RddV, carry, RssV, zero, carry, zero); \
663         tcg_gen_add2_i64(RddV, carry, RddV, carry, not_RttV, zero); \
664         tcg_gen_extrl_i64_i32(PxV, carry); \
665         gen_8bitsof(PxV, PxV); \
666     } while (0)
667 
668 /*
669  * Compare each of the 8 unsigned bytes
670  * The minimum is placed in each byte of the destination.
671  * Each bit of the predicate is set true if the bit from the first operand
672  * is greater than the bit from the second operand.
673  * r5:4,p1 = vminub(r1:0, r3:2)
674  */
675 #define fGEN_TCG_A6_vminub_RdP(SHORTCODE) \
676     do { \
677         TCGv left = tcg_temp_new(); \
678         TCGv right = tcg_temp_new(); \
679         TCGv tmp = tcg_temp_new(); \
680         tcg_gen_movi_tl(PeV, 0); \
681         tcg_gen_movi_i64(RddV, 0); \
682         for (int i = 0; i < 8; i++) { \
683             gen_get_byte_i64(left, i, RttV, false); \
684             gen_get_byte_i64(right, i, RssV, false); \
685             tcg_gen_setcond_tl(TCG_COND_GT, tmp, left, right); \
686             tcg_gen_deposit_tl(PeV, PeV, tmp, i, 1); \
687             tcg_gen_umin_tl(tmp, left, right); \
688             gen_set_byte_i64(i, RddV, tmp); \
689         } \
690     } while (0)
691 
692 #define fGEN_TCG_J2_call(SHORTCODE) \
693     gen_call(ctx, riV)
694 #define fGEN_TCG_J2_callr(SHORTCODE) \
695     gen_callr(ctx, RsV)
696 #define fGEN_TCG_J2_callrh(SHORTCODE) \
697     gen_callr(ctx, RsV)
698 
699 #define fGEN_TCG_J2_callt(SHORTCODE) \
700     gen_cond_call(ctx, PuV, TCG_COND_EQ, riV)
701 #define fGEN_TCG_J2_callf(SHORTCODE) \
702     gen_cond_call(ctx, PuV, TCG_COND_NE, riV)
703 #define fGEN_TCG_J2_callrt(SHORTCODE) \
704     gen_cond_callr(ctx, TCG_COND_EQ, PuV, RsV)
705 #define fGEN_TCG_J2_callrf(SHORTCODE) \
706     gen_cond_callr(ctx, TCG_COND_NE, PuV, RsV)
707 
708 #define fGEN_TCG_J2_loop0r(SHORTCODE) \
709     gen_loop0r(ctx, RsV, riV)
710 #define fGEN_TCG_J2_loop1r(SHORTCODE) \
711     gen_loop1r(ctx, RsV, riV)
712 #define fGEN_TCG_J2_loop0i(SHORTCODE) \
713     gen_loop0i(ctx, UiV, riV)
714 #define fGEN_TCG_J2_loop1i(SHORTCODE) \
715     gen_loop1i(ctx, UiV, riV)
716 #define fGEN_TCG_J2_ploop1sr(SHORTCODE) \
717     gen_ploopNsr(ctx, 1, RsV, riV)
718 #define fGEN_TCG_J2_ploop1si(SHORTCODE) \
719     gen_ploopNsi(ctx, 1, UiV, riV)
720 #define fGEN_TCG_J2_ploop2sr(SHORTCODE) \
721     gen_ploopNsr(ctx, 2, RsV, riV)
722 #define fGEN_TCG_J2_ploop2si(SHORTCODE) \
723     gen_ploopNsi(ctx, 2, UiV, riV)
724 #define fGEN_TCG_J2_ploop3sr(SHORTCODE) \
725     gen_ploopNsr(ctx, 3, RsV, riV)
726 #define fGEN_TCG_J2_ploop3si(SHORTCODE) \
727     gen_ploopNsi(ctx, 3, UiV, riV)
728 
729 #define fGEN_TCG_J2_endloop0(SHORTCODE) \
730     gen_endloop0(ctx)
731 #define fGEN_TCG_J2_endloop1(SHORTCODE) \
732     gen_endloop1(ctx)
733 #define fGEN_TCG_J2_endloop01(SHORTCODE) \
734     gen_endloop01(ctx)
735 
736 /*
737  * Compound compare and jump instructions
738  * Here is a primer to understand the tag names
739  *
740  * Comparison
741  *      cmpeqi   compare equal to an immediate
742  *      cmpgti   compare greater than an immediate
743  *      cmpgtiu  compare greater than an unsigned immediate
744  *      cmpeqn1  compare equal to negative 1
745  *      cmpgtn1  compare greater than negative 1
746  *      cmpeq    compare equal (two registers)
747  *      cmpgtu   compare greater than unsigned (two registers)
748  *      tstbit0  test bit zero
749  *
750  * Condition
751  *      tp0      p0 is true     p0 = cmp.eq(r0,#5); if (p0.new) jump:nt address
752  *      fp0      p0 is false    p0 = cmp.eq(r0,#5); if (!p0.new) jump:nt address
753  *      tp1      p1 is true     p1 = cmp.eq(r0,#5); if (p1.new) jump:nt address
754  *      fp1      p1 is false    p1 = cmp.eq(r0,#5); if (!p1.new) jump:nt address
755  *
756  * Prediction (not modelled in qemu)
757  *      _nt      not taken
758  *      _t       taken
759  */
760 #define fGEN_TCG_J4_cmpeq_tp0_jump_t(SHORTCODE) \
761     gen_cmpnd_cmp_jmp_t(ctx, 0, TCG_COND_EQ, RsV, RtV, riV)
762 #define fGEN_TCG_J4_cmpeq_tp0_jump_nt(SHORTCODE) \
763     gen_cmpnd_cmp_jmp_t(ctx, 0, TCG_COND_EQ, RsV, RtV, riV)
764 #define fGEN_TCG_J4_cmpeq_fp0_jump_t(SHORTCODE) \
765     gen_cmpnd_cmp_jmp_f(ctx, 0, TCG_COND_EQ, RsV, RtV, riV)
766 #define fGEN_TCG_J4_cmpeq_fp0_jump_nt(SHORTCODE) \
767     gen_cmpnd_cmp_jmp_f(ctx, 0, TCG_COND_EQ, RsV, RtV, riV)
768 #define fGEN_TCG_J4_cmpeq_tp1_jump_t(SHORTCODE) \
769     gen_cmpnd_cmp_jmp_t(ctx, 1, TCG_COND_EQ, RsV, RtV, riV)
770 #define fGEN_TCG_J4_cmpeq_tp1_jump_nt(SHORTCODE) \
771     gen_cmpnd_cmp_jmp_t(ctx, 1, TCG_COND_EQ, RsV, RtV, riV)
772 #define fGEN_TCG_J4_cmpeq_fp1_jump_t(SHORTCODE) \
773     gen_cmpnd_cmp_jmp_f(ctx, 1, TCG_COND_EQ, RsV, RtV, riV)
774 #define fGEN_TCG_J4_cmpeq_fp1_jump_nt(SHORTCODE) \
775     gen_cmpnd_cmp_jmp_f(ctx, 1, TCG_COND_EQ, RsV, RtV, riV)
776 
777 #define fGEN_TCG_J4_cmpgt_tp0_jump_t(SHORTCODE) \
778     gen_cmpnd_cmp_jmp_t(ctx, 0, TCG_COND_GT, RsV, RtV, riV)
779 #define fGEN_TCG_J4_cmpgt_tp0_jump_nt(SHORTCODE) \
780     gen_cmpnd_cmp_jmp_t(ctx, 0, TCG_COND_GT, RsV, RtV, riV)
781 #define fGEN_TCG_J4_cmpgt_fp0_jump_t(SHORTCODE) \
782     gen_cmpnd_cmp_jmp_f(ctx, 0, TCG_COND_GT, RsV, RtV, riV)
783 #define fGEN_TCG_J4_cmpgt_fp0_jump_nt(SHORTCODE) \
784     gen_cmpnd_cmp_jmp_f(ctx, 0, TCG_COND_GT, RsV, RtV, riV)
785 #define fGEN_TCG_J4_cmpgt_tp1_jump_t(SHORTCODE) \
786     gen_cmpnd_cmp_jmp_t(ctx, 1, TCG_COND_GT, RsV, RtV, riV)
787 #define fGEN_TCG_J4_cmpgt_tp1_jump_nt(SHORTCODE) \
788     gen_cmpnd_cmp_jmp_t(ctx, 1, TCG_COND_GT, RsV, RtV, riV)
789 #define fGEN_TCG_J4_cmpgt_fp1_jump_t(SHORTCODE) \
790     gen_cmpnd_cmp_jmp_f(ctx, 1, TCG_COND_GT, RsV, RtV, riV)
791 #define fGEN_TCG_J4_cmpgt_fp1_jump_nt(SHORTCODE) \
792     gen_cmpnd_cmp_jmp_f(ctx, 1, TCG_COND_GT, RsV, RtV, riV)
793 
794 #define fGEN_TCG_J4_cmpgtu_tp0_jump_t(SHORTCODE) \
795     gen_cmpnd_cmp_jmp_t(ctx, 0, TCG_COND_GTU, RsV, RtV, riV)
796 #define fGEN_TCG_J4_cmpgtu_tp0_jump_nt(SHORTCODE) \
797     gen_cmpnd_cmp_jmp_t(ctx, 0, TCG_COND_GTU, RsV, RtV, riV)
798 #define fGEN_TCG_J4_cmpgtu_fp0_jump_t(SHORTCODE) \
799     gen_cmpnd_cmp_jmp_f(ctx, 0, TCG_COND_GTU, RsV, RtV, riV)
800 #define fGEN_TCG_J4_cmpgtu_fp0_jump_nt(SHORTCODE) \
801     gen_cmpnd_cmp_jmp_f(ctx, 0, TCG_COND_GTU, RsV, RtV, riV)
802 #define fGEN_TCG_J4_cmpgtu_tp1_jump_t(SHORTCODE) \
803     gen_cmpnd_cmp_jmp_t(ctx, 1, TCG_COND_GTU, RsV, RtV, riV)
804 #define fGEN_TCG_J4_cmpgtu_tp1_jump_nt(SHORTCODE) \
805     gen_cmpnd_cmp_jmp_t(ctx, 1, TCG_COND_GTU, RsV, RtV, riV)
806 #define fGEN_TCG_J4_cmpgtu_fp1_jump_t(SHORTCODE) \
807     gen_cmpnd_cmp_jmp_f(ctx, 1, TCG_COND_GTU, RsV, RtV, riV)
808 #define fGEN_TCG_J4_cmpgtu_fp1_jump_nt(SHORTCODE) \
809     gen_cmpnd_cmp_jmp_f(ctx, 1, TCG_COND_GTU, RsV, RtV, riV)
810 
811 #define fGEN_TCG_J4_cmpeqi_tp0_jump_t(SHORTCODE) \
812     gen_cmpnd_cmpi_jmp_t(ctx, 0, TCG_COND_EQ, RsV, UiV, riV)
813 #define fGEN_TCG_J4_cmpeqi_tp0_jump_nt(SHORTCODE) \
814     gen_cmpnd_cmpi_jmp_t(ctx, 0, TCG_COND_EQ, RsV, UiV, riV)
815 #define fGEN_TCG_J4_cmpeqi_fp0_jump_t(SHORTCODE) \
816     gen_cmpnd_cmpi_jmp_f(ctx, 0, TCG_COND_EQ, RsV, UiV, riV)
817 #define fGEN_TCG_J4_cmpeqi_fp0_jump_nt(SHORTCODE) \
818     gen_cmpnd_cmpi_jmp_f(ctx, 0, TCG_COND_EQ, RsV, UiV, riV)
819 #define fGEN_TCG_J4_cmpeqi_tp1_jump_t(SHORTCODE) \
820     gen_cmpnd_cmpi_jmp_t(ctx, 1, TCG_COND_EQ, RsV, UiV, riV)
821 #define fGEN_TCG_J4_cmpeqi_tp1_jump_nt(SHORTCODE) \
822     gen_cmpnd_cmpi_jmp_t(ctx, 1, TCG_COND_EQ, RsV, UiV, riV)
823 #define fGEN_TCG_J4_cmpeqi_fp1_jump_t(SHORTCODE) \
824     gen_cmpnd_cmpi_jmp_f(ctx, 1, TCG_COND_EQ, RsV, UiV, riV)
825 #define fGEN_TCG_J4_cmpeqi_fp1_jump_nt(SHORTCODE) \
826     gen_cmpnd_cmpi_jmp_f(ctx, 1, TCG_COND_EQ, RsV, UiV, riV)
827 
828 #define fGEN_TCG_J4_cmpgti_tp0_jump_t(SHORTCODE) \
829     gen_cmpnd_cmpi_jmp_t(ctx, 0, TCG_COND_GT, RsV, UiV, riV)
830 #define fGEN_TCG_J4_cmpgti_tp0_jump_nt(SHORTCODE) \
831     gen_cmpnd_cmpi_jmp_t(ctx, 0, TCG_COND_GT, RsV, UiV, riV)
832 #define fGEN_TCG_J4_cmpgti_fp0_jump_t(SHORTCODE) \
833     gen_cmpnd_cmpi_jmp_f(ctx, 0, TCG_COND_GT, RsV, UiV, riV)
834 #define fGEN_TCG_J4_cmpgti_fp0_jump_nt(SHORTCODE) \
835     gen_cmpnd_cmpi_jmp_f(ctx, 0, TCG_COND_GT, RsV, UiV, riV)
836 #define fGEN_TCG_J4_cmpgti_tp1_jump_t(SHORTCODE) \
837     gen_cmpnd_cmpi_jmp_t(ctx, 1, TCG_COND_GT, RsV, UiV, riV)
838 #define fGEN_TCG_J4_cmpgti_tp1_jump_nt(SHORTCODE) \
839     gen_cmpnd_cmpi_jmp_t(ctx, 1, TCG_COND_GT, RsV, UiV, riV)
840 #define fGEN_TCG_J4_cmpgti_fp1_jump_t(SHORTCODE) \
841     gen_cmpnd_cmpi_jmp_f(ctx, 1, TCG_COND_GT, RsV, UiV, riV)
842 #define fGEN_TCG_J4_cmpgti_fp1_jump_nt(SHORTCODE) \
843     gen_cmpnd_cmpi_jmp_f(ctx, 1, TCG_COND_GT, RsV, UiV, riV)
844 
845 #define fGEN_TCG_J4_cmpgtui_tp0_jump_t(SHORTCODE) \
846     gen_cmpnd_cmpi_jmp_t(ctx, 0, TCG_COND_GTU, RsV, UiV, riV)
847 #define fGEN_TCG_J4_cmpgtui_tp0_jump_nt(SHORTCODE) \
848     gen_cmpnd_cmpi_jmp_t(ctx, 0, TCG_COND_GTU, RsV, UiV, riV)
849 #define fGEN_TCG_J4_cmpgtui_fp0_jump_t(SHORTCODE) \
850     gen_cmpnd_cmpi_jmp_f(ctx, 0, TCG_COND_GTU, RsV, UiV, riV)
851 #define fGEN_TCG_J4_cmpgtui_fp0_jump_nt(SHORTCODE) \
852     gen_cmpnd_cmpi_jmp_f(ctx, 0, TCG_COND_GTU, RsV, UiV, riV)
853 #define fGEN_TCG_J4_cmpgtui_tp1_jump_t(SHORTCODE) \
854     gen_cmpnd_cmpi_jmp_t(ctx, 1, TCG_COND_GTU, RsV, UiV, riV)
855 #define fGEN_TCG_J4_cmpgtui_tp1_jump_nt(SHORTCODE) \
856     gen_cmpnd_cmpi_jmp_t(ctx, 1, TCG_COND_GTU, RsV, UiV, riV)
857 #define fGEN_TCG_J4_cmpgtui_fp1_jump_t(SHORTCODE) \
858     gen_cmpnd_cmpi_jmp_f(ctx, 1, TCG_COND_GTU, RsV, UiV, riV)
859 #define fGEN_TCG_J4_cmpgtui_fp1_jump_nt(SHORTCODE) \
860     gen_cmpnd_cmpi_jmp_f(ctx, 1, TCG_COND_GTU, RsV, UiV, riV)
861 
862 #define fGEN_TCG_J4_cmpeqn1_tp0_jump_t(SHORTCODE) \
863     gen_cmpnd_cmp_n1_jmp_t(ctx, 0, TCG_COND_EQ, RsV, riV)
864 #define fGEN_TCG_J4_cmpeqn1_tp0_jump_nt(SHORTCODE) \
865     gen_cmpnd_cmp_n1_jmp_t(ctx, 0, TCG_COND_EQ, RsV, riV)
866 #define fGEN_TCG_J4_cmpeqn1_fp0_jump_t(SHORTCODE) \
867     gen_cmpnd_cmp_n1_jmp_f(ctx, 0, TCG_COND_EQ, RsV, riV)
868 #define fGEN_TCG_J4_cmpeqn1_fp0_jump_nt(SHORTCODE) \
869     gen_cmpnd_cmp_n1_jmp_f(ctx, 0, TCG_COND_EQ, RsV, riV)
870 #define fGEN_TCG_J4_cmpeqn1_tp1_jump_t(SHORTCODE) \
871     gen_cmpnd_cmp_n1_jmp_t(ctx, 1, TCG_COND_EQ, RsV, riV)
872 #define fGEN_TCG_J4_cmpeqn1_tp1_jump_nt(SHORTCODE) \
873     gen_cmpnd_cmp_n1_jmp_t(ctx, 1, TCG_COND_EQ, RsV, riV)
874 #define fGEN_TCG_J4_cmpeqn1_fp1_jump_t(SHORTCODE) \
875     gen_cmpnd_cmp_n1_jmp_f(ctx, 1, TCG_COND_EQ, RsV, riV)
876 #define fGEN_TCG_J4_cmpeqn1_fp1_jump_nt(SHORTCODE) \
877     gen_cmpnd_cmp_n1_jmp_f(ctx, 1, TCG_COND_EQ, RsV, riV)
878 
879 #define fGEN_TCG_J4_cmpgtn1_tp0_jump_t(SHORTCODE) \
880     gen_cmpnd_cmp_n1_jmp_t(ctx, 0, TCG_COND_GT, RsV, riV)
881 #define fGEN_TCG_J4_cmpgtn1_tp0_jump_nt(SHORTCODE) \
882     gen_cmpnd_cmp_n1_jmp_t(ctx, 0, TCG_COND_GT, RsV, riV)
883 #define fGEN_TCG_J4_cmpgtn1_fp0_jump_t(SHORTCODE) \
884     gen_cmpnd_cmp_n1_jmp_f(ctx, 0, TCG_COND_GT, RsV, riV)
885 #define fGEN_TCG_J4_cmpgtn1_fp0_jump_nt(SHORTCODE) \
886     gen_cmpnd_cmp_n1_jmp_f(ctx, 0, TCG_COND_GT, RsV, riV)
887 #define fGEN_TCG_J4_cmpgtn1_tp1_jump_t(SHORTCODE) \
888     gen_cmpnd_cmp_n1_jmp_t(ctx, 1, TCG_COND_GT, RsV, riV)
889 #define fGEN_TCG_J4_cmpgtn1_tp1_jump_nt(SHORTCODE) \
890     gen_cmpnd_cmp_n1_jmp_t(ctx, 1, TCG_COND_GT, RsV, riV)
891 #define fGEN_TCG_J4_cmpgtn1_fp1_jump_t(SHORTCODE) \
892     gen_cmpnd_cmp_n1_jmp_f(ctx, 1, TCG_COND_GT, RsV, riV)
893 #define fGEN_TCG_J4_cmpgtn1_fp1_jump_nt(SHORTCODE) \
894     gen_cmpnd_cmp_n1_jmp_f(ctx, 1, TCG_COND_GT, RsV, riV)
895 
896 #define fGEN_TCG_J4_tstbit0_tp0_jump_nt(SHORTCODE) \
897     gen_cmpnd_tstbit0_jmp(ctx, 0, RsV, TCG_COND_EQ, riV)
898 #define fGEN_TCG_J4_tstbit0_tp0_jump_t(SHORTCODE) \
899     gen_cmpnd_tstbit0_jmp(ctx, 0, RsV, TCG_COND_EQ, riV)
900 #define fGEN_TCG_J4_tstbit0_fp0_jump_nt(SHORTCODE) \
901     gen_cmpnd_tstbit0_jmp(ctx, 0, RsV, TCG_COND_NE, riV)
902 #define fGEN_TCG_J4_tstbit0_fp0_jump_t(SHORTCODE) \
903     gen_cmpnd_tstbit0_jmp(ctx, 0, RsV, TCG_COND_NE, riV)
904 #define fGEN_TCG_J4_tstbit0_tp1_jump_nt(SHORTCODE) \
905     gen_cmpnd_tstbit0_jmp(ctx, 1, RsV, TCG_COND_EQ, riV)
906 #define fGEN_TCG_J4_tstbit0_tp1_jump_t(SHORTCODE) \
907     gen_cmpnd_tstbit0_jmp(ctx, 1, RsV, TCG_COND_EQ, riV)
908 #define fGEN_TCG_J4_tstbit0_fp1_jump_nt(SHORTCODE) \
909     gen_cmpnd_tstbit0_jmp(ctx, 1, RsV, TCG_COND_NE, riV)
910 #define fGEN_TCG_J4_tstbit0_fp1_jump_t(SHORTCODE) \
911     gen_cmpnd_tstbit0_jmp(ctx, 1, RsV, TCG_COND_NE, riV)
912 
913 /* p0 = cmp.eq(r0, #7) */
914 #define fGEN_TCG_SA1_cmpeqi(SHORTCODE) \
915     do { \
916         TCGv p0 = tcg_temp_new(); \
917         gen_comparei(TCG_COND_EQ, p0, RsV, uiV); \
918         gen_log_pred_write(ctx, 0, p0); \
919     } while (0)
920 
921 #define fGEN_TCG_J2_jump(SHORTCODE) \
922     gen_jump(ctx, riV)
923 #define fGEN_TCG_J2_jumpr(SHORTCODE) \
924     gen_jumpr(ctx, RsV)
925 #define fGEN_TCG_J2_jumprh(SHORTCODE) \
926     gen_jumpr(ctx, RsV)
927 #define fGEN_TCG_J4_jumpseti(SHORTCODE) \
928     do { \
929         tcg_gen_movi_tl(RdV, UiV); \
930         gen_jump(ctx, riV); \
931     } while (0)
932 
933 #define fGEN_TCG_cond_jumpt(COND) \
934     do { \
935         TCGv LSB = tcg_temp_new(); \
936         COND; \
937         gen_cond_jump(ctx, TCG_COND_EQ, LSB, riV); \
938     } while (0)
939 #define fGEN_TCG_cond_jumpf(COND) \
940     do { \
941         TCGv LSB = tcg_temp_new(); \
942         COND; \
943         gen_cond_jump(ctx, TCG_COND_NE, LSB, riV); \
944     } while (0)
945 
946 #define fGEN_TCG_J2_jumpt(SHORTCODE) \
947     fGEN_TCG_cond_jumpt(fLSBOLD(PuV))
948 #define fGEN_TCG_J2_jumptpt(SHORTCODE) \
949     fGEN_TCG_cond_jumpt(fLSBOLD(PuV))
950 #define fGEN_TCG_J2_jumpf(SHORTCODE) \
951     fGEN_TCG_cond_jumpf(fLSBOLD(PuV))
952 #define fGEN_TCG_J2_jumpfpt(SHORTCODE) \
953     fGEN_TCG_cond_jumpf(fLSBOLD(PuV))
954 #define fGEN_TCG_J2_jumptnew(SHORTCODE) \
955     gen_cond_jump(ctx, TCG_COND_EQ, PuN, riV)
956 #define fGEN_TCG_J2_jumptnewpt(SHORTCODE) \
957     gen_cond_jump(ctx, TCG_COND_EQ, PuN, riV)
958 #define fGEN_TCG_J2_jumpfnewpt(SHORTCODE) \
959     fGEN_TCG_cond_jumpf(fLSBNEW(PuN))
960 #define fGEN_TCG_J2_jumpfnew(SHORTCODE) \
961     fGEN_TCG_cond_jumpf(fLSBNEW(PuN))
962 #define fGEN_TCG_J2_jumprz(SHORTCODE) \
963     fGEN_TCG_cond_jumpt(tcg_gen_setcondi_tl(TCG_COND_NE, LSB, RsV, 0))
964 #define fGEN_TCG_J2_jumprzpt(SHORTCODE) \
965     fGEN_TCG_cond_jumpt(tcg_gen_setcondi_tl(TCG_COND_NE, LSB, RsV, 0))
966 #define fGEN_TCG_J2_jumprnz(SHORTCODE) \
967     fGEN_TCG_cond_jumpt(tcg_gen_setcondi_tl(TCG_COND_EQ, LSB, RsV, 0))
968 #define fGEN_TCG_J2_jumprnzpt(SHORTCODE) \
969     fGEN_TCG_cond_jumpt(tcg_gen_setcondi_tl(TCG_COND_EQ, LSB, RsV, 0))
970 #define fGEN_TCG_J2_jumprgtez(SHORTCODE) \
971     fGEN_TCG_cond_jumpt(tcg_gen_setcondi_tl(TCG_COND_GE, LSB, RsV, 0))
972 #define fGEN_TCG_J2_jumprgtezpt(SHORTCODE) \
973     fGEN_TCG_cond_jumpt(tcg_gen_setcondi_tl(TCG_COND_GE, LSB, RsV, 0))
974 #define fGEN_TCG_J2_jumprltez(SHORTCODE) \
975     fGEN_TCG_cond_jumpt(tcg_gen_setcondi_tl(TCG_COND_LE, LSB, RsV, 0))
976 #define fGEN_TCG_J2_jumprltezpt(SHORTCODE) \
977     fGEN_TCG_cond_jumpt(tcg_gen_setcondi_tl(TCG_COND_LE, LSB, RsV, 0))
978 
979 #define fGEN_TCG_cond_jumprt(COND) \
980     do { \
981         TCGv LSB = tcg_temp_new(); \
982         COND; \
983         gen_cond_jumpr(ctx, RsV, TCG_COND_EQ, LSB); \
984     } while (0)
985 #define fGEN_TCG_cond_jumprf(COND) \
986     do { \
987         TCGv LSB = tcg_temp_new(); \
988         COND; \
989         gen_cond_jumpr(ctx, RsV, TCG_COND_NE, LSB); \
990     } while (0)
991 
992 #define fGEN_TCG_J2_jumprt(SHORTCODE) \
993     fGEN_TCG_cond_jumprt(fLSBOLD(PuV))
994 #define fGEN_TCG_J2_jumprtpt(SHORTCODE) \
995     fGEN_TCG_cond_jumprt(fLSBOLD(PuV))
996 #define fGEN_TCG_J2_jumprf(SHORTCODE) \
997     fGEN_TCG_cond_jumprf(fLSBOLD(PuV))
998 #define fGEN_TCG_J2_jumprfpt(SHORTCODE) \
999     fGEN_TCG_cond_jumprf(fLSBOLD(PuV))
1000 #define fGEN_TCG_J2_jumprtnew(SHORTCODE) \
1001     fGEN_TCG_cond_jumprt(fLSBNEW(PuN))
1002 #define fGEN_TCG_J2_jumprtnewpt(SHORTCODE) \
1003     fGEN_TCG_cond_jumprt(fLSBNEW(PuN))
1004 #define fGEN_TCG_J2_jumprfnew(SHORTCODE) \
1005     fGEN_TCG_cond_jumprf(fLSBNEW(PuN))
1006 #define fGEN_TCG_J2_jumprfnewpt(SHORTCODE) \
1007     fGEN_TCG_cond_jumprf(fLSBNEW(PuN))
1008 
1009 /*
1010  * New value compare & jump instructions
1011  * if ([!]COND(r0.new, r1) jump:t address
1012  * if ([!]COND(r0.new, #7) jump:t address
1013  */
1014 #define fGEN_TCG_J4_cmpgt_t_jumpnv_t(SHORTCODE) \
1015     gen_cmp_jumpnv(ctx, TCG_COND_GT, NsN, RtV, riV)
1016 #define fGEN_TCG_J4_cmpgt_t_jumpnv_nt(SHORTCODE) \
1017     gen_cmp_jumpnv(ctx, TCG_COND_GT, NsN, RtV, riV)
1018 #define fGEN_TCG_J4_cmpgt_f_jumpnv_t(SHORTCODE) \
1019     gen_cmp_jumpnv(ctx, TCG_COND_LE, NsN, RtV, riV)
1020 #define fGEN_TCG_J4_cmpgt_f_jumpnv_nt(SHORTCODE) \
1021     gen_cmp_jumpnv(ctx, TCG_COND_LE, NsN, RtV, riV)
1022 
1023 #define fGEN_TCG_J4_cmpeq_t_jumpnv_t(SHORTCODE) \
1024     gen_cmp_jumpnv(ctx, TCG_COND_EQ, NsN, RtV, riV)
1025 #define fGEN_TCG_J4_cmpeq_t_jumpnv_nt(SHORTCODE) \
1026     gen_cmp_jumpnv(ctx, TCG_COND_EQ, NsN, RtV, riV)
1027 #define fGEN_TCG_J4_cmpeq_f_jumpnv_t(SHORTCODE) \
1028     gen_cmp_jumpnv(ctx, TCG_COND_NE, NsN, RtV, riV)
1029 #define fGEN_TCG_J4_cmpeq_f_jumpnv_nt(SHORTCODE) \
1030     gen_cmp_jumpnv(ctx, TCG_COND_NE, NsN, RtV, riV)
1031 
1032 #define fGEN_TCG_J4_cmplt_t_jumpnv_t(SHORTCODE) \
1033     gen_cmp_jumpnv(ctx, TCG_COND_LT, NsN, RtV, riV)
1034 #define fGEN_TCG_J4_cmplt_t_jumpnv_nt(SHORTCODE) \
1035     gen_cmp_jumpnv(ctx, TCG_COND_LT, NsN, RtV, riV)
1036 #define fGEN_TCG_J4_cmplt_f_jumpnv_t(SHORTCODE) \
1037     gen_cmp_jumpnv(ctx, TCG_COND_GE, NsN, RtV, riV)
1038 #define fGEN_TCG_J4_cmplt_f_jumpnv_nt(SHORTCODE) \
1039     gen_cmp_jumpnv(ctx, TCG_COND_GE, NsN, RtV, riV)
1040 
1041 #define fGEN_TCG_J4_cmpeqi_t_jumpnv_t(SHORTCODE) \
1042     gen_cmpi_jumpnv(ctx, TCG_COND_EQ, NsN, UiV, riV)
1043 #define fGEN_TCG_J4_cmpeqi_t_jumpnv_nt(SHORTCODE) \
1044     gen_cmpi_jumpnv(ctx, TCG_COND_EQ, NsN, UiV, riV)
1045 #define fGEN_TCG_J4_cmpeqi_f_jumpnv_t(SHORTCODE) \
1046     gen_cmpi_jumpnv(ctx, TCG_COND_NE, NsN, UiV, riV)
1047 #define fGEN_TCG_J4_cmpeqi_f_jumpnv_nt(SHORTCODE) \
1048     gen_cmpi_jumpnv(ctx, TCG_COND_NE, NsN, UiV, riV)
1049 
1050 #define fGEN_TCG_J4_cmpgti_t_jumpnv_t(SHORTCODE) \
1051     gen_cmpi_jumpnv(ctx, TCG_COND_GT, NsN, UiV, riV)
1052 #define fGEN_TCG_J4_cmpgti_t_jumpnv_nt(SHORTCODE) \
1053     gen_cmpi_jumpnv(ctx, TCG_COND_GT, NsN, UiV, riV)
1054 #define fGEN_TCG_J4_cmpgti_f_jumpnv_t(SHORTCODE) \
1055     gen_cmpi_jumpnv(ctx, TCG_COND_LE, NsN, UiV, riV)
1056 #define fGEN_TCG_J4_cmpgti_f_jumpnv_nt(SHORTCODE) \
1057     gen_cmpi_jumpnv(ctx, TCG_COND_LE, NsN, UiV, riV)
1058 
1059 #define fGEN_TCG_J4_cmpltu_t_jumpnv_t(SHORTCODE) \
1060     gen_cmp_jumpnv(ctx, TCG_COND_LTU, NsN, RtV, riV)
1061 #define fGEN_TCG_J4_cmpltu_t_jumpnv_nt(SHORTCODE) \
1062     gen_cmp_jumpnv(ctx, TCG_COND_LTU, NsN, RtV, riV)
1063 #define fGEN_TCG_J4_cmpltu_f_jumpnv_t(SHORTCODE) \
1064     gen_cmp_jumpnv(ctx, TCG_COND_GEU, NsN, RtV, riV)
1065 #define fGEN_TCG_J4_cmpltu_f_jumpnv_nt(SHORTCODE) \
1066     gen_cmp_jumpnv(ctx, TCG_COND_GEU, NsN, RtV, riV)
1067 
1068 #define fGEN_TCG_J4_cmpgtui_t_jumpnv_t(SHORTCODE) \
1069     gen_cmpi_jumpnv(ctx, TCG_COND_GTU, NsN, UiV, riV)
1070 #define fGEN_TCG_J4_cmpgtui_t_jumpnv_nt(SHORTCODE) \
1071     gen_cmpi_jumpnv(ctx, TCG_COND_GTU, NsN, UiV, riV)
1072 #define fGEN_TCG_J4_cmpgtui_f_jumpnv_t(SHORTCODE) \
1073     gen_cmpi_jumpnv(ctx, TCG_COND_LEU, NsN, UiV, riV)
1074 #define fGEN_TCG_J4_cmpgtui_f_jumpnv_nt(SHORTCODE) \
1075     gen_cmpi_jumpnv(ctx, TCG_COND_LEU, NsN, UiV, riV)
1076 
1077 #define fGEN_TCG_J4_cmpgtu_t_jumpnv_t(SHORTCODE) \
1078     gen_cmp_jumpnv(ctx, TCG_COND_GTU, NsN, RtV, riV)
1079 #define fGEN_TCG_J4_cmpgtu_t_jumpnv_nt(SHORTCODE) \
1080     gen_cmp_jumpnv(ctx, TCG_COND_GTU, NsN, RtV, riV)
1081 #define fGEN_TCG_J4_cmpgtu_f_jumpnv_t(SHORTCODE) \
1082     gen_cmp_jumpnv(ctx, TCG_COND_LEU, NsN, RtV, riV)
1083 #define fGEN_TCG_J4_cmpgtu_f_jumpnv_nt(SHORTCODE) \
1084     gen_cmp_jumpnv(ctx, TCG_COND_LEU, NsN, RtV, riV)
1085 
1086 #define fGEN_TCG_J4_cmpeqn1_t_jumpnv_t(SHORTCODE) \
1087     gen_cmpi_jumpnv(ctx, TCG_COND_EQ, NsN, -1, riV)
1088 #define fGEN_TCG_J4_cmpeqn1_t_jumpnv_nt(SHORTCODE) \
1089     gen_cmpi_jumpnv(ctx, TCG_COND_EQ, NsN, -1, riV)
1090 #define fGEN_TCG_J4_cmpeqn1_f_jumpnv_t(SHORTCODE) \
1091     gen_cmpi_jumpnv(ctx, TCG_COND_NE, NsN, -1, riV)
1092 #define fGEN_TCG_J4_cmpeqn1_f_jumpnv_nt(SHORTCODE) \
1093     gen_cmpi_jumpnv(ctx, TCG_COND_NE, NsN, -1, riV)
1094 
1095 #define fGEN_TCG_J4_cmpgtn1_t_jumpnv_t(SHORTCODE) \
1096     gen_cmpi_jumpnv(ctx, TCG_COND_GT, NsN, -1, riV)
1097 #define fGEN_TCG_J4_cmpgtn1_t_jumpnv_nt(SHORTCODE) \
1098     gen_cmpi_jumpnv(ctx, TCG_COND_GT, NsN, -1, riV)
1099 #define fGEN_TCG_J4_cmpgtn1_f_jumpnv_t(SHORTCODE) \
1100     gen_cmpi_jumpnv(ctx, TCG_COND_LE, NsN, -1, riV)
1101 #define fGEN_TCG_J4_cmpgtn1_f_jumpnv_nt(SHORTCODE) \
1102     gen_cmpi_jumpnv(ctx, TCG_COND_LE, NsN, -1, riV)
1103 
1104 #define fGEN_TCG_J4_tstbit0_t_jumpnv_t(SHORTCODE) \
1105     gen_testbit0_jumpnv(ctx, NsN, TCG_COND_EQ, riV)
1106 #define fGEN_TCG_J4_tstbit0_t_jumpnv_nt(SHORTCODE) \
1107     gen_testbit0_jumpnv(ctx, NsN, TCG_COND_EQ, riV)
1108 #define fGEN_TCG_J4_tstbit0_f_jumpnv_t(SHORTCODE) \
1109     gen_testbit0_jumpnv(ctx, NsN, TCG_COND_NE, riV)
1110 #define fGEN_TCG_J4_tstbit0_f_jumpnv_nt(SHORTCODE) \
1111     gen_testbit0_jumpnv(ctx, NsN, TCG_COND_NE, riV)
1112 
1113 /* r0 = r1 ; jump address */
1114 #define fGEN_TCG_J4_jumpsetr(SHORTCODE) \
1115     do { \
1116         tcg_gen_mov_tl(RdV, RsV); \
1117         gen_jump(ctx, riV); \
1118     } while (0)
1119 
1120 /* if (p0.new) r0 = #0 */
1121 #define fGEN_TCG_SA1_clrtnew(SHORTCODE) \
1122     do { \
1123         tcg_gen_movcond_tl(TCG_COND_EQ, RdV, \
1124                            ctx->new_pred_value[0], tcg_constant_tl(0), \
1125                            RdV, tcg_constant_tl(0)); \
1126     } while (0)
1127 
1128 /* if (!p0.new) r0 = #0 */
1129 #define fGEN_TCG_SA1_clrfnew(SHORTCODE) \
1130     do { \
1131         tcg_gen_movcond_tl(TCG_COND_NE, RdV, \
1132                            ctx->new_pred_value[0], tcg_constant_tl(0), \
1133                            RdV, tcg_constant_tl(0)); \
1134     } while (0)
1135 
1136 #define fGEN_TCG_J2_pause(SHORTCODE) \
1137     do { \
1138         uiV = uiV; \
1139         tcg_gen_movi_tl(hex_gpr[HEX_REG_PC], ctx->next_PC); \
1140     } while (0)
1141 
1142 /* r0 = asr(r1, r2):sat */
1143 #define fGEN_TCG_S2_asr_r_r_sat(SHORTCODE) \
1144     gen_asr_r_r_sat(ctx, RdV, RsV, RtV)
1145 
1146 /* r0 = asl(r1, r2):sat */
1147 #define fGEN_TCG_S2_asl_r_r_sat(SHORTCODE) \
1148     gen_asl_r_r_sat(ctx, RdV, RsV, RtV)
1149 
1150 #define fGEN_TCG_SL2_jumpr31(SHORTCODE) \
1151     gen_jumpr(ctx, hex_gpr[HEX_REG_LR])
1152 
1153 #define fGEN_TCG_SL2_jumpr31_t(SHORTCODE) \
1154     gen_cond_jumpr31(ctx, TCG_COND_EQ, hex_pred[0])
1155 #define fGEN_TCG_SL2_jumpr31_f(SHORTCODE) \
1156     gen_cond_jumpr31(ctx, TCG_COND_NE, hex_pred[0])
1157 
1158 #define fGEN_TCG_SL2_jumpr31_tnew(SHORTCODE) \
1159     gen_cond_jumpr31(ctx, TCG_COND_EQ, ctx->new_pred_value[0])
1160 #define fGEN_TCG_SL2_jumpr31_fnew(SHORTCODE) \
1161     gen_cond_jumpr31(ctx, TCG_COND_NE, ctx->new_pred_value[0])
1162 
1163 /* Count trailing zeros/ones */
1164 #define fGEN_TCG_S2_ct0(SHORTCODE) \
1165     do { \
1166         tcg_gen_ctzi_tl(RdV, RsV, 32); \
1167     } while (0)
1168 #define fGEN_TCG_S2_ct1(SHORTCODE) \
1169     do { \
1170         tcg_gen_not_tl(RdV, RsV); \
1171         tcg_gen_ctzi_tl(RdV, RdV, 32); \
1172     } while (0)
1173 #define fGEN_TCG_S2_ct0p(SHORTCODE) \
1174     do { \
1175         TCGv_i64 tmp = tcg_temp_new_i64(); \
1176         tcg_gen_ctzi_i64(tmp, RssV, 64); \
1177         tcg_gen_extrl_i64_i32(RdV, tmp); \
1178     } while (0)
1179 #define fGEN_TCG_S2_ct1p(SHORTCODE) \
1180     do { \
1181         TCGv_i64 tmp = tcg_temp_new_i64(); \
1182         tcg_gen_not_i64(tmp, RssV); \
1183         tcg_gen_ctzi_i64(tmp, tmp, 64); \
1184         tcg_gen_extrl_i64_i32(RdV, tmp); \
1185     } while (0)
1186 
1187 #define fGEN_TCG_S2_insert(SHORTCODE) \
1188     do { \
1189         int width = uiV; \
1190         int offset = UiV; \
1191         if (width != 0) { \
1192             if (offset + width > 32) { \
1193                 width = 32 - offset; \
1194             } \
1195             tcg_gen_deposit_tl(RxV, RxV, RsV, offset, width); \
1196         } \
1197     } while (0)
1198 #define fGEN_TCG_S2_insert_rp(SHORTCODE) \
1199     gen_insert_rp(ctx, RxV, RsV, RttV)
1200 #define fGEN_TCG_S2_asr_r_svw_trun(SHORTCODE) \
1201     gen_asr_r_svw_trun(ctx, RdV, RssV, RtV)
1202 #define fGEN_TCG_A2_swiz(SHORTCODE) \
1203     tcg_gen_bswap_tl(RdV, RsV)
1204 
1205 /* Floating point */
1206 #define fGEN_TCG_F2_conv_sf2df(SHORTCODE) \
1207     gen_helper_conv_sf2df(RddV, tcg_env, RsV)
1208 #define fGEN_TCG_F2_conv_df2sf(SHORTCODE) \
1209     gen_helper_conv_df2sf(RdV, tcg_env, RssV)
1210 #define fGEN_TCG_F2_conv_uw2sf(SHORTCODE) \
1211     gen_helper_conv_uw2sf(RdV, tcg_env, RsV)
1212 #define fGEN_TCG_F2_conv_uw2df(SHORTCODE) \
1213     gen_helper_conv_uw2df(RddV, tcg_env, RsV)
1214 #define fGEN_TCG_F2_conv_w2sf(SHORTCODE) \
1215     gen_helper_conv_w2sf(RdV, tcg_env, RsV)
1216 #define fGEN_TCG_F2_conv_w2df(SHORTCODE) \
1217     gen_helper_conv_w2df(RddV, tcg_env, RsV)
1218 #define fGEN_TCG_F2_conv_ud2sf(SHORTCODE) \
1219     gen_helper_conv_ud2sf(RdV, tcg_env, RssV)
1220 #define fGEN_TCG_F2_conv_ud2df(SHORTCODE) \
1221     gen_helper_conv_ud2df(RddV, tcg_env, RssV)
1222 #define fGEN_TCG_F2_conv_d2sf(SHORTCODE) \
1223     gen_helper_conv_d2sf(RdV, tcg_env, RssV)
1224 #define fGEN_TCG_F2_conv_d2df(SHORTCODE) \
1225     gen_helper_conv_d2df(RddV, tcg_env, RssV)
1226 #define fGEN_TCG_F2_conv_sf2uw(SHORTCODE) \
1227     gen_helper_conv_sf2uw(RdV, tcg_env, RsV)
1228 #define fGEN_TCG_F2_conv_sf2w(SHORTCODE) \
1229     gen_helper_conv_sf2w(RdV, tcg_env, RsV)
1230 #define fGEN_TCG_F2_conv_sf2ud(SHORTCODE) \
1231     gen_helper_conv_sf2ud(RddV, tcg_env, RsV)
1232 #define fGEN_TCG_F2_conv_sf2d(SHORTCODE) \
1233     gen_helper_conv_sf2d(RddV, tcg_env, RsV)
1234 #define fGEN_TCG_F2_conv_df2uw(SHORTCODE) \
1235     gen_helper_conv_df2uw(RdV, tcg_env, RssV)
1236 #define fGEN_TCG_F2_conv_df2w(SHORTCODE) \
1237     gen_helper_conv_df2w(RdV, tcg_env, RssV)
1238 #define fGEN_TCG_F2_conv_df2ud(SHORTCODE) \
1239     gen_helper_conv_df2ud(RddV, tcg_env, RssV)
1240 #define fGEN_TCG_F2_conv_df2d(SHORTCODE) \
1241     gen_helper_conv_df2d(RddV, tcg_env, RssV)
1242 #define fGEN_TCG_F2_conv_sf2uw_chop(SHORTCODE) \
1243     gen_helper_conv_sf2uw_chop(RdV, tcg_env, RsV)
1244 #define fGEN_TCG_F2_conv_sf2w_chop(SHORTCODE) \
1245     gen_helper_conv_sf2w_chop(RdV, tcg_env, RsV)
1246 #define fGEN_TCG_F2_conv_sf2ud_chop(SHORTCODE) \
1247     gen_helper_conv_sf2ud_chop(RddV, tcg_env, RsV)
1248 #define fGEN_TCG_F2_conv_sf2d_chop(SHORTCODE) \
1249     gen_helper_conv_sf2d_chop(RddV, tcg_env, RsV)
1250 #define fGEN_TCG_F2_conv_df2uw_chop(SHORTCODE) \
1251     gen_helper_conv_df2uw_chop(RdV, tcg_env, RssV)
1252 #define fGEN_TCG_F2_conv_df2w_chop(SHORTCODE) \
1253     gen_helper_conv_df2w_chop(RdV, tcg_env, RssV)
1254 #define fGEN_TCG_F2_conv_df2ud_chop(SHORTCODE) \
1255     gen_helper_conv_df2ud_chop(RddV, tcg_env, RssV)
1256 #define fGEN_TCG_F2_conv_df2d_chop(SHORTCODE) \
1257     gen_helper_conv_df2d_chop(RddV, tcg_env, RssV)
1258 #define fGEN_TCG_F2_sfadd(SHORTCODE) \
1259     gen_helper_sfadd(RdV, tcg_env, RsV, RtV)
1260 #define fGEN_TCG_F2_sfsub(SHORTCODE) \
1261     gen_helper_sfsub(RdV, tcg_env, RsV, RtV)
1262 #define fGEN_TCG_F2_sfcmpeq(SHORTCODE) \
1263     gen_helper_sfcmpeq(PdV, tcg_env, RsV, RtV)
1264 #define fGEN_TCG_F2_sfcmpgt(SHORTCODE) \
1265     gen_helper_sfcmpgt(PdV, tcg_env, RsV, RtV)
1266 #define fGEN_TCG_F2_sfcmpge(SHORTCODE) \
1267     gen_helper_sfcmpge(PdV, tcg_env, RsV, RtV)
1268 #define fGEN_TCG_F2_sfcmpuo(SHORTCODE) \
1269     gen_helper_sfcmpuo(PdV, tcg_env, RsV, RtV)
1270 #define fGEN_TCG_F2_sfmax(SHORTCODE) \
1271     gen_helper_sfmax(RdV, tcg_env, RsV, RtV)
1272 #define fGEN_TCG_F2_sfmin(SHORTCODE) \
1273     gen_helper_sfmin(RdV, tcg_env, RsV, RtV)
1274 #define fGEN_TCG_F2_sfclass(SHORTCODE) \
1275     do { \
1276         TCGv imm = tcg_constant_tl(uiV); \
1277         gen_helper_sfclass(PdV, tcg_env, RsV, imm); \
1278     } while (0)
1279 #define fGEN_TCG_F2_sffixupn(SHORTCODE) \
1280     gen_helper_sffixupn(RdV, tcg_env, RsV, RtV)
1281 #define fGEN_TCG_F2_sffixupd(SHORTCODE) \
1282     gen_helper_sffixupd(RdV, tcg_env, RsV, RtV)
1283 #define fGEN_TCG_F2_sffixupr(SHORTCODE) \
1284     gen_helper_sffixupr(RdV, tcg_env, RsV)
1285 #define fGEN_TCG_F2_dfadd(SHORTCODE) \
1286     gen_helper_dfadd(RddV, tcg_env, RssV, RttV)
1287 #define fGEN_TCG_F2_dfsub(SHORTCODE) \
1288     gen_helper_dfsub(RddV, tcg_env, RssV, RttV)
1289 #define fGEN_TCG_F2_dfmax(SHORTCODE) \
1290     gen_helper_dfmax(RddV, tcg_env, RssV, RttV)
1291 #define fGEN_TCG_F2_dfmin(SHORTCODE) \
1292     gen_helper_dfmin(RddV, tcg_env, RssV, RttV)
1293 #define fGEN_TCG_F2_dfcmpeq(SHORTCODE) \
1294     gen_helper_dfcmpeq(PdV, tcg_env, RssV, RttV)
1295 #define fGEN_TCG_F2_dfcmpgt(SHORTCODE) \
1296     gen_helper_dfcmpgt(PdV, tcg_env, RssV, RttV)
1297 #define fGEN_TCG_F2_dfcmpge(SHORTCODE) \
1298     gen_helper_dfcmpge(PdV, tcg_env, RssV, RttV)
1299 #define fGEN_TCG_F2_dfcmpuo(SHORTCODE) \
1300     gen_helper_dfcmpuo(PdV, tcg_env, RssV, RttV)
1301 #define fGEN_TCG_F2_dfclass(SHORTCODE) \
1302     do { \
1303         TCGv imm = tcg_constant_tl(uiV); \
1304         gen_helper_dfclass(PdV, tcg_env, RssV, imm); \
1305     } while (0)
1306 #define fGEN_TCG_F2_sfmpy(SHORTCODE) \
1307     gen_helper_sfmpy(RdV, tcg_env, RsV, RtV)
1308 #define fGEN_TCG_F2_sffma(SHORTCODE) \
1309     gen_helper_sffma(RxV, tcg_env, RxV, RsV, RtV)
1310 #define fGEN_TCG_F2_sffma_sc(SHORTCODE) \
1311     gen_helper_sffma_sc(RxV, tcg_env, RxV, RsV, RtV, PuV)
1312 #define fGEN_TCG_F2_sffms(SHORTCODE) \
1313     gen_helper_sffms(RxV, tcg_env, RxV, RsV, RtV)
1314 #define fGEN_TCG_F2_sffma_lib(SHORTCODE) \
1315     gen_helper_sffma_lib(RxV, tcg_env, RxV, RsV, RtV)
1316 #define fGEN_TCG_F2_sffms_lib(SHORTCODE) \
1317     gen_helper_sffms_lib(RxV, tcg_env, RxV, RsV, RtV)
1318 
1319 #define fGEN_TCG_F2_dfmpyfix(SHORTCODE) \
1320     gen_helper_dfmpyfix(RddV, tcg_env, RssV, RttV)
1321 #define fGEN_TCG_F2_dfmpyhh(SHORTCODE) \
1322     gen_helper_dfmpyhh(RxxV, tcg_env, RxxV, RssV, RttV)
1323 
1324 /* Nothing to do for these in qemu, need to suppress compiler warnings */
1325 #define fGEN_TCG_Y4_l2fetch(SHORTCODE) \
1326     do { \
1327         RsV = RsV; \
1328         RtV = RtV; \
1329     } while (0)
1330 #define fGEN_TCG_Y5_l2fetch(SHORTCODE) \
1331     do { \
1332         RsV = RsV; \
1333     } while (0)
1334 #define fGEN_TCG_Y2_isync(SHORTCODE) \
1335     do { } while (0)
1336 #define fGEN_TCG_Y2_barrier(SHORTCODE) \
1337     do { } while (0)
1338 #define fGEN_TCG_Y2_syncht(SHORTCODE) \
1339     do { } while (0)
1340 #define fGEN_TCG_Y2_dcfetchbo(SHORTCODE) \
1341     do { \
1342         RsV = RsV; \
1343         uiV = uiV; \
1344     } while (0)
1345 
1346 #define fGEN_TCG_L2_loadw_aq(SHORTCODE)                 SHORTCODE
1347 #define fGEN_TCG_L4_loadd_aq(SHORTCODE)                 SHORTCODE
1348 
1349 /* Nothing to do for these in qemu, need to suppress compiler warnings */
1350 #define fGEN_TCG_R6_release_at_vi(SHORTCODE) \
1351     do { \
1352         RsV = RsV; \
1353     } while (0)
1354 #define fGEN_TCG_R6_release_st_vi(SHORTCODE) \
1355     do { \
1356         RsV = RsV; \
1357     } while (0)
1358 
1359 #define fGEN_TCG_S2_storew_rl_at_vi(SHORTCODE)          SHORTCODE
1360 #define fGEN_TCG_S4_stored_rl_at_vi(SHORTCODE)          SHORTCODE
1361 #define fGEN_TCG_S2_storew_rl_st_vi(SHORTCODE)          SHORTCODE
1362 #define fGEN_TCG_S4_stored_rl_st_vi(SHORTCODE)          SHORTCODE
1363 
1364 #define fGEN_TCG_J2_trap0(SHORTCODE) \
1365     do { \
1366         uiV = uiV; \
1367         tcg_gen_movi_tl(hex_gpr[HEX_REG_PC], ctx->pkt->pc); \
1368         TCGv excp = tcg_constant_tl(HEX_EXCP_TRAP0); \
1369         gen_helper_raise_exception(tcg_env, excp); \
1370     } while (0)
1371 #endif
1372 
1373 #define fGEN_TCG_A2_nop(SHORTCODE) do { } while (0)
1374 #define fGEN_TCG_SA1_setin1(SHORTCODE) tcg_gen_movi_tl(RdV, -1)
1375