xref: /openbmc/qemu/target/hexagon/gen_tcg.h (revision c85cad81)
1 /*
2  *  Copyright(c) 2019-2023 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, \
72                             hex_gpr[HEX_REG_CS0 + MuN]); \
73     } while (0)
74 #define GET_EA_pcr(SHIFT) \
75     do { \
76         TCGv ireg = tcg_temp_new(); \
77         tcg_gen_mov_tl(EA, RxV); \
78         gen_read_ireg(ireg, MuV, (SHIFT)); \
79         gen_helper_fcircadd(RxV, RxV, ireg, MuV, hex_gpr[HEX_REG_CS0 + MuN]); \
80     } while (0)
81 
82 /* Instructions with multiple definitions */
83 #define fGEN_TCG_LOAD_AP(RES, SIZE, SIGN) \
84     do { \
85         fMUST_IMMEXT(UiV); \
86         fEA_IMM(UiV); \
87         fLOAD(1, SIZE, SIGN, EA, RES); \
88         tcg_gen_movi_tl(ReV, UiV); \
89     } while (0)
90 
91 #define fGEN_TCG_L4_loadrub_ap(SHORTCODE) \
92     fGEN_TCG_LOAD_AP(RdV, 1, u)
93 #define fGEN_TCG_L4_loadrb_ap(SHORTCODE) \
94     fGEN_TCG_LOAD_AP(RdV, 1, s)
95 #define fGEN_TCG_L4_loadruh_ap(SHORTCODE) \
96     fGEN_TCG_LOAD_AP(RdV, 2, u)
97 #define fGEN_TCG_L4_loadrh_ap(SHORTCODE) \
98     fGEN_TCG_LOAD_AP(RdV, 2, s)
99 #define fGEN_TCG_L4_loadri_ap(SHORTCODE) \
100     fGEN_TCG_LOAD_AP(RdV, 4, u)
101 #define fGEN_TCG_L4_loadrd_ap(SHORTCODE) \
102     fGEN_TCG_LOAD_AP(RddV, 8, u)
103 
104 #define fGEN_TCG_L2_loadrub_pci(SHORTCODE)    SHORTCODE
105 #define fGEN_TCG_L2_loadrb_pci(SHORTCODE)     SHORTCODE
106 #define fGEN_TCG_L2_loadruh_pci(SHORTCODE)    SHORTCODE
107 #define fGEN_TCG_L2_loadrh_pci(SHORTCODE)     SHORTCODE
108 #define fGEN_TCG_L2_loadri_pci(SHORTCODE)     SHORTCODE
109 #define fGEN_TCG_L2_loadrd_pci(SHORTCODE)     SHORTCODE
110 
111 #define fGEN_TCG_LOAD_pcr(SHIFT, LOAD) \
112     do { \
113         TCGv ireg = tcg_temp_new(); \
114         tcg_gen_mov_tl(EA, RxV); \
115         gen_read_ireg(ireg, MuV, SHIFT); \
116         gen_helper_fcircadd(RxV, RxV, ireg, MuV, hex_gpr[HEX_REG_CS0 + MuN]); \
117         LOAD; \
118     } while (0)
119 
120 #define fGEN_TCG_L2_loadrub_pcr(SHORTCODE) \
121       fGEN_TCG_LOAD_pcr(0, fLOAD(1, 1, u, EA, RdV))
122 #define fGEN_TCG_L2_loadrb_pcr(SHORTCODE) \
123       fGEN_TCG_LOAD_pcr(0, fLOAD(1, 1, s, EA, RdV))
124 #define fGEN_TCG_L2_loadruh_pcr(SHORTCODE) \
125       fGEN_TCG_LOAD_pcr(1, fLOAD(1, 2, u, EA, RdV))
126 #define fGEN_TCG_L2_loadrh_pcr(SHORTCODE) \
127       fGEN_TCG_LOAD_pcr(1, fLOAD(1, 2, s, EA, RdV))
128 #define fGEN_TCG_L2_loadri_pcr(SHORTCODE) \
129       fGEN_TCG_LOAD_pcr(2, fLOAD(1, 4, u, EA, RdV))
130 #define fGEN_TCG_L2_loadrd_pcr(SHORTCODE) \
131       fGEN_TCG_LOAD_pcr(3, fLOAD(1, 8, u, EA, RddV))
132 
133 #define fGEN_TCG_L2_loadrub_pr(SHORTCODE)      SHORTCODE
134 #define fGEN_TCG_L2_loadrub_pbr(SHORTCODE)     SHORTCODE
135 #define fGEN_TCG_L2_loadrub_pi(SHORTCODE)      SHORTCODE
136 #define fGEN_TCG_L2_loadrb_pr(SHORTCODE)       SHORTCODE
137 #define fGEN_TCG_L2_loadrb_pbr(SHORTCODE)      SHORTCODE
138 #define fGEN_TCG_L2_loadrb_pi(SHORTCODE)       SHORTCODE
139 #define fGEN_TCG_L2_loadruh_pr(SHORTCODE)      SHORTCODE
140 #define fGEN_TCG_L2_loadruh_pbr(SHORTCODE)     SHORTCODE
141 #define fGEN_TCG_L2_loadruh_pi(SHORTCODE)      SHORTCODE
142 #define fGEN_TCG_L2_loadrh_pr(SHORTCODE)       SHORTCODE
143 #define fGEN_TCG_L2_loadrh_pbr(SHORTCODE)      SHORTCODE
144 #define fGEN_TCG_L2_loadrh_pi(SHORTCODE)       SHORTCODE
145 #define fGEN_TCG_L2_loadri_pr(SHORTCODE)       SHORTCODE
146 #define fGEN_TCG_L2_loadri_pbr(SHORTCODE)      SHORTCODE
147 #define fGEN_TCG_L2_loadri_pi(SHORTCODE)       SHORTCODE
148 #define fGEN_TCG_L2_loadrd_pr(SHORTCODE)       SHORTCODE
149 #define fGEN_TCG_L2_loadrd_pbr(SHORTCODE)      SHORTCODE
150 #define fGEN_TCG_L2_loadrd_pi(SHORTCODE)       SHORTCODE
151 
152 /*
153  * These instructions load 2 bytes and places them in
154  * two halves of the destination register.
155  * The GET_EA macro determines the addressing mode.
156  * The SIGN argument determines whether to zero-extend or
157  * sign-extend.
158  */
159 #define fGEN_TCG_loadbXw2(GET_EA, SIGN) \
160     do { \
161         TCGv tmp = tcg_temp_new(); \
162         TCGv byte = tcg_temp_new(); \
163         GET_EA; \
164         fLOAD(1, 2, u, EA, tmp); \
165         tcg_gen_movi_tl(RdV, 0); \
166         for (int i = 0; i < 2; i++) { \
167             gen_set_half(i, RdV, gen_get_byte(byte, i, tmp, (SIGN))); \
168         } \
169     } while (0)
170 
171 #define fGEN_TCG_L2_loadbzw2_io(SHORTCODE) \
172     fGEN_TCG_loadbXw2(fEA_RI(RsV, siV), false)
173 #define fGEN_TCG_L4_loadbzw2_ur(SHORTCODE) \
174     fGEN_TCG_loadbXw2(fEA_IRs(UiV, RtV, uiV), false)
175 #define fGEN_TCG_L2_loadbsw2_io(SHORTCODE) \
176     fGEN_TCG_loadbXw2(fEA_RI(RsV, siV), true)
177 #define fGEN_TCG_L4_loadbsw2_ur(SHORTCODE) \
178     fGEN_TCG_loadbXw2(fEA_IRs(UiV, RtV, uiV), true)
179 #define fGEN_TCG_L4_loadbzw2_ap(SHORTCODE) \
180     fGEN_TCG_loadbXw2(GET_EA_ap, false)
181 #define fGEN_TCG_L2_loadbzw2_pr(SHORTCODE) \
182     fGEN_TCG_loadbXw2(GET_EA_pr, false)
183 #define fGEN_TCG_L2_loadbzw2_pbr(SHORTCODE) \
184     fGEN_TCG_loadbXw2(GET_EA_pbr, false)
185 #define fGEN_TCG_L2_loadbzw2_pi(SHORTCODE) \
186     fGEN_TCG_loadbXw2(GET_EA_pi, false)
187 #define fGEN_TCG_L4_loadbsw2_ap(SHORTCODE) \
188     fGEN_TCG_loadbXw2(GET_EA_ap, true)
189 #define fGEN_TCG_L2_loadbsw2_pr(SHORTCODE) \
190     fGEN_TCG_loadbXw2(GET_EA_pr, true)
191 #define fGEN_TCG_L2_loadbsw2_pbr(SHORTCODE) \
192     fGEN_TCG_loadbXw2(GET_EA_pbr, true)
193 #define fGEN_TCG_L2_loadbsw2_pi(SHORTCODE) \
194     fGEN_TCG_loadbXw2(GET_EA_pi, true)
195 #define fGEN_TCG_L2_loadbzw2_pci(SHORTCODE) \
196     fGEN_TCG_loadbXw2(GET_EA_pci, false)
197 #define fGEN_TCG_L2_loadbsw2_pci(SHORTCODE) \
198     fGEN_TCG_loadbXw2(GET_EA_pci, true)
199 #define fGEN_TCG_L2_loadbzw2_pcr(SHORTCODE) \
200     fGEN_TCG_loadbXw2(GET_EA_pcr(1), false)
201 #define fGEN_TCG_L2_loadbsw2_pcr(SHORTCODE) \
202     fGEN_TCG_loadbXw2(GET_EA_pcr(1), true)
203 
204 /*
205  * These instructions load 4 bytes and places them in
206  * four halves of the destination register pair.
207  * The GET_EA macro determines the addressing mode.
208  * The SIGN argument determines whether to zero-extend or
209  * sign-extend.
210  */
211 #define fGEN_TCG_loadbXw4(GET_EA, SIGN) \
212     do { \
213         TCGv tmp = tcg_temp_new(); \
214         TCGv byte = tcg_temp_new(); \
215         GET_EA; \
216         fLOAD(1, 4, u, EA, tmp);  \
217         tcg_gen_movi_i64(RddV, 0); \
218         for (int i = 0; i < 4; i++) { \
219             gen_set_half_i64(i, RddV, gen_get_byte(byte, i, tmp, (SIGN)));  \
220         }  \
221     } while (0)
222 
223 #define fGEN_TCG_L2_loadbzw4_io(SHORTCODE) \
224     fGEN_TCG_loadbXw4(fEA_RI(RsV, siV), false)
225 #define fGEN_TCG_L4_loadbzw4_ur(SHORTCODE) \
226     fGEN_TCG_loadbXw4(fEA_IRs(UiV, RtV, uiV), false)
227 #define fGEN_TCG_L2_loadbsw4_io(SHORTCODE) \
228     fGEN_TCG_loadbXw4(fEA_RI(RsV, siV), true)
229 #define fGEN_TCG_L4_loadbsw4_ur(SHORTCODE) \
230     fGEN_TCG_loadbXw4(fEA_IRs(UiV, RtV, uiV), true)
231 #define fGEN_TCG_L2_loadbzw4_pci(SHORTCODE) \
232     fGEN_TCG_loadbXw4(GET_EA_pci, false)
233 #define fGEN_TCG_L2_loadbsw4_pci(SHORTCODE) \
234     fGEN_TCG_loadbXw4(GET_EA_pci, true)
235 #define fGEN_TCG_L2_loadbzw4_pcr(SHORTCODE) \
236     fGEN_TCG_loadbXw4(GET_EA_pcr(2), false)
237 #define fGEN_TCG_L2_loadbsw4_pcr(SHORTCODE) \
238     fGEN_TCG_loadbXw4(GET_EA_pcr(2), true)
239 #define fGEN_TCG_L4_loadbzw4_ap(SHORTCODE) \
240     fGEN_TCG_loadbXw4(GET_EA_ap, false)
241 #define fGEN_TCG_L2_loadbzw4_pr(SHORTCODE) \
242     fGEN_TCG_loadbXw4(GET_EA_pr, false)
243 #define fGEN_TCG_L2_loadbzw4_pbr(SHORTCODE) \
244     fGEN_TCG_loadbXw4(GET_EA_pbr, false)
245 #define fGEN_TCG_L2_loadbzw4_pi(SHORTCODE) \
246     fGEN_TCG_loadbXw4(GET_EA_pi, false)
247 #define fGEN_TCG_L4_loadbsw4_ap(SHORTCODE) \
248     fGEN_TCG_loadbXw4(GET_EA_ap, true)
249 #define fGEN_TCG_L2_loadbsw4_pr(SHORTCODE) \
250     fGEN_TCG_loadbXw4(GET_EA_pr, true)
251 #define fGEN_TCG_L2_loadbsw4_pbr(SHORTCODE) \
252     fGEN_TCG_loadbXw4(GET_EA_pbr, true)
253 #define fGEN_TCG_L2_loadbsw4_pi(SHORTCODE) \
254     fGEN_TCG_loadbXw4(GET_EA_pi, true)
255 
256 /*
257  * These instructions load a half word, shift the destination right by 16 bits
258  * and place the loaded value in the high half word of the destination pair.
259  * The GET_EA macro determines the addressing mode.
260  */
261 #define fGEN_TCG_loadalignh(GET_EA) \
262     do { \
263         TCGv tmp = tcg_temp_new(); \
264         TCGv_i64 tmp_i64 = tcg_temp_new_i64(); \
265         GET_EA;  \
266         fLOAD(1, 2, u, EA, tmp);  \
267         tcg_gen_extu_i32_i64(tmp_i64, tmp); \
268         tcg_gen_shri_i64(RyyV, RyyV, 16); \
269         tcg_gen_deposit_i64(RyyV, RyyV, tmp_i64, 48, 16); \
270     } while (0)
271 
272 #define fGEN_TCG_L4_loadalignh_ur(SHORTCODE) \
273     fGEN_TCG_loadalignh(fEA_IRs(UiV, RtV, uiV))
274 #define fGEN_TCG_L2_loadalignh_io(SHORTCODE) \
275     fGEN_TCG_loadalignh(fEA_RI(RsV, siV))
276 #define fGEN_TCG_L2_loadalignh_pci(SHORTCODE) \
277     fGEN_TCG_loadalignh(GET_EA_pci)
278 #define fGEN_TCG_L2_loadalignh_pcr(SHORTCODE) \
279     fGEN_TCG_loadalignh(GET_EA_pcr(1))
280 #define fGEN_TCG_L4_loadalignh_ap(SHORTCODE) \
281     fGEN_TCG_loadalignh(GET_EA_ap)
282 #define fGEN_TCG_L2_loadalignh_pr(SHORTCODE) \
283     fGEN_TCG_loadalignh(GET_EA_pr)
284 #define fGEN_TCG_L2_loadalignh_pbr(SHORTCODE) \
285     fGEN_TCG_loadalignh(GET_EA_pbr)
286 #define fGEN_TCG_L2_loadalignh_pi(SHORTCODE) \
287     fGEN_TCG_loadalignh(GET_EA_pi)
288 
289 /* Same as above, but loads a byte instead of half word */
290 #define fGEN_TCG_loadalignb(GET_EA) \
291     do { \
292         TCGv tmp = tcg_temp_new(); \
293         TCGv_i64 tmp_i64 = tcg_temp_new_i64(); \
294         GET_EA;  \
295         fLOAD(1, 1, u, EA, tmp);  \
296         tcg_gen_extu_i32_i64(tmp_i64, tmp); \
297         tcg_gen_shri_i64(RyyV, RyyV, 8); \
298         tcg_gen_deposit_i64(RyyV, RyyV, tmp_i64, 56, 8); \
299     } while (0)
300 
301 #define fGEN_TCG_L2_loadalignb_io(SHORTCODE) \
302     fGEN_TCG_loadalignb(fEA_RI(RsV, siV))
303 #define fGEN_TCG_L4_loadalignb_ur(SHORTCODE) \
304     fGEN_TCG_loadalignb(fEA_IRs(UiV, RtV, uiV))
305 #define fGEN_TCG_L2_loadalignb_pci(SHORTCODE) \
306     fGEN_TCG_loadalignb(GET_EA_pci)
307 #define fGEN_TCG_L2_loadalignb_pcr(SHORTCODE) \
308     fGEN_TCG_loadalignb(GET_EA_pcr(0))
309 #define fGEN_TCG_L4_loadalignb_ap(SHORTCODE) \
310     fGEN_TCG_loadalignb(GET_EA_ap)
311 #define fGEN_TCG_L2_loadalignb_pr(SHORTCODE) \
312     fGEN_TCG_loadalignb(GET_EA_pr)
313 #define fGEN_TCG_L2_loadalignb_pbr(SHORTCODE) \
314     fGEN_TCG_loadalignb(GET_EA_pbr)
315 #define fGEN_TCG_L2_loadalignb_pi(SHORTCODE) \
316     fGEN_TCG_loadalignb(GET_EA_pi)
317 
318 /*
319  * Predicated loads
320  * Here is a primer to understand the tag names
321  *
322  * Predicate used
323  *      t        true "old" value                  if (p0) r0 = memb(r2+#0)
324  *      f        false "old" value                 if (!p0) r0 = memb(r2+#0)
325  *      tnew     true "new" value                  if (p0.new) r0 = memb(r2+#0)
326  *      fnew     false "new" value                 if (!p0.new) r0 = memb(r2+#0)
327  */
328 #define fGEN_TCG_PRED_LOAD(GET_EA, PRED, SIZE, SIGN) \
329     do { \
330         TCGv LSB = tcg_temp_new(); \
331         TCGLabel *label = gen_new_label(); \
332         tcg_gen_movi_tl(EA, 0); \
333         PRED;  \
334         CHECK_NOSHUF_PRED(GET_EA, SIZE, LSB); \
335         tcg_gen_brcondi_tl(TCG_COND_EQ, LSB, 0, label); \
336         fLOAD(1, SIZE, SIGN, EA, RdV); \
337         gen_set_label(label); \
338     } while (0)
339 
340 #define fGEN_TCG_L2_ploadrubt_pi(SHORTCODE) \
341     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBOLD(PtV), 1, u)
342 #define fGEN_TCG_L2_ploadrubf_pi(SHORTCODE) \
343     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBOLDNOT(PtV), 1, u)
344 #define fGEN_TCG_L2_ploadrubtnew_pi(SHORTCODE) \
345     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBNEW(PtN), 1, u)
346 #define fGEN_TCG_L2_ploadrubfnew_pi(SHORTCODE) \
347     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBNEWNOT(PtN), 1, u)
348 #define fGEN_TCG_L2_ploadrbt_pi(SHORTCODE) \
349     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBOLD(PtV), 1, s)
350 #define fGEN_TCG_L2_ploadrbf_pi(SHORTCODE) \
351     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBOLDNOT(PtV), 1, s)
352 #define fGEN_TCG_L2_ploadrbtnew_pi(SHORTCODE) \
353     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBNEW(PtN), 1, s)
354 #define fGEN_TCG_L2_ploadrbfnew_pi(SHORTCODE) \
355     fGEN_TCG_PRED_LOAD({ fEA_REG(RxV); fPM_I(RxV, siV); }, \
356                        fLSBNEWNOT(PtN), 1, s)
357 
358 #define fGEN_TCG_L2_ploadruht_pi(SHORTCODE) \
359     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBOLD(PtV), 2, u)
360 #define fGEN_TCG_L2_ploadruhf_pi(SHORTCODE) \
361     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBOLDNOT(PtV), 2, u)
362 #define fGEN_TCG_L2_ploadruhtnew_pi(SHORTCODE) \
363     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBNEW(PtN), 2, u)
364 #define fGEN_TCG_L2_ploadruhfnew_pi(SHORTCODE) \
365     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBNEWNOT(PtN), 2, u)
366 #define fGEN_TCG_L2_ploadrht_pi(SHORTCODE) \
367     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBOLD(PtV), 2, s)
368 #define fGEN_TCG_L2_ploadrhf_pi(SHORTCODE) \
369     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBOLDNOT(PtV), 2, s)
370 #define fGEN_TCG_L2_ploadrhtnew_pi(SHORTCODE) \
371     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBNEW(PtN), 2, s)
372 #define fGEN_TCG_L2_ploadrhfnew_pi(SHORTCODE) \
373     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBNEWNOT(PtN), 2, s)
374 
375 #define fGEN_TCG_L2_ploadrit_pi(SHORTCODE) \
376     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBOLD(PtV), 4, u)
377 #define fGEN_TCG_L2_ploadrif_pi(SHORTCODE) \
378     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBOLDNOT(PtV), 4, u)
379 #define fGEN_TCG_L2_ploadritnew_pi(SHORTCODE) \
380     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBNEW(PtN), 4, u)
381 #define fGEN_TCG_L2_ploadrifnew_pi(SHORTCODE) \
382     fGEN_TCG_PRED_LOAD(GET_EA_pi, fLSBNEWNOT(PtN), 4, u)
383 
384 /* Predicated loads into a register pair */
385 #define fGEN_TCG_PRED_LOAD_PAIR(GET_EA, PRED) \
386     do { \
387         TCGv LSB = tcg_temp_new(); \
388         TCGLabel *label = gen_new_label(); \
389         tcg_gen_movi_tl(EA, 0); \
390         PRED;  \
391         CHECK_NOSHUF_PRED(GET_EA, 8, LSB); \
392         tcg_gen_brcondi_tl(TCG_COND_EQ, LSB, 0, label); \
393         fLOAD(1, 8, u, EA, RddV); \
394         gen_set_label(label); \
395     } while (0)
396 
397 #define fGEN_TCG_L2_ploadrdt_pi(SHORTCODE) \
398     fGEN_TCG_PRED_LOAD_PAIR(GET_EA_pi, fLSBOLD(PtV))
399 #define fGEN_TCG_L2_ploadrdf_pi(SHORTCODE) \
400     fGEN_TCG_PRED_LOAD_PAIR(GET_EA_pi, fLSBOLDNOT(PtV))
401 #define fGEN_TCG_L2_ploadrdtnew_pi(SHORTCODE) \
402     fGEN_TCG_PRED_LOAD_PAIR(GET_EA_pi, fLSBNEW(PtN))
403 #define fGEN_TCG_L2_ploadrdfnew_pi(SHORTCODE) \
404     fGEN_TCG_PRED_LOAD_PAIR(GET_EA_pi, fLSBNEWNOT(PtN))
405 
406 /* load-locked and store-locked */
407 #define fGEN_TCG_L2_loadw_locked(SHORTCODE) \
408     SHORTCODE
409 #define fGEN_TCG_L4_loadd_locked(SHORTCODE) \
410     SHORTCODE
411 #define fGEN_TCG_S2_storew_locked(SHORTCODE) \
412     SHORTCODE
413 #define fGEN_TCG_S4_stored_locked(SHORTCODE) \
414     SHORTCODE
415 
416 #define fGEN_TCG_STORE(SHORTCODE) \
417     do { \
418         TCGv HALF G_GNUC_UNUSED = tcg_temp_new(); \
419         TCGv BYTE G_GNUC_UNUSED = tcg_temp_new(); \
420         SHORTCODE; \
421     } while (0)
422 
423 #define fGEN_TCG_STORE_pcr(SHIFT, STORE) \
424     do { \
425         TCGv ireg = tcg_temp_new(); \
426         TCGv HALF G_GNUC_UNUSED = tcg_temp_new(); \
427         TCGv BYTE G_GNUC_UNUSED = tcg_temp_new(); \
428         tcg_gen_mov_tl(EA, RxV); \
429         gen_read_ireg(ireg, MuV, SHIFT); \
430         gen_helper_fcircadd(RxV, RxV, ireg, MuV, hex_gpr[HEX_REG_CS0 + MuN]); \
431         STORE; \
432     } while (0)
433 
434 #define fGEN_TCG_S2_storerb_pbr(SHORTCODE) \
435     fGEN_TCG_STORE(SHORTCODE)
436 #define fGEN_TCG_S2_storerb_pci(SHORTCODE) \
437     fGEN_TCG_STORE(SHORTCODE)
438 #define fGEN_TCG_S2_storerb_pcr(SHORTCODE) \
439     fGEN_TCG_STORE_pcr(0, fSTORE(1, 1, EA, fGETBYTE(0, RtV)))
440 
441 #define fGEN_TCG_S2_storerh_pbr(SHORTCODE) \
442     fGEN_TCG_STORE(SHORTCODE)
443 #define fGEN_TCG_S2_storerh_pci(SHORTCODE) \
444     fGEN_TCG_STORE(SHORTCODE)
445 #define fGEN_TCG_S2_storerh_pcr(SHORTCODE) \
446     fGEN_TCG_STORE_pcr(1, fSTORE(1, 2, EA, fGETHALF(0, RtV)))
447 
448 #define fGEN_TCG_S2_storerf_pbr(SHORTCODE) \
449     fGEN_TCG_STORE(SHORTCODE)
450 #define fGEN_TCG_S2_storerf_pci(SHORTCODE) \
451     fGEN_TCG_STORE(SHORTCODE)
452 #define fGEN_TCG_S2_storerf_pcr(SHORTCODE) \
453     fGEN_TCG_STORE_pcr(1, fSTORE(1, 2, EA, fGETHALF(1, RtV)))
454 
455 #define fGEN_TCG_S2_storeri_pbr(SHORTCODE) \
456     fGEN_TCG_STORE(SHORTCODE)
457 #define fGEN_TCG_S2_storeri_pci(SHORTCODE) \
458     fGEN_TCG_STORE(SHORTCODE)
459 #define fGEN_TCG_S2_storeri_pcr(SHORTCODE) \
460     fGEN_TCG_STORE_pcr(2, fSTORE(1, 4, EA, RtV))
461 
462 #define fGEN_TCG_S2_storerd_pbr(SHORTCODE) \
463     fGEN_TCG_STORE(SHORTCODE)
464 #define fGEN_TCG_S2_storerd_pci(SHORTCODE) \
465     fGEN_TCG_STORE(SHORTCODE)
466 #define fGEN_TCG_S2_storerd_pcr(SHORTCODE) \
467     fGEN_TCG_STORE_pcr(3, fSTORE(1, 8, EA, RttV))
468 
469 #define fGEN_TCG_S2_storerbnew_pbr(SHORTCODE) \
470     fGEN_TCG_STORE(SHORTCODE)
471 #define fGEN_TCG_S2_storerbnew_pci(SHORTCODE) \
472     fGEN_TCG_STORE(SHORTCODE)
473 #define fGEN_TCG_S2_storerbnew_pcr(SHORTCODE) \
474     fGEN_TCG_STORE_pcr(0, fSTORE(1, 1, EA, fGETBYTE(0, NtN)))
475 
476 #define fGEN_TCG_S2_storerhnew_pbr(SHORTCODE) \
477     fGEN_TCG_STORE(SHORTCODE)
478 #define fGEN_TCG_S2_storerhnew_pci(SHORTCODE) \
479     fGEN_TCG_STORE(SHORTCODE)
480 #define fGEN_TCG_S2_storerhnew_pcr(SHORTCODE) \
481     fGEN_TCG_STORE_pcr(1, fSTORE(1, 2, EA, fGETHALF(0, NtN)))
482 
483 #define fGEN_TCG_S2_storerinew_pbr(SHORTCODE) \
484     fGEN_TCG_STORE(SHORTCODE)
485 #define fGEN_TCG_S2_storerinew_pci(SHORTCODE) \
486     fGEN_TCG_STORE(SHORTCODE)
487 #define fGEN_TCG_S2_storerinew_pcr(SHORTCODE) \
488     fGEN_TCG_STORE_pcr(2, fSTORE(1, 4, EA, NtN))
489 
490 /* dczeroa clears the 32 byte cache line at the address given */
491 #define fGEN_TCG_Y2_dczeroa(SHORTCODE) SHORTCODE
492 
493 /* In linux-user mode, these are not modelled, suppress compiler warning */
494 #define fGEN_TCG_Y2_dcinva(SHORTCODE) \
495     do { RsV = RsV; } while (0)
496 #define fGEN_TCG_Y2_dccleaninva(SHORTCODE) \
497     do { RsV = RsV; } while (0)
498 #define fGEN_TCG_Y2_dccleana(SHORTCODE) \
499     do { RsV = RsV; } while (0)
500 #define fGEN_TCG_Y2_icinva(SHORTCODE) \
501     do { RsV = RsV; } while (0)
502 
503 /*
504  * allocframe(#uiV)
505  *     RxV == r29
506  */
507 #define fGEN_TCG_S2_allocframe(SHORTCODE) \
508     gen_allocframe(ctx, RxV, uiV)
509 
510 /* sub-instruction version (no RxV, so handle it manually) */
511 #define fGEN_TCG_SS2_allocframe(SHORTCODE) \
512     do { \
513         TCGv r29 = tcg_temp_new(); \
514         tcg_gen_mov_tl(r29, hex_gpr[HEX_REG_SP]); \
515         gen_allocframe(ctx, r29, uiV); \
516         gen_log_reg_write(ctx, HEX_REG_SP, r29); \
517     } while (0)
518 
519 /*
520  * Rdd32 = deallocframe(Rs32):raw
521  *     RddV == r31:30
522  *     RsV  == r30
523  */
524 #define fGEN_TCG_L2_deallocframe(SHORTCODE) \
525     gen_deallocframe(ctx, RddV, RsV)
526 
527 /* sub-instruction version (no RddV/RsV, so handle it manually) */
528 #define fGEN_TCG_SL2_deallocframe(SHORTCODE) \
529     do { \
530         TCGv_i64 r31_30 = tcg_temp_new_i64(); \
531         gen_deallocframe(ctx, r31_30, hex_gpr[HEX_REG_FP]); \
532         gen_log_reg_write_pair(ctx, HEX_REG_FP, r31_30); \
533     } while (0)
534 
535 /*
536  * dealloc_return
537  * Assembler mapped to
538  * r31:30 = dealloc_return(r30):raw
539  */
540 #define fGEN_TCG_L4_return(SHORTCODE) \
541     gen_return(ctx, RddV, RsV)
542 
543 /*
544  * sub-instruction version (no RddV, so handle it manually)
545  */
546 #define fGEN_TCG_SL2_return(SHORTCODE) \
547     do { \
548         TCGv_i64 RddV = get_result_gpr_pair(ctx, HEX_REG_FP); \
549         gen_return(ctx, RddV, hex_gpr[HEX_REG_FP]); \
550         gen_log_reg_write_pair(ctx, HEX_REG_FP, RddV); \
551     } while (0)
552 
553 /*
554  * Conditional returns follow this naming convention
555  *     _t                 predicate true
556  *     _f                 predicate false
557  *     _tnew_pt           predicate.new true predict taken
558  *     _fnew_pt           predicate.new false predict taken
559  *     _tnew_pnt          predicate.new true predict not taken
560  *     _fnew_pnt          predicate.new false predict not taken
561  * Predictions are not modelled in QEMU
562  *
563  * Example:
564  *     if (p1) r31:30 = dealloc_return(r30):raw
565  */
566 #define fGEN_TCG_L4_return_t(SHORTCODE) \
567     gen_cond_return(ctx, RddV, RsV, PvV, TCG_COND_EQ);
568 #define fGEN_TCG_L4_return_f(SHORTCODE) \
569     gen_cond_return(ctx, RddV, RsV, PvV, TCG_COND_NE)
570 #define fGEN_TCG_L4_return_tnew_pt(SHORTCODE) \
571     gen_cond_return(ctx, RddV, RsV, PvN, TCG_COND_EQ)
572 #define fGEN_TCG_L4_return_fnew_pt(SHORTCODE) \
573     gen_cond_return(ctx, RddV, RsV, PvN, TCG_COND_NE)
574 #define fGEN_TCG_L4_return_tnew_pnt(SHORTCODE) \
575     gen_cond_return(ctx, RddV, RsV, PvN, TCG_COND_EQ)
576 #define fGEN_TCG_L4_return_fnew_pnt(SHORTCODE) \
577     gen_cond_return(ctx, RddV, RsV, PvN, TCG_COND_NE)
578 
579 #define fGEN_TCG_SL2_return_t(SHORTCODE) \
580     gen_cond_return_subinsn(ctx, TCG_COND_EQ, hex_pred[0])
581 #define fGEN_TCG_SL2_return_f(SHORTCODE) \
582     gen_cond_return_subinsn(ctx, TCG_COND_NE, hex_pred[0])
583 #define fGEN_TCG_SL2_return_tnew(SHORTCODE) \
584     gen_cond_return_subinsn(ctx, TCG_COND_EQ, ctx->new_pred_value[0])
585 #define fGEN_TCG_SL2_return_fnew(SHORTCODE) \
586     gen_cond_return_subinsn(ctx, TCG_COND_NE, ctx->new_pred_value[0])
587 
588 /*
589  * Mathematical operations with more than one definition require
590  * special handling
591  */
592 #define fGEN_TCG_A5_ACS(SHORTCODE) \
593     do { \
594         gen_helper_vacsh_pred(PeV, cpu_env, RxxV, RssV, RttV); \
595         gen_helper_vacsh_val(RxxV, cpu_env, RxxV, RssV, RttV, \
596                              tcg_constant_tl(ctx->need_commit)); \
597     } while (0)
598 
599 #define fGEN_TCG_S2_cabacdecbin(SHORTCODE) \
600     do { \
601         TCGv p0 = tcg_temp_new(); \
602         gen_helper_cabacdecbin_pred(p0, RssV, RttV); \
603         gen_helper_cabacdecbin_val(RddV, RssV, RttV); \
604         gen_log_pred_write(ctx, 0, p0); \
605     } while (0)
606 
607 /*
608  * Approximate reciprocal
609  * r3,p1 = sfrecipa(r0, r1)
610  *
611  * The helper packs the 2 32-bit results into a 64-bit value,
612  * so unpack them into the proper results.
613  */
614 #define fGEN_TCG_F2_sfrecipa(SHORTCODE) \
615     do { \
616         TCGv_i64 tmp = tcg_temp_new_i64(); \
617         gen_helper_sfrecipa(tmp, cpu_env, RsV, RtV);  \
618         tcg_gen_extrh_i64_i32(RdV, tmp); \
619         tcg_gen_extrl_i64_i32(PeV, tmp); \
620     } while (0)
621 
622 /*
623  * Approximation of the reciprocal square root
624  * r1,p0 = sfinvsqrta(r0)
625  *
626  * The helper packs the 2 32-bit results into a 64-bit value,
627  * so unpack them into the proper results.
628  */
629 #define fGEN_TCG_F2_sfinvsqrta(SHORTCODE) \
630     do { \
631         TCGv_i64 tmp = tcg_temp_new_i64(); \
632         gen_helper_sfinvsqrta(tmp, cpu_env, RsV); \
633         tcg_gen_extrh_i64_i32(RdV, tmp); \
634         tcg_gen_extrl_i64_i32(PeV, tmp); \
635     } while (0)
636 
637 /*
638  * Add or subtract with carry.
639  * Predicate register is used as an extra input and output.
640  * r5:4 = add(r1:0, r3:2, p1):carry
641  */
642 #define fGEN_TCG_A4_addp_c(SHORTCODE) \
643     do { \
644         TCGv_i64 carry = tcg_temp_new_i64(); \
645         TCGv_i64 zero = tcg_constant_i64(0); \
646         tcg_gen_extu_i32_i64(carry, PxV); \
647         tcg_gen_andi_i64(carry, carry, 1); \
648         tcg_gen_add2_i64(RddV, carry, RssV, zero, carry, zero); \
649         tcg_gen_add2_i64(RddV, carry, RddV, carry, RttV, zero); \
650         tcg_gen_extrl_i64_i32(PxV, carry); \
651         gen_8bitsof(PxV, PxV); \
652     } while (0)
653 
654 /* r5:4 = sub(r1:0, r3:2, p1):carry */
655 #define fGEN_TCG_A4_subp_c(SHORTCODE) \
656     do { \
657         TCGv_i64 carry = tcg_temp_new_i64(); \
658         TCGv_i64 zero = tcg_constant_i64(0); \
659         TCGv_i64 not_RttV = tcg_temp_new_i64(); \
660         tcg_gen_extu_i32_i64(carry, PxV); \
661         tcg_gen_andi_i64(carry, carry, 1); \
662         tcg_gen_not_i64(not_RttV, RttV); \
663         tcg_gen_add2_i64(RddV, carry, RssV, zero, carry, zero); \
664         tcg_gen_add2_i64(RddV, carry, RddV, carry, not_RttV, zero); \
665         tcg_gen_extrl_i64_i32(PxV, carry); \
666         gen_8bitsof(PxV, PxV); \
667     } while (0)
668 
669 /*
670  * Compare each of the 8 unsigned bytes
671  * The minimum is placed in each byte of the destination.
672  * Each bit of the predicate is set true if the bit from the first operand
673  * is greater than the bit from the second operand.
674  * r5:4,p1 = vminub(r1:0, r3:2)
675  */
676 #define fGEN_TCG_A6_vminub_RdP(SHORTCODE) \
677     do { \
678         TCGv left = tcg_temp_new(); \
679         TCGv right = tcg_temp_new(); \
680         TCGv tmp = tcg_temp_new(); \
681         tcg_gen_movi_tl(PeV, 0); \
682         tcg_gen_movi_i64(RddV, 0); \
683         for (int i = 0; i < 8; i++) { \
684             gen_get_byte_i64(left, i, RttV, false); \
685             gen_get_byte_i64(right, i, RssV, false); \
686             tcg_gen_setcond_tl(TCG_COND_GT, tmp, left, right); \
687             tcg_gen_deposit_tl(PeV, PeV, tmp, i, 1); \
688             tcg_gen_umin_tl(tmp, left, right); \
689             gen_set_byte_i64(i, RddV, tmp); \
690         } \
691     } while (0)
692 
693 #define fGEN_TCG_J2_call(SHORTCODE) \
694     gen_call(ctx, riV)
695 #define fGEN_TCG_J2_callr(SHORTCODE) \
696     gen_callr(ctx, RsV)
697 #define fGEN_TCG_J2_callrh(SHORTCODE) \
698     gen_callr(ctx, RsV)
699 
700 #define fGEN_TCG_J2_callt(SHORTCODE) \
701     gen_cond_call(ctx, PuV, TCG_COND_EQ, riV)
702 #define fGEN_TCG_J2_callf(SHORTCODE) \
703     gen_cond_call(ctx, PuV, TCG_COND_NE, riV)
704 #define fGEN_TCG_J2_callrt(SHORTCODE) \
705     gen_cond_callr(ctx, TCG_COND_EQ, PuV, RsV)
706 #define fGEN_TCG_J2_callrf(SHORTCODE) \
707     gen_cond_callr(ctx, TCG_COND_NE, PuV, RsV)
708 
709 #define fGEN_TCG_J2_loop0r(SHORTCODE) \
710     gen_loop0r(ctx, RsV, riV)
711 #define fGEN_TCG_J2_loop1r(SHORTCODE) \
712     gen_loop1r(ctx, RsV, riV)
713 #define fGEN_TCG_J2_loop0i(SHORTCODE) \
714     gen_loop0i(ctx, UiV, riV)
715 #define fGEN_TCG_J2_loop1i(SHORTCODE) \
716     gen_loop1i(ctx, UiV, riV)
717 #define fGEN_TCG_J2_ploop1sr(SHORTCODE) \
718     gen_ploopNsr(ctx, 1, RsV, riV)
719 #define fGEN_TCG_J2_ploop1si(SHORTCODE) \
720     gen_ploopNsi(ctx, 1, UiV, riV)
721 #define fGEN_TCG_J2_ploop2sr(SHORTCODE) \
722     gen_ploopNsr(ctx, 2, RsV, riV)
723 #define fGEN_TCG_J2_ploop2si(SHORTCODE) \
724     gen_ploopNsi(ctx, 2, UiV, riV)
725 #define fGEN_TCG_J2_ploop3sr(SHORTCODE) \
726     gen_ploopNsr(ctx, 3, RsV, riV)
727 #define fGEN_TCG_J2_ploop3si(SHORTCODE) \
728     gen_ploopNsi(ctx, 3, UiV, riV)
729 
730 #define fGEN_TCG_J2_endloop0(SHORTCODE) \
731     gen_endloop0(ctx)
732 #define fGEN_TCG_J2_endloop1(SHORTCODE) \
733     gen_endloop1(ctx)
734 #define fGEN_TCG_J2_endloop01(SHORTCODE) \
735     gen_endloop01(ctx)
736 
737 /*
738  * Compound compare and jump instructions
739  * Here is a primer to understand the tag names
740  *
741  * Comparison
742  *      cmpeqi   compare equal to an immediate
743  *      cmpgti   compare greater than an immediate
744  *      cmpgtiu  compare greater than an unsigned immediate
745  *      cmpeqn1  compare equal to negative 1
746  *      cmpgtn1  compare greater than negative 1
747  *      cmpeq    compare equal (two registers)
748  *      cmpgtu   compare greater than unsigned (two registers)
749  *      tstbit0  test bit zero
750  *
751  * Condition
752  *      tp0      p0 is true     p0 = cmp.eq(r0,#5); if (p0.new) jump:nt address
753  *      fp0      p0 is false    p0 = cmp.eq(r0,#5); if (!p0.new) jump:nt address
754  *      tp1      p1 is true     p1 = cmp.eq(r0,#5); if (p1.new) jump:nt address
755  *      fp1      p1 is false    p1 = cmp.eq(r0,#5); if (!p1.new) jump:nt address
756  *
757  * Prediction (not modelled in qemu)
758  *      _nt      not taken
759  *      _t       taken
760  */
761 #define fGEN_TCG_J4_cmpeq_tp0_jump_t(SHORTCODE) \
762     gen_cmpnd_cmp_jmp_t(ctx, 0, TCG_COND_EQ, RsV, RtV, riV)
763 #define fGEN_TCG_J4_cmpeq_tp0_jump_nt(SHORTCODE) \
764     gen_cmpnd_cmp_jmp_t(ctx, 0, TCG_COND_EQ, RsV, RtV, riV)
765 #define fGEN_TCG_J4_cmpeq_fp0_jump_t(SHORTCODE) \
766     gen_cmpnd_cmp_jmp_f(ctx, 0, TCG_COND_EQ, RsV, RtV, riV)
767 #define fGEN_TCG_J4_cmpeq_fp0_jump_nt(SHORTCODE) \
768     gen_cmpnd_cmp_jmp_f(ctx, 0, TCG_COND_EQ, RsV, RtV, riV)
769 #define fGEN_TCG_J4_cmpeq_tp1_jump_t(SHORTCODE) \
770     gen_cmpnd_cmp_jmp_t(ctx, 1, TCG_COND_EQ, RsV, RtV, riV)
771 #define fGEN_TCG_J4_cmpeq_tp1_jump_nt(SHORTCODE) \
772     gen_cmpnd_cmp_jmp_t(ctx, 1, TCG_COND_EQ, RsV, RtV, riV)
773 #define fGEN_TCG_J4_cmpeq_fp1_jump_t(SHORTCODE) \
774     gen_cmpnd_cmp_jmp_f(ctx, 1, TCG_COND_EQ, RsV, RtV, riV)
775 #define fGEN_TCG_J4_cmpeq_fp1_jump_nt(SHORTCODE) \
776     gen_cmpnd_cmp_jmp_f(ctx, 1, TCG_COND_EQ, RsV, RtV, riV)
777 
778 #define fGEN_TCG_J4_cmpgt_tp0_jump_t(SHORTCODE) \
779     gen_cmpnd_cmp_jmp_t(ctx, 0, TCG_COND_GT, RsV, RtV, riV)
780 #define fGEN_TCG_J4_cmpgt_tp0_jump_nt(SHORTCODE) \
781     gen_cmpnd_cmp_jmp_t(ctx, 0, TCG_COND_GT, RsV, RtV, riV)
782 #define fGEN_TCG_J4_cmpgt_fp0_jump_t(SHORTCODE) \
783     gen_cmpnd_cmp_jmp_f(ctx, 0, TCG_COND_GT, RsV, RtV, riV)
784 #define fGEN_TCG_J4_cmpgt_fp0_jump_nt(SHORTCODE) \
785     gen_cmpnd_cmp_jmp_f(ctx, 0, TCG_COND_GT, RsV, RtV, riV)
786 #define fGEN_TCG_J4_cmpgt_tp1_jump_t(SHORTCODE) \
787     gen_cmpnd_cmp_jmp_t(ctx, 1, TCG_COND_GT, RsV, RtV, riV)
788 #define fGEN_TCG_J4_cmpgt_tp1_jump_nt(SHORTCODE) \
789     gen_cmpnd_cmp_jmp_t(ctx, 1, TCG_COND_GT, RsV, RtV, riV)
790 #define fGEN_TCG_J4_cmpgt_fp1_jump_t(SHORTCODE) \
791     gen_cmpnd_cmp_jmp_f(ctx, 1, TCG_COND_GT, RsV, RtV, riV)
792 #define fGEN_TCG_J4_cmpgt_fp1_jump_nt(SHORTCODE) \
793     gen_cmpnd_cmp_jmp_f(ctx, 1, TCG_COND_GT, RsV, RtV, riV)
794 
795 #define fGEN_TCG_J4_cmpgtu_tp0_jump_t(SHORTCODE) \
796     gen_cmpnd_cmp_jmp_t(ctx, 0, TCG_COND_GTU, RsV, RtV, riV)
797 #define fGEN_TCG_J4_cmpgtu_tp0_jump_nt(SHORTCODE) \
798     gen_cmpnd_cmp_jmp_t(ctx, 0, TCG_COND_GTU, RsV, RtV, riV)
799 #define fGEN_TCG_J4_cmpgtu_fp0_jump_t(SHORTCODE) \
800     gen_cmpnd_cmp_jmp_f(ctx, 0, TCG_COND_GTU, RsV, RtV, riV)
801 #define fGEN_TCG_J4_cmpgtu_fp0_jump_nt(SHORTCODE) \
802     gen_cmpnd_cmp_jmp_f(ctx, 0, TCG_COND_GTU, RsV, RtV, riV)
803 #define fGEN_TCG_J4_cmpgtu_tp1_jump_t(SHORTCODE) \
804     gen_cmpnd_cmp_jmp_t(ctx, 1, TCG_COND_GTU, RsV, RtV, riV)
805 #define fGEN_TCG_J4_cmpgtu_tp1_jump_nt(SHORTCODE) \
806     gen_cmpnd_cmp_jmp_t(ctx, 1, TCG_COND_GTU, RsV, RtV, riV)
807 #define fGEN_TCG_J4_cmpgtu_fp1_jump_t(SHORTCODE) \
808     gen_cmpnd_cmp_jmp_f(ctx, 1, TCG_COND_GTU, RsV, RtV, riV)
809 #define fGEN_TCG_J4_cmpgtu_fp1_jump_nt(SHORTCODE) \
810     gen_cmpnd_cmp_jmp_f(ctx, 1, TCG_COND_GTU, RsV, RtV, riV)
811 
812 #define fGEN_TCG_J4_cmpeqi_tp0_jump_t(SHORTCODE) \
813     gen_cmpnd_cmpi_jmp_t(ctx, 0, TCG_COND_EQ, RsV, UiV, riV)
814 #define fGEN_TCG_J4_cmpeqi_tp0_jump_nt(SHORTCODE) \
815     gen_cmpnd_cmpi_jmp_t(ctx, 0, TCG_COND_EQ, RsV, UiV, riV)
816 #define fGEN_TCG_J4_cmpeqi_fp0_jump_t(SHORTCODE) \
817     gen_cmpnd_cmpi_jmp_f(ctx, 0, TCG_COND_EQ, RsV, UiV, riV)
818 #define fGEN_TCG_J4_cmpeqi_fp0_jump_nt(SHORTCODE) \
819     gen_cmpnd_cmpi_jmp_f(ctx, 0, TCG_COND_EQ, RsV, UiV, riV)
820 #define fGEN_TCG_J4_cmpeqi_tp1_jump_t(SHORTCODE) \
821     gen_cmpnd_cmpi_jmp_t(ctx, 1, TCG_COND_EQ, RsV, UiV, riV)
822 #define fGEN_TCG_J4_cmpeqi_tp1_jump_nt(SHORTCODE) \
823     gen_cmpnd_cmpi_jmp_t(ctx, 1, TCG_COND_EQ, RsV, UiV, riV)
824 #define fGEN_TCG_J4_cmpeqi_fp1_jump_t(SHORTCODE) \
825     gen_cmpnd_cmpi_jmp_f(ctx, 1, TCG_COND_EQ, RsV, UiV, riV)
826 #define fGEN_TCG_J4_cmpeqi_fp1_jump_nt(SHORTCODE) \
827     gen_cmpnd_cmpi_jmp_f(ctx, 1, TCG_COND_EQ, RsV, UiV, riV)
828 
829 #define fGEN_TCG_J4_cmpgti_tp0_jump_t(SHORTCODE) \
830     gen_cmpnd_cmpi_jmp_t(ctx, 0, TCG_COND_GT, RsV, UiV, riV)
831 #define fGEN_TCG_J4_cmpgti_tp0_jump_nt(SHORTCODE) \
832     gen_cmpnd_cmpi_jmp_t(ctx, 0, TCG_COND_GT, RsV, UiV, riV)
833 #define fGEN_TCG_J4_cmpgti_fp0_jump_t(SHORTCODE) \
834     gen_cmpnd_cmpi_jmp_f(ctx, 0, TCG_COND_GT, RsV, UiV, riV)
835 #define fGEN_TCG_J4_cmpgti_fp0_jump_nt(SHORTCODE) \
836     gen_cmpnd_cmpi_jmp_f(ctx, 0, TCG_COND_GT, RsV, UiV, riV)
837 #define fGEN_TCG_J4_cmpgti_tp1_jump_t(SHORTCODE) \
838     gen_cmpnd_cmpi_jmp_t(ctx, 1, TCG_COND_GT, RsV, UiV, riV)
839 #define fGEN_TCG_J4_cmpgti_tp1_jump_nt(SHORTCODE) \
840     gen_cmpnd_cmpi_jmp_t(ctx, 1, TCG_COND_GT, RsV, UiV, riV)
841 #define fGEN_TCG_J4_cmpgti_fp1_jump_t(SHORTCODE) \
842     gen_cmpnd_cmpi_jmp_f(ctx, 1, TCG_COND_GT, RsV, UiV, riV)
843 #define fGEN_TCG_J4_cmpgti_fp1_jump_nt(SHORTCODE) \
844     gen_cmpnd_cmpi_jmp_f(ctx, 1, TCG_COND_GT, RsV, UiV, riV)
845 
846 #define fGEN_TCG_J4_cmpgtui_tp0_jump_t(SHORTCODE) \
847     gen_cmpnd_cmpi_jmp_t(ctx, 0, TCG_COND_GTU, RsV, UiV, riV)
848 #define fGEN_TCG_J4_cmpgtui_tp0_jump_nt(SHORTCODE) \
849     gen_cmpnd_cmpi_jmp_t(ctx, 0, TCG_COND_GTU, RsV, UiV, riV)
850 #define fGEN_TCG_J4_cmpgtui_fp0_jump_t(SHORTCODE) \
851     gen_cmpnd_cmpi_jmp_f(ctx, 0, TCG_COND_GTU, RsV, UiV, riV)
852 #define fGEN_TCG_J4_cmpgtui_fp0_jump_nt(SHORTCODE) \
853     gen_cmpnd_cmpi_jmp_f(ctx, 0, TCG_COND_GTU, RsV, UiV, riV)
854 #define fGEN_TCG_J4_cmpgtui_tp1_jump_t(SHORTCODE) \
855     gen_cmpnd_cmpi_jmp_t(ctx, 1, TCG_COND_GTU, RsV, UiV, riV)
856 #define fGEN_TCG_J4_cmpgtui_tp1_jump_nt(SHORTCODE) \
857     gen_cmpnd_cmpi_jmp_t(ctx, 1, TCG_COND_GTU, RsV, UiV, riV)
858 #define fGEN_TCG_J4_cmpgtui_fp1_jump_t(SHORTCODE) \
859     gen_cmpnd_cmpi_jmp_f(ctx, 1, TCG_COND_GTU, RsV, UiV, riV)
860 #define fGEN_TCG_J4_cmpgtui_fp1_jump_nt(SHORTCODE) \
861     gen_cmpnd_cmpi_jmp_f(ctx, 1, TCG_COND_GTU, RsV, UiV, riV)
862 
863 #define fGEN_TCG_J4_cmpeqn1_tp0_jump_t(SHORTCODE) \
864     gen_cmpnd_cmp_n1_jmp_t(ctx, 0, TCG_COND_EQ, RsV, riV)
865 #define fGEN_TCG_J4_cmpeqn1_tp0_jump_nt(SHORTCODE) \
866     gen_cmpnd_cmp_n1_jmp_t(ctx, 0, TCG_COND_EQ, RsV, riV)
867 #define fGEN_TCG_J4_cmpeqn1_fp0_jump_t(SHORTCODE) \
868     gen_cmpnd_cmp_n1_jmp_f(ctx, 0, TCG_COND_EQ, RsV, riV)
869 #define fGEN_TCG_J4_cmpeqn1_fp0_jump_nt(SHORTCODE) \
870     gen_cmpnd_cmp_n1_jmp_f(ctx, 0, TCG_COND_EQ, RsV, riV)
871 #define fGEN_TCG_J4_cmpeqn1_tp1_jump_t(SHORTCODE) \
872     gen_cmpnd_cmp_n1_jmp_t(ctx, 1, TCG_COND_EQ, RsV, riV)
873 #define fGEN_TCG_J4_cmpeqn1_tp1_jump_nt(SHORTCODE) \
874     gen_cmpnd_cmp_n1_jmp_t(ctx, 1, TCG_COND_EQ, RsV, riV)
875 #define fGEN_TCG_J4_cmpeqn1_fp1_jump_t(SHORTCODE) \
876     gen_cmpnd_cmp_n1_jmp_f(ctx, 1, TCG_COND_EQ, RsV, riV)
877 #define fGEN_TCG_J4_cmpeqn1_fp1_jump_nt(SHORTCODE) \
878     gen_cmpnd_cmp_n1_jmp_f(ctx, 1, TCG_COND_EQ, RsV, riV)
879 
880 #define fGEN_TCG_J4_cmpgtn1_tp0_jump_t(SHORTCODE) \
881     gen_cmpnd_cmp_n1_jmp_t(ctx, 0, TCG_COND_GT, RsV, riV)
882 #define fGEN_TCG_J4_cmpgtn1_tp0_jump_nt(SHORTCODE) \
883     gen_cmpnd_cmp_n1_jmp_t(ctx, 0, TCG_COND_GT, RsV, riV)
884 #define fGEN_TCG_J4_cmpgtn1_fp0_jump_t(SHORTCODE) \
885     gen_cmpnd_cmp_n1_jmp_f(ctx, 0, TCG_COND_GT, RsV, riV)
886 #define fGEN_TCG_J4_cmpgtn1_fp0_jump_nt(SHORTCODE) \
887     gen_cmpnd_cmp_n1_jmp_f(ctx, 0, TCG_COND_GT, RsV, riV)
888 #define fGEN_TCG_J4_cmpgtn1_tp1_jump_t(SHORTCODE) \
889     gen_cmpnd_cmp_n1_jmp_t(ctx, 1, TCG_COND_GT, RsV, riV)
890 #define fGEN_TCG_J4_cmpgtn1_tp1_jump_nt(SHORTCODE) \
891     gen_cmpnd_cmp_n1_jmp_t(ctx, 1, TCG_COND_GT, RsV, riV)
892 #define fGEN_TCG_J4_cmpgtn1_fp1_jump_t(SHORTCODE) \
893     gen_cmpnd_cmp_n1_jmp_f(ctx, 1, TCG_COND_GT, RsV, riV)
894 #define fGEN_TCG_J4_cmpgtn1_fp1_jump_nt(SHORTCODE) \
895     gen_cmpnd_cmp_n1_jmp_f(ctx, 1, TCG_COND_GT, RsV, riV)
896 
897 #define fGEN_TCG_J4_tstbit0_tp0_jump_nt(SHORTCODE) \
898     gen_cmpnd_tstbit0_jmp(ctx, 0, RsV, TCG_COND_EQ, riV)
899 #define fGEN_TCG_J4_tstbit0_tp0_jump_t(SHORTCODE) \
900     gen_cmpnd_tstbit0_jmp(ctx, 0, RsV, TCG_COND_EQ, riV)
901 #define fGEN_TCG_J4_tstbit0_fp0_jump_nt(SHORTCODE) \
902     gen_cmpnd_tstbit0_jmp(ctx, 0, RsV, TCG_COND_NE, riV)
903 #define fGEN_TCG_J4_tstbit0_fp0_jump_t(SHORTCODE) \
904     gen_cmpnd_tstbit0_jmp(ctx, 0, RsV, TCG_COND_NE, riV)
905 #define fGEN_TCG_J4_tstbit0_tp1_jump_nt(SHORTCODE) \
906     gen_cmpnd_tstbit0_jmp(ctx, 1, RsV, TCG_COND_EQ, riV)
907 #define fGEN_TCG_J4_tstbit0_tp1_jump_t(SHORTCODE) \
908     gen_cmpnd_tstbit0_jmp(ctx, 1, RsV, TCG_COND_EQ, riV)
909 #define fGEN_TCG_J4_tstbit0_fp1_jump_nt(SHORTCODE) \
910     gen_cmpnd_tstbit0_jmp(ctx, 1, RsV, TCG_COND_NE, riV)
911 #define fGEN_TCG_J4_tstbit0_fp1_jump_t(SHORTCODE) \
912     gen_cmpnd_tstbit0_jmp(ctx, 1, RsV, TCG_COND_NE, riV)
913 
914 /* p0 = cmp.eq(r0, #7) */
915 #define fGEN_TCG_SA1_cmpeqi(SHORTCODE) \
916     do { \
917         TCGv p0 = tcg_temp_new(); \
918         gen_comparei(TCG_COND_EQ, p0, RsV, uiV); \
919         gen_log_pred_write(ctx, 0, p0); \
920     } while (0)
921 
922 #define fGEN_TCG_J2_jump(SHORTCODE) \
923     gen_jump(ctx, riV)
924 #define fGEN_TCG_J2_jumpr(SHORTCODE) \
925     gen_jumpr(ctx, RsV)
926 #define fGEN_TCG_J2_jumprh(SHORTCODE) \
927     gen_jumpr(ctx, RsV)
928 #define fGEN_TCG_J4_jumpseti(SHORTCODE) \
929     do { \
930         tcg_gen_movi_tl(RdV, UiV); \
931         gen_jump(ctx, riV); \
932     } while (0)
933 
934 #define fGEN_TCG_cond_jumpt(COND) \
935     do { \
936         TCGv LSB = tcg_temp_new(); \
937         COND; \
938         gen_cond_jump(ctx, TCG_COND_EQ, LSB, riV); \
939     } while (0)
940 #define fGEN_TCG_cond_jumpf(COND) \
941     do { \
942         TCGv LSB = tcg_temp_new(); \
943         COND; \
944         gen_cond_jump(ctx, TCG_COND_NE, LSB, riV); \
945     } while (0)
946 
947 #define fGEN_TCG_J2_jumpt(SHORTCODE) \
948     fGEN_TCG_cond_jumpt(fLSBOLD(PuV))
949 #define fGEN_TCG_J2_jumptpt(SHORTCODE) \
950     fGEN_TCG_cond_jumpt(fLSBOLD(PuV))
951 #define fGEN_TCG_J2_jumpf(SHORTCODE) \
952     fGEN_TCG_cond_jumpf(fLSBOLD(PuV))
953 #define fGEN_TCG_J2_jumpfpt(SHORTCODE) \
954     fGEN_TCG_cond_jumpf(fLSBOLD(PuV))
955 #define fGEN_TCG_J2_jumptnew(SHORTCODE) \
956     gen_cond_jump(ctx, TCG_COND_EQ, PuN, riV)
957 #define fGEN_TCG_J2_jumptnewpt(SHORTCODE) \
958     gen_cond_jump(ctx, TCG_COND_EQ, PuN, riV)
959 #define fGEN_TCG_J2_jumpfnewpt(SHORTCODE) \
960     fGEN_TCG_cond_jumpf(fLSBNEW(PuN))
961 #define fGEN_TCG_J2_jumpfnew(SHORTCODE) \
962     fGEN_TCG_cond_jumpf(fLSBNEW(PuN))
963 #define fGEN_TCG_J2_jumprz(SHORTCODE) \
964     fGEN_TCG_cond_jumpt(tcg_gen_setcondi_tl(TCG_COND_NE, LSB, RsV, 0))
965 #define fGEN_TCG_J2_jumprzpt(SHORTCODE) \
966     fGEN_TCG_cond_jumpt(tcg_gen_setcondi_tl(TCG_COND_NE, LSB, RsV, 0))
967 #define fGEN_TCG_J2_jumprnz(SHORTCODE) \
968     fGEN_TCG_cond_jumpt(tcg_gen_setcondi_tl(TCG_COND_EQ, LSB, RsV, 0))
969 #define fGEN_TCG_J2_jumprnzpt(SHORTCODE) \
970     fGEN_TCG_cond_jumpt(tcg_gen_setcondi_tl(TCG_COND_EQ, LSB, RsV, 0))
971 #define fGEN_TCG_J2_jumprgtez(SHORTCODE) \
972     fGEN_TCG_cond_jumpt(tcg_gen_setcondi_tl(TCG_COND_GE, LSB, RsV, 0))
973 #define fGEN_TCG_J2_jumprgtezpt(SHORTCODE) \
974     fGEN_TCG_cond_jumpt(tcg_gen_setcondi_tl(TCG_COND_GE, LSB, RsV, 0))
975 #define fGEN_TCG_J2_jumprltez(SHORTCODE) \
976     fGEN_TCG_cond_jumpt(tcg_gen_setcondi_tl(TCG_COND_LE, LSB, RsV, 0))
977 #define fGEN_TCG_J2_jumprltezpt(SHORTCODE) \
978     fGEN_TCG_cond_jumpt(tcg_gen_setcondi_tl(TCG_COND_LE, LSB, RsV, 0))
979 
980 #define fGEN_TCG_cond_jumprt(COND) \
981     do { \
982         TCGv LSB = tcg_temp_new(); \
983         COND; \
984         gen_cond_jumpr(ctx, RsV, TCG_COND_EQ, LSB); \
985     } while (0)
986 #define fGEN_TCG_cond_jumprf(COND) \
987     do { \
988         TCGv LSB = tcg_temp_new(); \
989         COND; \
990         gen_cond_jumpr(ctx, RsV, TCG_COND_NE, LSB); \
991     } while (0)
992 
993 #define fGEN_TCG_J2_jumprt(SHORTCODE) \
994     fGEN_TCG_cond_jumprt(fLSBOLD(PuV))
995 #define fGEN_TCG_J2_jumprtpt(SHORTCODE) \
996     fGEN_TCG_cond_jumprt(fLSBOLD(PuV))
997 #define fGEN_TCG_J2_jumprf(SHORTCODE) \
998     fGEN_TCG_cond_jumprf(fLSBOLD(PuV))
999 #define fGEN_TCG_J2_jumprfpt(SHORTCODE) \
1000     fGEN_TCG_cond_jumprf(fLSBOLD(PuV))
1001 #define fGEN_TCG_J2_jumprtnew(SHORTCODE) \
1002     fGEN_TCG_cond_jumprt(fLSBNEW(PuN))
1003 #define fGEN_TCG_J2_jumprtnewpt(SHORTCODE) \
1004     fGEN_TCG_cond_jumprt(fLSBNEW(PuN))
1005 #define fGEN_TCG_J2_jumprfnew(SHORTCODE) \
1006     fGEN_TCG_cond_jumprf(fLSBNEW(PuN))
1007 #define fGEN_TCG_J2_jumprfnewpt(SHORTCODE) \
1008     fGEN_TCG_cond_jumprf(fLSBNEW(PuN))
1009 
1010 /*
1011  * New value compare & jump instructions
1012  * if ([!]COND(r0.new, r1) jump:t address
1013  * if ([!]COND(r0.new, #7) jump:t address
1014  */
1015 #define fGEN_TCG_J4_cmpgt_t_jumpnv_t(SHORTCODE) \
1016     gen_cmp_jumpnv(ctx, TCG_COND_GT, NsN, RtV, riV)
1017 #define fGEN_TCG_J4_cmpgt_t_jumpnv_nt(SHORTCODE) \
1018     gen_cmp_jumpnv(ctx, TCG_COND_GT, NsN, RtV, riV)
1019 #define fGEN_TCG_J4_cmpgt_f_jumpnv_t(SHORTCODE) \
1020     gen_cmp_jumpnv(ctx, TCG_COND_LE, NsN, RtV, riV)
1021 #define fGEN_TCG_J4_cmpgt_f_jumpnv_nt(SHORTCODE) \
1022     gen_cmp_jumpnv(ctx, TCG_COND_LE, NsN, RtV, riV)
1023 
1024 #define fGEN_TCG_J4_cmpeq_t_jumpnv_t(SHORTCODE) \
1025     gen_cmp_jumpnv(ctx, TCG_COND_EQ, NsN, RtV, riV)
1026 #define fGEN_TCG_J4_cmpeq_t_jumpnv_nt(SHORTCODE) \
1027     gen_cmp_jumpnv(ctx, TCG_COND_EQ, NsN, RtV, riV)
1028 #define fGEN_TCG_J4_cmpeq_f_jumpnv_t(SHORTCODE) \
1029     gen_cmp_jumpnv(ctx, TCG_COND_NE, NsN, RtV, riV)
1030 #define fGEN_TCG_J4_cmpeq_f_jumpnv_nt(SHORTCODE) \
1031     gen_cmp_jumpnv(ctx, TCG_COND_NE, NsN, RtV, riV)
1032 
1033 #define fGEN_TCG_J4_cmplt_t_jumpnv_t(SHORTCODE) \
1034     gen_cmp_jumpnv(ctx, TCG_COND_LT, NsN, RtV, riV)
1035 #define fGEN_TCG_J4_cmplt_t_jumpnv_nt(SHORTCODE) \
1036     gen_cmp_jumpnv(ctx, TCG_COND_LT, NsN, RtV, riV)
1037 #define fGEN_TCG_J4_cmplt_f_jumpnv_t(SHORTCODE) \
1038     gen_cmp_jumpnv(ctx, TCG_COND_GE, NsN, RtV, riV)
1039 #define fGEN_TCG_J4_cmplt_f_jumpnv_nt(SHORTCODE) \
1040     gen_cmp_jumpnv(ctx, TCG_COND_GE, NsN, RtV, riV)
1041 
1042 #define fGEN_TCG_J4_cmpeqi_t_jumpnv_t(SHORTCODE) \
1043     gen_cmpi_jumpnv(ctx, TCG_COND_EQ, NsN, UiV, riV)
1044 #define fGEN_TCG_J4_cmpeqi_t_jumpnv_nt(SHORTCODE) \
1045     gen_cmpi_jumpnv(ctx, TCG_COND_EQ, NsN, UiV, riV)
1046 #define fGEN_TCG_J4_cmpeqi_f_jumpnv_t(SHORTCODE) \
1047     gen_cmpi_jumpnv(ctx, TCG_COND_NE, NsN, UiV, riV)
1048 #define fGEN_TCG_J4_cmpeqi_f_jumpnv_nt(SHORTCODE) \
1049     gen_cmpi_jumpnv(ctx, TCG_COND_NE, NsN, UiV, riV)
1050 
1051 #define fGEN_TCG_J4_cmpgti_t_jumpnv_t(SHORTCODE) \
1052     gen_cmpi_jumpnv(ctx, TCG_COND_GT, NsN, UiV, riV)
1053 #define fGEN_TCG_J4_cmpgti_t_jumpnv_nt(SHORTCODE) \
1054     gen_cmpi_jumpnv(ctx, TCG_COND_GT, NsN, UiV, riV)
1055 #define fGEN_TCG_J4_cmpgti_f_jumpnv_t(SHORTCODE) \
1056     gen_cmpi_jumpnv(ctx, TCG_COND_LE, NsN, UiV, riV)
1057 #define fGEN_TCG_J4_cmpgti_f_jumpnv_nt(SHORTCODE) \
1058     gen_cmpi_jumpnv(ctx, TCG_COND_LE, NsN, UiV, riV)
1059 
1060 #define fGEN_TCG_J4_cmpltu_t_jumpnv_t(SHORTCODE) \
1061     gen_cmp_jumpnv(ctx, TCG_COND_LTU, NsN, RtV, riV)
1062 #define fGEN_TCG_J4_cmpltu_t_jumpnv_nt(SHORTCODE) \
1063     gen_cmp_jumpnv(ctx, TCG_COND_LTU, NsN, RtV, riV)
1064 #define fGEN_TCG_J4_cmpltu_f_jumpnv_t(SHORTCODE) \
1065     gen_cmp_jumpnv(ctx, TCG_COND_GEU, NsN, RtV, riV)
1066 #define fGEN_TCG_J4_cmpltu_f_jumpnv_nt(SHORTCODE) \
1067     gen_cmp_jumpnv(ctx, TCG_COND_GEU, NsN, RtV, riV)
1068 
1069 #define fGEN_TCG_J4_cmpgtui_t_jumpnv_t(SHORTCODE) \
1070     gen_cmpi_jumpnv(ctx, TCG_COND_GTU, NsN, UiV, riV)
1071 #define fGEN_TCG_J4_cmpgtui_t_jumpnv_nt(SHORTCODE) \
1072     gen_cmpi_jumpnv(ctx, TCG_COND_GTU, NsN, UiV, riV)
1073 #define fGEN_TCG_J4_cmpgtui_f_jumpnv_t(SHORTCODE) \
1074     gen_cmpi_jumpnv(ctx, TCG_COND_LEU, NsN, UiV, riV)
1075 #define fGEN_TCG_J4_cmpgtui_f_jumpnv_nt(SHORTCODE) \
1076     gen_cmpi_jumpnv(ctx, TCG_COND_LEU, NsN, UiV, riV)
1077 
1078 #define fGEN_TCG_J4_cmpgtu_t_jumpnv_t(SHORTCODE) \
1079     gen_cmp_jumpnv(ctx, TCG_COND_GTU, NsN, RtV, riV)
1080 #define fGEN_TCG_J4_cmpgtu_t_jumpnv_nt(SHORTCODE) \
1081     gen_cmp_jumpnv(ctx, TCG_COND_GTU, NsN, RtV, riV)
1082 #define fGEN_TCG_J4_cmpgtu_f_jumpnv_t(SHORTCODE) \
1083     gen_cmp_jumpnv(ctx, TCG_COND_LEU, NsN, RtV, riV)
1084 #define fGEN_TCG_J4_cmpgtu_f_jumpnv_nt(SHORTCODE) \
1085     gen_cmp_jumpnv(ctx, TCG_COND_LEU, NsN, RtV, riV)
1086 
1087 #define fGEN_TCG_J4_cmpeqn1_t_jumpnv_t(SHORTCODE) \
1088     gen_cmpi_jumpnv(ctx, TCG_COND_EQ, NsN, -1, riV)
1089 #define fGEN_TCG_J4_cmpeqn1_t_jumpnv_nt(SHORTCODE) \
1090     gen_cmpi_jumpnv(ctx, TCG_COND_EQ, NsN, -1, riV)
1091 #define fGEN_TCG_J4_cmpeqn1_f_jumpnv_t(SHORTCODE) \
1092     gen_cmpi_jumpnv(ctx, TCG_COND_NE, NsN, -1, riV)
1093 #define fGEN_TCG_J4_cmpeqn1_f_jumpnv_nt(SHORTCODE) \
1094     gen_cmpi_jumpnv(ctx, TCG_COND_NE, NsN, -1, riV)
1095 
1096 #define fGEN_TCG_J4_cmpgtn1_t_jumpnv_t(SHORTCODE) \
1097     gen_cmpi_jumpnv(ctx, TCG_COND_GT, NsN, -1, riV)
1098 #define fGEN_TCG_J4_cmpgtn1_t_jumpnv_nt(SHORTCODE) \
1099     gen_cmpi_jumpnv(ctx, TCG_COND_GT, NsN, -1, riV)
1100 #define fGEN_TCG_J4_cmpgtn1_f_jumpnv_t(SHORTCODE) \
1101     gen_cmpi_jumpnv(ctx, TCG_COND_LE, NsN, -1, riV)
1102 #define fGEN_TCG_J4_cmpgtn1_f_jumpnv_nt(SHORTCODE) \
1103     gen_cmpi_jumpnv(ctx, TCG_COND_LE, NsN, -1, riV)
1104 
1105 #define fGEN_TCG_J4_tstbit0_t_jumpnv_t(SHORTCODE) \
1106     gen_testbit0_jumpnv(ctx, NsN, TCG_COND_EQ, riV)
1107 #define fGEN_TCG_J4_tstbit0_t_jumpnv_nt(SHORTCODE) \
1108     gen_testbit0_jumpnv(ctx, NsN, TCG_COND_EQ, riV)
1109 #define fGEN_TCG_J4_tstbit0_f_jumpnv_t(SHORTCODE) \
1110     gen_testbit0_jumpnv(ctx, NsN, TCG_COND_NE, riV)
1111 #define fGEN_TCG_J4_tstbit0_f_jumpnv_nt(SHORTCODE) \
1112     gen_testbit0_jumpnv(ctx, NsN, TCG_COND_NE, riV)
1113 
1114 /* r0 = r1 ; jump address */
1115 #define fGEN_TCG_J4_jumpsetr(SHORTCODE) \
1116     do { \
1117         tcg_gen_mov_tl(RdV, RsV); \
1118         gen_jump(ctx, riV); \
1119     } while (0)
1120 
1121 /* if (p0.new) r0 = #0 */
1122 #define fGEN_TCG_SA1_clrtnew(SHORTCODE) \
1123     do { \
1124         tcg_gen_movcond_tl(TCG_COND_EQ, RdV, \
1125                            ctx->new_pred_value[0], tcg_constant_tl(0), \
1126                            RdV, tcg_constant_tl(0)); \
1127     } while (0)
1128 
1129 /* if (!p0.new) r0 = #0 */
1130 #define fGEN_TCG_SA1_clrfnew(SHORTCODE) \
1131     do { \
1132         tcg_gen_movcond_tl(TCG_COND_NE, RdV, \
1133                            ctx->new_pred_value[0], tcg_constant_tl(0), \
1134                            RdV, tcg_constant_tl(0)); \
1135     } while (0)
1136 
1137 #define fGEN_TCG_J2_pause(SHORTCODE) \
1138     do { \
1139         uiV = uiV; \
1140         tcg_gen_movi_tl(hex_gpr[HEX_REG_PC], ctx->next_PC); \
1141     } while (0)
1142 
1143 /* r0 = asr(r1, r2):sat */
1144 #define fGEN_TCG_S2_asr_r_r_sat(SHORTCODE) \
1145     gen_asr_r_r_sat(ctx, RdV, RsV, RtV)
1146 
1147 /* r0 = asl(r1, r2):sat */
1148 #define fGEN_TCG_S2_asl_r_r_sat(SHORTCODE) \
1149     gen_asl_r_r_sat(ctx, RdV, RsV, RtV)
1150 
1151 #define fGEN_TCG_SL2_jumpr31(SHORTCODE) \
1152     gen_jumpr(ctx, hex_gpr[HEX_REG_LR])
1153 
1154 #define fGEN_TCG_SL2_jumpr31_t(SHORTCODE) \
1155     gen_cond_jumpr31(ctx, TCG_COND_EQ, hex_pred[0])
1156 #define fGEN_TCG_SL2_jumpr31_f(SHORTCODE) \
1157     gen_cond_jumpr31(ctx, TCG_COND_NE, hex_pred[0])
1158 
1159 #define fGEN_TCG_SL2_jumpr31_tnew(SHORTCODE) \
1160     gen_cond_jumpr31(ctx, TCG_COND_EQ, ctx->new_pred_value[0])
1161 #define fGEN_TCG_SL2_jumpr31_fnew(SHORTCODE) \
1162     gen_cond_jumpr31(ctx, TCG_COND_NE, ctx->new_pred_value[0])
1163 
1164 /* Count trailing zeros/ones */
1165 #define fGEN_TCG_S2_ct0(SHORTCODE) \
1166     do { \
1167         tcg_gen_ctzi_tl(RdV, RsV, 32); \
1168     } while (0)
1169 #define fGEN_TCG_S2_ct1(SHORTCODE) \
1170     do { \
1171         tcg_gen_not_tl(RdV, RsV); \
1172         tcg_gen_ctzi_tl(RdV, RdV, 32); \
1173     } while (0)
1174 #define fGEN_TCG_S2_ct0p(SHORTCODE) \
1175     do { \
1176         TCGv_i64 tmp = tcg_temp_new_i64(); \
1177         tcg_gen_ctzi_i64(tmp, RssV, 64); \
1178         tcg_gen_extrl_i64_i32(RdV, tmp); \
1179     } while (0)
1180 #define fGEN_TCG_S2_ct1p(SHORTCODE) \
1181     do { \
1182         TCGv_i64 tmp = tcg_temp_new_i64(); \
1183         tcg_gen_not_i64(tmp, RssV); \
1184         tcg_gen_ctzi_i64(tmp, tmp, 64); \
1185         tcg_gen_extrl_i64_i32(RdV, tmp); \
1186     } while (0)
1187 
1188 #define fGEN_TCG_S2_insert(SHORTCODE) \
1189     do { \
1190         int width = uiV; \
1191         int offset = UiV; \
1192         if (width != 0) { \
1193             if (offset + width > 32) { \
1194                 width = 32 - offset; \
1195             } \
1196             tcg_gen_deposit_tl(RxV, RxV, RsV, offset, width); \
1197         } \
1198     } while (0)
1199 #define fGEN_TCG_S2_insert_rp(SHORTCODE) \
1200     gen_insert_rp(ctx, RxV, RsV, RttV)
1201 #define fGEN_TCG_S2_asr_r_svw_trun(SHORTCODE) \
1202     gen_asr_r_svw_trun(ctx, RdV, RssV, RtV)
1203 #define fGEN_TCG_A2_swiz(SHORTCODE) \
1204     tcg_gen_bswap_tl(RdV, RsV)
1205 
1206 /* Floating point */
1207 #define fGEN_TCG_F2_conv_sf2df(SHORTCODE) \
1208     gen_helper_conv_sf2df(RddV, cpu_env, RsV)
1209 #define fGEN_TCG_F2_conv_df2sf(SHORTCODE) \
1210     gen_helper_conv_df2sf(RdV, cpu_env, RssV)
1211 #define fGEN_TCG_F2_conv_uw2sf(SHORTCODE) \
1212     gen_helper_conv_uw2sf(RdV, cpu_env, RsV)
1213 #define fGEN_TCG_F2_conv_uw2df(SHORTCODE) \
1214     gen_helper_conv_uw2df(RddV, cpu_env, RsV)
1215 #define fGEN_TCG_F2_conv_w2sf(SHORTCODE) \
1216     gen_helper_conv_w2sf(RdV, cpu_env, RsV)
1217 #define fGEN_TCG_F2_conv_w2df(SHORTCODE) \
1218     gen_helper_conv_w2df(RddV, cpu_env, RsV)
1219 #define fGEN_TCG_F2_conv_ud2sf(SHORTCODE) \
1220     gen_helper_conv_ud2sf(RdV, cpu_env, RssV)
1221 #define fGEN_TCG_F2_conv_ud2df(SHORTCODE) \
1222     gen_helper_conv_ud2df(RddV, cpu_env, RssV)
1223 #define fGEN_TCG_F2_conv_d2sf(SHORTCODE) \
1224     gen_helper_conv_d2sf(RdV, cpu_env, RssV)
1225 #define fGEN_TCG_F2_conv_d2df(SHORTCODE) \
1226     gen_helper_conv_d2df(RddV, cpu_env, RssV)
1227 #define fGEN_TCG_F2_conv_sf2uw(SHORTCODE) \
1228     gen_helper_conv_sf2uw(RdV, cpu_env, RsV)
1229 #define fGEN_TCG_F2_conv_sf2w(SHORTCODE) \
1230     gen_helper_conv_sf2w(RdV, cpu_env, RsV)
1231 #define fGEN_TCG_F2_conv_sf2ud(SHORTCODE) \
1232     gen_helper_conv_sf2ud(RddV, cpu_env, RsV)
1233 #define fGEN_TCG_F2_conv_sf2d(SHORTCODE) \
1234     gen_helper_conv_sf2d(RddV, cpu_env, RsV)
1235 #define fGEN_TCG_F2_conv_df2uw(SHORTCODE) \
1236     gen_helper_conv_df2uw(RdV, cpu_env, RssV)
1237 #define fGEN_TCG_F2_conv_df2w(SHORTCODE) \
1238     gen_helper_conv_df2w(RdV, cpu_env, RssV)
1239 #define fGEN_TCG_F2_conv_df2ud(SHORTCODE) \
1240     gen_helper_conv_df2ud(RddV, cpu_env, RssV)
1241 #define fGEN_TCG_F2_conv_df2d(SHORTCODE) \
1242     gen_helper_conv_df2d(RddV, cpu_env, RssV)
1243 #define fGEN_TCG_F2_conv_sf2uw_chop(SHORTCODE) \
1244     gen_helper_conv_sf2uw_chop(RdV, cpu_env, RsV)
1245 #define fGEN_TCG_F2_conv_sf2w_chop(SHORTCODE) \
1246     gen_helper_conv_sf2w_chop(RdV, cpu_env, RsV)
1247 #define fGEN_TCG_F2_conv_sf2ud_chop(SHORTCODE) \
1248     gen_helper_conv_sf2ud_chop(RddV, cpu_env, RsV)
1249 #define fGEN_TCG_F2_conv_sf2d_chop(SHORTCODE) \
1250     gen_helper_conv_sf2d_chop(RddV, cpu_env, RsV)
1251 #define fGEN_TCG_F2_conv_df2uw_chop(SHORTCODE) \
1252     gen_helper_conv_df2uw_chop(RdV, cpu_env, RssV)
1253 #define fGEN_TCG_F2_conv_df2w_chop(SHORTCODE) \
1254     gen_helper_conv_df2w_chop(RdV, cpu_env, RssV)
1255 #define fGEN_TCG_F2_conv_df2ud_chop(SHORTCODE) \
1256     gen_helper_conv_df2ud_chop(RddV, cpu_env, RssV)
1257 #define fGEN_TCG_F2_conv_df2d_chop(SHORTCODE) \
1258     gen_helper_conv_df2d_chop(RddV, cpu_env, RssV)
1259 #define fGEN_TCG_F2_sfadd(SHORTCODE) \
1260     gen_helper_sfadd(RdV, cpu_env, RsV, RtV)
1261 #define fGEN_TCG_F2_sfsub(SHORTCODE) \
1262     gen_helper_sfsub(RdV, cpu_env, RsV, RtV)
1263 #define fGEN_TCG_F2_sfcmpeq(SHORTCODE) \
1264     gen_helper_sfcmpeq(PdV, cpu_env, RsV, RtV)
1265 #define fGEN_TCG_F2_sfcmpgt(SHORTCODE) \
1266     gen_helper_sfcmpgt(PdV, cpu_env, RsV, RtV)
1267 #define fGEN_TCG_F2_sfcmpge(SHORTCODE) \
1268     gen_helper_sfcmpge(PdV, cpu_env, RsV, RtV)
1269 #define fGEN_TCG_F2_sfcmpuo(SHORTCODE) \
1270     gen_helper_sfcmpuo(PdV, cpu_env, RsV, RtV)
1271 #define fGEN_TCG_F2_sfmax(SHORTCODE) \
1272     gen_helper_sfmax(RdV, cpu_env, RsV, RtV)
1273 #define fGEN_TCG_F2_sfmin(SHORTCODE) \
1274     gen_helper_sfmin(RdV, cpu_env, RsV, RtV)
1275 #define fGEN_TCG_F2_sfclass(SHORTCODE) \
1276     do { \
1277         TCGv imm = tcg_constant_tl(uiV); \
1278         gen_helper_sfclass(PdV, cpu_env, RsV, imm); \
1279     } while (0)
1280 #define fGEN_TCG_F2_sffixupn(SHORTCODE) \
1281     gen_helper_sffixupn(RdV, cpu_env, RsV, RtV)
1282 #define fGEN_TCG_F2_sffixupd(SHORTCODE) \
1283     gen_helper_sffixupd(RdV, cpu_env, RsV, RtV)
1284 #define fGEN_TCG_F2_sffixupr(SHORTCODE) \
1285     gen_helper_sffixupr(RdV, cpu_env, RsV)
1286 #define fGEN_TCG_F2_dfadd(SHORTCODE) \
1287     gen_helper_dfadd(RddV, cpu_env, RssV, RttV)
1288 #define fGEN_TCG_F2_dfsub(SHORTCODE) \
1289     gen_helper_dfsub(RddV, cpu_env, RssV, RttV)
1290 #define fGEN_TCG_F2_dfmax(SHORTCODE) \
1291     gen_helper_dfmax(RddV, cpu_env, RssV, RttV)
1292 #define fGEN_TCG_F2_dfmin(SHORTCODE) \
1293     gen_helper_dfmin(RddV, cpu_env, RssV, RttV)
1294 #define fGEN_TCG_F2_dfcmpeq(SHORTCODE) \
1295     gen_helper_dfcmpeq(PdV, cpu_env, RssV, RttV)
1296 #define fGEN_TCG_F2_dfcmpgt(SHORTCODE) \
1297     gen_helper_dfcmpgt(PdV, cpu_env, RssV, RttV)
1298 #define fGEN_TCG_F2_dfcmpge(SHORTCODE) \
1299     gen_helper_dfcmpge(PdV, cpu_env, RssV, RttV)
1300 #define fGEN_TCG_F2_dfcmpuo(SHORTCODE) \
1301     gen_helper_dfcmpuo(PdV, cpu_env, RssV, RttV)
1302 #define fGEN_TCG_F2_dfclass(SHORTCODE) \
1303     do { \
1304         TCGv imm = tcg_constant_tl(uiV); \
1305         gen_helper_dfclass(PdV, cpu_env, RssV, imm); \
1306     } while (0)
1307 #define fGEN_TCG_F2_sfmpy(SHORTCODE) \
1308     gen_helper_sfmpy(RdV, cpu_env, RsV, RtV)
1309 #define fGEN_TCG_F2_sffma(SHORTCODE) \
1310     gen_helper_sffma(RxV, cpu_env, RxV, RsV, RtV)
1311 #define fGEN_TCG_F2_sffma_sc(SHORTCODE) \
1312     gen_helper_sffma_sc(RxV, cpu_env, RxV, RsV, RtV, PuV)
1313 #define fGEN_TCG_F2_sffms(SHORTCODE) \
1314     gen_helper_sffms(RxV, cpu_env, RxV, RsV, RtV)
1315 #define fGEN_TCG_F2_sffma_lib(SHORTCODE) \
1316     gen_helper_sffma_lib(RxV, cpu_env, RxV, RsV, RtV)
1317 #define fGEN_TCG_F2_sffms_lib(SHORTCODE) \
1318     gen_helper_sffms_lib(RxV, cpu_env, RxV, RsV, RtV)
1319 
1320 #define fGEN_TCG_F2_dfmpyfix(SHORTCODE) \
1321     gen_helper_dfmpyfix(RddV, cpu_env, RssV, RttV)
1322 #define fGEN_TCG_F2_dfmpyhh(SHORTCODE) \
1323     gen_helper_dfmpyhh(RxxV, cpu_env, RxxV, RssV, RttV)
1324 
1325 /* Nothing to do for these in qemu, need to suppress compiler warnings */
1326 #define fGEN_TCG_Y4_l2fetch(SHORTCODE) \
1327     do { \
1328         RsV = RsV; \
1329         RtV = RtV; \
1330     } while (0)
1331 #define fGEN_TCG_Y5_l2fetch(SHORTCODE) \
1332     do { \
1333         RsV = RsV; \
1334     } while (0)
1335 #define fGEN_TCG_Y2_isync(SHORTCODE) \
1336     do { } while (0)
1337 #define fGEN_TCG_Y2_barrier(SHORTCODE) \
1338     do { } while (0)
1339 #define fGEN_TCG_Y2_syncht(SHORTCODE) \
1340     do { } while (0)
1341 #define fGEN_TCG_Y2_dcfetchbo(SHORTCODE) \
1342     do { \
1343         RsV = RsV; \
1344         uiV = uiV; \
1345     } while (0)
1346 
1347 #define fGEN_TCG_L2_loadw_aq(SHORTCODE)                 SHORTCODE
1348 #define fGEN_TCG_L4_loadd_aq(SHORTCODE)                 SHORTCODE
1349 
1350 /* Nothing to do for these in qemu, need to suppress compiler warnings */
1351 #define fGEN_TCG_R6_release_at_vi(SHORTCODE) \
1352     do { \
1353         RsV = RsV; \
1354     } while (0)
1355 #define fGEN_TCG_R6_release_st_vi(SHORTCODE) \
1356     do { \
1357         RsV = RsV; \
1358     } while (0)
1359 
1360 #define fGEN_TCG_S2_storew_rl_at_vi(SHORTCODE)          SHORTCODE
1361 #define fGEN_TCG_S4_stored_rl_at_vi(SHORTCODE)          SHORTCODE
1362 #define fGEN_TCG_S2_storew_rl_st_vi(SHORTCODE)          SHORTCODE
1363 #define fGEN_TCG_S4_stored_rl_st_vi(SHORTCODE)          SHORTCODE
1364 
1365 #define fGEN_TCG_J2_trap0(SHORTCODE) \
1366     do { \
1367         uiV = uiV; \
1368         tcg_gen_movi_tl(hex_gpr[HEX_REG_PC], ctx->pkt->pc); \
1369         TCGv excp = tcg_constant_tl(HEX_EXCP_TRAP0); \
1370         gen_helper_raise_exception(cpu_env, excp); \
1371     } while (0)
1372 #endif
1373