* Why on earth is this code giving me Segfaults?
@ 2019-09-29 4:04 Libo Zhou
2019-09-29 7:39 ` Libo Zhou
0 siblings, 1 reply; 2+ messages in thread
From: Libo Zhou @ 2019-09-29 4:04 UTC (permalink / raw)
To: qemu-devel
Hi All,
I have a custom ISA that's based on MIPS. The LW and SW instructions' opcodes are changed into 0x17(OPC_BGTZL) and 0x1F(OPC_SPECIAL3).
I have made the following changes in target/mips/translate.c:
@@ -29331,7 +29331,11 @@ static void decode_opc(CPUMIPSState *env, DisasContext *ctx)
decode_opc_special3(env, ctx);
}
#else
- decode_opc_special3(env, ctx);
+ if (ctx->insn_flags & INSN_MYISA) {
+ gen_st(ctx, OPC_SW, rt, rs, imm); /* OPC_MYISA_SW */
+ } else {
+ decode_opc_special3(env, ctx);
+ }
#endif
break;
case OPC_REGIMM:
@@ -29589,6 +29603,8 @@ static void decode_opc(CPUMIPSState *env, DisasContext *ctx)
}
/* OPC_BGTZC, OPC_BLTZC, OPC_BLTC */
gen_compute_compact_branch(ctx, op, rs, rt, imm << 2);
+ } else if (ctx->insn_flags & INSN_MYISA) {
+ gen_ld(ctx, OPC_LW, rt, rs, imm); /* OPC_MYISA_LW */
} else {
/* OPC_BGTZL */
gen_compute_branch(ctx, op, 4, rs, rt, imm << 2, 4);
I used gdbstub to singlestep my program, and I found that my sw instruction is working fine, but lw gives me a segfault. I have been stuck on this for a long while, since it looks like I only need to add that one line of gen_ld function. I also tried debugging QEMU wtih gdb, but the segfault wasn't thrown immediately after lw instruction like gdbstub does.
Does anyone have any advice?
Thanks,
Libo Zhou
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re:Why on earth is this code giving me Segfaults?
2019-09-29 4:04 Why on earth is this code giving me Segfaults? Libo Zhou
@ 2019-09-29 7:39 ` Libo Zhou
0 siblings, 0 replies; 2+ messages in thread
From: Libo Zhou @ 2019-09-29 7:39 UTC (permalink / raw)
To: qemu-devel
I'm copying and pasting all my diff below, just in case something else caused the segfault. Any help would be really appreciated guys.
diff --git a/target/mips/mips-defs.h b/target/mips/mips-defs.h
index bbf056a5..8a4acff3 100644
--- a/target/mips/mips-defs.h
+++ b/target/mips/mips-defs.h
@@ -50,6 +50,7 @@
#define INSN_LOONGSON2F 0x0002000000000000ULL
#define INSN_VR54XX 0x0004000000000000ULL
#define INSN_R5900 0x0008000000000000ULL
+#define INSN_MAOTU 0x0010000000000000ULL
/*
* bits 56-63: vendor-specific ASEs
*/
@@ -91,6 +92,9 @@
/* Wave Computing: "nanoMIPS" */
#define CPU_NANOMIPS32 (CPU_MIPS32R6 | ISA_NANOMIPS32)
+/* Sylincom: "maotu" */
+#define CPU_MAOTU (CPU_MIPS1 | INSN_MAOTU)
+
/* Strictly follow the architecture standard:
- Disallow "special" instruction handling for PMON/SPIM.
Note that we still maintain Count/Compare to match the host clock. */
diff --git a/target/mips/translate.c b/target/mips/translate.c
index ca628002..4e4b6e89 100644
--- a/target/mips/translate.c
+++ b/target/mips/translate.c
@@ -29234,7 +29234,11 @@ static void decode_opc(CPUMIPSState *env, DisasContext *ctx)
decode_opc_special3(env, ctx);
}
#else
- decode_opc_special3(env, ctx);
+ if (ctx->insn_flags & INSN_MAOTU) {
+ gen_st(ctx, OPC_SW, rt, rs, imm); /* OPC_MAOTU_SW */
+ } else {
+ decode_opc_special3(env, ctx);
+ }
#endif
break;
case OPC_REGIMM:
@@ -29484,7 +29488,7 @@ static void decode_opc(CPUMIPSState *env, DisasContext *ctx)
gen_compute_branch(ctx, op, 4, rs, rt, imm << 2, 4);
}
break;
- case OPC_BGTZC: /* OPC_BLTZC, OPC_BLTC, OPC_BGTZL */
+ case OPC_BGTZC: /* OPC_BLTZC, OPC_BLTC, OPC_BGTZL, OPC_MAOTU_LW */
if (ctx->insn_flags & ISA_MIPS32R6) {
if (rt == 0) {
generate_exception_end(ctx, EXCP_RI);
@@ -29492,6 +29496,9 @@ static void decode_opc(CPUMIPSState *env, DisasContext *ctx)
}
/* OPC_BGTZC, OPC_BLTZC, OPC_BLTC */
gen_compute_compact_branch(ctx, op, rs, rt, imm << 2);
+ } else if (ctx->insn_flags & INSN_MAOTU) {
+ /* OPC_MAOTU_LW */
+ gen_ld(ctx, OPC_LW, rt, rs, imm);
} else {
/* OPC_BGTZL */
gen_compute_branch(ctx, op, 4, rs, rt, imm << 2, 4);
diff --git a/target/mips/translate_init.inc.c b/target/mips/translate_init.inc.c
index 6d145a90..5b1a7cef 100644
--- a/target/mips/translate_init.inc.c
+++ b/target/mips/translate_init.inc.c
@@ -489,6 +489,10 @@ const mips_def_t mips_defs[] =
ASE_MT,
.mmu_type = MMU_TYPE_R4000,
},
+ {
+ .name = "maotu",
+ .insn_flags = CPU_MAOTU,
+ },
#if defined(TARGET_MIPS64)
{
.name = "R4000",
^ permalink raw reply related [flat|nested] 2+ messages in thread
end of thread, other threads:[~2019-09-29 7:41 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2019-09-29 4:04 Why on earth is this code giving me Segfaults? Libo Zhou
2019-09-29 7:39 ` Libo Zhou
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).