target/arm/cpu64: Use 32-bit GDBstub when running in 32-bit KVM mode
PATCH v2 1/1 target/arm/cpu64: Use 32-bit GDBstub when running in 32-bit KVM mode
When running under KVM, we may decide to run the CPU in 32-bit mode, by setting the 'aarch64=off' CPU option. In this case, we need to switch to the 32-bit version of the GDB stub too, so that GDB has the correct view of the CPU state. Without this, GDB debugging does not work at all, and errors out upon connecting to the target with a mysterious 'g' packet length error.
---
v2: refactor existing CPUClass::gdb_... member assignments for the
32-bit code so we can reuse it for the 64-bit code
target/arm/cpu.c | 16 +++++++++++----- target/arm/cpu.h | 2 ++ target/arm/cpu64.c | 3 +++ 3 files changed, 16 insertions(+), 5 deletions(-)
target/arm/cpu.c +11 -5
@@ -2049,6 +2049,15 @@ static const struct TCGCPUOps arm_tcg_ops = { };#endif /* CONFIG_TCG */ +void arm_cpu_class_gdb_init (CPUClass *cc) +{ + cc->gdb_read_register = arm_cpu_gdb_read_register; + cc->gdb_write_register = arm_cpu_gdb_write_register; + cc->gdb_num_core_regs =26 ; + cc->gdb_core_xml_file ="arm-core.xml" ; + cc->gdb_arch_name = arm_gdb_arch_name; +} +static void arm_cpu_class_init (ObjectClass *oc,void *data) {ARMCPUClass *acc =ARM_CPU_CLASS (oc);
@@ -2061,18 +2070,15 @@ static void arm_cpu_class_init(ObjectClass *oc, void *data)device_class_set_props (dc, arm_cpu_properties);device_class_set_parent_reset (dc, arm_cpu_reset, &acc->parent_reset); +arm_cpu_class_gdb_init (cc); + cc->class_by_name = arm_cpu_class_by_name; cc->has_work = arm_cpu_has_work; cc->dump_state = arm_cpu_dump_state; cc->set_pc = arm_cpu_set_pc; - cc->gdb_read_register = arm_cpu_gdb_read_register; - cc->gdb_write_register = arm_cpu_gdb_write_register;#ifndef CONFIG_USER_ONLY cc->sysemu_ops = &arm_sysemu_ops;#endif - cc->gdb_num_core_regs =26 ; - cc->gdb_core_xml_file ="arm-core.xml" ; - cc->gdb_arch_name = arm_gdb_arch_name; cc->gdb_get_dynamic_xml = arm_gdb_get_dynamic_xml; cc->gdb_stop_before_watchpoint = true; cc->disas_set_info = arm_disas_set_info;
target/arm/cpu.h +2 -0
@@ -1064,6 +1064,8 @@ int arm_gen_dynamic_svereg_xml(CPUState *cpu, int base_reg);*/ const char *arm_gdb_get_dynamic_xml (CPUState *cpu,const char *xmlname); +void arm_cpu_class_gdb_init (CPUClass *cc); +int arm_cpu_write_elf64_note (WriteCoreDumpFunction f,CPUState *cs,int cpuid,void *opaque); int arm_cpu_write_elf32_note(WriteCoreDumpFunction f, CPUState *cs,
target/arm/cpu64.c +3 -0
@@ -906,6 +906,7 @@ static bool aarch64_cpu_get_aarch64(Object *obj, Error **errp)static void aarch64_cpu_set_aarch64 (Object *obj,bool value,Error **errp) {ARMCPU *cpu =ARM_CPU (obj); +CPUClass *cc =CPU_GET_CLASS (obj);/* At this time, this property is only allowed if KVM is enabled. This * restriction allows us to avoid fixing up functionality that assumes a
@@ -919,6 +920,8 @@ static void aarch64_cpu_set_aarch64(Object *obj, bool value, Error **errp)return ; }unset_feature (&cpu->env,ARM_FEATURE_AARCH64 ); + + arm_cpu_class_gdb_init(cc) }else {set_feature (&cpu->env,ARM_FEATURE_AARCH64 ); }
Signature · 1 line
2.30.2
Quoted 11 lines · Ard Biesheuvel
When running under KVM, we may decide to run the CPU in 32-bit mode, by setting the 'aarch64=off' CPU option. In this case, we need to switch to the 32-bit version of the GDB stub too, so that GDB has the correct view of the CPU state. Without this, GDB debugging does not work at all, and errors out upon connecting to the target with a mysterious 'g' packet length error.
Cc Richard Henderson<[email protected]>Cc Peter Maydell<[email protected]>Cc Alex Bennee<[email protected]>Signed-off-by Ard Biesheuvel<[email protected]>
Quoted 8 lines
---
v2: refactor existingCPUClass::gdb_... member assignments for the
32-bit code so we can reuse it for the 64-bit codetarget/arm/cpu.c | 16 +++++++++++----- target/arm/cpu.h | 2 ++ target/arm/cpu64.c | 3 +++ 3 files changed, 16 insertions(+), 5 deletions(-)
r~
Quoted 7 lines · Ard Biesheuvel
When running under KVM, we may decide to run the CPU in 32-bit mode, by setting the 'aarch64=off' CPU option. In this case, we need to switch to the 32-bit version of the GDB stub too, so that GDB has the correct view of the CPU state. Without this, GDB debugging does not work at all, and errors out upon connecting to the target with a mysterious 'g' packet length error.
Quoted 7 lines
Cc Richard Henderson <[email protected]>Cc Peter Maydell <[email protected]>Cc Alex Bennee <[email protected]>Signed-off-by Ard Biesheuvel <[email protected]>---
v2: refactor existingCPUClass::gdb_... member assignments for the
32-bit code so we can reuse it for the 64-bit code
Way cleaner.
Quoted 11 lines · Ard Biesheuvel
When running under KVM, we may decide to run the CPU in 32-bit mode, by setting the 'aarch64=off' CPU option. In this case, we need to switch to the 32-bit version of the GDB stub too, so that GDB has the correct view of the CPU state. Without this, GDB debugging does not work at all, and errors out upon connecting to the target with a mysterious 'g' packet length error.
Cc Richard Henderson <[email protected]>Cc Peter Maydell <[email protected]>Cc Alex Bennee <[email protected]>Signed-off-by Ard Biesheuvel <[email protected]>
Signature · 1 line
Alex Bennée
Quoted 27 lines · Ard Biesheuvel
When running under KVM, we may decide to run the CPU in 32-bit mode, by setting the 'aarch64=off' CPU option. In this case, we need to switch to the 32-bit version of the GDB stub too, so that GDB has the correct view of the CPU state. Without this, GDB debugging does not work at all, and errors out upon connecting to the target with a mysterious 'g' packet length error.
Cc Richard Henderson <[email protected]>Cc Peter Maydell <[email protected]>Cc Alex Bennee <[email protected]>Signed-off-by Ard Biesheuvel <[email protected]>---
v2: refactor existingCPUClass::gdb_... member assignments for the
32-bit code so we can reuse it for the 64-bit codetarget/arm/cpu.c | 16 +++++++++++----- target/arm/cpu.h | 2 ++ target/arm/cpu64.c | 3 +++ 3 files changed, 16 insertions(+), 5 deletions(-)target/arm/cpu.c +0
-0@@ -2049,6 +2049,15 @@ static const struct TCGCPUOps arm_tcg_ops = { };
target/arm/cpu.c static const struct TCGCPUOps arm_tcg_ops = { · line 2059
Quoted 10 lines
target/arm/cpu.c +8
-0#endif /* CONFIG_TCG */ +void arm_cpu_class_gdb_init (CPUClass *cc) +{ + cc->gdb_read_register = arm_cpu_gdb_read_register; + cc->gdb_write_register = arm_cpu_gdb_write_register; + cc->gdb_num_core_regs =26 ; + cc->gdb_core_xml_file ="arm-core.xml" ; + cc->gdb_arch_name = arm_gdb_arch_name; +}
Most of these fields are not used by the gdbstub until runtime, but cc->gdb_num_core_regs is used earlier. In particular, in cpu_common_initfn() we copy that value into cpu->gdb_num_regs and cpu->gdb_num_g_regs (this happens at the CPU object's instance_init time, ie before the aarch64_cpu_set_aarch64 function is called), and these are the values that are then used when registering dynamic sysreg XML, coprocessor registers, etc.
target/arm/cpu64.c static bool aarch64_cpu_get_aarch64(Object *obj, Error **errp) · line 909
Quoted 7 lines
target/arm/cpu64.c +1
-0@@ -906,6 +906,7 @@ static bool aarch64_cpu_get_aarch64(Object *obj, Error **errp)static void aarch64_cpu_set_aarch64 (Object *obj,bool value,Error **errp) {ARMCPU *cpu =ARM_CPU (obj); +CPUClass *cc =CPU_GET_CLASS (obj);
This is called to change the property for a specific CPU object -- you can't change the values of the *class* here. (Consider a system with 2 CPUs, one of which has aarch64=yes and one of which has aarch64=no.)
target/arm/cpu64.c static void aarch64_cpu_set_aarch64(Object *obj, bool value, Error **errp) · line 924
Quoted 8 lines
target/arm/cpu64.c +2
-0/* At this time, this property is only allowed if KVM is enabled. This * restriction allows us to avoid fixing up functionality that assumes a @@ -919,6 +920,8 @@ static void aarch64_cpu_set_aarch64(Object *obj, bool value, Error **errp)return ; }unset_feature (&cpu->env,ARM_FEATURE_AARCH64 ); + +arm_cpu_class_gdb_init (cc)
This fails to compile because of the missing semicolon...
target/arm/cpu64.c static void aarch64_cpu_set_aarch64(Object *obj, bool value, Error **errp) · line 926
Quoted 2 lines
target/arm/cpu64.c +0
-0}else {set_feature (&cpu->env,ARM_FEATURE_AARCH64 );
If the user (admittedly slightly perversely) toggles the aarch64 flag from on to off to on again, we should reset the gdb function pointers to the aarch64 versions again.
target/arm/cpu64.c static void aarch64_cpu_set_aarch64(Object *obj, bool value, Error **errp) · line 927
Quoted 1 line
target/arm/cpu64.c +0
-0}
thanks
-- PMM
Quoted 40 lines · Peter Maydell
Quoted 27 lines · Ard Biesheuvel
When running under KVM, we may decide to run the CPU in 32-bit mode, by setting the 'aarch64=off' CPU option. In this case, we need to switch to the 32-bit version of the GDB stub too, so that GDB has the correct view of the CPU state. Without this, GDB debugging does not work at all, and errors out upon connecting to the target with a mysterious 'g' packet length error.
Cc Richard Henderson <[email protected]>Cc Peter Maydell <[email protected]>Cc Alex Bennee <[email protected]>Signed-off-by Ard Biesheuvel <[email protected]>---
v2: refactor existingCPUClass::gdb_... member assignments for the
32-bit code so we can reuse it for the 64-bit codetarget/arm/cpu.c | 16 +++++++++++----- target/arm/cpu.h | 2 ++ target/arm/cpu64.c | 3 +++ 3 files changed, 16 insertions(+), 5 deletions(-)target/arm/cpu.c +0
-0@@ -2049,6 +2049,15 @@ static const struct TCGCPUOps arm_tcg_ops = { };target/arm/cpu.c static const struct TCGCPUOps arm_tcg_ops = { · line 2059
Quoted 10 lines
target/arm/cpu.c +8
-0#endif /* CONFIG_TCG */ +void arm_cpu_class_gdb_init (CPUClass *cc) +{ + cc->gdb_read_register = arm_cpu_gdb_read_register; + cc->gdb_write_register = arm_cpu_gdb_write_register; + cc->gdb_num_core_regs =26 ; + cc->gdb_core_xml_file ="arm-core.xml" ; + cc->gdb_arch_name = arm_gdb_arch_name; +}
Quoted 9 lines
Most of these fields are not used by the gdbstub until runtime, but
cc->gdb_num_core_regsis used earlier. In particular, incpu_common_initfn()we copy that value intocpu->gdb_num_regsandcpu->gdb_num_g_regs(this happens at the CPU object'sinstance_inittime, ie before theaarch64_cpu_set_aarch64function is called), and these are the values that are then used when registering dynamic sysreg XML, coprocessor registers, etc.
Right.
Quoted 4 lines
Quoted 4 lines
target/arm/cpu64.c +0
-0@@ -906,6 +906,7 @@ static bool aarch64_cpu_get_aarch64(Object *obj, Error **errp)static void aarch64_cpu_set_aarch64 (Object *obj,bool value,Error **errp)
Quoted 9 lines
target/arm/cpu64.c static bool aarch64_cpu_get_aarch64(Object *obj, Error **errp) · line 909
Quoted 3 lines
target/arm/cpu64.c +1
-0{ARMCPU *cpu =ARM_CPU (obj); +CPUClass *cc =CPU_GET_CLASS (obj);This is called to change the property for a specific CPU object -- you can't change the values of the *class* here. (Consider a system with 2 CPUs, one of which has aarch64=yes and one of which has aarch64=no.)
So how is this fundamentally going to work then? Which GDB stub should we choose in such a case?
Quoted 11 lines
target/arm/cpu64.c static void aarch64_cpu_set_aarch64(Object *obj, bool value, Error **errp) · line 924
Quoted 8 lines
target/arm/cpu64.c +2
-0/* At this time, this property is only allowed if KVM is enabled. This * restriction allows us to avoid fixing up functionality that assumes a @@ -919,6 +920,8 @@ static void aarch64_cpu_set_aarch64(Object *obj, bool value, Error **errp)return ; }unset_feature (&cpu->env,ARM_FEATURE_AARCH64 ); + +arm_cpu_class_gdb_init (cc)This fails to compile because of the missing semicolon...
Oops, my bad. I spotted this locally as well but failed to fold it into the patch.
Quoted 7 lines
target/arm/cpu64.c static void aarch64_cpu_set_aarch64(Object *obj, bool value, Error **errp) · line 926
Quoted 2 lines
target/arm/cpu64.c +0
-0}else {set_feature (&cpu->env,ARM_FEATURE_AARCH64 );If the user (admittedly slightly perversely) toggles the aarch64 flag from on to off to on again, we should reset the gdb function pointers to the aarch64 versions again.
Ack.
So I can fix most of these issues, but the fundamental one remains, so I'll hold off on a v3 until we can settle that.
Thanks,
Ard.