| Message ID | 20230904161235.84651-4-philmd@linaro.org |
|---|---|
| State | Superseded |
| Headers | show |
| Series | (few more) Steps towards enabling -Wshadow | expand |
On Mon, 4 Sept 2023 at 17:12, Philippe Mathieu-Daudé <philmd@linaro.org> wrote: > > Per Peter Maydell analysis [*]: > > The hvf_vcpu_exec() function is not documented, but in practice > its caller expects it to return either EXCP_DEBUG (for "this was > a guest debug exception you need to deal with") or something else > (presumably the intention being 0 for OK). > > The hvf_sysreg_read() and hvf_sysreg_write() functions are also not > documented, but they return 0 on success, or 1 for a completely > unrecognized sysreg where we've raised the UNDEF exception (but > not if we raised an UNDEF exception for an unrecognized GIC sysreg -- > I think this is a bug). We use this return value to decide whether > we need to advance the PC past the insn or not. It's not the same > as the return value we want to return from hvf_vcpu_exec(). > > Retain the variable as locally scoped but give it a name that > doesn't clash with the other function-scoped variable. > > This fixes: > > target/arm/hvf/hvf.c:1936:13: error: declaration shadows a local variable [-Werror,-Wshadow] > int ret = 0; > ^ > target/arm/hvf/hvf.c:1807:9: note: previous declaration is here > int ret; > ^ > [*] https://lore.kernel.org/qemu-devel/CAFEAcA_e+fU6JKtS+W63wr9cCJ6btu_hT_ydZWOwC0kBkDYYYQ@mail.gmail.com/ > > Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> > --- Reviewed-by: Peter Maydell <peter.maydell@linaro.org> thanks -- PMM
diff --git a/target/arm/hvf/hvf.c b/target/arm/hvf/hvf.c index 486f90be1d..0715f8a01c 100644 --- a/target/arm/hvf/hvf.c +++ b/target/arm/hvf/hvf.c @@ -1933,16 +1933,16 @@ int hvf_vcpu_exec(CPUState *cpu) uint32_t rt = (syndrome >> 5) & 0x1f; uint32_t reg = syndrome & SYSREG_MASK; uint64_t val; - int ret = 0; + int sysreg_ret = 0; if (isread) { - ret = hvf_sysreg_read(cpu, reg, rt); + sysreg_ret = hvf_sysreg_read(cpu, reg, rt); } else { val = hvf_get_reg(cpu, rt); - ret = hvf_sysreg_write(cpu, reg, val); + sysreg_ret = hvf_sysreg_write(cpu, reg, val); } - advance_pc = !ret; + advance_pc = !sysreg_ret; break; } case EC_WFX_TRAP:
Per Peter Maydell analysis [*]: The hvf_vcpu_exec() function is not documented, but in practice its caller expects it to return either EXCP_DEBUG (for "this was a guest debug exception you need to deal with") or something else (presumably the intention being 0 for OK). The hvf_sysreg_read() and hvf_sysreg_write() functions are also not documented, but they return 0 on success, or 1 for a completely unrecognized sysreg where we've raised the UNDEF exception (but not if we raised an UNDEF exception for an unrecognized GIC sysreg -- I think this is a bug). We use this return value to decide whether we need to advance the PC past the insn or not. It's not the same as the return value we want to return from hvf_vcpu_exec(). Retain the variable as locally scoped but give it a name that doesn't clash with the other function-scoped variable. This fixes: target/arm/hvf/hvf.c:1936:13: error: declaration shadows a local variable [-Werror,-Wshadow] int ret = 0; ^ target/arm/hvf/hvf.c:1807:9: note: previous declaration is here int ret; ^ [*] https://lore.kernel.org/qemu-devel/CAFEAcA_e+fU6JKtS+W63wr9cCJ6btu_hT_ydZWOwC0kBkDYYYQ@mail.gmail.com/ Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> --- Peter, feel free to alter the commit description if it doesn't sound right. --- target/arm/hvf/hvf.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)