On 2024/08/02 16:51, Richard Henderson wrote:
On 7/30/24 19:40, Peter Maydell wrote:
From: Akihiko Odaki <akihiko.od...@daynix.com>
Any sysreg access results in an exception unless defined otherwise so
we should raise an exception by default.
Signed-off-by: Akihiko Odaki <akihiko.od...@daynix.com>
Reviewed-by: Peter Maydell <peter.mayd...@linaro.org>
Signed-off-by: Peter Maydell <peter.mayd...@linaro.org>
---
target/arm/hvf/hvf.c | 174 +++++++++++++++++++++----------------------
1 file changed, 85 insertions(+), 89 deletions(-)
This patch fails to compile:
https://gitlab.com/qemu-project/qemu/-/jobs/7489527235
../target/arm/hvf/hvf.c:1283:42: error: incompatible pointer types
passing 'uint64_t **' (aka 'unsigned long long **') to parameter of type
'uint64_t *' (aka 'unsigned long long *'); remove &
[-Werror,-Wincompatible-pointer-types]
if (hvf_sysreg_read_cp(cpu, reg, &val)) {
^~~~
../target/arm/hvf/hvf.c:1175:71: note: passing argument to parameter
'val' here
static bool hvf_sysreg_read_cp(CPUState *cpu, uint32_t reg, uint64_t *val)
^
1 error generated.
This snuck in while our Cirrus build minutes were exhausted, but it
suggests that the patch was never tested at all.
I have just submitted a fix. I'm sorry for causing a mess.
This is a problem in my workflow. I wrote the code on Linux, tested and
fixed it on macOS. However I forgot to synchronize the code between
macOS and Linux before sending it on Linux.
Regards,
Akihiko Odaki
