On 29.05.2023 14:13, Oleksii Kurochko wrote:
> --- a/xen/arch/riscv/traps.c
> +++ b/xen/arch/riscv/traps.c
> @@ -12,6 +12,31 @@
> #include <asm/processor.h>
> #include <asm/traps.h>
>
> +#define cast_to_bug_frame(addr) \
> + (const struct bug_frame *)(addr)
I can't find a use for this; should it be dropped or moved to some
later patch? In any event, if ti's intended to survive, it needs yet
another pair of parentheses.
> +/*
> + * Initialize the trap handling.
> + *
> + * The function is called after MMU is enabled.
> + */
> +void trap_init(void)
Is this going to be used for secondary processors as well? If not,
it will want to be __init.
> +{
> + /*
> + * When the MMU is off, addr varialbe will be a physical address
> otherwise
> + * it would be a virtual address.
> + *
> + * It will work fine as:
> + * - access to addr is PC-relative.
> + * - -nopie is used. -nopie really suppresses the compiler emitting
> + * code going through .got (which then indeed would mean using
> absolute
> + * addresses).
> + */
Is all of this comment still relevant not that you're running with
the MMU already enabled.
Jan
> + unsigned long addr = (unsigned long)&handle_trap;
> +
> + csr_write(CSR_STVEC, addr);
> +}
> +
> static const char *decode_trap_cause(unsigned long cause)
> {
> static const char *const trap_causes[] = {
