On 11.02.2026 10:47, Oleksii Kurochko wrote: > On 2/11/26 8:49 AM, Jan Beulich wrote: >> On 09.02.2026 17:52, Oleksii Kurochko wrote: >>> --- a/xen/arch/riscv/setup.c >>> +++ b/xen/arch/riscv/setup.c >>> @@ -32,6 +32,8 @@ >>> unsigned char __initdata cpu0_boot_stack[STACK_SIZE] >>> __aligned(STACK_SIZE); >>> >>> +struct csr_masks __ro_after_init csr_masks; >> setup.c would be nice to only have __init functions and __initdata data. >> Really up to now that's the case, and I wonder why the makefile doesn't >> leverage this by using setup.init.o in place of setup.o. This variable >> would likely better live elsewhere anyway, imo: Somewhere it's actually >> (going to be) used. > > I put it here as I wasn't able to find better place. If it is okay to have > it in domain.c then I'm okay to move this variable there.
If that's where it's going to be (mainly) used (as the next patch suggests), I see nothing speaking against you doing so. Jan
