On 10.07.2025 13:19, Oleksii Kurochko wrote: > On 7/8/25 3:58 PM, Jan Beulich wrote: >> On 07.07.2025 11:01, Oleksii Kurochko wrote: >>> +static struct aplic_priv aplic; >> Is this altered post-init? IOW can it be __ro_after_init? Or otherwise at >> least __read_mostly? > > Looking at the current downstream code there is, at least one case, where > aplic->regs > are changing (during vaplic_emulate_store()). > So __read_mostly would be better in this case. > >> With these two taken care of (one way or another): >> Acked-by: Jan Beulich<jbeul...@suse.com> >> If you clarify what is wanted / needed, I'm also happy to make adjustments >> while committing. > > It would be nice to do the following: > - Drop the inclusion of <asm/imsic.h> in aplic-priv.h. > - Add __read_mostly to the definition of the aplic variable.
Actually - no. Patch 4 adds a spin lock in the structure. That way it's definitely not mostly read. Jan
