On Thu, 26 Dec 2024 at 08:28, Hao Wu <wuhao...@google.com> wrote:
>
> These 2 values are different between NPCM7XX and NPCM8XX
> GCRs. So we add them to the class and assign different values
> to them.
>
> Signed-off-by: Hao Wu <wuhao...@google.com>
> ---
> hw/misc/npcm_gcr.c | 24 +++++++++++++++---------
> include/hw/misc/npcm_gcr.h | 13 +++++++++++--
> 2 files changed, 26 insertions(+), 11 deletions(-)
>
> @@ -156,10 +158,10 @@ static const struct MemoryRegionOps npcm_gcr_ops = {
> static void npcm7xx_gcr_enter_reset(Object *obj, ResetType type)
> {
> NPCMGCRState *s = NPCM_GCR(obj);
> + NPCMGCRClass *c = NPCM_GCR_GET_CLASS(obj);
>
> - QEMU_BUILD_BUG_ON(sizeof(s->regs) != sizeof(cold_reset_values));
> -
> - memcpy(s->regs, cold_reset_values, sizeof(s->regs));
> + memcpy(s->regs, c->cold_reset_values, c->nr_regs * sizeof(uint32_t));
Previously we had a compile-time assert that we were copying
the right amount of data. Making this a class-specific
thing means we can't assert at compile time any more, but
I think it's still worth a runtime assert that we aren't
copying more data than will fit into s->regs.
Otherwise
Reviewed-by: Peter Maydell <peter.mayd...@linaro.org>
thanks
-- PMM
