On 3/12/25 20:45, Richard Henderson wrote:
We were hiding a number of declarations from user-only, although it hurts nothing to allow them. The inlines for user-only are unused.Signed-off-by: Richard Henderson <richard.hender...@linaro.org> --- include/hw/intc/armv7m_nvic.h | 14 -------------- 1 file changed, 14 deletions(-) diff --git a/include/hw/intc/armv7m_nvic.h b/include/hw/intc/armv7m_nvic.h index 89fe8aedaa..7b9964fe7e 100644 --- a/include/hw/intc/armv7m_nvic.h +++ b/include/hw/intc/armv7m_nvic.h @@ -189,21 +189,7 @@ int armv7m_nvic_raw_execution_priority(NVICState *s); * @secure: the security state to test * This corresponds to the pseudocode IsReqExecPriNeg(). */ -#ifndef CONFIG_USER_ONLY bool armv7m_nvic_neg_prio_requested(NVICState *s, bool secure); -#else -static inline bool armv7m_nvic_neg_prio_requested(NVICState *s, bool secure) -{ - return false; -} -#endif -#ifndef CONFIG_USER_ONLY bool armv7m_nvic_can_take_pending_exception(NVICState *s); -#else -static inline bool armv7m_nvic_can_take_pending_exception(NVICState *s) -{ - return true; -} -#endif#endif
I'm a bit worried we might have regression when doing things this way. Maybe we could have a runtime check: config_user_mode() config_system_mode() And assert we are only in system.
