On 04.11.2021 16:24, Roger Pau Monné wrote: > It might be appropriate to switch pirq_dpci to: > > #define pirq_dpci(d, pirq) \ > ((is_hvm_domain(d) && (pirq)) ? &(pirq)->arch.hvm.dpci : NULL)
I don't see how this would help suppress the warning. > Or to an inline helper. I expect it's a macro because an inline function would cause header dependency issues. Plus both that or any other attempt to leverage some means to circumvent gcc recognizing the pattern is only going to help until they further enhance their recognition. Jan
