On Tue, Aug 19, 2025 at 07:19:53PM +0800, Huacai Chen wrote: > In __arch_prepare_bpf_trampoline(), retval_off is meaningful only when > save_ret is not 0, so the current logic is correct. But it may cause a > build warning: We pass uninitialized data to invoke_bpf_prog(), it's just that it isn't used. In the kernel, we would say that's not a bug if invoke_bpf_prog() is inlined. (The C standard doesn't differentiate between inlined and not inlined, passing uninitialized variables is always considered undefined behavior). UBSan will complain about the uninitialized variable at runtime as well depending on if it's inlined or not. It's not marked as an __always_inline function... regards, dan carpenter
