On 08/18/23 at 07:28pm, Paul E. McKenney wrote: > Hello! > > This morning's rcutorture testing on next-20230818 complained about > prepare_elf_headers() being defined but unused on several rcutorture > scenarios. The patch below makes rcutorture happy, but might or might > not be a proper fix. > > This is a resend adding a few more people on CC, given a possible > relationship to 9f1f399ca999 ("x86/crash: add x86 crash hotplug support"). > > Thoughts? > Thanks for reporting this, Paul. > > ------------------------------------------------------------------------ > > diff --git a/arch/x86/kernel/crash.c b/arch/x86/kernel/crash.c > index 1d0f824559fce..926c39e22387b 100644 > --- a/arch/x86/kernel/crash.c > +++ b/arch/x86/kernel/crash.c > @@ -127,7 +127,8 @@ void native_machine_crash_shutdown(struct pt_regs *regs) > crash_save_cpu(regs, safe_smp_processor_id()); > } > > -#if defined(CONFIG_KEXEC_FILE) || defined(CONFIG_CRASH_DUMP) > +#if (defined(CONFIG_KEXEC_FILE) || defined(CONFIG_CRASH_DUMP)) && defined(CONFIG_SMP) && defined(CONFIG_X86_LOCAL_APIC) Hi Eric, Now prepare_elf_headers() is needed for kexec_file_load and crash hotplug support of kexec_load, change it like this? #if defined(CONFIG_KEXEC_FILE) || defined(CONFIG_CRASH_HOTPLUG)