Re: mmotm 2022-02-11-15-07 uploaded (objtool: ftrace_likely_update)

[Peter Zijlstra <peterz@xxxxxxxxxxxxx>]

On Fri, Feb 11, 2022 at 06:32:06PM -0800, Randy Dunlap wrote:
> on x86_64:
> 
> $ gcc --version
> gcc (SUSE Linux) 7.5.0
> 
> 
> vmlinux.o: warning: objtool: fixup_bad_iret()+0x72: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: noist_exc_debug()+0x30: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: exc_nmi()+0x14e: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: poke_int3_handler()+0x55: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: __sev_get_ghcb()+0x1b: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: __sev_put_ghcb()+0x18: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: __sev_es_ist_exit()+0x2b: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: kernel_exc_vmm_communication()+0x4f: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: rcu_dynticks_eqs_exit()+0x2c: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: rcu_dynticks_eqs_enter()+0x29: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: rcu_eqs_exit.constprop.62()+0x1a: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: rcu_eqs_enter.constprop.63()+0x23: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: rcu_nmi_enter()+0x17: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: irqentry_nmi_enter()+0x36: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: irqentry_nmi_exit()+0xa1: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: enter_from_user_mode()+0x4f: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: syscall_enter_from_user_mode()+0x53: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: syscall_enter_from_user_mode_prepare()+0x4f: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: irqentry_enter_from_user_mode()+0x4f: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: context_tracking_recursion_enter()+0x67: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: __context_tracking_enter()+0x2c: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: fixup_bad_iret()+0x72: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: noist_exc_debug()+0x30: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: exc_nmi()+0x14e: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: poke_int3_handler()+0x55: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: __sev_get_ghcb()+0x1b: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: __sev_put_ghcb()+0x18: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: __sev_es_ist_exit()+0x2b: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: kernel_exc_vmm_communication()+0x4f: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: rcu_dynticks_eqs_exit()+0x2c: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: rcu_dynticks_eqs_enter()+0x29: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: rcu_eqs_exit.constprop.62()+0x1a: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: rcu_eqs_enter.constprop.63()+0x23: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: rcu_nmi_enter()+0x17: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: irqentry_nmi_enter()+0x36: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: irqentry_nmi_exit()+0xa1: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: enter_from_user_mode()+0x4f: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: syscall_enter_from_user_mode()+0x53: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: syscall_enter_from_user_mode_prepare()+0x4f: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: irqentry_enter_from_user_mode()+0x4f: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: context_tracking_recursion_enter()+0x67: call to ftrace_likely_update() leaves .noinstr.text section
> vmlinux.o: warning: objtool: __context_tracking_enter()+0x2c: call to ftrace_likely_update() leaves .noinstr.text section

Yes, TRACE_BRANCH_PROFILING and PROFILE_ALL_BRANCHES are fundamentally
broken and I have no intention of trying to fix them.

The moment we pull PTI into noinstr C code this will result in insta
boot fail.