linux-kernel - Re: [tip:WIP.x86/apic 40/43] arch/x86/kernel/irq.o:(__tracepoints+0x170): undefined reference to `trace_resched_ipi

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Fri, 25 Aug 2017 14:22:37 +0200 (CEST)
From:   Thomas Gleixner <tglx@...utronix.de>
To:     kbuild test robot <fengguang.wu@...el.com>
cc:     kbuild-all@...org, linux-kernel@...r.kernel.org, tipbuild@...or.com
Subject: Re: [tip:WIP.x86/apic 40/43] arch/x86/kernel/irq.o:(__tracepoints+0x170):
 undefined reference to `trace_resched_ipi_reg'

On Fri, 25 Aug 2017, kbuild test robot wrote:

> Hi Thomas,
> 
> It's probably a bug fix that unveils the link errors.
> 
> tree:   https://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git WIP.x86/apic
> head:   395b2c62edea5a2bb8a8786392feb12fae5917a7
> commit: 902f1bb88397dd6769c98b2177bbda134b3a1ef2 [40/43] x86/idt: Deinline setup functions
> config: i386-randconfig-s0-201734 (attached as .config)
> compiler: gcc-6 (Debian 6.2.0-3) 6.2.0 20160901
> reproduce:
>         git checkout 902f1bb88397dd6769c98b2177bbda134b3a1ef2
>         # save the attached .config to linux build tree
>         make ARCH=i386 
> 
> All errors (new ones prefixed by >>):
> 
> >> arch/x86/kernel/irq.o:(__tracepoints+0x170): undefined reference to `trace_resched_ipi_reg'
> >> arch/x86/kernel/irq.o:(__tracepoints+0x174): undefined reference to `trace_resched_ipi_unreg'
>    arch/x86/kernel/irq.o:(__tracepoints+0x184): undefined reference to `trace_resched_ipi_reg'
>    arch/x86/kernel/irq.o:(__tracepoints+0x188): undefined reference to `trace_resched_ipi_unreg'
> 

That link error is caused by the tracepoint code being stupid enough to
generate tracepoint stuff in an object file whether they are used or not.

The issue here is that the reschedule ipi tracepoint has an initializer
function which is only compiled in when CONFIG_SMP=y.

That tracepoint is not used in irq.o at all, but still the tracepoint macro
hell generates something for it:

objdump -t ../build-386/arch/x86/kernel/irq.o | grep resched
00000048 l     O _ftrace_events	00000004 __event_reschedule_exit
00000a00 l     O .data	0000004c event_reschedule_exit
0000004c l     O _ftrace_events	00000004 __event_reschedule_entry
00000a60 l     O .data	0000004c event_reschedule_entry
00000048 l     O __tracepoints_ptrs	00000004 __tracepoint_ptr_reschedule_exit
00000188 l     O __tracepoints_strings	00000010 __tpstrtab_reschedule_exit
0000004c l     O __tracepoints_ptrs	00000004 __tracepoint_ptr_reschedule_entry
00000198 l     O __tracepoints_strings	00000011 __tpstrtab_reschedule_entry
0000017c g     O __tracepoints	00000014 __tracepoint_reschedule_entry
00000000         *UND*	00000000 trace_resched_ipi_reg
00000168 g     O __tracepoints	00000014 __tracepoint_reschedule_exit
00000000         *UND*	00000000 trace_resched_ipi_unreg

Brilliant stuff that. I'll hide the tracepoint behind a CONFIG_SMP ifdef
for now.

Thanks,

	tglx