Re: [x86/setup] f5f3497cad: BUG: kernel boot crashed

[Ingo Molnar <mingo@kernel.org>]

[-- Attachment #1: Type: text/plain, Size: 2127 bytes --]


* Matt Fleming <matt@codeblueprint.co.uk> wrote:

> On Wed, 28 Oct, at 01:41:51PM, kernel test robot wrote:
> > FYI, we noticed the below changes on
> > 
> > https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master
> > commit f5f3497cad8c8416a74b9aaceb127908755d020a ("x86/setup: Extend low identity map to cover whole kernel range")
> > 
> > 
> > +------------------------------------------------+------------+------------+
> > |                                                | 8a53554e12 | f5f3497cad |
> > +------------------------------------------------+------------+------------+
> > | boot_successes                                 | 19         | 10         |
> > | boot_failures                                  | 2          | 9          |
> > | IP-Config:Auto-configuration_of_network_failed | 2          | 2          |
> > | BUG:kernel_boot_crashed                        | 0          | 7          |
> > +------------------------------------------------+------------+------------+
> > 
> > 
> > [    0.053410] smpboot: CPU0: GenuineIntel GenuineIntel QEMU Virtual CPU version 2.4.0QEMU Virtual CPU version 2.4.0 (family: 0x6, model: 0x6 (family: 0x6, model: 0x6, stepping: 0x3)
> > , stepping: 0x3)
> > [    0.056666] Performance Events: 
> > [    0.056666] Performance Events: no PMU driver, software events only.
> > no PMU driver, software events only.
> > [    0.060520] CPU 1 irqstacks, hard=8a832000 soft=8a834000
> > [    0.060520] CPU 1 irqstacks, hard=8a832000 soft=8a834000
> > [    0.061605] x86: Booting SMP configuration:
> > [    0.061605] x86: Booting SMP configuration:
> > [    0.062445] .... node  #0, CPUs:  
> > [    0.062445] .... node  #0, CPUs:         #1 #1
> > Elapsed time: 10
> > BUG: kernel boot crashed
> 
> Yikes! Yeah, I can reproduce this. It appears the STARTUP IPI code is
> triggering this crash in wakeup_secondary_cpu_via_init().
> 
> I'll take a look unless anyone else has any bright ideas?

Unfortunately this commit is now part of the v4.3 kernel, so we need a quick 
solution or a quick revert.

Thanks,

	Ingo