Re: [tip:x86/percpu] [x86/percpu] ca42563486: BUG:kernel_failed_in_early-boot_stage,last_printk:Booting_the_kernel(entry_offset:#)

[Philip Li <philip.li@intel.com>]

On Sat, Oct 07, 2023 at 11:36:26AM +0200, Uros Bizjak wrote:
> On Sat, Oct 7, 2023 at 8:16 AM kernel test robot <oliver.sang@intel.com> wrote:
> >
> >
> >
> > Hello,
> >
> > kernel test robot noticed "BUG:kernel_failed_in_early-boot_stage,last_printk:Booting_the_kernel(entry_offset:#)" on:
> >
> > commit: ca4256348660cb2162668ec3d13d1f921d05374a ("x86/percpu: Use C for percpu read/write accessors")
> > https://git.kernel.org/cgit/linux/kernel/git/tip/tip.git x86/percpu
> >
> > [test failed on linux-next/master 0f0fe5040de5e5fd9b040672e37725b046e312f0]
> >
> > in testcase: boot
> >
> > compiler: gcc-12
> > test machine: qemu-system-x86_64 -enable-kvm -cpu SandyBridge -smp 2 -m 16G
> >
> > (please refer to attached dmesg/kmsg for entire log/backtrace)
> >
> >
> > +--------------------------------------------------------------------------------------+------------+------------+
> > |                                                                                      | 9a462b9eaf | ca42563486 |
> > +--------------------------------------------------------------------------------------+------------+------------+
> > | boot_successes                                                                       | 13         | 0          |
> > | boot_failures                                                                        | 0          | 13         |
> > | BUG:kernel_failed_in_early-boot_stage,last_printk:Booting_the_kernel(entry_offset:#) | 0          | 13         |
> > +--------------------------------------------------------------------------------------+------------+------------+
> 
> Since this is a randconfig (x86_64-randconfig-006-20231006), does it
> mean that all other configs worked OK:

Thanks Uros. The previous report you receive are for build/compling test,
which means for these tested kconfigs, the kernel are built successfully.

For this report, we have found a possible boot issue on the reported kconfig
as compared to the parent commit. You can kindly do a check.

Thanks

> 
> i386                             allmodconfig   gcc
> i386                              allnoconfig   gcc
> i386                             allyesconfig   gcc
> i386         buildonly-randconfig-001-20231005   gcc
> i386         buildonly-randconfig-002-20231005   gcc
> i386         buildonly-randconfig-003-20231005   gcc
> i386         buildonly-randconfig-004-20231005   gcc
> i386         buildonly-randconfig-005-20231005   gcc
> i386         buildonly-randconfig-006-20231005   gcc
> i386                              debian-10.3   gcc
> i386                                defconfig   gcc
> i386                  randconfig-001-20231005   gcc
> i386                  randconfig-002-20231005   gcc
> i386                  randconfig-003-20231005   gcc
> i386                  randconfig-004-20231005   gcc
> i386                  randconfig-005-20231005   gcc
> i386                  randconfig-006-20231005   gcc
> 
> and
> 
> x86_64                            allnoconfig   gcc
> x86_64                           allyesconfig   gcc
> x86_64                              defconfig   gcc
> x86_64                randconfig-001-20231005   gcc
> x86_64                randconfig-002-20231005   gcc
> x86_64                randconfig-003-20231005   gcc
> x86_64                randconfig-004-20231005   gcc
> x86_64                randconfig-005-20231005   gcc
> x86_64                randconfig-006-20231005   gcc
> x86_64                          rhel-8.3-rust   clang
> x86_64                               rhel-8.3   gcc
> 
> are of interest to me. Assuming they are built with gcc-12, I wouldn't
> immediately blame the compiler for the failure. Due to the nature of
> the change, perhaps a weakness in the kernel has been found with some
> obscure config setting. As said, my default Fedora 39 kernel (6.5.5),
> built with gcc-13 works without any problems.
> 
> Also, does a successful report from yesterday [1] mean everything was OK?
> 
> [1] https://lore.kernel.org/lkml/202310060322.yeZgaj6Q-lkp@intel.com/
> 
> Uros.
>