Generate vmcore when kernel panic on E5500

[wengangzhu]

Hello,

     There is something wrong on our product , which caused kernel panic.  So we want to enable the kdump to generate the vmcore when kernel panic/oops happens next time, which will help us debug.

I tried the steps using on our X86 systems, but the vmcore is not generated when I triggered the panic manually via 'echo c > /proc/sysrq-trigger'

Do vmcore supported on E5500?

If yes, would you please provide some steps/document for me?

Our system is running on ppc32,  below is the system infomation on our product:

CPU:

root@E7-2:/proc/sys/kernel# cat /proc/cpuinfo
processor : 0
cpu : e5500
clock : 1200.000000MHz
revision : 2.1 (pvr 8024 1021)
bogomips : 62.50

processor : 1
cpu : e5500
clock : 1200.000000MHz
revision : 2.1 (pvr 8024 1021)
bogomips : 62.50

processor : 2
cpu : e5500
clock : 1200.000000MHz
revision : 2.1 (pvr 8024 1021)
bogomips : 62.50

processor : 3
cpu : e5500
clock : 1200.000000MHz
revision : 2.1 (pvr 8024 1021)
bogomips : 62.50

total bogomips : 250.00
timebase : 31250000
platform : CoreNet Calix
model : fsl,T1040RDB
Memory : 4087 MB

Memory is 4G

root@E7-2:/proc/sys/kernel# free
total used free shared buff/cache available
Mem: 3853244 1233464 873172 456712 1746608 2116580
Swap: 0 0 0

Linux verison:

root@E7-2:/proc/sys/kernel# uname -a
Linux E7-2 4.14.47+gb09b730 #1 SMP Wed Sep 7 22:04:55 UTC 2022 ppc GNU/Linux

Thanks

sky

[Oswalag]

Hi,

Below is a link for the Kdump feature and how to enable it within the kernel:
http://www.kernel.org/doc/Documentation/kdump/kdump.txt

Below is a list of required packages that must be installed on the system for Kdump to work.  Please note that your kernel must be compiled properly for these packages to work as expected.
Kdump          
Kexec-tools
Kernel-debuginfo * 

Best regards, 

[wengangzhu]

Hi Oswalag,

  On http://www.kernel.org/doc/Documentation/kdump/kdump.txt, there is no instruction for ppc32.  Can you send me the configurations (kernel config, boot args and kexec commands) for e5500 (32bits)?
Currently, my kernel configurations are:

CONFIG_KEXEC_CORE=y
CONFIG_KEXEC=y

CONFIG_SYSFS=y

CONFIG_DEBUG_INFO=y

CONFIG_PROC_VMCORE=y

CONFIG_RELOCATABLE=y

CONFIG_CRASH_DUMP=y

# CONFIG_ARCH_OPTIONAL_KERNEL_RWX is not set

#  CONFIG_ARCH_HAS_STRICT_KERNEL_RWX is not set

CONFIG_NONSTATIC_KERNEL=y

/usr/sbin/kexec -d --initrd=/BOOT/imgx/initramfs.u-boot -p /FLASH/persist/vmlinux  --append="imgmnt=/media/sda2 imgdir=imgx img=image.rootfs  console=ttyS1,115200 mem=16769020k  quiet crashkernel=256M irqpoll maxcpus=1 reset_devices runlevel=1 loglevel=8"

Stopped after the crash:

root@shasta:~# echo c > /proc/sysrq-trigger
[ 240.259053] sysrq: SysRq : Trigger a crash
[ 240.263202] Oops: Kernel access of bad area, sig: 11 [#1]
[ 240.268608] BE SMP NR_CPUS=8 CoreNet Calix
[ 240.272711] Modules linked in: i2c_dev ramoops reed_solomon pramfs m25p80 spidev linux_kernel_bde(O) uio_pdrv_clx(O) ucd9000 lm90 pmbus rtc_pcf8583 ltc2978 pmbus_core ds100br210(O) clx_shr_irq(O) kap(O) proc_fifo(O) sc16is7xx at24 lm75 si5345(O) spic(O) ltc4261 hwmon spi_fsl_espi
[ 240.297579] CPU: 0 PID: 5703 Comm: bash Tainted: G O 4.14.47+gb09b730 #1
[ 240.305422] task: e23c2940 task.stack: e15ce000
[ 240.309957] **bleep**: c053d0f4 LR: c053de70 CTR: c053d0d0
[ 240.315013] REGS: e15cfde0 TRAP: 0300 Tainted: G O (4.14.47+gb09b730)
[ 240.322855] MSR: 00029002 <CE,EE,ME> CR: 24022422 XER: 00000000
[ 240.329053] DEAR: 00000000 ESR: 00800000
[ 240.329053] GPR00: c053de70 e15cfe90 e23c2940 00000063 e6f7d484 e6f83f34 26278000 e15ce000
[ 240.329053] GPR08: c10d3a20 00000001 00000000 ffffffff 24022884 1011c4b8 00000000 00000000
[ 240.329053] GPR16: 00000000 10110000 10110000 100e1134 10116230 00000000 10110000 00000001
[ 240.329053] GPR24: 100f387c 00000002 00000001 c1089728 00000063 c1040000 00000000 c1089544
[ 240.366557] **bleep** [c053d0f4] sysrq_handle_crash+0x24/0x30
[ 240.371790] LR [c053de70] __handle_sysrq+0xd0/0x210
[ 240.376672] Call Trace:
[ 240.379118] [e15cfea0] [c053de70] __handle_sysrq+0xd0/0x210
[ 240.384700] [e15cfed0] [c053e58c] write_sysrq_trigger+0x4c/0x80
[ 240.390636] [e15cfee0] [c02b6c64] proc_reg_write+0x74/0xc0
[ 240.396133] [e15cfef0] [c0240604] vfs_write+0xd4/0x1f0
[ 240.401280] [e15cff10] [c02408d4] SyS_write+0x64/0x100
[ 240.406430] [e15cff40] [c0014318] ret_from_syscall+0x0/0x3c
[ 240.412012] --- interrupt: c01 at 0xfed6b04
[ 240.412012] LR = 0xfed6acc
[ 240.419245] Instruction dump:
[ 240.422214] 4bfffe2c 00000000 00000000 7c0802a6 90010004 60000000 9421fff0 3d40c10d
[ 240.429980] 39200001 912aa6d0 7c0004ac 39400000 <992a0000> 38210010 4e800020 7c0802a6
[ 240.437924] ---[ end trace 7a9fd673b916f2fe ]---
[ 240.449638]
[ 240.451178] Sending IPI to other CPUs
[ 240.455847] IPI complete
[ 240.458583] Bye!

[wengangzhu]

Anyone can help me on this issue.