system freeze without logs

[shiva_bhargav_y]

Hi,

We are using a i.mx 7 dual board with kernel version 4.1.46 with arm architecture, on it the we running Wi-Fi driver.

when we are trying to connect to Wi-Fi with hostapd num_max_sta = 1 and already 1 device connected to Wi-Fi without any problem and trying to connect 2nd device, after trying multiple times the entire system is freezing, to make the system working, we have to turn off power and power on again. before freeze happens no logs shown related to the freeze.

We don't see any logs and no kernel panic.

tried to install kexec and load secondary kernel but we are getting following error.

Could not find a free area of memory of 0x1000 bytes...
locate_hole failed

Tried to analyze the driver code didn't find anything.

Please suggest me how to get the logs or configure kexec to find the root of the freeze.

Thanks & Regards,

Shiva.

[shiva_bhargav_y]

Hi @Zhiming_Liu 

thanks for the reply,

locate_hole failed issue resolved, by giving the 128@25000 crashkernel

I m facing currently below error, please have look into it,

root@imx7d:~# kexec -d --type zImage -p zImage --append="'cat /proc/cmdline' maxcpus=1 reset_devices"
Try gzip decompression.
Try LZMA decompression.
lzma_decompress_file: read on zImage of 65536 bytes failed
kernel: 0x7676e008 kernel_size: 0x628c08
MEMORY RANGES
0000000080000000-00000000bfffffff (0)
zImage header: 0x016f2818 0x00000000 0x00628c08
zImage size 0x628c08, file size 0x628c08
zImage requires 0x00639c08 bytes
Reserved memory ranges
0000000083400000-00000000933fffff (0)
Coredump memory ranges
0000000080000000-00000000833fffff (0)
0000000093400000-00000000bfffffff (0)
kernel symbol _stext vaddr = 80008280
phys offset = 0x80000000, page offset = 80000000
Using 32-bit ELF core format
get_crash_notes_per_cpu: crash_notes addr = ab726568, size = 180
Elf header: p_type = 4, p_offset = 0xab726568 p_paddr = 0xab726568 p_vaddr = 0x0 p_filesz = 0xb4 p_memsz = 0xb4
get_crash_notes_per_cpu: crash_notes addr = ab732568, size = 180
Elf header: p_type = 4, p_offset = 0xab732568 p_paddr = 0xab732568 p_vaddr = 0x0 p_filesz = 0xb4 p_memsz = 0xb4
vmcoreinfo header: p_type = 4, p_offset = 0x80bd761c p_paddr = 0x80bd761c p_vaddr = 0x0 p_filesz = 0x1024 p_memsz = 0x1024
Elf header: p_type = 1, p_offset = 0x80000000 p_paddr = 0x80000000 p_vaddr = 0x80000000 p_filesz = 0x3400000 p_memsz = 0x3400000
Elf header: p_type = 1, p_offset = 0x93400000 p_paddr = 0x93400000 p_vaddr = 0x93400000 p_filesz = 0x2cc00000 p_memsz = 0x2cc00000
elfcorehdr 0x93300000-0x933003ff
crashkernel: [0x83400000 - 0x933fffff] (256M)
memory range: [0x80000000 - 0x833fffff] (52M)
memory range: [0x93400000 - 0xbfffffff] (716M)
kernel command line: "'cat /proc/cmdline' maxcpus=1 reset_devices elfcorehdr=0x93300000 mem=261120K"
Kernel: address=0x83408000 size=0x01f20c28
get_cells_size: #address-cells:1 #size-cells:1
cells_size_fitted: 93300000-933003ff
cells_size_fitted: 83400000-933fffff
DT : address=0x8532a000 size=0x0000b053
kexec_load: entry = 0x83408000 flags = 0x280001
nr_segments = 3
segment[0].buf = 0x7676e008
segment[0].bufsz = 0x628c08
segment[0].mem = 0x83408000
segment[0].memsz = 0x629000
segment[1].buf = 0x1557068
segment[1].bufsz = 0xb053
segment[1].mem = 0x8532a000
segment[1].memsz = 0xc000
segment[2].buf = 0x1556c60
segment[2].bufsz = 0x400
segment[2].mem = 0x93300000
segment[2].memsz = 0x1000

[Zhiming_Liu]

Hi @shiva_bhargav_y 

Please enlarge size of kdump reserved memory

Best Regards

Zhiming