CODEWITHSUNDEEP
linuxlinux-kernellinux-device-driverembedded-linux
amarkov
amarkov

Reputation: 19

Linux kernel crash call stack length

Would anyone be able to tell how I can easily increase the number of call stack functions reported on a Linux Kernel crash?

Currently I see: 

[<80100ca8>] (free_buffer_head) from [<80100d2c>] (try_to_free_buffers+0x7c/0xbc)
[<80100d2c>] (try_to_free_buffers) from [<800a9358>] (invalidate_inode_page+0x64/0x7c)
[<800a9358>] (invalidate_inode_page) from [<800a9454>] (invalidate_mapping_pages+0xe4/0x168)
[<800a9454>] (invalidate_mapping_pages) from [<80280f68>] (blkdev_ioctl+0x40c/0x910)
[<80280f68>] (blkdev_ioctl) from [<800e78a0>] (do_vfs_ioctl+0x3dc/0x59c)
[<800e78a0>] (do_vfs_ioctl) from [<800e7a94>] (SyS_ioctl+0x34/0x5c)
[<800e7a94>] (SyS_ioctl) from [<8000e460>] (ret_fast_syscall+0x0/0x30)

But I would really like to see even more output to get to the root of the error.

Many thanks

Upvotes: 1

Views: 233

Answers (1)

Zbynek Vyskovsky - kvr000
Zbynek Vyskovsky - kvr000

Reputation: 18865

I think you already see everything. SyS_ioctl is entry function from userspace call and free_buffer_head is function where it actually crashed. It will not provide the backtrace from userspace program as, first it can be application specific, secondly it's not relevant to kernel crash.

Upvotes: 2

Related Questions