Prabhakar K.
2009-09-24 21:04:40 UTC
Hi All -I have a router box on which I am trying to develop kexec based crash dump?collection as its very useful for us to get the dump from customer sites?where there is no console.
This is linux 2.6.14 based.
I am trying to do kexec into the same kernel(no separate dump capturing kernel).
I have kexec-tools and the kdump patch (kernel compiled with modified config).
Before I put my question, here is the set up details:
-- We have SMP support with 2 CPUs.-- It is i386 based, so I believe it is?
relocatable, and I dont need to modify the CONFIG_PHYSICAL, it is pointing?
to 0x100000, so I leave it as it is.-- I have enabled all the required in?config-- I am using the same kernel for collecting dump information as well
(so I have only one kernel image).-- I have added crashkernel=64M at 16M to?
my grub.conf and I am able to allocate 64MB space for the crash kernel.?cat?/proc/iomem shows this.
-- Before I do kexec -p for panic, I am trying to test with kexec -l and kexec -e,?if I can boot with the crash kernel
I do the following command from my root directory(/) where the?system.bzImage exists.
kexec -l system.bzImage --append= "1 irqpoll maxcpus=1 reset_devices root=/dev/hda2"
after that kexec -e
I get the message:
Starting new kerneland it hangs after that. I have to do hard reboot of the?router box.
So looking at some of the mailing lists, I have added --console-serial --debug,then I get :
I'm in purgatoryabcdefgh
Embedded BIOS Debugger Invalid Opcode Trap
Though I am skeptical about this debug, IMO, I must be missing something for?creating crash kernel.?May be like is allocating 64MB sufficient ? or Can?I have the same bzImage?which can boot normally and also with kexec fast boot ?
Your suggestions will be of great help.
Please let me know if any additional information is needed.
Please cc me as I'm not subscribed to the list.
Regards,Prabhakar
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.linux-foundation.org/pipermail/fastboot/attachments/20090924/3f7f5947/attachment.htm
This is linux 2.6.14 based.
I am trying to do kexec into the same kernel(no separate dump capturing kernel).
I have kexec-tools and the kdump patch (kernel compiled with modified config).
Before I put my question, here is the set up details:
-- We have SMP support with 2 CPUs.-- It is i386 based, so I believe it is?
relocatable, and I dont need to modify the CONFIG_PHYSICAL, it is pointing?
to 0x100000, so I leave it as it is.-- I have enabled all the required in?config-- I am using the same kernel for collecting dump information as well
(so I have only one kernel image).-- I have added crashkernel=64M at 16M to?
my grub.conf and I am able to allocate 64MB space for the crash kernel.?cat?/proc/iomem shows this.
-- Before I do kexec -p for panic, I am trying to test with kexec -l and kexec -e,?if I can boot with the crash kernel
I do the following command from my root directory(/) where the?system.bzImage exists.
kexec -l system.bzImage --append= "1 irqpoll maxcpus=1 reset_devices root=/dev/hda2"
after that kexec -e
I get the message:
Starting new kerneland it hangs after that. I have to do hard reboot of the?router box.
So looking at some of the mailing lists, I have added --console-serial --debug,then I get :
I'm in purgatoryabcdefgh
Embedded BIOS Debugger Invalid Opcode Trap
Though I am skeptical about this debug, IMO, I must be missing something for?creating crash kernel.?May be like is allocating 64MB sufficient ? or Can?I have the same bzImage?which can boot normally and also with kexec fast boot ?
Your suggestions will be of great help.
Please let me know if any additional information is needed.
Please cc me as I'm not subscribed to the list.
Regards,Prabhakar
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.linux-foundation.org/pipermail/fastboot/attachments/20090924/3f7f5947/attachment.htm