pstore looks very promising. I tried pstore with backup as efivars and it works. Although efivars works, I would like to use the pstore with ramoops and I am currently having difficulty in choosing the correct ramoops.mem_address. I enabled the related configs and using command line parameters as shown in the documentation https://www.kernel.org/doc/html/v4.14/admin-guide/ramoops.html Ex usage is: `mem=128M ramoops.mem_address=0x8000000 ramoops.ecc=1` 0x8000000 is the start of the kernel code section according to `/proc/iomem`. So, I need to find a different location. In /proc/iomem I can see a few sections of ram marked as "System RAM" and "Reserved" I tried to use these addresses but using them causes weird behavior and the system is freezing while booting up. *So, how do I select a 128 MB space on my ram and dedicate it for ramoops?* On Tue, Jun 1, 2021 at 5:40 PM Fox Chen <foxhlchen@gmail.com> wrote:
On Tue, Jun 1, 2021 at 2:07 PM manty kuma <mantykuma@gmail.com> wrote:
I am using debian , the logging system being used is rsyslogd. in /etc/rsyslog.d/default.conf we have rule as follows: ``` kern.* /var/log/kern.log ``` This rule I believe is the reason why all the kernel logs are being
redirected to /var/log/kern.log.
Also dmesg does not show anything. So, i am pretty sure that all the
kernel logs are being handled solely by `rsyslogd`
And this is just the default debian distribution without any customizations.
Thank you for the 'pstore' clue. I will explore it further.
On Tue, Jun 1, 2021 at 2:42 PM Greg KH <greg@kroah.com> wrote:
On Tue, Jun 01, 2021 at 02:33:50PM +0900, manty kuma wrote:
I just triggered a panic, expecting that the logs will be visible in `/var/log/kern.log` after reboot, but there are no logs present there.
I have never heard of kernel logs being written to that location, what tool do you have that does that and where is that documented?
Considering I have no access to the serial port, how do I know what
went
wrong?
When the kernel panics, it usually can not write to the disk, so it's a bit hard to save anything :)
That being said, there are ways the kernel can save the crash information, look into the "pstore" interface and see if that will work for your hardware platform (it requires hardware to store the information across boots.)
Also, check kdump, I think it can help as well.
good luck!
greg k-h
_______________________________________________ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org https://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies