I have 8 logical CPUs in my linux box. (As seen from /proc/cpuinfo). When I look in
/sys/kernel/debug/tracing/per_cpu
it shows 32 CPUs.
Why the difference? If it matters, I have hyperthreading disabled in the BIOS.
I have just built a custom kernel from source (obtained from kernel.org). I can boot from it fine on my local box. I now want to make it easy for a colleague to try this kernel out.
Is there an easy way to create a rpm that he can install? The instructions I found on googling assume that I configure and build my kernel from a kernel.src.rpm but this is not what I did.
Is it enough to replicate /lib/modules/$(uname -r) , /boot/vmlinuz , /boot/System.map-$(uname -r), /boot/initramfs-$(uname -r) to the new machine to be able to boot over there ?
Thanks.
Suppose a host computer has joined a multicast group 239.0.0.50 via IGMP 2.0. As per the IGMP 2.0 spec, when the host computer receives either a general query on 224.0.0.1 or a group wide query on 239.0.0.50, it is supposed to set a random delay timer and wait till the timeout before it sends out a membership report. If during this timeout period, it receives a membership report from some other host claiming to be a member of the group, then, there is no further need of sending out the membership report itself and it can just cancel the timer.
Is the above still the correct IGMP behavior for a host today? I suspect that the switch is timing me out (this happens in a certain prod environment and I will be able to confirm the suspicion only next week) and I heard a suggestion that I should always be sending out a membership report.
(Just for clarification, the host is not a standard/linux box, but a custom non-linux box)
Thanks
I have a PCIe (ver 2) linux system. I want to find out how many PCIe lanes were negotiated and are being used between my device and the host controller . (I do not know how many lanes my device supports or how many lanes the slot has)
I have a special lab machine (linux) and two different users will be using this machine. They need to log in with the same userid. I want to prevent both of them from logging in at the same time. How can I do that?
I was thinking of starting with something simple like this - Use the users source IP-address and add some checks in .bashrc :
If a user logs in from IPAddress1 and there is already somebody ssh'ed in from IPAddress2 then print a big fat message.