It is possible to trace all kernel funcs, but only if they are called from skb related functions (including indirect)? #560
Description
I am debugging TCP send, and I only see network stuff, but I wish to also see other things, like spinlocks, memory allocations, slab allocator, preemption disable sections, route table lookups, arp table checks, and updates, snmp counter and interface counter updates, tc-qdisc internal details, etc. Most of them do not have access to any skb, so simple pcap filter will simply reject them. But it should be possible to trace them, if any of the parents in a stack trace is traced.
Thanks.