since the functionality, speed, race conditions, etc of a CONFIG_KALLSYMS_LINE_LOCATIONS enabled kernel and a normal one are the same, it seems to me that the distros should have both kernels available, boot from the normal one (avoiding any overhead in the normal case, especially for virtual machines), but if someone is running into repeated problems, they can boot from the enabled version and then get the detailed info.
I wouldn't expect anyone to take the time to report a panic unless it's something that is happening repeatedly (and probably something they can trigger fairly reliably), so this approach should cover these situations nicely.