Messages in this thread | | | From | "Takashi Nishiie" <> | Subject | RE: Kernel marker has no performance impact on ia64. | Date | Tue, 3 Jun 2008 15:07:48 +0900 |
| |
Hello.
> One of my goal is to provide a mechanism that can feed both non-debug > and debug information to a generic tracing mechanism to allow > system-wide analysis of the kernel, both for production system and > kernel debugging.
I agree with the target. I understand there is a person who is the neat paranoiac that it doesn't want to leave the code for debugging for a final code. If it is a situation in which the code for debugging without united interfaces lies scattered, the opinion might be correct. It is an effective method only to set up the marker if the code for debugging without united interfaces can be appropriately set up with the marker. I think that the mechanism that can offer both non-debugging and debugging information to the generic pursuit mechanism to permit the analysis of the entire kernel system in the meaning of obtaining dynamic information for a long time not obtained by a static necropsy by kdump is indispensable in a mission-critical system.
Regards.
| |