Re: Linux RAS

2000-09-15 Thread Matt D. Robinson
I'd also want the default kernel build to create a symbol table namelist object that gets installed into $(INSTALL_PATH) that correlates to the kernel build. That way you build a symbol table mechanism for user-space applications that want more complete kernel debug information, but do it without

Re: Linux RAS

2000-09-15 Thread richardj_moore
To: Richard J Moore/UK/IBM@IBMGB cc: [EMAIL PROTECTED] Subject: Linux RAS On Fri, 15 Sep 2000 10:42:54 +0100, [EMAIL PROTECTED] wrote: >I think the case for the kernel debugger is better stated as the case for >RAS (Reliability, Serviceability and Availability) in the kernel [snip] &

Re: Linux RAS

2000-09-15 Thread Daniel Phillips
Keith Owens wrote: > * Standardize on tracking the System.map and .config with the kernel. There was a suggestion from Alan Cox that .config.gz be appended to bzImage, after the part that gets loaded into memory, to which I added the suggestion that System.map.gz also be appended. That about tak

Linux RAS

2000-09-15 Thread Keith Owens
On Fri, 15 Sep 2000 10:42:54 +0100, [EMAIL PROTECTED] wrote: >I think the case for the kernel debugger is better stated as the case for >RAS (Reliability, Serviceability and Availability) in the kernel [snip] >Customers knew we never sent >developers on site to debug OS/390 (or MVS as it was cal