Vivek Goyal <vgoyal at redhat.com> writes: > On Fri, May 27, 2011 at 11:16:34AM -0700, Eric W. Biederman wrote: >> "K.Prasad" <prasad at linux.vnet.ibm.com> writes: >> >> > Crash: Recognise slim coredumps and process new elf-note sections >> > >> > The Linux kernel will begin to support SlimDump for certain types of crashes >> > and the 'crash' tool needs to recognise them. For these types of coredumps, it >> > need not lookout for usual elf-structures and start gdb. Also process new >> > elf-note sections that contain additional information about the crash. >> >> I suppose patching crash make sense. >> >> Unfortunately crash doesn't work on 99% of the kernels I run so, so I >> stopped caring a while ago. > > Are these patched kernels? Otherwise Dave Anderson is pretty good about > keeping crash running across multiple kernels. Only slightly patched. But they are definitely kernels I have built myself. It might just be a logistics problem where if I want to use crash I need to upgrade it when I upgrade my kernels. I got frustrated a while ago trying to get something out of crash (and I'm afraid it shows), and I have just gone to extracting the dmesg. Which gives me all of the information I generally have time to process. Eric
