<!doctype html public "-//w3c//dtd html 4.0 transitional//en"> <html> Olaf Hering wrote: <blockquote TYPE=CITE>On Wed, Nov 08, Dave Anderson wrote: > Olaf Hering wrote: > > > crash 4.0-3.9 can not read a dump file from an ia64 box running 2.6.5. > > Is this supposed to work? > > > > crash -s boot/System.map-2.6.5-7.267-sn2 boot/Kerntypes-2.6.5-7.267-sn2 dump.3 > > Well, if it did work, it would be news to me. I've certainly never run crash > with a "Kerntypes" file as a substitute for a -g built vmlinux file. Thats all we have right now. And sadly, I have been told that gcc will generate different asm depending on what host the kernel is compiled (memory config etc.). So getting the exact vmlinux again may be a challenge with gcc3.2.</blockquote> <tt>Well, that's always been the case, and that's OK. In that case,</tt> <tt>all you have to do is compile the identical kernel source code</tt> <tt>with -g, but also pass the System.map file of the "real" dumped</tt> <tt>vmlinux on the command line, as in:</tt><tt></tt> <tt> # crash vmlinux-rebuilt-with-g System.map-of-dumped-vmlinux dumpfile</tt><tt></tt> <tt>That's the sole purpose of the System.map argument, which should</tt> <tt>*only* be used in such a case.</tt><tt></tt> <tt>What happens when you use a System.map command line argument is this:</tt><tt></tt> <tt>1. crash passes its embedded gdb the vmlinux-rebuilt-with-g file,</tt> <tt> which will most likely have many of its text addresses, and probably</tt> <tt> all of its data addresses "shifted" up, and therefore be incorrect.</tt> <tt> (But the data structure definitions, etc., are correct...)</tt><tt></tt> <tt>2. After that, crash takes the System.map file, which contains a</tt> <tt> list of the *correct* symbol values, goes in a back-door into gdb,</tt> <tt> and overwrites gdb's symbol list values with the correct values.</tt> <tt> That's what's happening when you see that init-time message:</tt><tt></tt> <tt> please wait... (patching 20920 gdb minimal_symbol values)</tt><tt></tt> <tt>We had to do that for RHEL 2.1, which at the time did not have a</tt> <tt>kernel debuginfo package, so we always had to rebuild specific</tt> <tt>kernel versions with -g for crash sessions. With RHEL3 and beyond,</tt> <tt>the -g built vmlinux is part of each kernel's debuginfo package,</tt> <tt>so from that point on, there has never been a need to use a</tt> <tt>System.map file.</tt><tt></tt> <tt>Dave</tt> </html>