Bug 243442

Summary: kdump may not work on x86-64, might need to push the 5.1 kdump kernel in the RT yum repo
Product: Red Hat Enterprise MRG Reporter: Tim Burke <tburke>
Component: realtime-kernelAssignee: Guy Streeter <streeter>
Status: CLOSED NOTABUG QA Contact:
Severity: low Docs Contact:
Priority: low    
Version: 1.0CC: anderson, nobody
Target Milestone: ---   
Target Release: ---   
Hardware: All   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2007-06-08 19:34:30 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Bug Depends On: 242648    
Bug Blocks:    

Description Tim Burke 2007-06-08 18:19:49 UTC 
Bugzilla Bug #242648: kdump broken on x86_64

describes a fix in 5.1 that is needed to get the kdump kernel working on some
systems. 

in the interests of keeping the set of rhel-rt packages as lean as possible, it
may be necessary to provide the RHEL5.1 (beta?) kdump kernel through the yum
repo for RHEL-RT.

the howto needs to be updated with instructions
Comment 1 Dave Anderson 2007-06-08 18:51:31 UTC 
Note that that patch should not be necessary, and in fact won't apply to 
kernel-rt-2.6.21-25.el5rt.  The upstream kernel, and therefore
the -rt kernel, is less restrictive when checking the "crashkernel="
command line argument, just making sure that they're not the same
value:

  #ifdef CONFIG_KEXEC
        if (crashk_res.start != crashk_res.end) {
                reserve_bootmem_generic(crashk_res.start,
                        crashk_res.end - crashk_res.start + 1);
        }
#endif

whereas the broken, but more restrictive, RHEL5.1 code does this:

        if ((crashk_res.start < crashk_res.end) &&
            (crashk_res.end <= (max_low_pfn << PAGE_SHIFT))) {
                reserve_bootmem_generic(crashk_res.start,
                                        crashk_res.end - crashk_res.start + 1);
        }
 
where the RHEL5.1 patch replaces "max_low_pfn" with "end_pfn" for x86_64.

So if you just leave things as they are, the -rt kernel should load the
secondary (RHEL5.1) kernel OK...   ;-)
Comment 2 Guy Streeter 2007-06-08 19:34:07 UTC 
This patch affects the boot kernel (it involves the parsing of the crashkernel=
option), not the kdump kernel. The RHEL5 kernel can be used as a kdump kernel
even with this bug.
The patch which introduced the bug has not been carried into the RT kernel, so
it does not affect the crashkernel= parsing in the RT kernel.