=Comment: #0================================================= JOSHUA A. TRIPLETT <josht.com> - 2008-07-22 16:07 EDT To make kdump work, the main kernel must have memory reserved for the kdump kernel. It reserves this memory during the early boot process, based on the "crashkernel" option in the kernel command line. This option generally looks similar to "crashkernel=128M@32M". The first number specifies the amount of memory to reserve, and the second number specifies the offset in physical memory to reserve this at. However, some values for the offset will cause kdump not to work on some systems; for instance, some systems fail with "crashkernel=128M@16M", but work with "crashkernel=128M@32M". This makes it difficult to provide a standard kdump configuration for all platforms, and requires additional documentation to address. Git commit 32105f7fd8faa7bc3d101dcc3eabc0ae1ac375a7 ("x86: find offset for crashkernel reservation automatically") in mainline linux-2.6 fixes this problem. It allows a crashkernel parameter like "crashkernel=128M", without an offset, and automatically detects a usable offset. (This requires a relocatable kdump kernel, which MRG currently uses.) This patch appears non-invasive. Including it would simplify kdump configuration and documentation. Please consider including it in future versions of the MRG kernel, and updating rt-setup-kdump to remove the offset from the crashkernel parameter. =Comment: #1================================================= JOSHUA A. TRIPLETT <josht.com> - 2008-07-24 12:47 EDT http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff_plain;h=32105f7fd8faa7bc3d101dcc3eabc0ae1ac375a7
------- Comment From sripathik.com 2009-06-29 11:02 EDT------- Saw this patch in 2.6.29 based MRG kernels. So MRG got this patch automatically. We can close this bug. ------- Comment From sripathik.com 2009-06-29 11:02 EDT------- To RH: We are closing this bug on IBM side.
closing as well