From mboxrd@z Thu Jan 1 00:00:00 1970 From: Andrew Cooper Subject: [PATCH 0 of 4 RFC] Kexec path alteration Date: Tue, 4 Dec 2012 18:16:07 +0000 Message-ID: Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xen.org Errors-To: xen-devel-bounces@lists.xen.org To: xen-devel@lists.xen.org Cc: Andrew Cooper List-Id: xen-devel@lists.xenproject.org I discovered a triple fault bug in the crash kernel while working on the wider reentrant NMI/MCE issues discussed on the past few weeks on the list. As this is quite a substantial change, I present it here as an RFC ahead of the longer series for reentrant behaviour (which I am still working on). I present patches 2 thru 4 simply as debugging aids for the bug fixed in patch 1. I am primarily rally concerned about RFC feedback from patch 1, although comments on patches 2 and 3 will not go amis, as they are expected to be in the longer reentrant series. Patch 4 is a debugkey '1' which deliberately creates reentrant NMIs and stack corruption. It is not intended for actually committing into the codebase. I am at a loss to explain why the triple fault is actually occurring; I failed to get anything from early printk from the crash kernel, but did not attempt to debug the extra gubbins which kexec puts in the blob. The fault however is completely deterministic and can be verified by commenting out the call to enable_nmis() in machine_kexec() My setup is xen-unstable, 2.6.32-classic based dom0 and kdump kernels with kexec-tools 2.0.2 ~Andrew