Messages in this thread | | | Date | Fri, 12 Jun 2009 13:55:18 -0400 | From | Theodore Tso <> | Subject | Re: [PATCH 1/5] HWPOISON: define VM_FAULT_HWPOISON to 0 when feature is disabled |
| |
On Fri, Jun 12, 2009 at 05:36:20PM +0200, Ingo Molnar wrote: > > The data corruption has not caused real hurt yet, and can be > > isolated to prevent future accesses. So it makes sense to just > > kill the impacted process(es). > > Dunno, this just looks like a license to allow more crappy hardware, > hm? I'm all for _logging_ errors, but hwpoison is not about that: it > is about allowing the hardware to limp along in 'enterprise' setups, > with a (false looking) 'guarantee' that everything is fine.
This should be tunable; in some cases, logging it is the right thing to do; I imagine that in the case of the desktop OS, the user would appreciate being given *some* chance to save the document he or she has spent the past hour working on before the system goes down "hard and fast".
In other cases, the sysadmin is using a high-availability setup in an enterprise deployment, and there he or she would want the system to immediately shutdown so the hot standby can take over.
- Ted
| |