Please always cc the bug address when replying to bug-related mails. On Sat, 2011-07-16 at 09:54 -0400, Frank McCormick wrote: > On 11/07/11 12:12 AM, Ben Hutchings wrote: > > On Tue, 2011-07-05 at 17:29 -0400, Frank wrote: > >> On 05/07/11 04:23 PM, Ben Hutchings wrote: > >>> On Tue, Jul 05, 2011 at 02:48:26PM -0400, Frank M wrote: > >>>> I find it strange that the PAE kernel is the ONLY one which gives me > >>>> any trouble. > >>>> I guess it could be hardware related, but 2.6.38-1 and 2.6.38-2 have both been > >>>> booting fine for months. > >>>> Is there something about the PAE series that would uncover hardware faults which > >>>> may have existed for a long time ? > >>> > >>> Have you tried booting those earlier versions recently? > > >> > >> Yes I've been booting them everyday for the past month or so - the > >> PAE series is the first time in years, literally, that I have had > >> problems like this. > > > > We have provided kernel packages using PAE for a long time, previously > > labelled as '686-bigmem'. You can test linux-image-2.6.38-2-686-bigmem > > to check whether this bug is related to use of PAE. However, I think it > > is probably due to a change between Linux 2.6.38 and 2.6.39. > > Well I **think** the bigmem kernel now is actually the PAE kernel. I > installed the bigmem kernel and it turned out to be the troublesome PAE > version. Right. It has been renamed because it is now the only '686' flavour, but has different hardware requirements from the old '686' flavour. I believe the processor in your computer should support this, but could you confirm what model you have (look at the file /proc/cpuinfo)? > I still don't understand why the changes created this problem for me. > The strange thing is on multiple tries, on a regular boot and on > maintenance boots, the computer will sometimes get to different points > before it gives up. The boots usually end up with the num lock and caps > lock lights flashing. This sounds like there is some sort of hardware fault, but I can't see why it would only occur when using PAE. It is just possible that the circuitry for PAE is faulty, but I think that is only a very small part of the chip. Please do consider the suggestions in <http://www.bitwizard.nl/sig11/>. Ben. > I guess what I'll do is remove the PAE kernel again and continue with > the previous series. Long term I don't know what I'll do - I do know > that Ubuntu kernels have not yet been a problem. So if the working > kernel in my machine gets too far behind I have a choice. -- Ben Hutchings Absolutum obsoletum. (If it works, it's out of date.) - Stafford Beer
Attachment:
signature.asc
Description: This is a digitally signed message part