Linux Kernel Network Subsystem Patching

[Mel Wilson]

On Tue, 2014-01-21 at 16:08 -0500, Lennart Sorensen wrote:
> On Sat, Jan 11, 2014 at 09:28:38PM -0500, Mel Wilson wrote:
> > Or worse.  The system here is an MSI motherboard G41M-P26.  From the day
> > I got it, under load (large compiles, for instance), it would eventually
> > lock up.  ISTR I needed to cycle power to get it down.  Resource monitor
> > would show "both" processors pinned at 100%.  When I'd hunted the web
> > and found out what Hyperthreading was, and turned it off, the machine
> > got stable, and has been ever since.
> 
> Code that dead locks because the coder didn't know how to do
> multithreading properly is not hyperthreading's fault.  Hyperthreading
> just makes it MUCH easier to hit such a deadlock if the code is wrong
> to begin with.  Any multi processor system would increase the chances
> of hitting such bugs.

I wasn't really worried about fault, just getting a gcc build of my
bignum-embedded library to go to completion.  That couldn't happen then,
and now with hyperthreading turned of in the BIOS options, the system is
totally smooth and satisfactory.

	Mel.


--
The Toronto Linux Users Group.      Meetings: http://gtalug.org/
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://gtalug.org/wiki/Mailing_lists