Alan Cox <alan@xxxxxxxxxxxxxxxxxxx> wrote:Paul's fix is actually in 2.6.15-git9. This version seems to work (downloaded 20x more data than usually required to trigger the bug).
On Iau, 2006-01-12 at 22:43 -0800, Andrew Morton wrote:
That sounds like the bug that Paul fixed.infinitely replicated several last packets. events/0 consumes all the CPU. tty buffering revamping patch is the obvious candidate, but I haven't tried to revert it yet.Darn, I hadn't thought of that. Yes tty-revamp might be the culprit.
That means "please test 2.6.15-git10" ;)