On Thu, 29 May 2008, Andy Furniss wrote:

? wrote:

quoted

On Sat, 3 May 2008, Andy Furniss wrote:

quoted

Ilpo J?rvinen wrote:

quoted

Since I've not yet found any other path that could lead to it except the
yesterday's false-positive, getting the write queue state more
accurately
captured by the debug patch on the very first occassion might help.
Here's
less spammy version of the debug patch (I added one printout as well to
get
more usefulness to transitional state reported).

OK, increased CONFIG_LOG_BUF_SHIFT from 14 to 18 and reversed old /
applied
new patch.

Any luck so far?

Finally - looks like I got two together.

It doesn't start with WARNING - but dmesg isn't full.

This looks ok trace this time :-), I changed the ordering of the lines so 
that warning was supposed to appear only after the wq print. Now it's just 
the hard par remaining, ie., I must figure out what was the state of TCP 
before the event the event... :-).

If you still get additional events (you could if a particular torrent 
peer is there having the same network behavior), please post them as well 
if they reveal something that helps exclude cases (if I don't right away 
get the idea :-)).

-- 
 i.