network slowness/freez-up since update 10/11

[Jon Noack]

On 10/13/04 19:08, Andrey Chernov wrote:
> On Wed, Oct 13, 2004 at 04:58:16PM -0700, Kris Kennaway wrote:
>> On Thu, Oct 14, 2004 at 03:44:12AM +0400, Andrey Chernov wrote:
>>> Even with its locking mess it works perfectly in -current until
>>> late August.
>> 
>> You mean, until rwatson changed the default to debug.mpsafenet=1? :-)
> 
> Your guess is precisely right! :-)
> 
> (IMHO making such commit without testing major drivers such as if_de was 
> wrong step)

Considering its scope, I believe the mpsafenet project has been handled 
quite well.  Those involved went to great lengths (i.e. did a lot of 
testing) to ensure networking was usable throughout the process, and for 
the most part they succeeded.  There are still a few rough edges (like 
if_de locking), but they're being worked on...

>> debug.mpsafenet=0 is the known workaround, already mentioned in
>> this thread. It hides the locking bugs by running the driver under Giant.
> 
> Thanx, I already figure that out. Now building new kernel...

That last sentence may be superfluous, but just in case:
There's no need to build a new kernel for this -- debug.mpsafenet is a 
tunable and can be set in /boot/loader.conf or from the boot loader 
(still requires a reboot, though).

Jon