The second NetBSD Bugathon was even more successful than the first: three times as much people (both users and developers) showed up, and together they fixed more than 300 PR's on two days. See Elad's report and the Bugathon's homepage for the details.
To read all comments associated with this story, please click here.
Member since:2005-08-06
Pardon me for possibly missing out on progress of the teams here, but I didn't think FreeBSD completely dropped the M:N model, they rather develop in parallell?
Maybe I'm mistaken here but I thought that was the case, just like with the schedulers? (ULE/KSE).
Either way, I'd be delighted to hear about any NetBSD attempt on this, would, if even half successful bring a possibly brilliant newsitem to Osnews =).
Oh, and finally, if someone could clear up my mental mess with what FreeBSD is trying to do I'd be delighted.
Member since:2006-03-15
"Pardon me for possibly missing out on progress of the teams here, but I didn't think FreeBSD completely dropped the M:N model, they rather develop in parallell"
Don't take this as an authoritative answer. IMHO, while they have not abandoned M:N threading, the new push is to get 1:1 working and working fast. In fact, there was a major discussion on the list a few months ago regarding this and it looks like FreeBSD 7 will by default use 1:1 (libthread) rather than M:N (libpthread). However, because of the pluggable nature of threading for FreeBSD, development of M:N threading will go on in the background by a few developers, especially since it seems they are rather adamant of using M:N.
Member since:
2006-03-15
How is threading progressing in NetBSD? I believe that FreeBSD attempted to implement M:N threading but failed and is now moving towards 1:1. Sun tried to do it before FreeBSD attemped it but also failed. Linux devs did not even try to do it siting the complexity. However, NetBSD is now trying to go where others failed. Is there progress in removing Giant Lock? How is this coming along when compared to FreeBSD?