Hopefully this test patch should fix all the problems with BFS 404 on 2.6.39:
bfs404-test9.patch
Ubuntu Packages : grab the 2.6.39-ck1-3 package
Please report back if you haven't already! Thanks to everyone who has tested so far! Your feedback has been absolutely essential on this weird and wonderful bug.
I have two machines.
ReplyDeleteThose earlier patches couldn't work on both.
After applying test9, the two machines work well so far.
Running 2.6.39-ck1-3 on a MacbookPro3,1 on Ubuntu 10.04 ... so far so good!
ReplyDelete(P.s. - To get an initramdisk, I had to do update initramfs -c -k 2.6.39-ck1-3; Is this documented somewhere? Here's a heads-up to my fellow kernel newbies at least!.)
Runs without hassle, as most times on a mac Mini - 64bit.
ReplyDeleteWorking well for me with Chrome, Rhythmbox, VirtualBox on a quad-core Xeon with Nvidia proprietary drivers + BFQ.
ReplyDeleteThanks, ck.
Does the 2.6.39.1 release contain any code relevant to this problem?
ReplyDeleteNot as far as I can see. This does appear to be unique to an interaction with BFS and the new block plugging code. Although test9 does so far seem to fix all the remaining problems. There is a small possibility this could happen on mainline as well, but is much harder to hit due to design differences.
ReplyDelete@ck - we will keep banging away at test9. I'm assuming that given some stable run time (3-4 days ?) you'll call test9 ck2 and stable?
ReplyDeleteThanks graysky. That's precisely the plan.
ReplyDeleteToo bad it's such a small population. Oh the power: 1 person reporting prolems = 1k? 5k? :)
ReplyDeleteSo far so good here with test9. Since all of the previous test patches that worked well for others were still fucked off for me, I imagine if there's any remaining breakage here I'll run into it pretty soon. ;)
ReplyDelete