Linux 2.6.33-rc1 Kernel Released

[Kano]

Well when you benchmark the compile time bfs is usally a bit faster.

[skirst]

Well when you benchmark the compile time bfs is usally a bit faster.

Maybe so. Never tested it. Qt takes quite a while to compile Based on the results of kernel compile times posted by Con, I'd gauge that I could potentially save anywhere from 0 to 1 minute. Here's to that extra minute goofing off I will trust mainline a little more for stability, whether it's there or not. I thank Con for his hard work, it seems to me BFS made a difference and served it's purpose very well. And perhaps, BFS's work is not yet complete. We shall see.

[trapDoor]

...Also where is reiser4 . I want to see Linus bitching for reiser4 inclusion on next widow merge like he did with nouveau...

I'm glad that kernel developers are focused on improving btrfs and there is no Linus' bitching required to push it!
Regarding reiser4. I used to use it. Very efficient and decent fs but since it's not developed anymore (?) and as far as I know it didn't approach the stage to be called stable, I don't think Linus would like to include that 'dead' project in the kernel.

[kraftman]

It seems Con has made his point. I agree (based on my experience), there is no need for BFS in mainline at this point.

That's it.

@Kano

Well when you benchmark the compile time bfs is usally a bit faster.

Yes, but it suffer in other things like big drop in fps when playing games and listening to the music same time (at least version I tried).

[hmmm]

I seem to get 2.6.33-rc1 from ubuntu's mainline kernel failing to boot into gdm ...

[zika]

I seem to get 2.6.33-rc1 from ubuntu's mainline kernel failing to boot into gdm ...

If You have ATI r600 or r700 put /lib/firmware/2.6.33-999-generic/radeon/R600_rlc.bin or /lib/firmware/2.6.33-999-generic/radeon/R700_rlc.bin in place... It did not work for me either (with KMS forced) before that either. Witout KMS it worked. (ATI HD3650).

[uwgandalf]

Did they fix the eternal low responsiveness when doing IO already?

I'm getting so fed up with it, and there's no kernel maintainer who's addressing this bug.

http://bugzilla.kernel.org/show_bug.cgi?id=12309#c360
https://bugs.launchpad.net/ubuntu/+s...22/+bug/131094

Can someone with mainstream hardware (sata; dual-core, 2GB) perform a read test with tiobench 64MB and 32 threads, and still open iotop and use firefox in a usable manner?

[frantaylor]

What is the preferred distribution to use for testing new kernels?

[rvdboom]

Indeed, I compiled the kernel source as I always do and this time at reboot, I was requested a firmware or something.
I thought they were compiled in the kernel, weren't they?

[bridgman]

All new microcode is going into the /lib/firmware folder for distribution rather than being built directly into the kernel. I believe the latest microcode file (rlc) needs to go into the initrd so that drm can find it at boot.

It is still possible to build the firmware into the kernel yourself (ie on your own system) but I think the overall direction for the kernel is increased use of lib/firmware and initrd.