Linux 2.6.31 Kernel Is Ready With -RC8 Release

[phoronix]

Phoronix: Linux 2.6.31 Kernel Is Ready With -RC8 Release

Linus Torvalds has announced the release of the Linux 2.6.31-rc8 kernel this evening, which he expects will be the last release candidate before this quarterly Linux kernel update goes gold in just over one week. Linux 2.6.31-rc8 doesn't bring any particularly interesting fixes, but just small corrections around the kernel. In the Linux 2.6.31-rc8 release announcement, Linus mentions he will be gone for the next week, but is planning on releasing the Linux 2.6.31 kernel in time for Labor Day...

http://www.phoronix.com/vr.php?view=NzQ4Ng

[Tinuva]

Well I for one can't wait for 2.6.31 to hit final, that means I will finally have a somewhat opensource driver for my X-Fi that ships with the kernel instead of manually installing it with each upgrade.

Also I hope Arch will get it soon after it is final, they usually don't take too long.

[entropy]

So, that's actually a bad thing for ATI R6xx+ owners, as the coming merge window for 2.6.32 is too early then, I guess?
Some days ago I read (IRC iirc) that the R6xx+ code is not ready yet for merging...

Hope I'm wrong though.

[DragonSA]

Is anyone else unnerved to see 8 Release Candidates? In my opinion a release candidate should be a we-think-this-is-ready-but-to-be-on-the-safe-side...

With 8 RCs it means the developers thought they had a winning kernel 7 times and got it wrong, perhaps they should have released a few more beta's? All the other projects that I am aware of have 1, maybe 2 and at the most 3 RCs, certainly not 8!

If I don't understand the Linux way of releasing please enlighten me.

[deanjo]

Is anyone else unnerved to see 8 Release Candidates? In my opinion a release candidate should be a we-think-this-is-ready-but-to-be-on-the-safe-side...

With 8 RCs it means the developers thought they had a winning kernel 7 times and got it wrong, perhaps they should have released a few more beta's? All the other projects that I am aware of have 1, maybe 2 and at the most 3 RCs, certainly not 8!

If I don't understand the Linux way of releasing please enlighten me.

Different projects have their own interpretation of levels of development. Depending on the base of what you consider Alpha / Beta etc a RC status may be justified. For example if the linux kernel considers 2.5 series of kernel the alpha or beta then incremental updates of the kernel can be considered RC's as they are minor updates when compared to the initial development of the 2.6 kernel. Other projects simply use a build number instead of a Alpha/Beta/RC status.

[nanonyme]

With 8 RCs it means the developers thought they had a winning kernel 7 times and got it wrong, perhaps they should have released a few more beta's? All the other projects that I am aware of have 1, maybe 2 and at the most 3 RCs, certainly not 8!

Actually throughout the release candidate process bugs are fixed. You do understand probably that the kernel is the single most important part of the operating system security-wise and it's important that it goes through very thorough screening?

[highlandsun]

-rc8 always hangs in udev during bootup for me. I'm guessing it probably works fine for everyone else otherwise there'd be more complaints by now... I built it by applying the rc6-rc7 incremental and rc7-rc8 on top of my previous rc6 tree.

In rc6 the ath9k wifi driver keeps hanging with DMA errors. Every driver since then in the wireless-testing branch has done the same. I think the last usable one I had was rc5, and that one gets a weak signal and still drops the connection occasionally (even though other computers nearby show 100% signal strength and no problems). I didn't ever run 2.6.30, my last kernel with solid wifi was 2.6.29.6.

[smitty3268]

If I don't understand the Linux way of releasing please enlighten me.

You don't If it makes you feel better, think of RC1-3 as alphas, 4-7 as betas, and 8+ as the actual release candidates. I think they said they skip the alpha and beta names because no one was testing them, but by changing the name to a RC all of a sudden people thought it was OK to run. So it's nothing but a marketing ploy, really - the releases always take around 3 months with 8 or so RCs, so it's not like it's difficult to figure out what shape the kernel is in at any point in time.

[Craig73]

I only just tried a recent release candidate, and on my Inspiron 700M which uses Intel 855GME just goes black on boot... I guess I should dig around for a fix to see if there is a regression or my settings are wrong

[combuster]

I've been testing 2.6.31 since rc1 and there was no hickups on my dell vostro 1310, everything worked as it should, from the start it gave me better performance of my x3100...