On Mon, 08 Aug 2011 09:16:02 +0000, Jedibeeftrix wrote:
> RC1 of kernel 3.1 has arrived and given the usual 3 month release
> schedule it should arrive just a week after RC1 of opensuse 12.1.
>
> ‘[Phoronix] Linux 3.1-rc1 Kernel: A “Pretty Normal Release”’
> (http://www.phoronix.com/scan.php?page=news_item&px=OTc2NA)
>
> How about it?
I would be surprised if they used it given the timing - I guess it
depends on whether it’s stable or not when it’s released.
I imagine if it weren’t stable, the openSUSE project would be blamed for
using an unstable (or not fully tested) kernel, and that wouldn’t be a
good thing.
On 08/08/2011 11:37 AM, Jim Henderson wrote:
> On Mon, 08 Aug 2011 09:16:02 +0000, Jedibeeftrix wrote:
>
>> RC1 of kernel 3.1 has arrived and given the usual 3 month release
>> schedule it should arrive just a week after RC1 of opensuse 12.1.
>>
>> ‘[Phoronix] Linux 3.1-rc1 Kernel: A “Pretty Normal Release”’
>> (http://www.phoronix.com/scan.php?page=news_item&px=OTc2NA)
>>
>> How about it?
>
> I would be surprised if they used it given the timing - I guess it
> depends on whether it’s stable or not when it’s released.
>
> I imagine if it weren’t stable, the openSUSE project would be blamed for
> using an unstable (or not fully tested) kernel, and that wouldn’t be a
> good thing.
There will be a feature freeze on 12.1 after MS4. As 3.1 will not be anywhere
close to stable in 2 weeks, using it will be out of the question. That said, I’m
using 3.1-rc1 and finding it stable, but my setup is not nearly as exotic as
many others. I found a couple of problems with the kernels in the pre-rc1 merge,
but patches for those were already available.
On 08/09/2011 10:16 AM, Jedibeeftrix wrote:
>
> dale14846;2373223 Wrote:
>> is in the kernel vanilla repository
>> ‘Index of /repositories/Kernel:/vanilla’
>> (http://download.opensuse.org/repositories/Kernel:/vanilla/)
>> so one might consider it in the pipe
>
> cheers, i certainly hope so.
What will 3.1 have that people are so desperate to get it?
From what I have read, it doesn’t look like the Linux 3.1 kernel is set to fix up any of the purported power regressions.
But there could be some graphic driver improvements which will make our support tasks easier:
Nouveau Fermi support no longer requires manually loading micro-code after extracting it from the binary driver (i.e. out of the box GeForce 400/500 support).
Some time back we had a number of rants from Intel Poulsbo graphic device users, moaning about the very poor GNU/Linux support for that hardware.
My wife is purchasing a laptop with Intel Sandybridge architecture, so it would be nice to see superior support for that hardware.
Still, I would not use the word ‘desperate’ and IMHO one needs to trade off the benefit of the risk of late 3.1 introduction. If my memory serves me correctly, late kernel update introductions in an openSUSE release cause more problems than naught, hence while I would like to see 3.1, its a bit risky to put it in and is quite possibly not worth the risk, especially with Tumbleweed-12.1 likely to be available.
i realise this is a standard response to those suspected of suffering from versionnumberitus, but you could have answered your own question by clicking on the link in the OP.
The best way IMHO to answer in such cases (granted it can be hard at times) is to simply answer the question, and neither criticize the response nor the answer nor the person who answers the answerer, nor the person who answers the x who answers the y, who answers the z … (ad infinitum) … < says me as the pot calling the kettle black > … :X
… anyway, thanks all your forum participation. Its most appreciated even if we don’t all have the same views ! …
Installed kernel 3.1-rc1, and updated same today. A few rough edges, such as AppArmor startup failure, and recovery from suspension to RAM. Most significant initial observation is decreased processor usage and processor temperatures, ceteris parabus, compared with kernel 3.0.0-8.1.
Running on battery was also surprising: while the power management for the Intel Arrandale/Ironlake is still incomplete, running on battery showed a significant (23%) increase in battery life, ceteris parabus, to prior kernel. (I was immediately suspicious, so I repeated the test: Firefox with 4 windows, 74 tabs, 3 tabs with active graphics (MLB “Gameday”), one YouTube video/music, and a real-time sports-score monitor). I wonder if the reports concerning power regressions are bearing some results?
One downside: dual monitors die a rapid death, so a little investigation is in order.
One new (to me) early boot message:
BUG: scheduling while atomic: swapper/0/0x10000002
needs research. Also, the (recoverable) “black screen” is still present, while kernel 3.1-rc1 does not appear to present a splash screen.
Wrt graphics aspects, when testing the latest kernel, do you also use the latest Mesa (such as the one currently in factory) ?
I’ve read ‘hints’ that it may be broken for some (which is why it has not migrated its way to Tumbleweed, and possibly not yet to the 12.1 milestone releases) but I’m most definitely not clear that is the case.
As far as breaking, I experience problems regularly with the pairing of xorg-x11-server and xorg-x11-driver-video. I believe that the problem(s) are associated with the Intel driver (specifically, the Ironlake support), as described is https://bugzilla.novell.com/show_bug.cgi?id=704052. The circumvention (at leat, in my case) is the regression of xorg-x11-driver-video, as described in the bug.