kernel 2.6.0-test4 out!

Message

lucida · Post by **lucida** » Sat Aug 23, 2003 3:01 am

Compiling now..hope it can fix the palm sync problem on test3.

Jevin Sweval · Post by **Jevin Sweval** » Sat Aug 23, 2003 3:29 am

omg... i just went from test1'mm1 to test3'mm3 two days ago.

my luck

how long does it take for the mm kernel to get updated in portage?

Dominique · Post by **Dominique** » Sat Aug 23, 2003 4:01 am

Hello

This new test kernel is awesome for me. All my devices now Work !!
and it's lightning very very fast.

Thanks to all kernel developpers for this great JOB...

LJ · Post by LJ » Sat Aug 23, 2003 7:28 am

Dominique wrote:Hello

This new test kernel is awesome for me. All my devices now Work !!
and it's lightning very very fast.

Thanks to all kernel developpers for this great JOB...

Are you talking about 2.6 in general or test4-mm1 specifically?

Anyhow, does anyone know if the premptive scheduling has been fixed? Mplayer still takes up assloads of CPU time to play normal mpeg4 movies (like 26% time when it should be taking 5%)

Lovechild · Post by **Lovechild** » Sat Aug 23, 2003 7:42 am

and cpumask_t is merged... yay for that patch..

TheCoop · Post by **TheCoop** » Sat Aug 23, 2003 9:08 am

does anyone know when reiser4 will be officially merged?

Lovechild · Post by **Lovechild** » Sat Aug 23, 2003 9:27 am

TheCoop wrote:does anyone know when reiser4 will be officially merged?

That will take a while... I run the snapshot patch and it ain't all that pretty at present - plus it needs LOTS of testing, my bet is that within a month it will be merged in -mm or some other testing branch and be included in mainstream after the release of 2.6.0 - development seems to be hitting a small wall, plus there's a lot of work going into making sure the CPU usage is lower.

There are snapshot to try, and it's stable enough for usage on a desktop - but Hans and the guys are aiming for rock stability for inclusion - since this is a fully atomic filesystem and an otherwise new and innovative feature filled release. I run it, and I'm happy with it.

lucida · Post by **lucida** » Sat Aug 23, 2003 9:34 am

Can i upgrage from v3 to v4 without losing data? have no idea at present.

Lovechild · Post by **Lovechild** » Sat Aug 23, 2003 10:16 am

lucida wrote:Can i upgrage from v3 to v4 without losing data? have no idea at present.

the easy answer: no

the hard answer: probably

if you have a seperate partition or some spare space you could incrementaliy convert your partition. There are tools to help you do this, I think someone at namesys wrote it so I would expect a future version to support r3->r4 conversion. This is however a tad dangerous.

jufoa · Post by **jufoa** » Sat Aug 23, 2003 10:17 am

These new test kernels are coming out faster than i manage to install them =)

Have you guys noticed, that some programs will not compile if the linux symlink in /usr/src is pointing to these test kernels.

Yarrick · Post by **Yarrick** » Sat Aug 23, 2003 10:18 am

compiling now.. hopefully it handles the swap better

AlterEgo · Post by **AlterEgo** » Sat Aug 23, 2003 11:03 am

Findings so far:
1) My tux-bootlogo (vesa-fb; missing since 2.5.70) is back

2) Sensor readouts (missing since test-2) are back

3) My alsasound (82xx) has started behaving badly and is looping the last half-second of everything I play forever: console beeps are making my crazy, and I cannot figure out why?

blofeld · Post by **blofeld** » Sat Aug 23, 2003 11:17 am

Well here it doesn't boot when ACPI is enabled (was working before with all 2.6-test* kernels). It hangs after initializing
the USB devices or right before enabling ALSA (which is compiled into the kernel - not module).

Motherboard: ABIT KG7-RAID

w/o ACPI it works fine.

Coming from 2.6-test3-mm2 I must say, the scheduler is by far not so good for interactivity than the kernels with Con's patches applied. So I'm eagerly waiting for mm1

wishkah · Post by **wishkah** » Sat Aug 23, 2003 1:42 pm

jufoa wrote:These new test kernels are coming out faster than i manage to install them =)

Have you guys noticed, that some programs will not compile if the linux symlink in /usr/src is pointing to these test kernels.

Only experienced problems when the symlink is missing/is wrong and you try to emerge nvidia-kernel stuff...

Hopefully test4 clears up some sound probs (alsa, intel 8x0) and keyboard issues I had or I might just get back to test1, that was the best kernel I ever used.

mslay · Post by **mslay** » Sat Aug 23, 2003 2:49 pm

got tested it here, no problems so far with nVidia GeForce 2 card and maestro3 sound driver.

under heavy load(SAP R3 with Oracle, regeneration of all SAP ABAPs) no problems!

looks like the fastest linux kernel I've seen so far...
Good work, kernel developers!!!!

x20vmk4x · Post by **x20vmk4x** » Sat Aug 23, 2003 2:55 pm

i dont know if this is too off-topic, but i cant get the patch from test3 to test4 to work. i did what the kernel-howto on tldp.org said, but it cant find any of the files to be patched. i guess ill download the full source, but that kinda seems like a waste :-/

TIA
bryan

Wedge_ · Post by **Wedge_** » Sat Aug 23, 2003 3:02 pm

blofeld wrote:Well here it doesn't boot when ACPI is enabled (was working before with all 2.6-test* kernels). It hangs after initializing
the USB devices or right before enabling ALSA (which is compiled into the kernel - not module).

Same for me, nForce2 board.

stefanwa · Post by **stefanwa** » Sat Aug 23, 2003 3:10 pm

Wedge_ wrote:
blofeld wrote:Well here it doesn't boot when ACPI is enabled (was working before with all 2.6-test* kernels). It hangs after initializing
the USB devices or right before enabling ALSA (which is compiled into the kernel - not module).
Same for me, nForce2 board.

I have an nForce2 board with ACPI enabled and have none of the problems you mentioned... ??

wishkah · Post by **wishkah** » Sat Aug 23, 2003 3:32 pm

Ooooh, man. test4 doesn't even boot, although I use the same basic .config I used with test1, which worked fine. It gives no kernel panic, it just "stops" and does nothing after the message "using antipacitory scheduling elevator" (not sure about second word

) After that test1 gives out something about a disk-drive controller, I think... the notebook has an external USB floppy that is not plugged in, might this cause the problem?

Info:
Toshiba 5200-701, acpi-only, p4-m 1.9 GHz, gf4 mx 460, 512 MB DDR, 2x 40 GB (toshiba+IBM), no "experimental" stuff enabled in kernel-config.

I checked out the kernel mailing list, didn't find something similar...

Superman53142 · Post by **Superman53142** » Sat Aug 23, 2003 4:03 pm

AlterEgo wrote:Findings so far:
1) My tux-bootlogo (vesa-fb; missing since 2.5.70) is back
2) Sensor readouts (missing since test-2) are back
3) My alsasound (82xx) has started behaving badly and is looping the last half-second of everything I play forever: console beeps are making my crazy, and I cannot figure out why?

The tux logo and sensor readings were back in test3-mm3

But I too now have the problem with via82xx screwing up. It seems to only occur for me when piping through esound. Going directly to alsa is fine...

EDIT: Appearantly there were numerous updates of ALSA in test4, so one of those broke it for via82xx soundcards :-/

agent_jdh · Post by **agent_jdh** » Sat Aug 23, 2003 4:07 pm

blofeld wrote:Well here it doesn't boot when ACPI is enabled (was working before with all 2.6-test* kernels). It hangs after initializing
the USB devices or right before enabling ALSA (which is compiled into the kernel - not module).

Motherboard: ABIT KG7-RAID

w/o ACPI it works fine.

Coming from 2.6-test3-mm2 I must say, the scheduler is by far not so good for interactivity than the kernels with Con's patches applied. So I'm eagerly waiting for mm1

Try it with ACPI but using the boot option pci=noacpi

You still get ACPI functionality (e.g. halt on power button press), but it doesn't use ACPI to enumerate PCI IRQs.

There have been considerable changes with SMP/ACPI/APIC stuff recently that have broken stuff.

NiklasH · Post by **NiklasH** » Sat Aug 23, 2003 4:08 pm

littleendian wrote:Ooooh, man. test4 doesn't even boot, although I use the same basic .config I used with test1, which worked fine. It gives no kernel panic, it just "stops" and does nothing after the message "using antipacitory scheduling elevator" (not sure about second word ) After that test1 gives out something about a disk-drive controller, I think... the notebook has an external USB floppy that is not plugged in, might this cause the problem?

Info:
Toshiba 5200-701, acpi-only, p4-m 1.9 GHz, gf4 mx 460, 512 MB DDR, 2x 40 GB (toshiba+IBM), no "experimental" stuff enabled in kernel-config.

I checked out the kernel mailing list, didn't find something similar...

Have you tried 'pci=noacpi' or 'acpi=off' in the kernel params? That fixed my lockup.

Grue · Post by **Grue** » Sat Aug 23, 2003 5:07 pm

Seems to work pretty much ok otherwise but I cannot get nvidia-driver to work, module compiles fine and I can even insmod it into kernel, but cannot start X...

(**) NVIDIA(0): Depth 24, (--) framebuffer bpp 32
(==) NVIDIA(0): RGB weight 888
(==) NVIDIA(0): Default visual is TrueColor
(==) NVIDIA(0): Using gamma correction (1.0, 1.0, 1.0)
(**) NVIDIA(0): Option "NvAGP" "2"
(**) NVIDIA(0): Option "CursorShadow"
(**) NVIDIA(0): enabling cursor shadow
(**) NVIDIA(0): Use of AGPGART requested
(--) NVIDIA(0): Linear framebuffer at 0xC8000000
(--) NVIDIA(0): MMIO registers at 0xD9000000
(EE) NVIDIA(0): Failed to initialize the NVIDIA kernel module!
(EE) NVIDIA(0): *** Aborting ***

I tried other agp modes too with no luck, nvidia module works ok with 2.4 kernel so..

Any ideas what to try ?

agent_jdh · Post by **agent_jdh** » Sat Aug 23, 2003 5:15 pm

Grue wrote:Seems to work pretty much ok otherwise but I cannot get nvidia-driver to work, module compiles fine and I can even insmod it into kernel, but cannot start X...

(**) NVIDIA(0): Depth 24, (--) framebuffer bpp 32
(==) NVIDIA(0): RGB weight 888
(==) NVIDIA(0): Default visual is TrueColor
(==) NVIDIA(0): Using gamma correction (1.0, 1.0, 1.0)
(**) NVIDIA(0): Option "NvAGP" "2"
(**) NVIDIA(0): Option "CursorShadow"
(**) NVIDIA(0): enabling cursor shadow
(**) NVIDIA(0): Use of AGPGART requested
(--) NVIDIA(0): Linear framebuffer at 0xC8000000
(--) NVIDIA(0): MMIO registers at 0xD9000000
(EE) NVIDIA(0): Failed to initialize the NVIDIA kernel module!
(EE) NVIDIA(0): *** Aborting ***

I tried other agp modes too with no luck, nvidia module works ok with 2.4 kernel so..

Any ideas what to try ?

Did you compile agpgart+support for your mobo chipset into the kernel or as a module? NvAGP option of 2 suggests that you want to use agpgart instead of the nvidia kernel agp driver.

Maybe there's a bug in agpgart in the 2.6 test kernels. You could try the nvidia kernel module, emerge nvidia-kernel and switch the NvAGP optoin to 1 instead of 2. And make sure agpgart isn't compiled into the kernel or loaded as a module if you decide to use the nvidia module.

x20vmk4x · Post by **x20vmk4x** » Sat Aug 23, 2003 5:16 pm

make sure your /usr/src/linux link is pointing to the right kernel dir, then

ACCEPT_KEYWORDS="~x86" emerge -pv nvidia-kernel nvidia-glx

still doesnt work w/test4, tho