Kernels, kernels, kernels

brain0 · 2006-10-15 13:11:13

Morning people. I have just moved kernel26suspend2, with Nigel Cunningham's awesome suspend2 2.2.8.2 patch for better software suspend abilities, to the [extra] repository. Along with it, I moved the following external kernel modules:

ati-fglrx-suspend2
fcpci-suspend2
fcpcmcia-suspend2
intel-536ep-suspend2
intel-537-suspend2
ipw3945-suspend2
lirc-suspend2
madwifi-suspend2
ndiswrapper-suspend2
nvidia-suspend2
nvidia-legacy-suspend2
rt2500-suspend2
wlan-ng26-suspend2

I didn't add these modules for now, if anybody wants them, please submit a feature request in the bugtracker or post in this thread (Tom, if you want you can add and maintain truecrypt yourself)

ltmodem-suspend2
rt2x00-suspend2
slmodem-suspend2
tiacx-suspend2
truecrypt-suspend2

In other news, I will start work on kernel26mm today. It has been lying around for over a year without any updates. The -mm tree, maintained by Andrew Morton, provides some features that may or may not make it to the official Linus tree. It may be unstable and incompatible to some external kernel modules, but people who really want to be bleeding-edge can get it from the [extra] repository soon. I will update you in this thread.

iphitus · 2006-10-15 13:13:30

add rt2x00 if you want, ill maintain it.

Lone_Wolf · 2006-10-15 13:30:35

let's see : stock, beyond, ck, suspend2 .
that makes 4 official kernels.

I had the impression arch devs wanted to reduce the number of kernels ?

brain0 · 2006-10-15 13:30:40

done

brain0 · 2006-10-15 14:03:54

Lone_Wolf wrote:

let's see : stock, beyond, ck, suspend2 .
that makes 4 official kernels.
I had the impression arch devs wanted to reduce the number of kernels ?

Make it 5, we will also get -mm. But then, it should be enough IMO. I don't have a problem with the number of kernels, as long as I maintain them.

Anonymo · 2006-10-15 15:02:58

brain0, you're awesome.

brain0 · 2006-10-15 22:04:52

Okay, it is done: I brought kernel26mm back up to speed. The current version is 2.6.19-rc1-mm1, which has quite some changes to the current linux kernels:

* ext4dev added, reiser4 included
* most likely other features I don't know about
* new libata. I have to say something about this. New libata is not only in this -mm kernel, but will also be merged in 2.6.19/20. This means that there are now two types of ide drivers: old ide drivers and new libata-based "pata" ide drivers.

This means the following:

* If you boot from sata you NEED mkinitcpio 0.5.8 or later. You also need to use the "ata" hook instead of the "sata" hook for now. Actually, you can use "ata sata", so your mkinitcpio.conf will work for both the stock and the -mm kernel.
* If you boot from ide, you can leave everything as is. If you like, you can replace "ide" with "ata" in your mkinitcpio configuration (this - again - requires mkinitcpio 0.5.8 or later). But if you do this, all your ide devices will show up as sdX instead of hdX. If you do this and have both sata and ide, you may run into trouble with the controllers randomly switching their order, so you have to define it by explicitly loading the modules in the right order. (note for jmicron users: jmicron is only available in libata, not in ide)
* Please, please, please, only use this if know what you are doing. And leave another kernel as fallback. Thank you so much.

I only compiled a few external modules for now:

* lirc-mm
* nvidia-mm
* ati-fglrx-mm

I added the first two because I need them on the machine where I test this, and the last because it is likely to be used. Right now, I am not planning to add all available module packages for -mm, but only add some on demand, as -mm is only going to have a few users anyway. So if you are planning to use -mm and need a specific module, then post here or add a feature request in the bugtracker.

If I forgot to include a feature/option in -mm that you want to use, please tell me. Have fun.

Pierre · 2006-10-15 22:20:07

Nice work, but I think that all those experimental kernels should not stay in current nor extra but in unstable or perhaps testing.

brain0 · 2006-10-15 22:24:16

Pierre wrote:

Nice work, but I think that all those experimental kernels should not stay in current nor extra but in unstable or perhaps testing.

I was already in extra, so I decided to leave it there. The rest of the kernels is not unstable at all, so there is no problem. I don't think there won't be more kernels added now anyway.

Pierre · 2006-10-15 22:34:51

ok, but I do not think anyone used mm; it was quite outdated. So this might be a chance to move it to unstable. (the other kernels might stay in extra)

Some day we`ll need a seperate repository for all those kernels and modules. ;-)

Exclamation · 2006-10-16 02:35:57

Thanks brain0, i used to use mm before it became outdated. I will try it out when I get some time. Just 1 req, could you add an ndiswrapper module?

brain0 · 2006-10-16 08:36:48

yes, I'll add it if it builds.

EDIT: I just saw that phrakture already added ndiswrapper-mm earlier. Thanks

phrakture · 2006-10-16 14:42:36

brain0 wrote:

yes, I'll add it if it builds.
EDIT: I just saw that phrakture already added ndiswrapper-mm earlier. Thanks

Yep, but CVS was misbehaving so I couldn't upgrade ndiswrapper-suspend2 or ndiswrapper-ck to v 1.26. I gave up as it was like 2am

shadowhand · 2006-10-16 16:05:00

This brings up a point I've been meaning to ask: Why isn't suspend2 in the kernel26 package? I don't see a reason for a kernel package with nothing different except suspend2.

phrakture · 2006-10-16 16:14:09

shadowhand wrote:

This brings up a point I've been meaning to ask: Why isn't suspend2 in the kernel26 package? I don't see a reason for a kernel package with nothing different except suspend2.

Because the stock kernel is as closely vanilla as it can get. That has always been a philosophical point. It doesn't have suspend2 for the same reason that linus hasn't merged suspend2 - no one really feels it's entirely stable.

On a personal note, I never suspend my machines, so I have no need for it. Desktop users also probably have no need for it. Rather than force an unaccepted (by the kernel mainline) patch down someone's throat, it's better to provide the option to use it or not. This is what we have here.

hussam · 2006-10-16 16:44:39

phrakture, actually there are uses for suspend on desktops. I suspend my machines when for example there is a thunderstorm so if the electricity goes out, I don't lose my work.
In any case the suspend in the stock kernel works fine but the suspend2 scripts are a bit more powerful.

Romashka · 2006-10-16 16:57:39

Aa phrakture already said, the main reason for suspend2 not in kernel26 is that kernel26 (and most packages in Arch) tries to be as close to vanilla as possible.
I don't see a problem with the number of kernels and that kernel26suspend2 is kernel26+suspend2 only.

Lets see on this from another side.
There is kernel26 which is vanilla + very few patches.
There's also kernel26beyond, which has suspend2 and much more.
Why there was a reason for kernel26suspend2? Because there were some users that need suspend2 but don't need all other features from kernel26bayond. No problem.
So why not merge suspend2 into kernel26? The reason is the same! Because there are users that don't want any additional features in kernel26, especially when they don't use them.
Maybe Linux 3.0 will be more modular, but until that time we have different kernels/patchsets for different feature sets. And I don't see big problem here.

Thikasabrik · 2006-10-16 17:07:56

Suspend is a very nice thing for desktop users like me. I have my power button set to suspend the system to disk (after locking the screen). When I power on I am dumped straight to my desktop (after entering my password) in much less time than it takes to fully boot the system and with my email client and news-reader ready and running. I can also pick up web browsing where I leave off. Of course, I assume that a desktop user does occasionally *want* to power off the machine... notably the case if it is in his/her bedroom

PS: network-manager means all network configuration is automagically taken care of - it will take down interfaces and even switch epiphany into offline mode via a dbus call when gnome-power-manager calls for a suspend. How handy.

PPS: for me standard swsusp works just as well as suspend2

shadowhand · 2006-10-16 18:35:29

phrakture wrote:

shadowhand wrote:
This brings up a point I've been meaning to ask: Why isn't suspend2 in the kernel26 package? I don't see a reason for a kernel package with nothing different except suspend2.
Because the stock kernel is as closely vanilla as it can get. That has always been a philosophical point. It doesn't have suspend2 for the same reason that linus hasn't merged suspend2 - no one really feels it's entirely stable.

Have you looked at the patchlist for kernel26?

Really, I don't see what kind of effect having it there, doing nothing does. afaik, the suspend2 guy spends a lot of time making sure that suspend2 doesn't change the way the kernel works, until suspend2 is invoked.

Romashka · 2006-10-16 18:58:50

shadowhand wrote:

Have you looked at the patchlist for kernel26?
Really, I don't see what kind of effect having it there, doing nothing does. afaik, the suspend2 guy spends a lot of time making sure that suspend2 doesn't change the way the kernel works, until suspend2 is invoked.

Yes, I looked. And I don't see the reason for "LOL" here.

acpi-dsdt-initrd-v0.8.2-2.6.18.patch
r1000-build.patch
jmicron-ide.patch
wacom-volito.patch

So...??? Does these patches add some new functionality? :?

BTW, why r1000-build.patch and wacom-volito.patch is in kernel26 but not in kernel26suspend2? Shouldn't kernel26suspend2 be just the same as kernel26 + suspend2 patch?

brain0 · 2006-10-16 19:11:15

r1000 is in kernel26suspend2. wacom has only been added recently, and I didn't feel like adding it.

Romashka · 2006-10-16 19:31:16

brain0 wrote:

r1000 is in kernel26suspend2. wacom has only been added recently, and I didn't feel like adding it.

Oh, didn't notice r1000, sorry.

Maybe suspend2 will be finally included in vanilla kernel, so there will be no suspend2-related discussions at all.

Recent 2.6.1x kernels tend to contain much features merged each release.
There will be libata and ecryptfs merged into 2.6.19 and there were even plans to merge Reiser4 finally in 2.6.20 (but I don't know how it turns now, because of Reiser's problems with the police).
All 2.6.x series is in much more development than 2.4.xx few years ago.

phrakture · 2006-10-16 19:34:54

Romashka wrote:

BTW, why r1000-build.patch and wacom-volito.patch is in kernel26 but not in kernel26suspend2? Shouldn't kernel26suspend2 be just the same as kernel26 + suspend2 patch?

Good question. Also, if those are modular, we should probably build them outside the kernel too.

Dusty · 2006-10-16 20:00:17

Hey guys,

This thread seems a little like a developer's forum... don't you guys usually discuss development stuff in a top-secret closet in a hidden cave at the top of some unknown mountain somewhere?

I'm not sure how relevant certain portions of this discussion are to the community (e.g: who adopts what packages). Some things should likely be in the bug tracker instead of the forum too (e.g: missing patches on alternative kernels).

Dusty

brain0 · 2006-10-16 23:55:53

Dusty, it was intended to informational only, but some developer discussions may have sneaked in.

Arch Linux

#1 2006-10-15 13:11:13

Kernels, kernels, kernels

#2 2006-10-15 13:13:30

Re: Kernels, kernels, kernels

#3 2006-10-15 13:30:35

Re: Kernels, kernels, kernels

#4 2006-10-15 13:30:40

Re: Kernels, kernels, kernels

#5 2006-10-15 14:03:54

Re: Kernels, kernels, kernels

#6 2006-10-15 15:02:58

Re: Kernels, kernels, kernels

#7 2006-10-15 22:04:52

Re: Kernels, kernels, kernels

#8 2006-10-15 22:20:07

Re: Kernels, kernels, kernels

#9 2006-10-15 22:24:16

Re: Kernels, kernels, kernels

#10 2006-10-15 22:34:51

Re: Kernels, kernels, kernels

#11 2006-10-16 02:35:57

Re: Kernels, kernels, kernels

#12 2006-10-16 08:36:48

Re: Kernels, kernels, kernels

#13 2006-10-16 14:42:36

Re: Kernels, kernels, kernels

#14 2006-10-16 16:05:00

Re: Kernels, kernels, kernels

#15 2006-10-16 16:14:09

Re: Kernels, kernels, kernels

#16 2006-10-16 16:44:39

Re: Kernels, kernels, kernels

#17 2006-10-16 16:57:39

Re: Kernels, kernels, kernels

#18 2006-10-16 17:07:56

Re: Kernels, kernels, kernels

#19 2006-10-16 18:35:29

Re: Kernels, kernels, kernels

#20 2006-10-16 18:58:50

Re: Kernels, kernels, kernels

#21 2006-10-16 19:11:15

Re: Kernels, kernels, kernels

#22 2006-10-16 19:31:16

Re: Kernels, kernels, kernels

#23 2006-10-16 19:34:54

Re: Kernels, kernels, kernels

#24 2006-10-16 20:00:17

Re: Kernels, kernels, kernels

#25 2006-10-16 23:55:53

Re: Kernels, kernels, kernels

Board footer