2.6.10-bk3-ck1-nitro2 "cd / && rm -rf *"

[discomfitor]

Two notes before I get into stuff:
1) Do not attempt the command in the title. I really shouldn't have to say this, but I know that if I don't, some idiot will either tell me to because they almost did it, or some bigger idiot might actually do it. Save yourself the trouble. Don't delete all your files and blame me.
2) There is no second note.

Code: Select all

2.6.10 #newest kernel version
-ck1 #con kolivas' patchset
-ITERAID #Feel the excitement of new hardware support!
-alsa 1.0.7 #(finally!)
-reiser4 #reiser4 support
-vesafb-tng #new fb driver
-fbsplash #new version of bootsplash
-config_hz #change cpu frequency
-chmp #change usb mouse polling rate
cflags-selection-for-nitro #change cflags for kernel
hostap #hostAP drivers
=============================
+bk3 #latest snapshot
+ac2 #alan cox's patchset
+ckdev #latest ck snapshot (not including supermount)
+alsa 1.0.7 #includes several misc fixes which may or may not affect things
+reiser4 #newest snapshot with extra patches; newer than morph-sources for once!
+vesafb-tng #slight changes which may or may not affect anything
+fbsplash #slight changes which may or may not affect anything
+chmp #new version of mouse polling patch
+sk98lin driver #sk98lin network driver
+parport fix #(http://lkml.org/lkml/2004/12/26/2)
+inotify
+swsusp2+win4lin #see note

Extra:

CONFIG_IOPRIO_WRITE
See that option? Do NOT enable it. Bad things happen if you do. I'm not even sure if you can find it in the menuconfig or the xconfig, but if you do, leave it alone. Perhaps in time it will grow mature enough to be used, but now it just breaks things.

y0zza: I looked into the time-sliced cfq, but it was a decision between that and a newer update from Con Kolivas. Suffice to say that it didn't make the cut this time.
Robin79: sk98lin driver is in!

To users of it8212: it compiles. I tested it.
To users of reiser4: it compiles. I tested it.
To users of swsusp2 and win4lin: it compiles. I tested it.
To users of this ebuild: it works. I tested it.

With all that having been said, I think this is my finest work to date. I fired up rhythmbox and it opened INSTANTLY, completely disregarding its normal policy of taking a second or two. Other things seem to be faster too. Best of all, I wasn't rushed on this one, so I had time to test it properly.

Oh, and I know I said that I would never again put in the swsusp2+win4lin patch. The newest one went on like warm butter though, so I really had no choice.

Oh, and there's a broken out directory on the site now. I can't guarantee that it will always be completely up-to-date, nor can I assure you that EVERY patch I used is in there. But most of them are in there.

I've been told that the orinoco driver works in this version, but I can't verify that. Therefore, the following may or may not apply.

Official orinoco statement:
I am not going to be trying to patch orinoco drivers in again until they update the patch. It's just too much of a hassle, and it breaks the drivers for other things when I try. That's not to say that seppe won't when he gets back, because he most likely has his own policy about such things. For now though, you'll have to do it yourself.

How to get orinoco working:
Simply copy the hermes and orinoco files from a previous kernel with working drivers. These files will be located in linux/drivers/net/wireless/ (where linux is the name of your kernel). Copy them into the same place in this kernel tree, and it will work.

If someone wants to post links to hermes+orinoco files that work for this, I will make a kernel patch (or you can do that part yourself) that will make it easier to get working in the future. It will still need to be manually applied though.

To learn how to use the new mouse polling patch, go here.

For the ati fix, go here.

For a cool automotive background picture made by M@rijn, go here.

Changed hosting location thanks to PickledOnion!
Ebuild
Patch

[desertstalker]

Nice patch, good to see that you've included swsusp2,

However when i use swsusp2 i get errors from the usb device drivers uhci-hcd and ehci-hcd that do not appear normally when using other kernels with swsusp2.

I get this in dmesg

Code: Select all

usb 2-1: device descriptor read/64, error -71

and the device does not initialise

Any ideas, i realize that it is probbably not your fault

thanks

EDIT: it only happens with USB mass storage devices, my mouse works fine, ill check the low performance usb driver.

EDIT2: Id oesnt only happen after suspending with swsusp2 but also after a clean boot, it is a driver problem (not usb-storage) as the ub driver doesnt work either.

It seems there may be some problems with the usb in this release.

Thanks again

[anir]

Great Work!

[r00tzz]

Sorry, but I'll try this only after the hollydays...

Happy New Year to everybody!!

as soon as I test it, I'll post here...

[Mandos]

As was the case with the last nitro, your links return a connection refused error. Could someone else mirror the ebuild and patch please?

[Pink]

Darckness: If you want a permanent and reliable home for the patches and ebuilds, let me know and I'll host them for you. I'll give you an ftp account to a nitro directory.

Until then:

ebuild here
patch here

HTH

[anir]

And another mirror: http://home.arcor.de/_aero/nitrosources/

[Rainmaker]

cool patchset (again )

cool title too. I tried it, but it somehow didn't work

Code: Select all

root@Medusa$ cd / && rm -rf *                                          (~)
zsh: sure you want to delete all the files in / [yn]? 

What should I answer?

Thanks! Compiling now...

[Mpemba Effect]

Good stuff Darckness! I'm running r1 and it's awesome, however is it just me or is the "chmp" mouse polling patch missing from r2?

[pleusicles]

Thanks for the new nitro Darckness!

I've got two problems when trying to compile it. First, I couldn't build floppy support as module:

Code: Select all

CC [M]  drivers/block/floppy.o
drivers/block/floppy.c: In function `init_module':
drivers/block/floppy.c:4598: error: parse error before "UTS_RELEASE"
make[2]: *** [drivers/block/floppy.o] Error 1
make[1]: *** [drivers/block] Error 2
make: *** [drivers] Error 2

It works if compiled into the kernel. (I enabled io fairness in the cfq scheduler - this is the only change from my working 2.6.10-nitro1 config - maybe it's related?)

My second problem is related to reiser4; it compiles (as module), but warns about two undefined symbols (find_get_pages, find_get_pages_tag). Adding the two corresponding EXPORT_SYMBOL lines to mm/filemap.c solved the problem. (Credits go to Mika's blog and google)

Here's the diff for 2.6.10-nitro2:

Code: Select all

--- /data/filemap.c     2004-12-30 14:26:23.000000000 +0100
+++ /usr/src/linux-2.6.10-nitro2/mm/filemap.c   2004-12-30 14:10:19.000000000 +0100
@@ -619,6 +619,8 @@
        return ret;
 }

+EXPORT_SYMBOL(find_get_pages);
+
 /*
  * Like find_get_pages, except we only return pages which are tagged with
  * `tag'.   We update *index to index the next page for the traversal.
@@ -640,6 +642,8 @@
        return ret;
 }

+EXPORT_SYMBOL(find_get_pages_tag);
+
 /*
  * Same as grab_cache_page, but do not wait if the page is unavailable.
  * This is intended for speculative data generators, where the data can

[COiN3D]

Btw, what is a good value for config_hz? Or, whats the best value for a desktop / gamer system?

[eikketk]

Thanks for the new nitro Darckness!

I've got two problems when trying to compile it. First, I couldn't build floppy support as module:

Code: Select all

CC [M]  drivers/block/floppy.o
drivers/block/floppy.c: In function `init_module':
drivers/block/floppy.c:4598: error: parse error before "UTS_RELEASE"
make[2]: *** [drivers/block/floppy.o] Error 1
make[1]: *** [drivers/block] Error 2
make: *** [drivers] Error 2

It works if compiled into the kernel. (I enabled io fairness in the cfq scheduler - this is the only change from my working 2.6.10-nitro1 config - maybe it's related?)

I fixed this for -ikke3. Easy fix, just revert the -bk2 change (i.e. add #include <linux/version.h> in drivers/char/floppy.c or so, see kernel.org's diff viewer)

My second problem is related to reiser4; it compiles (as module), but warns about two undefined symbols (find_get_pages, find_get_pages_tag). Adding the two corresponding EXPORT_SYMBOL lines to mm/filemap.c solved the problem. (Credits go to Mika's blog and google)

Here's the diff for 2.6.10-nitro2:

Code: Select all

--- /data/filemap.c     2004-12-30 14:26:23.000000000 +0100
+++ /usr/src/linux-2.6.10-nitro2/mm/filemap.c   2004-12-30 14:10:19.000000000 +0100
@@ -619,6 +619,8 @@
        return ret;
 }

+EXPORT_SYMBOL(find_get_pages);
+
 /*
  * Like find_get_pages, except we only return pages which are tagged with
  * `tag'.   We update *index to index the next page for the traversal.
@@ -640,6 +642,8 @@
        return ret;
 }

+EXPORT_SYMBOL(find_get_pages_tag);
+
 /*
  * Same as grab_cache_page, but do not wait if the page is unavailable.
  * This is intended for speculative data generators, where the data can

AFAIK only find_get_pages_tag needs an aditional export? Also fixed this for -ikke*

[COiN3D]

Hey, where is the chmp-patch? I didn't find it!

[seppe]

Just want to say that nitro-sources has it's own domain name now: http://www.nitro-sources.org

I will try to update everything there, but my exams are coming closer

[discomfitor]

Alright, I just updated (technically it's still uploading, but it should be done in about five minutes and I have to leave so...) the patch (delete the digest and remerge it to get it again) with the new reiser4 update from namesys, chmp (felt kinda retarded when I realized I'd just forgotten about it) the fixes listed in this thread, and a couple other fixes I made to get things compiling when I was testing it all.

Thanks for pointing stuff out, and good luck seppe.

Also, I just changed the location of the patch and ebuild to PickledOnion's site. It should be accessible to everyone now.

[Mpemba Effect]

Yay mouse polling is back in However there seems to be something wrong with sk98lin. I recieve the following compile error.

Code: Select all

  LD      drivers/media/dvb/ttusb-dec/built-in.o
  LD      drivers/media/dvb/built-in.o
  LD      drivers/media/radio/built-in.o
  LD      drivers/media/video/built-in.o
  LD      drivers/media/built-in.o
  LD      drivers/misc/built-in.o
  CC      drivers/net/sk98lin/skge.o
drivers/net/sk98lin/skge.c: In function `sk98lin_resume':
drivers/net/sk98lin/skge.c:678: error: too many arguments to function `pci_restore
_state'
drivers/net/sk98lin/skge.c: In function `sk98lin_suspend':
drivers/net/sk98lin/skge.c:755: error: too many arguments to function `pci_save_st
ate'
make[3]: *** [drivers/net/sk98lin/skge.o] Error 1
make[2]: *** [drivers/net/sk98lin] Error 2
make[1]: *** [drivers/net] Error 2
make: *** [drivers] Error 2
#

[Arek75]

Ebuild D/L == 404

--Arek
(james@verge-rpg.com)

[seppe]

Ebuild D/L == 404

--Arek
(james@verge-rpg.com)

Try this one instead. It looks like the ebuild hosted on PickledOnion's site hasn't changed the URL where the patch file is located yet.

I haven't' tested it, because I'm on an evil OS now.

[MaxDamage]

Scroll down some posts to find a mirror.

[rmh3093]

I dont know if this was noted somewhere but the orinoco problem seems to be fixed in 2.6.10-r2, I havent tried .15 modules yet or the kernel patch but the in kernel drivers compiled.

[AgenT]

Ebuild D/L == 404

--Arek
(james@verge-rpg.com)

Mirrors are listed in posts above. They work

UPDATE: BAD ADVICE!! (see next post)

[discomfitor]

Don't use any of the mirrors. Unless the people have updated them recently, they have older versions of the patch. I fixed the url in the ebuild so that it will work now.

Also, I fixed the sk98lin driver so it compiles. Here's the patch if you don't want to remerge everything:

Code: Select all

diff -urN ../../linux/drivers/net/sk98lin/skge.c drivers/net/sk98lin/skge.c
--- ../../linux/drivers/net/sk98lin/skge.c      2004-12-30 03:07:03.000000000 -0500
+++ drivers/net/sk98lin/skge.c  2004-12-30 20:07:56.358200512 -0500
@@ -675,7 +675,7 @@
 
        /* Set the power state to D0 */
        pci_set_power_state(pdev, 0);
-       pci_restore_state(pdev, pAC->PciState);
+       pci_restore_state(pdev);
 
        /* Set the adapter power state to D0 */
        SkPciReadCfgWord(pAC, PCI_PM_CTL_STS, &PmCtlSts);
@@ -752,7 +752,7 @@
 
        pci_enable_wake(pdev, 3, 1);
        pci_enable_wake(pdev, 4, 1);    /* 4 == D3 cold */
-       pci_save_state(pdev, pAC->PciState);
+       pci_save_state(pdev);
        pci_set_power_state(pdev, state); /* set the state */
 
        return 0;

I updated the patch on the site accordingly. Anyone who gets it after this message is posted will get the fixed patch.

EDIT: I also updated the link in the first post to work (see next post).

Just call me Clark.

[AgenT]

The url given in the original post (for ebuild) does not work due to missing forward-slash. Fixed link:

download

[feld]

my sound problem with alsa still exists in your nitro, but is a LOT less severe. It's tolerable now, but audio is still not clear like it should be. just thought u should know.



-Feld

[discomfitor]

Just updated the patch again, adding the new bk3 patch and the ac2 patch. Forgot to mention that I also snuck some other updates from ckdev into the last version with the sk98lin update.

Feld: link me to your kernel config. I have the same card and it works fine, so you must have something else enabled that's causing the problem.