2.6.8.1-nitro6 "Say no to no say"

[seppe]

Here is the new nitro sources!

Applied patches (applies on the 2.6.8.1 Linux kernel)

Code: Select all

from_2.6.8.1_to_staircase8.0.bz2 
schedrange.diff 
schedbatch2.4.diff 
schediso2.5.diff 
sched-adjust-p4gain
mapped_watermark.diff 
defaultcfq.diff 
config_hz.diff 
akpm-latency-fix.patch
9000-SuSE-117-writeback-lat.patch 
cddvd-cmdfilter-drop.patch 
cool-spinlocks-i386.diff 
bio_uncopy_user-mem-leak.patch
bio_uncopy_user2.diff
ioport-latency-fix-2.6.8.1.patch 
supermount-ng204.diff.bz2 
fbsplash-0.9-r5-2.6.8-rc3.patch.bz2
make-tree_lock-an-rwlock.patch.bz2 
invalidate_inodes-speedup.patch 
2.6.8.1-mm2-reiser4.diff.bz2 
s8.0_s8.1
mapped_watermark_fix.diff 
sc_mw.diff 
vesafb-tng-0.9-rc4-r2-2.6.8.1.patch
lirc-26.patch
menuconfig-NAME-v1.0.patch
squashfs2.0-patch
gcloop-2.6-20040527.patch
kernel-events-rml-2.6.8-rc2-6.patch
write-barriers.patch
lufs-0.9.7-2.6.0-test9.patch
iteraid_1.44.diff
omnibook-2.6.8-rc2-bk2.diff
config-nr-tty-devices.patch
cdfs-2.6.3a.diff
packet-2.6.8.patch 
orinoco-0.13e-patch.diff
cflags-selection.patch
ipw2100-2.6.8-patch
acx100-2.6.8-rc2-bk2.diff
hid_poll_interval-2.6.0test7.diff
BadRAM-2.6.5.2.patch
acerhk.patch
mki-adapter26_1_3_7.patch
Kernel-Win4Lin3-2.6.8.1.patch

Download
Ebuild
bz2

What's new?
* Updated to CK5, so swap problems should be fixed now
* Win4Lin is back
* 1g Low Mem patch is gone (Win4Lin doesn't like it)
* Acer hotkeys support
* USB Mouse Hack
* vesafb is updated
* BadRam is back

WARNING
As usual, nitro-sources isn't tested very well. It compiles and works on my machine without problems, but that's it. I haven't tested everything (like reiser4 or vesafb-tng for example), so things could always be broken.

Webpage and RSS Feed
A webpage and RSS feed are born!

Final notes
I wanted to add voluntary preemption, but when I finally fixed all rejects (and it were a LOT with this in) and rebooted, I got endless errors flying over my screen. So I guess it's not enough mature yet.
Oh and join #nitro-sources on irc.freenode.net if you want to help us or if you want more info. And feel free to suggest new patches.


Have fun with it, I know I did

[Bols]

Code: Select all

 ______________ 
< 2.6.8-nitro6 >
 -------------- 
        \   ^__^
         \  (oo)\_______
            (__)\       )\/\
                ||----w |
                ||     ||

These nitro-sources rocks... as ever!
Thanks a lot seppe!! You do a great job.

[Robin79]

im compiling it right now I hope it will be as good as usual!

[sobers_2002]

okay just compiled .......seems to be working fine..............just a note

i alwayz get this error from xmms when i close it (in nitro-5,6) that xmms has segfaulted.........though there are no real time problems with that but still i hope if it could be fixed........other than that it simply rocks

PS:- i haven;t been able to rectify my age old problem with text files on reiser4.....whenever i edit/create a file in pico/vim it waits for a while b4 retirning to console, saying writing.... in pico..........i have tried all available kernels so i don;t think it could be a kernel problem but then plz help if nyone knows

[Crisp]

Hi Seppe,

Great job; I did encounter some problems that I am sharing here, perhaps it is of any help for your next release. For the 2.6.8.1-r6 nitro source, I had to run a

Code: Select all

chmod a+x /usr/src/linux/scripts/* -R

before make menuconfig would work. Somehow, the execute permissions in the scripts dir were not all set.

WARNING: Kernel amateur fix-ups ahead; use at your own risk!

To compile the actual kernel, I had to

Code: Select all

ln -s /usr/src/linux/include/asm-i386/asm_offsets.h /usr/src/linux/include/asm/asm_offsets.h

since the include/asm/asm_offset.h file was missing. I also had to patch fs/reiser4/debug.c because it complained about a wrong type:

fs/reiser4/debug.c:48: error: conflicting types for `cond_resched'
include/linux/sched.h error: previous declaration of `cond_resched'
make[2]: *** [fs/reiser4/debug.o] Error 1
make[1]: *** [fs/reiser4] Error 2
make: *** [fs] Error 2

So I edited fs/reiser4/debug.c to read

Code: Select all

/* extern void cond_resched(void); */
extern int cond_resched(void);

at the top of that file. That was not the only fix, I also modified drivers/ide/ide-io.c because of an error

drivers/ide/ide-io.c: In function `__ide_end_request':
drivers/ide/ide-io.c:134: error: `flags' undeclared (first use in this function)
drivers/ide/ide-io.c:134: error: (Each undeclared identifier is reported only once
drivers/ide/ide-io.c:134: error: for each function it appears in.)
make[2]: *** [drivers/ide/ide-io.o] Error 1
make[1]: *** [drivers/ide] Error 2
make: *** [drivers] Error 2

so I added the flags declaration in the __end_request function

Code: Select all

static int __ide_end_request(ide_drive_t *drive, struct request *rq,
                             int uptodate, int nr_sectors)
{
        int ret = 1;
        unsigned long flags;

        BUG_ON(!(rq->flags & REQ_STARTED));
...

Now I am off to try to get framebuffer console support working on my SuperSavage chip. More information such as kernel config is of course available if necessary .

[Robin79]

Works great here no problems so far will see if it stays so!

[Safrax]

Is it possible to back out the framebuffer patches? I don't need it and I cant get the kernel to compile without it.

[butters]

Buried in the threads for the previous couple nitro-sources releases is a patch to allow nitro to compile without vesafb. Worked for me

[butters]

Why does the ebuild insist on replacing the older nitro-sources instead of just adding the new? It never used to behave like this. The --noreplace option doesn't do the trick either. It always wants to upgrade nitro-sources instead of just installing the new sources.

I looked at recent nitro-sources ebuilds, and nothing rang a bell. Any ideas?

[Rainmaker]

Looks good, let's try it. Hope it will resolve some of the issues I've been having with the previous one...

I know there will always be someone whos complaining, but why is 1G out?? Win4Lin sucks anyway

Seriously, isn;t there a way to get them both in? Couldn't you alter the win4lin patch so it doesn't use LowMem from > 0xffff (or whatever the big 896MB limit is)

*edit: Even stranger: why is there still a configure option for lowmem? Is it in or out?

[Rainmaker]

Why does the ebuild insist on replacing the older nitro-sources instead of just adding the new? It never used to behave like this. The --noreplace option doesn't do the trick either. It always wants to upgrade nitro-sources instead of just installing the new sources.

I looked at recent nitro-sources ebuilds, and nothing rang a bell. Any ideas?

It doesn't remove your old sources. It's got to do with slotting, so it indicates the correct action (correct me if I'm wrong)

[phranzee]

lirc_serial does not work. afair it was "invalid arguments" problem.

[Rainmaker]

Hi Seppe,

Great job; I did encounter some problems that I am sharing here, perhaps it is of any help for your next release. For the 2.6.8.1-r6 nitro source, I had to run a

Code: Select all

chmod a+x /usr/src/linux/scripts/* -R

No problem starting xconfig, didn't try menuconfig though... Aren't you trying to run make menuconfig as a normal user, while the directory is owned by root?

before make menuconfig would work. Somehow, the execute permissions in the scripts dir were not all set.

WARNING: Kernel amateur fix-ups ahead; use at your own risk!

To compile the actual kernel, I had to

Code: Select all

ln -s /usr/src/linux/include/asm-i386/asm_offsets.h /usr/src/linux/include/asm/asm_offsets.h

since the include/asm/asm_offset.h file was missing. I also had to patch fs/reiser4/debug.c because it complained about a wrong type:

fs/reiser4/debug.c:48: error: conflicting types for `cond_resched'
include/linux/sched.h error: previous declaration of `cond_resched'
make[2]: *** [fs/reiser4/debug.o] Error 1
make[1]: *** [fs/reiser4] Error 2
make: *** [fs] Error 2

So I edited fs/reiser4/debug.c to read

Code: Select all

/* extern void cond_resched(void); */
extern int cond_resched(void);

Reiser4 compiles cleanly here...

at the top of that file. That was not the only fix, I also modified drivers/ide/ide-io.c because of an error

drivers/ide/ide-io.c: In function `__ide_end_request':
drivers/ide/ide-io.c:134: error: `flags' undeclared (first use in this function)
drivers/ide/ide-io.c:134: error: (Each undeclared identifier is reported only once
drivers/ide/ide-io.c:134: error: for each function it appears in.)
make[2]: *** [drivers/ide/ide-io.o] Error 1
make[1]: *** [drivers/ide] Error 2
make: *** [drivers] Error 2

so I added the flags declaration in the __end_request function

Code: Select all

static int __ide_end_request(ide_drive_t *drive, struct request *rq,
                             int uptodate, int nr_sectors)
{
        int ret = 1;
        unsigned long flags;

        BUG_ON(!(rq->flags & REQ_STARTED));
...

Now I am off to try to get framebuffer console support working on my SuperSavage chip. More information such as kernel config is of course available if necessary .

I'm not done compiling yet, but didn't see this one pass by yet (I saw reiser4 getting compiled)

[Wi1d]

Thanks and great work Seppe. My machine seems to like nitro6 quite well.

[MrDooM]

Is there a Bootsplash Patch like in the older Love Sources too?
I started to like the nitro sources, but i miss my nice booting screen.


Greetz MrDooM

[kepreon]

Is it just me or does it show 1GB Low Memory Support six times?

Linux Kernel v2.6.8-nitro6 "Say no to no say" Configuration
──────────────────────────────────────────────────────────────────────────────
┌────────────────────── Processor type and features ──────────────────────┐
│ Arrow keys navigate the menu. <Enter> selects submenus --->. │
│ Highlighted letters are hotkeys. Pressing <Y> includes, <N> excludes, │
│ <M> modularizes features. Press <Esc><Esc> to exit, <?> for Help. │
│ Legend: [*] built-in [ ] excluded <M> module < > module capable │
│ ┌^(-)─────────────────────────────────────────────────────────────────┐ │
│ │[*] 1Gb Low Memory Support │ │
│ │[*] Work around bad spots in RAM (NEW) │ │
│ │[*] 1Gb Low Memory Support │ │
│ │[*] 1Gb Low Memory Support │ │
│ │[*] 1Gb Low Memory Support │ │
│ │[*] 1Gb Low Memory Support │ │
│ │[*] 1Gb Low Memory Support │ │
│ │[ ] Math emulation │ │
│ │[*] MTRR (Memory Type Range Register) support │ │
│ │[ ] Boot from EFI support (EXPERIMENTAL)

[Boohbah]

Is there a Bootsplash Patch like in the older Love Sources too?
I started to like the nitro sources, but i miss my nice booting screen.


Greetz MrDooM

Check out gensplash:
http://dev.gentoo.org/~spock/

[Crisp]

Reiser4 compiles cleanly here...
...
I'm not done compiling yet, but didn't see this one pass by yet (I saw reiser4 getting compiled)

I have the impression that it has something to do with an experimental version that I unpacked yesterday (before the gentoo ebuild was released), or that my laptop configuration/kernel config is peculiar enough to let some modules collide ...

Anyway, it works smoothly now, if I have some problems with the next nitro release, I'll delve in and let you guys know.

[Wi1d]

Is it just me or does it show 1GB Low Memory Support six times?

I hadn't noticed before but I've got it 6x too.

[zerojay]

I don't know what the problem is, but for some reason, my box has been completely allergic to all nitro and ck patched kernels. Each and every single one of them gives me a kernel oops whenever I run any emerge commands whatsoever, even unprivileged users. I'm hoping that maybe someone might be able to help track down and fix this problem.

Here's the kernel oops, which I had to write down so ignore any capitalization/formatting errors.

Code: Select all

Oops: 0002 [#1]
Modules linked in:
CPU:          0
EIP:    0060:[<c01131a8>]        Not tainted
EFLAGS: 00010082         (2.6.8.1-nitro6)
EIP is at enqueue_task+0x28/0x40
eax:80000079  ebx:c8204910   ecx: c04f3948   edx:c04f3540
esi:c04f3948   edi:00000001   ebp:c6309f5c    esp:c6309f54
ds:007b   es:007b   ss:0068
Process emerge (pid:12385, threadinfo=c6308000 task=c82048f0)
Stack: c82048f0 00000000 c6309f70 c01139c9 c6309fc4 20000001 00000000 00000000 c011d109 c6309fc4 c010a615 c0415548 20000001 00000000 c0106f70 00000000 00000000 c04baa00 c0415548 c010722d 00000000 c6309fc4 c0415548 4a76eddc
Call Trace:
  [<c01139c9>] scheduler_tick+0x169/0x270
  [<c011d109>] do_timer+0x29/0xd0
  [<c010a615>] timer_interrupt+0x25/0x110
  [<c0106f70>] handle_IRQ_event+0x30/0x60
  [<c010722d>] do_IRQ+0x6d/0xe0
  [<c01059ac>] common_interrupt+0x18/0x20
Code: 0f a6 42 2c 8b 1c 24 8b 74 24 04 89 ec 5d c3 89 f6 8d bc 27
 <0>Kernel panic: Fatal exception in interrupt
In interrupt handler - not syncing

Here's my .config file.
http://darkstalker.newcon.org/.config

I guess it's interrupt or scheduler related. Any ideas of what I should do here? Maybe Con might know something?

By the way, I have the multiple 1GB RAM entries as well.

[kepreon]

The kernel finished compiling, which it did cleanly, and it seems to work fine upon first boot, no errors!

kevin@kepreon kevin $ uname -r
2.6.8-nitro6

I just tried nitro5 a few days ago and after running cpuburn and timing an emerge on both nitro5 and gentoo-dev-sources-r1, the emerge was much more responsive on nitro5 and finished in about half the time as the gentoo dev sources. After trying the vanilla 2.6 sources, gentoo's 2.6 sources, and the latest mm-sources, I liked nitro the best. I'm a convert

[G3n2]

omfg 500hz patch owns ik hou van je seppe , je't aime !! \o/

[G3n2]

when doing :

ebuild nitro-sources-bla install , it looks for the file in :

/usr/portage/usr/nitro-sources/ instead of in /usr/portage/nitro-sources/

a fix would be great

i had to mkdir /usr/portage/usr/nitro-sources/ and copy the ebuild over there

[d0nju4n]

yay! I check the forums every day for new nitro sources. Grabbing them now!

[G3n2]

Code: Select all

 ______________ 
< 2.6.8-nitro6 >
 -------------- 
        \   ^__^
         \  (oo)\_______
            (__)\       )\/\
                ||----w |
                ||     ||

These nitro-sources rocks... as ever!
Thanks a lot seppe!! You do a great job.

is there a way to do the cow or whatever on bootup ?
i know its possible , can u tell me how Bols ?

i mean at the prompt , when everything's done loading , it shows the nice 1337 cow \o/ , then the login prompt

thx for any tips u can give me

-G3n2