[nitro] _discontinued_

[Tiger683]

2nd nitro-sources-2.6.16 "Big Deal" release series

Disclaimer
This patchset is in no way affiliated with the Gentoo Foundation and so it's
use does not imply any more or less official support from the gentoo developers.
Do NOT, EVER, post bugs regarding a kernel patched to nitro-sources to
_bugs.gentoo.org_. Post your problems in this thread instead.


Announcement:

Beginning with tomorrow, the nitro git repository will go away,
furthermore i will make a normal patch release for historical purposes and to
let ppl still be able to emerge it, until the new sources come along (secrit )


Before you proceed, emerge dev-util/git with useflags "webdav curl" and then emerge cogito..

If you used previous overlay, update it (ebuild AND eclass!) and remove /usr/portage/distfiles/git-src directory before proceeding with emerging this release series !!!!!!!!!

This release fetches a seed tarball from http server and then updates from a git-repository.


What is it?
Custom kernel patchset, based upon ck-patchset from Con Kolivas and
many additional features usually requested by users.

For details on included things go here:
GIT SUMMARY

NEW COMMITS SINCE 2.6.16-nitro-git4 (MOST RECENT FIRST):
*** release 2.6.16-nitro-git12
*** suspend2 fix (take 2)
*** suspend2 fix (take 1)
*** release 2.6.16-nitro-git11
*** update to 2.6.16-ck2
*** update to Suspend2-2.2.2.1
*** release 2.6.16-nitro-git10
*** update base vanilla to 2.6.16.1
*** Another suspend2 vs. gcc4 fix
*** release 2.6.16-nitro-git9
*** Reiser4 - Bring back export of page_cache_readahead as EXPORT_SYMBOL_GPL
(aka. fix reiser4 compilation as module)
*** Add suspend/resume hooks to pata_via and pata_sil680
*** release 2.6.16-nitro-git8
*** new-sched-domain-for-representing-multi-core.patch
*** Reiser4 update
*** release 2.6.16-nitro-git7
*** Fix compilation with CONFIG_SUSPEND2=n
*** release 2.6.16-nitro-git6
*** fix compilation of Suspend2 with gcc4
*** release 2.6.16-nitro-git5
*** Update Suspend2 to 2.2.1.2
** If you see new commits since last time you emerged, you should reemerge to get them.

Howto install

gentoo:
Simply unpack the attached tarball into an overlay,redigest the nitro-sources-2.6.16.ebuild and emerge nitro-sources
Use a separate overlay plz!!!

other distros:
cd /usr/src

mkdir linux-2.6.16-nitro

rm linux

ln -s linux-2.6.16-nitro linux

wget http://nitro-sources.generation.no/seed-1.tar.bz2 --user-agent=Opera
wget http://nitro-sources.generation.no/seed-1.md5sum --user-agent=Opera
md5sum -c seed-1.md5sum

cd linux

tar xvjf ../seed-1.tar.bz2

#this updates to latest git changes:
cg-update

make menuconfig
<compile & install>

And now, even more important part:

nitro_overlay-2.tar.bz2
I recommend to have a separate overlay for this to not mix it with standard ebuilds and to avoid digest problems!

Separate files:
nitro-sources-2.6.16.ebuild
cogito.eclass

SHOWTIME !!!!

cheers

T

PS:

VMWare users probably still need to put these two files into /opt/vmware/workstation/lib/modules/source, but backup the original ones first
vmmon.tar
vmnet.tar

MADWIFI users, use this overlay to install svn-version of madwifi-{tools,driver}, as portage version
seems incompatible
madwifi-tools-9999.ebuild
madwifi-driver-9999.ebuild

[Tiger683]

Just as a quick update:

The eclass now not only downloads a seed for less first-time-traffic, but also checks md5sum for ensuring security,
supports resuming of interrupted seed tarball download and is pretty smart in overall manner

If you for example would like to pass any additional options to the wget command for seed download, you can use the
ECOGITO_WGET_PARM environment variable.

For example to limit the download speed to 40K so that it peacefully runs in background, just do

Code: Select all

ECOGITO_WGET_PARM="--limit-rate=40.0k" emerge nitro-sources

instead of simple "emerge nitro-sources".



PS: Please post all bugs in the eclass/ebuild as well as any improvements for them here.

we also need testing for the non-gentoo release way!!!!!!

cheers

T

[Zephyrus]

I think that there's something strange in a link in your first post, maybe a typographical error.
I read:

Code: Select all

nitro_overlay-2.tar.bz2 
I recommend to have a separate overlay for this to not mix it with standard ebuilds and to avoid digest problems! 

But actually the link is to nitro_overlay-1.tar.bz2, which seems to be the old overlay (at least the name is the same).
The two files have also two different md5sums.

Apart from that, i would like to thank you for your excellent work .
Compiling right now.

[jannis]

emerging worked perfectly, fetched the seed-tarball, updated via git (using the eclass and the ebuild, not the overlay-tbz)
compiling right now on my bleeding-edge-system

*UPDATE*: Compiled without any problem here (gcc-4.1 and suspend2 enabled), 2.6.16-nitro-git6 running without any problem =D

[nesl247]

I get the following error when compiling. No I don't have suspend or suspend2 compiled as a module, nor directly into the kernel.

Code: Select all

kernel/built-in.o: In function `swsusp_save': undefined reference to `suspend_post_context_save'

[Zephyrus]

I get the following error when compiling. No I don't have suspend or suspend2 compiled as a module, nor directly into the kernel.

Code: Select all

kernel/built-in.o: In function `swsusp_save': undefined reference to `suspend_post_context_save'

Same error here.
Are you using an amd64 system?

[nesl247]

No I am not.

Here is my emerge --info if it helps due to CFLAGS or gcc-4.1

Code: Select all

Portage 2.1_pre6-r6 (default-linux/x86/2006.0, gcc-4.1.0, glibc-2.4-r1, 2.6.16-nitro_git3 i686)
=================================================================
System uname: 2.6.16-nitro_git3 i686 AMD Athlon(tm) 64 Processor 3300+
Gentoo Base System version 1.12.0_pre16
distcc 2.18.3 i686-pc-linux-gnu (protocols 1 and 2) (default port 3632) [disabled]
ccache version 2.4 [enabled]
dev-lang/python:     2.4.2-r1
sys-apps/sandbox:    1.2.17
sys-devel/autoconf:  2.13, 2.59-r7
sys-devel/automake:  1.4_p6, 1.5, 1.6.3, 1.7.9-r1, 1.8.5-r3, 1.9.6-r2
sys-devel/binutils:  2.16.91.0.6, 2.16.91.0.7
sys-devel/libtool:   1.5.22
virtual/os-headers:  2.6.11-r3
ACCEPT_KEYWORDS="x86 ~x86"
AUTOCLEAN="yes"
CBUILD="i686-pc-linux-gnu"
CFLAGS="-O2 -march=k8 -pipe -ftracer -fweb -frename-registers -fomit-frame-pointer -mfpmath=sse -fno-ident -freorder-blocks-and-partition"
CHOST="i686-pc-linux-gnu"
CONFIG_PROTECT="/etc /usr/kde/2/share/config /usr/kde/3/share/config /usr/share/config /var/qmail/control"
CONFIG_PROTECT_MASK="/etc/gconf /etc/env.d"
CXXFLAGS="-O2 -march=k8 -pipe -ftracer -fweb -frename-registers -fomit-frame-pointer -mfpmath=sse -fno-ident -freorder-blocks-and-partition -fvisibility-inlines-hidden"
DISTDIR="/usr/portage/distfiles"
FEATURES="autoconfig ccache confcache distlocks metadata-transfer parallel-fetch sandbox sfperms strict"
GENTOO_MIRRORS="http://distfiles.gentoo.org http://distro.ibiblio.org/pub/linux/distributions/gentoo"
LDFLAGS="-Wl,-O1 -Wl,--sort-common -Wl,-Bdirect -Wl,-hashvals -Wl,-zdynsort"
PKGDIR="/usr/portage/packages"
PORTAGE_TMPDIR="/var/tmp"
PORTDIR="/usr/portage"
PORTDIR_OVERLAY="/usr/local/portage /usr/local/overlays/emission-guide /usr/local/overlays/xgl-coffee"
SYNC="rsync://rsync.gentoo.org/gentoo-portage"
USE="x86 3dfx 3dnow X a52 alsa apm avi bash-completion beagle berkdb bitmap-fonts cairo cli crypt ctype cups dba dbus dri dvd eds emboss encode esd expat fam fastbuild firefox flac foomaticdb force-cgi-redirect fortran ftp gd gdbm gif glibc-omitfp glitz gnome gpm gstreamer gtk gtk2 hal imlib inotify ipv6 java jpeg kde libg++ libnotify libsexy libwww mad memlimit mikmod mmext mmx mono motif mp3 mpeg ncurses nls nptl nptlonly nsplugin nvidia ogg opengl oss pam pcre pdflib perl png posix python qt quicktime readline sdl session simplexml soap sockets spell spl sqlite sqlite3 sse sse2 ssl tcpd theora tokenizer truetype truetype-fonts type1-fonts udev vorbis win32codecs xml xsl xv xvid zlib elibc_glibc input_devices_keyboard input_devices_mouse input_devices_evdev kernel_linux userland_GNU video_cards_nvidia"
Unset:  ASFLAGS, CTARGET, EMERGE_DEFAULT_OPTS, LANG, LC_ALL, LINGUAS, MAKEOPTS

[jannis]

suggestion: have you tried enabling suspend2?

*EDIT*: I know, it's not a fix, it's a workaround and I agree with nesl247 that it shouldn't have to be enabled if you don't want it.

[Tiger683]

I corrected the first post, thx zephyrus

About the error: can i have a config? (plz only a link, pasting it here makes things unreadable)

[nesl247]

That works, but you shouldn't be required to have suspend2 if you don't want it.

Edit: Tiger here you go: http://nesl247.pastebin.com/621606 - Now it includes suspend2 to fix it.

[brot]

Code: Select all

  LD      init/built-in.o
  LD      .tmp_vmlinux1
kernel/built-in.o: In function `swsusp_save':
: undefined reference to `suspend_post_context_save'
make: *** [.tmp_vmlinux1] Error 1

Same here, seems that we have to enable suspend2.

(but the ebuild works perfect... as ever )

--- edit:

ok, i will never ever let the answer window open for 1,5 h and then post, i promise...
I synced with git, and am now running git6. Everything compiles, i just have to reboot to see if it works, too

--- edit2:

kernel works, and so does everything like tv-card, vmware, the nvidia driver and my soundcard...
The new release system is really great.

Thanks for your work..


brot

[philicitas]

Same here, will try to compile with enabled suspend2

edit: compiled without errors.

[Zephyrus]

It seems that enabling the power management support (CONFIG_PM=y) triggers the compilation of kernel/power/snapshot.c which contains the reference to suspend_post_context_save function of swsuspend2.
Instead the vanilla snapshot.c doesn't contain suspend_post_context_save, so maybe the nitro file got patched by the swsupend2 patch.

(i'm not an expert, this is only an hypothesis, maybe it's all wrong )

[Tiger683]

Zephyrus: correct.

Everyone: git7 has the fix commited, please update (reemerge for gentoo, cg-update everyone else)

EDIT: i also updated the announcement for non-gentoo users to use faked user agent with wget, as the host
we have the seed on doesn't like wget's user-agent id (securitywise nice feature, but we need this working )
The eclass and ebuild already implement this feature (and a lot more )

[_sys/sid]

Code: Select all

  LD      init/built-in.o
  LD      .tmp_vmlinux1
kernel/built-in.o: In function `swsusp_save':
: undefined reference to `suspend_post_context_save'
make: *** [.tmp_vmlinux1] Error 1

[Tiger683]

Code: Select all

  LD      init/built-in.o
  LD      .tmp_vmlinux1
kernel/built-in.o: In function `swsusp_save':
: undefined reference to `suspend_post_context_save'
make: *** [.tmp_vmlinux1] Error 1

As above post states, i just commited a fix for this.

[Zephyrus]

Zephyrus: correct.

Everyone: git7 has the fix commited, please update (reemerge for gentoo, cg-update everyone else)

EDIT: i also updated the announcement for non-gentoo users to use faked user agent with wget, as the host
we have the seed on doesn't like wget's user-agent id (securitywise nice feature, but we need this working )
The eclass and ebuild already implement this feature (and a lot more )

Nice.
The new version compiles and runs fine.

[rmh3093]

im loving how this -git ebuild is working.... can you throw in sched-new-sched-domain-for-representing-multi-core.patch from -mm sources, it wont hurt nothing

[Tiger683]

im loving how this -git ebuild is working.... can you throw in sched-new-sched-domain-for-representing-multi-core.patch from -mm sources, it wont hurt nothing

Thx

I will take a VERY critical look at the patches in current -mm, a must have will be a reiser4 update, it has some important bugfix,
with that i will also try the patch you said (if it won't hurt con's staircase, I'll throw it in)

EDIT: done, but you better say it works

[buthus]

it works realy great, thanks Tiger!

[jwj]

Thanks Tiger, git7 runs very nice on my notebook.

[dR0PS]

emerging splashutils fails on this kernel, this time it's not the addressing failure known emerging this onto rc3-nitro1.

Code: Select all

  MAKE    libjpeg
In Datei, eingefügt von /usr/lib/klibc/include/stdio.h:9,
                    von jinclude.h:46,
                    von jcapimin.c:20:
/usr/lib/klibc/include/stdarg.h:9:25: stdarg.h: Datei oder Verzeichnis nicht gefunden
In file included from jinclude.h:46,
                 from jcapimin.c:20:
/usr/lib/klibc/include/stdio.h:100: Fehler: Syntaxfehler vor "va_list"
/usr/lib/klibc/include/stdio.h:102: Fehler: Syntaxfehler vor "va_list"
/usr/lib/klibc/include/stdio.h:104: Fehler: Syntaxfehler vor "va_list"
/usr/lib/klibc/include/stdio.h:106: Fehler: Syntaxfehler vor "va_list"
/usr/lib/klibc/include/stdio.h:108: Fehler: Syntaxfehler vor "va_list"
/usr/lib/klibc/include/stdio.h:119: Fehler: Syntaxfehler vor "va_list"
make[1]: *** [jcapimin.o] Error 1
make: *** [jpeglib] Error 2

any hints ?

thx, dR0PS

[Tiger683]

any hints ?

thx, dR0PS

Try reemerging klibc as hinted in the no-sources thread.
The problem is known, a fix is not known.

You can feel free to email spock and tell him to fix his software,
or at least keep it up2date.



T

[dR0PS]

Hi Tiger!

my [*] Marvell Yukon Chipset / SysKonnect SK-98xx net-card (sk98lin)
falled into very slow performance 'state' i got this effect on some kernels in the past
last time i got this problem is some months ago
performance is ~400-500kb/s on 100MB/FullDuplex!

Maybe it has something todo with IRQs shared or address space used, dunno...

I don't know if this has something to do with your flavoured kernel, but it rised up again with this one
Everything was okay with .15-nitro3 and .16-rc3-nitro1

//EDIT: i doublechecked this, falling back to .16-rc3-nitro1 gave me back fullspeed on my network-card, so this must be a git only behave

[orphen]

When complied kernel with "page_cache_readahead", it said that

Code: Select all

*** Warning: "page_cache_readahead" [fs/reiser4/reiser4.ko] undefined!

And

Code: Select all

WARNING: /lib/modules/2.6.16-nitro-git8/kernel/fs/reiser4/reiser4.ko needs unknown symbol page_cache_readahead

in the make modules_install.

Boot with this kernel, it said

Code: Select all

modprobe: FATAL: Error inserting reiser4 (/lib/modules/2.6.16-nitro-git8/kernel/fs/reiser4/reiser4.ko): Unknown symbol in module, or unknown parameter (see dmesg)

And stopped soon.

When complied kernel with NOT "page_cache_readahead", it had the same error above.