Old posts from the morph-sources thread

[xeno666]

Being generally lazy I still hadn't tried patching by hand and again I'm ahead doing so: the newest morph works without a hitch - yardbird, whatever have you done it fixed the weird behavior - many thanks and keep up the good job

[ahorn]

reiser4 fix for umount freezes.. /fs/reiser4/plugins/object.c

Code: Select all

--- object.c~   2005-03-23 12:30:52.000000000 +0300 
 +++ object.c    2005-03-23 13:31:39.000000000 +0300 
 @@ -976,6 +976,7 @@ 
          * FIXME: this resembles generic_delete_inode 
          */ 
         list_del_init(&object->i_list); 
 +       list_del_init(&object->i_sb_list); 
         object->i_state |= I_FREEING; 
         inodes_stat.nr_inodes--; 
         spin_unlock(&inode_lock);

edit: ftp://ftp.namesys.com/pub/reiser4-for-2 ... -fix.patch

[yardbird]

reiser4 fix for umount freezes.. /fs/reiser4/plugins/object.c

Code: Select all

--- object.c~   2005-03-23 12:30:52.000000000 +0300 
 +++ object.c    2005-03-23 13:31:39.000000000 +0300 
 @@ -976,6 +976,7 @@ 
          * FIXME: this resembles generic_delete_inode 
          */ 
         list_del_init(&object->i_list); 
 +       list_del_init(&object->i_sb_list); 
         object->i_state |= I_FREEING; 
         inodes_stat.nr_inodes--; 
         spin_unlock(&inode_lock);

edit: ftp://ftp.namesys.com/pub/reiser4-for-2 ... -fix.patch

Thanks for the link ahorn, I updated the website.

[yardbird]

Hi folks,

a new release is out. It includes reiser4, cfq-ts and general bugfixes. Update from previous versions is recommended. Let me know how it goes

Cheers.

[mayday147]

2 cents question: Which sched is supposed to be faster? nicksched or staircase ?

[Illissius]

2 cents question: Which sched is supposed to be faster? nicksched or staircase ?

both

[ahorn]

damn namsys guys waiting for morph6 ..

[allex87]

For some weird reason I cannot seem to be able to emerge morph sources. I get the following error:

Code: Select all

idaho morph-sources # emerge morph-sources
Calculating dependencies ...done!
>>> emerge (1 of 1) sys-kernel/morph-sources-2.6.11-r5 to /

!!! Digest verification Failed:
!!!    /usr/portage/distfiles/linux-2.6.11.tar.bz2
!!! Reason: Filesize does not match recorded size

Why is that? I tried to re-download linux-2.6.11.tar.bz2 several times but I still get the error.

Anyone?


Thanks,
Alex.

[yardbird]

2 cents question: Which sched is supposed to be faster? nicksched or staircase ?

Personally I'd go with staircase, but this is rather a personal choice and is a very ... flammable ... topic I say staircase because IMHO it is much more supported (there is the mailing list of -ck kernels where you can ask for help) and it has faster development cycles. Because of this, it is more tested and tweaked than nicksched, and from my non-scientific benchmark (watching a movie while emerging a kernel) is behaves better interactivity-wise. That said, probably you'd better try both and decide by yourself

allex87: it seems like your kernel tarball is corrupted. Try to re-download it.

EDIT: sorry allex87, I did not read your message entirely You should try to erase all the files in /usr/local/portage/sys-kernel/morph-sources/, re-copy there the ebuild and recreate the digest. Also before doing that remove the kernel tarball in /usr/portage/distfiles, so that you are sure the right tarball is downloaded.

[mayday147]

yarbird: 10x for your answer . I'll try both of them.

allex87:

For some weird reason I cannot seem to be able to emerge morph sources. I get the following error:

Code: Select all

idaho morph-sources # emerge morph-sources
Calculating dependencies ...done!
>>> emerge (1 of 1) sys-kernel/morph-sources-2.6.11-r5 to /

!!! Digest verification Failed:
!!!    /usr/portage/distfiles/linux-2.6.11.tar.bz2
!!! Reason: Filesize does not match recorded size

Why is that? I tried to re-download linux-2.6.11.tar.bz2 several times but I still get the error.

Anyone?


Thanks,
Alex.

Try digesting the sources:

Code: Select all

emerge morph-sources  --digest

[allex87]

yarbird: 10x for your answer . I'll try both of them.

allex87:

For some weird reason I cannot seem to be able to emerge morph sources. I get the following error:

Code: Select all

idaho morph-sources # emerge morph-sources
Calculating dependencies ...done!
>>> emerge (1 of 1) sys-kernel/morph-sources-2.6.11-r5 to /

!!! Digest verification Failed:
!!!    /usr/portage/distfiles/linux-2.6.11.tar.bz2
!!! Reason: Filesize does not match recorded size

Why is that? I tried to re-download linux-2.6.11.tar.bz2 several times but I still get the error.

Anyone?


Thanks,
Alex.

Try digesting the sources:

Code: Select all

emerge morph-sources  --digest

Thank you all for the replies.

I did to the digest for it. As one post up suggested, I removed the kernel tarrball and re-created the digest.

That did the trick

Morph-sources emerging right now, I'll post here how things go



Alex.

[mayday147]

yarbird: Can you please update your morph-overlay to the latest sysvinit and kexec-tools?

[fallow]

2 cents question: Which sched is supposed to be faster? nicksched or staircase ?

cpu scheduler doesnt provide a speed. this is provided by your CPU for ex and via nice level for a not so good example. yeah somebody can said that indirectly maybe , but which scheduler is faster ? , this sounds like which scheduler code is faster but this is making a less of sense , You can write scheduler in a pure asm if Your are good coder but this is not a point

Once again - Is no any magic in this

cheers.

[Roamer]

I cannot get ipw2200 to start with morph5. It worked flawlessly with morph2 (which is my current kernelbuild).
I copied the .config from morph2 to morph5 and recompiled.
Then reemerge ipw2200.
That compiled cleanly (atleast there is no errormessages)
But when I modprobe the newly built ipw2200 module I get a segmentation fault.

Any clues as to where I can start to look for the culprit?
I am by no means an expert to these things, but I have some experience with C and am not scared off by applying patches or manually editing patches to get them applied cleanly (If I was I would probably not be messing with morph )


Great Patchset by the way!

--
- Roamer

[Roamer]

Nevermind!!

This is obiously a known bug with the 2.6.11.1 and above kernels.
(https://sourceforge.net/tracker/index.p ... tid=650334)

Ah well.....

--
- Roamer

[bluemx]

Merged from Gentoo on AMD64. --Maedhros

Code: Select all

# Copyright 1999-2004 Gentoo Foundation 
# Distributed under the terms of the GNU General Public License v2 
# $Header: $ 

UNIPATCH_LIST="${DISTDIR}/patch-${KV_FULL}.bz2" 
K_PREPATCHED="yes" 
UNIPATCH_STRICTORDER="yes" 

NR=$(echo ${PR} | cut -c 2-) 
SC_VER="10.6" 
#XBOX_VER="20050102" 
ORINOCO_VER="0.13e-SN-6" 
HOSTAP_VER="0.3.7" 
MH_VER="2" 
MPPE_VER="1.3" 
V4L_VER="2" 

K_NOUSENAME="yes" 
ETYPE="sources" 
inherit kernel-2 eutils 
detect_version 
IUSE="hostap orinoco touchpad-fixes sk98 staircase mppe-mppc updated-bluetooth nicksched updated-v4l" 

DESCRIPTION="Full sources for the Stock Linux kernel plus various enhancements for desktop and laptop users." 
HOMEPAGE="http://morph-sources.sf.net" 
SRC_URI="${KERNEL_URI} 
    mirror://sourceforge/${PN}/patch-${PV}-morph${NR}.bz2 
    mirror://sourceforge/${PN}/extra-patches-${PV}-morph${NR}.tar.bz2 
    http://osdn.dl.sourceforge.net/sourceforge/morph-sources/patch-${PV}-morph${NR}.bz2 
    http://osdn.dl.sourceforge.net/sourceforge/morph-sources/extra-patches-${PV}-morph${NR}.tar.bz2" 

KEYWORDS="~x86 ~amd64" 

src_compile() { 

    if use staircase && use nicksched 
    then 
        eerror "Cannot use both staircase AND nicksched, please choose just one." 
        eerror "Aborting." 
        die 
    fi 

    cd ${WORKDIR} 
    tar xjvf ${DISTDIR}/extra-patches-${PV}-morph${NR}.tar.bz2 
    cd ${S} 

    EPATCH_OPTS="-p1  --no-backup-if-mismatch" 

    if use staircase 
    then 
        epatch ../extra-patches/${PV}-staircase-${SC_VER}.diff 
    fi 
    if use nicksched 
    then 
        epatch ../extra-patches/${PV}-nicksched.patch 
    fi 
    if use mppe-mppc 
    then 
        epatch ../extra-patches/linux-${PV}-mppe-mppc-${MPPE_VER}.patch 
    fi 
    #if use xbox 
    #then 
    #    epatch ../extra-patches/xbox-sources-${PV}-${XBOX_VER}-morph${NR}.patch 
    #fi 
    if use updated-bluetooth 
    then 
        epatch ../extra-patches/patch-${PV}-mh${MH_VER}.patch 
    fi 
    if use updated-v4l 
    then 
        epatch ../extra-patches/v4l-${PV}-${V4L_VER}.diff 
    fi     
    if use touchpad-fixes 
    then 
        epatch ../extra-patches/touchpad-fixes-${PV}.diff 
    fi 
    if use sk98 
    then 
        epatch ../extra-patches/sk98-${PV}.patch 
    fi 
    if use orinoco 
        then 
                epatch ../extra-patches/orinoco-${ORINOCO_VER}.patch 
        fi 
    if use hostap 
            then 
                epatch ../extra-patches/hostap-${HOSTAP_VER}.patch 
        fi 
} 

pkg_postinst() { 
    postinst_sources 

        #if use xbox 
        #then 
        #        ewarn "" 
        #        ewarn "WARNING: The FATX driver is currently horribly broken. Writing to a FATX partition with this kernel will corrupt it." 
        #        ewarn "" 
        #fi 
     
    if use mppe-mppc 
    then 
        ewarn "" 
        ewarn "PLEASE NOTE:" 
        ewarn "The mppe-mppc patch contains a patented algorithm. To avoid legal problems US citizens" 
        ewarn "should NOT use this patch. You have been warned. More details here:" 
        ewarn "http://www.polbox.com/h/hs001/" 
    fi 
     
    ewarn "" 
    ewarn "IMPORTANT:" 
    ewarn "This is a experimental kernel version, I'm not responsible for breaking your system" 
    ewarn "ALWAYS keep a second stable and bootable kernel apart in your boot manager" 
    ewarn "" 
     
} 
 

how to used this ebuild

[yardbird]

Hi,

2.6.11-morph6 is out. It sports an updated reiser4 patch and some fixes for the touchpad fixes introduced in the previous release Also -ac patch was updated to -ac6. This should behave quite well, so try hard to break it Head to the website for download.

bluemx: This link ha nice info on how to setup 3rd party ebuilds and portage overlays. It should answer your question.

[_sys/sid]

Framebuffer does not work!!!

Code: Select all

Graphics support  --->
	[*] Support for frame buffer devices
	...
	<*>   VGA 16-color graphics support
	<*>   VESA VGA graphics support
		VESA driver type (vesafb)  --->
	...
	[*] Support for the framebuffer splash 

Code: Select all

# Gentoo & MORPH Sources
title GNU/Linux Gentoo 1.4.16 (Kernel 2.6.11-morph6)
kernel (hd0,2)/kernel-2.6.11-morph6 ro root=/dev/hda6 video=vesafb:ywrap,pmipal,1280x1024-16@60 splash=verbose,theme:livecd-2005.0 vga=794
initrd (hd0,2)/initrd-2.6.11

emerge splashutils (0.9.1)

Code: Select all

  CC      splash.o
splash.c: In function `main':
splash.c:218: error: `FBIOSPLASH_SETMODE' undeclared (first use in this function)
splash.c:218: error: (Each undeclared identifier is reported only once
splash.c:218: error: for each function it appears in.)
splash.c:218: error: `FB_SPLASH_MODE_SILENT' undeclared (first use in this function)
splash.c:218: error: `FB_SPLASH_MODE_VERBOSE' undeclared (first use in this function)
splash.c:239: error: `FBIOSPLASH_GETMODE' undeclared (first use in this function)
make: *** [splash.o] Error 1

[ahorn]

splashutils 1.1.9.3 ? if not, you need them. unmask them.

[_sys/sid]

What I have to do for unmask splashutils pkg ?
I've Removed line from /usr/portage/profiles/package.mask but I don't like this method...

tnks

[ahorn]

put >=splashutils-0.9.1 in /etc/portage/package.unmask and don't forget to set the keyword.

[_sys/sid]

With splashutils-1.1.9.3 the framebuffer works
And the system was faster with nicksched

Tnks

[jubo]

Is there a place where all of the USE flags and their effects are explained for morph-sources?

[ahorn]

Is there a place where all of the USE flags and their effects are explained for morph-sources?

Look to the morph webpage and check the news. some use-flags are hidden in older news, of course.
You got a good idea, yardbird should do a use-flag explanation list.

[HardenCoonor]

You got a good idea, yardbird should do a use-flag explanation list.

Probably. It is maybe just a collecting of information from this thread, or the changelog posted on the 1st page.

While you are at it(or not), yardbird, i would like to ask whether to rethink the way you release your files on sourceforge. I think that the ebuild-file and the main patch might be put into one archive, maybe including the digest. It is just, in my case for example, i always have to bother sourceforge for a 1kb-file, and have to click through certain dialogs to get it, besides getting the patch and the extra-patch.
Since you add certain feature/use-flags, i cannot recycle the older ebuild. I could probably, but i only know this afterwards.


Besides that, i am still satisfied with morph-sources, that reiser4-error in morph4 made me switch to ext3 for /, and any kernel-related crash that showed up since that in my logs was artsd-related. Good reasons fro me to stick with morph.