Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
2.6.7_rc3-love1 aka 'Feels like love'
View unanswered posts
View posts from last 24 hours

Goto page Previous  1, 2, 3, 4  Next  
Reply to topic    Gentoo Forums Forum Index Unsupported Software
View previous topic :: View next topic  
Author Message
HomerSimpson
l33t
l33t


Joined: 25 Jan 2003
Posts: 869
Location: Ohio, USA

PostPosted: Sat Jun 12, 2004 4:10 pm    Post subject: Reply with quote

When I try to build love-rc3 I get this error:
Code:
make[1]: `arch/i386/kernel/asm-offsets.s' is up to date.
  CHK     include/asm-i386/asm_offsets.h
  CHK     include/linux/compile.h
  CC      fs/reiser4/txnmgr.o
fs/reiser4/txnmgr.c: In function `replace_page_in_mapping':
fs/reiser4/txnmgr.c:3727: error: structure has no member named `page_lock'
fs/reiser4/txnmgr.c:3740: warning: implicit declaration of function `___add_to_page_cache'
fs/reiser4/txnmgr.c:3742: error: structure has no member named `page_lock'
fs/reiser4/txnmgr.c: In function `swap_jnode_pages':
fs/reiser4/txnmgr.c:3772: error: structure has no member named `page_lock'
fs/reiser4/txnmgr.c:3784: error: structure has no member named `page_lock'
fs/reiser4/txnmgr.c: In function `create_copy_and_replace':
fs/reiser4/txnmgr.c:4143: warning: implicit declaration of function `pte_chain_lock'
fs/reiser4/txnmgr.c:4146: warning: implicit declaration of function `try_to_unmap'
fs/reiser4/txnmgr.c:4147: error: `SWAP_AGAIN' undeclared (first use in this function)
fs/reiser4/txnmgr.c:4147: error: (Each undeclared identifier is reported only once
fs/reiser4/txnmgr.c:4147: error: for each function it appears in.)
fs/reiser4/txnmgr.c:4150: error: `SWAP_FAIL' undeclared (first use in this function)
fs/reiser4/txnmgr.c:4158: warning: implicit declaration of function `pte_chain_unlock'
make[2]: *** [fs/reiser4/txnmgr.o] Error 1
make[1]: *** [fs/reiser4] Error 2
make: *** [fs] Error 2

I tried patching the kernel with
Code:
bzcat reiser4-2.6.7-rc3-love1-2004.06.11.bz2 | patch -p1 in /usr/src/linux
as in OneOfOne's post but I get
Code:
patch: **** read error : Is a directory
I am new to patching but I guess I understand why it gives me this error. /usr/src/linux is a directory and not a file to patch. The patch itself notes the file but it is not in my directory tree.

I apologize for asking such a noob question but what is the verbose syntax I need to use to patch the kernel? I am hoping this will get rid of my build error.

Thx for the help and thx for the Love-sources
_________________
The strong must protect the Sweet.
Back to top
View user's profile Send private message
freelight
Apprentice
Apprentice


Joined: 12 Jun 2004
Posts: 295
Location: NYC, NY, USA

PostPosted: Sat Jun 12, 2004 4:33 pm    Post subject: Reply with quote

HomerSimpson wrote:

I tried patching the kernel with
Code:
bzcat reiser4-2.6.7-rc3-love1-2004.06.11.bz2 | patch -p1 in /usr/src/linux
as in OneOfOne's post but I get
Code:
patch: **** read error : Is a directory
I am new to patching but I guess I understand why it gives me this error. /usr/src/linux is a directory and not a file to patch. The patch itself notes the file but it is not in my directory tree.

I made this mistake too before I read the line again. Here are the commands, verbatum (assuming your love-sources are in /usr/src/linux, and you've already put the patch file there):
Code:
cd /usr/src/linux
bzcat reiser4-2.6.7-rc3-love1-2004.06.11.bz2 | patch -p1


ps: just rebooted with this reiser4 patch, and it works great! (My entire system, except /boot, is reiser4.)
Back to top
View user's profile Send private message
HomerSimpson
l33t
l33t


Joined: 25 Jan 2003
Posts: 869
Location: Ohio, USA

PostPosted: Sat Jun 12, 2004 6:52 pm    Post subject: Reply with quote

Thx :oops: Now my kernel is patched but I still get this error when I build:
Code:
fs/reiser4/txnmgr.c: In function `replace_page_in_mapping':
fs/reiser4/txnmgr.c:3727: error: structure has no member named `page_lock'
fs/reiser4/txnmgr.c:3740: warning: implicit declaration of function `___add_to_page_cache'
fs/reiser4/txnmgr.c:3742: error: structure has no member named `page_lock'
fs/reiser4/txnmgr.c: In function `swap_jnode_pages':
fs/reiser4/txnmgr.c:3772: error: structure has no member named `page_lock'
fs/reiser4/txnmgr.c:3784: error: structure has no member named `page_lock'
fs/reiser4/txnmgr.c: In function `create_copy_and_replace':
fs/reiser4/txnmgr.c:4143: warning: implicit declaration of function `pte_chain_lock'
fs/reiser4/txnmgr.c:4146: warning: implicit declaration of function `try_to_unmap'
fs/reiser4/txnmgr.c:4147: error: `SWAP_AGAIN' undeclared (first use in this function)
fs/reiser4/txnmgr.c:4147: error: (Each undeclared identifier is reported only once
fs/reiser4/txnmgr.c:4147: error: for each function it appears in.)
fs/reiser4/txnmgr.c:4150: error: `SWAP_FAIL' undeclared (first use in this function)
fs/reiser4/txnmgr.c:4158: warning: implicit declaration of function `pte_chain_unlock'
make[2]: *** [fs/reiser4/txnmgr.o] Error 1
make[1]: *** [fs/reiser4] Error 2
make: *** [fs] Error 2
Any ideas on how to fix this?

I am building now without reiserfs4.

Thx
_________________
The strong must protect the Sweet.
Back to top
View user's profile Send private message
Lowspirit
Apprentice
Apprentice


Joined: 31 Jul 2002
Posts: 258
Location: Northern Sweden

PostPosted: Sat Jun 12, 2004 10:58 pm    Post subject: Reply with quote

Works like a charm *pets new favorite kernel*

Could be the renicing tip though, cause the removal of the temporary freezes really makes a difference, don't know about the rest.
_________________
Gentoo | AMD X2 3800+ 2GB RAM | Kernel 2.6.30 . ReiserFS . CFQ . GCC4.3.3 | Firefox 3.5 | Gnome 2.26 w/ Compiz-Fusion
"Penguins are the only fish that can fly"
Back to top
View user's profile Send private message
luqas
Guru
Guru


Joined: 02 Jan 2004
Posts: 588
Location: /US/Texas/Beaumont

PostPosted: Sat Jun 12, 2004 11:41 pm    Post subject: Reply with quote

Question OoO: Since I have heard a couple people see performance using the mm scheduler instead of Nick's, are you going to put Nick's back in?
Back to top
View user's profile Send private message
LinuxRocks
Guru
Guru


Joined: 27 Nov 2003
Posts: 397
Location: New Mexico

PostPosted: Sun Jun 13, 2004 3:14 am    Post subject: Reply with quote

Interesting. I emerged this new love sources, and when I boot, it locks up on some messages that says BOOTING 2/15 processors. I only have 2 procs and SMP is enabled in the config file. Im using a .config from my Love 4 build that works great.

Anyone else seeing this?

Also, if you need more info, let me know.

Thanks!!!

Joe
Back to top
View user's profile Send private message
freelight
Apprentice
Apprentice


Joined: 12 Jun 2004
Posts: 295
Location: NYC, NY, USA

PostPosted: Sun Jun 13, 2004 3:27 am    Post subject: Reply with quote

When you enable SMP, you get to choose how many CPUs you want support for. If you only have 2, make sure you set it to 2.
Back to top
View user's profile Send private message
LinuxRocks
Guru
Guru


Joined: 27 Nov 2003
Posts: 397
Location: New Mexico

PostPosted: Sun Jun 13, 2004 3:30 am    Post subject: Reply with quote

ok.. Im makeing that change and recompileing... Before, I left it to the default setting of 8.

Thanks!!!

Joe
Back to top
View user's profile Send private message
OneOfOne
Guru
Guru


Joined: 28 May 2003
Posts: 368

PostPosted: Sun Jun 13, 2004 4:38 am    Post subject: Reply with quote

HomerSimpson wrote:
Thx :oops: Now my kernel is patched but I still get this error when I build:
Code:
fs/reiser4/txnmgr.c: In function `replace_page_in_mapping':
fs/reiser4/txnmgr.c:3727: error: structure has no member named `page_lock'
fs/reiser4/txnmgr.c:3740: warning: implicit declaration of function `___add_to_page_cache'
fs/reiser4/txnmgr.c:3742: error: structure has no member named `page_lock'
fs/reiser4/txnmgr.c: In function `swap_jnode_pages':
fs/reiser4/txnmgr.c:3772: error: structure has no member named `page_lock'
fs/reiser4/txnmgr.c:3784: error: structure has no member named `page_lock'
fs/reiser4/txnmgr.c: In function `create_copy_and_replace':
fs/reiser4/txnmgr.c:4143: warning: implicit declaration of function `pte_chain_lock'
fs/reiser4/txnmgr.c:4146: warning: implicit declaration of function `try_to_unmap'
fs/reiser4/txnmgr.c:4147: error: `SWAP_AGAIN' undeclared (first use in this function)
fs/reiser4/txnmgr.c:4147: error: (Each undeclared identifier is reported only once
fs/reiser4/txnmgr.c:4147: error: for each function it appears in.)
fs/reiser4/txnmgr.c:4150: error: `SWAP_FAIL' undeclared (first use in this function)
fs/reiser4/txnmgr.c:4158: warning: implicit declaration of function `pte_chain_unlock'
make[2]: *** [fs/reiser4/txnmgr.o] Error 1
make[1]: *** [fs/reiser4] Error 2
make: *** [fs] Error 2
Any ideas on how to fix this?

I am building now without reiserfs4.

Thx

well looks like you enabled too much stuff ;)
as a general rule, dont enable anything under reiser4 except "large key support".

about the renicing tip, I suppose it works to an extent, but the key thing now is dont try to "over" stress it (compiling stuff while watch movies is ok, while compiling 4 things while playing ut2k4 and running sati is not ok :P)
but atleast i can upgrade my glibc now :D
peace
Back to top
View user's profile Send private message
OneOfOne
Guru
Guru


Joined: 28 May 2003
Posts: 368

PostPosted: Sun Jun 13, 2004 4:53 am    Post subject: Reply with quote

Dryre wrote:
Question OoO: Since I have heard a couple people see performance using the mm scheduler instead of Nick's, are you going to put Nick's back in?

well sure, when a new version comes out.
but in 2.6.7 so far default scheduler runs better than nick's.

peace
Back to top
View user's profile Send private message
Peacefaker
Tux's lil' helper
Tux's lil' helper


Joined: 07 Jan 2004
Posts: 117
Location: Vaasa, Finland

PostPosted: Sun Jun 13, 2004 12:46 pm    Post subject: Reply with quote

Sorry, I've been away for a couple of days but here it is. Nvidia still goes boom and
Code:

bash-2.05b# cat .config | grep -i regparm
# CONFIG_REGPARM is not set
bash-2.05b# cat .config | grep -i 4k
# CONFIG_4KSTACKS is not set


If that was what OneOfOne meant in his post. However I'm happy with the current kernel and can wait this one out.
Back to top
View user's profile Send private message
()
l33t
l33t


Joined: 25 Nov 2002
Posts: 610

PostPosted: Sun Jun 13, 2004 2:06 pm    Post subject: Reply with quote

The last nvidia-kernel in portage works for me.
_________________
to be concerned is good
Back to top
View user's profile Send private message
Gentree
Watchman
Watchman


Joined: 01 Jul 2003
Posts: 5350
Location: France, Old Europe

PostPosted: Sun Jun 13, 2004 3:02 pm    Post subject: Reply with quote

What is regarded as "heavy use" as far a the r4 pbs are concerned.

I have /portage as r4 and have just compiled xorg-x11 with no trouble on 2.6.6-love4 , would this be the kind of use likely to cause a lockup?

If so the comparison may be helpful / enlightening.

It's nice to have the "nice -15" work-around but does seem to indicate a serious flaw somewhere in the coding if the disk I/O has to be massages like that to prevent it locking up. A bit worrying. :?
Back to top
View user's profile Send private message
Jowilly
Guru
Guru


Joined: 22 Jun 2002
Posts: 319

PostPosted: Sun Jun 13, 2004 3:54 pm    Post subject: Reply with quote

Please put supermount back in.

I have now updated everything to submount as it seems the way to go. It mounts my cd properly, but cannot unmount it.

I cannot even unmount it from the shell, it says the device is busy.
eject doesn't work either.

The only way I have to eject the damn cd is to reboot.

It's nice to try beta things, but there's no reason to remove stable stuff.

Please put supermount back in, it was working perfectly.

Thanks.
Back to top
View user's profile Send private message
OneOfOne
Guru
Guru


Joined: 28 May 2003
Posts: 368

PostPosted: Sun Jun 13, 2004 4:32 pm    Post subject: Reply with quote

Gentree wrote:
What is regarded as "heavy use" as far a the r4 pbs are concerned.

I have /portage as r4 and have just compiled xorg-x11 with no trouble on 2.6.6-love4 , would this be the kind of use likely to cause a lockup?

If so the comparison may be helpful / enlightening.

It's nice to have the "nice -15" work-around but does seem to indicate a serious flaw somewhere in the coding if the disk I/O has to be massages like that to prevent it locking up. A bit worrying. :?

heavy usage here = high cpu load + high reiser4 fs usage (mainly /var/tmp/portage)
also the lock up under heavy load is rather a new bug so I'm looking into it with Zam.
so if you have /var/tmp/portage on a sperate not-reiser4 partition you're safe.

@Jowilly try ivman please, *mount will be removed very soon., ivman pretty good. if you need help with it, join us on irc.

peace
Back to top
View user's profile Send private message
HomerSimpson
l33t
l33t


Joined: 25 Jan 2003
Posts: 869
Location: Ohio, USA

PostPosted: Sun Jun 13, 2004 5:01 pm    Post subject: Reply with quote

OneOfOne wrote:
HomerSimpson wrote:
Thx :oops: Now my kernel is patched but I still get this error when I build:
Code:
fs/reiser4/txnmgr.c: In function `replace_page_in_mapping':
fs/reiser4/txnmgr.c:3727: error: structure has no member named `page_lock'
fs/reiser4/txnmgr.c:3740: warning: implicit declaration of function `___add_to_page_cache'
fs/reiser4/txnmgr.c:3742: error: structure has no member named `page_lock'
fs/reiser4/txnmgr.c: In function `swap_jnode_pages':
fs/reiser4/txnmgr.c:3772: error: structure has no member named `page_lock'
fs/reiser4/txnmgr.c:3784: error: structure has no member named `page_lock'
fs/reiser4/txnmgr.c: In function `create_copy_and_replace':
fs/reiser4/txnmgr.c:4143: warning: implicit declaration of function `pte_chain_lock'
fs/reiser4/txnmgr.c:4146: warning: implicit declaration of function `try_to_unmap'
fs/reiser4/txnmgr.c:4147: error: `SWAP_AGAIN' undeclared (first use in this function)
fs/reiser4/txnmgr.c:4147: error: (Each undeclared identifier is reported only once
fs/reiser4/txnmgr.c:4147: error: for each function it appears in.)
fs/reiser4/txnmgr.c:4150: error: `SWAP_FAIL' undeclared (first use in this function)
fs/reiser4/txnmgr.c:4158: warning: implicit declaration of function `pte_chain_unlock'
make[2]: *** [fs/reiser4/txnmgr.o] Error 1
make[1]: *** [fs/reiser4] Error 2
make: *** [fs] Error 2
Any ideas on how to fix this?

I am building now without reiserfs4.

Thx

well looks like you enabled too much stuff ;)
as a general rule, dont enable anything under reiser4 except "large key support".

about the renicing tip, I suppose it works to an extent, but the key thing now is dont try to "over" stress it (compiling stuff while watch movies is ok, while compiling 4 things while playing ut2k4 and running sati is not ok :P)
but atleast i can upgrade my glibc now :D
peace

Yep! That is what is was. Thanks. :)
_________________
The strong must protect the Sweet.
Back to top
View user's profile Send private message
Gentree
Watchman
Watchman


Joined: 01 Jul 2003
Posts: 5350
Location: France, Old Europe

PostPosted: Sun Jun 13, 2004 5:42 pm    Post subject: Reply with quote

OneOfOne wrote:
Gentree wrote:
What is regarded as "heavy use" as far a the r4 pbs are concerned.

I have /portage as r4 and have just compiled xorg-x11 with no trouble on 2.6.6-love4 , would this be the kind of use likely to cause a lockup?

If so the comparison may be helpful / enlightening.

It's nice to have the "nice -15" work-around but does seem to indicate a serious flaw somewhere in the coding if the disk I/O has to be massages like that to prevent it locking up. A bit worrying. :?

heavy usage here = high cpu load + high reiser4 fs usage (mainly /var/tmp/portage)
also the lock up under heavy load is rather a new bug so I'm looking into it with Zam.
so if you have /var/tmp/portage on a sperate not-reiser4 partition you're safe.


peace


Sorry, I should have been more accurate. What I meant was /usr/portage and my $PORTAGE_TMPDIR is under this structure so all emerge and tmpdir is on r4 and fine under 2.6.6-love4.

this should provide a point of reference in tracking down the pb.

HTH.
8)
Back to top
View user's profile Send private message
polle
Veteran
Veteran


Joined: 28 Feb 2003
Posts: 1498
Location: Belgium

PostPosted: Sun Jun 13, 2004 6:51 pm    Post subject: Reply with quote

I have problems with the latest love-sources (the ones from 2.6.7), they compile fine, booting from cold start ok, but when I do a simple reboot,my system hangs at "starting cupsd", ok but that's a minor problem.
with the 2.6.7 sources I can compile the nvidia kernel (without errors) but when I try modprobe nvidia, I get an error of nvidia module.
I just copied the .config file of my lover sources 2.6.6-love4 to the directory of 2.6.7 and did oldconfig.
I'm using nvidia-kernel 4496-r3, it worked fine with anterior love-sources (last was 2.6.6-love4)
For the moment I reverted using 2.6.6-love4, but I ould like to use the 2.6.7 kernel as well, what I'm doing wrong?
Back to top
View user's profile Send private message
ejohnson
Tux's lil' helper
Tux's lil' helper


Joined: 20 Jan 2004
Posts: 140

PostPosted: Sun Jun 13, 2004 7:01 pm    Post subject: Reply with quote

polle & peacefaker -

I had those problems with nvidia-kernel as well in the past and with this kernel. It seems that nvidia's drivers are getting nasty with the newer kernels and gcc releases. Recompiling both the kernel and nvidia-kernel with gcc-3.3.3 fixed my seg faults when loading the nvidia module. You might give that a try.
Back to top
View user's profile Send private message
polle
Veteran
Veteran


Joined: 28 Feb 2003
Posts: 1498
Location: Belgium

PostPosted: Sun Jun 13, 2004 7:11 pm    Post subject: Reply with quote

I just compiled both my kernel and nvidia-kernel with the latest stable gcc, so that's not the problem I believe.
Back to top
View user's profile Send private message
Cerement
Guru
Guru


Joined: 14 Jun 2003
Posts: 404

PostPosted: Sun Jun 13, 2004 7:17 pm    Post subject: Reply with quote

Possible workaround for some people using nVidia:

Symptoms: kernel builds fine, nvidia driver builds fine, modprobe nvidia is successful and returns no error, but startx switches to a black screen and locks up keyboard and mouse, required Alt+SysRq+K to be able to switch back to a console and then kill -9 X -- modprobe -r nvidia and then switching to nv driver worked okay

If this workaround applies to you: 1) You are using the hotplug scripts at boot runlevel (as mentioned in gentoo's udev guide) -- these are in the package sys-apps/hotplug and applies to version 20040401, and -- 2) during the boot scripts (* Starting xxx [ ok ]), you see the phrase Disabling IRQ #X, and 3) cat /proc/driver/nvidia/cards/0 shows your AGP slot using the same IRQ

The workaround: Currently it looks like PCI Hotplugging is the culprit and for some reason determines that the AGP IRQ is spurious and therefore disables it -- since udev only requires sys-apps/hotplug-base, not sys-apps/hotplug, it is safe to rc-update del hotplug -- if you rely on hotplug to load any modules (USB drivers especially) then add those to autoload for now

Hope this helps ... :?
Back to top
View user's profile Send private message
polle
Veteran
Veteran


Joined: 28 Feb 2003
Posts: 1498
Location: Belgium

PostPosted: Sun Jun 13, 2004 7:23 pm    Post subject: Reply with quote

thanks for the reply, but i already get an error when I modprobe nvidia, something about invalid module format
Back to top
View user's profile Send private message
DaMouse
Apprentice
Apprentice


Joined: 18 Jul 2003
Posts: 233

PostPosted: Sun Jun 13, 2004 7:46 pm    Post subject: Reply with quote

Con Kolivas made a nifty schedular and also the autoswap stuff so I thought I'd pop a patch against love1 to give it staircase.

http://www.evilrealms.net/~damouse/kernel-stuff/patch-2.6.7-rc3-love1-cons-work.bz2

-DaMouse
_________________
I am the dark lord DaMouse I eat souls and wear the dressing gown of evil.
Back to top
View user's profile Send private message
Safrax
Guru
Guru


Joined: 23 Apr 2002
Posts: 422

PostPosted: Sun Jun 13, 2004 9:13 pm    Post subject: Reply with quote

Anyone else noticed strange swap behavior with this kernel? According to top I have 256MB of ram free but yet I've got 86mb of crap in swap. Have the kernel developers gone insane and started trying to copy Microsoft's memory mismanagement or something?
Back to top
View user's profile Send private message
fromooze
Guru
Guru


Joined: 07 Mar 2004
Posts: 360
Location: Salamanca!

PostPosted: Sun Jun 13, 2004 9:33 pm    Post subject: Reply with quote

Hi!

As I'm very happy with mm-sources, I think about trying love-sources, but me keyboard doesn't works, and the keyboard-ps2 patch doesn't work to me :( patch says line 11 is wrong... may be because I do it wrong :

Quote:

bash-2.05b# patch -p1 < fix-ps2-keyboard.patch
patching file drivers/Makefile
patch: **** malformed patch at line 11:
-


Thanks.
_________________
Xiau!

JID: fromoze@jabberes.org/jabber.org
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Unsupported Software All times are GMT
Goto page Previous  1, 2, 3, 4  Next
Page 3 of 4

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum