You are not logged in.

#26 2004-08-19 07:55:45

tpowa
Developer
From: Lauingen , Germany
Registered: 2004-04-05
Posts: 2,322

Re: kernel26mm 2.6.8.1.mmX

from nvidia readme:
Q: My system hangs when vt-switching if I have rivafb enabled.

A: Using both rivafb and the NVIDIA kernel module at the same time is
   currently broken.  In general, using two independent software drivers
   to drive the same piece of hardware is a bad idea.

perhaps video=vesa as boot parameter works didn't test it
found that in kernel-parameters documentation

i'll test it now give you feedback when it's done

Offline

#27 2004-08-19 08:06:51

tpowa
Developer
From: Lauingen , Germany
Registered: 2004-04-05
Posts: 2,322

Re: kernel26mm 2.6.8.1.mmX

the problem will only arise
if you load the rivafb module

if you don't touch it it works at least for me

by the way:
is it usefull to have the fb part as modules?
i thought the kernel needs the module at the beginning of the boot process.
The root partitions is mounted later so a ramdisk would be necessary to achive this

Offline

#28 2004-08-19 08:15:42

dp
Member
From: Zürich, Switzerland
Registered: 2003-05-27
Posts: 3,378
Website

Re: kernel26mm 2.6.8.1.mmX

tpowa wrote:

the problem will only arise
if you load the rivafb module

if you don't touch it it works at least for me

good to know - thx for checking

tpowa wrote:

by the way:
is it usefull to have the fb part as modules?
i thought the kernel needs the module at the beginning of the boot process.
The root partitions is mounted later so a ramdisk would be necessary to achive this

maybe it would be better to have them inside the kernel, but as modules, the user sees them in lsmod and knows that they are loaded (i dont want to surprise people by just like that using accelerated fb)

at least for me, programms using fb work much faster with radeonfb instead of vesa :-)


The impossible missions are the only ones which succeed.

Offline

#29 2004-08-19 08:33:48

tpowa
Developer
From: Lauingen , Germany
Registered: 2004-04-05
Posts: 2,322

Re: kernel26mm 2.6.8.1.mmX

do you load the fb module in rc.conf or by a ramdisk?

if you decide to include the fb stuff into the kernel then you have to use the video=vesa,radeonfb,... option

Offline

#30 2004-08-19 08:41:54

Mr Green
Forum Fellow
From: U.K.
Registered: 2003-12-21
Posts: 5,893
Website

Re: kernel26mm 2.6.8.1.mmX

Building kernel module:
   executing: 'cd ./usr/src/nv; make module SYSSRC=/lib/modules/2.6.8.1-mm1/bui
   ld SYSOUT=/lib/modules/2.6.8.1-mm1/build'...
   Your kernel was configured to include rivafb support as
   a loadable kernel module.

   The rivafb driver conflicts with the NVIDIA driver; the
   NVIDIA kernel module will still be built and installed,
   but be aware that the NVIDIA driver will not be able to
   function properly if the rivafb module is loaded!

   *** Failed rivafb module sanity check, but continuing! ***

HTH

dumb question but what do I do to correct this. ?


Mr Green

Offline

#31 2004-08-19 08:47:47

dp
Member
From: Zürich, Switzerland
Registered: 2003-05-27
Posts: 3,378
Website

Re: kernel26mm 2.6.8.1.mmX

tpowa wrote:

do you load the fb module in rc.conf or by a ramdisk?

if you decide to include the fb stuff into the kernel then you have to use the video=vesa,radeonfb,... option

it simply loades automatically - i pass the kernel only "vga=792" and the rest works magically :-)

(no rc.conf and by ramdisk i dont know what way you mean)


The impossible missions are the only ones which succeed.

Offline

#32 2004-08-19 08:49:26

dp
Member
From: Zürich, Switzerland
Registered: 2003-05-27
Posts: 3,378
Website

Re: kernel26mm 2.6.8.1.mmX

Mr Green wrote:
Building kernel module:
   executing: 'cd ./usr/src/nv; make module SYSSRC=/lib/modules/2.6.8.1-mm1/bui
   ld SYSOUT=/lib/modules/2.6.8.1-mm1/build'...
   Your kernel was configured to include rivafb support as
   a loadable kernel module.

   The rivafb driver conflicts with the NVIDIA driver; the
   NVIDIA kernel module will still be built and installed,
   but be aware that the NVIDIA driver will not be able to
   function properly if the rivafb module is loaded!

   *** Failed rivafb module sanity check, but continuing! ***

HTH

dumb question but what do I do to correct this. ?

if you use video=vesa, the rivafb will not load - however, if all nvidia users use the native drivers, i will remove the kernel-own fb-one from the kernel


The impossible missions are the only ones which succeed.

Offline

#33 2004-08-19 09:33:47

tpowa
Developer
From: Lauingen , Germany
Registered: 2004-04-05
Posts: 2,322

Re: kernel26mm 2.6.8.1.mmX

with ramdisk i think of initrd for modules like reiserfs scsi etc.

i passed no kernel argument to the kernel and it doesn't load rivafb automatically perhaps this is only for older nvidia cards
i have a geforce 4 ti

the warning message tells you only that there will be trouble if you use the module ;-)

Offline

#34 2004-08-19 09:37:55

dp
Member
From: Zürich, Switzerland
Registered: 2003-05-27
Posts: 3,378
Website

Re: kernel26mm 2.6.8.1.mmX

tpowa wrote:

with ramdisk i think of initrd for modules like reiserfs scsi etc.

oh .. ok, no, i don0t have them there

tpowa wrote:

i passed no kernel argument to the kernel and it doesn't load rivafb automatically perhaps this is only for older nvidia cards
i have a geforce 4 ti

the warning message tells you only that there will be trouble if you use the module ;-)

ok - rivafb stays for now


The impossible missions are the only ones which succeed.

Offline

#35 2004-08-19 11:07:26

Mr Green
Forum Fellow
From: U.K.
Registered: 2003-12-21
Posts: 5,893
Website

Re: kernel26mm 2.6.8.1.mmX

Well with append="video=vesa" 

Nvidia drivers have compiled with warnings .....

Currently working in 2.6.8.1-mm1 see how it goes ...

thanks guys  tongue


Mr Green

Offline

#36 2004-08-19 11:12:10

dp
Member
From: Zürich, Switzerland
Registered: 2003-05-27
Posts: 3,378
Website

Re: kernel26mm 2.6.8.1.mmX

Mr Green wrote:

Well with append="video=vesa" 

Nvidia drivers have compiled with warnings .....

Currently working in 2.6.8.1-mm1 see how it goes ...

thanks guys  tongue

the kernel itself compiles with around 11 warnings under gcc 3.4.x, warnings are in most cases no problem, as long as there are no errors :-)


The impossible missions are the only ones which succeed.

Offline

#37 2004-08-19 11:19:50

dp
Member
From: Zürich, Switzerland
Registered: 2003-05-27
Posts: 3,378
Website

Re: kernel26mm 2.6.8.1.mmX

2.6.8.1.mm2 is out and compiling right now - they solved the mem leak while burning audio !!!

EDIT: but most probably burning audio still broken:
http://kerneltrap.org/node/view/3674


The impossible missions are the only ones which succeed.

Offline

#38 2004-08-19 12:43:03

dp
Member
From: Zürich, Switzerland
Registered: 2003-05-27
Posts: 3,378
Website

Re: kernel26mm 2.6.8.1.mmX

the rivafb is an old problem - that's why nvidia itself warns people:

http://bugzilla.kernel.org/show_bug.cgi?id=728


The impossible missions are the only ones which succeed.

Offline

#39 2004-08-19 15:24:13

colnago
Member
From: Victoria, BC
Registered: 2004-03-25
Posts: 438

Re: kernel26mm 2.6.8.1.mmX

I was worried a bit due to the bumps in 2.6.8 since 2.6.7 was so very smooth.  I have been trying your 2.6.8.1-mm2 build on these problems and I find all is back to normal now.   I did not test the wine compatibility problem a previous poster had, and I rarely burn audio so who knows, but everything else is great.  Thank you very much for the quick update.

Offline

#40 2004-08-19 15:56:29

cmf
Member
Registered: 2003-10-18
Posts: 86

Re: kernel26mm 2.6.8.1.mmX

Planning on including Reiser4 support seen as it's in mm2 now?

If you do i think it will be time for an Arch reinstall!

Offline

#41 2004-08-19 17:09:51

dp
Member
From: Zürich, Switzerland
Registered: 2003-05-27
Posts: 3,378
Website

Re: kernel26mm 2.6.8.1.mmX

cmf wrote:

Planning on including Reiser4 support seen as it's in mm2 now?

If you do i think it will be time for an Arch reinstall!

it should be compiled in - but i cannot test it because of lack of free harddisk space :-( (have a 40gb hdd in 5 partitions and all are 96% or fuller)


The impossible missions are the only ones which succeed.

Offline

#42 2004-08-19 17:27:28

Mr Green
Forum Fellow
From: U.K.
Registered: 2003-12-21
Posts: 5,893
Website

Re: kernel26mm 2.6.8.1.mmX

clear your pacman cache

df -hT

HTH


Mr Green

Offline

#43 2004-08-19 23:41:02

tranquility
Member
From: Portugal
Registered: 2004-08-06
Posts: 136

Re: kernel26mm 2.6.8.1.mmX

Hi. I just installed kernel26mm and could not get it to find my / . When it boots, the my SATA controller is nowhere to be seen, and I've tried using root=/dev/sda5 and nothing. Here's my grub entry for this kernel:

# (1) Arch Linux kernel mm
title  Arch Linux mm [/boot/vmlinuz26]
root   (hd0,4)
kernel /boot/vmlinuz26mm root=/dev/scsi/host0/bus0/target0/lun0/part5 ro hdg=noprobe vga=773 devfs=nomount

I'm having a deja-vu, since I'm having the exact same problem with kernel26-2.6.8, and I still couldn't get it to find my root partition. Problem solved though, I'm now booting kernel26-2.6.7 which works fine.
Hope it helps. Joao.

Offline

#44 2004-08-19 23:59:27

dp
Member
From: Zürich, Switzerland
Registered: 2003-05-27
Posts: 3,378
Website

Re: kernel26mm 2.6.8.1.mmX

Mr Green wrote:

clear your pacman cache

df -hT

HTH

everything clean as possible - cleaning the cvs trees and everything gives me ~1gb of free on / and ~800mb of free on /home :-)


The impossible missions are the only ones which succeed.

Offline

#45 2004-08-20 00:00:20

dp
Member
From: Zürich, Switzerland
Registered: 2003-05-27
Posts: 3,378
Website

Re: kernel26mm 2.6.8.1.mmX

tranquility wrote:

Hi. I just installed kernel26mm and could not get it to find my / . When it boots, the my SATA controller is nowhere to be seen, and I've tried using root=/dev/sda5 and nothing. Here's my grub entry for this kernel:

# (1) Arch Linux kernel mm
title  Arch Linux mm [/boot/vmlinuz26]
root   (hd0,4)
kernel /boot/vmlinuz26mm root=/dev/scsi/host0/bus0/target0/lun0/part5 ro hdg=noprobe vga=773 devfs=nomount

I'm having a deja-vu, since I'm having the exact same problem with kernel26-2.6.8, and I still couldn't get it to find my root partition. Problem solved though, I'm now booting kernel26-2.6.7 which works fine.
Hope it helps. Joao.

what adapter/chipset do you have?


The impossible missions are the only ones which succeed.

Offline

#46 2004-08-20 14:16:44

tranquility
Member
From: Portugal
Registered: 2004-08-06
Posts: 136

Re: kernel26mm 2.6.8.1.mmX

dp wrote:

what adapter/chipset do you have?

From 2.6.7 dmesg:

 $ dmesg  | grep Si
SiI3112 Serial ATA: IDE controller at PCI slot 0000:01:0b.0
SiI3112 Serial ATA: chipset revision 2
SiI3112 Serial ATA: 100% native mode on irq 217
    ide2: MMIO-DMA , BIOS settings: hde:pio, hdf:pio
    ide3: MMIO-DMA , BIOS settings: hdg:pio, hdh:pio
hde: Maxtor 6Y160M0, ATA DISK drive

Offline

#47 2004-08-20 14:38:50

dp
Member
From: Zürich, Switzerland
Registered: 2003-05-27
Posts: 3,378
Website

Re: kernel26mm 2.6.8.1.mmX

SiI3112 uses CONFIG_SCSI_SATA_SIL=y and this is enabled in kernel26 and kernel26mm - so you should get it working

however, i'm not an expert - maybe someone else has more infos

where i searched:
http://www.linuxmafia.com/faq/Hardware/sata.html


The impossible missions are the only ones which succeed.

Offline

#48 2004-08-20 14:47:11

tpowa
Developer
From: Lauingen , Germany
Registered: 2004-04-05
Posts: 2,322

Re: kernel26mm 2.6.8.1.mmX

NEWS from the NVIDIA drivers

problem with mm2
only works if you compile it fresh and then start X
else it will crash the System

new kernel mm3
breaks nvidia drivers completly unable to compile the drivers

Offline

#49 2004-08-20 14:59:56

dp
Member
From: Zürich, Switzerland
Registered: 2003-05-27
Posts: 3,378
Website

Re: kernel26mm 2.6.8.1.mmX

tpowa wrote:

NEWS from the NVIDIA drivers

problem with mm2
only works if you compile it fresh and then start X
else it will crash the System

new kernel mm3
breaks nvidia drivers completly unable to compile the drivers

http://kerneltrap.org/node/view/3680

- the mm3 fixed the audio burning :-)

- but the new scheduler seems to break nvidia compiling :-( --- what output do you have from compiling? if it is something related to the scheduler, you can post a bug to http://bugzilla.kernel.org (Alternate Trees, -> mm) or post on the http://kerneltrap.org/node/view/3680 article


The impossible missions are the only ones which succeed.

Offline

#50 2004-08-20 15:07:10

tpowa
Developer
From: Lauingen , Germany
Registered: 2004-04-05
Posts: 2,322

Re: kernel26mm 2.6.8.1.mmX

NVIDIA: calling KBUILD...
   make -C /lib/modules/2.6.8.1-mm3/build 
   KBUILD_SRC=/usr/src/linux-2.6.8.1-mm3      KBUILD_VERBOSE=1
   KBUILD_CHECK= KBUILD_EXTMOD="/tmp/selfgz3408/NVIDIA-Linux-x86-1.0-6111-pkg0/
   usr/src/nv"
           -f /usr/src/linux-2.6.8.1-mm3/Makefile modules
   mkdir -p /tmp/selfgz3408/NVIDIA-Linux-x86-1.0-6111-pkg0/usr/src/nv/.tmp_vers
   ions
   make -f /usr/src/linux-2.6.8.1-mm3/scripts/Makefile.build obj=/tmp/selfgz340
   8/NVIDIA-Linux-x86-1.0-6111-pkg0/usr/src/nv
   echo #define NV_COMPILER "`cc -v 2>&1 | tail -n 1`" > /tmp/selfgz3408/NVI
   DIA-Linux-x86-1.0-6111-pkg0/usr/src/nv/nv_compiler.h
     cc -Wp,-MD,/tmp/selfgz3408/NVIDIA-Linux-x86-1.0-6111-pkg0/usr/src/nv/.nv.o
   .d -nostdinc -iwithprefix include -D__KERNEL__ -Iinclude -Iinclude2 -I/usr/s
   rc/linux-2.6.8.1-mm3/include  -I/tmp/selfgz3408/NVIDIA-Linux-x86-1.0-6111-pk
   g0/usr/src/nv -Wall -Wstrict-prototypes -Wno-trigraphs -fno-strict-aliasing
   -fno-common -pipe -msoft-float -mpreferred-stack-boundary=2 -fno-unit-at-a-t
   ime -march=i686 -I/usr/src/linux-2.6.8.1-mm3/include/asm-i386/mach-default -
   Iinclude/asm-i386/mach-default -O2 -fomit-frame-pointer -Wdeclaration-after-
   statement  -I/tmp/selfgz3408/NVIDIA-Linux-x86-1.0-6111-pkg0/usr/src/nv -Wall
   -Wimplicit -Wreturn-type -Wswitch -Wformat -Wchar-subscripts -Wparentheses -
   Wpointer-arith -Wno-multichar -Werror -O -fno-common -MD -Wno-cast-qual -Wno
   -error -D_LOOSE_KERNEL_NAMES -D__KERNEL__ -DMODULE -DNTRM -D_GNU_SOURCE -D_L
   OOSE_KERNEL_NAMES -D__KERNEL__ -DMODULE -DNV_MAJOR_VERSION=1 -DNV_MINOR_VERS
   ION=0 -DNV_PATCHLEVEL=6111 -DNV_UNIX -DNV_LINUX -DNV_INT64_OK -DNVCPU_X86 -U
   DEBUG -U_DEBUG -DNDEBUG -DNV_CHANGE_PAGE_ATTR_PRESENT -DNV_CLASS_SIMPLE_PRES
   ENT -DMODULE -
   DKBUILD_BASENAME=nv -DKBUILD_MODNAME=nvidia -c -o /tmp/selfgz3408/NVIDIA-Lin
   ux-x86-1.0-6111-pkg0/usr/src/nv/.tmp_nv.o /tmp/selfgz3408/NVIDIA-Linux-x86-1
   .0-6111-pkg0/usr/src/nv/nv.c
   In file included from include/linux/list.h:7,
                    from include/linux/wait.h:23,
                    from include/asm/semaphore.h:41,
                    from include/linux/sched.h:19,
                    from include/linux/module.h:10,
                    from /tmp/selfgz3408/NVIDIA-Linux-x86-1.0-6111-pkg0/usr/src
   /nv/nv-linux.h:52,
                    from /tmp/selfgz3408/NVIDIA-Linux-x86-1.0-6111-pkg0/usr/src
   /nv/nv.c:14:
   include/linux/prefetch.h: In function `prefetch_range':
   include/linux/prefetch.h:62: Warnung: Zeiger auf Typen »void *« in Arithmeti
   k verwendet
   /tmp/selfgz3408/NVIDIA-Linux-x86-1.0-6111-pkg0/usr/src/nv/nv.c: In function
   `nv_acpi_event':
   /tmp/selfgz3408/NVIDIA-Linux-x86-1.0-6111-pkg0/usr/src/nv/nv.c:3697: Fehler:
   »PM_SAVE_STATE« nicht deklariert (erste Benutzung in dieser Funktion)
   /tmp/selfgz3408/NVIDIA-Linux-x86-1.0-6111-pkg0/usr/src/nv/nv.c:3697: Fehler:
   (Jeder nicht deklarierte Bezeichner wird nur einmal aufgeführt
   /tmp/selfgz3408/NVIDIA-Linux-x86-1.0-6111-pkg0/usr/src/nv/nv.c:3697: Fehler:
   für jede Funktion in der er auftritt.)
   make[4]: *** [/tmp/selfgz3408/NVIDIA-Linux-x86-1.0-6111-pkg0/usr/src/nv/nv.o
   ] Fehler 1
   make[3]: *** [_module_/tmp/selfgz3408/NVIDIA-Linux-x86-1.0-6111-pkg0/usr/src
   /nv] Fehler 2
   make[2]: *** [modules] Fehler 2
   NVIDIA: left KBUILD.
   nvidia.ko failed to build!
   make[1]: *** [module] Fehler 1
   make: *** [module] Fehler 2
-> Error.

Offline

Board footer

Powered by FluxBB