1.0-8756 and linux-kernel 2.6.16 remap_page_range bug resurfaces! - Page 3

blue666ma · 04-22-06, 02:01 PM

Okay, so I tried manually editing the Makefile.kbuild file and commenting out these if statements:

Quote:

# ifeq ($(shell $(CONFTEST) pci_get_class), 1)
# EXTRA_CFLAGS += -DNV_PCI_GET_CLASS_PRESENT
# endif

Quote:

ifeq ($(shell $(CONFTEST) remap_pfn_range), 1)
EXTRA_CFLAGS += -DNV_REMAP_PFN_RANGE_PRESENT
#else
# REMAP_PAGE_RANGE := $(shell $(CONFTEST) remap_page_range)

# ifeq ($(REMAP_PAGE_RANGE), 5)
# EXTRA_CFLAGS += -DNV_REMAP_PAGE_RANGE_5_PRESENT
# endif

# ifeq ($(REMAP_PAGE_RANGE), 4)
# EXTRA_CFLAGS += -DNV_REMAP_PAGE_RANGE_4_PRESENT
# endif
endif

Somehow, it still compiled with these two CFLAGS in it. Dunno where I can manually disable it. Is there anyway to ditch part of the make file and use my own CFLAGS? Why does it use the march=k8 flag instead of march=amd64??

blue666ma · 04-25-06, 01:06 PM

Ran tcupdate and completely recompiled the entire toolchain on my gentoo box:

Quote:

[ebuild R ] sys-apps/portage-2.0.54
[ebuild R ] sys-kernel/linux-headers-2.6.11-r2
[ebuild R ] sys-kernel/linux-headers-2.6.11-r2
[ebuild R ] sys-libs/glibc-2.3.5-r2
[ebuild R ] sys-devel/binutils-config-1.8-r7
[ebuild R ] sys-devel/binutils-2.16.1
[ebuild R ] sys-devel/gcc-config-1.3.13-r1
[ebuild R ] sys-devel/gcc-3.4.5
[ebuild R ] sys-libs/glibc-2.3.5-r2
[ebuild R ] sys-devel/binutils-2.16.1
[ebuild R ] sys-devel/gcc-3.4.5
[ebuild R ] sys-libs/glibc-2.3.5-r2
[ebuild R ] sys-devel/binutils-2.16.1
[ebuild R ] sys-devel/gcc-3.4.5
[ebuild R ] sys-apps/portage-2.0.54

Absolutely no effect.... I also ran emerge -uDN system && emerge -uDN world afterwards; pci_find_class and remap_page_range symbols still being compiled into the nvidia kernel module. Also tried switching away from distcc to compile, same thing. No known solution.

blue666ma · 04-26-06, 10:37 AM

P.S. energyman76b,

I am using the ebuilds ("I use gentoo portage to manage this install"), because the ebuilds manually issue "make install"s and as such do not hang like the nvidia installer does. When I said, "he installer issues a warning about remap_page_range as an unknown symbol" I mean to say that gcc issues the warning while compiling the ebuild. As for the kernel config, I still don't see a reason to use yours, it has nothing to do with the problem. I have MTRR built in on all my configs, and I don't load nvidiafb or nv drivers, although, even if I did, the both the ebuild and the installer use conftest.sh which checks this and dies if either is found.

Can no one help???

energyman76b · 04-26-06, 11:43 AM

And I DON'T GET THE RRORS USING THE EBUILDS!

SO IT IS NOT THE DRIVER

IT IS NOT GCC

IT IS THE KERNEL!

SO TRYING A DIFFERENT KERNEL CONFIG, ONE THAT DOES NOT SHOW THAT ERROR (MINE), MAY BE HELPFULL

But obviously you are too thick to understand that.

guero61 · 04-30-06, 10:11 PM

All the angst, all the anger!

I've just recently rebuilt my amd64 system (Presario R3000) from bare metal - 2.6.16 headers, gentoo-sources, etc. Definitely been around the block with Gentoo - since 1.3 (yes, I actually have a 1.3a CD).

I'd like to verify the reported issue, but on the sound module as well; I've used both the ebuild and run the standard NVidia installer and come up with the same results - loading the nvsound module produces a failure "Unknown symbol remap_page_range". This issue did not exist prior to my re-building the system in 64-bit native and against the 2.6.16 headers (as opposed to the 2.6.11 headers most gentoo-ers use) and glibc-2.4. Pre-rebuild, I was running 2.6.16-r1 headers, 32-bit, glibc-2.3.6-r3, and 2.6.11 headers.

I think it has to do with either glibc or the headers; which, I don't know, but it's endemic to the current NVIDIA drivers being used with what seems to be very current setups. I haven't seen it with either of my NVidia cards yet, but...

netllama · 04-30-06, 10:40 PM

If by sound module, you mean the nvsound NFORCE driver, then yes, this is a known bug. You should be able to use the ALSA driver instead.

Thanks,
Lonni

energyman76b · 04-30-06, 11:10 PM

Quote:

Originally Posted by guero61

All the angst, all the anger!

I've just recently rebuilt my amd64 system (Presario R3000) from bare metal - 2.6.16 headers, gentoo-sources, etc. Definitely been around the block with Gentoo - since 1.3 (yes, I actually have a 1.3a CD).

I'd like to verify the reported issue, but on the sound module as well; I've used both the ebuild and run the standard NVidia installer and come up with the same results - loading the nvsound module produces a failure "Unknown symbol remap_page_range". This issue did not exist prior to my re-building the system in 64-bit native and against the 2.6.16 headers (as opposed to the 2.6.11 headers most gentoo-ers use) and glibc-2.4. Pre-rebuild, I was running 2.6.16-r1 headers, 32-bit, glibc-2.3.6-r3, and 2.6.11 headers.

I think it has to do with either glibc or the headers; which, I don't know, but it's endemic to the current NVIDIA drivers being used with what seems to be very current setups. I haven't seen it with either of my NVidia cards yet, but...

pff.. I am using gentoo since 1.0

but:
I am using kernel 2.6.16.9 and prior that 2.6.16.6, 2.6.16
(not gentoo kernels, vanilla one)
glibc 2.4
gcc 4.1 (since some days)
2.6.16 headers
Amd64 (~amd64 to be correct)
64bit
8756 drivers

And I don't get the error.

blue666ma · 05-01-06, 03:47 PM

I'm gonna resurect this thread once my finals are done in 9 days, but until then, I'll let anyone who's interested put this in the think tank:

emerge --info (yes, i'm now running the beyond2 sources)

Code:

Portage 2.0.54 (default-linux/amd64/2005.1, gcc-3.4.5, glibc-2.3.5-r2, 2.6.16-beyond2 x86_64)
==========================================================
System uname: 2.6.16-beyond2 x86_64 AMD Athlon(tm) 64 Processor 3000+
Gentoo Base System version 1.6.14
distcc 2.18.3 x86_64-pc-linux-gnu (protocols 1 and 2) (default port 3632) [enabled]
ccache version 2.3 [enabled]
dev-lang/python:     2.3.5, 2.4.2
sys-apps/sandbox:    1.2.12
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-r1
sys-devel/binutils:  2.16.1
sys-devel/libtool:   1.5.22
virtual/os-headers:  2.6.11-r2
ACCEPT_KEYWORDS="amd64"
AUTOCLEAN="yes"
CBUILD="x86_64-pc-linux-gnu"
CFLAGS="-march=athlon64 -O2 -pipe"
CHOST="x86_64-pc-linux-gnu"
CXXFLAGS="-march=athlon64 -O2 -pipe"
DISTDIR="/usr/portage/distfiles"
FEATURES="autoaddcvs autoconfig ccache distcc distlocks sandbox sfperms strict 
userpriv usersandbox"
GENTOO_MIRRORS="http://gentoo.mirrored.ca/ http://mirrors.acm.cs.rpi.edu/gentoo/ http://distro.ibiblio.org/pub/linux/distributions/gentoo/"
MAKEOPTS="-j2"
PKGDIR="/usr/portage/packages"
PORTAGE_TMPDIR="/var/tmp"
PORTDIR="/usr/portage"
PORTDIR_OVERLAY="/usr/local/portage"
SYNC="rsync://rsync.gentoo.org/gentoo-portage"
USE="X a52 aac aalib acl acpi aim alsa amd64 apache2 apm arts audiofile avi
bash berkdb bindist bitmap-fonts bonobo bzip2 cdr clamav cli crypt cups curl
deltas directfb divx4linux dri dv dvd dvdr dvdread eds emboss emul encode esd
exif expat fam fbcon ffmpeg flash foomaticdb fortran ftp gdbm gif glut gnome
gphoto2 gpm gstreamer gtk gtk2 gtkhtml hal idn imagemagick imlib ipv 
isdnlog jack java javascript jikes joystick jpeg junit kde kdeenablefinal kdex
lcms libwww lzw lzw-tiff mad matroska mhash mime ming mng mozilla mp3
mpeg mysql mysqli ncurses nls nocd nptl ogg oggvorbis opengl oscar oss pam
pcre pdflib perl php pic png posix ppds pppd python qt quicktime readline
reflection ruby samba sdl session slang sockets socks5 spell spl ssl symlink
szip tcpd theora threads tiff truetype truetype-fonts type1-fonts udev unicode
usb vorbis xine xinerama xml xml2 xmms xorg xpm xv xvid zlib
userland_GNU kernel_linux elibc_glibc"
Unset:  ASFLAGS, CTARGET, INSTALL_MASK, LANG, LC_ALL, LDFLAGS, LINGUAS

glibc: 2.3.5-r2
gcc: 3.4.5
linux-headers: 2.6.11-r2
binutils: 2.6.1

blue666ma · 05-03-06, 11:33 AM

I'm starting to think that the linux-headers are the issue...

guero61 · 05-03-06, 11:52 PM

Quote:

Originally Posted by energyman76b

pff.. I am using gentoo since 1.0

...

And I don't get the error.

I see how it is... go stealing my thunder...

I do mean the nforce driver; I know I can use the ALSA, but I found it a very curious coincidence that I might stumble on these forums again in search for an answer for an odd question - and seem to have found at least validation, but in a roundabout manner.

@energyman - one thing I'm definitely _not_ doing yet is gcc-4.1; the 64-bit is my wife's machine, and I'd hate to jack it up messing around w/GCC. Says the guy running everything else foolishly...

energyman76b · 05-04-06, 09:47 AM

Hi,

I only switched some days ago - and with gcc 3.4.5 I did not get the error too

Hm, I don't recommed it either - compiling takes a lot longer (between some minutes, like 67 instead of 65 to a lot of minutes like 10 instead of 3 ...).

melchoir · 05-10-06, 09:02 PM

I had this same problem. I tried a couple of different things, most anything suggested here. My fix ended up being moving to a 2.6.15 kernel (2.6.15.7). I see some other people that posted here are using 2.6.16.x kernels so I'm not sure where the problem truly lies but this did work.

04-26-06, 10:37 AM	#27
blue666ma Registered User Join Date: Apr 2006 Posts: 18	Re: 1.0-8756 and linux-kernel 2.6.16 remap_page_range bug resurfaces! P.S. energyman76b, I am using the ebuilds ("I use gentoo portage to manage this install"), because the ebuilds manually issue "make install"s and as such do not hang like the nvidia installer does. When I said, "he installer issues a warning about remap_page_range as an unknown symbol" I mean to say that gcc issues the warning while compiling the ebuild. As for the kernel config, I still don't see a reason to use yours, it has nothing to do with the problem. I have MTRR built in on all my configs, and I don't load nvidiafb or nv drivers, although, even if I did, the both the ebuild and the installer use conftest.sh which checks this and dies if either is found. Can no one help???

04-26-06, 11:43 AM	#28
energyman76b Registered User Join Date: Dec 2002 Location: Clausthal/Germany Posts: 1,104	Re: 1.0-8756 and linux-kernel 2.6.16 remap_page_range bug resurfaces! And I DON'T GET THE RRORS USING THE EBUILDS! SO IT IS NOT THE DRIVER IT IS NOT GCC IT IS THE KERNEL! SO TRYING A DIFFERENT KERNEL CONFIG, ONE THAT DOES NOT SHOW THAT ERROR (MINE), MAY BE HELPFULL But obviously you are too thick to understand that.

04-30-06, 10:11 PM	#29
guero61 Registered User Join Date: Apr 2003 Location: Behind you Posts: 5	Re: 1.0-8756 and linux-kernel 2.6.16 remap_page_range bug resurfaces! All the angst, all the anger! I've just recently rebuilt my amd64 system (Presario R3000) from bare metal - 2.6.16 headers, gentoo-sources, etc. Definitely been around the block with Gentoo - since 1.3 (yes, I actually have a 1.3a CD). I'd like to verify the reported issue, but on the sound module as well; I've used both the ebuild and run the standard NVidia installer and come up with the same results - loading the nvsound module produces a failure "Unknown symbol remap_page_range". This issue did not exist prior to my re-building the system in 64-bit native and against the 2.6.16 headers (as opposed to the 2.6.11 headers most gentoo-ers use) and glibc-2.4. Pre-rebuild, I was running 2.6.16-r1 headers, 32-bit, glibc-2.3.6-r3, and 2.6.11 headers. I think it has to do with either glibc or the headers; which, I don't know, but it's endemic to the current NVIDIA drivers being used with what seems to be very current setups. I haven't seen it with either of my NVidia cards yet, but...

04-30-06, 10:40 PM	#30
netllama NVIDIA Corporation Join Date: Dec 2004 Posts: 8,763	Re: 1.0-8756 and linux-kernel 2.6.16 remap_page_range bug resurfaces! If by sound module, you mean the nvsound NFORCE driver, then yes, this is a known bug. You should be able to use the ALSA driver instead. Thanks, Lonni

05-03-06, 11:33 AM	#33
blue666ma Registered User Join Date: Apr 2006 Posts: 18	Re: 1.0-8756 and linux-kernel 2.6.16 remap_page_range bug resurfaces! I'm starting to think that the linux-headers are the issue...

05-04-06, 09:47 AM	#35
energyman76b Registered User Join Date: Dec 2002 Location: Clausthal/Germany Posts: 1,104	Re: 1.0-8756 and linux-kernel 2.6.16 remap_page_range bug resurfaces! Hi, I only switched some days ago - and with gcc 3.4.5 I did not get the error too Hm, I don't recommed it either - compiling takes a lot longer (between some minutes, like 67 instead of 65 to a lot of minutes like 10 instead of 3 ...).

05-10-06, 09:02 PM	#36
melchoir Registered User Join Date: May 2006 Posts: 1	Re: 1.0-8756 and linux-kernel 2.6.16 remap_page_range bug resurfaces! I had this same problem. I tried a couple of different things, most anything suggested here. My fix ended up being moving to a 2.6.15 kernel (2.6.15.7). I see some other people that posted here are using 2.6.16.x kernels so I'm not sure where the problem truly lies but this did work.