| View previous topic :: View next topic |
| Author |
Message |
ntraft
n00b
Joined: 02 Apr 2006
Posts: 9
Location: New Orleans
|
 Posted: Sun Apr 02, 2006 5:19 pm Post subject: Kernel REFUSES to support unloading modules! [mostly SOLVED] |
 |
|
Hello, first-time Linux user and Gentoo forums poster here...
I'm attempting to install the latest Gentoo kernel (2.6.15-r1) on my Toshiba Satellite 5205-S505. I've run into problems and errors every step of the way. I've been on the same problem for a week now and there's something very odd which might be related which I have been trying to ignore ... my kernel modules are all kinds of screwed up!
At first, I compiled the kernel without support for rmmod, and it ran fine except that at startup the machine spits out a big list of warnings that look just like these 4 lines (I had to copy by hand, this isn't the full list):
| Code: |
/etc/modules/aliases: line 45: alias: char-major-195: not found
/etc/modules/aliases: line 45: alias: nvidia: not found
/etc/modules/aliases: line 46: alias: /dev/nvidiactl: not found
/etc/modules/aliases: line 46: alias: char-major-195: not found
* Failed to calculate module dependencies |
And then I get the final warning you see there, but then the modules go on to load just fine, and I am presented w/ my login. But THEN, because I've been having trouble getting EITHER the synaptics touchpad OR the USB mouse to recognize, I decided I needed to be able to rmmod and modprobe (i've been having trouble with psmouse module too if anyone can help with that. Also, the nvidia driver won't even load -- I'm stuck using nv right now. You see my luck?).
So anyway, here's what happens when I try to reconfigure my kernel (using menuconfig) and recompile it ("# make && make modules_install"). When I try to boot up ... or, in fact, any time I try to use modprobe ... it FAILS and FREEZES, with the following message:
| Code: |
Oops: 0002 [#1]
PREEMPT SMP
Modules linked in:
CPU: 0
EIP: 0060:[<c01371d0>]
EFLAGS: 00010246 (2.6.15-gentoo-r1)
EIP is at module_unload_init+0xf/0x51
eax: e08571a0 ebx: e084f000 ecx: e0856ca0 edx: 00000007
esi: 000002f8 edi: e0851380 ebp: 00000000 esp: df389f00
ds: 007b es: 007b ss: 0068
Process modprobe (pid:1910, threadinfo=df388000 task=df7f5030)
Stack: c01390e3 e0856ca0 e084f000 e0851380 e08512d4 00000000 00003ecd 00000000
08058000 c01514a8 dfcd0680 00000320 08058000 df7f5030 c0132632 df7b9320
00000000 e0856ca0 00000000 00000000 00000000 00000000 00000000 00000007
Call Trace:
[<c01390e3>] load_module+0x738/0xbd4
[<c01514a8>] __handle_mm_fault+0xd4/0x28b
[<c0132632>] autoremove_wake_function+0x0/0x57
[<c0139608>] sys_init_module+0x66/0x239
[<c0102e6f>] sysenter_past_esp+0x54/0x75
Code: 83 88 00 00 00 89 04 24 e8 6b 2c 01 00 c7 83 88 00 00 00 00 00 00 00 83 c4 04 5b c3 8b 4c 24 04 ba 07 00 00 00 8d 81 00 05 00 00 <89> 81 00 05 00 00 89 81 04 05 00 00 89 c8 c7 80 00 01 00 00 00
/etc/init.d/modules: line 53: 1910 Segmentation fault modprobe -q${x} ${modargs[${i}]} >&/dev/null
* Failed to load svgalib_helper
* Loading module sata_sis ...
|
This happens no matter what I put in /etc/modules.autoload.d/kernel-2.6 ... although the module "libcrc32c" loads and unloads perfectly, I haven't seen any other exceptions (though I haven't tried every module). After this segfault, the console I'm using freezes, or if it's during startup, like the code above, it prints the next line claiming to load the second module, then just hangs there. Sometimes, if I've already logged in, I can switch to other consoles (I mean, the ones using the alt+F keys) and the other consoles will work sporadically.
Like I said, I'm on a Toshiba Satellite laptop with the latest gentoo-sources kernel. I compiled the kernel pretty normally, following the gentoo handbook step-by-step, only enabling support for what I think I need, and also some support for Toshiba hardware & power management extras.
So following the advent of this error, I went into my system with the liveCD and commented out every line of modules.autoload.d/kernel-2.6, and it boots up fine when there are no mods to be loaded. I then went to /usr/src/linux to recompile my kernel the way it was before, without unloading of modules support. Now it works fine once again, though I continue to get the 'alias' errors.
What does it all mean?
I have no clue whether this will help me with my more immediate problems with mouse and graphics support, but as both are partly modularized, maybe?? I've already tried every other solution known to man and mountebank. In any case, I WANT RMMOD TO WORK! HELP!!!
-- Neil
_________________
I'm about ready for a more traditional debugging tool. Got a sledgehammer?
Last edited by ntraft on Sun May 14, 2006 8:20 am; edited 6 times in total |
|
| Back to top |
|
 |
cchapman
Guru
Joined: 16 Jan 2003
Posts: 440
Location: Fremont, NE
|
| Posted: Sun Apr 02, 2006 7:49 pm Post subject: |
|
|
Unless you have a SMP box i would disable SMP support... it seems to muck things up alot on Uni processor boxen...
Also if just that doesnt work disbale premempt also. |
|
| Back to top |
|
|
Nuage
n00b
Joined: 04 Sep 2003
Posts: 39
|
| Posted: Mon Apr 03, 2006 6:26 pm Post subject: |
|
|
Hello,
I have the exact same problem. When I have the module unloading enabled in the kernel, then my computer crashes the same way. It's especially unconvenient as ALSA is unloading modules at shutdown, and my computer can never be stoped normaly.
For now I have disabled the kernel module, but has anyone found out where this bug comes from ?
I don't have SMP, and I have kernel 2.6.15-r1 too. I used to make the kernel module unlading to work on other kernels, but I could not say exactly when it did stop to work.
Thx,
Nuage |
|
| Back to top |
|
|
Etherealflaim
n00b
Joined: 22 Dec 2005
Posts: 38
|
| Posted: Mon Apr 03, 2006 6:34 pm Post subject: |
|
|
I can't say i've seen the problem, but in answer to Nuage, look at what modules ALSA Is using and compile those in. Then ALSA won't need to unload any modules on shutdown. I avoid using kernel modules like the plague, though I am forced to for things like nvidia-kernel, kqemu, and madwifi-driver.
As to your nvidia driver not loading and needing to use nv, did you remember to remerge nvidia-kernel after you recompiled your kernel?
That's all i've got for now, hope it helps!
~Kyle
[EDIT]
If you don't want to have to remember what needs to be remerged after a kernel recompile, check out the module-rebuild package. | Code: | | module-rebuild rebuild |
[/EDIT]
_________________
Gentoo / AMD64 / PPC / XFCE4 / X11R7
"If it weren't for the last minute, nothing would ever get done!" |
|
| Back to top |
|
|
ntraft
n00b
Joined: 02 Apr 2006
Posts: 9
Location: New Orleans
|
| Posted: Fri Apr 07, 2006 4:59 pm Post subject: |
|
|
OK then...
Sorry I haven't gotten to write back on what's happening to this machine; I'm a student and have been very busy with other things. So, situations have improved but there's still alot going haywire, so it'd be awesome if anyone is still paying attention to this post for some help...
2 things... perhaps related, perhaps not.
First thing is: MODULE SUPPORT:
So, I removed the multiprocessing and preemptible-kernel support from my kernel, and things seemed worse than before. This was back on Sunday and I don't quite remember what errors I had, but I put PREEMPT back in, but kept SMP out, and recompiled the kernel, and now I CAN have rmmod support without any crashes. NOW, the deal is:
- for some reason, svgalib-helper now fails to load. haven't looked into this yet
- nvidia gave some warnings at some point that it needs certain "spinlocks" and things, which sound like multi-processing features, but who knows whether that matters at all...
- the alias "not found" errors are ALL STILL THERE ... coincidentally, the same set of errors pops up when i run sbin/NVmakedevices.sh
Any ideas?
And the second thing I mentioned ... GRAPHICS SUPPORT:
I think I may be able to fix mouse support, however, as for nvidia...
After recompiling my kernel all those times I rebuilt X11 with "dlloader" support in my USE tags -- I hadn't included it before, and I found out that nvidia needs it? So I rebuilt xorg-x11 and then unmerged and remerged nvidia-glx and nvidia-kernel with dlloader enabled. The nvidia kernel module seems to load fine. NVmakedevices seems to spit out the alias errors mentioned above, and then exit. I don't even know what that script is supposed to do, so I don't know if it's working right.
And as for startx, as always it works with driver "nv", but HERE's what happens when I use "nvidia" driver instead, and load extension "glx" instead of "dri":
| Code: | Using vt 7
NVIDIA: Failed to set MTRR 0xd8000000, 128M (write-combining)
Symbol fbWinPrivateIndex from module /usr/lib/modules/drivers/nvidia_drv.o is unresolved!
Symbol fbWinPrivateIndex from module /usr/lib/modules/drivers/nvidia_drv.o is unresolved!
Symbol fbWinPrivateIndex from module /usr/lib/modules/drivers/nvidia_drv.o is unresolved!
Required symbol xf86InternetEDID from module /usr/lib/modules/drivers/nvidia_drv.o is unresolved!
Required symbol xf86PrindEDID from module /usr/lib/modules/drivers/nvidia_drv.o is unresolved!
Symbol fbPictureInit from module /usr/lib/modules/drivers/nvidia_drv.o is unresolved!
Symbol fbValidateGC from module /usr/lib/modules/drivers/nvidia_drv.o is unresolved!
Symbol fbGCPrivateIndex from module /usr/lib/modules/drivers/nvidia_drv.o is unresolved!
Symbol fbCreateGC from module /usr/lib/modules/drivers/nvidia_drv.o is unresolved!
Symbol fbCreateWindow from module /usr/lib/modules/drivers/nvidia_drv.o is unresolved!
Symbol fbWinPrivateIndex from module /usr/lib/modules/drivers/nvidia_drv.o is unresolved!
Symbol fbWinPrivateIndex from module /usr/lib/modules/drivers/nvidia_drv.o is unresolved!
Symbol fbCloseScreen from module /usr/lib/modules/drivers/nvidia_drv.o is unresolved!
Fatal server error:
Some required symbols were unresolved
|
Here's the /var/log/Xorg.0.log:
It's only the important part, b/c I had to copy it by hand, so let me know if you'd like any more information...
| Code: | (--) Chipset NVIDIA GPU found
(II) resource ranges after xf86ClaimFixedResources() call:
[a bunch of memory ranges, 28 to be exact]
(II) resource ranges after probing:
[33 mem ranges]
(II) Setting vga for screen 0.
(**) NVIDIA(0): Depth 16, (--) framebuffer bpp 16
(==) NVIDIA(0): RGB weight 565
(==) NVIDIA(0): Default visual is TrueColor
(==) NVIDIA(0): Using gamma correction (1.0, 1.0, 1.0)
(--) NVIDIA(0): Linear framebuffer at 0xD8000000
(--) NVIDIA(0): MMIO registers at 0xFD000000
# Here's where it used to fail before, and give an error of "Screen(s) found, but none have a usable configuration"
# That was before I incorporated dlloader. Now, it goes on:
(II) NVIDIA(0): NVIDIA GPU detected as: GeForce 460 Go
(--) NVIDIA(0): VideoBIOS: 04.17.00.59.d3
(--) NVIDIA(0): Interlaced video modes are supported on this GPU
(II) NVIDIA(0): Detected AGP rate: 4X
(--) NVIDIA(0): VideoRAM: 32768 kBytes
(II) NVIDIA(0): Connected display device(s): DFP-0
(--) NVIDIA(0): Display device DFP-0: maximum pixel clock at 8 bpp: 350 MHz
(--) NVIDIA(0): Display device DFP-0: maximum pixel clock at 16 bpp: 350 MHz
(--) NVIDIA(0): Display device DFP-0: maximum pixel clock at 32 bpp: 350 MHz
(II) Loading sub module "ddc"
(II) LoadModule: "ddc"
(II) Loading /usr/lib/modules/libddc.so
(II) Module ddc: vendor="X.Org Foundation"
compiled for 6.8.2, module version = 1.0.0
ABI class: X.Org Driver version 0.7
Symbol fbWinPrivateIndex from module /usr/lib/modules/drivers/nvidia_drv.o is unresolved!
Symbol fbWinPrivateIndex from module /usr/lib/modules/drivers/nvidia_drv.o is unresolved!
Symbol fbWinPrivateIndex from module /usr/lib/modules/drivers/nvidia_drv.o is unresolved!
Required symbol xf86InternetEDID from module /usr/lib/modules/drivers/nvidia_drv.o is unresolved!
Required symbol xf86PrindEDID from module /usr/lib/modules/drivers/nvidia_drv.o is unresolved!
Symbol fbPictureInit from module /usr/lib/modules/drivers/nvidia_drv.o is unresolved!
Symbol fbValidateGC from module /usr/lib/modules/drivers/nvidia_drv.o is unresolved!
Symbol fbGCPrivateIndex from module /usr/lib/modules/drivers/nvidia_drv.o is unresolved!
Symbol fbCreateGC from module /usr/lib/modules/drivers/nvidia_drv.o is unresolved!
Symbol fbCreateWindow from module /usr/lib/modules/drivers/nvidia_drv.o is unresolved!
Symbol fbWinPrivateIndex from module /usr/lib/modules/drivers/nvidia_drv.o is unresolved!
Symbol fbWinPrivateIndex from module /usr/lib/modules/drivers/nvidia_drv.o is unresolved!
Symbol fbCloseScreen from module /usr/lib/modules/drivers/nvidia_drv.o is unresolved!
Fatal server error:
Some required symbols were unresolved
|
Also important: command "eselect set opengl nvidia" fails!
Thanks always for your suggestions.
-- Neil
_________________
I'm about ready for a more traditional debugging tool. Got a sledgehammer? |
|
| Back to top |
|
|
Etherealflaim
n00b
Joined: 22 Dec 2005
Posts: 38
|
| Posted: Fri Apr 07, 2006 5:45 pm Post subject: |
|
|
| ntraft wrote: | ...
Also important: command "eselect set opengl nvidia" fails!
... |
What was the error message? That seems like it would be a fairly substantial problem.
As for the mtrr problem, post your /proc/mtrr. All lines listed there should say write-combining or something... NONE should say uncacheable. If they do, then you've got some BIOS tweaking to do. See the gentoo-wiki for more details (I think it's in the nvidia dri page).
_________________
Gentoo / AMD64 / PPC / XFCE4 / X11R7
"If it weren't for the last minute, nothing would ever get done!" |
|
| Back to top |
|
|
ntraft
n00b
Joined: 02 Apr 2006
Posts: 9
Location: New Orleans
|
| Posted: Fri Apr 07, 2006 7:16 pm Post subject: HWSetup |
|
|
Okay, as for the eselect error:
cancel that! I'm an idiot, I was typing the command syntax wrong. It does change to nvidia opengl fine.
but when i "cat /proc/mtrr", I get:
| Code: | reg00: base=0xfeda0000 (4077MB), size= 128KB: write-back, count=1
reg01: base=0xfff80000 (4095MB), size= 512KB: write-protect, count=1
reg02: base=0x00000000 ( 0MB), size= 512KB: write-back, count=1
|
... keep in mind there may be errors in this; I have to copy everything by hand...
but i guess that means they're not write-combining? I have done everything it says to do in "HOWTO nVidia Drivers" @ gentoo-wiki.com, and you're right, it does describe how to change your MTRR settings in BIOS from "continuous" to "discrete" ...
HOWEVER, my Toshiba Satellite 5205-S505 is one of a few "legacy-free" models, described in http://www.samspublishing.com/articles/article.asp?p=435192&seqNum=4&rl=1 as follows:
| Quote: | | These are "legacy-free" models, and they do not incorporate a BIOS Setup program in the motherboard ROM. The only way to run BIOS Setup is to use a Windows-based BIOS Setup utility called HWSetup. |
But the "HWSetup" utility included in my computer is HORRIBLY simple and insufficient. It is devoid of options, except for a few acpi and boot priority options, and has NOTHING to say for MTRR settings! What the hell!
Anyone know of any other utilities I can use? Am I even on the right track?
_________________
I'm about ready for a more traditional debugging tool. Got a sledgehammer? |
|
| Back to top |
|
|
|
|
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
|
|
Kernel & Hardware
