Go Back   nV News Forums > Linux Support Forums > NVIDIA Linux

Newegg Daily Deals

Reply
 
Thread Tools
Old 04-10-09, 06:37 AM   #13
MiLK_MaN
Registered User
 
Join Date: Jan 2009
Posts: 4
Default Re: 8200: Can using NVIDIA driver cause SATA errors as below? Ubuntu 8.10

I have an Asus motherboard, so I don't think its related to a specific motherboard I'm afraid.

I too raised a bug report, and havent seen a single thing at all back. I'm not going to go out and spend some more money to get a separate graphics card when its claimed the integrated graphics is sufficient for what I'm trying to achieve.
MiLK_MaN is offline   Reply With Quote
Old 05-16-09, 09:28 AM   #14
Haggy
Registered User
 
Join Date: May 2009
Posts: 13
Default Re: 8200: Can using NVIDIA driver cause SATA errors as below? Ubuntu 8.10

Hi,

I just registered solely because of this issue. I want to roll it up again since it present in even the latest nvidia drivers (i tested 180.51 and 185.18.08). A few seconds after X starts up (i run xbmc directly after) dmesg logs sata errors as shown above. Afterwards my home partition (only this one!) gets remounted read-only. I can't remember having those issues with 185.19 but i'll give it a try again and report back.

My system specs:
- ASUS M3N78-EM with 8300 IGP
- ArchLinux x86_64 with 4GB RAM

Any hints? Is nvidia aware of this?
Haggy is offline   Reply With Quote
Old 05-16-09, 05:11 PM   #15
MiLK_MaN
Registered User
 
Join Date: Jan 2009
Posts: 4
Default Re: 8200: Can using NVIDIA driver cause SATA errors as below? Ubuntu 8.10

I ended up putting Windows on to see if it was a Linux only issue, and the issue also occurred in Windows.

I first thought it was the hard drive, so I replaced that, issue actually got worse. Swapped SATA cables, no difference.

I've now ripped out the motherboard and sent it back to Gigabyte. My brother works at a computer repair shop, and I gave him a buzz, and it turns out that while its certainly not commong for a SATA controller to be dodgey on a motherboard, it does happen.

I did a bit of reading around from google searches, and there has been 1 or 2 responses of a replacement motherboard fixing the issue, so here's hoping.
MiLK_MaN is offline   Reply With Quote
Old 05-16-09, 09:14 PM   #16
pcjason
Registered User
 
Join Date: Dec 2007
Posts: 17
Default Re: 8200: Can using NVIDIA driver cause SATA errors as below? Ubuntu 8.10

Quote:
Originally Posted by MiLK_MaN View Post
I ended up putting Windows on to see if it was a Linux only issue, and the issue also occurred in Windows.

I first thought it was the hard drive, so I replaced that, issue actually got worse. Swapped SATA cables, no difference.

I've now ripped out the motherboard and sent it back to Gigabyte. My brother works at a computer repair shop, and I gave him a buzz, and it turns out that while its certainly not commong for a SATA controller to be dodgey on a motherboard, it does happen.

I did a bit of reading around from google searches, and there has been 1 or 2 responses of a replacement motherboard fixing the issue, so here's hoping.
Interesting. I could have sworn that I tested this with Windows XP and the latest version of the Windows drivers + MPC-HC to play HD content and I did not see this issue. I'm tempted to try installing Windows and running some OpenGL applications just to see if I can reproduce this, but honestly I'm not hopeful. The plain fact is, though, that there are no issues when using the VESA driver and running X. The issues only appear when using the nvidia driver and running X.
pcjason is offline   Reply With Quote
Old 05-17-09, 03:17 AM   #17
Haggy
Registered User
 
Join Date: May 2009
Posts: 13
Default Re: 8200: Can using NVIDIA driver cause SATA errors as below? Ubuntu 8.10

Hmm, thanks for the update. That's my last hope - but i still cannot figure out why this happens only while running X / while the nvidia kernel module is loaded. I gave 185.19 a try and first it seemed that fixed the problem, but now i have it once again, though not that frequently. However, what i did notice is that nvidia switches interrupts more or less randomly and once it got an interrupt alone (IRQ22) and i THINK, that was a stable situation.
Haggy is offline   Reply With Quote
Old 05-17-09, 10:49 AM   #18
ledoc
Registered User
 
Join Date: Jan 2008
Posts: 330
Default Re: 8200: Can using NVIDIA driver cause SATA errors as below? Ubuntu 8.10

Maybe, if it's related to interrupts, you could try putting the nvidia driver on MSI with NVreg_EnableMSI=1?
ledoc is offline   Reply With Quote
Old 05-17-09, 01:02 PM   #19
Haggy
Registered User
 
Join Date: May 2009
Posts: 13
Default Re: 8200: Can using NVIDIA driver cause SATA errors as below? Ubuntu 8.10

Thanks, but that made it even worse (though i did not see the sata issue). Seconds after X started up the whole machine locked up hard flooding the kernel log with several errors and only powercycling brought it back. I just disabled NvAGP in xorg.conf and for now it's stable, but i guess just after sending this post it breaks again
Haggy is offline   Reply With Quote
Old 05-17-09, 02:33 PM   #20
zkmyth
Registered User
 
Join Date: Apr 2009
Posts: 10
Default Re: 8200: Can using NVIDIA driver cause SATA errors as below? Ubuntu 8.10

Haggy:

If it is an IRQ sharing problem then you should be able to get the nvidia driver on it's own irq by enabling MSI on the other devices that support it (not the nvidia module it MSI support is unstable currently) and then controlling when the remaining modules/drivers are loaded.

I have managed to get my ubuntu64 system to boot consistently with the nvidia driver on it's own irq with 2 different kernels (2.6.27 and 2.6.29 other should work too)

The basic ideas are to put all of the important modules in /etc/modprobe.d/blacklist then put them into /etc/modules in the desired order and finally run "update-initramfs -u" to generate a new initrd image.

For my M3N78 Pro system I did the following

Enable MSI
snd_hda_intel -> add "options snd_hda_intel enable_msi=1" to /etc/modprobe.d/options
forcedeth should do it automatically
ahci should do it automatically as well

/etc/modprobe.d/blacklist has the following at the bottom of it ( note: do not put nvidia in the blacklist it will never load if you do)
blacklist forcedeth
blacklist snd_hda_intel
blacklist ehci_hcd
blacklist uhci_hcd
blacklist ohci_hcd

/etc/modules has
ahci
snd_hda_intel
ehci_hcd
forcedeth
nvidia
uhci_hcd
ohci_hcd

Note the order here it is important ( although the nvidia module seems to ignore the order )

Then Update the initrd image and reboot. It may take you a few tries to get the order right.

The kernel seems to load the drivers/modules and assign them an IRQ, after the module has started to load it switches to MSI and drops the IRQ.
However the kernel assigns IRQs in the order 23,22,21,20 23,22,21,20 ...(for my pcie bus) and even though the modules switches to MSI the kernel continues to assign the next IRQ in order.
So you need to control when the kernel loads the module to ensure that the nvidia device gets an IRQ that is shared with a device that will switch to MSI and thus get the nvidia driver on a free IRQ.

Very confusing but for me it has lead to a stable system, it only took me 9 months to figure it all out and make it work.

For me it was todo with the usb modules since my system has 4 usb buses and they want to take 4 IRQs, kinda a problem when the pcie bus only has 4 IRQs. So I needed to get the usb buses to share IRQs. This is doable since 2 usb are ehci (USB2) and 2 are ohci (USB1) so I controlled when ehci was loaded and when ohci was loaded to get them to share some IRQs.

You can analyze what order things are loading for you like this

just after booting run

"dmesg | grep IRQ"

now look at the IRQs 23-20, you should see a line like this ( perhaps with a different module name )
ahci 0000:00:09.0: PCI INT A -> Link[APSI] -> GSI 23 (level, low) -> IRQ 23

Ignore the IRQs below 20 since they are the PCI bus. If you make a list of IRQs (23,22,21,20) and what modules are using them it might help clear things up.

once you do that you can compare it to the output of "cat /proc/interrupts" and figure out when to get the modules to load.

see also http://www.nvnews.net/vbulletin/showthread.php?t=123583

a couple of alternate ideas would be to disable the USB in the bios and see if your system is stable. If you still need USB you can use a PCI usb card it will use IRQs below 20. I have tested this and it works too.
zkmyth is offline   Reply With Quote

Old 05-17-09, 02:52 PM   #21
Haggy
Registered User
 
Join Date: May 2009
Posts: 13
Default Re: 8200: Can using NVIDIA driver cause SATA errors as below? Ubuntu 8.10

Wow, thanks for that really extensive explanation. Thank you! Seems i'm not the only one with troubles on this kind of board (it's a ASUS M3N78-EM for me). I'll give it a try as soon as it breaks again (currently stable for ~ 1 hour with pci=nomsi and 3 threads of bonnie++ hammering the sata bus). Will report back if that helps.
Haggy is offline   Reply With Quote
Old 05-18-09, 04:03 AM   #22
Haggy
Registered User
 
Join Date: May 2009
Posts: 13
Default Re: 8200: Can using NVIDIA driver cause SATA errors as below? Ubuntu 8.10

Still stable with pci=nomsi for now. I'm quite happy since that parameter also fixed my resume/suspend issues with this board. without it resuming always ended in a readonly filesystem and again sata errors in my kernel log.

haggy@aereogramme ~/files/build/xbmc/xbmc-svn $ cat /proc/interrupts
CPU0 CPU1
0: 51 61 IO-APIC-edge timer
1: 0 2 IO-APIC-edge i8042
4: 16 2953 IO-APIC-edge lirc_serial
7: 1 0 IO-APIC-edge
8: 0 184 IO-APIC-edge rtc0
9: 0 0 IO-APIC-fasteoi acpi
12: 0 4 IO-APIC-edge i8042
20: 5817 1048939 IO-APIC-fasteoi ehci_hcd:usb2, eth0
21: 659 111452 IO-APIC-fasteoi ehci_hcd:usb1, HDA Intel
22: 2518 2692732 IO-APIC-fasteoi ohci_hcd:usb4, nvidia
23: 226 52124 IO-APIC-fasteoi ahci, ohci_hcd:usb3
NMI: 0 0 Non-maskable interrupts
LOC: 515499 2971393 Local timer interrupts
RES: 49149 15118 Rescheduling interrupts
CAL: 1440348 22079 Function call interrupts
TLB: 68454 14268 TLB shootdowns
TRM: 0 0 Thermal event interrupts
THR: 0 0 Threshold APIC interrupts
SPU: 0 0 Spurious interrupts
ERR: 1
MIS: 0
Haggy is offline   Reply With Quote
Old 06-04-09, 07:19 AM   #23
Haggy
Registered User
 
Join Date: May 2009
Posts: 13
Default Re: 8200: Can using NVIDIA driver cause SATA errors as below? Ubuntu 8.10

Damn it. I got the sata errors again, but i noticed that changing the APM level of the hard drive does somehow have an effect. If i set my drive (WD3200BEVT) to 192, the sata errors don't occur (at least not for several hours), if i disable APM (with hdparm -B255) i get the sata errors again after a few minutes. Sata controllers are running AHCI mode.

EDIT: I have to say that with an APM level of less than 255 my Load_Cycle_Count goes through the roof. My drive is at ~30000 cycles and it's only ~4 months old!
Haggy is offline   Reply With Quote
Old 06-08-09, 05:34 AM   #24
Haggy
Registered User
 
Join Date: May 2009
Posts: 13
Default Re: 8200: Can using NVIDIA driver cause SATA errors as below? Ubuntu 8.10

And another update. It seems a bios upgrade to 0604 for my ASUS M3N78-EM has fixed the issues. I haven't seen the error in ~8h uptime now - before it already appeared after a few minutes. Also, setting APM level for my harddrive now seems to work with this BIOS revision, as my load cycle count does not increase anymore.
Haggy is offline   Reply With Quote
Reply


Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump


Similar Threads
Thread Thread Starter Forum Replies Last Post
nvidia driver quality decrease? legluondunet NVIDIA Linux 17 06-26-12 10:03 AM
GT650M not recognized by 295.59 Driver on Ubuntu 12.04 hvontres NVIDIA Linux 13 06-22-12 01:33 PM
Need Help Installing NVIDIA Tesla M2070Q in Linux RHEL5 Ferianto85 NVIDIA Linux 0 05-18-12 08:35 PM
Getting the proprietary nvidia driver to run with Debian 3.0 r0 (woody) Katchina404 NVIDIA Linux 9 01-12-03 08:49 AM
nvidia + 2.4.19 gentoo = messed up ulukay NVIDIA Linux 18 11-01-02 11:58 PM

All times are GMT -5. The time now is 06:48 AM.


Powered by vBulletin® Version 3.7.1
Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Copyright 1998 - 2014, nV News.