| View previous topic :: View next topic |
| Author |
Message |
Konsti
l33t
Joined: 10 Dec 2002
Posts: 691
|
 Posted: Fri May 27, 2005 8:39 am Post subject: does somebody also run gentoo on SN85G4V2? |
 |
|
I have urgent trouble with my forcedeth ethernet device on this Shuttle XPC Barebone.
Does somebody run gentoo 64Bit with a recent kernel and has Problems with ethernet or not?
Konsti |
|
| Back to top |
|
 |
96140
Retired Dev
Joined: 23 Jan 2005
Posts: 1324
|
| Posted: Fri May 27, 2005 8:46 am Post subject: |
|
|
Have you updated the BIOS yet?
Many more recent Shuttles have numerous hardware troubles, especially with motherboards and NICs, right out of the box. But AFAIK, they're typically solved by downloading a BIOS update from the Shuttle website, saving it to a floppy, and booting the disk to flash the BIOS.
From what I've seen in many, many places around the 'net the hardware troubles are identical for users of any OS, whether Linux or Gentoo.
But you might also stop by the AMD64 forum to see if users there have any ideas. |
|
| Back to top |
|
|
Konsti
l33t
Joined: 10 Dec 2002
Posts: 691
|
| Posted: Fri May 27, 2005 8:50 am Post subject: |
|
|
Hm, I will look at their website. I have not Flashed the BIOS yet.
I have no Floppy, no Cables for it and such... But I will try it 
I did not post into amd64 problem because somebody might run 32Bit on it and tell his NIC works like a charm so it may be a 64Bit driver Problem in forcedeth because the author has not much experience with 64Bit or whatsoever,,, |
|
| Back to top |
|
|
96140
Retired Dev
Joined: 23 Jan 2005
Posts: 1324
|
| Posted: Fri May 27, 2005 8:56 am Post subject: |
|
|
| Eh, you should be able to still boot from the CD drive (I know, a waste of a whole CD-R for one BIOS update). I think you can even use a USB stick, if you have one. I was just browsing the Shuttle FAQs today, and there is a section on flashing the BIOS. |
|
| Back to top |
|
|
Konsti
l33t
Joined: 10 Dec 2002
Posts: 691
|
| Posted: Fri May 27, 2005 9:02 am Post subject: |
|
|
| Interesting, will try that USB Stick stuff... |
|
| Back to top |
|
|
JuNix
Apprentice
Joined: 05 Mar 2003
Posts: 224
Location: Sheffield
|
| Posted: Fri May 27, 2005 12:00 pm Post subject: |
|
|
Hi , let us know how this BIOS upgrade went. I have the same hardware, and if I have a lot of sockets open using the network with high traffic utilisation, and try burning a dvd, for example, I start to get horrendous packet loss, until finally the shuttle goes unreachable.
If I unplug the ethernet cable, and re-insert it, this seems to "reset" the eth0 port on the shuttle and the problem goes away. I wondered if the forcedeth reversed engineered driver had problems.
I may try the BIOS upgrade. |
|
| Back to top |
|
|
Konsti
l33t
Joined: 10 Dec 2002
Posts: 691
|
| Posted: Fri May 27, 2005 12:28 pm Post subject: |
|
|
That sounds like my Problem. with what Kernel Versions does this Problem occur?
I will post my effords in dos USB booting soon  |
|
| Back to top |
|
|
JuNix
Apprentice
Joined: 05 Mar 2003
Posts: 224
Location: Sheffield
|
| Posted: Fri May 27, 2005 1:37 pm Post subject: |
|
|
I'm running
| Code: | shuttle $ uname -a
Linux shuttle 2.6.11-gentoo-r3 #3 Tue Apr 5 13:28:08 BST 2005 x86_64 AMD Athlon(tm) 64 Processor 3000+ AuthenticAMD GNU/Linux |
I'm just about to upgrade to 2.6.11-gentoo-r7, flash my BIOS, and upgrade the firmware on my NEC DVD-+RW drive, in an attempt to fix all my darned problems. If that doesn't fix it, I'm going to slap an Intel EtherExpressPro 100 network card in. That's got to be one network card which has no driver problems!
One day, my machine will just work without network and buffer underrun problems.... Grrrr
 |
|
| Back to top |
|
|
JuNix
Apprentice
Joined: 05 Mar 2003
Posts: 224
Location: Sheffield
|
| Posted: Sat May 28, 2005 8:43 am Post subject: |
|
|
| JuNix wrote: | | I'm just about to upgrade to 2.6.11-gentoo-r7, flash my BIOS |
Well that made things a thousand times worse. What was an intermittent problem that happened every couple of days, is a permanent problem which happens after 2 minutes of uptime.
I think I'll spend £20 on a PCI network card |
|
| Back to top |
|
|
JuNix
Apprentice
Joined: 05 Mar 2003
Posts: 224
Location: Sheffield
|
| Posted: Sat May 28, 2005 10:47 am Post subject: |
|
|
I've been scouring the internet for things to try to fix this forcedeth problem. I'll try to put everything I've found into here. If anybody knows anything else please feel free to add.
This bit is so that this post comes up when people search for "shuttle SN85G4 forcedeth problem"
I haven't tried any of this yet, so I'll let you know.
People say that it is better using forcedeth as a kernel module, instead of compiled in. People say this helps (?)
Also try disabling the timer interrupt
It's possible that you're running an old version of forcedeth. Look in /usr/src/linux/drivers/net/forcedeth.c and compare the version number with the latest release. The forcedeth site is at http://www.hailfinger.org/carldani/linux/patches/forcedeth/
How about this comment in forcedeth.c
| Code: | /*
* If your nic mysteriously hangs then try to reduce the limits
* to 1/0: It might be required to set NV_TX_LASTPACKET in the
* last valid ring entry. But this would be impossible to
* implement - probably a disassembly error.
*/
#define TX_LIMIT_STOP 63
#define TX_LIMIT_START 62 |
So how about changing these values and recompiling?
Try using nvnet (the closed source driver from nvidia) as a kernel module instead of forcedeth, but people say this still locks up.
Some people say that thie nVidia driver just doesn't work on their box at all
try from the command line, ip -s link, to get information about interface errors
emerge iptraf, to test your link under complete saturation.
[edit] try taking firewire support OUT from the kernel. |
|
| Back to top |
|
|
Konsti
l33t
Joined: 10 Dec 2002
Posts: 691
|
| Posted: Sun May 29, 2005 2:42 pm Post subject: |
|
|
I am actually going mad with this.
Yesterday I flashed the BIOs and indeed the ethernet Interface was unusable after that.
Completely broken. Most often the Connect and Speed LEDs on Interface and switch flashed on..off once a second. I considered declaring the hardware broken already.
Today I switch this beast on and Ethernet works. After a couple of minutes it broke again.
rmmod/modprobe did help once, after breaking again I needed to powercycle.
What the fuck? Is that hardware broken? How should I explain this the distributor in a warranty case so he could reproduce this? Why is rmmod/modprobe fixing hardware? Should I install windows to reproduce it there? I don't know...
In 2.6.12-rc4 there is Version 31.
How about this comment in forcedeth.c
| JuNix wrote: |
#define TX_LIMIT_STOP 63
#define TX_LIMIT_START 62
|
So this should become
| Code: |
#define TX_LIMIT_STOP 0
#define TX_LIMIT_START 1
|
or what? I will try that.
| JuNix wrote: |
Try using nvnet (the closed source driver from nvidia) as a kernel module instead of forcedeth, but people say this still locks up.
Some people say that thie nVidia driver just doesn't work on their box at all
|
This one breaks more often or completely indeed, tried it already. |
|
| Back to top |
|
|
JuNix
Apprentice
Joined: 05 Mar 2003
Posts: 224
Location: Sheffield
|
| Posted: Mon May 30, 2005 12:49 am Post subject: |
|
|
For F***s sake I couldn't get it to work. 
I am now the proud owner of one Intel PCI Etherexpress Pro/100 network card. That solved it 
I've put up with months of problems network-wise. I thought I had a faulty router. Now it's rock solid. ))) |
|
| Back to top |
|
|
Konsti
l33t
Joined: 10 Dec 2002
Posts: 691
|
| Posted: Mon May 30, 2005 12:24 pm Post subject: |
|
|
Damn! Yesterday the NIC run fine. I am soooo unsure if Hardware or Software is to blame, I am going mad.
I can't stick an an additional NIC 'cause the Heatsink of the GeForce is in the way *argh*
K |
|
| Back to top |
|
|
JuNix
Apprentice
Joined: 05 Mar 2003
Posts: 224
Location: Sheffield
|
| Posted: Mon May 30, 2005 1:44 pm Post subject: |
|
|
Konsti: Probably then you should downgrade the BIOS. I wish I'd made a backup of my BIOS but I don't have a FAT32 or NTFS partition to do it onto, so I just upgraded.
I've not researched it, but maybe ndiswrapper would let you load the Windows NDIS driver?
From the nVidia website..... http://download.nvidia.com/XFree86/nforce/amd64/1.0-0301/KnownProblems.html
| Code: | Network and other devices randomly stop working when ACPI is enabled
This problem may be caused by an incorrect ACPI table entry that causes the timer interrupt to be incorrectly configured.
If the kernel console boot trace (viewable using dmesg) contains messages such as these:
..MP-BIOS bug: 8254 timer not connected to IOAPIC
...trying to set up timer (IRQ0) through the 8259A . failed.
...trying to set up timer as Virtual Wire IRQ... failed.
...trying to set up timer as ExtINT IRQ... works.
then the incorrect ACPI table entry is present. On 2.6 kernels, this can be worked around by specifying the 'acpi_skip_timer_override' boot line option. An alternative workaround is to disable ACPI in the BIOS or by using the 'acpi=off' boot line option. |
How about getting the very latest shiny version of nvnet from http://www.nvidia.com/object/linux_nforce_amd64_1.0-0301.html
not that I've tried this myself though.....
Well, we're not on our owns - it doesn't work under FreeBSD either.
http://mail-index.netbsd.org/port-amd64/2003/12/15/0000.html
I hope something works for you Konsti |
|
| Back to top |
|
|
Konsti
l33t
Joined: 10 Dec 2002
Posts: 691
|
| Posted: Mon May 30, 2005 2:12 pm Post subject: |
|
|
Your Tips are interesting , thanks 
I will try them out. Reading that I assume the hardware is not broken, many people have these problems. I will try the ACPI stuff and again the nvnet Module.
But the patches on http://www.hailfinger.org/carldani/linux/patches/forcedeth/ are all older then the provided forcedeth.c in recent vanilla Kernels (.12 and .11 have version 31). |
|
| Back to top |
|
|
JuNix
Apprentice
Joined: 05 Mar 2003
Posts: 224
Location: Sheffield
|
|
| Back to top |
|
|
Konsti
l33t
Joined: 10 Dec 2002
Posts: 691
|
| Posted: Thu Jun 02, 2005 9:08 am Post subject: |
|
|
Yes, I updated the BIOS on saturday and told it was no help increasing stability.
I also upgraded the kernel to vanilla-2.6.12-rc5 this evening with no help, may be I did not powercycle while updating or such or I pinged the server to the wrong time, because its CPU broke int the time between Friday and saturday so it was probably unstable before (I pulled the fans last week from 12V to 7V with no hitch, but the waether went mad to the weekend, 20°C are normal here, it became 35°C). I was in a hurry and went drinking beer then
Whatsoever...
On sunday I fixed up my server and started to hack again and suspected the NIC to break after a couple of minutes copying a lot stoff over it. After that I wanted to test nvnet binary module, acpi settings, old kernels (i.e. 2.6.7), watch out for BIOS bugs...
You know what? The NIC did not break from Sunday until now once! WTF?
As I pointed out I am still totally unsure, if hardware or software is to blame and I wanted to find out, but starting investigating it went 100% stable? I dont get it. So may be 2.6.12-rc5 fixed stuff?
Due to the lack of time I was to wheenie to boot an old, known to break kernel yet :/
I don't believe really.
If hardware is to blame I suspect the connection between southbridge and PHY, I have possibilities to check that at work, but as said, it won't brake anymore, I coupied tons of iso images and movies and such...
I will refer to the fan stuff in the fan thread. |
|
| Back to top |
|
|
Konsti
l33t
Joined: 10 Dec 2002
Posts: 691
|
| Posted: Wed Jun 22, 2005 8:06 am Post subject: |
|
|
Well, my BB run stable for 2 Weeks, now it is getting nearly unuseable.
I have a question to the people having such issues:
Is the Problem reproduceable using the nvnet module from nvidia instead of forcedeth?
As a last ressort I am trying another OS (free bsd bootcd or opensolaris) to get a final answer if hardware is to blame....
regards, Konsti |
|
| Back to top |
|
|
|
Other Things Gentoo
