Is there a problem with the forcedeth driver not being compatible with
the Asus K8N-VM motherboard? I installed Fedora Core 5 and the Ethernet
doesn't want to work. I installed the latest FC5 kernel which is some
flavor og 2.6.16 and it still doesn't work. The FC4 CD and rescue disk
don't work either. Windows XP however does work so I know that hardware
is good.
lspci says the hardware is an nVidia MCP51 ethernet controller. What am
I missing?
Thanks in advance.
Marc Perkel wrote:
> Is there a problem with the forcedeth driver not being compatible with
> the Asus K8N-VM motherboard? I installed Fedora Core 5 and the
> Ethernet doesn't want to work. I installed the latest FC5 kernel which
> is some flavor og 2.6.16 and it still doesn't work. The FC4 CD and
> rescue disk don't work either. Windows XP however does work so I know
> that hardware is good.
>
> lspci says the hardware is an nVidia MCP51 ethernet controller. What
> am I missing?
>
> Thanks in advance.
>
Also - why is it that I can load a 3 year old version of Windows XP on
this motherboad and it just work but I load a modern Linux Kernel and it
can't find the Ethernet card?
On Sat, 2006-05-27 23:07:39 -0700, Marc Perkel <[email protected]> wrote:
> Marc Perkel wrote:
[forcedeth not working correctly]
> Also - why is it that I can load a 3 year old version of Windows XP on
> this motherboad and it just work but I load a modern Linux Kernel and it
> can't find the Ethernet card?
That is either because Windows XP is totally superior to Linux, or
because the vendor (NVidia in this case) wrote a Windows driver, but
they didn't write something for Linux _and_ didn't publish any specs
for this ethernet controller. So it took some time to reverse
engineer the Windows driver...
MfG, JBG
--
Jan-Benedict Glaw [email protected] . +49-172-7608481 _ O _
"Eine Freie Meinung in einem Freien Kopf | Gegen Zensur | Gegen Krieg _ _ O
für einen Freien Staat voll Freier Bürger" | im Internet! | im Irak! O O O
ret = do_actions((curr | FREE_SPEECH) & ~(NEW_COPYRIGHT_LAW | DRM | TCPA));
On Sunday 28 May 2006 06:11, Marc Perkel wrote:
> Is there a problem with the forcedeth driver not being compatible with
> the Asus K8N-VM motherboard? I installed Fedora Core 5 and the Ethernet
> doesn't want to work. I installed the latest FC5 kernel which is some
> flavor og 2.6.16 and it still doesn't work. The FC4 CD and rescue disk
> don't work either. Windows XP however does work so I know that hardware
> is good.
>
> lspci says the hardware is an nVidia MCP51 ethernet controller. What am
> I missing?
A decent bug report. Please go to http://bugzilla.kernel.org/ and publish your
lspci, /proc/interrupts and dmesg and if you're sure it's forcedeth, put the
maintainer on CC.
--
Cheers,
Alistair.
Third year Computer Science undergraduate.
1F2 55 South Clerk Street, Edinburgh, UK.
On Sunday 28 May 2006 18:54, Alistair John Strachan wrote:
> On Sunday 28 May 2006 06:11, Marc Perkel wrote:
> > Is there a problem with the forcedeth driver not being compatible with
> > the Asus K8N-VM motherboard? I installed Fedora Core 5 and the Ethernet
> > doesn't want to work. I installed the latest FC5 kernel which is some
> > flavor og 2.6.16 and it still doesn't work. The FC4 CD and rescue disk
> > don't work either. Windows XP however does work so I know that hardware
> > is good.
> >
> > lspci says the hardware is an nVidia MCP51 ethernet controller. What am
> > I missing?
>
> A decent bug report. Please go to http://bugzilla.kernel.org/ and publish
> your lspci, /proc/interrupts and dmesg and if you're sure it's forcedeth,
> put the maintainer on CC.
Five minutes of research and I found this:
http://www.asus.com.tw/products4.aspx?modelmenu=2&model=952&l1=3&l2=14&l3=245
It indicates that the board, although using an nForce 410 chipset, does NOT
use NVIDIA ethernet, but (quote) "Relatek RTL8201CL external PHY"; you're
probably trying the wrong driver and this is the source of the problem.
This info could be wrong, but I thought it might help.
--
Cheers,
Alistair.
Third year Computer Science undergraduate.
1F2 55 South Clerk Street, Edinburgh, UK.
On Sun, May 28, 2006 at 07:20:02PM +0100, Alistair John Strachan wrote:
> Five minutes of research and I found this:
>
> http://www.asus.com.tw/products4.aspx?modelmenu=2&model=952&l1=3&l2=14&l3=245
>
> It indicates that the board, although using an nForce 410 chipset, does NOT
> use NVIDIA ethernet, but (quote) "Relatek RTL8201CL external PHY"; you're
> probably trying the wrong driver and this is the source of the problem.
>
> This info could be wrong, but I thought it might help.
The linux driver is for the MAC, which is in the chipset. The PHY is
external, and it doens't particularly matter which one it is in general.
The driver almost never has to care. I know I don't have to tell the
pcnet32 driver that I am using a broadcom 5221 PHY with it. It works
just fine because the MAC is an AMD 972 which is one fo the chips the
pcnet32 driver operates.
Len Sorensen
On Monday 29 May 2006 15:14, Lennart Sorensen wrote:
> On Sun, May 28, 2006 at 07:20:02PM +0100, Alistair John Strachan wrote:
> > Five minutes of research and I found this:
> >
> > http://www.asus.com.tw/products4.aspx?modelmenu=2&model=952&l1=3&l2=14&l3
> >=245
> >
> > It indicates that the board, although using an nForce 410 chipset, does
> > NOT use NVIDIA ethernet, but (quote) "Relatek RTL8201CL external PHY";
> > you're probably trying the wrong driver and this is the source of the
> > problem.
> >
> > This info could be wrong, but I thought it might help.
>
> The linux driver is for the MAC, which is in the chipset. The PHY is
> external, and it doens't particularly matter which one it is in general.
> The driver almost never has to care. I know I don't have to tell the
> pcnet32 driver that I am using a broadcom 5221 PHY with it. It works
> just fine because the MAC is an AMD 972 which is one fo the chips the
> pcnet32 driver operates.
Since this document reveals no information about such a MAC device, it is
still unknown whether forcedeth is truly to blame. From the silence of the
original poster I'm inclined to assume he's fixed the problem..
I was suspicious because in contra to the original post, it is NOT the case
that Windows XP (even SP2) will automatically detect and support NVIDIA
ethernet; a third party driver must be installed before the device is usable.
Realtek chips, however, tend to work out of the box..
--
Cheers,
Alistair.
Third year Computer Science undergraduate.
1F2 55 South Clerk Street, Edinburgh, UK.
On Mon, May 29, 2006 at 05:46:01PM +0100, Alistair John Strachan wrote:
> Since this document reveals no information about such a MAC device, it is
> still unknown whether forcedeth is truly to blame. From the silence of the
> original poster I'm inclined to assume he's fixed the problem..
Well the web page for the mainboard does say Integrated 10/100 Mb MAC,
which would be part of the nvidia chipset. Being 100Mbit certainly
means it should work with forcedeth. At least if the kernel is new
enough to recognize the chipset.
> I was suspicious because in contra to the original post, it is NOT the case
> that Windows XP (even SP2) will automatically detect and support NVIDIA
> ethernet; a third party driver must be installed before the device is usable.
> Realtek chips, however, tend to work out of the box..
Well I don't know if XP added support for nvidia's network feature in
any of the service packs. I do know it works with forcedeth on all the
nvidia boards I haved used, although I have only used their 100Mbit
versions. There seemed to be a problem with 2.6.8 on some boards where
the nvidia MAC would not be detected, but manually loading forcedeth
worked anyhow. Later kernels seemed to have no such problem.
Len Sorensen
Jan-Benedict Glaw wrote:
> On Sat, 2006-05-27 23:07:39 -0700, Marc Perkel <[email protected]> wrote:
>> Marc Perkel wrote:
> [forcedeth not working correctly]
>> Also - why is it that I can load a 3 year old version of Windows XP on
>> this motherboad and it just work but I load a modern Linux Kernel and it
>> can't find the Ethernet card?
>
> That is either because Windows XP is totally superior to Linux, or
> because the vendor (NVidia in this case) wrote a Windows driver, but
> they didn't write something for Linux _and_ didn't publish any specs
> for this ethernet controller. So it took some time to reverse
> engineer the Windows driver...
>
> MfG, JBG
>
I can concur that the forcedeth is unreliable on nvidia based motherboards.
I have a ethernet device that works with forcedeth.
0000:00:0a.0 Bridge: nVidia Corporation CK804 Ethernet Controller (rev a3)
0000:00:0a.0 0680: 10de:0057 (rev a3)
Subsystem: 147b:1c12
Flags: 66MHz, fast devsel, IRQ 11
Memory at fe02f000 (32-bit, non-prefetchable) [size=4K]
I/O ports at fc00 [size=8]
Capabilities: [44] Power Management version 2
It works in that it can actually send and receive packets.
The problems are:
1) one cannot rmmod the forcedeth module. Even after ifdown etc.
2) the machine hangs randomly.
Before someone asks, the MB has no serial port, so no stack trace available.
3) The netconsole fails to function with it.
I have installed a standard PCI based intel ethernet card, and only use
that. Without the forcedeth loaded, no hangs since.
So, although I can confirm that there are certainly problems with the
forcedeth driver, without a serial port, I am at a loss at how I might
help diagnose the problem and fix it.
The only help this email is, is to confirm that if someone is raising
problems with the forcedeth driver, it is most likely to be a truthful
report and if they also have a serial port, they might even be able to
give some good diagnosis output.
James
On Monday 29 May 2006 18:48, James Courtier-Dutton wrote:
[snip irrelevant Other OS discussion]
>
> I can concur that the forcedeth is unreliable on nvidia based motherboards.
> I have a ethernet device that works with forcedeth.
> 0000:00:0a.0 Bridge: nVidia Corporation CK804 Ethernet Controller (rev a3)
> 0000:00:0a.0 0680: 10de:0057 (rev a3)
> Subsystem: 147b:1c12
> Flags: 66MHz, fast devsel, IRQ 11
> Memory at fe02f000 (32-bit, non-prefetchable) [size=4K]
> I/O ports at fc00 [size=8]
> Capabilities: [44] Power Management version 2
>
> It works in that it can actually send and receive packets.
> The problems are:
> 1) one cannot rmmod the forcedeth module. Even after ifdown etc.
> 2) the machine hangs randomly.
> Before someone asks, the MB has no serial port, so no stack trace
> available. 3) The netconsole fails to function with it.
File a bug report. What kernel version did you last test?
I have that exact ethernet controller (CK804) and cannot reproduce any of the
problems you have listed on 2.6.17-rc5.
--
Cheers,
Alistair.
Third year Computer Science undergraduate.
1F2 55 South Clerk Street, Edinburgh, UK.
Hi James,
On Monday, 29. May 2006 19:48, James Courtier-Dutton wrote:
> I can concur that the forcedeth is unreliable on nvidia based motherboards.
> I have a ethernet device that works with forcedeth.
> 0000:00:0a.0 Bridge: nVidia Corporation CK804 Ethernet Controller (rev a3)
> 0000:00:0a.0 0680: 10de:0057 (rev a3)
> Subsystem: 147b:1c12
> Flags: 66MHz, fast devsel, IRQ 11
> Memory at fe02f000 (32-bit, non-prefetchable) [size=4K]
> I/O ports at fc00 [size=8]
> Capabilities: [44] Power Management version 2
>
> It works in that it can actually send and receive packets.
> The problems are:
> 1) one cannot rmmod the forcedeth module. Even after ifdown etc.
> 2) the machine hangs randomly.
> Before someone asks, the MB has no serial port, so no stack trace available.
> 3) The netconsole fails to function with it.
>
> I have installed a standard PCI based intel ethernet card, and only use
> that. Without the forcedeth loaded, no hangs since.
>
> So, although I can confirm that there are certainly problems with the
> forcedeth driver, without a serial port, I am at a loss at how I might
> help diagnose the problem and fix it.
This sounds like http://bugzilla.kernel.org/show_bug.cgi?id=6480
Maybe we can help to resolve this. I already stored a lot of info in
the bugzilla entry.
Could you please describe your setup further? I'm interested on
the devices behind your nvidia NIC, since we only have it at one customers
mail server behind a SonicWall, so we cannot really try a lot.
For all other customers it works. I'm a bit lost on the cause.
I've seen 4% collisions, which might reduce performance,
but should not stop the transmitter forever.
PS: CC'ed some more relevant addresses to increase awareness.
Regards
Ingo Oeser
Marc Perkel wrote:
> Is there a problem with the forcedeth driver not being compatible with
> the Asus K8N-VM motherboard? I installed Fedora Core 5 and the Ethernet
> doesn't want to work. I installed the latest FC5 kernel which is some
> flavor og 2.6.16 and it still doesn't work. The FC4 CD and rescue disk
> don't work either. Windows XP however does work so I know that hardware
> is good.
>
> lspci says the hardware is an nVidia MCP51 ethernet controller. What am
> I missing?
I had similar issues on A8N-VM. The device would stop working after some
packets initially got through and would not recover.
As I recall 'noapic' made it work reliably. nvidia has a linux binary
driver for the chipset which you can try (you probably already use the
binary video driver?).
--
Eyal Lebedinsky ([email protected]) <http://samba.org/eyal/>
attach .zip as .dat
> Is there a problem with the forcedeth driver not being compatible with the
> Asus K8N-VM motherboard? I installed Fedora Core 5 and the Ethernet doesn't
I have the ASUS K8N-VM CSM here(different variant), and it works "fine"
with the forcedeth driver. Throughput is fairly slow though, I average
only about 25 meg/second with scp, on an Athlon64 3700+ talking to a dual
opteron 265.
Also, jumbo frames are apparently not supported.
I may add on a pci-e network card at some point.
Mike
Kindly ignore my previous posting, I have the A8N-VM CSM. Though it's
still an MCP51 on the board, it is a gigabit controller.
James Courtier-Dutton wrote:
> I can concur that the forcedeth is unreliable on nvidia based motherboards.
> I have a ethernet device that works with forcedeth.
> 0000:00:0a.0 Bridge: nVidia Corporation CK804 Ethernet Controller (rev a3)
> 0000:00:0a.0 0680: 10de:0057 (rev a3)
> Subsystem: 147b:1c12
> Flags: 66MHz, fast devsel, IRQ 11
> Memory at fe02f000 (32-bit, non-prefetchable) [size=4K]
> I/O ports at fc00 [size=8]
> Capabilities: [44] Power Management version 2
>
> It works in that it can actually send and receive packets.
> The problems are:
> 1) one cannot rmmod the forcedeth module. Even after ifdown etc.
> 2) the machine hangs randomly.
> Before someone asks, the MB has no serial port, so no stack trace available.
> 3) The netconsole fails to function with it.
>
> I have installed a standard PCI based intel ethernet card, and only use
> that. Without the forcedeth loaded, no hangs since.
>
> So, although I can confirm that there are certainly problems with the
> forcedeth driver, without a serial port, I am at a loss at how I might
> help diagnose the problem and fix it.
Watchdog timer? Compile with the appropriate timer(s) and set it with
wdog or similar. Of course it may not be "hung enough" but it's worth a
try, I would think.
>
> The only help this email is, is to confirm that if someone is raising
> problems with the forcedeth driver, it is most likely to be a truthful
> report and if they also have a serial port, they might even be able to
> give some good diagnosis output.
--
Bill Davidsen <[email protected]>
Obscure bug of 2004: BASH BUFFER OVERFLOW - if bash is being run by a
normal user and is setuid root, with the "vi" line edit mode selected,
and the character set is "big5," an off-by-one errors occurs during
wildcard (glob) expansion.