Re: eepro100 - Linux vs. FreeBSD

Re: eepro100 - Linux vs. FreeBSD

[Micah Gorrell]

As stated in a number of previous messages to this list many people have had
serious problems with the eepro100 driver in 2.4.  These problems where not
there in 2.2 and it is not a select few machines showing this so I very much
doubt that it is a configuration problem.  I assume that the intel driver
would prolly fix all of these issues but its not ready for 2.4 yet and its
not GPL so no one wants to use it.  If there is a good driver that is GPL'ed
lets use it.  I am not up to the task of porting it myself but I would be
glad to help in any way that I can.  I do write code, I'm just not familiar
enough with the linux kernel.

Micah
___
The irony is that Bill Gates claims to be making a stable operating system
and Linus Torvalds claims to be trying to take over the world
-----Original Message-----
From: "Richard B. Johnson" <root@chaos.analogic.com>
To: "Craig I. Hagan" <hagan@cih.com>
Cc: "Romain Kang" <romain@kzsu.stanford.edu>; <linux-kernel@vger.kernel.org>
Date: Monday, January 29, 2001 10:50 AM
Subject: Re: eepro100 - Linux vs. FreeBSD


>On Mon, 29 Jan 2001, Craig I. Hagan wrote:
>
>> > One approach to the endless eepro100 headaches would be to port
>> > the FreeBSD if_fxp driver to Linux.  After all, drivers have been
>> > ported between these OSs before; e.g., the aic7xxx SCSI adapter.
>> > However, I see no evidence that this has been attempted.  Can
>> > someone tell me what I'm obviously missing?
>>
>> Had I my druthers, i'd see the intel e100 driver brought into the kernel.
It
>> seems to work quite well with the eepro100 boards.
>>
>
>Two of my Linux machines use the Intel Ethernet controller on the
>motherboard. These are both SMP machines. I have never, ever, had
>any problems with the eepro100 driver that handles these chips.
>
>I spite of the fact that the driver loops in the ISR, and does other
>things that show poor design, it works so I have not done anything
>to it. "If it ain't broke, don't fix it..."
>
>So, if you have problems with using on-board Intel chip, it's
>unlikely that it's a driver problem. If you have cards on the PCI
>bus, the driver doesn't "know" any difference (PCI is PCI even if
>it's not in a connector). You may find that the problem is caused
>by PCI (mis)configuration since recent kernels use internal PCI
>code. You may find that some bus master device does not have its
>latency set correctly so it's taking over the bus. This can cause
>problems with any high-activity device on the bus, such as a
>network device.
>
>Cheers,
>Dick Johnson
>
>Penguin : Linux version 2.4.0 on an i686 machine (799.53 BogoMips).
>
>"Memory is like gasoline. You use it up when you are running. Of
>course you get it all back when you reboot..."; Actual explanation
>obtained from the Micro$oft help desk.
>
>
>-
>To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
>the body of a message to majordomo@vger.kernel.org
>Please read the FAQ at http://www.tux.org/lkml/
>
>

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/

Re: eepro100 - Linux vs. FreeBSD

[Andrey Savochkin]

On Mon, Jan 29, 2001 at 11:06:11AM -0700, Micah Gorrell wrote:
> As stated in a number of previous messages to this list many people have had
> serious problems with the eepro100 driver in 2.4.  These problems where not
> there in 2.2 and it is not a select few machines showing this so I very much
> doubt that it is a configuration problem.  I assume that the intel driver
> would prolly fix all of these issues but its not ready for 2.4 yet and its
[snip]

In the first place, the "no resource" problem is a hardware one.
As far as I understand, it's a buggy (or undocumented) timing requirement
for some revisions.
This problem showed with any kernel, 2.2 or 2.4, until a workaround was
developed.  On a single computer suffering from that problem it showed not on
every boot, but about in 30 percents.  That's why the reports were different.
So, the kernel version is irrelevant to this problem.

Best regards
					Andrey V.
					Savochkin
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/

Re: eepro100 - Linux vs. FreeBSD

[Micah Gorrell]

I haven't tried 2.4.1 yet but I will be soon (prolly today) and I will let
you know if I still see problems.

Micah
___
The irony is that Bill Gates claims to be making a stable operating system
and Linus Torvalds claims to be trying to take over the world
-----Original Message-----
From: "Davide Libenzi" <davidel@xmail.virusscreen.com>
To: "Micah Gorrell" <angelcode@myrealbox.com>; "Andrey Savochkin"
<saw@saw.sw.com.sg>
Cc: "Romain Kang" <romain@kzsu.stanford.edu>;
<linux-kernel@vger.kernel.org>; <root@chaos.analogic.com>; "Craig I. Hagan"
<hagan@cih.com>
Date: Tuesday, January 30, 2001 9:15 AM
Subject: Re: eepro100 - Linux vs. FreeBSD


>On Tuesday 30 January 2001 08:14, Micah Gorrell wrote:
>> I have been running 2.2 on many machines since its release and have
updated
>> to the latest version of 2.2 many times.  All of these machines have an
>> eepro100 and I never saw a single problem with any of them.  I updated
most
>> of my machines to 2.4 over the course of a week and within a day of
>> updating each of them showed the problem.  This may be pure chance but it
>> sounds to me as if it is a difference with the 2.4 kernel.
>
>I had the same problem on my dual PIII with a dual eepro100 NIC.
>2.4.1-pre12 solved the problem ( don't ask me why :) ).
>
>
>
>- Davide
>-
>To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
>the body of a message to majordomo@vger.kernel.org
>Please read the FAQ at http://www.tux.org/lkml/
>
>

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/

Re: eepro100 - Linux vs. FreeBSD

[Davide Libenzi]

On Tuesday 30 January 2001 08:14, Micah Gorrell wrote:
> I have been running 2.2 on many machines since its release and have updated
> to the latest version of 2.2 many times.  All of these machines have an
> eepro100 and I never saw a single problem with any of them.  I updated most
> of my machines to 2.4 over the course of a week and within a day of
> updating each of them showed the problem.  This may be pure chance but it
> sounds to me as if it is a difference with the 2.4 kernel.

I had the same problem on my dual PIII with a dual eepro100 NIC.
2.4.1-pre12 solved the problem ( don't ask me why :) ).



- Davide
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/

Re: eepro100 - Linux vs. FreeBSD

[Micah Gorrell]

I have been running 2.2 on many machines since its release and have updated
to the latest version of 2.2 many times.  All of these machines have an
eepro100 and I never saw a single problem with any of them.  I updated most
of my machines to 2.4 over the course of a week and within a day of updating
each of them showed the problem.  This may be pure chance but it sounds to
me as if it is a difference with the 2.4 kernel.

Micah
___
The irony is that Bill Gates claims to be making a stable operating system
and Linus Torvalds claims to be trying to take over the world
-----Original Message-----
From: "Andrey Savochkin" <saw@saw.sw.com.sg>
To: "Micah Gorrell" <angelcode@myrealbox.com>
Cc: "Romain Kang" <romain@kzsu.stanford.edu>;
<linux-kernel@vger.kernel.org>; <root@chaos.analogic.com>; "Craig I. Hagan"
<hagan@cih.com>
Date: Tuesday, January 30, 2001 12:35 AM
Subject: Re: eepro100 - Linux vs. FreeBSD


>On Mon, Jan 29, 2001 at 11:06:11AM -0700, Micah Gorrell wrote:
>> As stated in a number of previous messages to this list many people have
had
>> serious problems with the eepro100 driver in 2.4.  These problems where
not
>> there in 2.2 and it is not a select few machines showing this so I very
much
>> doubt that it is a configuration problem.  I assume that the intel driver
>> would prolly fix all of these issues but its not ready for 2.4 yet and
its
>[snip]
>
>In the first place, the "no resource" problem is a hardware one.
>As far as I understand, it's a buggy (or undocumented) timing requirement
>for some revisions.
>This problem showed with any kernel, 2.2 or 2.4, until a workaround was
>developed.  On a single computer suffering from that problem it showed not
on
>every boot, but about in 30 percents.  That's why the reports were
different.
>So, the kernel version is irrelevant to this problem.
>
>Best regards
> Andrey V.
> Savochkin
>
>

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/

Re: eepro100 - Linux vs. FreeBSD

[Richard B. Johnson]

On Mon, 29 Jan 2001, Sergey Kubushin wrote:

> On Mon, 29 Jan 2001, Richard B. Johnson wrote:
> 
> > Two of my Linux machines use the Intel Ethernet controller on the
> > motherboard. These are both SMP machines. I have never, ever, had
> > any problems with the eepro100 driver that handles these chips.
> >
> > I spite of the fact that the driver loops in the ISR, and does other
> > things that show poor design, it works so I have not done anything
> > to it. "If it ain't broke, don't fix it..."
> >
> > So, if you have problems with using on-board Intel chip, it's
> > unlikely that it's a driver problem. If you have cards on the PCI
> > bus, the driver doesn't "know" any difference (PCI is PCI even if
> > it's not in a connector). You may find that the problem is caused
> > by PCI (mis)configuration since recent kernels use internal PCI
> > code. You may find that some bus master device does not have its
> > latency set correctly so it's taking over the bus. This can cause
> > problems with any high-activity device on the bus, such as a
> > network device.
> 
> The older chips (e.g. 82557) work fine. The problem arises when you have the
> newer 82559's. They do work, however, if the power management for eepro100
> is enabled in kernel config. It definitely means that those chips are
> underinitialized (or overinitialized :)) when it's not.
> 
> ---

Ah HA! Thanks for helping to get the word out. So it's new new-fangled
EPA stuff that's mucking them up. I suppose if you save a microwatt
here and a microwatt there, eventually you are talking about keeping
California on-line ;).

grep CONFIG_EEPRO100  ./.config
CONFIG_EEPRO100=m
CONFIG_EEPRO100_PM=y

So those who are having problems should try turning on power managment
as above.

Cheers,
Dick Johnson

Penguin : Linux version 2.4.0 on an i686 machine (799.53 BogoMips).

"Memory is like gasoline. You use it up when you are running. Of
course you get it all back when you reboot..."; Actual explanation
obtained from the Micro$oft help desk.


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/

Re: eepro100 - Linux vs. FreeBSD

[Udo A. Steinberg]

Sergey Kubushin wrote:
> 
> The older chips (e.g. 82557) work fine. The problem arises when you have the
> newer 82559's. They do work, however, if the power management for eepro100
> is enabled in kernel config. It definitely means that those chips are
> underinitialized (or overinitialized :)) when it's not.

Andrey posted a patch last week, which obviously fixes the 82559 problems.
It's in Linus' latest 2.4.1-pre release too. I have an 82559 and with the
patch there've been no issues here yet - so things are looking good so far.

I suggest that instead of having 3 drivers (eepro100, e100, freebsd), people
should just work together, look at the goodies of each driver and merge them
into one perfect driver.

Regards,
-Udo.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/

Re: eepro100 - Linux vs. FreeBSD

[Sergey Kubushin]

On Mon, 29 Jan 2001, Richard B. Johnson wrote:

> Two of my Linux machines use the Intel Ethernet controller on the
> motherboard. These are both SMP machines. I have never, ever, had
> any problems with the eepro100 driver that handles these chips.
>
> I spite of the fact that the driver loops in the ISR, and does other
> things that show poor design, it works so I have not done anything
> to it. "If it ain't broke, don't fix it..."
>
> So, if you have problems with using on-board Intel chip, it's
> unlikely that it's a driver problem. If you have cards on the PCI
> bus, the driver doesn't "know" any difference (PCI is PCI even if
> it's not in a connector). You may find that the problem is caused
> by PCI (mis)configuration since recent kernels use internal PCI
> code. You may find that some bus master device does not have its
> latency set correctly so it's taking over the bus. This can cause
> problems with any high-activity device on the bus, such as a
> network device.

The older chips (e.g. 82557) work fine. The problem arises when you have the
newer 82559's. They do work, however, if the power management for eepro100
is enabled in kernel config. It definitely means that those chips are
underinitialized (or overinitialized :)) when it's not.

---
Sergey Kubushin				Sr. Unix Administrator
CyberBills, Inc.			Phone:	702-567-8857
874 American Pacific Dr,		Fax:	702-567-8890
Henderson, NV, 89014

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/

Re: eepro100 - Linux vs. FreeBSD

[Richard B. Johnson]

On Mon, 29 Jan 2001, Craig I. Hagan wrote:

> > One approach to the endless eepro100 headaches would be to port
> > the FreeBSD if_fxp driver to Linux.  After all, drivers have been
> > ported between these OSs before; e.g., the aic7xxx SCSI adapter.
> > However, I see no evidence that this has been attempted.  Can
> > someone tell me what I'm obviously missing?
> 
> Had I my druthers, i'd see the intel e100 driver brought into the kernel. It
> seems to work quite well with the eepro100 boards.
> 

Two of my Linux machines use the Intel Ethernet controller on the
motherboard. These are both SMP machines. I have never, ever, had
any problems with the eepro100 driver that handles these chips.

I spite of the fact that the driver loops in the ISR, and does other
things that show poor design, it works so I have not done anything
to it. "If it ain't broke, don't fix it..."

So, if you have problems with using on-board Intel chip, it's
unlikely that it's a driver problem. If you have cards on the PCI
bus, the driver doesn't "know" any difference (PCI is PCI even if
it's not in a connector). You may find that the problem is caused
by PCI (mis)configuration since recent kernels use internal PCI
code. You may find that some bus master device does not have its
latency set correctly so it's taking over the bus. This can cause
problems with any high-activity device on the bus, such as a
network device.

Cheers,
Dick Johnson

Penguin : Linux version 2.4.0 on an i686 machine (799.53 BogoMips).

"Memory is like gasoline. You use it up when you are running. Of
course you get it all back when you reboot..."; Actual explanation
obtained from the Micro$oft help desk.


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/

Re: eepro100 - Linux vs. FreeBSD

[Craig I. Hagan]

> One approach to the endless eepro100 headaches would be to port
> the FreeBSD if_fxp driver to Linux.  After all, drivers have been
> ported between these OSs before; e.g., the aic7xxx SCSI adapter.
> However, I see no evidence that this has been attempted.  Can
> someone tell me what I'm obviously missing?

Had I my druthers, i'd see the intel e100 driver brought into the kernel. It
seems to work quite well with the eepro100 boards.

-- craig

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/

eepro100 - Linux vs. FreeBSD

[Romain Kang]

Dumb question:

I've been following the freebsd-hackers list for a while, and in
that domain, the Intel NICs are the preferred interfaces because
they perform well and are very stable.

One approach to the endless eepro100 headaches would be to port
the FreeBSD if_fxp driver to Linux.  After all, drivers have been
ported between these OSs before; e.g., the aic7xxx SCSI adapter.
However, I see no evidence that this has been attempted.  Can
someone tell me what I'm obviously missing?

Romain Kang                             Disclaimer: I speak for myself alone,
romain@kzsu.stanford.edu                except when indicated otherwise.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/