[tulip] Tulip IRQ problems on FIC / Gericom laptops solved

Luis Falcon lfalcon@thymbra.com
Thu Apr 11 01:08:02 2002


--=-HzglFk/ZAchIgq4rXU0O
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

Many thanks to Dominik Brodowski and Brian Rittmeyer who pointed us and
developed a patch for this problem !

I have to say that in order to get it running w/o any problems, I had to
recompile the kernel without the cardbus controller, since I got random
problems there (both shared the same IRQ and they were not happy.)=20

Dominik pointed that might be a IRQ balancing problem, and at this
point, I don't know what's the status on the driver.


Regards,
Luis



On Fri, 2002-04-05 at 17:10, Arndt Schoenewald wrote:

    Hi all,
   =20
    there were a couple of postings during the last few months from people
    trying to get the builtin Accton EN2242 network card of their Gericom
    (1st Supersonic M6-T, Supersonic GPRS, Supersonic2) or FIC (A360, A380)
    laptops to work. (The same problems were observed by people who owned
    other OEM versions of this laptop, such as the Medion MD 9703, which ar=
e
    equipped with a Realtek 8139 chip.)
   =20
    The symptoms were that the NIC either did not get an IRQ assigned at al=
l,
    or (with an ACPI kernel patch) it would get an IRQ, but transmitting an=
d
    receiving packets would still fail, with driver error messages like:
   =20
        NETDEV WATCHDOG: eth0: transmit timed out
        eth0: Transmit timed out, status fc67c017, CSR12 00000000, resettin=
g...
   =20
    Well, I have the pleasure to announce that these problems are finally
    resolved in the latest ACPI kernel patch acpi-20020404, which can now
    be obtained from http://sourceforge.net/projects/acpi/ for the kernel
    versions 2.4.18 and 2.5.7.
   =20
    The problems were not related to the tulip driver, they were caused by=20
    improper IRQ activation.
   =20
    Best regards,
    Arndt
   =20
   =20
    Q: Why doesn't the Accton EN2242 network card get an IRQ with stock Lin=
ux
       kernels?
   =20
    A: These laptops totally rely on ACPI for configuration (the BIOS does
       not have a valid PCI IRQ routing table), and the ACPI implementation
       of current standard Linux kernels is not fully functional yet.
   =20
    Q: Why did it work with the acpi-20011218 kernel patch (plus updates),
       but not with later versions?
   =20
    A: Patch acpi-20011218 did not yet include IRQ assignment; this feature
       was contributed by Kai Germaschewski through of a patch posted to th=
e
       ACPI list in January 2002. With these patches the laptops worked fin=
e,
       but unfortunately a subtile aspect of Kai's patch got lost during it=
s
       integration into the ACPI mainline (the call to eisa_set_level_irq()=
).
       Many thanks to Dominik Brodowski for tracking this problem down and
       fixing it!
   =20
    --=20
     /////    Quelltext AG -- Professional Software Services
    //   //   Arndt Sch=F6newald <arndt@quelltext.com>, CEO
    //   //   Ostenhellweg 31, 44135 Dortmund, Germany
    //  \\/   Tel +49 231 9503750, Fax +49 231 9503751
     ////\\   Web http://quelltext.com
    _______________________________________________
    tulip mailing list, tulip@scyld.com
    To change to digest mode or unsubscribe visit

http://www.scyld.com/mailman/listinfo/tulip

   =20

--=-HzglFk/ZAchIgq4rXU0O
Content-Type: text/html; charset=utf-8

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 TRANSITIONAL//EN">
<HTML>
<HEAD>
  <META HTTP-EQUIV="Content-Type" CONTENT="text/html; CHARSET=UTF-8">
  <META NAME="GENERATOR" CONTENT="GtkHTML/1.0.1">
</HEAD>
<BODY>
Many thanks to Dominik Brodowski and Brian Rittmeyer who pointed us and developed a patch for this problem !
<BR>

<BR>
I have to say that in order to get it running w/o any problems, I had to recompile the kernel without the cardbus controller, since I got random problems there (both shared the same IRQ and they were not happy.) 
<BR>

<BR>
Dominik pointed that might be a IRQ balancing problem, and at this point, I don't know what's the status on the driver.
<BR>

<BR>

<BR>
Regards,
<BR>
Luis
<BR>

<BR>

<BR>

<BR>
On Fri, 2002-04-05 at 17:10, Arndt Schoenewald wrote:
    <BLOCKQUOTE>
<PRE><FONT COLOR="#737373"><FONT SIZE="3"><I>Hi all,</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I></FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>there were a couple of postings during the last few months from people</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>trying to get the builtin Accton EN2242 network card of their Gericom</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>(1st Supersonic M6-T, Supersonic GPRS, Supersonic2) or FIC (A360, A380)</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>laptops to work. (The same problems were observed by people who owned</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>other OEM versions of this laptop, such as the Medion MD 9703, which are</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>equipped with a Realtek 8139 chip.)</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I></FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>The symptoms were that the NIC either did not get an IRQ assigned at all,</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>or (with an ACPI kernel patch) it would get an IRQ, but transmitting and</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>receiving packets would still fail, with driver error messages like:</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I></FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>    NETDEV WATCHDOG: eth0: transmit timed out</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>    eth0: Transmit timed out, status fc67c017, CSR12 00000000, resetting...</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I></FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>Well, I have the pleasure to announce that these problems are finally</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>resolved in the latest ACPI kernel patch acpi-20020404, which can now</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>be obtained from http://sourceforge.net/projects/acpi/ for the kernel</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>versions 2.4.18 and 2.5.7.</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I></FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>The problems were not related to the tulip driver, they were caused by </FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>improper IRQ activation.</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I></FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>Best regards,</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>Arndt</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I></FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I></FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>Q: Why doesn't the Accton EN2242 network card get an IRQ with stock Linux</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>   kernels?</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I></FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>A: These laptops totally rely on ACPI for configuration (the BIOS does</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>   not have a valid PCI IRQ routing table), and the ACPI implementation</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>   of current standard Linux kernels is not fully functional yet.</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I></FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>Q: Why did it work with the acpi-20011218 kernel patch (plus updates),</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>   but not with later versions?</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I></FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>A: Patch acpi-20011218 did not yet include IRQ assignment; this feature</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>   was contributed by Kai Germaschewski through of a patch posted to the</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>   ACPI list in January 2002. With these patches the laptops worked fine,</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>   but unfortunately a subtile aspect of Kai's patch got lost during its</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>   integration into the ACPI mainline (the call to eisa_set_level_irq()).</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>   Many thanks to Dominik Brodowski for tracking this problem down and</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>   fixing it!</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I></FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>-- </FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I> /////    Quelltext AG -- Professional Software Services</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>//   //   Arndt Sch&#246;newald &lt;arndt@quelltext.com&gt;, CEO</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>//   //   Ostenhellweg 31, 44135 Dortmund, Germany</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>//  \\/   Tel +49 231 9503750, Fax +49 231 9503751</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I> ////\\   Web http://quelltext.com</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>_______________________________________________</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>tulip mailing list, tulip@scyld.com</FONT></FONT></I>
<FONT COLOR="#737373"><FONT SIZE="3"><I>To change to digest mode or unsubscribe visit</FONT></FONT></I></PRE>
    </BLOCKQUOTE>
<A HREF="http://www.scyld.com/mailman/listinfo/tulip"><FONT SIZE="3"><I>http://www.scyld.com/mailman/listinfo/tulip</FONT></I></A>
    <BLOCKQUOTE>
<PRE></PRE>
    </BLOCKQUOTE>
</BODY>
</HTML>

--=-HzglFk/ZAchIgq4rXU0O--