<!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/3.2.5">
</HEAD>
<BODY>
On Mon, 2006-03-06 at 10:51 +0000, abraham.manuatg wrote:
<BLOCKQUOTE TYPE=CITE>
<PRE>
<FONT COLOR="#000000">Matt Pratt wrote:</FONT>
<FONT COLOR="#000000">> On 2/27/06, *Matt Pratt* <<A HREF="mailto:mattpratt.au@gmail.com">mattpratt.au@gmail.com</A> </FONT>
<FONT COLOR="#000000">> <mailto:<A HREF="mailto:mattpratt.au@gmail.com">mattpratt.au@gmail.com</A>>> wrote:</FONT>
<FONT COLOR="#000000">></FONT>
<FONT COLOR="#000000">></FONT>
<FONT COLOR="#000000">> Non-working case:</FONT>
<FONT COLOR="#000000">></FONT>
<FONT COLOR="#000000">> bttv0: Bt878 (rev 17) at 0000:02:0c.0, irq: 169, latency: 132,</FONT>
<FONT COLOR="#000000">> mmio: 0xf9000000</FONT>
<FONT COLOR="#000000">> bttv0: using: *** UNKNOWN/GENERIC *** [card=0,autodetected]</FONT>
<FONT COLOR="#000000">> bttv0: gpio: en=00000000, out=00000000 in=00fffffe [init]</FONT>
<FONT COLOR="#000000">></FONT>
<FONT COLOR="#000000">> Which seems to indicate that the pci_read_config_word() of</FONT>
<FONT COLOR="#000000">> PCI_SUBSYSTEM_ID and PCI_SUBSYSTEM_VENDOR_ID is failing.</FONT>
<FONT COLOR="#000000">></FONT>
<FONT COLOR="#000000">> Does this indicate larger PCI problems, like a conflict between</FONT>
<FONT COLOR="#000000">> the boards?</FONT>
<FONT COLOR="#000000">></FONT>
<FONT COLOR="#000000">> Can some ACPI fiddling help here?</FONT>
<FONT COLOR="#000000">></FONT>
<FONT COLOR="#000000">></FONT>
<FONT COLOR="#000000">></FONT>
<FONT COLOR="#000000">> So even lspci -x was showing that the Subsystem Id and Subsustem </FONT>
<FONT COLOR="#000000">> Vendor Id were being cleared to zeros in the PCI config space of this </FONT>
<FONT COLOR="#000000">> card. But only when the two cards were in the machine. No amount of </FONT>
<FONT COLOR="#000000">> slot swapping seemed to make a difference. Upgrade of motherboard </FONT>
<FONT COLOR="#000000">> firmware made no difference. noacpi, lacpi, pci=routeirq, pci=noacpi </FONT>
<FONT COLOR="#000000">> made no difference. Went to the effort of putting the cards into </FONT>
<FONT COLOR="#000000">> another machine. They just worked. Eventually tried acpi=off (which I </FONT>
<FONT COLOR="#000000">> thought noacpi would have done the same thing) and they both worked!</FONT>
<FONT COLOR="#000000">></FONT>
<FONT COLOR="#000000">It all sounds very scary/crazy. If you can get in some logs on that, </FONT>
<FONT COLOR="#000000">lspci -vvv in both cases would be a bit interesting.</FONT>
<FONT COLOR="#000000">Manu</FONT>
</PRE>
</BLOCKQUOTE>
<TABLE CELLSPACING="0" CELLPADDING="0" WIDTH="100%">
<TR>
<TD>
-- <BR>
Martin Petrovsky <<A HREF="mailto:mpetrovsky@swiftdsl.com.au">mpetrovsky@swiftdsl.com.au</A>>
</TD>
</TR>
</TABLE>
</BODY>
</HTML>