linux-mips
[Top] [All Lists]

Re: i/o and memory space enable bits in PCI-PCI bridge

To: Pavel Kiryukhin <savl@dev.rtsoft.ru>
Subject: Re: i/o and memory space enable bits in PCI-PCI bridge
From: Ralf Baechle <ralf@linux-mips.org>
Date: Tue, 7 Dec 2004 21:54:17 +0100
Cc: linux-mips@linux-mips.org
In-reply-to: <41B608FD.7070209@dev.rtsoft.ru>
Original-recipient: rfc822;linux-mips@linux-mips.org
References: <41B608FD.7070209@dev.rtsoft.ru>
Sender: linux-mips-bounce@linux-mips.org
User-agent: Mutt/1.4.1i
On Tue, Dec 07, 2004 at 10:48:13PM +0300, Pavel Kiryukhin wrote:

> Can somebody give me a hint:  - what part of 2.6 (mips) code is 
> responsible for setting i/o and memory space enable bits in PCI-PCI 
> bridge config. space command register?
> On my board those bits are not set after bridge is configured.
> Currently I'm using the following change in "pcibios_enable_resources" 
> to work with devices behind the bridge.
> --- arch/mips/pci/pci.c_org    2004-12-06 18:20:50.000000000 +0300
> +++ arch/mips/pci/pci.c    2004-12-06 18:21:22.000000000 +0300
> @@ -164,7 +164,7 @@
> 
>    pci_read_config_word(dev, PCI_COMMAND, &cmd);
>    old_cmd = cmd;
> -    for(idx=0; idx<6; idx++) {
> +    for(idx=0; idx<=PCI_BRIDGE_RESOURCES; idx++) {
>        /* Only set up the requested stuff */
>        if (!(mask & (1<<idx)))
>            continue;
> 
> but I think there should be some legal way I missed.

We had a discussion a while ago where somebody did suggest an afair
identical patch a while ago.  The problem with this patch is it fixes
things for a few platforms and breaks things for a few others.

  Ralf

<Prev in Thread] Current Thread [Next in Thread>