Next-in-Thread Next Message

None A strange problem 

Forum: PCMCIA Installation and Configuration Issues
Re: Sad "Bad address mapping" on Sony N505VE (toby cabot)
Date: 2000, Jun 19
From: David Hinds <dhinds@pcmcia.sourceforge.org>

I don't understand how something like this could develop all of a
sudden.  Does it matter if you cold boot linux, versus warm booting
from windows?  If the reported bridge mapping address is 0x000000dc,
that would certainly be a problem; the BIOS is supposed to fill this
in, and it is putting it in space used by your system RAM.

The pcmcia_core module has a parameter, cb_mem_base, that lets you
override the bridge address selected by the BIOS.  You could try using
that as a workaround, assuming this isn't just a symptom of something
more serious.  My first guess was "hardware failure" until you said it
still works under Windows.

-- Dave

Next-in-Thread Next Message

Messages Inline: 0 1

Sad you're not kidding

Re: None A strange problem (David Hinds)
Date: 2000, Jun 20
From: toby cabot toby

Dave,

It sure is a strange problem. I can't for the life of me figure out what changed between "working fine" for the last five months since the machine was new and "FUBAR" now. I've tried warm-booting from Windows, cold boot, remove battery and power then re-boot, and nothing seems to work (except for win98, which works fine, which really pisses me off).

I scanned my messages log and found that the bridge mapping used to be 0x68000000 when things were working. So I tried setting the cb_mem_base to 0x68000000 but then get "Bad bridge mapping at 0x680000dc!".

I've also tried kernel 2.2.16 with pcmcia-cs 3.1.16 with the same results.

I looked for BIOS upgrades on the Sony website and no luck.

I guess I could try a pre-2.4 kernel but I'm not usually that cutting-edge.

Regards,
Toby

None Maybe a very particular hardware fault

Re: Sad you're not kidding (toby cabot)
Date: 2000, Jun 20
From: David Hinds <dhinds@pcmcia.sourceforge.org>

> I scanned my messages log and found that the bridge mapping used to be
> 0x68000000 when things were working. So I tried setting the cb_mem_base
> to 0x68000000 but then get "Bad bridge mapping at 0x680000dc!".

This smells like a hardware failure.  The PCMCIA drivers are writing
0x68000000 to a register, and it is reading back as 0x680000dc.  Maybe
Windows works because it only writes to this register and never relies
on reading it back.

Just for kicks, you could try editing modules/i82365.c, find the line
that says:

 pci_readl(s, PCI_BASE_ADDRESS_0, &s->cb_phys);

and add right after that:

 s->cb_phys &= ~0x0fff;

-- Dave

Agree looks like you're right

Re: None Maybe a very particular hardware fault (David Hinds)
Date: 2000, Jun 21
From: toby cabot toby

Dave,

Great suggestion! When we meet the beer is on me.

By hard-wiring i82365.c I'm now back online. I went for the full-bore hardcoded memory address:

s->cb_phys = 0x68000000;

Thanks,
Toby

PS. I'm also seeing the dropped ethernet interrupts after a suspend/resume. I'm not a driver programmer but if you need someone to guinea pig some new code on a N505VE please give me a shout.

A strange problem


Add Message to: "A strange problem"

Members Subscribe Admin Mode Show Frames Help for HyperNews at pcmcia-cs.sourceforge.net 1.10