Date: Wed, 17 Mar 1999 01:20:03 -0500 (EST) From: "John W. DeBoskey" <jwd@unx.sas.com> To: wes@softweyr.com (Wes Peters), freebsd-hackers@freebsd.org Cc: sascht@unx.sas.com (Charles Tudor) Subject: Re: fxp driver causing lockup Message-ID: <199903170620.BAA54317@bb01f39.unx.sas.com> In-Reply-To: From Wes Peters at "Mar 16, 1999 7:39:43 am"
next in thread | raw e-mail | index | archive | help
Hi,
I beleive you are correct. The following code fragment from
if_fxp.c is the lockup:
/*
* Start the config command/DMA.
*/
fxp_scb_wait(sc);
CSR_WRITE_4(sc, FXP_CSR_SCB_GENERAL, vtophys(&cbp->cb_status));
CSR_WRITE_1(sc, FXP_CSR_SCB_COMMAND, FXP_SCB_COMMAND_CU_START);
/* ...and wait for it to complete. */
while (!(cbp->cb_status & FXP_CB_STATUS_C));
according to doc at Dell's website (they have a machine with this
problem), the bios is not enabling dma on the adapter board. I beleive
the person I copied on this mail can provide the precise web page
that documents this.
According to them, it is possible to discreetly turn on dma on
the adapter. This is were someone with the appropriate doc needs
to jump in...
Basically, the driver needs to check that dma is enabled and
enable it if not.
Just my 0.02 cents,
John
> Andy Doran wrote:
> >
> > Hi all,
> > I've got weird problem with an Intel Etherexpress Pro 100. The cards work
> > fine, in fact great, with most of my FreeBSD boxes, all of which run
> > the 3.x branch updated weekly over cvsup.
> >
> > The trouble is, when I do an 'ifconfig fxp0 $some_address' on one
> > particular machine (a Tulip P-II, real junk) the machine locks up. The
> > keyboard still works (i.e. can toggle numlock light), and trying to switch
> > to another VT causes the machine to generate a never ending beep. I guess
> > it's getting stuck at spl???() somewhere.
> >
> > Setting the media type works, and the card works fine on other
> > machines. I've tried changing the slot, and using a generic 3.x kernel
> > too, but no luck. A 2.2.6 kernel works fine, no lockups. This card shares
> > an IRQ with a 3c905 and a S3 ViRGE/GX2.
> >
> > Unfortunately I can't hack at this because it's a production box, and has
> > to be up 24/7 almost always. Any ideas?
>
> I've seen the same problem on a Toshiba Equuium with on-board Pro 100+.
> The failure happens when the first DMA in the attach routine never
> completes. I'll be looking into it this week; I finally have some time
> to debug this problem and two other network cards for when it the machine
> has to work.
>
> I've saved a couple of recent messages about BIOS not initializing PCI
> cards; I'm wondering if that might be the cause of the problem. This
> machine worked fine with 2.2.x, so I'll be looking at what changed
> between 2.2.7 and 3.0/3.1.
>
> - --
> "Where am I, and what am I doing in this handbasket?"
>
> Wes Peters Softweyr LLC
> http://www.softweyr.com/~softweyr wes@softweyr.com
>
>
> To Unsubscribe: send mail to majordomo@FreeBSD.org
> with "unsubscribe freebsd-hackers" in the body of the message
>
> ------------------------------
>
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199903170620.BAA54317>