Date: Thu, 3 Jun 1999 13:48:25 -0400 (EDT) From: Larry Lile <lile@stdio.com> To: Phil Regnauld <regnauld@ftf.net> Cc: freebsd-tokenring@FreeBSD.ORG Subject: Re: Strange behavior with OC-3117, 4.0-C Message-ID: <Pine.BSF.4.05.9906031339020.10273-100000@heathers.stdio.com> In-Reply-To: <19990603134535.55646@ns.int.ftf.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, 3 Jun 1999, Phil Regnauld wrote:
> Wood, Richard writes:
> > > -----Original Message-----
> > > From: Phil Regnauld [mailto:regnauld@ftf.net]
> > > Sent: Friday, May 28, 1999 8:57 AM
> > > To: freebsd-tokering@freebsd.org
> > > Subject: Strange behavior with OC-3117, 4.0-C
> >
> > <huge snip>
<more snippage>
> > to restart it, a powerdown was the only option.
>
> I solved this problem by reinstalling 19990421-CURRENT --
> something between 19990421 and 19990530 had broken the DMA or
> something else.
>
> Larry Lile said it might be the new bus code.
Sorry, I was being paranoid. It doesn't look like that had any effect.
> My adapter works fine now, except for regular
> receive error(3)
> messages on the console (and corresponding errors in netstat -i)
Maybe there was some other change to the IP stack? I don't see any
obvious changes in the driver code that would have caused this to be
any more broken than it already was. There is a problem that I need
to address, and will. I still don't understand why things are happening
the way they are.
For example:
ftp works fine, mostly 1 or 2 mbufs in each packet chain. Most telnet
packets are the same 1 or 2 mbufs in each packet. However as a couple
of people have noted if you do a "ls" that returns a lot of data it will
hang and this is becuase I am getting roughly 1500 bytes of data
(perfectly normal and acceptable) but it is being sent in a chain of 12
to 36 mbufs! I do not understand why this data has not been collapsed
into a mbuf (and or a mbuf cluster) at some point further into the stack.
I don't beleive that it has always been this way previously.
Never the less, I have a bug and I will fix it. Could someone look into
what might have changed in the IP stack? I am really pressed for time
right now.
Larry Lile
lile@stdio.com
[Changes to if_oltr.c between 19990421 and -current]
*** if_oltr.c Wed Mar 10 12:45:26 1999
--- /sys/contrib/dev/oltr/if_oltr.c Tue May 18 21:57:52 1999
***************
*** 30,36 ****
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
! * $Id: if_oltr.c,v 1.2 1999/03/10 17:45:26 julian Exp $
*/
--- 30,36 ----
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
! * $Id: if_oltr.c,v 1.5 1999/05/09 17:07:24 peter Exp $
*/
***************
*** 259,265 ****
NULL
};
! DATA_SET(pcidevice_set, oltr_device);
int pci_cards = 0;
#endif /* NPCI */
--- 259,265 ----
NULL
};
! COMPAT_PCI_DRIVER(oltr_pci, oltr_device);
int pci_cards = 0;
#endif /* NPCI */
***************
*** 355,361 ****
if (!is->id_irq)
is->id_irq = (1 << sc->config->interruptlevel); /* Claim our interrupt */
is->id_intr = (inthand2_t *)oltr_intr;
- register_intr(ffs(is->id_irq) - 1, is->id_id, is->id_ri_flags, is->id_intr, &net_imask, is->id_unit);
if ((is->id_drq == 0xffffffff) && (sc->config->dmalevel != TRLLD_DMA_PIO))
is->id_drq = sc->config->dmalevel; /* Claim our dma channel */
printf("oltr%d: <%s> [%6D]\n", is->id_unit, AdapterName[sc->config->type], sc->config->macaddress, ":");
--- 355,360 ----
***************
*** 374,380 ****
if (is->id_irq == 0)
is->id_irq = (1 << sc->config->interruptlevel); /* Claim our interrupt */
is->id_intr = (inthand2_t *)oltr_intr;
- register_intr(ffs(is->id_irq) - 1, is->id_id, is->id_ri_flags, is->id_intr, &net_imask, is->id_unit);
if ((is->id_drq == 0xffffffff) && (sc->config->dmalevel != TRLLD_DMA_PIO))
is->id_drq = sc->config->dmalevel; /* Claim our dma channel */
printf("oltr%d: <%s> [%6D]\n", is->id_unit, AdapterName[sc->config->type], sc->config->macaddress, ":");
--- 373,378 ----
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-tokenring" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.4.05.9906031339020.10273-100000>