Hello,
This is a problem I reported on July 19; unfortunately, because of a typo,
the subject header did not appear. Briefly, during traffic through a PCMCIA
card (Ethernet, WiFi, or SCSI adapter), both PCMCIA and USB time out if there
is any USB activity at the time (such as moving a mouse). Otherwise they can
coexist for days. After the timeout PCMCIA cannot be restarted until the USB
modules are removed. This makes USB essentially unusable in this machine, as
there is always some PC card in use. In 2.2.xx, this issue only showed up with
CardBus cards.
I apologize for reposting, but I think it's a real problem that may only
show up in older, slow machines, which is why I haven't seen any references to
it, and the absence of a subject in my original post probably made it useless.
Details are in the archived original message, e.g. in
http://www.uwsg.indiana.edu/hypermail/linux/kernel/0207.2/0884.html
If anybody has any ideas, please CC also to [email protected] and thanks in
advance.
Vassili Papavassiliou
--
Vassili Papavassiliou E-mail: [email protected]
NMSU, Physics Dept. Phone: (505) 646-1310
Las Cruces, NM 88003 Fax: (505) 646-1934
On Wednesday 24 July 2002 02.33, Vassili Papavassiliou wrote:
> Hello,
> This is a problem I reported on July 19; unfortunately, because of a
typo,
> the subject header did not appear. Briefly, during traffic through a PCMCIA
> card (Ethernet, WiFi, or SCSI adapter), both PCMCIA and USB time out if
there
> is any USB activity at the time (such as moving a mouse). Otherwise they can
> coexist for days. After the timeout PCMCIA cannot be restarted until the USB
> modules are removed. This makes USB essentially unusable in this machine, as
> there is always some PC card in use. In 2.2.xx, this issue only showed up
with
> CardBus cards.
>
> I apologize for reposting, but I think it's a real problem that may only
> show up in older, slow machines, which is why I haven't seen any references
to
> it, and the absence of a subject in my original post probably made it
useless.
> Details are in the archived original message, e.g. in
> http://www.uwsg.indiana.edu/hypermail/linux/kernel/0207.2/0884.html
>
> If anybody has any ideas, please CC also to [email protected] and thanks in
> advance.
> Vassili Papavassiliou
Your logs shows several
"Jul 18 14:58:33 localhost kernel: wvlan_cs: This is a PrismII card, not a
Wavelan IEEE card :-( "
I have found that the best driver for PrismII cards is the orinoco, this is my
> lsmod
orinoco_cs 4328 1
orinoco 29408 0 [orinoco_cs]
hermes 3328 0 [orinoco_cs orinoco]
ds 6368 1 [orinoco_cs]
yenta_socket 8384 1
pcmcia_core 37984 0 [orinoco_cs ds yenta_socket]
...
I do also use USB, memory card with quite big transfers. And I have never seen
this problem - I will look for it...
/RogerL
--
Roger Larsson
Skellefte?
Sweden
|
| Your logs shows several
| "Jul 18 14:58:33 localhost kernel: wvlan_cs: This is a PrismII card, not a
| Wavelan IEEE card :-( "
|
| ...
|
| I do also use USB, memory card with quite big transfers. And I have never seen
| this problem - I will look for it...
|
| /RogerL
|
Thank you for taking the time to respond. I showed the error logs for this
card and driver as an example; however I see the same problem (simultaneous
disconnects) using a 3Com 589D Ethernet card, an Adaptec 1480A SCSI adapter
and even a PCMCIA modem (Gateway). All work fine until I move a USB mouse
while a lot of data flows through the PCMCIA card (small transfers do not
automatically hang the two). I don't have other USB devices with me at the
moment to try, but I believe I have seen the effect in the past with others.
I'd be happy to attach more logs if someone wants to poke, and also to try
suggestions for different settings etc. Thanks again,
Vassili
--
Vassili Papavassiliou E-mail: [email protected]
NMSU, Physics Dept. Phone: (505) 646-1310
Las Cruces, NM 88003 Fax: (505) 646-1934