From: Michael Frank <mhf@linuxmail.org>
To: Greg KH <greg@kroah.com>
Subject: Re: [linux-usb-devel] 2.6.0-test4 - PL2303 OOPS - see also 2.4.22: OOPS on disconnect PL2303 adapter
Date: Sat, 6 Sep 2003 10:31:19 +0800
User-Agent: KMail/1.5.2
Cc: linux-kernel <linux-kernel@vger.kernel.org>,
       linux-usb-devel@lists.sourceforge.net
References: <200309020139.08248.mhf@linuxmail.org> <200309031432.17209.mhf@linuxmail.org> <20030905230852.GA18196@kroah.com>
In-Reply-To: <20030905230852.GA18196@kroah.com>
MIME-Version: 1.0
Content-Disposition: inline
Message-Id: <200309060932.47136.mhf@linuxmail.org>
Content-Type: text/plain;
  charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 3372
Lines: 106

On Saturday 06 September 2003 07:08, Greg KH wrote:
> On Wed, Sep 03, 2003 at 02:32:16PM +0800, Michael Frank wrote:
> > On Wednesday 03 September 2003 07:52, Greg KH wrote:
> > > Try the patch below and let me know if this solves it for you or not.
> >
> > If it is meant to reset the buffers, it has _no_ effect.
> >
> > Some more observations:
> >
> > Besides it just stopping without obvious reason:
> >
> > 1) It does not like when something is typed on cu and not received by the
> > serial port side connected to PL2303 (CTS low). It tends to hang and the
> > trouble starts....
> >
> > Sep  3 12:52:15 mhfl2 kernel: ttyUSB0: 1 input overrun(s)
> > Sep  3 12:54:30 mhfl2 last message repeated 2 times
>
> Hm, what is causing this?
>

I don't understand why it get's Input overruns when it sends a single key.
"Input" seems not to have any problem.

Could there be an event meant for output misrouted to input - messing things
up?

> That is probably why cu is getting confused, right?

I think so. Once this message shows up, it is essentially unusable.

> > plug in
> > Sep  3 12:55:47 mhfl2 kernel: hub 1-0:0: debounce: port 2: delay 100ms stable 4 status 0x101
> > Sep  3 12:55:48 mhfl2 kernel: hub 1-0:0: new USB device on port 2, assigned address 3
> > Sep  3 12:55:48 mhfl2 kernel: usb 1-2: device not accepting address 3, error -110

> That's showing either you don't have good pci interrupt routing going
> on, or a messed up device.

Interrupts - no, I copied gigabytes to/from USB hard disk, 
eth0 and yenta on PCI are fine too.

Device - how to verify?

Could it be a "misunderstanding" between device and driver?
 - driver (seing new device) want's  to assign address to device
 - device (plugged in, reset), sees a line status changed and 
   want's to send respective event to driver

> > _disconnect_ serial port side of PL2303
> >
> > plug in - OK
> > Sep  3 12:56:07 mhfl2 kernel: usbserial 1-2:0: PL-2303 converter detected
> > Sep  3 12:56:07 mhfl2 kernel: usb 1-2: PL-2303 converter now attached to
> > ttyUSB0 (or usb/tts/0 for devfs)
>
> Heh, ok, it looks like you have a wierd device.

Could it be that - device (plugged in, reset), sees _no_ line status changed
and has nothing to send. 

Perhaps there is a sequencing problem somewhere, and it works by chance.

>> After a while it hang again, this time unloaded USB _without_ exit cu

> Hm, how can you do this?  There should be a reference on the pl2303
> driver as you have the port open.  Or are you just removing the host
> controller driver here?

It's not loaded on boot, but only when needed. The scripts:

usb1)
  if [ ! -e /proc/bus/usb ]; then
    echo Loading USB
    modprobe usbcore
    mount -t usbdevfs usbdevfs /proc/bus/usb
    modprobe ohci_hcd
    modprobe sd_mod
    modprobe pl2303 
    modprobe lp
  fi
  ;;

usb0)
  echo Unloading USB
  rmmod  usb-storage sd_mod scsi-mod
  rmmod pl2303 usbserial
  rmmod  lp parport
  rmmod  ohci_hcd 
  umount usbdevfs
  rmmod  usbcore
  ;;

Perhaps this is too dumb and I should do some checking along the way,
however joe user should be unable to oops things up...


Regards
Michael


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/