Tom Rini <[EMAIL PROTECTED]> writes:

> >   o 2.4.10 with my hacked version of se401.c version 0.23 "p1"
> 
> What changes did you have to make?

o The hacking was 8 months ago when I was getting 2.4.9-ac12 working
  for the first time -- my apologies if I'm a little fuzzy about it
  now.

o My notes show that I got a patch against se401-0.23 from Jeroen
  Vreeken.  I think I remember that I had to make some simple change
  to get it to compile with the 2.4.9-ac12 tree.

o I'm attaching "0.23p1.diff", which is `diff se401-0.23{,p1}/se401.c`.
  Past that, I made some trivial changes to the LED handling, and that's
  what I'm using now.


> Question, do you have the hotplug package installed, (and if so, is
> CONFIG_HOTPLUG=y set in your kernel?  If not, that would explain why you
> have to modprobe the drivers by hand.)

o Yes, it is set.  My guess is that I'm having problems because the
  hotplug scripts were installed with SuSE 8.0 (kernel 2.4.18), and
  I'm booting 2.4.10 on the system.


o More: Since my last post, I got the same USB crash I'd been getting
  about once a week on the old RedHat/2.4.9-ac12 system.  I posted
  about this on 4/1/2002, but didn't get any replies.  USB stops
  working, and syslog fills up with several gigabytes per hour of:

    kernel: [cd108120] link (0d108092) element (0cd6c000)
    kernel:   0: [ccd6c000] link (0cd6c030) e0 Stalled CRC/Timeo
BitStuff Length=7ff MaxLen=3f DT1 EndPt=1 Dev=5, PID=69(IN)
(buf=0d3ce000)
    kernel:   1: [ccd6c030] link (0cd6c060) e3 SPD Active Length=0
MaxLen=3f DT0 EndPt=1 Dev=5, PID=69(IN) (buf=0d3ce040)
    kernel:   2: [ccd6c060] link (0cd6c090) e3 SPD Active Len
    ...
    kernel: [skipped 52 active TD's]
    kernel:   63: [ccd6cbd0] link (00000001) e3 SPD IOC Active Length=0
MaxLen=3f DT0 EndPt=1 Dev=5, PID=69(IN) (buf=0d3cefc0)


o To fix the problem, I wrote watchdog scripts which restart the USB
  system when it crashes.  In this current thread, I wrote:

> o This works.  I'm not sure what in the order of operations (modules,
>   hardware plugging) is important.  I'll experiment booting with
>   everything plugged in.

o On my new 2.4.10 kernel, the watchdog USB restarter failed.  The
  only way I got back running was to do the modprobes/insmod with the
  webcams unplugged, and then plug them back in.  I haven't figured
  out how to write a script to do that.  :-)


--
MARK
[EMAIL PROTECTED]
213a214
>       if (se401->inturb) {
215a217,218
>               se401->buttonpressed=0;
>       }
1407c1410
< static int se401_init(struct usb_se401 *se401)
---
> static int se401_init(struct usb_se401 *se401, int snapshot)
1465a1469
>       if (snapshot) {
1481a1486,1488
>       } else {
>               se401->inturb=NULL;
>       }
1501a1509
>       int snapshot;
1512a1521
>               snapshot=1;
1515a1525
>               snapshot=0;
1518a1529
>               snapshot=1;
1521a1533
>               snapshot=1;
1524a1537
>               snapshot=0;
1550c1563
<         if (se401_init(se401)) {
---
>         if (se401_init(se401, snapshot)) {

Reply via email to