Subject: PROBLEM: EHCI Driver hard-locks system when USB devices plugged. (2.4.21 & 2.6.0-test2)
Date: Monday 11 August 2003 03.46
From: Alexander Markley <[EMAIL PROTECTED]>
To: [EMAIL PROTECTED]


I'm sure you guys in this list just thrill with delight when a bug
report comes in. Keeping that in mind, I have attempted to make this the

That reaction comes more with patches, actually ... ;)


most complete bug-report possible. I am using the template for bug
reports that is in the root of the source tree. (If there is too much
here, sorry... Just following the bug report template.)

Then there's also the USB FAQ, which tells you among other things that we need to know about the device being plugged in ... in your case, it's pretty clear that the device is part of the problem, since lots of people plug USB devices in without any trouble.

Or maybe not ... you've also got IRQ values over 16, which are
often a problem.  When you plug your device into a USB port
with just the UHCI driver loaded, does it work?  If so, make
your bug report include /proc/bus/usb/devices output.  If not,
try the usual:  boot with "pci=noacpi" or "acpi=off", disable
APIC usage, etc.

If that doesn't work for you, more info is needed.

What'd be ideal then is to build your kernel with CONFIG_USB_DEBUG,
and use some console other than X11 so you can collect more information.
Come to think of it, I can't really imagine any way to remotely debug
such a problem without knowing basics like whether this was an oops
or some other kind of problem.

For example, boot into runlevel 2, alt-sysrq-8 to make sure messages
go to the console, plug it in.  The last messages coming to the
console are important; send them to the list.  (Or if you know how
to use a serial console from another box, use minicom or something
similar to capture even more information ...)  Tell whether you
still get responses from alt-sysrq commands in the "hang".

Don't use 2.4.21 EHCI; use at least 2.4.22-pre1.

Of random interest, there's clearly a problem with the 2.4.22 code
setting cacheline sizes.  It shouldn't matter, but I'm slightly
surprised that's still not fixed.

- Dave





-------------------------------------------------------
This SF.Net email sponsored by: Free pre-built ASP.NET sites including
Data Reports, E-commerce, Portals, and Forums are available now.
Download today and enter to win an XBOX or Visual Studio .NET.
http://aspnet.click-url.com/go/psa00100003ave/direct;at.aspnet_072303_01/01
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel

Reply via email to