Bug 119815 - acpi: Interrupt routing for e100 broken
Summary: acpi: Interrupt routing for e100 broken
Keywords:
Status: CLOSED UPSTREAM
Alias: None
Product: Fedora
Classification: Fedora
Component: kernel
Version: rawhide
Hardware: i386
OS: Linux
medium
medium
Target Milestone: ---
Assignee: Arjan van de Ven
QA Contact:
URL:
Whiteboard:
Depends On:
Blocks: FC2Target
TreeView+ depends on / blocked
 
Reported: 2004-04-02 10:41 UTC by Alex Kiernan
Modified: 2007-11-30 22:10 UTC (History)
1 user (show)

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Clone Of:
Environment:
Last Closed: 2004-04-08 16:53:29 UTC
Type: ---
Embargoed:

Attachments (Terms of Use)
/proc/interrupts from a 2.6.3-2.1.253.2.1smp kernel (the working one) (1.34 KB, text/plain)
2004-04-02 19:14 UTC, Alex Kiernan 		no flags Details
/proc/interrupts from a 2.6.4-1.303smp kernel (a broken one) (1.26 KB, text/plain)
2004-04-03 10:20 UTC, Alex Kiernan 		no flags Details
Comparison between startup messages (25.11 KB, text/plain)
2004-04-05 20:33 UTC, Alex Kiernan 		no flags Details
Boot messages from 2.6.4-1.305 (31.09 KB, text/plain)
2004-04-06 12:41 UTC, Alex Kiernan 		no flags Details
View All

Description Alex Kiernan 2004-04-02 10:41:13 UTC 
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.6)
Gecko/20040312 Epiphany/1.1.12

Description of problem:
Using 2.6.4-1.300 with a dual e100 card, no interrupts get delivered
and the card doesn't work - the card gets routed to to 14.

Falling back to 2.6.3-2.1.253.2.1, the interrupt gets routed to 138,
interrupts get delivered and the card works.

Version-Release number of selected component (if applicable):
kernel-smp-2.6.4-1.300

How reproducible:
Always

Steps to Reproduce:
1. Install & Configure e100 card
2.
3.
    

Additional info:
Comment 1 Sahil Verma 2004-04-02 17:56:25 UTC 
Please post /proc/interrupts from both kernels.
Comment 2 Alex Kiernan 2004-04-02 19:14:08 UTC 
Created attachment 99079 [details]
/proc/interrupts from a 2.6.3-2.1.253.2.1smp kernel (the working one)

eth0 and eth1 are e100s on a dual card, eth2 is an e1000
Comment 3 Alex Kiernan 2004-04-03 10:20:43 UTC 
Created attachment 99085 [details]
/proc/interrupts from a 2.6.4-1.303smp kernel (a broken one)

eth0, eth1 are the dual e100, eth2 the e1000
Comment 4 Alex Kiernan 2004-04-05 09:06:33 UTC 
Swapping the dual e100 for a quad tulip has the same problems, so it 
doesn't look like the problem is directly related to the NIC in use.
Comment 5 Arjan van de Ven 2004-04-05 09:09:45 UTC 
what kind of motherboard is this ?
Comment 6 Alex Kiernan 2004-04-05 09:17:42 UTC 
Its a Dell Precision 450 w/ dual 2.6GHz Xeons - according to Dell's 
web site its an Intel E7505 chipset with 533MHz system bus.
Comment 7 Sahil Verma 2004-04-05 17:30:25 UTC 
Any irq related errors in /var/log/messages?
Comment 8 Alex Kiernan 2004-04-05 20:00:41 UTC 
Only lots of:

Apr  2 08:43:26 alexk kernel: NETDEV WATCHDOG: eth1: transmit timed 
out
Apr  2 08:43:26 alexk kernel: e100: eth1: e100_watchdog: link up, 
100Mbps, full-duplex

Nothing else I can see.
Comment 9 Alex Kiernan 2004-04-05 20:33:35 UTC 
Created attachment 99122 [details]
Comparison between startup messages

This is the result of diffing the two sets of startup messages
Comment 10 Sahil Verma 2004-04-06 02:19:43 UTC 
2.6.4 is causing a horrible interrupt storm, maybe the ehci-hcd driver
is buggy. Please boot with that module disabled and post dmesg.
Comment 11 Alex Kiernan 2004-04-06 12:41:57 UTC 
Created attachment 99137 [details]
Boot messages from 2.6.4-1.305

This is with ehci-hcd disabled - unfortunately its no better in respect of the
interrupt problem - the dual e100 still doesn't work.
Comment 12 Alex Kiernan 2004-04-07 07:43:17 UTC 
Just rebuilt 2.6.4-1.305 with the patch from attachment id=2481 from
http://bugme.osdl.org/show_bug.cgi?id=2409 - this fixes the problem.
Note You need to log in before you can comment on or make changes to this bug.