NetBSD Problem Report #21472

Received: (qmail 24984 invoked by uid 605); 5 May 2003 23:46:50 -0000
Message-Id: <20030505234637.E4018269D@rangerover.13thmonkey.org>
Date: Tue,  6 May 2003 01:46:37 +0200 (CEST)
From: imago@13thmonkey.org
Sender: gnats-bugs-owner@netbsd.org
Reply-To: imago@13thmonkey.org
To: gnats-bugs@gnats.netbsd.org
Subject: USB device interferes with IDE device; interrupt lossage
X-Send-Pr-Version: 3.95

>Number:         21472
>Category:       port-alpha
>Synopsis:       USB device interferes with IDE device; interrupt lossage
>Confidential:   no
>Severity:       critical
>Priority:       medium
>Responsible:    port-alpha-maintainer
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Mon May 05 23:47:00 +0000 2003
>Closed-Date:    
>Last-Modified:  Sun Sep 03 01:18:24 +0000 2006
>Originator:     Reinoud Zandijk
>Release:        NetBSD 1.5 till at least 1.6R
>Organization:
NetBSD

>Environment:
NetBSD 1.6-current kernel of 20030505 on an DEC Alpha 500au. Big IDE disc, 
no SCSI discs. No USB devices.

>Description:
When the `ohci* at pci? dev? function?' is enabled in the kernel 
configuration IDE interrupts seem to be intercepted/delivered at the USB 
device that doesnt understand them and rejects them. This gives rise to IDE 
timeouts, USB sheduling overruns (with NO DEVICES!!!) etc.

I've dug into this before and in conversations it seemed that the ohci
driver isnt fully Alpha aware in that the bus-barriers are not allways done
right. The last time i tried it still had this problem so nowadays i've 
disabled ohci by default.

>How-To-Repeat:
Boot a kernel with ohci enabled.

>Fix:
Disable ohci and disable usb at ohci
>Release-Note:
>Audit-Trail:

From: "Nathan J. Williams" <nathanw@wasabisystems.com>
To: imago@13thmonkey.org
Cc: gnats-bugs@gnats.netbsd.org
Subject: Re: port-alpha/21472: USB device interferes with IDE device; interrupt lossage
Date: 06 May 2003 13:36:24 -0400

 imago@13thmonkey.org writes:

 > I've dug into this before and in conversations it seemed that the ohci
 > driver isnt fully Alpha aware in that the bus-barriers are not allways done
 > right. The last time i tried it still had this problem so nowadays i've 
 > disabled ohci by default.

 Since the bus barriers are currently implicit in all of the Alpha bus
 space operations, and I've been successfully using USB and IDE on my
 alphas for some time, I think the problem is more subtle than you
 suggest.

 Can you post dmesg for the system in quesion, with ohci included?

         - Nathan

From: Reinoud Zandijk <reinoud@netbsd.org>(by way of Erik E. Fair)
To: NetBSD GNATS Problem Report Tracking System <gnats-bugs@gnats.netbsd.org>
Cc:  
Subject: Re: port-alpha/21472: USB device interferes with IDE device;
 interrupt lossage
Date: Tue, 6 May 2003 16:32:40 -0700

 Hiya Nathan,

 hmm... i got another datapoint and i dont like it :-/ ... it can completely
 crash my Alpha and kill everything including keyboard and reset butten etc.

 I've added the kernel message upto the freeze. There may be more but those
 were not saved in the crash.

 Cheers,
 Reinoud

 May  6 22:44:30 starbuck /netbsd: scsibus0: waiting 2 seconds for 
 devices to settle...
 May  6 22:44:30 starbuck /netbsd: root on wd0a dumps on wd0b
 May  6 22:44:30 starbuck /netbsd: root file system type: ffs
 May  6 22:44:29 starbuck savecore: no core dump
 May  6 22:44:30 starbuck /netbsd: ohci0: 1 scheduling overruns
 May  6 22:44:31 starbuck mountd[239]: "/usr/tmp	-maproot", line 3: 
 can't get address info for host oboema.reinoud.kasbah
 May  6 22:44:35 starbuck timed[495]: slave to kabel208244
 May  6 22:44:43 starbuck /netbsd: ohci0: 1 scheduling overruns
 May  6 22:44:51 starbuck /netbsd: tlp0: transmit underrun; new 
 threshold: 96/256 bytes
 May  6 22:45:57 starbuck /netbsd: ohci0: 1 scheduling overruns
 May  6 22:45:58 starbuck /netbsd: ohci0: 3 scheduling overruns
 May  6 23:08:29 starbuck /netbsd: cd0: no disk label
 May  6 23:09:40 starbuck /netbsd: ohci0: 1 scheduling overruns
 May  6 23:09:40 starbuck /netbsd: wd0f: DMA error writing fsbn 
 68397200 of 68397200-68397215 (wd0 bn 70982720; cn 70419 tn 5 sn 53), 
 retrying
 May  6 23:09:40 starbuck /netbsd: stray isa irq 14
 May  6 23:09:40 starbuck /netbsd: wd0: soft error (corrected)
 May  6 23:24:56 starbuck syslogd: restart
 May  6 23:24:56 starbuck /netbsd: hip 21143 Ethernet, pass 3.0
 May  6 23:24:56 starbuck /netbsd: tlp0: interrupting at dec 550 irq 0
 May  6 23:24:56 starbuck /netbsd: tlp0: DEC , Ethernet address 
 00:00:f8:76:43:f5
 May  6 23:24:56 starbuck /netbsd: nsphy0 at tlp0 phy 5: DP83840 
 10/100 media interface, rev. 1
 May  6 23:24:56 starbuck /netbsd: nsphy0: 10baseT, 10baseT-FDX, 
 100baseTX, 100baseTX-FDX, auto
 May  6 23:24:56 starbuck /netbsd: tlp0: 10baseT, 10baseT-FDX, 10base2, 10base5

From: Manuel Bouyer <bouyer@antioche.lip6.fr>
To: imago@13thmonkey.org
Cc: gnats-bugs@gnats.netbsd.org
Subject: Re: port-alpha/21472: USB device interferes with IDE device; interrupt lossage
Date: Wed, 14 May 2003 14:40:14 +0200

 On Tue, May 06, 2003 at 01:46:37AM +0200, imago@13thmonkey.org wrote:
 > 
 > >Number:         21472
 > >Category:       port-alpha
 > >Synopsis:       USB device interferes with IDE device; interrupt lossage
 > >Confidential:   no
 > >Severity:       critical
 > >Priority:       medium
 > >Responsible:    port-alpha-maintainer
 > >State:          open
 > >Class:          sw-bug
 > >Submitter-Id:   net
 > >Arrival-Date:   Mon May 05 23:47:00 UTC 2003
 > >Closed-Date:
 > >Last-Modified:
 > >Originator:     Reinoud Zandijk
 > >Release:        NetBSD 1.5 till at least 1.6R
 > >Organization:
 > NetBSD
 > 	
 > >Environment:
 > NetBSD 1.6-current kernel of 20030505 on an DEC Alpha 500au. Big IDE disc, 
 > no SCSI discs. No USB devices.
 > 
 > >Description:
 > When the `ohci* at pci? dev? function?' is enabled in the kernel 
 > configuration IDE interrupts seem to be intercepted/delivered at the USB 
 > device that doesnt understand them and rejects them. This gives rise to IDE 
 > timeouts, USB sheduling overruns (with NO DEVICES!!!) etc.

 On what kind of alpha ?
 For DS10 it's known that the USB ports is not functionnal, and the latest
 SRM disable the USB devices completely.

 --
 Manuel Bouyer, LIP6, Universite Paris VI.           Manuel.Bouyer@lip6.fr
      NetBSD: 24 ans d'experience feront toujours la difference
 --

From: Manuel Bouyer <bouyer@antioche.lip6.fr>
To: Reinoud Zandijk <reinoud@netbsd.org>
Cc: gnats-bugs@gnats.netbsd.org
Subject: Re: port-alpha/21472: USB device interferes with IDE device; interrupt lossage
Date: Wed, 14 May 2003 15:14:03 +0200

 On Wed, May 14, 2003 at 03:04:21PM +0200, Reinoud Zandijk wrote:
 > Hiya Manuel,
 > 
 > On Wed, May 14, 2003 at 02:40:14PM +0200, Manuel Bouyer wrote:
 > > > >Category:       port-alpha
 > > > >Synopsis:       USB device interferes with IDE device; interrupt lossage
 > 
 > > > >Environment:
 > > > NetBSD 1.6-current kernel of 20030505 on an DEC Alpha 500au. Big IDE disc, 
 > > > no SCSI discs. No USB devices.
 > 
 > > On what kind of alpha ?
 > > For DS10 it's known that the USB ports is not functionnal, and the latest
 > > SRM disable the USB devices completely.
 > 
 > Sorry i didnt specify it enough : its an DEC Alpha Personal Work Station
 > 500au with the DECchip 2117x Core Logic Chipset (Pyxis), pass 1.
 > 
 > I hope this helps you more ... feel free to ask!

 Can you post a complete boot message ?

 --
 Manuel Bouyer, LIP6, Universite Paris VI.           Manuel.Bouyer@lip6.fr
      NetBSD: 24 ans d'experience feront toujours la difference
 --
Responsible-Changed-From-To: port-alpha-maintainer->mycroft 
Responsible-Changed-By: mycroft 
Responsible-Changed-When: Tue Jun 29 03:46:55 UTC 2004 
Responsible-Changed-Why:  
. 
State-Changed-From-To: open->feedback 
State-Changed-By: mycroft 
State-Changed-When: Sat Jul 3 00:54:23 UTC 2004 
State-Changed-Why:  
This is now partly fixed.  The IDE works, even in the presence of ohci. 
However, ohci does not yet work. 

From: Reinoud Zandijk <reinoud@netbsd.org>
To: gnats-bugs@gnats.netbsd.org
Cc:  
Subject: Re: port-alpha/21472
Date: Mon, 5 Jul 2004 16:28:21 +0200

 Dear Charles,

 On Sat, Jul 03, 2004 at 12:55:40AM -0000, mycroft@netbsd.org wrote:
 > Synopsis: USB device interferes with IDE device; interrupt lossage
 > 
 > State-Changed-From-To: open->feedback
 > State-Changed-By: mycroft
 > State-Changed-When: Sat Jul 3 00:54:23 UTC 2004
 > State-Changed-Why: 
 > This is now partly fixed.  The IDE works, even in the presence of ohci.
 > However, ohci does not yet work.

 The result on my DEC PWS 500au is that the IDE is still giving problems; 
 lots of spurious IRQ 14's. The only changes on the GENERIC kernel are :

 Index: GENERIC
 ===================================================================
 RCS file: /cvsroot/src/sys/arch/alpha/conf/GENERIC,v
 retrieving revision 1.246
 diff -u -r1.246 GENERIC
 --- GENERIC     28 Jun 2004 21:07:47 -0000      1.246
 +++ GENERIC     5 Jul 2004 14:24:33 -0000
 @@ -150,7 +150,7 @@

  # System V-like shared memory
  options        SYSVSHM
 -#options       SHMMAXPGS=1024          # 1024 pages is the default
 +options        SHMMAXPGS=16384         # 1024 pages is the default

  # p1003.1b semaphore support
  options        P1003_1B_SEMAPHORE
 @@ -159,7 +159,7 @@
  options        LKM

  # Disable kernel security.
 -#options       INSECURE
 +options        INSECURE

  # Misc. options
  options        EISAVERBOSE             # recognize "unknown" EISA devices
 ===================================================================

 This results in the following kernel lines: 

 .....
 ohci0 at pci0 dev 7 function 3: Contaq Microsystems 82C693 PCI-ISA Bridge (rev. 0x00)
 ohci0: interrupting at isa irq 10
 ohci0: OHCI version 1.0, legacy support
 usb0 at ohci0: USB revision 1.0
 uhub0 at usb0
 uhub0: Contaq Microsys OHCI root hub, class 9/0, rev 1.00/1.00, addr 1
 uhub0: 2 ports with 2 removable, self powered
 .....
 wd0: drive supports PIO mode 4, DMA mode 2, Ultra-DMA mode 5 (Ultra/100)
 wd0(cypide0:0:0): using PIO mode 4, DMA mode 2 (using DMA data transfers)
 cd0(cypide0:0:1): using PIO mode 4, DMA mode 2 (using DMA data transfers)
 atapibus1 at atabus1: 2 targets
 cd1 at atapibus1 drive 0: <SAMSUNG CD-R/RW DRIVE SW-224B, , VE002R4> cdrom 
 removable
 cd1: 32-bit data port
 cd1: drive supports PIO mode 4, DMA mode 2
 cd1(cypide1:0:0): using PIO mode 4
 stray isa irq 14
 root on wd0a dumps on wd0b
 stray isa irq 14
 stray isa irq 14; stopped logging
 root file system type: ffs
 nfs_send: proc botch: rep 0x0 arg 0x0 curproc 0xfffffc0001c28280
 ohci0: 1 scheduling overruns
 ohci0: 1 scheduling overruns
 ....
 ohci0: 1 scheduling overruns
 ohci0: 2 scheduling overruns
 ......

 While compiling another version of the disc, it got hung completely.... a 
 power cycle was needed to get it working again.

 Cheers,
 Reinoud

State-Changed-From-To: feedback->open 
State-Changed-By: reinoud 
State-Changed-When: Mon Jul 5 14:32:53 UTC 2004 
State-Changed-Why:  
As the latest mail indicates, the problem is still there. 

From: Reinoud Zandijk <reinoud@netbsd.org>
To: gnats-bugs@netbsd.org
Cc:  
Subject: Re: port-alpha/21472
Date: Thu, 5 Aug 2004 16:44:29 +0200

 This PR# seems to be similar to kern/16379. When running a -current kernel, 
 the USB works fine now apart from ONE sheduling overrun with nothing 
 attached (!) but otherwise just the IRQ 14 still so i am hessitant to try 
 it out intesively.

 Attaching an USB mouse works fine (see 16379) so i think the interference
 with the IDE interface is gone now but the spurious IRQ 14's keep on
 coming.  Eventually the spurious IRQ 14's crash the machine when ohci not 
 compiled in-kernel.
Responsible-Changed-From-To: mycroft->port-alpha-maintainer
Responsible-Changed-By: wiz@netbsd.org
Responsible-Changed-When: Sun, 03 Sep 2006 01:18:24 +0000
Responsible-Changed-Why:
Back to role account, mycroft doesn't have commit access any longer.


>Unformatted:

NetBSD Home
NetBSD PR Database Search

(Contact us) $NetBSD: query-full-pr,v 1.39 2013/11/01 18:47:49 spz Exp $
$NetBSD: gnats_config.sh,v 1.8 2006/05/07 09:23:38 tsutsui Exp $
Copyright © 1994-2007 The NetBSD Foundation, Inc. ALL RIGHTS RESERVED.