Bug 149200

Summary: kernel freezes soon after it switches to fbdev on iBook G4
Product: [Fedora] Fedora Reporter: Gabriel Schulhof <gabrielschulhof>
Component: kernelAssignee: David Woodhouse <dwmw2>
Status: CLOSED ERRATA QA Contact: Brian Brock <bbrock>
Severity: medium Docs Contact:
Priority: medium    
Version: 4CC: davej, keifer.miller, overholt, redhat, wtogami
Target Milestone: ---   
Target Release: ---   
Hardware: powerpc   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2005-10-20 14:30:20 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:

Description Gabriel Schulhof 2005-02-21 00:54:33 UTC
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.3)
Gecko/20041020

Description of problem:
Here's the screen it freezes on:
+----------+
|          |
| Penguin  |
|  Logo    |
|          |
|          |
|          |
+----------+
Apple USB OHCI 0001:10:19.0 disabled by firmware
usbcore: registered new driver usbfs
usbcore: registered new driver hub
Thermal assist unit not available
Registering PowerMac CPU frequency driver
Low: 707 Mhz, High: 931 Mhz, Boot: 707 Mhz
audit: initializing knetlink socket (disabled)
audit(1108919135.453:0): initialized
VFS: Disk quotas dquot_6.5.1
Dquot-cache hash table entries: 1024 (order 0, 4096 bytes)
SELinux:  Registering netfilter hooks
Initializing Cryptographic API
ksign: Installing public key data
Loading keyring
- Added public key D5ABE686CE6001A6
- User ID: Red Hat, Inc. (Kernel Module GPG key)
PCI: Enabling device 0000:00:10.0 (0006 -> 0007)
radeonfb (0000:00:10.0): Invalid ROM signature 0 should be0xaa55
radeonfb: Retreived PLL infos from Open Firmware
radeonfb: Reference=27.00 MHz (RefDiv=12) Memory=190.00 Mhz,
System=183.00 MHz
radeonfb: PLL min 12000 max 35000
radeonfb: Monitor 1 type LCD found
radeonfb: EDID probed
radeonfb: Monitor 2 type not found
Console: switching to colour frame buffer device 128x48
radeonfb: Dynamic Clock Power Management enabled
Registered "mnca" backlight controller, level: 15/15
radeonfb (0000:00:10.0): ATI Radeon \c
Generic RTC Driver v1.07
Macintosh non-volatile memory driver v1.1
Linux agpgart interface v0.100  (c) Dave Jones
agpgart: Detected Apple UniNorth 2 chipset
agpgart: Maximum main memory to use for agp memory: 565M
agpgart: configuring for size idx: 4
agpgart: AGP aperture is 16M @ 0x0
[drm] Initialized drm 1.0.0.20040925
Serial: 8250/16550 driver $Revision: 1.90 $ 76 ports, IRQ sharing enabled


Version-Release number of selected component (if applicable):
kernel-2.6.10-1.1148_FC4

How reproducible:
Always

Steps to Reproduce:
Well, almost every time. I have to try 3-4-5 times before it boots.
1. Boot 1148 on an iBook G4

Actual Results:  kernel freezes as described above

Expected Results:  Normal boot ...

Additional info:

This problem has been present since I started testing FC for PPC
(kernel 2.6.9-1.667, IIRC).

Comment 1 Gabriel Schulhof 2005-03-03 16:19:29 UTC
Well, things are going from bad to worse. With
kernel-2.6.10-1.1162_FC4 it freezes even before it starts the frame
buffer. This time, it freezes every time (tried it about 8-9 times in
a row). It doesn't get past this point:

Welcome to Linux, kernel 2.6.10-1.1162_FC4

linked at        : 0xc0000000
frame buffer at  : 0x9c000000  (phys), 0xd0000000  (log)
klimit           : 0xc0405888
MSR              : 0x00003030
HID0             : 0x8410c0bc

pmac_init(): exit
id mach(): done
MMU:enter
MMU: hw init
hash:enter
hash:find piece
hash:patch
hash:done
MMU:mapin
MMU:setio
MMU:edxit
setup_arch: enter
setup_arch: bootmem
arch: exit
openpic: enter
openpic: timer
openpic: external
openpic: spurious
openpic: exit

Comment 2 Gabriel Schulhof 2005-03-03 16:22:12 UTC
P.S.: If there are speling errors, it's because i copied this
information by hand.

s/0xd0000000/0xd0008000/

Comment 3 Gabriel Schulhof 2005-05-15 01:58:52 UTC
Well, the problem went away for a few kernel versions, but, as of
2.6.11-1.1305_FC4, the problem is back. Please fix it !

Comment 4 Gabriel Schulhof 2005-05-20 04:43:29 UTC
The problem, as originally stated, persists when using kernel 2.6.11-1.1319_FC4.

Comment 5 David Woodhouse 2005-05-20 09:16:18 UTC
The real problem is with catching the machine check when we access nonexistent
I/O addresses. I'll see if we can track that down -- can you show the full
contents of /proc/cpuinfo?

It was originally triggered by the serial driver, but that is now not probing
for PC serial ports so it doesn't happen.

In 1.1305 we turned on CONFIG_ISA (and with it CONFIG_ISAPNP) for the benefit of
RS/6000 users. That's why it broke then.

We've now turned off CONFIG_ISAPNP, in 1.1327 and later. Please test that. If it
doesn't work we'll turn CONFIG_ISA off again, and for RS/6000 we'll either have
to have a separate kernel or we'll have to fix your _real_ underlying problem first.

Comment 6 Gabriel Schulhof 2005-05-20 17:06:51 UTC
/proc/cpuinfo:

processor       : 0
cpu             : 7455, altivec supported
clock           : 707MHz
revision        : 0.3 (pvr 8001 0303)
bogomips        : 710.65
machine         : PowerBook6,3
motherboard     : PowerBook6,3 MacRISC3 Power Macintosh
detected as     : 287 (iBook G4)
pmac flags      : 0000001b
L2 cache        : 256K unified
memory          : 640MB
pmac-generation : NewWorld


Comment 7 Gabriel Schulhof 2005-05-20 17:35:42 UTC
Well, 1.1327 is not yet available via yum, but 1.1323 booted successfully twice
in a row.

Comment 8 David Woodhouse 2005-05-24 10:44:04 UTC
*** Bug 152385 has been marked as a duplicate of this bug. ***

Comment 9 Gabriel Schulhof 2005-05-25 07:49:00 UTC
kernel-2.6.11-1.1341 freezes as described at the top of this thread.

Comment 10 David Woodhouse 2005-05-25 08:52:54 UTC
OK, then we need to turn CONFIG_ISA back off. Done in 2.6.11-1.1356, which will
hopefully squeeze its way into the FC4 release. 

Comment 11 Dave Jones 2005-05-27 00:12:02 UTC
This did make it into the FC4 release kernel.


Comment 12 Gabriel Schulhof 2005-06-07 20:21:13 UTC
Well, as of kernel-2.6.11-1.1369_FC4, the kernel /still/ freezes occasionally
(approximately 1/4 of the time) as described at the beginning of this bug.

Comment 13 Dave Jones 2005-10-06 05:03:33 UTC
How does the latest errata kernel fare ?


Comment 14 Andrew Overholt 2005-10-19 23:03:07 UTC
The latest errata kernel no longer boots for me.  I updated some stuff the other
day (from FC4 updates ... it was just a few things like PostgreSQL but I can't
remember what else) and now it won't boot.  I get a hang at "Red Hat nash
version 4.2.15 starting".  The cursor is blinking but there is no disk activity
and it's not doing anything.  Before I get to this point, I can hear some disk
activity but then it stops.  Caps Lock and Num Lock function and I can type
things, but the boot appears to just be stopped.

Comment 15 David Woodhouse 2005-10-20 14:30:20 UTC
I think the current kernel should have the 8250 bug fixed though. What you
describe in comment #14 sounds like something entirely different. 

I'll commit the 8250 patch to rawhide (it was only in FC4 before) and close this
bug.

Comment 16 Andrew Overholt 2005-10-20 15:30:08 UTC
David helped me on IRC and it turns out that my problem was due to the fact that
I had updated to rawhide glibc (by way of gcc) but not rawhide prelink.  Thanks,
David!