Bug 672214

Summary: Transparent hugepages broken in 2.6.38-rc
Product: [Fedora] Fedora Reporter: Paul Bolle <pebolle>
Component: kernelAssignee: Kernel Maintainer List <kernel-maint>
Status: CLOSED RAWHIDE QA Contact: Fedora Extras Quality Assurance <extras-qa>
Severity: unspecified Docs Contact:
Priority: unspecified    
Version: rawhideCC: gansalmon, itamar, jonathan, kernel-maint, kmcmartin, madhu.chinakonda, thomasbelvin
Target Milestone: ---   
Target Release: ---   
Hardware: i686   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2011-01-29 17:35:51 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Attachments:
Description Flags
output of alt-sysrq-t when (apparently) hung none

Description Paul Bolle 2011-01-24 12:35:49 UTC 
Description of problem:
kernel 2.6.38-0.rc2.git0.1.fc15.i686 hangs at boot

Version-Release number of selected component (if applicable):
2.6.38-0.rc2.git0.1

How reproducible:
Always

Steps to Reproduce:
1. boot
2.
3.
  
Actual results:
hangs

Expected results:
runs

Additional info:
0) The last lines of the log (ie, the lines appearing on another machine after the serial port has been reset) are:
[...]
[   11.411257] Synaptics Touchpad, model: 1, fw: 5.9, id: [...], caps: 0x884793/0x0/0x0
[   11.428040] serio: Synaptics pass-through port at isa0060/serio1/input0
[   11.483357] input: SynPS/2 Synaptics TouchPad as /devices/platform/i8042/serio1/input/input4
[   11.674908] yenta_cardbus 0000:02:00.0: CardBus bridge found [1014:0552]
[   11.681731] yenta_cardbus 0000:02:00.0: Using INTVAL to route CSC interrupts to PCI
[   11.689473] yenta_cardbus 0000:02:00.0: Routing CardBus interrupts to PCI
[   11.696362] yenta_cardbus 0000:02:00.0: TI: mfunc 0x01d21b22, devctl 0x64
[   11.780043] ehci_hcd 0000:00:1d.7: PCI INT D disabled
[   11.796036] ehci_hcd 0000:00:1d.7: power state changed by ACPI to D3
[   11.951868] yenta_cardbus 0000:02:00.0: ISA IRQ mask 0x04a8, PCI irq 11
[   11.958564] yenta_cardbus 0000:02:00.0: Socket status: 30000086
[   11.964836] yenta_cardbus 0000:02:00.0: pcmcia: parent PCI bridge window: [io  0x4000-0x8fff]
[   11.973589] pcmcia_socket pcmcia_socket0: cs: IO port probe 0x4000-0x8fff: excluding 0x4000-0x40ff 0x4400-0x44ff 0x4800-0x48ff 0x4c00-0x4cff 0x8000-0x803f
[   12.004161] yenta_cardbus 0000:02:00.0: pcmcia: parent PCI bridge window: [mem 0xc0200000-0xcfffffff]
[   12.013537] pcmcia_socket pcmcia_socket0: cs: memory probe 0xc0200000-0xcfffffff: excluding 0xc0200000-0xc09fffff 0xc3a00000-0xcc1fffff 0xcfa00000-0xd01fffff
[   12.028246] yenta_cardbus 0000:02:00.0: pcmcia: parent PCI bridge window: [mem 0xe8000000-0xefffffff pref]
[   12.038326] pcmcia_socket pcmcia_socket0: cs: memory probe 0xe8000000-0xefffffff: excluding 0xe8000000-0xefffffff
[   12.050031] yenta_cardbus 0000:02:00.1: CardBus bridge found [1014:0552]
[   12.056837] yenta_cardbus 0000:02:00.1: Using INTVAL to route CSC interrupts to PCI
[   12.064611] yenta_cardbus 0000:02:00.1: Routing CardBus interrupts to PCI
[   12.071495] yenta_cardbus 0000:02:00.1: TI: mfunc 0x01d21b22, devctl 0x64
[   12.304467] yenta_cardbus 0000:02:00.1: ISA IRQ mask 0x04a8, PCI irq 11
[   12.311207] yenta_cardbus 0000:02:00.1: Socket status: 30000086
[   12.317590] yenta_cardbus 0000:02:00.1: pcmcia: parent PCI bridge window: [io  0x4000-0x8fff]
[   12.326244] pcmcia_socket pcmcia_socket1: cs: IO port probe 0x4000-0x8fff: excluding 0x4000-0x40ff 0x4400-0x44ff 0x4800-0x48ff 0x4c00-0x4cff 0x8000-0x803f
[   12.361214] yenta_cardbus 0000:02:00.1: pcmcia: parent PCI bridge window: [mem 0xc0200000-0xcfffffff]
[   12.370665] pcmcia_socket pcmcia_socket1: cs: memory probe 0xc0200000-0xcfffffff: excluding 0xc0200000-0xc09fffff 0xc3a00000-0xcc1fffff 0xcfa00000-0xd01fffff
[   12.386455] yenta_cardbus 0000:02:00.1: pcmcia: parent PCI bridge window: [mem 0xe8000000-0xefffffff pref]
[   12.396728] pcmcia_socket pcmcia_socket1: cs: memory probe 0xe8000000-0xefffffff: excluding 0xe8000000-0xefffffff
[   12.506781] dracut: Scanning devices sda2  for LVM volume groups 
[   12.554438] dracut: Reading all physical volumes. This may take a while...
[   12.562243] dracut: Found volume group "VolGroup00" using metadata type lvm2

1) No more messages appear after that. The machine needs to be power cycled to use it again.

2) Feel free to ask for more details.
Comment 1 Chuck Ebbert 2011-01-25 18:05:04 UTC 
How long did you wait before resetting the machine?
Comment 2 Kyle McMartin 2011-01-25 18:30:35 UTC 
Add sysrq_always_enabled to the boot arguments and then when it hangs
hit alt-sysrq-t and see where it's hung. It's already in userspace at this point...
Comment 3 Paul Bolle 2011-01-25 19:58:41 UTC 
(In reply to comment #1)
> How long did you wait before resetting the machine?

Some minutes.
Comment 4 Kyle McMartin 2011-01-25 20:09:23 UTC 
What graphics chipset in that machine? Is it nouveau?
Comment 5 Paul Bolle 2011-01-25 20:26:59 UTC 
(In reply to comment #4)
> What graphics chipset in that machine? Is it nouveau?

No, some radeon variant (it's a 2003/2004 IBM ThinkPad T41). Can't get the details right now. Happens in runlevel 3 too, by the way.
Comment 6 Paul Bolle 2011-01-25 20:35:12 UTC 
Created attachment 475266 [details]
output of alt-sysrq-t when (apparently) hung

(In reply to comment #2)
> Add sysrq_always_enabled to the boot arguments and then when it hangs
> hit alt-sysrq-t and see where it's hung. It's already in userspace at this
> point...

See (enormous) attachment. Have fun!
Comment 7 Paul Bolle 2011-01-25 20:40:24 UTC 
> (In reply to comment #4) 
> Can't get the details right now.

In a previous bug report I reported this chipset as a
    Radeon RV250 [Mobility FireGL 9000]
Comment 8 Kyle McMartin 2011-01-26 13:13:11 UTC 
Sorry, can you try the 2.6.38-0.rc2.git3.2 package I built last night, that should fix things for now.
Comment 9 Paul Bolle 2011-01-26 13:43:17 UTC 
(In reply to comment #8)
> Sorry, can you try the 2.6.38-0.rc2.git3.2 package I built last night, that
> should fix things for now.

0) Will try that shorty. 

1) I just logged in to tell 2.6.38-0.rc2.git1.3 (which yum pulled a short while ago) displays a similar issue (for what that's worth).
Comment 10 Paul Bolle 2011-01-26 14:05:05 UTC 
(In reply to comment #8)
> Sorry, can you try the 2.6.38-0.rc2.git3.2 package I built last night, that
> should fix things for now.

That's correct: that package seems to have fixed things.
Comment 11 Chuck Ebbert 2011-01-26 18:25:20 UTC 
Leave this bug open until the proper fix appears. (Disabling transparent hugepages is only a workaround.)
Comment 12 Paul Bolle 2011-01-28 09:17:55 UTC 
0) The changelog of kernel-2.6.38-0.rc2.git5.1.fc15.i686 claims:
    * Wed Jan 26 2011 Kyle McMartin <kmcmartin> 2.6.38-0.rc2.git5.1
    - [...]
    - [x86] Re-enable TRANSPARENT_HUGEPAGE, should be fixed by cacf061c.

    [...]

1) kernel-2.6.38-0.rc2.git5.1.fc15.i686 seems to boot just fine. Fixed?