Bug 14356

Summary: NFS dead after series of 10-50 suspend-resume
Product: Power Management Reporter: warpme (piotr.oniszczuk)
Component: Hibernation/SuspendAssignee: power-management_other
Severity: blocking CC: florian, lenb, rjw
Priority: P1    
Hardware: All   
OS: Linux   
Kernel Version:, Subsystem:
Regression: Yes Bisected commit-id:
Bug Depends on:    
Bug Blocks: 7216, 12398    

Description warpme 2009-10-10 09:46:55 UTC
System is diskless multimedia terminal (minimyth).
Issue is: dead Eth port after series of S3 suspend-resume (10-50 cycles) when NFS is used as share. 
With CIFS all looks OK. works OK (700 cycles tested)

I applied patch for bug 11156 as without it system randomly hangs at resume stage.

dmsg reports:
Oct 10 10:17:25 (none) user.err kernel: ACPI: I/O resource nForce2_smbus [0x2e00-0x2e3f] conflicts with ACPI region SM00 [0x2e00-0x2e3f]
Oct 10 10:17:25 (none) user.info kernel: ACPI: Device needs an ACPI driver
Oct 10 10:17:25 (none) user.info kernel: Linux agpgart interface v0.103
Oct 10 10:17:25 (none) user.err kernel: nForce2_smbus 0000:00:01.1: Error probing SMB2.
Comment 1 Rafael J. Wysocki 2009-10-10 21:13:29 UTC
If everything is correct with CIFS, I guess the problem is somehow related to NFS.
Comment 2 Len Brown 2009-10-13 01:39:42 UTC
Can you confirm that we understand?
CIFS is okay, and NFS is failing, and thus the network interface is
okay and the issues is specific to NFS?
Comment 3 warpme 2009-10-13 16:45:25 UTC
I have following results for
System is disk/kbd-less network booted appliance so I can control it only by: IR (lirc), VNC and telnet. 

CIFS: 500 S3 cycles: OK

-no 11156patch: random hang on S3 resume (black screen, Eth dead)
-with 11156patch: no Eth after 1-40 S3 cycles, but rest seems to be fully functional.

Exactly the same system but with - S3 with CIFS/NFS are OK
Comment 4 Florian Mickler 2010-12-08 09:11:14 UTC
Is this still a problem on current kernels?
Comment 5 warpme 2010-12-08 18:53:30 UTC
Hi !

It looks that with 2.6.33 and newer all is OK, albeit I'm not able to tell whivh kernel verssion fixed problem.
Comment 6 Florian Mickler 2010-12-08 20:13:28 UTC
I'm closing this as unreproducible then, thanks. 

If this pops up again, please reopen.