[Bug 399816] New: Cannot suspend to ram on AMD TL-64 in smp mode
https://bugzilla.novell.com/show_bug.cgi?id=399816 Summary: Cannot suspend to ram on AMD TL-64 in smp mode Product: openSUSE 11.0 Version: RC 1 Platform: x86-64 OS/Version: openSUSE 11.0 Status: NEW Severity: Normal Priority: P5 - None Component: Kernel AssignedTo: bnc-team-screening@forge.provo.novell.com ReportedBy: clarkt@cnsp.com QAContact: qa@suse.de Found By: Beta-Customer Since the 2.25 kernel, it has been impossible to suspend to ram or disk. The system hangs on the vt screen during the suspend process. Only way to recover is a hard poweroff. Using nosmp, suspends now work. 10.3 works fine on this laptop in suspend to ram or disk modes with both cpus up. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=399816 Robert Vojcik <rvojcik@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |rvojcik@novell.com AssignedTo|bnc-team-screening@forge.provo.novell.com |seife@novell.com -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=399816 User seife@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=399816#c1 Stefan Seyfried <seife@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |fseidel@novell.com, pavel@novell.com Status|NEW |ASSIGNED --- Comment #1 from Stefan Seyfried <seife@novell.com> 2008-06-27 09:34:21 MDT --- Pavel, any idea on how to debug this? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=399816 User pavel@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=399816#c2 --- Comment #2 from Pavel Machek <pavel@novell.com> 2008-06-30 17:43:44 MDT --- ..well, we could do git bisect ;-). Can you add no_console_suspend, and try various echo test > /sys/power/disk (then disk > state) and testproc > disk (then disk > state) ? Can you try with minimal modules? (init 1) -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=399816 User clarkt@cnsp.com added comment https://bugzilla.novell.com/show_bug.cgi?id=399816#c3 --- Comment #3 from Clark Tompsett <clarkt@cnsp.com> 2008-07-01 20:15:15 MDT --- Tried the first part and it did work at level 1. testproc was not found. I have found reports of others having suspend problems with the 2.6.25 kernel in other distributions. I also found entries on the kernel mailing list dealing with AMD chips, iommu and suspend. This may need to put on hold until more information is posted on the kernel mailing list. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=399816 User clarkt@cnsp.com added comment https://bugzilla.novell.com/show_bug.cgi?id=399816#c4 --- Comment #4 from Clark Tompsett <clarkt@cnsp.com> 2008-08-17 17:15:45 MDT --- This problem is due to changes made between 2.6.24 and 2.6.25+ kernels. Modules now have to be unloaded in order to have any suspend (ram or disk) to work. Modules that need to be unloaded are -- button ohci_hcd ehci_hcd These modules do not allow the system powerdown during suspend actions. This may affect other machines. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=399816 User seife@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=399816#c5 Stefan Seyfried <seife@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |rwysocki@novell.com AssignedTo|seife@novell.com |pavel@novell.com Status|ASSIGNED |NEW --- Comment #5 from Stefan Seyfried <seife@novell.com> 2008-08-18 04:30:39 MDT --- ok, so it is a plain kernel bug => Pavel :-) -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=399816 User pavel@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=399816#c6 Pavel Machek <pavel@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |ASSIGNED --- Comment #6 from Pavel Machek <pavel@novell.com> 2008-08-19 01:42:58 MDT --- Well, suspend works for me with those modules. Do you really need to unload all 3? Could you narrow it down a bit more using bisection? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=399816 User clarkt@cnsp.com added comment https://bugzilla.novell.com/show_bug.cgi?id=399816#c7 --- Comment #7 from Clark Tompsett <clarkt@cnsp.com> 2008-08-19 17:45:09 MDT --- ehci_hci is the only module that needs to be unloaded. Need information on how to use bisection to see what changed between 2.6.24 and 25. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=399816 User rjw@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=399816#c8 --- Comment #8 from Rafael Wysocki <rjw@novell.com> 2008-08-21 14:30:54 MDT --- 1. Clone the mainline kernel repository: $ git clone \ git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git linux-2.6 (this will create subdirectory 'linux-2.6' in your current directory and download the repository contents into it) $ cd linux-2.6 2. Start bisection: $ git bisect start $ git bisect good v2.6.24 $ git bisect bad v2.6.25 3. Build the kernel, test it and if it fails do $ git bisect bad Otherwise do $ git bisect good 4. Repeat 3. until there are no more commits to test. If you're lucky, git will tell you which commit causes the problem. (see also: http://linux.yyz.us/git-howto.html) -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=399816 User rjw@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=399816#c9 --- Comment #9 from Rafael Wysocki <rjw@novell.com> 2008-08-21 14:32:25 MDT --- Still, after you've done 1. above, please build the kernel and test it. There is a chance it has the problem fixed. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=399816 User pavel@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=399816#c10 Pavel Machek <pavel@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |oneukum@novell.com Summary|Cannot suspend to ram on AMD TL-64 in smp mode |ehci_hci breaks suspend --- Comment #10 from Pavel Machek <pavel@novell.com> 2008-08-22 04:40:13 MDT --- Oliver, if you have any ideas... seems like ehci problem is here... -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=399816 User oneukum@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=399816#c11 --- Comment #11 from Oliver Neukum <oneukum@novell.com> 2008-08-25 02:01:25 MDT --- This is new. If it hangs the obvious things to try would be the watchdog or sysrq-t. If it works with nosmp it looks like a spinlock deadlock. I'd suggest the watchdog approach. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=399816 User clarkt@cnsp.com added comment https://bugzilla.novell.com/show_bug.cgi?id=399816#c12 --- Comment #12 from Clark Tompsett <clarkt@cnsp.com> 2008-08-25 10:34:33 MDT --- Please explain watchdog, where do I find it and how to use it to get the information needed. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=399816 User clarkt@cnsp.com added comment https://bugzilla.novell.com/show_bug.cgi?id=399816#c13 --- Comment #13 from Clark Tompsett <clarkt@cnsp.com> 2008-08-26 11:22:28 MDT --- cLoaded 11.1A1 into a different partition. With the 2.6.26 kernel, the system would suspend with out unloading any modules. Sync'ed with the Factory repo this morning and installed the updates for 11.1A2 with the 2.6.27-rc3 kernel. The system is still able to suspend without having to unload any modules. What ever was broken in 2.6.25 has been fixed in 2.6.26/27-rc3. But, a new problem has shown up. S2ram -- system goes to sleep. Pressing the power button, system starts up and boots. It does not try to start from ram. S2disk is also broken. Shuts down and should have saved and image to the swap file. Rebooting, the save image is not read (if it was created) and the system does a normal boot. (This was tested on a 2nd machine with the same results). -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=399816 User pavel@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=399816#c14 Pavel Machek <pavel@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |WONTFIX --- Comment #14 from Pavel Machek <pavel@novell.com> 2008-09-01 08:21:38 MDT --- Now you got me confused... Anyway, if ehci_hcd problem is fixed in 2.6.27-rc3, there's probably no pressing need to debug/bisect it. (But if you do bisect it, we can probably fix that on 2.6.26.) If you want to debug s2ram/s2disk on 2.6.27-rc3, that's probably best done on vanilla kernel from kernel.org, debugging done on lkml or bugzilla.kernel.org. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
participants (1)
-
bugzilla_noreply＠novell.com