Comment 9 for bug 1065840

Joe, I think we're clearly dealing with the same bug as yours #48951.

From what I've found on people having this same problem :

-> Joe's bug #48951 on the kernel tracker (+ the various contributors on it)
-> here : http://forums.fedoraforum.org/showthread.php?t=283005
-> here : https://bugs.archlinux.org/task/31833
-> here : http://forums.opensuse.org/english/get-technical-help-here/applications/479271-suspend-s2ram-issues-12-2-nvidia-platform.html
-> and here : http://www.linux.org.ru/forum/linux-hardware/8359386

... I've noticed that all these people had one thing in common in their hardware : an NVidia chipset => sata_nv driver.

Some other people share the "Some devices failed to suspend" error message, but on other devices. The non stopping hard drives case seems to happen, AFAIK, only on NVidia chipsets.

In my case, as I previously described in my (too long) first post, behavior is even weirder :
- one machine with 4 hdd
- Ubuntu Studio 10.10 x64 installed on one physical disk (Hitachi 160GB)
- Ubuntu Studio 12.10 x64 installed on another one (HP-labeled VB0250EAVER 250GB - origin Seagate)
- suspend works OK when booted from 10.10 (see attached log)
- only one disk enters suspend when booted from 12.10 (the one holding 10.10 install), the others fail to suspend (see second log after this post .... sorry machine renamed in 12.10 install)

Note: the "FLUSH CACHE EXT" error on the HP disk may be due to a firmware bug. I'll try to update HP firmware to at least get this bug out of the way, so don't take this error into account for now .... although I find it really hard to believe that a firmware bug could be OS-dependent !

Well, gathering all this, I'd say that the problem lies in the sata_nv driver, either the way sata_nv interprets PM requests or the way sata_nv talks to disks.

I've found no matching package name for potential affectation of the bug, is sata_nv maintained directly by NVidia ?