Per bug filed upstream: http://bugzilla.kernel.org/show_bug.cgi?id=13353 Hibernate/suspend would still work if net.eth0 (e100) were stopped prior to hibernate/suspend (using hibernate-script /etc option to stop the service) using =sys-kernel/gentoo-sources-2.6.29-r4. As soon as I upgraded to =sys-kernel/gentoo-sources-2.6.29-r5, hibernate now hangs even after net.eth0 (e100) is stopped. Reproducible: Always The apparent main cause of the upstream bug appears to be an extreme amount of code rewrites/patches to the e100.c file when compared to 2.6.28 tree to the 2.6.29 tree. (Many include firmware code rewrites.) Hibernate worked w/o problems in 2.6.28. Hibernate stops on initial 2.6.29 releases. But a workaround of stopping the net.eth0 prior to hibernate worked. From what I see, gentoo-sources-2.6.29-r5 includes a patch to "enable PCI resources". <shrugs> ... whatever is going on here, I'm seeing a lot of upgrading of code all at once without reasonable testing prior to moving on! ... in my opinion.
can you try gentoo-sources-2.6.30-r4?... seems they added/patched some suspend-related code in e100.c...
I don't see gentoo-sources-2.6.30-r4 yet, but thanks for the heads up. Will try when I see it. (From what I'm both seeing & hearing, 2.6.30 needs a lot of lov'n.)
http://www.kernel.org/pub/linux/kernel/v2.6/ChangeLog-2.6.30.1 http://www.kernel.org/pub/linux/kernel/v2.6/ChangeLog-2.6.30.2 The above two change logs state nothing concerning e100.c Just e1000.c in the *.1 patch However, the *.2 patch, the very first entry makes mention of fixing a compile time option with gcc-4.1. Both changelogs make patches concerning suspend. Just wanted to clear the air of Comment #1. I test them as they become available.
Yea. I have three boxes here, and =sys-kernel/gentoo-sources-2.6.29-r4 was solid with hibernating with in kernel suspend/hibernate with NVidia Proprietory/Xorg started. As soon as =sys-kernel/gentoo-sources-2.6.29-r5 hit, and all newer version, can no longer hibernate while Xorg & Nvidia Proprietary driver are loaded.
there another changes on e100 in the 2.6.30-r5 try to reproduce with it http://kernel.org/pub/linux/kernel/v2.6/ChangeLog-2.6.30.5
Will watch the upstream bug
Dunno what's to watch. It's already solved upstream within the later git kernel releases. But I'm guessing, you are implying we should watch to see if the next stable release of 2.6.30.* will have the incorporated patch ... or if 2.6.31 release might be soon. ;-) ... basically what I'm doing right now (after masking all 2.6.30 versions and sticking with 2.6.29 for my SMP box).