The r8169 driver in kernel 2.6.38 will sometimes delay resume from suspend-to-RAM by 60 s when the rtl8168d firmware is not present. This affects chips which need no external firmware as well. Maybe a dependency on the sys-kernel/linux-firmware package could be introduced to the kernel source packages. Possibly behind a "firmware" USE flag so it is not forced on users who don't want/need it.
I'm not sure that's a good workaround. Maybe a message, or notice, or something like that, which would warn r8169 users about the delay issue when no firmware is used, would be better, until/if the issue is fixed. r8169 devs are aware of this problem. From the commit of message of [1]: "It will not avoid the 60 seconds delay when: - there is no firmware - the driver is loaded and the device is not up before a suspend/resume" I think, a bug report should be filed upstream, to check if the devs are going to fix this soon. [1] http://git.kernel.org/?p=linux/kernel/git/torvalds/linux2.6.git;a=commit;h=f1e02ed109df5f99abf942b8ccc99960cb09dd38
Created attachment 270777 [details, diff] r8169_fix_firmware_suspend.patch Just saw a discussion in LKML about this. There's a patch proposed which probably fixes the issue, although not fully tested yet. Attached the patch if anyone wants to try it out.
Commit 953a12cc2889d1be92e80a2d0bab5ffef4942300 upstream should fix this. Can anyone test it and confirm that the delay problem is gone? The patch doesn't apply cleanly to 2.6.38 sources, but if it actually fixes the problem, mabye I can backport it. Although 2.6.39 will be released soon, I think.
Created attachment 275007 [details, diff] Patch for linux-2.6.38-gentoo-r6 I confirm upstream commit 953a12cc2889d1be92e80a2d0bab5ffef4942300 solves that issue, I've attached the patch I created for my sys-kernel/gentoo-sources-2.6.38-r6 Cheers, Patrick
2.6.39 released, and added to portage. Closing.