On Fri, 27 Apr 2012, Sarah Sharp wrote: > Hi Alan, > > I noticed your commit to fix S3 hangs on ASUS machines with an EHCI host > controller: > I've seen a whole bunch of users complaining about what look like > similar issues. Most of them come across this blog post and use the > script to unbind both the EHCI and xHCI drivers: > > http://thecodecentral.com/2011/01/18/fix-ubuntu-10-10-suspendhibernate-not-working-bug I've seen different scripts that just had the sysfs paths hardcoded. > I'm wondering if several other computer brands are effected by a similar > buggy BIOS/firmware. It's hard to tell, since unbinding the drivers can > paper over many different kinds of symptoms. Yep. So far I haven't heard about it on anything but ASUS machines. And only those with chipsets in the 6 Series/C200 Series. > Colin is working with someone who is running Ubuntu on a Lenovo U300s, > who also sees a hang on suspend. The user said installing the script to > unbind both the EHCI and xHCI drivers fixed the suspend hang. The bug > report is here: > > https://bugs.launchpad.net/ubuntu/+source/linux/+bug/904261 > > At this point, we actually don't know if xHCI or EHCI is to blame, since > the script unbinds both drivers. I didn't think about that fact when I > was trying various fixes (all of which didn't work). > > Colin, could you have the user unbind the EHCI driver only, attempt to > suspend, and see if they can suspend and resume successfully? Then > reboot, unbind the xHCI driver only, and see if suspend works? > > If just unbinding the EHCI driver helps, then you should roll an Ubuntu > kernel with Alan's patch and see if suspend just works with the patch. > > Alan, do you remember if any of the ASUS machines that were effected by > this bug also had xHCI? I'm wondering if the buggy firmware needs both > host controllers in D0 on suspend, or just the EHCI host controller. At least one of the affected machines _does_ have xHCI, and the bug reporter did try suspending with xHCI bound and EHCI unbound. It worked. > If it needs both hosts, then I need to make a similar patch for the xHCI > driver. However, most of the laptops include the NEC host controller > chip, and not all the laptops with an NEC host have this buggy suspend > behavior. So maybe we would have to key off BIOS names? > > Alan, do you know what BIOS vendor/version the ASUS machine that had > this issue was running? There were several machines involved, with slightly varying PCI IDs on the controllers. Some of them had two EHCI controllers and some had only one. BIOSes tested included versions 213 and 219 from ASUS's web site; I don't know the vendor. Alan Stern -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
