On Wed, 2014-07-30 at 17:10 +0300, Andy Shevchenko wrote: > On Tue, 2014-07-22 at 12:19 +0100, Scott Ashcroft wrote: > > On Tue, 2014-07-22 at 11:52 +0300, Andy Shevchenko wrote: > > [] > > > > commit c78b0830667a7e7c1f0ca65b76b33166a84806b3 "ACPI / LPSS: custom power domain for LPSS" > > > > I'm running 3.16-rc5 so that commit is there. > > I'm currently running the packaged kernel from Debian experimental as building anything was difficult until I figured out which module to blacklist. > > I'll be able to build kernels if we need to try some patches. > > Good to hear this, but first one. > > We used to have a bug when LPSS devices went back from suspend with > wrong settings. It should be fixed there. Apparently you have something > new we can't reproduce on the development boards. > > > > > I'd like some advice on how this should be fixed: > > > > > > > > I see that some very similar chipsets would cause the > > > > i2c_designware-pci driver to be loaded where as mine loads the > > > > platform device. > > > > > > You laptop seems to use ACPI, meanwhile other PCI mode for some devices > > > inside (LPSS). > > (*) > > > > > > > > So is the correct fix to get some PCI ids and settings added to the > > > > PCI driver? > > > > > > I'm pretty sure you don't need this. > > I was pretty sure that was the case as I can't see a PCI device which would match but just wanted to check. > > If I understood you correctly in (*), your laptop using ACPI, that > explains why you have platform driver loaded instead of PCI one. > > > > > Or should I be looking to add a DMI based blacklist to stop the i2c > > > > bus being detected at all. > > So, I'm a bit confused now. Which module you blacklisted eventually? > dw_dmac or i2c-designware-platform? At the moment I've blacklisted dw_dmac which stops the crashes. I was speculating on what sort of fix might get accepted in the kernel. Stopping the i2c bus being detected based on DMI looked like it might be easier than trying to force dw_dmac not to load. > > > > Or is this just a bug in the dw_dmac module. If so how can I get any > > > > debug out of it during the rmmod? > > I don't think you may get much from the rmmod stage. > Most of debug messages are for runtime: enabling them via > dw_dmac_core.dyndbg in kernel cmdline (if built-in), or modprobe > dw_dmac_core dyndbg. You have to have CONFIG_DYNAMIC_DEBUG=y in the > kernel configuration. I'll try turning on some debug to see if it shows anything useful. Does having the dw_dmac module loaded actually provide anything useful on laptop? Cheers, Scott -- To unsubscribe from this list: send the line "unsubscribe dmaengine" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
