On Fri, Sep 22, 2017 at 02:27:57PM +0200, Takashi Iwai wrote: > On Fri, 22 Sep 2017 10:04:53 +0200, > Johannes Stezenbach wrote: > > > > In conclusion the BIOS bug is that it enables these clocks at boot, > > and the correct quirk is to make clk-pmc-atom.c disable them during init, > > but leave the _PRx handling for the used clocks (CLK3) to ACPI. > > Well, I can't say which is the buggy behavior, but at least, I agree > with Rafael that white/blacklisting is likely the way to go. > > One thing I'm wondering is whether it's specific to device or specific > to platform. The ASUS one that showed a regression Johannes and I > have is the Cherrytrail, while the device showed the hang-at-boot was > Baytrail, IIRC. If I understand correctly, the clk-pmc-atom.c clocks are not used by the SoC itself but provided for devices added to system by OEM. So I don't think its platform (BYT vs. CHT) specific. Meanwhile I have doubts about d31fd43c0f9a41, though. The original issue mentioned by Carlo: https://www.spinics.net/lists/platform-driver-x86/msg12092.html >>> We have a quirk downstream in place where we basically modified the >>> r8169 driver to claim the 25MHz pmc_plt_clk_4 clock, and this seems to >>> work fine, but we really want to find a more upstreamable and >>> definitive solution. It seems to me that this quirk to claim pmc_plt_clk_4 would actually be a better solution than just keeping clocks enabled. And you'd need to handle the clock during suspend for S0ix to work, if Baytrail pmc_plt_clk_4 has the same S0ix blocking behaviour as Cherrytrail CLK3. IOW, the driver which uses the pmc_plt_clk_4 clock needs to handle it if ACPI DSDT doesn't handle it via _PRx. Thanks, Johannes -- To unsubscribe from this list: send the line "unsubscribe linux-acpi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html