I bought a laptop several days ago, and begin to build gentoo for it, after struggling for days, I finally configured the kernel properly for my hardware, but some strange behavior is still there and I have no way to concel that. first I tried the stable version gentoo-sources 4.1.15, configured with the configuration of kernel of version 4.0.5 that I have used for my old laptop, with only necessary hardware drivers modified, but it print so many warnings or errors, I have to configure it from 0 again. Finaly some the problem seems to have gone, and I tried to start x but the terminal just flick and I have no choice but poweroff. startx report error no screen at that time, but as I see there is no configuration mistake with the X. then I switched to kernel 4.5.2, with the same configuration( directly load the configuration), startx works but gdm login screen just come into noresponding status after I input the password of the user. soon I found that when this happend, the command lspci also doesn't work(no responding and can't stop, after several minutes a kernel panic prints and also some ACPI ERRORS), what's the most amazing is that this problem occurs randomly, the chance of the lspci works is larger if I have poweroffed the computer for long time. After looking into dmesg, modified and recompiled the kernel for several times, I found that the culprit seems to be nouveau module loaed by the kernel, If I blacklist nouveau, nearly all the problem went away, and everything works, as there is intel integrated GPU I can still use X. Besides I also found the problem of nvcvideo, iwlwifi, i915 which may be a bug. For nvcvideo, kernel v4.4.8 is all right but v4.5.2 with the same configuration print "Entity type for entity Extension 4 was not initialized!" and etc. on boot. for kernel v4.4.8, so long as I select the usb_gadget module in kernel the same message will also be printed on boot, but for kernel v4.5.2, even if I deselect the gadget module, the message is still there, while they toad the same kernel coniguration. For iwlwifi and i915, in different kernel it nseems to require different firmware, that's why I report th bug here for kernel, as I think the module iwlwifi and i915 should explicitly declare the firmware it need to load when configure the kernel. And the firmware iwlwifi they required are not included in the package linux-firmware even for ~amd64(there is iwlwifi-7265D-21.ucode but no iwlwifi-7265D-20.ucode which is require in kernel v4.5.2), but it's in the git version of linux-firmware. Reproducible: Always Steps to Reproduce: 1.compile the kernel with the configurations 2.blacklist nouveau or not, reboot 3.execute lspci or start gdm and login(you'd better try only lspci so that you can have the chance to login and log dmesg
Created attachment 432002 [details] emerge --info
Created attachment 432004 [details] lspci -v
Created attachment 432006 [details] gentoo-sources-4.4.8 configuration load from kernel v4.5.2 with USB_GADGET deselected and EXTRA_FIRMWARE modified, module tpm or tpm_crb deselected too as it caused a panic like this for kernel 4.4.8: resource sanity check: requesting [mem 0xfed40040-0xfed4103f], which spans more than MSFT0101:00 [mem 0xfed40000-0xfed4087f] [ 14.417282] ------------[ cut here ]------------ [ 14.417311] WARNING: CPU: 2 PID: 4624 at arch/x86/mm/ioremap.c:198 __ioremap_caller+0x2d4/0x350() [ 14.417334] Info: mapping multiple BARs. Your kernel is fine. [ 14.417335] Modules linked in: [ 14.417344] tpm_crb(+) hci_uart acpi_cpufreq tpm_tis tpm btintel i2c_hid pinctrl_sunrisepoint pinctrl_intel mei_me mei i2c_i801 wmi shpchp [ 14.417392] CPU: 2 PID: 4624 Comm: systemd-udevd Not tainted 4.4.8-gentoo-rt #1 [ 14.417412] Hardware name: Notebook N550RC/N550RC, BIOS 5.11 02/17/2016 [ 14.417435] 0000000000000000 ffffffff813b39d9 ffff88037088fb00 ffffffff81f8a3e1 [ 14.417460] ffffffff8108bacc ffffc90001f80040 ffff88037088fb50 ffffc90001f80040 [ 14.417485] 0000000000002000 0000000000000040 ffffffff8108bb47 ffffffff81f843c0 [ 14.417510] Call Trace: [ 14.417521] [<ffffffff813b39d9>] ? dump_stack+0x46/0x5d or a sanity check and devm_*remap (several lines on boot, not remembered) for kernel 4.5.2.
Created attachment 432008 [details] gentoo-sources-4.5.2 configuration
Created attachment 432010 [details] dmesg-4.4.8.log
Created attachment 432012 [details] dmesg for kernel v 4.5.2 with nouveau blacklist cancelled same kernel of v4.5.2, only delete nouveau blacklist in grub2, when logined, lspci don't work as described in the problem description.
Created attachment 432014 [details] dmesg for kernel v 4.5.2
for the ACPI Warning: ACPI Warning: SystemIO range 0x000000000000F040-0x000000000000F05F conflicts with OpRegion 0x000000000000F040-0x000000000000F04F (\_SB.PCI0.SBUS.SMBI) (20160108/utaddress-255) I still have no idea what the problem is, it seems to be related to module wmi but as I have tried deselect drm and wmi, there is still a warning like this. it's now always printed on boot, it's a little awesome.
This doesn't look like you're reporting a bug, but like you have several problems that you don't know how to fix. Please use our support forums, mailing lists and/or IRC channels for that.
(In reply to Jeroen Roovers from comment #9) > This doesn't look like you're reporting a bug, but like you have several > problems that you don't know how to fix. Please use our support forums, > mailing lists and/or IRC channels for that. Sorry for that, but I think the problem is indeed a bug, of course it's the problem that I want to solve and now have no way to do that, but I think it's first a bug as the modules in kernel just caused the issues, and same module for kernel of different version have different behaviour(the newer one seems not working fine), so it's reasonable to be regarded as bugs and should be right for here. As for the acpi warning "SystemIO range ...", I'm sincerely sorry, it might be a bug related to bios which shouldn't have reported here.
(In reply to rtgiskard from comment #10) > (In reply to Jeroen Roovers from comment #9) > > This doesn't look like you're reporting a bug, but like you have several > > problems that you don't know how to fix. Please use our support forums, > > mailing lists and/or IRC channels for that. > > Sorry for that, but I think the problem is indeed a bug, of course it's the > problem that I want to solve and now have no way to do that, but I think > it's first a bug as the modules in kernel just caused the issues, and same > module for kernel of different version have different behaviour(the newer > one seems not working fine), so it's reasonable to be regarded as bugs and > should be right for here. > > As for the acpi warning "SystemIO range ...", I'm sincerely sorry, it might > be a bug related to bios which shouldn't have reported here. it seems that I'm not very clear about that, if you're sure that this is not a bug, please tell me! so that I have chance to deal with it, thank you very much!
did you set your VIDEO_CARDS variable?
(In reply to Alan Neilan from comment #12) > did you set your VIDEO_CARDS variable? yes, of cource，nouveau is included，I think it maight be for my mainboard，which is Sunrise Point HM170 for intel skylake CPU，the support may not be well enough. And I got to understand why Torvalds said "fuck you" to nvidia, when I found the crash is caused by the module nouveau or nvidia.
i'm pretty sure you're only supposed to have either nouveau or nvidia, not both
(In reply to Alan Neilan from comment #14) > i'm pretty sure you're only supposed to have either nouveau or nvidia, not > both Indeed，at least not for the same time， but I'd like to have a choice of booting with nouveau or nvidia driver，from gentoo wiki, I have managed to do that within a single kernel with one of the driver blacklisted，of course it works for my old laptop. for the new notebook，neither of them works, even if I have only one of them installed, not only blacklist. Finally I found that blacklist them have the same effect with uninstall, so I have them both installed now.
