Bug 141741

Created attachment 225671 [details]
DMESG on 4.5.5

Created attachment 225681 [details]
DMESG on 4.6.4

Created attachment 225691 [details]
LSPCI on 4.5.5

Created attachment 225701 [details]
LSPCI on 4.6.4

Created attachment 225711 [details]
XRANDR 4.5.5

Created attachment 225721 [details]
XRANDR 4.6.4

Can you bisect?

??

What Michel means is using "git bisect" to find which patch introduced the problem.

E.g. you do "git bisect start v4.6.4 v4.5.5" on the kernel source and git will jump into the middle between those two versions.

Then you compile the kernel and test it. If it works you do "git bisect good" in the kernel source directory, if it doesn't work you do "git bisect bad".

Git will send narrow down which patch was the bad one who introduced the problem.

I'll be glad to help but I don't know where to start.
Do you have some documentation ?

Thanks
Michaël

(In reply to Michael Schenaerts from comment #10)
> I'll be glad to help but I don't know where to start.
> Do you have some documentation ?
> 
> Thanks
> Michaël

This might be helpful: https://fedoraproject.org/wiki/Building_a_custom_kernel

Thx, I'll try right away

I did the bisect with no luck. Evrything was running fine

I built the 4.6 and even the 4.7 with the config files from the 4.5 I set the other options to default and... it works.

I'll try to build them with the 4.6 config as soon as I have the time.

Created attachment 239331 [details]
dmesg | egrep 'drm|radeon' on 4.6.0-kali1-amd64

Can you bisect?

It may be the same problem as Bug 120321.

At my PC the latest Antergos live-image does work normal although it has kernel 4.7.4. The installed version and every other distros i tried with the distro or vanilla kernel version has the problem with a kernel version of 4.6 and higher.

(In reply to Lukas P from comment #16)
> It may be the same problem as Bug 120321.
> 
> At my PC the latest Antergos live-image does work normal although it has
> kernel 4.7.4. The installed version and every other distros i tried with the
> distro or vanilla kernel version has the problem with a kernel version of
> 4.6 and higher.

Can you bisect?

I can say that I am also affected by this problem. Same errors as the ones posted here and the ones posted in https://bugzilla.kernel.org/show_bug.cgi?id=120321

I tried with different OS's with kernel >4.5 (Antergos and Fedora) and all of them give the same results. Using the fallback kernel 4.5 for Fedora gives no problems whatsoever

I did a bisect and it reported the following commit as the first bad one:
30b5b8808c12bcd947dd474980482561b69c1bcb

diff --git a/drivers/pci/Kconfig b/drivers/pci/Kconfig
index a1f37db..209292e 100644
--- a/drivers/pci/Kconfig
+++ b/drivers/pci/Kconfig
@@ -128,4 +128,5 @@ config PCI_HYPERV
           The PCI device frontend driver allows the kernel to import arbitrary
           PCI devices from a PCI backend to support PCI driver domains.
 
+source "drivers/pci/hotplug/Kconfig"
 source "drivers/pci/host/Kconfig"

If I remove the changed line in the source of v4.6 everything works normally.

Affected use here as well.
Toshiba Satellite c850 on Arch

kernel 4.6+ fails during boot journalctl shows this

[drm:radeon_get_bios [radeon]] *ERROR* Unable to locate a BIOS ROMradeon 0000:01:00.0: Fatal error during GPU init
[drm] radeon: finishing device.

(In reply to Lukas P from comment #19)
> I did a bisect and it reported the following commit as the first bad one:
> 30b5b8808c12bcd947dd474980482561b69c1bcb
> 
> diff --git a/drivers/pci/Kconfig b/drivers/pci/Kconfig
> index a1f37db..209292e 100644
> --- a/drivers/pci/Kconfig
> +++ b/drivers/pci/Kconfig
> @@ -128,4 +128,5 @@ config PCI_HYPERV
>            The PCI device frontend driver allows the kernel to import
> arbitrary
>            PCI devices from a PCI backend to support PCI driver domains.
>  
> +source "drivers/pci/hotplug/Kconfig"
>  source "drivers/pci/host/Kconfig"
> 
> If I remove the changed line in the source of v4.6 everything works normally.

I'm not an expert, but removing the inclusion of drivers/pci/hotplug/Kconfig equals to disabling all the HOTPLUG_PCI_* options in the config ("Support for PCI Hotplug" in menuconfig). You could try disabling those one by one to see which one causes the problem.

Hello,

You're right Tero Roponen, it is the tree that hides the forest !

I recently opened bug #175391 ( https://bugzilla.kernel.org/show_bug.cgi?id=175391 ) and it was reported by Michael Schenaerts that it was a potential duplicate of this one (#141741).

Thanks to my determination to inquire in that matter and close this bug, we got what happened wrong and Bjorn Helgaas wrote a fix.

Bug #175391 Bjorn's patch made it in both mainline (4.9-rc5) and GregKH's -stable (4.8.9).

I invite you to then carefully read the full thread of bug #175391.

In hope this helps.

Sincerely,
Valentin

Michael, Jose, Lukas, David, Socrates, could you test this with v4.9?  It includes a change that is related: 
http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=16d917b130d7 ("PCI: Don't attempt to claim shadow copies of ROM").

If the problem still occurs on v4.9, please attach the complete dmesg log.

Closing on the assumption that the patch mentioned in comment #23 fixes the problem.

If the problem still occurs on v4.9 or later, please reopen this and attach the complete dmesg log.