-----Original Message----- From: Holger Hoffstätte [mailto:holger@applied-asynchrony.com] Sent: Monday, November 12, 2018 11:42 AM To: Schmauss, Erik erik.schmauss@intel.com; Greg Kroah-Hartman gregkh@linuxfoundation.org; linux-kernel@vger.kernel.org Cc: stable@vger.kernel.org; Jean-Marc Lenoir archlinux@jihemel.com; Wysocki, Rafael J rafael.j.wysocki@intel.com Subject: Re: [PATCH 4.19 025/361] ACPICA: AML interpreter: add region addresses in global list during initialization
Hello Erik,
thanks for responding.
On 11/12/18 8:01 PM, Schmauss, Erik wrote:
As probably expected this patch causes problems. In my case one server can no longer load the nct6775 hwmon module, which means the fan cannot be monitored, and therefore my monitoring system promptly starts spamming me with alerts that my fan has failed - which is of course
not true.
--snip-- Nov 12 18:08:56 tux kernel: nct6775: Found NCT6776D/F or compatible chip at 0x2e:0x290 Nov 12 18:08:56 tux kernel: ACPI Warning: SystemIO range 0x0000000000000295-0x0000000000000296 conflicts with OpRegion 0x0000000000000290-0x Nov 12 18:08:56 tux kernel: ACPI: If an ACPI driver is available for this device, you should use it instead of the native driver --snip--
This is certainly caused by my old BIOS and its broken ACPI implementation, however since it's working perfectly fine otherwise I see no reason to replace it. That being said, I must be able to monitor my fan, so for now reverting the patch immediately "fixed" the problem for me - the fan entries appeared in sysfs again after
successfully loading the module.
Idea, workarounds or patches welcome.
Is there a firmware update available for this machine? If so, you may want to do a firmware update.
The motherboard is from 2011 and I already installed the last available BIOS from 2013 a few years ago. That fixed exactly nothing. :) This was my last motherboard from Asus ever, but alas, it's still here and otherwise working, so..
Also, what is the behavior in kernels 4.16 or older?
Hi,
Everything just worked - yes, with ACPI spew, but.. ¯(ツ)/¯ I understand that the patch might well be "technically correct" and OK for mainline/current hardware and will try playing with acpi=.. kernel flags.
It's strange how the driver would load in 4.16 and fail to load with this patch. I would like to make sure that the driver load behavior is not being changed from 4.16 or earlier.
Could you send a working dmesg from 4.16 (or earlier) and a failing dmesg with this stable branch using the same kernel command line parameters for both?
I don't need too much of the spew if it looks the same. I'm more interested in the initializations and the warning message.
Erik
thanks, Holger