Bug 84721

Summary: [NVC1] Nvidia Geforce GT 630 using nouveau on 3.16 kernel. dangerous Fan speed
Product: xorg Reporter: r004 <roo4.id>
Component: Driver/nouveauAssignee: Nouveau Project <nouveau>
Status: NEW --- QA Contact: Xorg Project Team <xorg-team>
Severity: normal    
Priority: medium CC: aplattner, fdsfgs, martin.peres, matheus4551, vnd
Version: unspecified   
Hardware: x86-64 (AMD64)   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features:
Attachments:
Description Flags
dmesg
none
vbois of gt630 by the "using /sys:" method
none
VBIOS of GT 730
none
VBIOS nVidia GT 630 none

Description r004 2014-10-06 15:02:56 UTC
Hello; 
I am using gentoo sources kernel version 3.16 -r1 both hardened and normal kernel.
when booting to thee kernel the fan starts speeding and making very frightening sound. as if it is going to fly :)


x86_64 box.

GRAPHICS:

02:00.0 VGA compatible controller: NVIDIA Corporation GF108 [GeForce GT 630] (rev a1) (prog-if 00 [VGA controller])
	Subsystem: ZOTAC International (MCO) Ltd. Device 6199
	Flags: bus master, fast devsel, latency 0, IRQ 16
	Memory at f6000000 (32-bit, non-prefetchable) [size=16M]
	Memory at e8000000 (64-bit, prefetchable) [size=128M]
	Memory at f0000000 (64-bit, prefetchable) [size=32M]
	I/O ports at e000 [size=128]
	[virtual] Expansion ROM at f7000000 [disabled] [size=512K]
	Capabilities: [60] Power Management version 3
	Capabilities: [68] MSI: Enable- Count=1/1 Maskable- 64bit+
	Capabilities: [78] Express Endpoint, MSI 00
	Capabilities: [b4] Vendor Specific Information: Len=14 <?>
	Capabilities: [100] Virtual Channel
	Capabilities: [128] Power Budgeting <?>
	Capabilities: [600] Vendor Specific Information: ID=0001 Rev=1 Len=024 <?>
	Kernel driver in use: nvidia
	Kernel modules: nouveau, nvidia
Comment 1 r004 2014-10-06 15:04:28 UTC
the above info is extracted from porteus linux that uses binary driver. because I can't stay in the gentoo os with faulted kernel.
Comment 2 Aaron Plattner 2014-10-07 15:31:04 UTC
Please attach a kernel log from the nouveau driver in the bad state.
Comment 3 r004 2014-10-16 16:11:21 UTC
Created attachment 107935 [details]
dmesg

dmesg
Comment 4 Ilia Mirkin 2014-10-16 16:21:01 UTC
Please attach your vbios.

http://nouveau.freedesktop.org/wiki/DumpingVideoBios/
Comment 5 Ilia Mirkin 2014-10-16 16:23:52 UTC
Oh, and also:

sensors (when booted with nouveau)
dmesg (when booted with nouveau, adding nouveau.debug=debug to the kernel command line, and letting it run for at least a minute or two... looking for "FAN target request:" messages)
Comment 6 r004 2014-10-16 16:32:48 UTC
The same problem with FatDog 7.00b1 64bit.
Comment 7 Martin Peres 2014-12-23 21:57:46 UTC
(In reply to r004 from comment #6)
> The same problem with FatDog 7.00b1 64bit.

Hello, I am willing to fix this problem. Please send us your vbios, as Ilia asked. I cannot do anything without it :s
Comment 8 r004 2015-01-23 23:17:19 UTC
Created attachment 112752 [details]
vbois of gt630 by the "using /sys:" method

as requested.
if it is not the correct dump tell me and I will fix it.
but please tell me how to do it if possible.
Comment 9 r004 2015-01-23 23:18:39 UTC
Sorry for the dely but I wasn't sure if I had to stay that long in that env. but I did it because I need nouveau for hardened kernels
Comment 10 Krzysztof Katowicz-Kowalewski 2015-01-25 23:04:25 UTC
As far as I've checked, the problem appeared in 3.15 kernel. I've tried hardened-sources-3.14.29 and there's no problem with the fan, at least in auto mode. But since you change the mode to manual (echo 1 > pwm1_enable), the fan gets mad as in 3.15 kernel branch. Related info from me: https://forums.gentoo.org/viewtopic-t-1009318.html
Comment 11 r004 2015-01-26 00:19:31 UTC
(In reply to Krzysztof Katowicz-Kowalewski from comment #10)
> As far as I've checked, the problem appeared in 3.15 kernel. I've tried
> hardened-sources-3.14.29 and there's no problem with the fan, at least in
> auto mode. But since you change the mode to manual (echo 1 > pwm1_enable),
> the fan gets mad as in 3.15 kernel branch. Related info from me:
> https://forums.gentoo.org/viewtopic-t-1009318.html

I didn't do anything to the kernel and I did change nothing in settings.
the tests I have done are all in the live usb env and all from the 3.16 and above kernels.
 voild-linux-x86_64-01-2015-enlightenment.iso
 fatdog-x86_64-7.00b1.iso
 latest lileblue.iso
 ... and so on
IT MAY BE statred at 3.15 but all of live cd/usb (s) that I used were 3.16 or above I assumed it started by then.
Comment 12 Krzysztof Katowicz-Kowalewski 2015-01-26 22:10:01 UTC
@r004, I had the same problem with my system and I've made little investigation. Indeed the problem appeared in 3.15 and it's still there in latest 3.18 kernel. For the quick and easy work-around use kernel 3.14 or lower.
Comment 13 Ilia Mirkin 2015-01-26 23:13:43 UTC
(In reply to Krzysztof Katowicz-Kowalewski from comment #12)
> @r004, I had the same problem with my system and I've made little
> investigation. Indeed the problem appeared in 3.15 and it's still there in
> latest 3.18 kernel. For the quick and easy work-around use kernel 3.14 or
> lower.

Identifying the commit that broke things between 3.14 and 3.15 would go a long way to solving this issue.

git bisect start v3.15 v3.14 -- drivers/gpu/drm/nouveau

should limit the bisect to just nouveau.
Comment 14 Krzysztof Katowicz-Kowalewski 2015-01-29 20:26:01 UTC
After 7 tries... here's the culprit:

localhost linux # git bisect good
0e994d645627bb67088ae4860e9a0295b123f7b0 is the first bad commit
commit 0e994d645627bb67088ae4860e9a0295b123f7b0
Author: Martin Peres <martin.peres@labri.fr>
Date:   Wed Feb 19 01:04:56 2014 +0100
 
    drm/nouveau/therm: let the vbios decide on the automatic fan management mode
   
    This should fix automatic fan management on fermi cards who do not have
    0x46 entries in the thermal table.
   
    On my nve6, the blob sets the default linear range from 40°C to 100°C
    but my nvcf's default values are 40°C to 85°C. Let's keep 85 as a default
    for everyone.
   
    Signed-off-by: Martin Peres <martin.peres@labri.fr>
    Tested-by: Timothée Ravier <tim@siosm.fr>
    Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
 
:040000 040000 1a0d1ec08c6ffaf36f923a2b68c9f91e945bb299 96bf5f6667528a22493d3a6d4cec6b11ce09414d M      drivers
Comment 15 Tobias Klausmann 2015-01-29 21:39:05 UTC
(In reply to Krzysztof Katowicz-Kowalewski from comment #14)
> 0e994d645627bb67088ae4860e9a0295b123f7b0 is the first bad commit
> commit 0e994d645627bb67088ae4860e9a0295b123f7b0
> Author: Martin Peres <martin.peres@labri.fr>
> Date:   Wed Feb 19 01:04:56 2014 +0100
>  
>     drm/nouveau/therm: let the vbios decide on the automatic fan management
> mode


Following to this there was a conversation on #nouveau:

Reverting a part of the above noted patch:

-	/* starting from fermi, fan management is always linear */
-	if (nv_device(bios)->card_type >= NV_C0 &&
-		fan->fan_mode == NVBIOS_THERM_FAN_OTHER) {
-		fan->fan_mode = NVBIOS_THERM_FAN_LINEAR;
-	}

hides the bug away again.
Comment 16 Krzysztof Katowicz-Kowalewski 2015-02-02 21:19:52 UTC
Created attachment 113073 [details]
VBIOS of GT 730

I've added my gfx card vbios for further investigation.
Comment 17 K.-P. Schrage 2015-03-04 17:44:49 UTC
(In reply to r004 from comment #1)

You card seems to be similar if not identical to mine, so you might be interested in the discussions and Martin's proposals further down in Bug 80901.
Comment 18 matheus4551 2017-02-25 20:42:07 UTC
Still happening as 4.11 kernel. I have to use the closed driver or the fan noise will fry my nerves.
Comment 19 Javier Fernandez 2018-11-27 09:02:54 UTC
Hi there,

im using kernel 4.18 (Ubuntu), and nouveau 1.0.15

fan speed is high here too

Here is my VBIOS
Comment 20 Javier Fernandez 2018-11-27 09:05:13 UTC
(In reply to Javier Fernandez from comment #19)
> Hi there,
> 
> im using kernel 4.18 (Ubuntu), and nouveau 1.0.15
> 
> fan speed is high here too
> 
> Here is my VBIOS

well, i cant attach it!

i get this error message: "The content type is invalid. Valid types must be of the form foo/bar where foo is one of application, audio, image, message, model, multipart, text, video and bar must not contain any special characters (such as "=", "?", ...)."
Comment 21 Javier Fernandez 2018-11-27 09:06:21 UTC
Created attachment 142627 [details]
VBIOS nVidia GT 630

vbios

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.