On 25.07.2014, Klaus-Peter Schrage wrote: It takes no time when you already have a complete kernel tree :-) You can just apply the patch, type "make" and your're done within a minute. Otherwise, if you plan to recompile a few times more, setting up ccache could be a good idea. https://ccache.samba.org/ Well done! Unfortunately, I have no nvidia graphics to test your work. But I'm shure the ones with problems here in this thread will give you some feedback.

On 24.07.2014 20:15, Lars E. Pettersson wrote: You can directly ask the developers, right. Developers, developers, developers! :) $ grep -R -w 0x46 drivers/gpu/drm/nouveau/ drivers/gpu/drm/nouveau/nouveau_connector.c: nv_device(drm->device)->chipset >= 0x46) drivers/gpu/drm/nouveau/core/engine/device/nv40.c: case 0x46: drivers/gpu/drm/nouveau/core/engine/graph/ctxnv40.c: case 0x46: drivers/gpu/drm/nouveau/core/engine/graph/ctxnv40.c: case 0x46: drivers/gpu/drm/nouveau/core/engine/graph/ctxnv40.c: if (device->chipset == 0x46 || device->chipset == 0x47 || drivers/gpu/drm/nouveau/core/engine/graph/ctxnv40.c: if (device->chipset == 0x46) drivers/gpu/drm/nouveau/core/engine/graph/ctxnv40.c: case 0x46: drivers/gpu/drm/nouveau/core/engine/graph/ctxnv40.c: case 0x46: drivers/gpu/drm/nouveau/core/engine/graph/ctxnv40.c: case 0x46: drivers/gpu/drm/nouveau/core/engine/graph/ctxnv40.c: case 0x46: drivers/gpu/drm/nouveau/core/engine/graph/nv40.c: case 0x46: drivers/gpu/drm/nouveau/core/engine/graph/nv40.c: case 0x46: /* G72 */ drivers/gpu/drm/nouveau/core/engine/graph/nv40.c: case 0x46: drivers/gpu/drm/nouveau/core/subdev/therm/nv40.c: case 0x46: drivers/gpu/drm/nouveau/core/subdev/fb/nv46.c: .base.base.handle = NV_SUBDEV(FB, 0x46), drivers/gpu/drm/nouveau/core/subdev/i2c/anx9805.c: while (nv_rdi2cr(mast, port->addr, 0x46) & 0x10) { drivers/gpu/drm/nouveau/core/subdev/vm/nvc0.c: 0x28, 0x29, 0xff, 0xff, 0xff, 0xff, 0x46, 0xff, /* 0x40 */ drivers/gpu/drm/nouveau/core/subdev/vm/nvc0.c: 0xff, 0x46, 0x46, 0x46, 0x46, 0xff, 0xff, 0xff, /* 0x50 */ drivers/gpu/drm/nouveau/core/subdev/bios/pll.c: * maxN2 (0x46) and 10 * maxM2 (0x4), so set maxN2 to 0x28 and drivers/gpu/drm/nouveau/core/subdev/bios/therm.c: case 0x46: drivers/gpu/drm/nouveau/core/include/subdev/bios/conn.h: DCB_CONNECTOR_DP = 0x46, drivers/gpu/drm/nouveau/dispnv04/nvreg.h:# define NV_CIO_CRE_RCR 0x46 drivers/gpu/drm/nouveau/dispnv04/tvmodesnv17.c: 0x21, 0xf6, 0x94, 0x46, 0x0, 0x0, 0xb, 0x18, Perhaps this one is related to this topic? :) https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/driv... developers fan club

On 24.07.2014, Klaus-Peter Schrage wrote: This would be a perfect chance to update your kernel skills :-) The offending patch (as suggested here; I don't have any nvidia graphics and thus can't verify the effect) reverts cleanly from latest -stable ("cat patch.diff -p1 -R") http://tinyurl.com/mz4vsr8 I guess the Fedora kernel .src.rpm compiles just fine doing a "rpmbuild -bb" on the .spec file. So this should not be of any problem. According to git, 3.14 has been released 30 mar 2014. The offending patch has been committed 26 mar 2014..

Am 24.07.2014 16:35, schrieb poma: Thank you, Mr. Fantomas, for your thorough analysis, which gave some insight even to me! But I can refrain from trying a kernel recompile which I last did a decade or so ago (although, as far as I can remember, this is not such a kind of rocket science as one might guess from the recent thread on "new kernel & rebooting"). So I'll wait until a clever patch finds its way into the (Fedora-)kernel. Meanwhile I stick to the 3.14-kernel.

On 24.07.2014 14:31, Klaus-Peter Schrage wrote: https://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git/tree... static void nouveau_therm_fan_set_defaults(struct nouveau_therm *therm) { struct nouveau_therm_priv *priv = (void *)therm; priv->fan->bios.pwm_freq = 0; priv->fan->bios.min_duty = 0; priv->fan->bios.max_duty = 100; priv->fan->bios.bump_period = 500; priv->fan->bios.slow_down_period = 2000; /*XXX: talk to mupuf */ #if 0 priv->fan->bios.linear_min_temp = 40; priv->fan->bios.linear_max_temp = 85; #endif } https://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git/tree... static void nouveau_therm_fan_set_defaults(struct nouveau_therm *therm) { struct nouveau_therm_priv *priv = (void *)therm; priv->fan->bios.pwm_freq = 0; priv->fan->bios.min_duty = 0; priv->fan->bios.max_duty = 100; priv->fan->bios.bump_period = 500; priv->fan->bios.slow_down_period = 2000; priv->fan->bios.linear_min_temp = 40; priv->fan->bios.linear_max_temp = 85; } "drm/nouveau/therm: let the vbios decide on the automatic fan management mode" https://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git/comm... 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. Here refers to NVCF, your is NV108 (GK208) i.e. GeForce GT 630 http://nouveau.freedesktop.org/wiki/CodeNames/ The fan is really LOUD all the time. How do I make it more quiet? http://nouveau.freedesktop.org/wiki/FAQ/#index25h3 ... In the mean time, if you are interested, you can read PowerManagement and/or contact mupuf on the #nouveau freenode IRC channel. fan fan club

Am 24.07.2014 12:48, schrieb poma: The error report is somewhat ambiguous in the respect, other fans might be affected too. Yes output of sensors with 3.15.6-200.fc20.x86_64 (for reasons I don't know, fan speed of the graphics card GeForce GT 630 is not reported): acpitz-virtual-0 Adapter: Virtual device temp1: +27.8°C (crit = +105.0°C) temp2: +29.8°C (crit = +105.0°C) nouveau-pci-0100 Adapter: PCI adapter temp1: +35.0°C (high = +95.0°C, hyst = +3.0°C) (crit = +105.0°C, hyst = +5.0°C) (emerg = +135.0°C, hyst = +5.0°C) coretemp-isa-0000 Adapter: ISA adapter Physical id 0: +33.0°C (high = +80.0°C, crit = +100.0°C) Core 0: +32.0°C (high = +80.0°C, crit = +100.0°C) Core 1: +32.0°C (high = +80.0°C, crit = +100.0°C) output of sensors with 3.14.8-200.fc20.x86_64 (no noise): acpitz-virtual-0 Adapter: Virtual device temp1: +27.8°C (crit = +105.0°C) temp2: +29.8°C (crit = +105.0°C) nouveau-pci-0100 Adapter: PCI adapter temp1: +34.0°C (high = +95.0°C, hyst = +3.0°C) (crit = +105.0°C, hyst = +5.0°C) (emerg = +135.0°C, hyst = +5.0°C) coretemp-isa-0000 Adapter: ISA adapter Physical id 0: +31.0°C (high = +80.0°C, crit = +100.0°C) Core 0: +30.0°C (high = +80.0°C, crit = +100.0°C) Core 1: +29.0°C (high = +80.0°C, crit = +100.0°C)

On 07/24/14 15:52, poma wrote: ... I really must say that I do not understand why this change was made. Obviously it breaks the fan management for some (me included). Did it solve it for others? And what is 0x46 entries? This is my card (a Quadro 2000): 01:00.0 VGA compatible controller: NVIDIA Corporation GF106GL [Quadro 2000] (rev a1) (prog-if 00 [VGA controller]) Subsystem: NVIDIA Corporation Device 084a Flags: bus master, fast devsel, latency 0, IRQ 45 Memory at f4000000 (32-bit, non-prefetchable) [size=32M] Memory at e0000000 (64-bit, prefetchable) [size=128M] Memory at e8000000 (64-bit, prefetchable) [size=64M] I/O ports at e000 [size=128] Expansion ROM at f6000000 [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: nouveau Kernel modules: nouveau NVC0 family (Fermi) NVC3 (GF106) GeForce GT (440, 445M, 545, 555M, 630M, 635M), GTS 450, GTX 460M, Quadro 2000 (D), 2000M Lars -- Lars E. Pettersson <lars(a)homer.se&gt; http://www.sm6rpz.se/