Re: drm/nouveau contiguous bo check produces lots of output

Re: drm/nouveau contiguous bo check produces lots of output

[Maarten Lankhorst]

op 11-11-13 16:31, Jerry Cooperstein schreef:
> Hi:
>
> commit bd9c5a2016307164c419c5e24a46921c10e620a0
>
> (drm/nouveau: require contiguous bo for framebuffer)
> produces about 2000 lines of
>
> nouveau E[     DRM] framebuffer requires contiguous bo
>
> on system boot and more on shutdown, but I don't see other
> negative effects.  I tried deploying a trivial patch removing the check and saw
> no other problems.
>
> Is this a problem you have heard about?
>
> My hardware is an NVIDIA Geoforce G98 (8400 GS), lspci -v gives:
>
> 01:00.0 VGA compatible controller: NVIDIA Corporation G98 [GeForce 8400 GS] (rev a1) (prog-if 00 [VGA controller])
>     Subsystem: ASUSTeK Computer Inc. Device 8278
>     Flags: bus master, fast devsel, latency 0, IRQ 16
>     Memory at fd000000 (32-bit, non-prefetchable) [size=16M]
>     Memory at d0000000 (64-bit, prefetchable) [size=256M]
>     Memory at fa000000 (64-bit, non-prefetchable) [size=32M]
>     I/O ports at cc00 [size=128]
>     Expansion ROM at fe9e0000 [disabled] [size=128K]
>     Capabilities: [60] Power Management version 3
>     Capabilities: [68] MSI: Enable- Count=1/1 Maskable- 64bit+
>     Capabilities: [78] Express Endpoint, MSI 00
>     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
>
> This appeared in the 3.12 rc kernels and is in 3.12.0
>
> If there is anything I can do to help you debug this please let me know.
> I don't understand the drm code well enough to dive in deep on my own.
It's a paranoia patch and any bo should not be created with NOUVEAU_GEM_TILE_NONCONTIG flag.
Even if it happens to work it will fail on any sufficiently fragmented vram.

What driver versions do you use?

~Maarten

Re: drm/nouveau contiguous bo check produces lots of output

[Jerry Cooperstein]

On 11/11/2013 09:57 AM, Maarten Lankhorst wrote:
> op 11-11-13 16:31, Jerry Cooperstein schreef:
>> Hi:
>>
>> commit bd9c5a2016307164c419c5e24a46921c10e620a0
>>
>> (drm/nouveau: require contiguous bo for framebuffer)
>> produces about 2000 lines of
>>
>> nouveau E[     DRM] framebuffer requires contiguous bo
>>
>> on system boot and more on shutdown, but I don't see other
>> negative effects.  I tried deploying a trivial patch removing the check and saw
>> no other problems.
>>
>> Is this a problem you have heard about?
>>
>> My hardware is an NVIDIA Geoforce G98 (8400 GS), lspci -v gives:
>>
>> 01:00.0 VGA compatible controller: NVIDIA Corporation G98 [GeForce 8400 GS] (rev a1) (prog-if 00 [VGA controller])
>>      Subsystem: ASUSTeK Computer Inc. Device 8278
>>      Flags: bus master, fast devsel, latency 0, IRQ 16
>>      Memory at fd000000 (32-bit, non-prefetchable) [size=16M]
>>      Memory at d0000000 (64-bit, prefetchable) [size=256M]
>>      Memory at fa000000 (64-bit, non-prefetchable) [size=32M]
>>      I/O ports at cc00 [size=128]
>>      Expansion ROM at fe9e0000 [disabled] [size=128K]
>>      Capabilities: [60] Power Management version 3
>>      Capabilities: [68] MSI: Enable- Count=1/1 Maskable- 64bit+
>>      Capabilities: [78] Express Endpoint, MSI 00
>>      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
>>
>> This appeared in the 3.12 rc kernels and is in 3.12.0
>>
>> If there is anything I can do to help you debug this please let me know.
>> I don't understand the drm code well enough to dive in deep on my own.
> It's a paranoia patch and any bo should not be created with NOUVEAU_GEM_TILE_NONCONTIG flag.
> Even if it happens to work it will fail on any sufficiently fragmented vram.
>
> What driver versions do you use?
>
> ~Maarten
>
>

I'm just using vanilla 3.12 kernel. Been using the nouveau driver
for a long time, no problems until this printout due to the
contiguous check.  Not using any of the "nvidia" related modules etc.
CONFIG_AGP is set.

lsmod | grep nouveau

nouveau               895494  2
ttm                    67720  1 nouveau
drm_kms_helper         38367  1 nouveau
drm                   258311  4 nouveau,ttm,drm_kms_helper
i2c_algo_bit            5122  1 nouveau
mxm_wmi                 1769  1 nouveau
wmi                     9036  2 nouveau,mxm_wmi

no other hardware other than built in intel hda sound, ahci disk,
and marvel GB NIC's.  Asus motherboard.

-- 
============================================================
  Jerry Cooperstein, PhD
  Training Program Director
  Linux Foundation
  coop@linuxfoundation.org
  (608) 848-1507 (Madison, WI, USA)
============================================================

Re: drm/nouveau contiguous bo check produces lots of output

[Maarten Lankhorst]

op 11-11-13 17:05, Jerry Cooperstein schreef:
> On 11/11/2013 09:57 AM, Maarten Lankhorst wrote:
>> op 11-11-13 16:31, Jerry Cooperstein schreef:
>>> Hi:
>>>
>>> commit bd9c5a2016307164c419c5e24a46921c10e620a0
>>>
>>> (drm/nouveau: require contiguous bo for framebuffer)
>>> produces about 2000 lines of
>>>
>>> nouveau E[     DRM] framebuffer requires contiguous bo
>>>
>>> on system boot and more on shutdown, but I don't see other
>>> negative effects.  I tried deploying a trivial patch removing the check and saw
>>> no other problems.
>>>
>>> Is this a problem you have heard about?
>>>
>>> My hardware is an NVIDIA Geoforce G98 (8400 GS), lspci -v gives:
>>>
>>> 01:00.0 VGA compatible controller: NVIDIA Corporation G98 [GeForce 8400 GS] (rev a1) (prog-if 00 [VGA controller])
>>>      Subsystem: ASUSTeK Computer Inc. Device 8278
>>>      Flags: bus master, fast devsel, latency 0, IRQ 16
>>>      Memory at fd000000 (32-bit, non-prefetchable) [size=16M]
>>>      Memory at d0000000 (64-bit, prefetchable) [size=256M]
>>>      Memory at fa000000 (64-bit, non-prefetchable) [size=32M]
>>>      I/O ports at cc00 [size=128]
>>>      Expansion ROM at fe9e0000 [disabled] [size=128K]
>>>      Capabilities: [60] Power Management version 3
>>>      Capabilities: [68] MSI: Enable- Count=1/1 Maskable- 64bit+
>>>      Capabilities: [78] Express Endpoint, MSI 00
>>>      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
>>>
>>> This appeared in the 3.12 rc kernels and is in 3.12.0
>>>
>>> If there is anything I can do to help you debug this please let me know.
>>> I don't understand the drm code well enough to dive in deep on my own.
>> It's a paranoia patch and any bo should not be created with NOUVEAU_GEM_TILE_NONCONTIG flag.
>> Even if it happens to work it will fail on any sufficiently fragmented vram.
>>
>> What driver versions do you use?
>>
>> ~Maarten
>>
>>
>
> I'm just using vanilla 3.12 kernel. Been using the nouveau driver
> for a long time, no problems until this printout due to the
> contiguous check.  Not using any of the "nvidia" related modules etc.
> CONFIG_AGP is set.
Sorry I meant userspace, libdrm, mesa, xserver-xorg-video-nouveau? Using Xorg, mir or wayland?