[Nouveau] Atomic modesetting + DisplayPort MST

[Ben Skeggs]

On 11/14/2016 06:17 AM, Pierre Moreau wrote:
> Hello Ben,
> 
> You can find below my results; there is a slight over-representation of 1st gen
> Tesla cards, but… that’s what I have. :-D No regressions observed, apart from:
> 
> * screen rotation on G80, MCP79
> * resuming on G86
Hey Pierre,

Thanks for testing all this!  Notes below.

> 
> I’ll retest those cards with your linux-4.10 branch but without the atomic + DP
> serie.
> 
> Pierre
> 
> 
> Tested: linux-4.10 (HEAD at b27add13f "drm/nouveau/fifo/gf100-: protect channel preempt with subdev mutex)
> Against: 4.8.7
> Fullscreen game: Portal
> (FAIL is: it is not working, but neither did it before.)
> 
> 
> G80:
> * Basic modesetting: OK
> * VT switching:      OK
> * Changing rotation: REGRESSION (disp: ERROR 5 [INVALID_STATE]
ACK, I'll look into this one.

> * Suspend:           OK
> * Hotplugging:       -
> * Fullscreen game:   FAIL (fail ttm_validate, vaildate: -12) (different on 4.8.7, but still failing)
> 
> G84:
> * Basic modesetting: OK
> * VT switching:      OK
> * Changing rotation: OK
> * Suspend:           FAIL (DRM: base-0: timeout)
> * Hotplugging:       OK
> * Fullscreen game:   OK
> 
> G86:
> * Basic modesetting: OK
> * VT switching:      OK
> * Changing rotation: OK
> * Suspend:           REGRESSION (PGRAPH TLB flush idle timeout fail)
Hm, I wouldn't expect something like this to be related.  Is this
reproducible (works reliably before, fails reliably after)?

> * Hotplugging:       -
> * Fullscreen game:   FAIL (gr: TRAP_PROP - TP 0 - 00000040 [RT_FAULT] - Address 004bed0000) (does not work im 4.8.7 either, though there was no trap, just FAIL ttm_validate & cie)
> 
> G92:
> * Basic modesetting: OK
> * VT switching:      OK
> * Changing rotation: OK
> * Suspend:           OK
> * Hotplugging:       -
> * Fullscreen game:   OK
> 
> G94:
> * Basic modesetting: OK
> * VT switching:      OK
> * Changing rotation: OK
> * Suspend:           OK
> * Hotplugging:       -
> * Fullscreen game:   OK
> 
> G96:
> * Basic modesetting: OK
> * VT switching:      OK
> * Changing rotation: OK
> * Suspend:           OK
> * Hotplugging:       -
> * Fullscreen game:   OK
> 
> G98:
> * Basic modesetting: OK
> * VT switching:      OK
> * Changing rotation: OK
> * Suspend:           OK
> * Hotplugging:       OK
> * Fullscreen game:   FAIL (fail ttm_validate, vaildate: -12)
> 
> MCP79:
> * Basic modesetting: OK
> * VT switching:      OK
> * Changing rotation: REGRESSION (no freeze, no errors, cursor was correctly updated, but screen does not update though switching to TTY works just fine)
I don't have the hardware for this, but hopefully whatever I uncover on
G80 will help too :)

Thanks again,
Ben.

> * Suspend:           OK
> * Hotplugging:       -
> * Fullscreen game:   OK
> 
> GT200:
> * Basic modesetting: OK
> * VT switching:      OK
> * Changing rotation: OK
> * Suspend:           FAIL (PGRAPH TLB idle timeout fail)
> * Hotplugging:       -
> * Fullscreen game:   OK
> 
> GF100:
> * Basic modesetting: OK
> * VT switching:      OK
> * Changing rotation: OK
> * Suspend:           OK
> * Hotplugging:       -
> * Fullscreen game:   OK
> 
> GF114:
> * Basic modesetting: OK
> * VT switching:      OK
> * Changing rotation: OK
> * Suspend:           OK
> * Hotplugging:       -
> * Fullscreen game:   OK
> 
> GK208:
> * Basic modesetting: OK
> * VT switching:      OK
> * Changing rotation: OK
> * Suspend:           OK
> * Hotplugging:       OK
> * Fullscreen game:   OK
> 
> GM206:
> * Basic modesetting: OK
> * VT switching:      OK
> * Changing rotation: OK
> * Suspend:           OK
> * Hotplugging:       OK
> * Fullscreen game:   OK
> 
> 
> 
> On 07:41 pm - Nov 04 2016, Ben Skeggs wrote:
>> Hey all,
>>
>> I've just pushed out the initial Nouveau support for $subject \o/
>>
>> As the atomic modesetting transition is basically a rewrite of the KMS
>> portion of the driver, I would be very grateful for any additional
>> testing that people could provide (even as simple as just booting and
>> making sure you get a display is valuable).
>>
>> There's a Linux tree at
>> https://github.com/skeggsb/linux/tree/linux-4.10, and the code also
>> exists in my usual development tree[1] on the master branch (which is
>> currently based on Dave's drm-next tree for Linux 4.10).
>>
>> There's another branch (devel-kms) in the same repository, which is the
>> same code on top of what's currently Linux 4.9.  If you have problems
>> with the 4.10 code, I'd definitely be interested in seeing if they exist
>> on the 4.9 branch too.
>>
>> Thanks in advance!
>> Ben.
>>
>> [1] https://github.com/skeggsb/nouveau
>>
> 
> 
> 
> 
>> _______________________________________________
>> Nouveau mailing list
>> Nouveau at lists.freedesktop.org
>> https://lists.freedesktop.org/mailman/listinfo/nouveau
> 

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 801 bytes
Desc: OpenPGP digital signature
URL: <https://lists.freedesktop.org/archives/nouveau/attachments/20161116/7a4f8441/attachment-0001.sig>