CSMT patch

[Emanuele Oriani]

Hi Wine Devs/Stefan,

Can you please give us an update in regards to the CSMT patch?

Do you plan to integrate it in mainline? If not "officially", would be 
possible to have it in as an experimental code-path (default off)?
What are the obstacles of making this happen? What is your view?

Again, I believe this patch is superior to just NVIDIA Threaded 
Optimizations; on applications such as Starcraft II I get +150% FPS and 
performance is comparable to windows (if not better).
I believe many other games would benefit from this great improvement.

Cheers,
Emanuele

[Sebastian Lackner]

On 24.12.2014 10:20, Emanuele Oriani wrote:
> Hi Wine Devs/Stefan,
>
> Can you please give us an update in regards to the CSMT patch?
>
> Do you plan to integrate it in mainline? If not "officially", would be possible to have it in as an experimental code-path (default off)?
> What are the obstacles of making this happen? What is your view?
>
> Again, I believe this patch is superior to just NVIDIA Threaded Optimizations; on applications such as Starcraft II I get +150% FPS and performance is comparable to windows (if not better).
> I believe many other games would benefit from this great improvement.
>
> Cheers,
> Emanuele

Hi Emanuele,

I am not directly involved in d3d development, but this is the last information from the mailinglist:

Quote by Stefan Dösinger:
> I stopped working on upstreaming the command stream after you [=Henri Verbeet] made a
> d3d10/11-like resource interface the de-facto requirement for merging
> location management in a (sub-) resource class
> (https://www.winehq.org/pipermail/wine-patches/2013-October/127312.html,
> and later on a series that introduced a subresource class in the wrong
> way). I worked on moving away some surface and volume methods as time
> permitted but gave up on upstreaming the CS until you're done with
> d3d10/d2d.

( https://www.winehq.org/pipermail/wine-devel/2014-October/105376.html )

Since d3d10/d2d is still far away from being complete this most likely means that CSMT will not be included anytime soon.
We from the Wine Staging team decided to include the patches anyway (in Wine >= 1.7.33), and allow to enable/disable the whole CSMT patchset at runtime using winecfg. Fedora users do not even have to install a third party Wine version because our patchset is included in the official one.

See:
* https://github.com/wine-compholio/wine-staging
* https://github.com/wine-compholio/wine-staging/wiki/CSMT
* http://www.wine-staging.com/install.html

The main difference between our patchset and Stefan Dösingers official CSMT patchset out there: When CSMT is disabled in our version, the wined3d code is (almost) identical to upstream Wine. We decided for this approach because for some apps we noticed regressions, even when CSMT was turned off using the registry switch. While adding the patchset we also fixed some obviously incorrect memory accesses:
https://github.com/wine-compholio/wine-staging/blob/master/patches/wined3d-C...

@Stefan Dösinger: I am not sure if you have already seen my latest answer on https://github.com/wine-compholio/wine-staging/issues/201, would be nice if you guys could cooperate to simplify the task of providing a CSMT and a non-CSMT wined3d version. :)

Best regards,
Sebastian

[Stefan Dösinger]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Am 2014-12-24 um 16:50 schrieb Sebastian Lackner:
> @Stefan Dösinger: I am not sure if you have already seen my latest
>  answer on 
> https://github.com/wine-compholio/wine-staging/issues/201, would
> be nice if you guys could cooperate to simplify the task of
> providing a CSMT and a non-CSMT wined3d version. :)

I have seen it, but I've been on holidays for a while. I guess I can
move the patch, but I don't think upstreaming it makes much sense on
its own.

Regarding the memory accesses: I thought I fixed them (or rather,
Matteo fixed them in CrossOver, and I thought I cherry-picked the
fixes into my Wine tree)
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2
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=TOzQ
-----END PGP SIGNATURE-----

[Emanuele Oriani]

Stefan/Wine Devs,

Do you think the CSMT patch will ever make "official" wine?

Cheers

On 25/12/14 20:03, Stefan Dösinger wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Am 2014-12-24 um 16:50 schrieb Sebastian Lackner:
> > @Stefan Dösinger: I am not sure if you have already seen my latest
> >   answer on
> > https://github.com/wine-compholio/wine-staging/issues/201, would
> > be nice if you guys could cooperate to simplify the task of
> > providing a CSMT and a non-CSMT wined3d version. :)
>
> I have seen it, but I've been on holidays for a while. I guess I can
> move the patch, but I don't think upstreaming it makes much sense on
> its own.
>
> Regarding the memory accesses: I thought I fixed them (or rather,
> Matteo fixed them in CrossOver, and I thought I cherry-picked the
> fixes into my Wine tree)
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v2
>
> iQIcBAEBAgAGBQJUnG2OAAoJEN0/YqbEcdMwsIcP/3bNVsO5QMq1tphRLt3Q5Kbz
> EX0kfVKUVsf6QLsico9CyCRJQPr7lkwINWfGYXjud+LLB3pWUwQFrvTB92Q7MS9E
> xdEUU6ZI+Vc7FZLekBjJCroOG8U/v2j8pFF0pGP/iYxIJxn8FIeR1IQ/i5Qlvf4W
> KyUSqEWcFgevr/3FLA7FUgOpjnPdQyavmTBNYLS1TLVqtRcnOJLSWCnhTFwZ4KzN
> XKletUU0SLT10iYjTuvmMXP60SrMBlAqeEOE5p6buLKUZNt6FvN0xPcajL6tEYeH
> Zj3tnP9BnU4a9FLJTqzucOpOU2TtOC749gFIkqLUgyU+seYJfoVGAJxqXPmt1DIq
> 32znlSMchiDGv5lWkEfDwpllHEEV4GFhIzddyX1YqqFaI6Dh8r153yAYvo9fR3Fp
> EKxRAbykIWQPvMbxJJbGbc0k/D+HouCL1yQdP4NdqXN2ps47kwIRXDIyWBY91EVh
> t5QU6ZsB0cjCoPlvdcW7GBMP3ARsqrq6yPEyCZRI+ZS2rIE8jOh+oAS/fBzP7BnE
> XQqHk9otvhFhpqXtG5RKkF9QsCKgzF1d25/o0BqxuLM/5aDez5fW458hnxtSBc3O
> 4Zd184KHvpqxI2Bb7Dxy6FscXTRdaGpcqf2MiHbp8Ju386EfBNtld5Oy6Gv+apGh
> DShS8HE+1ppZkzq6IrR+
> =TOzQ
> -----END PGP SIGNATURE-----

[Stefan Dösinger]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Am 2014-12-27 um 12:45 schrieb Emanuele Oriani:
> Stefan/Wine Devs,
>
> Do you think the CSMT patch will ever make "official" wine?

Yes
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2
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=ImTm
-----END PGP SIGNATURE-----

[Ruslan Kabatsayev]

Hello all,

On Sat, Dec 27, 2014 at 7:29 PM, Stefan Dösinger
 wrote:
> > Do you think the CSMT patch will ever make "official" wine?
>
> Yes

Are you still interested in user feedback for you patch? I'm in doubt
because of your mail to Henry Verbeet:
> I stopped working on upstreaming the command stream after you [=Henri Verbeet] made a
> d3d10/11-like resource interface the de-facto requirement...

Regards,
Ruslan

[Stefan Dösinger]

Am 28.12.2014 um 18:11 schrieb Ruslan Kabatsayev :
> Are you still interested in user feedback for you patch? I'm in doubt
> because of your mail to Henry Verbeet:

Sure, I'm always reading them. I can't promise that I'll fix anything, as the more important task is upstreaming it, and this will require some major changes anyway.