Stefan Dösinger stefan@codeweavers.com writes:
+/* Certain applications(Steam) complain if we report an outdated driver version. In general,
- reporting a driver version is moot because we are not the Windows driver, and we have different
- bugs, features, etc.
- Some drivers(ATI, Nvidia Linux drivers) have a shared codebase with the Windows driver, so we're
- reading the Linux driver version and report it to the app. Some drivers(Apple drivers, Mesa) don't
- have the version match, so we overwrite the OpenGL driver version with a hardcoded one. Obviously
- the hardcoded driver version is outdated once a new Windows driver is released, so we'll have to
- keep them in sync manually.
- If a card is not found in this table, the gl driver version is reported
- */
+struct driver_version_information {
- WORD vendor; /* reported PCI card vendor ID */
- WORD card; /* reported PCI card device ID */
- BOOL loword_matches_apple; /* Are Mac and Win driver versions matching? */
- BOOL loword_matches_linux; /* Are Lin and Win driver versions matching? */
The world is not limited to Mac and Linux, this sort of thing can't scale.
Am Mittwoch, 6. Februar 2008 11:56:40 schrieb Alexandre Julliard:
The world is not limited to Mac and Linux, this sort of thing can't scale.
Official drivers by Nvidia and ATI are limited to Mac, Linux and Windows. Nvidia has Solaris and FreeBSD drivers drivers as well, so maybe the Linux name should be renamed to non_apple.
I don't think we can come up with scheme that will scale to all future developments. In the near future we maybe should add Mesa to that list(open source ATI driver), but if a future event does big unforseen changes to the driver landscape then we might have to change this table.
Stefan Dösinger stefan@codeweavers.com writes:
Am Mittwoch, 6. Februar 2008 11:56:40 schrieb Alexandre Julliard:
The world is not limited to Mac and Linux, this sort of thing can't scale.
Official drivers by Nvidia and ATI are limited to Mac, Linux and Windows. Nvidia has Solaris and FreeBSD drivers drivers as well, so maybe the Linux name should be renamed to non_apple.
I don't think we can come up with scheme that will scale to all future developments. In the near future we maybe should add Mesa to that list(open source ATI driver), but if a future event does big unforseen changes to the driver landscape then we might have to change this table.
I don't see the point of going through all that trouble to return the real driver version on some platforms but not others. Since we have the infrastructure to fake it why not always use that? We'll have to keep it up to date anyway.
Am Mittwoch, 6. Februar 2008 12:47:37 schrieb Alexandre Julliard:
I don't see the point of going through all that trouble to return the real driver version on some platforms but not others. Since we have the infrastructure to fake it why not always use that? We'll have to keep it up to date anyway.
For example, users with matching driver versions will get a correct warning from Steam if their driver is outdated.
Stefan Dösinger stefan@codeweavers.com writes:
Am Mittwoch, 6. Februar 2008 12:47:37 schrieb Alexandre Julliard:
I don't see the point of going through all that trouble to return the real driver version on some platforms but not others. Since we have the infrastructure to fake it why not always use that? We'll have to keep it up to date anyway.
For example, users with matching driver versions will get a correct warning from Steam if their driver is outdated.
And what are the consequences of an outdated driver? Isn't that something that wined3d should be working around anyway? What happens on platforms where version numbers don't match, do things silently break?
Stefan Dösinger stefan@codeweavers.com writes:
Am Mittwoch, 6. Februar 2008 12:47:37 schrieb Alexandre Julliard:
I don't see the point of going through all that trouble to return the real driver version on some platforms but not others. Since we have the infrastructure to fake it why not always use that? We'll have to keep it up to date anyway.
For example, users with matching driver versions will get a correct
warning
from Steam if their driver is outdated.
And what are the consequences of an outdated driver? Isn't that something that wined3d should be working around anyway? What happens on platforms where version numbers don't match, do things silently break?
-- Alexandre Julliard julliard@winehq.org
Personally I don't think we should report the real driver version to the games as we aren't the same as the Windows Direct3D drivers. Second driver version parsing is a lot of work. Just on Linux nvidia already has two different versioning mechanisms, second on e.g. OSX the version string again has a different layout. Faking the driver numbers is the best thing to do in my opinion.
Roderick
Am Mittwoch, 6. Februar 2008 14:23:02 schrieb Roderick Colenbrander:
Stefan Dösinger stefan@codeweavers.com writes:
Am Mittwoch, 6. Februar 2008 12:47:37 schrieb Alexandre Julliard:
I don't see the point of going through all that trouble to return the real driver version on some platforms but not others. Since we have the infrastructure to fake it why not always use that? We'll have to keep it up to date anyway.
For example, users with matching driver versions will get a correct
warning
from Steam if their driver is outdated.
And what are the consequences of an outdated driver? Isn't that something that wined3d should be working around anyway? What happens on platforms where version numbers don't match, do things silently break?
-- Alexandre Julliard julliard@winehq.org
Personally I don't think we should report the real driver version to the games as we aren't the same as the Windows Direct3D drivers. Second driver version parsing is a lot of work. Just on Linux nvidia already has two different versioning mechanisms, second on e.g. OSX the version string again has a different layout. Faking the driver numbers is the best thing to do in my opinion.
Ok, I'll change the patches then to always fake the driver version
-
Alexandre Julliard -
Roderick Colenbrander -
Stefan Dösinger