[Dri-devel] *_drv.c patch, more kudos

2001-04-06 Thread Nicholas Leippe
The latest trunk needs these 1 liner fixes otherwise DRM(sg_cleanup)() ends up as an unresolved symbol rather than an inlined function in both .o and _drv.o for mga, tdfx, gamma, and i810. Both r128 and radeon already had it. add: #include "drm_scatter.h" -- kudos -- This latest pull gets me

Re: [Dri-devel] *_drv.c patch, more kudos

2001-04-06 Thread Alan Hourihane
On Fri, Apr 06, 2001 at 02:09:16AM -0600, Nicholas Leippe wrote: > The latest trunk needs these 1 liner fixes otherwise > DRM(sg_cleanup)() ends up as an unresolved symbol > rather than an inlined function in both .o and > _drv.o for mga, tdfx, gamma, and i810. Both > r128 and radeon already ha

RE: [Dri-devel] cut-and-paste bug in host.def

2001-04-06 Thread Champigny, Michael
While we are on this topic, the flags for Alpha are too restrictive. It should be the more conservative "-mcpu=ev5". By setting it to ev6, older Alpha's will not be able to run the compiled code. ev5 is chosen since it will allow the compiler to schedule instructions better on EV5 and greater proc

Re: [Dri-devel] cut-and-paste bug in host.def

2001-04-06 Thread Alan Hourihane
On Fri, Apr 06, 2001 at 09:42:38AM -0400, Champigny, Michael wrote: > While we are on this topic, the flags for Alpha are too restrictive. > It should be the more conservative "-mcpu=ev5". By setting it to ev6, > older Alpha's will not be able to run the compiled code. ev5 is chosen > since it wil

RE: [Dri-devel] cut-and-paste bug in host.def

2001-04-06 Thread Champigny, Michael
Then it's an accident waiting to happen. Distributions will use the host.def shipped with XFree86 and you'll only get a bug report. I'm just trying to save you some work. It's a 1 byte change, sheesh. -Michael -Original Message- From: Alan Hourihane [mailto:[EMAIL PROTECTED]] Sent: Frida

Re: [Dri-devel] cut-and-paste bug in host.def

2001-04-06 Thread Michel Dänzer
"Champigny, Michael" wrote: > > Then it's an accident waiting to happen. Distributions will use > the host.def shipped with XFree86 and you'll only get a bug report. XFree86 has no host.def, it's a local file from that point of view. -- Earthling Michel Dänzer (MrCooper)\ Debian GNU/Lin

Re: [Dri-devel] cut-and-paste bug in host.def

2001-04-06 Thread AMD-760MP
In light of this info, never mind what I just wrote. - Original Message - From: "Michel Dänzer" <[EMAIL PROTECTED]> To: "Champigny, Michael" <[EMAIL PROTECTED]> Cc: <[EMAIL PROTECTED]> Sent: Friday, 06 April, 2001 10:12 Subject: Re: [Dri-devel] cut-and-paste bug in host.def > "Champigny

Re: [Dri-devel] cut-and-paste bug in host.def

2001-04-06 Thread Kevin E Martin
Actually, XFree86 does not ship with this host.def file, because it is never merged into the XFree86 tree. This particular host.def file is provided as a convenience for us, the DRI developers and the people who test our early development code. The host.def files that are used to build the binar

Re: [Dri-devel] cut-and-paste bug in host.def

2001-04-06 Thread Alan Hourihane
On Fri, Apr 06, 2001 at 09:58:31AM -0400, Champigny, Michael wrote: > Then it's an accident waiting to happen. Distributions will use > the host.def shipped with XFree86 and you'll only get a bug report. > I'm just trying to save you some work. It's a 1 byte change, sheesh. > host.def does NOT ex

Re: [Dri-devel] cut-and-paste bug in host.def

2001-04-06 Thread AMD-760MP
I have to concur with Mr. Champigny on this one. I've always been taught that template configuration files should be lowest common denominator, so that those with a clue can wring out the performance, and those without clues don't waste the time of those with. Of course, on the other hand, I can't

[Dri-devel] kernel headers for drm module building

2001-04-06 Thread Trond Eivind Glomsrød
When building the DRI tree with DRM module for the kernel, the default location to look for headers should be (this is the way Linus started recommending last summer, if memory serves): /lib/modules/`uname -r`/build/include Right now, it looks like it only searches through /usr/src/linux-`unam

Re: [Dri-devel] kernel headers for drm module building

2001-04-06 Thread Alan Hourihane
On Fri, Apr 06, 2001 at 11:04:20AM -0400, Trond Eivind Glomsrød wrote: > When building the DRI tree with DRM module for the kernel, the default > location to look for headers should be (this is the way Linus started > recommending last summer, if memory serves): > > /lib/modules/`uname -r`/build/

[Dri-devel] Support quad-buffering for active stereo

2001-04-06 Thread David Geldreich
Hello all, I am wondering if there is any plan to support quad-buffering in DRI. GL_FRONT_LEFT GL_BACK_LEFT GL_FRONT_RIGHT GL_BACK_RIGHT One of the problems is keeping the flipping of LEFT/RIGHT in sync with the vertical retrace. If you miss one, you will have right/left view swaped

[Dri-devel] RADEON VE patch

2001-04-06 Thread Buddy Smith
Hi. This is a patch i made to add the RADEON VE chipset to the DRI source. It's available at http://foobar.resnet.gatech.edu/~nullset/radeon_ve-DRI-patch I did not post it to the patch list, because i'm not sure if the format is correct. i generated it using diff -u -r It changes: xc/lib/GL/

Re: [Dri-devel] kernel headers for drm module building

2001-04-06 Thread Joseph Carter
On Fri, Apr 06, 2001 at 11:04:20AM -0400, Trond Eivind Glomsr?d wrote: > When building the DRI tree with DRM module for the kernel, the default > location to look for headers should be (this is the way Linus started > recommending last summer, if memory serves): > > /lib/modules/`uname -r`/build/