Re: Debian's XFree86 4.2 and SiS vs. DRI
Michel Dänzer wrote: Is there common code inside sis_dri.so? Yes, the common code is part of all the code. :) OK. (I don't consider my question that stupid :) as common code could reside in other - common - modules as well) I assume as regards 4.3, you mean the inclusion of Mesa4, right? Well, no one of the DRI folks seems to care about the SiS driver, so this is hopeless anyway. It's not the job of the DRI folks to maintain a driver for hardware they don't have. That's exactly what I have been doing for a year now. Implementing support for hardware I don't have (SiS 315, 550, 650, 740, etc) I don't know of any DRI developers having any SiS hardware, and I think this is even harder to do for 3D than it is for 2D. No offence, it probably is. I didn't actually exspect the DRI folks to develope a new driver for yet unsupported chipsets. I just think that porting the existing SiS driver (with support for the 300 series) to Mesa4 API can't be impossible, even without the hardware. And dropping an API without some sort of compatibility layer is not an attitude I especially appreciate. BTW: In the meantime, I have received a number of success reports for SiS-DRI with 4.2.1. These people even used a sis_dri.so binary which was compiled with Debian's source... 8:O Thomas -- Thomas Winischhofer Vienna/Austria mailto:[EMAIL PROTECTED]http://www.winischhofer.net/
Re: Debian's XFree86 4.2 and SiS vs. DRI
Michel Dänzer wrote: Is there common code inside sis_dri.so? Yes, the common code is part of all the code. :) OK. (I don't consider my question that stupid :) as common code could reside in other - common - modules as well) I assume as regards 4.3, you mean the inclusion of Mesa4, right? Well, no one of the DRI folks seems to care about the SiS driver, so this is hopeless anyway. It's not the job of the DRI folks to maintain a driver for hardware they don't have. That's exactly what I have been doing for a year now. Implementing support for hardware I don't have (SiS 315, 550, 650, 740, etc) I don't know of any DRI developers having any SiS hardware, and I think this is even harder to do for 3D than it is for 2D. No offence, it probably is. I didn't actually exspect the DRI folks to develope a new driver for yet unsupported chipsets. I just think that porting the existing SiS driver (with support for the 300 series) to Mesa4 API can't be impossible, even without the hardware. And dropping an API without some sort of compatibility layer is not an attitude I especially appreciate. BTW: In the meantime, I have received a number of success reports for SiS-DRI with 4.2.1. These people even used a sis_dri.so binary which was compiled with Debian's source... 8:O Thomas -- Thomas Winischhofer Vienna/Austria mailto:[EMAIL PROTECTED]http://www.winischhofer.net/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Debian's XFree86 4.2 and SiS vs. DRI
On Don, 2003-01-02 at 01:54, Thomas Winischhofer wrote: > Michel Dänzer wrote: > > >>There were no significant changes in the code either, as I saw in my > >>comparison. (Mesa is of the same version, too.) > > > > I guess it's still possible that the interface between the common client > > side DRI code and the drivers was changed in a way that renders the SiS > > driver non-working. > > Is there common code inside sis_dri.so? Yes, the common code is part of all the code. :) > >>I assume as regards 4.3, you mean the inclusion of Mesa4, right? Well, > >>no one of the DRI folks seems to care about the SiS driver, so this is > >>hopeless anyway. > > > > > > It's not the job of the DRI folks to maintain a driver for hardware they > > don't have. > > That's exactly what I have been doing for a year now. Implementing > support for hardware I don't have (SiS 315, 550, 650, 740, etc) I don't know of any DRI developers having any SiS hardware, and I think this is even harder to do for 3D than it is for 2D. -- Earthling Michel Dänzer (MrCooper)/ Debian GNU/Linux (powerpc) developer XFree86 and DRI project member / CS student, Free Software enthusiast
Re: Debian's XFree86 4.2 and SiS vs. DRI
Michel Dänzer wrote: On Die, 2002-12-31 at 10:31, Thomas Winischhofer wrote: Michel Dänzer wrote: Is SiS DRI even supposed to work in 4.2.x? I know that it most definitely won't work in 4.3.0 because it hasn't been maintained in a while. It is, I assume. Assumptions tend to be flawed. :) Have there been success reports with the 4.2.x sis_dri.so? That was my question... > If so, you should look at diffs between upstream and the Debian packages in lib/GL and extras/Mesa. I'll do that asap. There were no significant changes in the code either, as I saw in my comparison. (Mesa is of the same version, too.) I guess it's still possible that the interface between the common client side DRI code and the drivers was changed in a way that renders the SiS driver non-working. Is there common code inside sis_dri.so? I assume as regards 4.3, you mean the inclusion of Mesa4, right? Well, no one of the DRI folks seems to care about the SiS driver, so this is hopeless anyway. It's not the job of the DRI folks to maintain a driver for hardware they don't have. That's exactly what I have been doing for a year now. Implementing support for hardware I don't have (SiS 315, 550, 650, 740, etc) Thomas -- Thomas Winischhofer Vienna/Austria mailto:[EMAIL PROTECTED] *** http://www.winischhofer.net
Re: Debian's XFree86 4.2 and SiS vs. DRI
On Don, 2003-01-02 at 01:54, Thomas Winischhofer wrote: > Michel Dänzer wrote: > > >>There were no significant changes in the code either, as I saw in my > >>comparison. (Mesa is of the same version, too.) > > > > I guess it's still possible that the interface between the common client > > side DRI code and the drivers was changed in a way that renders the SiS > > driver non-working. > > Is there common code inside sis_dri.so? Yes, the common code is part of all the code. :) > >>I assume as regards 4.3, you mean the inclusion of Mesa4, right? Well, > >>no one of the DRI folks seems to care about the SiS driver, so this is > >>hopeless anyway. > > > > > > It's not the job of the DRI folks to maintain a driver for hardware they > > don't have. > > That's exactly what I have been doing for a year now. Implementing > support for hardware I don't have (SiS 315, 550, 650, 740, etc) I don't know of any DRI developers having any SiS hardware, and I think this is even harder to do for 3D than it is for 2D. -- Earthling Michel Dänzer (MrCooper)/ Debian GNU/Linux (powerpc) developer XFree86 and DRI project member / CS student, Free Software enthusiast -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Debian's XFree86 4.2 and SiS vs. DRI
Michel Dänzer wrote: On Die, 2002-12-31 at 10:31, Thomas Winischhofer wrote: Michel Dänzer wrote: Is SiS DRI even supposed to work in 4.2.x? I know that it most definitely won't work in 4.3.0 because it hasn't been maintained in a while. It is, I assume. Assumptions tend to be flawed. :) Have there been success reports with the 4.2.x sis_dri.so? That was my question... > If so, you should look at diffs between upstream and the Debian packages in lib/GL and extras/Mesa. I'll do that asap. There were no significant changes in the code either, as I saw in my comparison. (Mesa is of the same version, too.) I guess it's still possible that the interface between the common client side DRI code and the drivers was changed in a way that renders the SiS driver non-working. Is there common code inside sis_dri.so? I assume as regards 4.3, you mean the inclusion of Mesa4, right? Well, no one of the DRI folks seems to care about the SiS driver, so this is hopeless anyway. It's not the job of the DRI folks to maintain a driver for hardware they don't have. That's exactly what I have been doing for a year now. Implementing support for hardware I don't have (SiS 315, 550, 650, 740, etc) Thomas -- Thomas Winischhofer Vienna/Austria mailto:[EMAIL PROTECTED] *** http://www.winischhofer.net -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Debian's XFree86 4.2 and SiS vs. DRI
On Die, 2002-12-31 at 10:31, Thomas Winischhofer wrote: > Michel Dänzer wrote: > > Is SiS DRI even supposed to work in 4.2.x? I know that it most > > definitely won't work in 4.3.0 because it hasn't been maintained in a > > while. > > It is, I assume. Assumptions tend to be flawed. :) Have there been success reports with the 4.2.x sis_dri.so? If so, you should look at diffs between upstream and the Debian packages in lib/GL and extras/Mesa. > Otherwise how should one explain why just exchanging > sis_dri.so makes it work? The logical consequence can only be that there > were no changes in the DR infrastructure. No, that file contains all the client side DRI code. > There were no significant changes in the code either, as I saw in my > comparison. (Mesa is of the same version, too.) I guess it's still possible that the interface between the common client side DRI code and the drivers was changed in a way that renders the SiS driver non-working. > I assume as regards 4.3, you mean the inclusion of Mesa4, right? Well, > no one of the DRI folks seems to care about the SiS driver, so this is > hopeless anyway. It's not the job of the DRI folks to maintain a driver for hardware they don't have. -- Earthling Michel Dänzer (MrCooper)/ Debian GNU/Linux (powerpc) developer XFree86 and DRI project member / CS student, Free Software enthusiast
Re: Debian's XFree86 4.2 and SiS vs. DRI
On Die, 2002-12-31 at 10:31, Thomas Winischhofer wrote: > Michel Dänzer wrote: > > Is SiS DRI even supposed to work in 4.2.x? I know that it most > > definitely won't work in 4.3.0 because it hasn't been maintained in a > > while. > > It is, I assume. Assumptions tend to be flawed. :) Have there been success reports with the 4.2.x sis_dri.so? If so, you should look at diffs between upstream and the Debian packages in lib/GL and extras/Mesa. > Otherwise how should one explain why just exchanging > sis_dri.so makes it work? The logical consequence can only be that there > were no changes in the DR infrastructure. No, that file contains all the client side DRI code. > There were no significant changes in the code either, as I saw in my > comparison. (Mesa is of the same version, too.) I guess it's still possible that the interface between the common client side DRI code and the drivers was changed in a way that renders the SiS driver non-working. > I assume as regards 4.3, you mean the inclusion of Mesa4, right? Well, > no one of the DRI folks seems to care about the SiS driver, so this is > hopeless anyway. It's not the job of the DRI folks to maintain a driver for hardware they don't have. -- Earthling Michel Dänzer (MrCooper)/ Debian GNU/Linux (powerpc) developer XFree86 and DRI project member / CS student, Free Software enthusiast -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Debian's XFree86 4.2 and SiS vs. DRI
Michel Dänzer wrote: Is SiS DRI even supposed to work in 4.2.x? I know that it most definitely won't work in 4.3.0 because it hasn't been maintained in a while. It is, I assume. Otherwise how should one explain why just exchanging sis_dri.so makes it work? The logical consequence can only be that there were no changes in the DR infrastructure. There were no significant changes in the code either, as I saw in my comparison. (Mesa is of the same version, too.) I assume as regards 4.3, you mean the inclusion of Mesa4, right? Well, no one of the DRI folks seems to care about the SiS driver, so this is hopeless anyway. 2) Any clue on why this happens? Does running glxinfo with LIBGL_DEBUG=verbose give a hint? Unfortunately no. Output looks exactly the same with the 4.1 (=working) and 4.2 (=non-working) version of sis_dri.so. The only difference is that with 4.2, Direct Rendering says "no" (hence "Indirect Rendering")... > Otherwise, you may want to post to dri-devel. Allright, thanks. Thomas -- Thomas Winischhofer Vienna/Austria mailto:[EMAIL PROTECTED]http://www.winischhofer.net/
Re: Debian's XFree86 4.2 and SiS vs. DRI
Michel Dänzer wrote: Is SiS DRI even supposed to work in 4.2.x? I know that it most definitely won't work in 4.3.0 because it hasn't been maintained in a while. It is, I assume. Otherwise how should one explain why just exchanging sis_dri.so makes it work? The logical consequence can only be that there were no changes in the DR infrastructure. There were no significant changes in the code either, as I saw in my comparison. (Mesa is of the same version, too.) I assume as regards 4.3, you mean the inclusion of Mesa4, right? Well, no one of the DRI folks seems to care about the SiS driver, so this is hopeless anyway. 2) Any clue on why this happens? Does running glxinfo with LIBGL_DEBUG=verbose give a hint? Unfortunately no. Output looks exactly the same with the 4.1 (=working) and 4.2 (=non-working) version of sis_dri.so. The only difference is that with 4.2, Direct Rendering says "no" (hence "Indirect Rendering")... > Otherwise, you may want to post to dri-devel. Allright, thanks. Thomas -- Thomas Winischhofer Vienna/Austria mailto:[EMAIL PROTECTED]http://www.winischhofer.net/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Debian's XFree86 4.2 and SiS vs. DRI
On Mon, 2002-12-30 at 16:13, Thomas Winischhofer wrote: > I finally updated to 4.2.1 (-4) from 4.1.0.1 and found that DRI on the > SiS chipsets does not work anymore. > > My box is running kernel 2.4.19 (with xfs patch, otherwise vanilla) and > the latest XFree 4.2.1-4 packages from unstable. > > The (XFree) SiS driver enables DRI correctly, but glxgears says "Direct > Rendering: No". Consequently, indirect rendering is used. > > Replacing the 4.2.1-4 xlibsmesa3 package with the one from stable > (4.1.0-16) makes DRI work again. To be more exact, replacing 4.2.1's > sis_dri.so with the 4.1 version is enough. > > A comparison of the source between the 4.1 and 4.2 version (as regards > sis_dri.so) shows as good as no changes, except in the (I)Makefile(s). > > 1) Has anyone managed to make DRI on the SiS chipsets work with Debian's > 4.2 packages? Is SiS DRI even supposed to work in 4.2.x? I know that it most definitely won't work in 4.3.0 because it hasn't been maintained in a while. > 2) Any clue on why this happens? Does running glxinfo with LIBGL_DEBUG=verbose give a hint? Otherwise, you may want to post to dri-devel. -- Earthling Michel Dänzer (MrCooper)/ Debian GNU/Linux (powerpc) developer XFree86 and DRI project member / CS student, Free Software enthusiast
Re: Debian's XFree86 4.2 and SiS vs. DRI
On Mon, 2002-12-30 at 16:13, Thomas Winischhofer wrote: > I finally updated to 4.2.1 (-4) from 4.1.0.1 and found that DRI on the > SiS chipsets does not work anymore. > > My box is running kernel 2.4.19 (with xfs patch, otherwise vanilla) and > the latest XFree 4.2.1-4 packages from unstable. > > The (XFree) SiS driver enables DRI correctly, but glxgears says "Direct > Rendering: No". Consequently, indirect rendering is used. > > Replacing the 4.2.1-4 xlibsmesa3 package with the one from stable > (4.1.0-16) makes DRI work again. To be more exact, replacing 4.2.1's > sis_dri.so with the 4.1 version is enough. > > A comparison of the source between the 4.1 and 4.2 version (as regards > sis_dri.so) shows as good as no changes, except in the (I)Makefile(s). > > 1) Has anyone managed to make DRI on the SiS chipsets work with Debian's > 4.2 packages? Is SiS DRI even supposed to work in 4.2.x? I know that it most definitely won't work in 4.3.0 because it hasn't been maintained in a while. > 2) Any clue on why this happens? Does running glxinfo with LIBGL_DEBUG=verbose give a hint? Otherwise, you may want to post to dri-devel. -- Earthling Michel Dänzer (MrCooper)/ Debian GNU/Linux (powerpc) developer XFree86 and DRI project member / CS student, Free Software enthusiast -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Debian's XFree86 4.2 and SiS vs. DRI
I finally updated to 4.2.1 (-4) from 4.1.0.1 and found that DRI on the SiS chipsets does not work anymore. My box is running kernel 2.4.19 (with xfs patch, otherwise vanilla) and the latest XFree 4.2.1-4 packages from unstable. The (XFree) SiS driver enables DRI correctly, but glxgears says "Direct Rendering: No". Consequently, indirect rendering is used. Replacing the 4.2.1-4 xlibsmesa3 package with the one from stable (4.1.0-16) makes DRI work again. To be more exact, replacing 4.2.1's sis_dri.so with the 4.1 version is enough. A comparison of the source between the 4.1 and 4.2 version (as regards sis_dri.so) shows as good as no changes, except in the (I)Makefile(s). 1) Has anyone managed to make DRI on the SiS chipsets work with Debian's 4.2 packages? 2) Any clue on why this happens? After a year of explaining SiS- and SiS-DRI related stuff to people I feel somewhat stupid running into such a problem myself now... Thomas -- Thomas Winischhofer Vienna/Austria mailto:[EMAIL PROTECTED]http://www.winischhofer.net/
Debian's XFree86 4.2 and SiS vs. DRI
I finally updated to 4.2.1 (-4) from 4.1.0.1 and found that DRI on the SiS chipsets does not work anymore. My box is running kernel 2.4.19 (with xfs patch, otherwise vanilla) and the latest XFree 4.2.1-4 packages from unstable. The (XFree) SiS driver enables DRI correctly, but glxgears says "Direct Rendering: No". Consequently, indirect rendering is used. Replacing the 4.2.1-4 xlibsmesa3 package with the one from stable (4.1.0-16) makes DRI work again. To be more exact, replacing 4.2.1's sis_dri.so with the 4.1 version is enough. A comparison of the source between the 4.1 and 4.2 version (as regards sis_dri.so) shows as good as no changes, except in the (I)Makefile(s). 1) Has anyone managed to make DRI on the SiS chipsets work with Debian's 4.2 packages? 2) Any clue on why this happens? After a year of explaining SiS- and SiS-DRI related stuff to people I feel somewhat stupid running into such a problem myself now... Thomas -- Thomas Winischhofer Vienna/Austria mailto:[EMAIL PROTECTED]http://www.winischhofer.net/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]