Re: [GATOS]DRI problem
Status: Beta
Brought to you by:
volodya
From: Vladimir D. <vo...@mi...> - 2003-04-16 01:02:49
|
On Tue, 15 Apr 2003, Cookie wrote: > Hi, > > I do not think I will help you with the problem since I have the same > :-(. I think it is not motherbord hw related. For sure it is something > with 4.3.0 which is installed with my rh90. > > I am curious if there are any successfull reports about r128 and ati.2 > binaries working with original rh90 distributed 4.3.0 XFree86. > There was other reports of lockups. For now I would suggest you try recompiling ati.2 from source. The last Redhat X.0 distributions (6.0, 7.0, 8.0 and 9.0) have *all* had compatibility problems with GATOS. GATOS works with hardware and software on a very intimate level. * linux kernel should be very close to what km is dependent against. (certainly not have patches applied that change the number of arguments of common functions) * XFree86 should be compiled from exactly the same code as ati.2 drivers were developed against, with the *same* compiler (current GATOS binaries use gcc 2.95) and the same configuration flags (the latter is responsible for binaries not working on BSDs as is) * the compiler should not be too different from the one used to develop ati.2 and km. When accessing hardware certain code is required to behave *exactly* as developer wants. Different compiler might use different optimizations and queitly change an order of execution thinking (for example) that a=3D1; b=3D1; is the same as b=3D1; a=3D1; - for hardware the order you write values in matters. The way these dependencies are currently handled is by using "standard" versions - i.e. versions released by the developers of applications, be it XFree86, Linux kernel or Tcl/Tk. This has an advantage of being able to ask developers of these applications for help if an unclear issue arises. This puts distributions like Redhat X.0at a disadvantage, as Redhat uses them to preview new technology by taking unfinished or not thoroughly tested code and "brushing it up" so that it performs reasonably well. This results in Redhat X.0 codebase being significantly different from what the developers are using.. The reason Slackware is mentioned is that Patrick uses unmodified packages to form his distribution and the only difference is in particular settings used (like md5 shadow passwords instead of PAM). This provides a very stable base to work against. If someone wants a particular patch applied they remove the regular package and install that software themselves - knowning exactly what the different with "standard" version is. best Vladimir Dergachev > > > > If I set it to 16 or 24, X freezes up and I can only remoting log in to > > my machine, I get the XFree86.0.log file shown down below.. > > It seems I am not the first to have this problem: > > > http://www.mail-archive.com/dri...@li.../msg09877.html > > > > I have an ATI Rage 128 GL RE (PCI). Kernel is 2.4.20. I'm using Xfree > > 4.3.0. I've tried using r128.o from the 2.4.20 kernel and from the > > drm-kernel (cvs latest).. I'm using ati.2 (cvs latest). > > > > Also, I found in dmesg that mtrr was loaded before agpgart which is > > before r128.o: > > > > mtrr: v1.40 (20010327) Richard Gooch (rg...@at...) > > mtrr: detected mtrr type: Intel > > Linux agpgart interface v0.99 (c) Jeff Hartmann > > agpgart: Maximum main memory to use for agp memory: 816M > > agpgart: Detected Via Apollo Pro KT400 chipset > > agpgart: AGP aperture is 128M @ 0xf0000000 > > > > But, I have an AMD motherboard (KT400). > > mtrr type: Intel > > Is this correct? > > > (EE) R128(0): Idle timed out, resetting engine... > > (EE) R128(0): R128CCEWaitForIdle: CCE reset -1007 > > (EE) R128(0): R128CCEWaitForIdle: CCE start -1007 > > .... > > This repeats forever..... > > Mr.Cookie > mailto: mrc...@ce... > > ---------------------------------- > E-mail akceler=E1tor, spam filtr, bezpe=E8nostn=ED agent. Rychle a bezpe= =E8n=EC s Centrum.cz E-mail! http://home.centrum.cz/reklama/email/email_vyh= ody.html > > > > > > ------------------------------------------------------- > This sf.net email is sponsored by:ThinkGeek > Welcome to geek heaven. > http://thinkgeek.com/sf > _______________________________________________ > Gatos-devel mailing list > Gat...@li... > https://lists.sourceforge.net/lists/listinfo/gatos-devel > |