Dear Martin: I hope you know that we sit in awe and admiration of your great effort and accomplishment in developing all this Xorg driver support, and that it will be a great benefit to many who wish to continue using these venerable machines in an uptodate desktop environment.
Stuart Kreitman Sun X group Martin Bochnig wrote: >> - SPARC graphics support limited to the devices >> supported by the >> Xorg drivers we're getting from Martin Bochnig. >> This is more >> likely to cover the older machines that "hobbiests" >> have than >> the latest Ultra workstations, but that may be a >> better fit for >> Indiana at first anyway. >> > > Hi, > > I initially wanted to release this stuff last year, then in January, then on > March 31st, then before April 30th, then "sometimes soon", then in [...] and > then "before September hits". > Now we have October here (Mon Oct 1 01:20:39 MEST 2007). > Except for the first week of September, when I had been traveling through > wild wild Eastern Europe, I have continued to work on improving my private > fox-gate in *exclusive* mode, every day, every night. No friends. No cinema. > No job. NOTHING else. It's not so time consuming to add a few client > programs, even had-writing pkgdefs which do involve some more complexity. But > rather to ensure, that the Xorg Foundation written drivers && my diffs > against them really work on every pci bridge (bus scanning !!!) and with > every of the now supported SPARC frame buffers. Then comes another hurdle: > dtlogin doesn't properly start in a few configurations such as on the U80 > with whatever frame buffer. And pci frame buffers like the PGX32 don't work > at all on the U80, aperture mmap() related problem, in this precise way only > present on the U80. If you happen to have an Ultra80, be glad to hear that > the now supported UPA frame buffers (afb and ffb) do work there, except that > dtlogin cannot b! e u > sed. > I have fought against the most weird SIGSEGV's and SIGBUS's. System > lock-up's. AFAR CPU-errors. Kernel panics. Most (happy) individuals have > thankfully never seen things like that. > But I have claimed victory against most of those things. Which costed > substantial amounts of time, effort and money, and health. Even UPA frame > buffers are found now (bus scanning). That stuff had been written for Linux > circa in 2000/2001. But no-one ever got these drivers to function on Solaris. > Also the TADPOLE laptops (with m64 chips) should function now, anyways my > SPARCLE500SX does. That's not only a Solaris premiere, but new in general: > According to SPARC-*BSD's, -Linux's release notes and man pages, Xorg had > been considered completely unsupported on the TADPOLE's, by any OS. > The changes that I made are not very big in length/number of lines, but those > things have been a street-fight for every line, if not letter. > > The following systems are known to work now (mostly a question of whether the > pci bridge is supported by sparcPci.c, where pci is present) : > > U1 (UPA) : FFB1/+) > U2 (UPA) : FFB1/+) > U3 Mobile Workstation (the ones based onTadpole SPARCLE) : Onboard PGX64 > U5/U10: Onboard PGX8, Onboard PGX24, Plugged in PGX8, PGX32, PGX64, XVR-100 > U30: (UPA) : all fitting Creator and Elite3D models, XVR-100, other pci-fb's > may not work > U60: (UPA) : all fitting Creator and Elite3D models, PGX8, PGX32, PGX64, > XVR-100 > U80: (UPA) : all fitting Creator and Elite3D models, PGX32 does *not* work, > rest untested > SunBlade100/150: Onboard-PGX64, Plugged in PGX8, PGX32, PGX64, XVR-100 > (never press Stop+A when Xorg is running !!!) > SunBlade1000/2000: all fitting Creator and Elite3D models, Plugged in PGX8, > PGX32, PGX64, XVR-100 > SF280R/Netra20: Exclusively the PGX32 (same board as SB1000/2000, *but* > different OBP-sub-image selected on boot, therefore ...) > > Otherwise, the following cards/chipsets do have Xorg driver support now, but > may not work on every system. E.g. the SB1500 Red/Silver && SB2500 Red/Silver > && U25 && U45 && T1000 && T2000 have pci-bridges that sparcPci.c does not > support as of now. Plus the CG6 doesn't work yet, as I did get UPA scanning > to function, but not legacy Sbus scanning: > > #UPA > > NAME *** Elite3D-m3 / Elite3D-m6 / AFB Series 1 > CHIPSET sunffb > SERVER SVGA > DRIVER sunffb > > NAME *** Elite3D-m3 / Elite3D-m6 / AFB Series 2 > CHIPSET sunffb > SERVER SVGA > DRIVER sunffb > > NAME *** Elite3D-m6 / AFB (Ultra 2, 450) > CHIPSET sunffb > SERVER SVGA > DRIVER sunffb > > NAME *** Creator Series 1 (FFB 67MHz clock, Ultra 1E, 2) > CHIPSET sunffb > SERVER SVGA > DRIVER sunffb > > NAME *** Creator3D Series 1 (FFB 67MHz clock, Ultra 1E, 2) > CHIPSET sunffb > SERVER SVGA > DRIVER sunffb > > NAME *** Creator3D Series 1 (FFB 75MHz clock, Ultra 1E, 2) > CHIPSET sunffb > SERVER SVGA > DRIVER sunffb > > NAME *** Creator Series 2 (FFB2 Ultra 30, 60) > CHIPSET sunffb > SERVER SVGA > DRIVER sunffb > > NAME *** Creator3D Series 2 (FFB2 Ultra 1E, 2) > CHIPSET sunffb > SERVER SVGA > DRIVER sunffb > NAME *** Creator3D Series 2 (FFB2 Ultra 30, 60) > CHIPSET sunffb > SERVER SVGA > DRIVER sunffb > > NAME *** Creator Series 3 (FFB2+ Ultra 10, 30, 60) > CHIPSET sunffb > SERVER SVGA > DRIVER sunffb > > NAME *** Creator3D Series 3 (FFB2+ Ultra 2) > CHIPSET sunffb > SERVER SVGA > DRIVER sunffb > > NAME *** Creator3D Series 3 (FFB2+ Ultra 10, 30, 60, Blade 1000/2000) > CHIPSET sunffb > SERVER SVGA > DRIVER sunffb > > > #SBUS > > NAME *** GX 8-Bit Color Frame Buffer > CHIPSET suncg6 > SERVER SVGA > DRIVER suncg6 > > NAME *** GXplus 8-Bit Color Frame Buffer > CHIPSET suncg6 > SERVER SVGA > DRIVER suncg6 > > NAME *** TurboGX 8-Bit Color Frame Buffer > CHIPSET suncg6 > SERVER SVGA > DRIVER suncg6 > > NAME *** TurboGXplus 8-Bit Color Frame Buffer > CHIPSET suncg6 > SERVER SVGA > DRIVER suncg6 > > > #PCI > > NAME *** PGX 8-Bit Color Frame Buffer (ATI Mach64 3D RAGE II 2MB) > CHIPSET ati > SERVER Mach64 > DRIVER ati > LINE VideoRam 4096 > NOCLOCKPROBE > > NAME *** PGX24 8/24-Bit Onboard Color Frame Buffer (ATI Mach64 3D Rage Pro > 4MB SGRAM) > CHIPSET ati > SERVER Mach64 > DRIVER ati > LINE VideoRam 4096 > NOCLOCKPROBE > > NAME *** PGX32 8/24-Bit - Techsource Raptor 8p (Permedia2 8MB) > CHIPSET PERMEDIA 2 > SERVER 3DLabs > DRIVER glint > LINE VideoRam 8192 > NOCLOCKPROBE > > NAME *** PGX64 8/24-Bit Color Frame Buffer (ATI RAGE XL 8 MB) > CHIPSET ati > SERVER Mach64 > DRIVER ati > LINE VideoRam 8192 > NOCLOCKPROBE > > NAME *** Ultra[tm] 5/10 Workstation Onboard PGX (8-Bit) > CHIPSET ati > SERVER Mach64 > DRIVER ati > LINE VideoRam 2048 > NOCLOCKPROBE > > NAME *** Ultra[tm] 5/10 Workstation Onboard PGX24 (8/24-Bit) > CHIPSET ati > SERVER Mach64 > DRIVER ati > LINE VideoRam 4096 > NOCLOCKPROBE > > NAME *** Sun Blade[tm] 100 Workstation Onboard PGX64 (8/24-Bit) > CHIPSET ati > SERVER Mach64 > DRIVER ati > LINE VideoRam 8192 > NOCLOCKPROBE > > NAME *** Sun Blade[tm] 150 Workstation Onboard PGX64 (8/24-Bit) > CHIPSET ati > SERVER Mach64 > DRIVER ati > LINE VideoRam 8192 > NOCLOCKPROBE > > NAME *** Sun Ultra[tm] 3 Mobile Workstation Onboard PGX64 (8/24-Bit) > CHIPSET ati > SERVER Mach64 > DRIVER ati > LINE VideoRam 8192 > NOCLOCKPROBE > > NAME *** Tadpole SPARCle[tm] 500SX Mobile Workstation Onboard PGX64 (8/24-Bit) > CHIPSET ati > SERVER Mach64 > DRIVER ati > LINE VideoRam 8192 > NOCLOCKPROBE > > NAME *** XVR-100 Graphics Accelerator 24-Bit (32MB ATI Radeon 7000 MAC > Edition 3.3/5V 32Bit 33/66MHz) > CHIPSET radeon > SERVER SVGA > DRIVER radeon > LINE VideoRam 32768 > LINE option "vgaaccess" "false" ### needed by SPARC_snv_4x++ for some reason > LINE option "noint10" "true" ### needed by SPARC_snv_4x++ for some reason > > NAME *** XVR-100 Graphics Accelerator 24-Bit (64MB ATI Radeon 7000 MAC > Edition 3.3/5V 32Bit 33/66MHz RoHS:Y) > CHIPSET radeon > SERVER SVGA > DRIVER radeon > LINE VideoRam 65536 > LINE option "vgaaccess" "false" ### needed by SPARC_snv_4x++ for some reason > LINE option "noint10" "true" ### needed by SPARC_snv_4x++ for some reason > > > #PCI Express > > NAME *** XVR-200 Sun Fire T2000 (x1 PCI Express MatroxGraphics G55-MDDE32LPDF > RoHS:Y) > CHIPSET mgag550 > SERVER SVGA > DRIVER mga > LINE VideoRam 32768 > LINE # Option "mga_sdram" > NOCLOCKPROBE > > > NAME *** XVR-300 Graphics Accelerator (ATI FireMV 2200 x16 PCI Express RoHS:Y) > CHIPSET radeon > SERVER SVGA > DRIVER radeon > LINE VideoRam 131072 > LINE option "vgaaccess" "false" ### needed by SPARC_snv_4x++ for some reason > LINE option "noint10" "true" ### needed by SPARC_snv_4x++ for some reason > > > ### Don't wonder, I could have shortened the file dramatically by using SEE > references. ### > ### But I plan to set up _specific_ values for each config, lots of testing > involved. ### > > > ############## > ############## > > > The following packages are today's snapshot. They still feature many bugs, in > many cases packaging related (most things hand-edited). Then a few missing > symbols, due to wrong lib version, missing symlinks (which cannot be created > during pkgadd'ing, because of conflicts with existing plain files or real > directories). > Then: The xkb stuff thought to replace Sun's (non-redistributable) > implementation of SUNW-xkb does not properly detect most Sun-kbds (not even > when Sun's xkbtable.map is copied over), and hence is useless (only putting > xmodpap into ~/.xinitrc would help you, but not when dtlogin is used to start > Xorg up). Etc. etc. etc. > Please do *not* install those pkg's on any non-experimental box, only onto a > test machine. > Don't (yet) expect being conveniently driven through all the steps of getting > Xorg to work on your box (i.e. to replace Xsun completely, as I do on all of > my boxes). > > This is only a proof-of-concept release to actually give you something to > test. > And to give you a first impression of how nice many things will be on SPARC, > as soon the those last nasty (pretty much high-level) bugs are removed soon. > > Here are the pkgs (built on snv_70_SPARC, but also work down on snv_60 as I > just verified, this time no major libs upgrade is preventing this, so you can > take SXCR_snv_60 or higher) : > > http://www.martux.org/xorg/fox-gate/20070930sun/bin_pkgadd/Xorg7.2.0_opensol_sparc_snv70_binary_pkgadd__fox-gate_20070930.tar.gz > http://www.martux.org/xorg/fox-gate/20070930sun/bin_pkgadd/MD5SUMS.txt > > ( You may also find the *old* README and INSTALL files useful: > http://www.martux.org/xorg/OLD/20070503__7.2.0/bin_pkgadd_packages/INSTALL > http://www.martux.org/xorg/OLD/20070503__7.2.0/bin_pkgadd_packages/READ_ME.ascii > ) > > Src release currently not posssible, as too many temporary files are still > polluting my tree here at home. Before I clean that up (in order to create a > global diff, upgarde to today's FOX and push things against mercurial), I > first remove as many of the pkgdefs related bugs, as possible. > > > Regards, > %martin > > p.s Thanks too those who have encouraged me to continue :-) > I didn't yet come to reading my mountain of emails for almost a week, I > always simply use jive to post, what I'm doing right now, too. > pp.s.: All errors, bugs, orthographic or grammar mistakes are > copyright-protected. > Good night ... > > > This message posted from opensolaris.org > _______________________________________________ > xwin-discuss mailing list > xwin-discuss at opensolaris.org >
