summaryrefslogtreecommitdiff
path: root/src
AgeCommit message (Collapse)Author
2007-03-24[mach64] Use Mach64Chipsets[] instead of ATIChipNames[].George Sapountzis
atimisc is PCI-only now, we can get the chip name with xf86TokenToString().
2007-03-24[mach64] Set pATI->ChipRevision correctly, instead of overriding.George Sapountzis
There is no need to override pATI->ChipRevision for GX/CX, as it is only reported with a printf.
2007-03-23Drop probing by driver name from "Device" section.George Sapountzis
atimisc is PCI-only now, so we only need to scan the PCI space.
2007-03-23Fold FillIn() back to Probe().George Sapountzis
2007-03-23Drop now unneeded _X_EXPORT's.George Sapountzis
2007-03-23Move {atimach64,r128,radeon}_probe.c from ati to subdrivers.George Sapountzis
Subdrivers are now loaded from the wrapper Probe function rather than at screen creation time. The wrapper Identify callback only prints chip families now, chip lists are printed when a subdriver is loaded. This also avoids duplication of subdriver Identify callbacks. Unknown radeons should still get a list of known radeon and then probe fails... Probe for atimisc last to avoid needless loading in most cases (r128, radeon).
2007-03-20Fix advertised minimum minor version of the DRI module.Michel Dänzer
We don't automatically require bumped minor versions.
2007-03-20radeon: Only sync to hardware when really necessary with EXA.Michel Dänzer
In particular, don't sync again after accelerated DownloadFromScreen, which syncs implicitly. This avoids calling into the kernel when it's not necessary, which can be relevant in some situations.
2007-03-20radeon: Unify code to release the CP.Michel Dänzer
2007-03-20radeon: Minor BlockHandler cleanups.Michel Dänzer
Don't flush indirect buffer in BlockHandler; it's done in LeaveServer. Also set the EXA engine mode to unknown only at the end.
2007-03-20radeon: Disable CP line acceleration on RV280s.Michel Dänzer
There have been several reports of stability issues with things like the xscreensaver hack hypercube.
2007-03-20radeon: Make sure 3D clients will re-upload textures to video RAM after LeaveVT.Michel Dänzer
Walk the SAREA texList and bump the age of every active object, so their owners will consider them kicked out when they grab the HW lock next time.
2007-03-15atimisc: Fix strict aliasing violation flagged by -Wall.Michel Dänzer
2007-03-15radeon: Fix build warnings.Michel Dänzer
2007-03-08Move atioption.c from ati to atimisc.George Sapountzis
2007-03-08Move atiprobe.c from ati to atimisc.George Sapountzis
2007-03-04[mach64] Bug 5586: overlay scaler limited to 720 pixels on ATI Rage Pro.George Sapountzis
2007-03-04[mach64] Xv: use single surface and encoding structs.George Sapountzis
2007-03-04radeon: add option to set pci aperture size from config fileDave Airlie
This requires a drm > 1.26 to work
2007-02-25[mach64] Consolidate adjustments of mode timings, part 3.George Sapountzis
Cosmetic.
2007-02-25[mach64] Consolidate adjustments of mode timings, part 2.George Sapountzis
Factor out to seperate function.
2007-02-25[mach64] Consolidate adjustments of mode timings, part 1.George Sapountzis
Move LCD block from atimode.c to atimach64.c .
2007-02-21add missing part of last commit...Roland Scheidegger
2007-02-21fix alignment issues with planar yuv and a bug with packed uyvyRoland Scheidegger
respect that all source planar yuv planes are already dword aligned. Some attempts to fix up odd widths and odd heights (which are a bit strange for 4:2:0 formats). They still don't quite work 100% correctly (at the borders) but neither do they with packed yuv formats. While here, fix totally broken packed UYVY format by inserting missing break... Both bugs reported by Felipe Contreras.
2007-02-20radeon: move some i2c registers out into common register fileDave Airlie
These regs are more useful than just multimedia
2007-02-16Add "DRI" optionAlex Deucher
This option allows you to disable the DRI per card. It also removes the "RN50Force3D" option as it is now covered by this option. RN50 users should set this to TRUE if they want to force the DRI on.
2007-02-13[mach64] PreInit: maxPitch, minor cosmetic.George Sapountzis
2007-02-13[mach64] PreInit: pitchInc is local.George Sapountzis
2007-02-13[mach64] PreInit: Use goto bail.George Sapountzis
Match ATILock - ATIUnlock.
2007-02-12[mach64] Enable RENDER acceleration on the Pro variants.George Sapountzis
2007-02-05Consolidate atixv.c w/ atimach64xv.c, part 2.George Sapountzis
Move {Initialize,Close}XVideo from atixv.c to atimach64xv.c .
2007-02-05Consolidate atixv.c w/ atimach64xv.c, part 1.George Sapountzis
Merge ATIXVInitializeAdaptor() in ATIMach64XVInitialiseAdaptor().
2007-02-05Consolidate aticursor.c w/ atimach64cursor.cGeorge Sapountzis
2007-02-05Consolidate atiaccel.c w/ atimach64accel.c, part 2.George Sapountzis
Merge ATIInitializeAcceleration() in ATIMach64AccelInit().
2007-02-05Consolidate atiaccel.c w/ atimach64accel.c, part 1.George Sapountzis
Move ATIResizeOffscreenLinear() from atiaccel.c to atimach64xv.c and make it static. It is only used by XV memory management functions.
2007-02-05Drop non-PCI support from atividmem.cGeorge Sapountzis
- we require a PCI device (pVideo != NULL) - map the VGA framebuffer with xf86MapDomainMemory()
2007-02-05Minor refactoring of pATI->Block0Base computation.George Sapountzis
2007-02-05Cleanup checks for pATI->PCIInfo.George Sapountzis
Fixes a bug where shadowfb would not be enabled by default in the non-accel case because pATI->PCIInfo is now set after ATIProcessOptions() is called.
2007-02-05Merge branch 'mach64-pci-1'George Sapountzis
2007-01-30add option to force the DRI on for RN50 chipsAlex Deucher
Allow user to force the DRI on for RN50 chips. 3D is not guaranteed to work on these chips, however in some cases it does. fixes bug 9802.
2007-01-28Drop pATI->Chipset, no longer used.mach64-pci-1George Sapountzis
2007-01-28Add option "probe_sparse" to force probing sparse I/O bases.George Sapountzis
2007-01-28Sane probe, based on work by Luc Verhaegen <libv@skynet.be>.George Sapountzis
2007-01-28Drop checking of sparse I/O bases conflictsGeorge Sapountzis
This will be handled with an option somehow.
2007-01-28Minor refactoring of module (un)loading for TV_OUT.George Sapountzis
2007-01-28Drop support for non-programmable clocks.George Sapountzis
From README.ati: Clocks for supported programmable clock generators: The driver currently supports all programmable clock generators known to exist on Mach64 adapters. Clocks for unsupported programmable clock generators: This case is unlikely to occur, but is documented for the sake of completeness. Thus: - check for (pATI->ProgrammableClock > ATI_CLOCK_FIXED) && (pATI->ProgrammableClock < ATI_CLOCK_MAX) - drop "probe_clocks" option - pATIHW->ClockUnmap is no longer used - pATIHW->ClockMap is only used with NewHW.crtc which is always ATI_CRTC_MACH64 and has the identity map, so drop it - (pATI->ProgrammableClock != ATI_CLOCK_INTERNAL) => (pATI->depth <= 8)
2007-01-28Banked memory is no longer needed.George Sapountzis
- (pATI->BankInfo.BankSize = 0) in all cases, cull pATI->BankInfo - only keep the minimal pATIHW.SetBank interface for save/restore - clean ATISwap() a little, (NewHW.crtc != ATI_CRTC_VGA) - (UseSmallApertures == TRUE) <=> pATI->VGAAdapter
2007-01-28Always use the linear aperture.George Sapountzis
- drop (pATI->OptionLinear == FALSE) - AcceleratorVideoRAM is always set, i.e. VGAVideoRAM is not used - pATI->LinearBase is always set - xf86LinearVidMem() is now checked in atipreinit() for both CPIO and MMIO
2007-01-28We always use the accelerator CRTC within the xserver.George Sapountzis
- cull (pATI->NewHW.crtc != ATI_CRTC_MACH64).
2007-01-28Clean pATI->Adapter and pATI->VGAAdapter.George Sapountzis
- cull (pATI->Adapter != ATI_ADAPTER_MACH64) - treat pATI->VGAAdapter as Bool