profile/ivi/libdrm.git
19 years agoRemove the AGP requirement from Makefile and Kconfig for MGA. Remove the
Ian Romanick [Wed, 29 Jun 2005 23:20:30 +0000 (23:20 +0000)]
Remove the AGP requirement from Makefile and Kconfig for MGA. Remove the
    AGP requirement from Kconfig for SIS. There never was a requirement in
    Makefile, and Eric Anholt confirms that the Makefile was correct.

19 years agoadd remaining calls
Alan Hourihane [Wed, 29 Jun 2005 13:19:47 +0000 (13:19 +0000)]
add remaining calls

19 years agoadd i915_ioc32.c
Alan Hourihane [Wed, 29 Jun 2005 13:13:22 +0000 (13:13 +0000)]
add i915_ioc32.c

19 years agosilence warning
Alan Hourihane [Wed, 29 Jun 2005 13:00:29 +0000 (13:00 +0000)]
silence warning

19 years agofix some warnings from cross compiler
Dave Airlie [Wed, 29 Jun 2005 12:02:18 +0000 (12:02 +0000)]
fix some warnings from cross compiler

19 years agomake r128/mga compile properly on sparc cross-compiler
Dave Airlie [Wed, 29 Jun 2005 11:56:42 +0000 (11:56 +0000)]
make r128/mga compile properly on sparc cross-compiler

19 years agoMove to linux specific directory
Alan Hourihane [Wed, 29 Jun 2005 11:36:37 +0000 (11:36 +0000)]
Move to linux specific directory

19 years agoadd mga and r128 32/64 bits
Dave Airlie [Wed, 29 Jun 2005 11:22:39 +0000 (11:22 +0000)]
add mga and r128 32/64 bits
This is Egberts code, ported to Pauls framework by me..

19 years agothese don't need reclaim buffers their release functions handle it
Dave Airlie [Wed, 29 Jun 2005 03:06:47 +0000 (03:06 +0000)]
these don't need reclaim buffers their release functions handle it

19 years agoBug in conversion from old DRM to core DRM....
Dave Airlie [Wed, 29 Jun 2005 02:57:18 +0000 (02:57 +0000)]
Bug in conversion from old DRM to core DRM....

19 years agoGet the BSD DRM compiling again after MGA and mapping changes. Note that
Eric Anholt [Wed, 29 Jun 2005 02:54:19 +0000 (02:54 +0000)]
Get the BSD DRM compiling again after MGA and mapping changes. Note that
    i915 has been disabled for the moment, rather than working around
    linux-specific code in the shared dir.

19 years ago- Remove drm_initmap and replace its usage with drm_addmap. This reduces
Eric Anholt [Tue, 28 Jun 2005 20:58:34 +0000 (20:58 +0000)]
- Remove drm_initmap and replace its usage with drm_addmap. This reduces
    code duplication, and it also hands you the map pointer so you don't
    need to re-find it.
- Remove the permanent maps flag. Instead, for register and framebuffer
    maps, we always check whether there's already a map of that type and
    offset around. Move the Radeon map initialization into presetup (first
    open) so it happens again after every takedown.
- Remove the split cleanup of maps between driver takedown (last close) and
    cleanup (module unload). Instead, always tear down maps on takedown,
    and drivers can recreate them on first open.
- Make MGA always use addmap, instead of allocating consistent memory in
    the PCI case and then faking up a map for it, which accomplished nearly
    the same thing, in a different order. Note that the maps are exposed to
    the user again: we may want to expose a flag to avoid this, but it's
    not a security concern, and saves us a lot of code.
- Remove rmmaps in the MGA driver. Since the function is only called during
    takedown anyway, we can let them die a natural death.
- Make removal of maps happen in one function, which is called by both
    drm_takedown and drm_rmmap_ioctl.
Reviewed by: idr (previous revision) Tested on: mga (old/new/pci dma),
    radeon, savage

19 years agoadd compat code from Paul Mackerras
Dave Airlie [Tue, 28 Jun 2005 13:02:20 +0000 (13:02 +0000)]
add compat code from Paul Mackerras

19 years agoAdd compat to Makefile
Dave Airlie [Tue, 28 Jun 2005 12:52:41 +0000 (12:52 +0000)]
Add compat to Makefile

19 years agoAdd drm and radeon 32/64-bit compat code from Paul Mackerras
Dave Airlie [Tue, 28 Jun 2005 12:50:15 +0000 (12:50 +0000)]
Add drm and radeon 32/64-bit compat code from Paul Mackerras

19 years agoGiven that BenH says using the sysdev approach for DRM is bogus, I'll yank
Alan Hourihane [Tue, 28 Jun 2005 08:03:33 +0000 (08:03 +0000)]
Given that BenH says using the sysdev approach for DRM is bogus, I'll yank
    the code for it, rather than introducing something that isn't going to
    work 100% of the time.

19 years agoComment out the .resume function as without programming their dependent
Alan Hourihane [Tue, 28 Jun 2005 07:34:49 +0000 (07:34 +0000)]
Comment out the .resume function as without programming their dependent
    registers things tend to lock up in certain situations.
The BIOS repost will fix things up.

19 years agoReverse the pm_message_t patch for now, it appears that the 2.6.12 release
Alan Hourihane [Mon, 27 Jun 2005 15:17:12 +0000 (15:17 +0000)]
Reverse the pm_message_t patch for now, it appears that the 2.6.12 release
    didn't have it.

19 years agoCheck for 2.6.12 suspend/resume changes (pm_message_t)
Alan Hourihane [Mon, 27 Jun 2005 12:39:02 +0000 (12:39 +0000)]
Check for 2.6.12 suspend/resume changes (pm_message_t)

19 years agoCheck dev_priv
Alan Hourihane [Mon, 27 Jun 2005 11:41:40 +0000 (11:41 +0000)]
Check dev_priv

19 years agoSave state of registers for suspend/resume.
Alan Hourihane [Mon, 27 Jun 2005 11:39:44 +0000 (11:39 +0000)]
Save state of registers for suspend/resume.

19 years agoremoved dev->sysdev_registered
Jon Smirl [Sun, 26 Jun 2005 13:31:15 +0000 (13:31 +0000)]
removed dev->sysdev_registered

19 years agoRef count the sysdev class to support multiple DRM cards
Jon Smirl [Fri, 24 Jun 2005 22:48:16 +0000 (22:48 +0000)]
Ref count the sysdev class to support multiple DRM cards

19 years agoMake sysdev class only register when fbdev detected
Jon Smirl [Fri, 24 Jun 2005 21:50:40 +0000 (21:50 +0000)]
Make sysdev class only register when fbdev detected

19 years agoMore err path clean up for drm_pm Add mandatory sysdev shutdown function
Jon Smirl [Fri, 24 Jun 2005 19:31:06 +0000 (19:31 +0000)]
More err path clean up for drm_pm Add mandatory sysdev shutdown function

19 years agoFix the sysdev approach for power management.
Alan Hourihane [Fri, 24 Jun 2005 09:28:50 +0000 (09:28 +0000)]
Fix the sysdev approach for power management.
We need to use the container_of() call to access our device private.

19 years agoFix suspend problem when i915 private hasn't been initialized yet.
Alan Hourihane [Fri, 24 Jun 2005 08:56:12 +0000 (08:56 +0000)]
Fix suspend problem when i915 private hasn't been initialized yet.

19 years agoFix drm_memory_debug.c to compile, doesn't seem to be working Clean up
Jon Smirl [Thu, 23 Jun 2005 05:29:16 +0000 (05:29 +0000)]
Fix drm_memory_debug.c to compile, doesn't seem to be working Clean up
    error return path in drm_stub.c

19 years agoGet the power management hooks into the right place so that everything gets
Jon Smirl [Wed, 22 Jun 2005 05:49:56 +0000 (05:49 +0000)]
Get the power management hooks into the right place so that everything gets
    freed correctly.

19 years agoChange initialization of savage register access to _DRM_READ_ONLY. Flags of
Jon Smirl [Tue, 21 Jun 2005 02:20:02 +0000 (02:20 +0000)]
Change initialization of savage register access to _DRM_READ_ONLY. Flags of
    zero does not mean no privs, instead it grants write access
    _DRM_READ_ONLY only applies to non-root users. Problem is only in CVS,
    initmaps are not in the kernel yet.

19 years agoChange initialization of radeon register access to _DRM_READ_ONLY. Flags of
Jon Smirl [Mon, 20 Jun 2005 15:40:48 +0000 (15:40 +0000)]
Change initialization of radeon register access to _DRM_READ_ONLY. Flags of
    zero does not mean no privs, instead it grants write access
    _DRM_READ_ONLY only applies to non-root users. Problem is only in CVS,
    initmaps are not in the kernel yet.

19 years agoRemove I2C support from radeon driver. Same support is available from
Jon Smirl [Sun, 19 Jun 2005 04:15:58 +0000 (04:15 +0000)]
Remove I2C support from radeon driver. Same support is available from
    radeonfb.

19 years agofix up drm_alloc_agp to take a dev arg and not pass crappy agpgart around
Dave Airlie [Fri, 17 Jun 2005 09:09:17 +0000 (09:09 +0000)]
fix up drm_alloc_agp to take a dev arg and not pass crappy agpgart around

19 years agoFix 810/830 build
Jon Smirl [Fri, 17 Jun 2005 04:47:30 +0000 (04:47 +0000)]
Fix 810/830 build

19 years agoForce AGP always for Intel chipsets.
Alan Hourihane [Thu, 16 Jun 2005 19:58:00 +0000 (19:58 +0000)]
Force AGP always for Intel chipsets.
Fixes bug #3552

19 years agoRe-sync (and correct!) shared/mga_drm.h with shared-core/mga_drm.h.
Ian Romanick [Wed, 15 Jun 2005 17:47:33 +0000 (17:47 +0000)]
Re-sync (and correct!) shared/mga_drm.h with shared-core/mga_drm.h.

19 years agoAdds support for PCI cards to MGA DRM
Ian Romanick [Tue, 14 Jun 2005 22:34:11 +0000 (22:34 +0000)]
Adds support for PCI cards to MGA DRM
This patch adds serveral new ioctls and a new query to get_param query to
    support PCI MGA cards.
Two ioctls were added to implement interrupt based waiting. With this
    change, the client-side driver no longer needs to map the primary DMA
    region or the MMIO region. Previously, end-of-frame waiting was done by
    busy waiting in the client-side driver until one of the MMIO registers
    (the current DMA pointer) matched a pointer to the end of primary DMA
    space. By using interrupts, the busy waiting and the extra mappings are
    removed.
A third ioctl was added to bootstrap DMA. This ioctl, which is used by the
    X-server, moves a *LOT* of code from the X-server into the kernel. This
    allows the kernel to do whatever needs to be done to setup DMA buffers.
    The entire process and the locations of the buffers are hidden from
    user-mode.
Additionally, a get_param query was added to differentiate between G4x0
    cards and G550 cards. A gap was left in the numbering sequence so that,
    if needed, G450 cards could be distinguished from G400 cards. According
    to Ville Syrjälä, the G4x0 cards and the G550 cards handle
    anisotropic filtering differently. This seems the most compatible way
    to let the client-side driver know which card it's own. Doing this very
    small change now eliminates the need to bump the DRM minor version
    twice.
http://marc.theaimsgroup.com/?l=dri-devel&m=106625815319773&w=2
A number of ioctl handlers in linux-core were also modified so that they
    could be called in-kernel. In these cases, the in-kernel callable
    version kept the existing name (e.g., drm_agp_acquire) and the ioctl
    handler added _ioctl to the name (e.g., drm_agp_acquire_ioctl).
This patch also replaces the drm_agp_do_release function with
    drm_agp_release. drm_agp_release (drm_core_agp_release in the previous
    patch) is very similar to drm_agp_do_release, and I saw no reason to
    have both.
This commit *breaks the build* on BSD. Eric said that he would make the
    required updates to the BSD side soon.
Xorg bug: 3259 Reviewed by: Eric Anholt

19 years agofix up drm.h for C++ include as this can be included in user programs
Dave Airlie [Sat, 11 Jun 2005 10:08:39 +0000 (10:08 +0000)]
fix up drm.h for C++ include as this can be included in user programs

19 years agoadd dragonfly #define from Xorg CVS
Dave Airlie [Fri, 10 Jun 2005 01:21:23 +0000 (01:21 +0000)]
add dragonfly #define from Xorg CVS

19 years agoSynchronize with the shared-core version.
Ian Romanick [Thu, 9 Jun 2005 21:24:34 +0000 (21:24 +0000)]
Synchronize with the shared-core version.

19 years agoCompletely re-initialize DMA settings
Ian Romanick [Thu, 9 Jun 2005 21:18:56 +0000 (21:18 +0000)]
Completely re-initialize DMA settings
There were two problems. First, the 'warp' and 'primary' pointers weren't
    cleared, so mga_do_cleanup_dma, which gets called multiple times, would
    try to ioremapfree them multiple times. This resulted in the new error
    messages to syslog. The second problem was the, since the dev_private
    structure isn't reallocated and cleaned out in mga_do_init_dma, when
    the server is reloaded idle-waits would wait for impossible values.
I have given this patch some more riggorous testing. This includes:
- Load module, start server, run GL app, stop server, unload module.
- Load module, start server, run GL app, stop server, unload module, reload
    module, restart server, run GL app.
- Load module, start server, run GL app, stop server, restart server, run
    GL app, stop server, unload module.
In all three cases, everything worked as expected. Please let me know if
    there are any further regressions with this patch.
Xorg bug: 3408 Reported by: Chris Rankin

19 years agodrm_mem_init should be done at core load, not driver init
Jon Smirl [Thu, 9 Jun 2005 13:22:12 +0000 (13:22 +0000)]
drm_mem_init should be done at core load, not driver init

19 years agofix some issues with radeon interrupt handling
Dave Airlie [Mon, 6 Jun 2005 11:35:43 +0000 (11:35 +0000)]
fix some issues with radeon interrupt handling
From: Dave Airlie + Benjamin Herrenschmidt

19 years agoFix copyrights
Alan Hourihane [Mon, 6 Jun 2005 09:18:44 +0000 (09:18 +0000)]
Fix copyrights

19 years agoAdd a few more bits of Tonnerre's NetBSD port (Still need to deal with the
Eric Anholt [Mon, 6 Jun 2005 06:45:41 +0000 (06:45 +0000)]
Add a few more bits of Tonnerre's NetBSD port (Still need to deal with the
    device attachment).

19 years agomisc cleanup patch from Adrian Bunk
Dave Airlie [Sat, 4 Jun 2005 06:18:11 +0000 (06:18 +0000)]
misc cleanup patch from Adrian Bunk

19 years agoRemove warnings about code mixed with declerations..
Dave Airlie [Sat, 4 Jun 2005 06:14:58 +0000 (06:14 +0000)]
Remove warnings about code mixed with declerations..

19 years agoMinor clean-ups. The drm_mga_fullscreen_t stuff has been dead for a looong
Ian Romanick [Fri, 3 Jun 2005 22:53:32 +0000 (22:53 +0000)]
Minor clean-ups. The drm_mga_fullscreen_t stuff has been dead for a looong
    time. Remove it and an out-dated comment.

19 years agoMove the deallocation of dev_private. Since dev_private is allocated when
Ian Romanick [Fri, 3 Jun 2005 22:45:21 +0000 (22:45 +0000)]
Move the deallocation of dev_private. Since dev_private is allocated when
    the driver is loaded and is always expected to be around, it should
    only be deallocated when the driver is unloaded.
Xorg bug: #3408 Reported by: Chris Rankin

19 years agoBugzilla #3217: Create a new __drm_pci_free which is used internally in
Eric Anholt [Sat, 28 May 2005 20:36:22 +0000 (20:36 +0000)]
Bugzilla #3217: Create a new __drm_pci_free which is used internally in
    linux-core to free pci memory without freeing the structure. Linux-core
    internals often create pci dma handle structures on the stack due to
    the lack of a drm_local_map_t to store them in properly. Fix the
    original drm_pci_free to actually free the dma handle structure instead
    of leaking it.
Signed-off-by: Sergey Vlasov <vsu@altlinux.ru>
19 years agoBugzilla #3217: The size field in the new dma handle structure was
Eric Anholt [Sat, 28 May 2005 20:25:04 +0000 (20:25 +0000)]
Bugzilla #3217: The size field in the new dma handle structure was
    uninitialized, and its use in drm_pci_free later resulted in panics.
Signed-off-by: Sergey Vlasov <vsu@altlinux.ru>
19 years agoActually check for drm_fb_loaded before trying to initialize the sysdev
Alan Hourihane [Sat, 28 May 2005 00:08:53 +0000 (00:08 +0000)]
Actually check for drm_fb_loaded before trying to initialize the sysdev
    approach.

19 years agoRe-implement the power management.
Alan Hourihane [Sat, 28 May 2005 00:00:08 +0000 (00:00 +0000)]
Re-implement the power management.
There's two choices when fb is or isn't loaded as we treat ourselves as a
    PCI driver in the latter case.
If we are a PCI driver, then register the suspend/resume functions
    directly. If not, then we register as a sysdev and pick up the
    suspend/resume actions and pump them down into a generic *power
    function.
It'll be nice when this little mess is sorted out with regard to being a
    real PCI driver ;-/

19 years agoModify drm_driver::device_is_agp to return a tri-state value to indicate
Ian Romanick [Fri, 27 May 2005 23:42:11 +0000 (23:42 +0000)]
Modify drm_driver::device_is_agp to return a tri-state value to indicate
    that a device absolutely is, absolutely is not, or may or may not be
    AGP. Modify the i915 DRM to use this to force all i9x5 devices to be
    "AGP" (even the PCI-e devices).
Reported by: Lukas Hejtmanek

19 years agoadd radeon registers from VHA code these are the "unknown" registers
Dave Airlie [Fri, 27 May 2005 07:23:44 +0000 (07:23 +0000)]
add radeon registers from VHA code these are the "unknown" registers

19 years agoVIA:
Thomas Hellstrom [Mon, 23 May 2005 20:56:54 +0000 (20:56 +0000)]
VIA:
64-bit fixes. Bumped driver date and patchlevel.

19 years agoRefactor the routines for "calculating" the size of the WARP microcode. The
Ian Romanick [Sun, 22 May 2005 04:36:33 +0000 (04:36 +0000)]
Refactor the routines for "calculating" the size of the WARP microcode. The
    two routines (one for G200 and one for G400) were replaced with static
    const variables and a single function that returns the correct size.
    The code to generate an error if the allocated WARP region is too small
    was refactored from mga_warp_install_{g200,g400}_microcode to
    mga_wrap_install_microcode.
mga_warp_microcode_size is global to the MGA DRM because it will soon be
    used by code in another file.

19 years agoForgot to bump the patchlevel and driver date on last commit.
Ian Romanick [Sat, 21 May 2005 02:31:08 +0000 (02:31 +0000)]
Forgot to bump the patchlevel and driver date on last commit.

19 years agoChange the MGA initialization and cleanup a bit. The dev_private structure
Ian Romanick [Sat, 21 May 2005 02:27:51 +0000 (02:27 +0000)]
Change the MGA initialization and cleanup a bit. The dev_private structure
    is now allocated (and partially filled in) by the new
    mga_driver_preinit function.
This allows the driver to detect the type of card (i.e., G200 class vs.
    G400 class) on its own. The chipset value passed to mga_dma_init is now
    ignored. This same technique is used by the radeon DRM.
As a result of this, mga_driver_pretakedown was converted to
    mga_driver_postcleanup. This routine gets called in some other places
    than might be expected, and it sets the dev_private pointer to NULL.
    That little gem took over an hour to track down. :(

19 years agoRefactor common, boilerplate ioctl code from drm_addbufs_* functions into
Ian Romanick [Fri, 20 May 2005 00:17:40 +0000 (00:17 +0000)]
Refactor common, boilerplate ioctl code from drm_addbufs_* functions into
    drm_addbufs. This makes the code more like the BSD code, and makes the
    drm_addbufs_* functions callable in-kernel.
Reviewed by: Dave Airlie

19 years agoAdd i945G pci ids to drm
Dave Airlie [Wed, 18 May 2005 06:35:16 +0000 (06:35 +0000)]
Add i945G pci ids to drm
From: Christopher Allen Wing <wingc@engin.umich.edu> Signed-off-by: Dave
    Airlie <airlied@linux.ie>

19 years agoSavage doesn't require AGP any more. Enable build even without CONFIG_AGP.
Felix Kuehling [Tue, 17 May 2005 02:08:02 +0000 (02:08 +0000)]
Savage doesn't require AGP any more. Enable build even without CONFIG_AGP.

19 years agoAdded device_is_agp callback to drm_driver. This function is called by the
Ian Romanick [Mon, 16 May 2005 17:37:10 +0000 (17:37 +0000)]
Added device_is_agp callback to drm_driver. This function is called by the
    platform-specific drm_device_is_agp function. Added implementation of
    this function the the Linux-specific portion of the MGA driver to
    detect PCI G450 cards. Added code to the Linux-specific portion of the
    generic DRM layer to not initialize AGP infrastructure if the card is
    not AGP (this matches what already existed in BSD).
Bumped the driver date and the driver patch-level for MGA.
This mostly fixes bugzilla #3248. The BSD side still needs an
    implementation of mga_driver_device_is_agp.

19 years agowrap config.h include with defined for KERNEL
Dave Airlie [Sun, 15 May 2005 10:19:21 +0000 (10:19 +0000)]
wrap config.h include with defined for KERNEL

19 years agoVIA: Fix for oops when AGP ring-buffer initialization is called and there
Thomas Hellstrom [Sun, 8 May 2005 20:33:04 +0000 (20:33 +0000)]
VIA: Fix for oops when AGP ring-buffer initialization is called and there
    is no AGP memory acquired.

19 years agoConvert NOMAN to the new preferred spelling NO_MAN to quiet warnings.
Eric Anholt [Tue, 26 Apr 2005 15:27:58 +0000 (15:27 +0000)]
Convert NOMAN to the new preferred spelling NO_MAN to quiet warnings.

19 years agoClean up some const qualifier cast warnings.
Eric Anholt [Tue, 26 Apr 2005 06:03:39 +0000 (06:03 +0000)]
Clean up some const qualifier cast warnings.

19 years agoConvert BSD code to mostly use bus_dma, the dma abstraction for dealing
Eric Anholt [Tue, 26 Apr 2005 05:19:11 +0000 (05:19 +0000)]
Convert BSD code to mostly use bus_dma, the dma abstraction for dealing
    with IOMMUs and such. There is one usage of the forbidden vtophys()
    left in drm_scatter.c which will be fixed up soon. This required a KPI
    change for drm_pci_alloc/free() to return/use a drm_dma_handle_t that
    keeps track of os-specific bits, rather than just passing around the
    vaddr/busaddr/size.
Submitted by: Tonnerre Lombard (partially) Tested on: FreeBSD: Rage128
    AGP/PCI Linux: Savage4 AGP/PCI

19 years agoUse msleep instead of tsleep to drop the DRM device lock and avoid a
Eric Anholt [Tue, 26 Apr 2005 05:11:20 +0000 (05:11 +0000)]
Use msleep instead of tsleep to drop the DRM device lock and avoid a
    sleep-with- mutex held. This probably ought to be an os-independent
    sleep function ala DRM_USLEEP.

19 years agoFix a panic on startup with non-initmapping drivers by assigning the
Eric Anholt [Sun, 24 Apr 2005 19:09:09 +0000 (19:09 +0000)]
Fix a panic on startup with non-initmapping drivers by assigning the
    drm_ioremap return value to the map handle again.
Submitted by: Tor Egge, tegge at freebsd dot org

19 years agoA fix for a locking bug which is triggered when a client tries to lock with
Thomas Hellstrom [Wed, 20 Apr 2005 18:50:49 +0000 (18:50 +0000)]
A fix for a locking bug which is triggered when a client tries to lock with
    flag DMA_QUIESCENT (typically the X server), but gets interrupted by a
    signal. The locking IOCTL should then return an error, but if
    DMA_QUIESCENT succeeds it returns 0, and the client falsely thinks it
    has the lock. In addition The client waits for DMA_QUISCENT and
    possibly DMA_READY without having the lock.

19 years agoVIA: Interrupt handler bugfixes. Bumped patchlevel to 2.6.2.
Thomas Hellstrom [Wed, 20 Apr 2005 10:16:52 +0000 (10:16 +0000)]
VIA: Interrupt handler bugfixes. Bumped patchlevel to 2.6.2.

19 years agoRevert last commit, it affect via things
Dave Airlie [Tue, 19 Apr 2005 00:31:16 +0000 (00:31 +0000)]
Revert last commit, it affect via things

19 years agoVIA:
Thomas Hellstrom [Mon, 18 Apr 2005 08:26:00 +0000 (08:26 +0000)]
VIA:
1) Security fix: Stopped possible MMIO access to PCI DMA area for the
    unichrome Pro.
2) Fixed an odd cpu usage problem by padding small AGP DMA submissions.
    Bumped patchlevel.

19 years agoUse /*- to begin license blocks in BSD code to reduce diffs against FreeBSD
Eric Anholt [Sat, 16 Apr 2005 03:02:52 +0000 (03:02 +0000)]
Use /*- to begin license blocks in BSD code to reduce diffs against FreeBSD
    CVS.

19 years agoFix build on FreeBSD-current, thanks to jhb@.
Eric Anholt [Wed, 13 Apr 2005 04:20:08 +0000 (04:20 +0000)]
Fix build on FreeBSD-current, thanks to jhb@.

19 years agomirror changes made in main tree.. just happened to be doing this myself
Dave Airlie [Mon, 4 Apr 2005 04:08:29 +0000 (04:08 +0000)]
mirror changes made in main tree.. just happened to be doing this myself

19 years agoMissing file via_video.c commited.
Thomas Hellstrom [Tue, 29 Mar 2005 09:48:12 +0000 (09:48 +0000)]
Missing file via_video.c commited.

19 years agoAdd skeletal imagine driver (but don't build it yet).
Adam Jackson [Tue, 29 Mar 2005 01:47:38 +0000 (01:47 +0000)]
Add skeletal imagine driver (but don't build it yet).

19 years agoVia updates:
Thomas Hellstrom [Mon, 28 Mar 2005 21:21:42 +0000 (21:21 +0000)]
Via updates:
New PCI command parser. Moved from via_dma.c to via_verifier.c so functions
    with similar functionality are close to eachother.
Moved video related functions to via_video.c, which might be extended in
    the future, as new video functionality is added.
New device-specific generic IRQ IOCTL, similar to the general VBLANK IOCTL,
    but with support for multiple device IRQ sources and functionality.
Support for Unichrome Pro PM800/CN400 video DMA commands in verifier and
    PCI parser.
Support for Unichrome Pro PM800/CN400 HQV IRQs in the new generic IRQ
    IOCTL.
Bumped minor. New version 2.6.0.

19 years agoDescription: Rather than use custom code in DRM_WAIT_ON() to do exactly
Dave Airlie [Sun, 27 Mar 2005 07:05:28 +0000 (07:05 +0000)]
Description: Rather than use custom code in DRM_WAIT_ON() to do exactly
    what wait_event_interruptible_timeout() does, use the function and just
    change the return values appropriately.
Signed-off-by: Nishanth Aravamudan <nacc@us.ibm.com>
19 years agoSome ia64 platforms may not support write combining on all type of memory,
Dave Airlie [Sun, 27 Mar 2005 06:58:12 +0000 (06:58 +0000)]
Some ia64 platforms may not support write combining on all type of memory,
    so we need to consult the EFI memory map before we try to set the write
    combine attribute of a page. This patch will try to map a page write
    combined if it's not an AGP page and the EFI memory map says it's ok,
    otherwise it falls back to a regular, uncached mapping. Can someone
    please apply this to the drm tree?
From: Jesse Barnes

19 years agoDon't know who set the i915GM PCI ID incorrectly but it should be 0x2592
Alan Hourihane [Fri, 25 Mar 2005 13:44:47 +0000 (13:44 +0000)]
Don't know who set the i915GM PCI ID incorrectly but it should be 0x2592

19 years agoFrom Xorg CVS realign the i915_drv.h
Dave Airlie [Fri, 25 Mar 2005 13:16:38 +0000 (13:16 +0000)]
From Xorg CVS realign the i915_drv.h
From: Tungsten Graphics Signed-off-by: Dave Airlie <airlied@linux.ie>

19 years agouse linux-core for 2.6 not this stuff..
Dave Airlie [Fri, 25 Mar 2005 13:06:53 +0000 (13:06 +0000)]
use linux-core for 2.6 not this stuff..

19 years agofix up AGP multi-head support for kernel 2.6.12
Dave Airlie [Fri, 25 Mar 2005 09:48:34 +0000 (09:48 +0000)]
fix up AGP multi-head support for kernel 2.6.12

19 years agoverify_area is deprecated, replaced by access_ok. Seems I missed this one
Dave Airlie [Fri, 25 Mar 2005 09:47:36 +0000 (09:47 +0000)]
verify_area is deprecated, replaced by access_ok. Seems I missed this one
    when I did the big overall conversion.
Signed-off-by: Jesper Juhl <juhl-lkml@dif.dk>
19 years agofix bug with XFree86 4.3 on core drm
Dave Airlie [Fri, 25 Mar 2005 09:05:10 +0000 (09:05 +0000)]
fix bug with XFree86 4.3 on core drm
Signed-off-by: Dave Airlie <airlied@linux.ie>
19 years ago3dfx DRM depends on PCI
Dave Airlie [Fri, 25 Mar 2005 09:02:20 +0000 (09:02 +0000)]
3dfx DRM depends on PCI
Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
19 years agoFor 2.6.12 and 2.6.11.x:
Dave Airlie [Fri, 25 Mar 2005 09:01:27 +0000 (09:01 +0000)]
For 2.6.12 and 2.6.11.x:
Remove incorrect "drm_"-prefix from parameter description.
Signed-off-by: Magnus Damm <damm@opensource.se>
19 years agoAdd support for production version of ATI RN50/ES1000. (ATI Technologies
Michel Daenzer [Fri, 18 Mar 2005 22:51:43 +0000 (22:51 +0000)]
Add support for production version of ATI RN50/ES1000. (ATI Technologies
    Inc.)

19 years agoadd R200_EMIT_PP_TRI_PERF_CNTL packet to support brilinear filtering on
Roland Scheidegger [Tue, 15 Mar 2005 22:12:30 +0000 (22:12 +0000)]
add R200_EMIT_PP_TRI_PERF_CNTL packet to support brilinear filtering on
    r200

19 years agovia changes:
Thomas Hellstrom [Mon, 14 Mar 2005 22:50:21 +0000 (22:50 +0000)]
via changes:
1. Initialize futex locks to zero on device init.
2. Remove some stray defines from via_drm.h
3. Prepare via_drm.h for drm client inclusion. The goal is to share a
    common file with common definitions.
4. Sync shared / shared-core via_drm.h
5. Bump minor, because of the futex lock initialization.

19 years agoCommand DMA optimizations:
Felix Kuehling [Sun, 13 Mar 2005 02:16:10 +0000 (02:16 +0000)]
Command DMA optimizations:
- don't waste DMA memory when small command buffers are flushed
- minimized padding with noops
- slightly simplified faked DMA flushing

19 years agoFix for bug 2673 from Egbert Eich - memset the version
Dave Airlie [Tue, 8 Mar 2005 23:47:11 +0000 (23:47 +0000)]
Fix for bug 2673 from Egbert Eich - memset the version

19 years agoTracked down random lockups related to command DMA that occurred in Quake3
Felix Kuehling [Mon, 7 Mar 2005 12:19:49 +0000 (12:19 +0000)]
Tracked down random lockups related to command DMA that occurred in Quake3
    after ~5min: buffer aging was subtly broken. Part of this may have also
    affected vertex DMA buffer aging and client-side texture heap aging,
    though with less fatal consequences. Bumped minor version and driver
    date.

19 years agominor pointer cast typo
Dave Airlie [Mon, 7 Mar 2005 09:40:18 +0000 (09:40 +0000)]
minor pointer cast typo

19 years agoAdded support for command DMA on Savage4-based hardware. Unfortunately
Felix Kuehling [Sun, 6 Mar 2005 01:41:06 +0000 (01:41 +0000)]
Added support for command DMA on Savage4-based hardware. Unfortunately
    command and vertex DMA don't work at the same time. Command DMA
    performance is superior and works with all vertex formats. Bumped minor
    version and driver date.

19 years agoadd some idct type 3 packets for reference
Dave Airlie [Fri, 4 Mar 2005 03:04:37 +0000 (03:04 +0000)]
add some idct type 3 packets for reference

19 years agoUse wrap counter to extend 16-bit hardware event tags to 32-bit logical
Felix Kuehling [Wed, 23 Feb 2005 16:34:30 +0000 (16:34 +0000)]
Use wrap counter to extend 16-bit hardware event tags to 32-bit logical
    event counters in the SAVAGE_EVENT_EMIT/WAIT ioctls. This is needed for
    reliable client-side texture heap aging. Slightly simplified DMA buffer
    aging while at it. Bumped minor version and driver date.