profile/ivi/libdrm.git
17 years agoRemove broken crtc enable checks, radeon does it slightly differently
Jesse Barnes [Sat, 16 Jun 2007 00:06:46 +0000 (17:06 -0700)]
Remove broken crtc enable checks, radeon does it slightly differently
(this makes get_vblank_counter return an actual value).

17 years agoFirst cut at radeon support for the vblank rework.
Jesse Barnes [Fri, 15 Jun 2007 18:21:57 +0000 (11:21 -0700)]
First cut at radeon support for the vblank rework.

17 years agoRemove DRIVER_IRQ_VBL(2).
Michel Dänzer [Fri, 15 Jun 2007 09:01:51 +0000 (11:01 +0200)]
Remove DRIVER_IRQ_VBL(2).

If the driver doesn't support vertical blank interrupts, it won't call
drm_vblank_init(), and dev->num_crtcs will be 0.

Also fix an off-by-one test against dev->num_crtcs.

17 years agoMake vblank waitqueue per CRTC.
Michel Dänzer [Fri, 15 Jun 2007 08:49:16 +0000 (10:49 +0200)]
Make vblank waitqueue per CRTC.

17 years agoWake up vblank waitqueue in drm_handle_vblank().
Michel Dänzer [Fri, 15 Jun 2007 08:25:50 +0000 (10:25 +0200)]
Wake up vblank waitqueue in drm_handle_vblank().

17 years agoi915: Fix tests for vblank interrupts being enabled on CRTC by X server.
Michel Dänzer [Fri, 15 Jun 2007 08:21:44 +0000 (10:21 +0200)]
i915: Fix tests for vblank interrupts being enabled on CRTC by X server.

17 years agoReturn current sequence number to userspace after blocking wait for vblank.
Michel Dänzer [Fri, 15 Jun 2007 08:12:23 +0000 (10:12 +0200)]
Return current sequence number to userspace after blocking wait for vblank.

17 years agoFix memory leaks in vblank error paths.
Michel Dänzer [Fri, 15 Jun 2007 08:10:33 +0000 (10:10 +0200)]
Fix memory leaks in vblank error paths.

Also use drm_calloc instead of drm_alloc and memset, and use the size of the
struct instead of the size of the pointer for allocation...

17 years agoComment new vblank routines and fixup several issues:
Jesse Barnes [Thu, 14 Jun 2007 18:32:31 +0000 (11:32 -0700)]
Comment new vblank routines and fixup several issues:
  - use correct refcount variable in get/put routines
  - extract counter update from drm_vblank_get
  - make signal handling callback per-crtc
  - update interrupt handling logic, drivers should use drm_handle_vblank
  - move wakeup and counter update logic to new drm_handle_vblank routine
  - fixup usage of get/put in light of counter update extraction
  - fix longstanding bug in signal code, update pending counter only
    *after* we're sure we'll setup signal handling

17 years agoRemove unnecessary (and uncommented!) read barrier from the interrupt
Jesse Barnes [Tue, 12 Jun 2007 23:29:09 +0000 (16:29 -0700)]
Remove unnecessary (and uncommented!) read barrier from the interrupt
path.  It doesn't appear to serve any useful purpose.

17 years agoUpdate vblank code:
Jesse Barnes [Tue, 12 Jun 2007 20:35:41 +0000 (13:35 -0700)]
Update vblank code:
  - move pre/post modeset ioctl to core
  - fixup i915 buffer swap
  - fix outstanding signal count code
  - create new core vblank init routine
  - test (works with glxgears)
  - simplify i915 interrupt handler

17 years agoInitial checkin of vblank rework. Code attempts to reduce the number
Jesse Barnes [Tue, 12 Jun 2007 17:44:21 +0000 (10:44 -0700)]
Initial checkin of vblank rework.  Code attempts to reduce the number
of vblank interrupt in order to save power.

17 years agouse krh's idr mods to remove lists from idr code
Dave Airlie [Sun, 10 Jun 2007 05:40:10 +0000 (15:40 +1000)]
use krh's idr mods to remove lists from idr code

17 years agor300: Added the CP maximum fetch size and ring rptr update variables.
Oliver McFadden [Fri, 8 Jun 2007 19:40:57 +0000 (19:40 +0000)]
r300: Added the CP maximum fetch size and ring rptr update variables.

17 years agooops must fix this properly at some point
Dave Airlie [Thu, 7 Jun 2007 08:45:00 +0000 (18:45 +1000)]
oops must fix this properly at some point

17 years agodrm: fix radeon setparam alignment issues on 32/64-bit
Dave Airlie [Thu, 7 Jun 2007 08:40:41 +0000 (18:40 +1000)]
drm: fix radeon setparam alignment issues on 32/64-bit

17 years agor300: Small correction to the previous commit.
Oliver McFadden [Tue, 5 Jun 2007 19:19:42 +0000 (19:19 +0000)]
r300: Small correction to the previous commit.

17 years agor300: Document more of the RADEON_RBBM_STATUS register.
Alex Deucher [Tue, 5 Jun 2007 19:05:49 +0000 (19:05 +0000)]
r300: Document more of the RADEON_RBBM_STATUS register.

17 years agoAdd support for the G33, Q33, and Q35 chipsets.
Wang Zhenyu [Tue, 5 Jun 2007 18:15:29 +0000 (11:15 -0700)]
Add support for the G33, Q33, and Q35 chipsets.

These require that the status page be referenced by a pointer in GTT, rather
than phsyical memory.  So, we have the X Server allocate that memory and tell
us the address, instead.

17 years agoremove include of linux ioctl32.h from drm drivers
Dave Airlie [Tue, 5 Jun 2007 08:14:54 +0000 (18:14 +1000)]
remove include of linux ioctl32.h from drm drivers

17 years agonouveau: fix RAMHT wrapping
Maurice van der Pot [Mon, 4 Jun 2007 00:49:30 +0000 (10:49 +1000)]
nouveau: fix RAMHT wrapping

17 years agoradeon: refine irq acking for vbl on crtc 2
Dave Airlie [Sun, 3 Jun 2007 08:30:52 +0000 (18:30 +1000)]
radeon: refine irq acking for vbl on crtc 2

17 years agoRevert "drm: add new drm_wait_on function to replace macro"
root [Sun, 3 Jun 2007 08:12:28 +0000 (18:12 +1000)]
Revert "drm: add new drm_wait_on function to replace macro"

This reverts commit 6e860d08d0f5b1e9a2d711aaf9fd6b982aa8039e.

As I said not a good plan - this macro will have to stay for now,
trying to do the vbl code with the inline was a bit messy - may need specialised
drm wait on functions

17 years agoRevert "move i915 to new drm_wait_on function"
root [Sun, 3 Jun 2007 08:11:44 +0000 (18:11 +1000)]
Revert "move i915 to new drm_wait_on function"

This reverts commit feb68037784ac09e333a321d294fdb2d8c57a4c8.

This was a bad idea, the macro is actually a bit harder to convert
to a static for the other use cases

17 years agoradeon: add support for vblank on crtc2
Dave Airlie [Sun, 3 Jun 2007 06:28:21 +0000 (16:28 +1000)]
radeon: add support for vblank on crtc2

This add support for CRTC2 vblank on radeon similiar to the i915 support

17 years agodrm: fixup initialisation of list heads and idr
Dave Airlie [Fri, 1 Jun 2007 09:00:24 +0000 (19:00 +1000)]
drm: fixup initialisation of list heads and idr

17 years agoi915: Add support for 945GME chip
Wang Zhenyu [Wed, 30 May 2007 08:25:49 +0000 (16:25 +0800)]
i915: Add support for 945GME chip

17 years agoi915: Add support for 965GME/GLE chip.
Wang Zhenyu [Wed, 30 May 2007 08:24:42 +0000 (16:24 +0800)]
i915: Add support for 965GME/GLE chip.

17 years agoUpdate a bunch of FreeBSD port code.
Jung-uk Kim [Tue, 15 May 2007 20:35:33 +0000 (13:35 -0700)]
Update a bunch of FreeBSD port code.

Tested on r200/r300.  i915 updates still remain to be done.

17 years agoreformatting, clean-ups
Brian [Tue, 29 May 2007 20:56:17 +0000 (14:56 -0600)]
reformatting, clean-ups

17 years agoClean-ups and reformatting.
Brian [Tue, 29 May 2007 20:54:00 +0000 (14:54 -0600)]
Clean-ups and reformatting.

Use 4-space indentation consistently.
Replace occurances of:
if (cond) code;
with:
if (cond)
code;
to facilitate putting breakpoints on code.

17 years agodrm: move context handling code to use linux idr
Dave Airlie [Sat, 26 May 2007 22:44:38 +0000 (08:44 +1000)]
drm: move context handling code to use linux idr

17 years agodrm: convert drawable handling to use Linux idr
Dave Airlie [Sat, 26 May 2007 21:26:52 +0000 (07:26 +1000)]
drm: convert drawable handling to use Linux idr

This cleans this code up a lot and uses the generic Linux idr which is
designed for this.

Signed-off-by: Dave Airlie <airlied@linux.ie>
17 years agodrm: spinlock initializer cleanup
Thomas Gleixner [Fri, 25 May 2007 19:20:59 +0000 (05:20 +1000)]
drm: spinlock initializer cleanup

17 years agodrm: make sure the drawable code doesn't call malloc(0).
Michel Dänzer [Fri, 25 May 2007 18:35:54 +0000 (04:35 +1000)]
drm: make sure the drawable code doesn't call malloc(0).

Signed-off-by: Michel Dänzer <michel@tungstengraphics.com>
Signed-off-by: Dave Airlie <airlied@linux.ie>
17 years agoradeon: add other IGP chipsets
Dave Airlie [Fri, 25 May 2007 18:02:55 +0000 (04:02 +1000)]
radeon: add other IGP chipsets

17 years agoRevert "drm/ttm: cleanup mm_ioctl ioctls to be separate ioctls."
Dave Airlie [Fri, 25 May 2007 17:48:08 +0000 (03:48 +1000)]
Revert "drm/ttm: cleanup mm_ioctl ioctls to be separate ioctls."

This reverts commit 3fdef0dc2000308b16907b95f637c60acde80a74.

ditto not on master yet

17 years agoRevert "drm/ttm: cleanup most of fence ioctl split out"
Dave Airlie [Fri, 25 May 2007 17:47:48 +0000 (03:47 +1000)]
Revert "drm/ttm: cleanup most of fence ioctl split out"

This reverts commit 3dfc1400e9fc58c69292d7cf7c2e1653fa5e6991.

this shouldn't have gone on master yet

17 years agowhitespace fixups from kernel
Dave Airlie [Tue, 8 May 2007 05:43:29 +0000 (15:43 +1000)]
whitespace fixups from kernel

17 years agoati_pcigart: cut to 80 chars
Dave Airlie [Tue, 8 May 2007 05:09:48 +0000 (15:09 +1000)]
ati_pcigart: cut to 80 chars

17 years agodrm/ttm: cleanup most of fence ioctl split out
Dave Airlie [Sun, 6 May 2007 01:35:11 +0000 (11:35 +1000)]
drm/ttm: cleanup most of fence ioctl split out

17 years agodrm/ttm: cleanup mm_ioctl ioctls to be separate ioctls.
Dave Airlie [Sun, 6 May 2007 01:17:30 +0000 (11:17 +1000)]
drm/ttm: cleanup mm_ioctl ioctls to be separate ioctls.

This is the first bunch of ioctls

17 years agodrm: cleanup use of Linux list handling macros
Dave Airlie [Fri, 25 May 2007 18:26:24 +0000 (04:26 +1000)]
drm: cleanup use of Linux list handling macros

This makes the drms use of the list handling macros a lot cleaner
and more along the lines of how they should be used.

17 years agonouveau: don't build the module by default
Stephane Marchesin [Thu, 17 May 2007 15:35:14 +0000 (17:35 +0200)]
nouveau: don't build the module by default

17 years agor300: Added my comments into r300_reg.h.
Oliver McFadden [Sun, 13 May 2007 16:18:54 +0000 (16:18 +0000)]
r300: Added my comments into r300_reg.h.

17 years agor300: Synchronized R300 register defines file.
Oliver McFadden [Sun, 13 May 2007 07:53:40 +0000 (07:53 +0000)]
r300: Synchronized R300 register defines file.

Just moved the indent control comments so that indent doesn't try to change
anything.

17 years agonouveau : nv10 graph move clipping value to per channel init
Matthieu Castet [Sat, 12 May 2007 13:35:39 +0000 (15:35 +0200)]
nouveau : nv10 graph move clipping value to per channel init

17 years agonouveau : nv10 graph clipping values were forgoten in ddx to drm commit
Matthieu Castet [Sat, 12 May 2007 13:16:37 +0000 (15:16 +0200)]
nouveau : nv10 graph clipping values were forgoten in ddx to drm commit

17 years agoAllow vblank interrupts to remain disabled across VT switch.
Keith Packard [Mon, 7 May 2007 13:07:48 +0000 (09:07 -0400)]
Allow vblank interrupts to remain disabled across VT switch.

i915_driver_irq_postinstall was forcing vblank interrupts to pipe A when
called with vblank interrupts disabled. This caused vblank interrupts to be
accidentally re-enabled when VT switching the X server. Instead, start the
driver with vblank interrupts enabled on pipe A to support older X servers,
but then leave control over the state to the X server if it is able to do so.

17 years agor300: Synchronized R300 register defines file.
Oliver McFadden [Wed, 9 May 2007 18:31:31 +0000 (18:31 +0000)]
r300: Synchronized R300 register defines file.

17 years agor300: Synchronized R300 register defines file.
Oliver McFadden [Wed, 9 May 2007 15:22:09 +0000 (15:22 +0000)]
r300: Synchronized R300 register defines file.

17 years agonouveau : fix fifo context size for nv10
Matthieu Castet [Tue, 8 May 2007 19:18:02 +0000 (21:18 +0200)]
nouveau : fix fifo context size for nv10

17 years agolibdrm: fix typo on ttm fence interface
Dave Airlie [Tue, 8 May 2007 07:51:21 +0000 (17:51 +1000)]
libdrm: fix typo on ttm fence interface

17 years agos/OJBECTS/OBJECTS/
Brian [Sun, 6 May 2007 16:09:59 +0000 (10:09 -0600)]
s/OJBECTS/OBJECTS/

17 years agor300: Synchronize the register file from Mesa.
Oliver McFadden [Sun, 6 May 2007 12:47:03 +0000 (12:47 +0000)]
r300: Synchronize the register file from Mesa.

17 years agor300: Use the defined names for known registers.
Oliver McFadden [Sun, 6 May 2007 12:35:16 +0000 (12:35 +0000)]
r300: Use the defined names for known registers.

17 years agodrm: remove unused define DRM_HAS_TTM
Dave Airlie [Sun, 6 May 2007 00:50:10 +0000 (10:50 +1000)]
drm: remove unused define DRM_HAS_TTM

17 years agoFix buffer object list freeing.
Thomas Hellstrom [Wed, 2 May 2007 15:13:32 +0000 (17:13 +0200)]
Fix buffer object list freeing.

17 years agoFix userspace ABI breakage from 3c384a9ad5f964709a237cfe035ea5d6df2da5fa.
Michel Dänzer [Tue, 1 May 2007 15:03:55 +0000 (17:03 +0200)]
Fix userspace ABI breakage from 3c384a9ad5f964709a237cfe035ea5d6df2da5fa.

17 years agoradeon: Don't mess up page flipping when a file descriptor is closed.
Michel Dänzer [Sun, 29 Apr 2007 10:17:57 +0000 (12:17 +0200)]
radeon: Don't mess up page flipping when a file descriptor is closed.

There can still be other contexts that may use page flipping later on, so don't
just unilaterally 'clean it up', which could lead to the wrong page being
displayed, e.g. when running 3D apps with a GLX compositing manager such as
compiz using page flipping.

17 years agomove i915 to new drm_wait_on function
Dave Airlie [Sat, 28 Apr 2007 05:07:22 +0000 (15:07 +1000)]
move i915 to new drm_wait_on function

17 years agodrm: add new drm_wait_on function to replace macro
Dave Airlie [Sat, 28 Apr 2007 05:05:20 +0000 (15:05 +1000)]
drm: add new drm_wait_on function to replace macro

17 years agoremove register usage
Dave Airlie [Sat, 28 Apr 2007 05:04:49 +0000 (15:04 +1000)]
remove register usage

17 years agoremove DRM_GETSAREA and replace with drm_getsarea function
Dave Airlie [Sat, 28 Apr 2007 04:49:27 +0000 (14:49 +1000)]
remove DRM_GETSAREA and replace with drm_getsarea function

17 years agoRevert "bug 7092 : add pci ids for mach64 in Dell poweredge 4200"
George Sapountzis [Thu, 26 Apr 2007 11:16:51 +0000 (14:16 +0300)]
Revert "bug 7092 : add pci ids for mach64 in Dell poweredge 4200"

This reverts commit 255f3e6f76dfd267a14765dd1293229184298d89.

Rage IIc does not have a vertex setup engine.

17 years agolinux: minor indent.
George Sapountzis [Thu, 26 Apr 2007 11:16:28 +0000 (14:16 +0300)]
linux: minor indent.

17 years agofreebsd: remove stray apperance of IN_MODULE.
George Sapountzis [Thu, 26 Apr 2007 11:16:13 +0000 (14:16 +0300)]
freebsd: remove stray apperance of IN_MODULE.

The xserver no longer uses the libc-wrapper.

17 years agolibdrm: remove HAVE_XORG_CONFIG_H and XFree86LOADER ifdef's.
George Sapountzis [Thu, 26 Apr 2007 11:15:55 +0000 (14:15 +0300)]
libdrm: remove HAVE_XORG_CONFIG_H and XFree86LOADER ifdef's.

We no longer import libdrm in the xserver.

17 years agoupdate create_lk_drm.sh script
Dave Airlie [Thu, 26 Apr 2007 07:54:06 +0000 (17:54 +1000)]
update create_lk_drm.sh script

17 years agowhitespace cleanup
Dave Airlie [Thu, 26 Apr 2007 06:25:43 +0000 (16:25 +1000)]
whitespace cleanup

17 years agoremove some whitespace violation
Dave Airlie [Thu, 26 Apr 2007 06:22:28 +0000 (16:22 +1000)]
remove some whitespace violation

17 years agodrm: nopfn warning on 64-bit
Dave Airlie [Thu, 26 Apr 2007 06:22:11 +0000 (16:22 +1000)]
drm: nopfn warning on 64-bit

17 years agodrm/ttm: allow drm_buffer_object_create to be used in-kernel
Jesse Barnes [Thu, 26 Apr 2007 06:09:24 +0000 (16:09 +1000)]
drm/ttm: allow drm_buffer_object_create to be used in-kernel

17 years agoAdd new buffer object type for kernel allocations that don't initially have a user...
Jesse Barnes [Wed, 11 Apr 2007 19:51:52 +0000 (12:51 -0700)]
Add new buffer object type for kernel allocations that don't initially have a user mapping.
(cherry picked from commit 2e21779992bd5026d8ec4dea52466377dbe5a0ed)

17 years agoMore detailed instructions, tips.
Brian [Wed, 25 Apr 2007 20:52:29 +0000 (14:52 -0600)]
More detailed instructions, tips.

17 years agonouveau: fix wacky pci id
Stephane Marchesin [Mon, 23 Apr 2007 20:37:01 +0000 (22:37 +0200)]
nouveau: fix wacky pci id

17 years agoInitialize rwlock using rwlock_init to appease lockdep validator.
Kristian Høgsberg [Fri, 20 Apr 2007 22:06:31 +0000 (18:06 -0400)]
Initialize rwlock using rwlock_init to appease lockdep validator.

17 years agoAvoid a fence timeout problem when a signal is pending.
Thomas Hellstrom [Thu, 19 Apr 2007 21:25:14 +0000 (23:25 +0200)]
Avoid a fence timeout problem when a signal is pending.

17 years agoFix via compile error
Thomas Hellstrom [Fri, 20 Apr 2007 13:49:31 +0000 (15:49 +0200)]
Fix via compile error
(Reported by Benno Schulenberg.)

17 years agoSimplify the ttm backend interface and the agp ttm backend.
Thomas Hellstrom [Wed, 18 Apr 2007 14:33:28 +0000 (16:33 +0200)]
Simplify the ttm backend interface and the agp ttm backend.

17 years agoDon't always free up memory space when we unpin buffers.
Thomas Hellstrom [Tue, 17 Apr 2007 12:15:37 +0000 (14:15 +0200)]
Don't always free up memory space when we unpin buffers.

17 years agoFix buffer object reference problems.
Thomas Hellstrom [Tue, 17 Apr 2007 08:53:19 +0000 (10:53 +0200)]
Fix buffer object reference problems.
(Reported by Dave Airlie).

17 years agovia: Make sure we flush write-combining using a follow-up read.
Thomas Hellstrom [Tue, 17 Apr 2007 06:58:23 +0000 (08:58 +0200)]
via: Make sure we flush write-combining using a follow-up read.

17 years agoAdd a code comment.
Thomas Hellstrom [Tue, 17 Apr 2007 06:46:45 +0000 (08:46 +0200)]
Add a code comment.

17 years agoRequire the hardware lock for buffer creation
Thomas Hellstrom [Mon, 16 Apr 2007 14:23:05 +0000 (16:23 +0200)]
Require the hardware lock for buffer creation
(since that implies a validate).
Fix drm_bo_wait_unfenced error messages and codes.
Fix some return codes from libdrm.

17 years agonouveau: nv10 per channel init from ddx
Matthieu Castet [Tue, 10 Apr 2007 21:19:29 +0000 (23:19 +0200)]
nouveau: nv10 per channel init from ddx

17 years agors480: Renamed some unknown registers. See dri-devel list.
Oliver McFadden [Mon, 9 Apr 2007 23:23:40 +0000 (23:23 +0000)]
rs480: Renamed some unknown registers. See dri-devel list.

17 years agonouveau: NV46 support
Ben Skeggs [Mon, 9 Apr 2007 13:20:26 +0000 (23:20 +1000)]
nouveau: NV46 support

17 years agoradeon: bump version for IGPGART support
Dave Airlie [Mon, 9 Apr 2007 12:00:34 +0000 (22:00 +1000)]
radeon: bump version for IGPGART support

17 years agoradeon: add support for reverse engineered xpress200m
Dave Airlie [Mon, 9 Apr 2007 11:52:59 +0000 (21:52 +1000)]
radeon: add support for reverse engineered xpress200m

The IGPGART setup code was traced using mmio-trace on fglrx by myself
and Phillip Ezolt <phillipezolt@gmail.com> on dri-devel.

This code doesn't let the 3D driver work properly as the card has no
vertex shader support.

Thanks to Matthew Garrett + Ubuntu for providing me some hardware to do this
work on.

17 years agofixup install target, not sure what I was smokin...
Dave Airlie [Fri, 6 Apr 2007 21:21:05 +0000 (07:21 +1000)]
fixup install target, not sure what I was smokin...

17 years agoi915: use breadcrumb macro everywhere
Dave Airlie [Sun, 1 Apr 2007 06:30:52 +0000 (16:30 +1000)]
i915: use breadcrumb macro everywhere

17 years agonouveau: make a note about a bit that breaks some cards
Ben Skeggs [Thu, 5 Apr 2007 17:24:39 +0000 (03:24 +1000)]
nouveau: make a note about a bit that breaks some cards

17 years agonouveau: Power up all card units by default on startup.
Ben Skeggs [Thu, 5 Apr 2007 17:03:59 +0000 (03:03 +1000)]
nouveau: Power up all card units by default on startup.

17 years agoadd an install target to the drm modules makefile
Dave Airlie [Thu, 5 Apr 2007 01:18:00 +0000 (11:18 +1000)]
add an install target to the drm modules makefile

17 years agoFix user object reference when caller is not object creator.
Thomas Hellstrom [Tue, 3 Apr 2007 08:54:23 +0000 (10:54 +0200)]
Fix user object reference when caller is not object creator.
(Reported by Dave Airlie).

17 years agoAdd a fence flush event to each fence-signaled check when lazy-waiting
Thomas Hellstrom [Tue, 3 Apr 2007 08:24:06 +0000 (10:24 +0200)]
Add a fence flush event to each fence-signaled check when lazy-waiting

to make sure we don't lose any sequence numbers if, for some reason,
they don't generate an IRQ.

17 years agoMake sure we ack irqs before we read a breadcrumb so that
Thomas Hellstrom [Tue, 3 Apr 2007 08:21:41 +0000 (10:21 +0200)]
Make sure we ack irqs before we read a breadcrumb so that

breadcrumb updates that occur _AFTER_ we've read the breadcrumb really
generates a new IRQ.

17 years agoEvicted no-move buffers can get lost if they end up in another
Thomas Hellstrom [Fri, 30 Mar 2007 12:14:26 +0000 (14:14 +0200)]
Evicted no-move buffers can get lost if they end up in another

memory type than local.

17 years agoFix an oops when trying to clean a not yet initialized memory type.
Thomas Hellstrom [Fri, 30 Mar 2007 10:23:22 +0000 (12:23 +0200)]
Fix an oops when trying to clean a not yet initialized memory type.