Commit Graph

1805 Commits (2dd3c039fdbe86db4043abcc69babf768edb3ab8)

Author SHA1 Message Date
Stephane Marchesin b509abe413 Fix the fifo context size on nv10, nv20 and nv30. 2006-10-13 22:35:22 +02:00
Ben Skeggs 4988074794 Fix some randomness in activating a second channel on NV40 (odd GET/PUT vals). Ch 1 GET now advances, but no ctx_switch. 2006-10-14 06:57:49 +11:00
Stephane Marchesin a9c6c3f21d Oops. 2006-10-12 21:18:55 +02:00
Stephane Marchesin 7ef44b2b8d Still more work on the context switching code. 2006-10-12 17:31:49 +02:00
Thomas Hellstrom 540c64c378 Bugfixes. 2006-10-12 16:10:47 +02:00
Thomas Hellstrom 10150df02b Simplify the AGP backend interface somewhat.
Fix buffer bound caching policy changing, Allow
on-the-fly changing of caching policy on bound buffers if the hardware
supports it.

Allow drivers to use driver-specific AGP memory types for TTM AGP pages.
Will make AGP drivers much easier to migrate.
2006-10-12 12:09:16 +02:00
Stephane Marchesin a749d9d5b4 More work on the context switch code. Still doesn't work. I'm mostly convinced it's an initialization issue. 2006-10-12 01:08:15 +02:00
Thomas Hellstrom 3070389367 Compatibility code for 2.6.15-2.6.18. It is ugly but a little comfort is that
it will go away in the mainstream kernel.
Some bugfixes, mainly in error paths.
2006-10-11 22:21:01 +02:00
Thomas Hellstrom f2db76e2f2 Big update:
Adapt for new functions in the 2.6.19 kernel.
Remove the ability to have multiple regions in one TTM.
   This simplifies a lot of code.
Remove the ability to access TTMs from user space.
   We don't need it anymore without ttm regions.
Don't change caching policy for evicted buffers. Instead change it only
   when the buffer is accessed by the CPU (on the first page fault).
   This tremendously speeds up eviction rates.
Current code is safe for kernels <= 2.6.14.
Should also be OK with 2.6.19 and above.
2006-10-11 13:40:35 +02:00
Stephane Marchesin dd473411f8 Context switching work.
Added preliminary support for context switches (triggers the interrupts, but hangs after the switch ; something's not quite right yet).
Removed the PFIFO_REINIT ioctl. I hope it's that a good idea...
Requires the upcoming commit to the DDX.
2006-10-11 00:28:15 +02:00
Thomas Hellstrom c58574c605 Use a nopage-based approach to fault in pfns. 2006-10-10 10:37:26 +02:00
Roland Scheidegger a9f57a2b9c only allow specific type-3 packets to pass the verifier instead of all for r100/r200 as others might be unsafe (r300 already does this), and add checking for these we need but aren't safe. Check the RADEON_CP_INDX_BUFFER packet on both r200 and r300 as it isn't safe neither. 2006-10-10 02:24:19 +02:00
Thomas Hellstrom cee659afb5 Get rid of all ugly PTE hacks. 2006-10-03 12:08:07 +02:00
George Sapountzis c9e3aa961e Bug 6242: [mach64] Use private DMA buffers, part #4.
mach64_state.c: convert the DRM_MACH64_BLIT ioctl to submit a pointer to
user-space memory rather than a DMA buffer index, similar to DRM_MACH64_VERTEX.

This change allows the DDX to map the DMA buffers read-only and eliminate a
security problem where a client can alter the contents of the DMA buffer after
submission to the DRM.

This change also affects the DRI/DRM interface. Performace-wise, it basically
affects PCI mode where I get a ~12% speedup for some Mesa demos I tested.
This is mainly due to eliminating an ioctl for allocating the DMA buffer.

mach64_dma.c: move the responsibility for allocating memory for the DMA ring
in PCI mode to the DDX.

This change affects the DDX/DRM interface and unifies a couple of PCI/AGP code
paths for ring memory in the DRM.

Bump the mach64 DRM version major and date.
2006-10-02 22:47:26 +03:00
George Sapountzis f3deef730d Bug 6242: [mach64] Use private DMA buffers, part #3.
Add DRM_PCI_BUFFER_RO flag for mapping PCI DMA buffer read-only. An additional
flag is needed, since PCI DMA buffers do not have an associated map.
2006-10-02 22:47:23 +03:00
George Sapountzis 25760c30d4 Bug 6242: [mach64] Use private DMA buffers, part #2.
Factor out from mach64_dma_dispatch_vertex() the code to reclaim an unsed
buffer, in preperation for using it in mach64_dma_dispatch_blit() also.
2006-10-02 22:47:19 +03:00
George Sapountzis eea150e776 Bug 6242: [mach64] Use private DMA buffers, part #1.
Factor out from mach64_freelist_get() the code to reclaim a completed buffer,
this is to improve readability for me.
2006-10-02 22:47:14 +03:00
George Sapountzis d1b31a228b Bug 6209: [mach64] AGP DMA buffers not mapped correctly.
Map the DMA buffers from the same linear address as the vertex bufs. If
dev->agp_buffer_token is not set, mach64 drm maps the DMA buffers from
linear address 0x0.
2006-10-02 22:46:54 +03:00
Michel Dänzer 16be6ba63a Fix type of second argument to spin_lock_irqsave().
(cherry picked from f6238cf624 commit)
2006-10-02 15:42:07 +02:00
Michel Dänzer f6238cf624 Fix type of second argument to spin_lock_irqsave(). 2006-10-02 15:33:19 +02:00
Thomas Hellstrom eacedf41a6 Make the user_token 44-bit for TTMs, and have them occupy a unique file space
starting at 0x00100000000. This will hopefully allow us to use
unmap_mapping_range(). Note that user-space will need
64-bit file offset support.
2006-10-02 15:06:35 +02:00
Thomas Hellstrom a31046b873 Add a buffer object manager for TTM maps. 2006-10-02 14:03:15 +02:00
Thomas Hellstrom d85b99435f Allow for 44 bit user-tokens (or drm_file offsets) 2006-10-02 13:49:43 +02:00
Thomas Hellstrom 418b81c65c Add a comment to previos commit. 2006-10-02 13:37:54 +02:00
Thomas Hellstrom c6be27401f Trap and be verbose about a deadlock that occurs with AIGLX and drivers that
use drm_reclaim_buffers_locked().
2006-10-02 13:34:30 +02:00
Felix Kühling 58a23d193f drm_rmdraw: Declare id and idx as signed so testing for < 0 works as intended.
(cherry picked from d583899681 commit)
2006-10-02 11:21:10 +02:00
Michel Dänzer 3a16e615ca Make locked tasklet handling more robust.
Initialize the spinlock unconditionally when struct drm_device is filled in,
and return early in drm_locked_tasklet() if the driver doesn't support IRQs.
2006-10-02 11:04:42 +02:00
Felix Kühling d583899681 drm_rmdraw: Declare id and idx as signed so testing for < 0 works as intended. 2006-10-02 10:50:40 +02:00
Thomas Hellstrom 8e908eaf50 Bump driver date. 2006-09-29 14:21:51 +02:00
Michel Dänzer 17a640419a i915: Only schedule vblank tasklet if there are scheduled swaps pending.
This fixes issues on X server startup with versions of xf86-video-intel that
enable the IRQ before they have a context ID.
(cherry picked from 7af93dd984 commit)
2006-09-29 12:55:09 +02:00
Michel Dänzer 48367fdfe6 i915: Only initialize IRQ fields in postinstall, not the PIPE_SET ioctl.
Some other minor changes in preparation for actually disabling user interrupts.
2006-09-29 12:55:09 +02:00
Michel Dänzer 3620a3ec85 i915: Bump minor again to differentiate from vsync changes. 2006-09-29 12:55:09 +02:00
Michel Dänzer 390184df92 i915: Avoid mis-counting vblank interrupts when they're only enabled for pipe A.
It looks like 'after a while', I915REG_INT_IDENTITY_R for some reason always has
VSYNC_PIPEB_FLAG set in the interrupt handler, even though pipe B is disabled.
So we only increase dev->vbl_received if the corresponding bit is also set in
dev->vblank_pipe.
(cherry picked from 881ba56992 commit)
2006-09-29 12:55:09 +02:00
Michel Dänzer c0bff9f9cd i915: Bump minor for swap scheduling ioctl and secondary vblank support.
(cherry picked from 2627131e5d commit)
2006-09-29 12:55:09 +02:00
Michel Dänzer 0a7d9edcfb i915_vblank_swap: Add support for DRM_VBLANK_NEXTONMISS.
(cherry picked from 0356fe260d commit)
2006-09-29 12:55:09 +02:00
Michel Dänzer c47ebd9707 Only return EBUSY after we've established we need to schedule a new swap.
(cherry picked from 50a0284a61 commit)
2006-09-29 12:55:09 +02:00
Michel Dänzer cef0f24325 Core vsync: Don't clobber target sequence number when scheduling signal.
It looks like this would have caused signals to always get sent on the next
vertical blank, regardless of the sequence number.
(cherry picked from cf6b2c5299 commit)
2006-09-29 12:55:08 +02:00
Michel Dänzer ed82172378 Core vsync: Add flag DRM_VBLANK_NEXTONMISS.
When this flag is set and the target sequence is missed, wait for the next
vertical blank instead of returning immediately.
(cherry picked from 89e323e490 commit)
2006-09-29 12:55:08 +02:00
Michel Dänzer c4c47a7eac Fix 'sequence has passed' condition in i915_vblank_swap().
(cherry picked from 7f09f957d9 commit)
2006-09-29 12:55:08 +02:00
Michel Dänzer f9aa4f5973 Add SAREA fileds for determining which pipe to sync window buffer swaps to.
(cherry picked from c2bdb76814 commit)
2006-09-29 12:55:08 +02:00
Michel Dänzer 316e736768 Add definition of DRM_VBLANK_SECONDARY.
(cherry picked from 84b38b63f0 commit)
2006-09-29 12:55:08 +02:00
Michel Dänzer 4a3d270862 Make handling of dev_priv->vblank_pipe more robust.
Initialize it to default value if it hasn't been set by the X server yet.

In i915_vblank_pipe_set(), only update dev_priv->vblank_pipe and call
i915_enable_interrupt() if the argument passed from userspace is valid to avoid
corrupting dev_priv->vblank_pipe on invalid arguments.
(cherry picked from 87c57cba1a commit)
2006-09-29 12:55:08 +02:00
Michel Dänzer 1f3493f65b DRM_I915_VBLANK_SWAP ioctl: Take drm_vblank_seq_type_t instead of pipe number.
Handle relative as well as absolute target sequence numbers.

Return error if target sequence has already passed, so userspace can deal with
this situation as it sees fit.

On success, return the sequence number of the vertical blank when the buffer
swap is expected to take place.

Also add DRM_IOCTL_I915_VBLANK_SWAP definition for userspace code that may want
to use ioctl() instead of drmCommandWriteRead().
(cherry picked from d5a0f10751 commit)
2006-09-29 12:55:08 +02:00
Michel Dänzer 00531cecad Change first valid DRM drawable ID to be 1 instead of 0.
This makes it easier for userspace to know when it needs to allocate an ID.

Also free drawable information memory when it's no longer needed.
(cherry picked from df7551ef73 commit)
2006-09-29 12:55:08 +02:00
Michel Dänzer 7d487602a3 Add copyright notice.
(cherry picked from d04751face commit)
2006-09-29 12:55:08 +02:00
Michel Dänzer da75d59cd6 i915: Add ioctl for scheduling buffer swaps at vertical blanks.
This uses the core facility to schedule a driver callback that will be called
ASAP after the given vertical blank interrupt with the HW lock held.
(cherry picked from 257771fa29 commit)
2006-09-29 12:55:08 +02:00
Michel Dänzer d7389a9758 Locking and memory management fixes.
(cherry picked from 23d2833aaa commit)
2006-09-29 12:55:08 +02:00
Michel Dänzer ca3a1b5ec4 Drop tasklet locked driver callback when uninstalling IRQ.
(cherry picked from b9f3009160 commit)
2006-09-29 12:55:08 +02:00
Michel Dänzer f93e482269 Export drm_get_drawable_info symbol from core.
(cherry picked from 43f8675534 commit)
2006-09-29 12:55:08 +02:00
Michel Dänzer 67e88e5628 Hook up DRM_IOCTL_UPDATE_DRAW ioctl.
(cherry picked from 98a8950458 commit)
2006-09-29 12:55:08 +02:00