summaryrefslogtreecommitdiff
path: root/linux-core
AgeCommit message (Collapse)Author
2007-02-08Simplify external ttm page allocation.Thomas Hellstrom
Implement a memcpy fallback for copying between buffers.
2007-02-08Update memory compatibility tests.Thomas Hellstrom
Now only pinned buffers are broken.
2007-02-07Checkpoint commit.Thomas Hellstrom
Flag handling and memory type selection cleanup. glxgears won't start.
2007-02-07Fix a stray unlock_kernel() in drm_vm.cThomas Hellstrom
Add a file for memory move helpers, drm_bo_move.c Implement generic memory move. Cached, no_move and unmapped memory temporarily broken.
2007-02-06Simplify pci map vs no pci map choice.Thomas Hellstrom
2007-02-06Implement a drm_mem_reg_t substructure in the buffer object type.Thomas Hellstrom
2007-02-06Implement a policy for selecting memory types.Thomas Hellstrom
2007-02-05i915: Add copy-blit operation.Thomas Hellstrom
2007-02-02Make also later kernels work with buffer object vmThomas Hellstrom
and clean up some function names.
2007-02-02Make vm handle buffer objects instead of ttm objects.Thomas Hellstrom
Remove ttm objects. Make vm aware of PCI memory type buffer objects. (Only works for pre 2.6.16 kernels for now).
2007-02-01Fix missing ttm_open_vma call from previous commit.Thomas Hellstrom
Honour the ttm backend cant-use-aperture flag.
2007-02-01Prepare for removal of the ttm_object type.Thomas Hellstrom
2007-02-01Protect drm_mmap against disappearing maps.Thomas Hellstrom
The map lists and hash tables are protected using dev->struct_mutex, but drm_mmap strangely never locked this mutex.
2007-01-31memory manager: Make device driver aware of different memory types.Thomas Hellstrom
Memory types are either fixed (on-card or pre-bound AGP) or not fixed (dynamically bound) to an aperture. They also carry information about: 1) Whether they can be mapped cached. 2) Whether they are at all mappable. 3) Whether they need an ioremap to be accessible from kernel space. In this way VRAM memory and, for example, pre-bound AGP appear identical to the memory manager. This also makes support for unmappable VRAM simple to implement.
2007-01-31Fix an error-path oops.Thomas Hellstrom
2007-01-30Add a buffer object transfer function.Thomas Hellstrom
Creates a placeholder for the old buffer contents when it is transfered to / from static memory like VRAM.
2007-01-30Clean up buffer object destruction somewhat.Thomas Hellstrom
2007-01-29Use pre-defined list_splice function.Thomas Hellstrom
2007-01-29s/buf/bo/ for consistency.Thomas Hellstrom
2007-01-29Some cleanup. A buffer object should only have one active memory type.Thomas Hellstrom
2007-01-25Remove a scary error printed when we were leaking memory caches.Thomas Hellstrom
We don't use memory caches anymore... Fix memory accounting initialization to only use low or DMA32 memory.
2007-01-13nouveau: nv20 graph ctx switch.Matthieu Castet
Untested...
2007-01-13nouveau: first step to make graph ctx worksMatthieu Castet
It is still not working, but now we could use some 3D commands without needed to run nvidia blob before.
2007-01-13nouveau: add missing symlinkDave Airlie
2007-01-12nouveau: get nv30 context switching to work.Jeremy Kolb
* Pulled in some registers from nv10reg.h. Needed for context switching. * Filled in nv30 graphics context (based on nv40_graph.c). * Figure out nv30 context table, set up on context creation. Allows the cards automatic switching to work.
2007-01-10Revert "nouveau: Fill in context_init for nv10-nv3x."Dave Airlie
This reverts ac076cb9aff976e8fae567cfa82a898bfc2712e8 commit. Well it didn't do anything interesting...
2007-01-09nouveau: Fill in context_init for nv10-nv3x.Jeremy Kolb jkolb@brandeis.edu
Fill in the context with the values from PFIFO_CACH1. This should work from nv10 through the nv30 series.
2007-01-10nouveau: Don't use DRIVER_USE_MTRR, we already setup our own mtrr over vram.Stephane Marchesin
2007-01-09ttm: make ttm alloc/free into alloc_pages/free_pagesDave Airlie
Add a vmalloc flag to the page flags
2007-01-08add export symbol for memory managerDave Airlie
2007-01-08i830: complete fix for i830 mapsDave Airlie
2007-01-08drm: remove drm_follow_page, and drm_ioremap and ioremapfreeChristoph Hellwig
This comes from the Re: [patch] paravirt: isolate module ops on lkml It needs some testing, please report any regressions caused. Signed-off-by: Dave Airlie <airlied@linux.ie>
2007-01-08fixup i810/i830 to use drm_core_ioremap instead of drm_ioremapDave Airlie
2007-01-05nouveau: Add an mtrr over the whole FBStephane Marchesin
2007-01-02linux-core: Make git ignore generated module symbol version files.Michel Dänzer
2007-01-02nouveau: Add nv40-specific PGRAPH code, not hooked up yet.Ben Skeggs
2007-01-01make build against 2.6.20 hopefullyDave Airlie
2007-01-01fixup permission along line of kernelDave Airlie
2006-12-27Proper allocation of AGP pages for ttms.Thomas Hellstrom
2006-12-21Bug #9120.Thomas Hellstrom
Require at least agpgart version 0.102 for the AGP TTM backend. This should hopefully avoid crashes when the wrong agpgart driver is installed.
2006-12-21Improve memory manager accounting printout formatting.Thomas Hellstrom
2006-12-21Fix buggy aligned allocations.Thomas Hellstrom
2006-12-20Remove the stupid root_node field from the core memory manager.Thomas Hellstrom
2006-12-20Replace vmalloc_32.Thomas Hellstrom
The vmalloc_32 function together with the memset to clear the new pages are replaced with a vmalloc_user. A pre-2.6.18 compat vmalloc_user is added. Please replace any breakage on machines with > 1GB of memory.
2006-12-20Some via PCI posting flushes.Thomas Hellstrom
2006-12-20Merge branch 'nouveau-1'Dave Airlie
2006-12-20fixup symlinks via MakefileDave Airlie
2006-12-20add nouveau symlinks via gitDave Airlie
2006-12-19Security fix. Zero pages before they are handed to user space.Thomas Hellstrom
TTM pages were not cleared when allocated and handed to user space. Sensitive information may leak.
2006-12-19Security fix. Zero pages before they are handed to user space.Thomas Hellstrom
Shared memory areas were not cleared when they are allocated and handed to user space. Sensitive information may leak.