summaryrefslogtreecommitdiff
path: root/bsd-core/drm_bufs.c
AgeCommit message (Collapse)Author
2006-12-01Core build fix for BSD.Michel Dänzer
2005-11-08Correct another LOR issue with resource allocation. This leaves theEric Anholt
drm_get_resource_* resource allocation a little racy, but they're getting called at either X Server startup or driver load, so it's serialized anyway.
2005-11-08Initial port of savage to FreeBSD for the AGP and !ShadowStatus case. AddsEric Anholt
drm_mtrr_{add,del} for handling the MTRR setup. Still has a LOR issue with DRM_VERIFYAREA_READ/DRM_COPY_FROM_USER_UNCHECKED in savage_bci.c -- this won't work with the fine-grained locking in use, and just doing a single copyin to a temporary will probably work fine. Also note that the module leaks approximately 4 kb on unload.
2005-11-08Correct a LOR related to the PCI resource allocations by simply moving theEric Anholt
uninit to unload time rather than lastclose.
2005-11-08Correct a recursion on non-recursive mutex in drm_addmap from radeon'sEric Anholt
firstopen, by making drm_addmap require the drm device lock to be held. Also, make matching of kernel maps match linux by requiring shm matches to have the contains_lock flag set if the offset doesn't match.
2005-08-26- Don't try to allocate mappings of less than a PAGE_SIZE in MGA DMA code.Eric Anholt
- Comment out the "is this mapping/bufs in allocated AGP" bits in BSD because they break mga (which uses AGP allocation that doesn't track entries). It's not a security issue when we still have the related ioctls marked root-only. - Apply some power-of-two alignment restrictions to hopefully avoid some panicing in bad cases of drm_pci_alloc() on FreeBSD. - Add verbosity to some error handling that I found useful while debugging.
2005-08-26Previously, drm_get_resource_start() and drm_get_resource_len() wouldEric Anholt
allocate the resource RF_ACTIVE, pull out the appropriate value, and return it. However, allocating large framebuffers RF_ACTIVE would run the system out of KVA, and this also left open the possibility of the resource getting moved after getting the offset. Instead, when either of these are called, allocate the resource if it isn't allocated already (non-RF_ACTIVE) and store it in the DRM device, to be cleaned up on lastclose.
2005-08-26[1] Fix BSD DRM for the nonroot changes. [2] Don't attempt to acquire theEric Anholt
DMA lock in a non-DMA driver, as it will be uninitialized. Submitted by: [1] jkim (minor changes by me)
2005-08-05Rename the driver hooks in the DRM to something a little moreEric Anholt
understandable: preinit -> load postinit -> (removed) presetup -> firstopen postsetup -> (removed) open_helper -> open prerelease -> preclose free_filp_priv -> postclose pretakedown -> lastclose postcleanup -> unload release -> reclaim_buffers_locked version -> (removed) postinit and version were replaced with generic code in the Linux DRM (drivers now set their version numbers and description in the driver structure, like on BSD). postsetup wasn't used at all. Fixes the savage hooks for initializing and tearing down mappings at the right times. Testing involved at least starting X, running glxgears, killing glxgears, exiting X, and repeating. Tested on: FreeBSD (g200, g400, r200, r128) Linux (r200, savage4)
2005-06-29Get the BSD DRM compiling again after MGA and mapping changes. Note thatEric Anholt
i915 has been disabled for the moment, rather than working around linux-specific code in the shared dir.
2005-06-06Add a few more bits of Tonnerre's NetBSD port (Still need to deal with theEric Anholt
device attachment).
2005-04-26Convert BSD code to mostly use bus_dma, the dma abstraction for dealingEric Anholt
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
2005-04-24Fix a panic on startup with non-initmapping drivers by assigning theEric Anholt
drm_ioremap return value to the map handle again. Submitted by: Tor Egge, tegge at freebsd dot org
2005-04-16Use /*- to begin license blocks in BSD code to reduce diffs against FreeBSDEric Anholt
CVS.
2005-04-13Fix build on FreeBSD-current, thanks to jhb@.Eric Anholt
2005-02-07Restore a debugging message to DRM_DEBUG instead of DRM_ERROR.Eric Anholt
2005-02-05- Implement drm_initmap, and extend it with the resource number to helpEric Anholt
FreeBSD. Add drm_get_resource_{start|len} so linux-specific stuff doesn't need to be in shared code. - Fix mach64 build by using __DECONST to work around passing a const pointer to useracc, which is unfortunately not marked const. - Get rid of a lot of maplist code by not having dev->maplist be a pointer, and by sticking the link entries directly in drm_local_map_t rather than having a separate structure for the linked list. - Factor out map uninit and removal into its own routine, rather than duplicating in both drm_takedown() and drm_rmmap(). - Hook up more driver functions, and correct FreeBSD-specific bits of radeon_cp.c, making radeon work. - Baby steps towards using bus_space as we should.
2004-11-07Now that the memory debug code is gone, and all 3 BSDs have M_ZERO, stopEric Anholt
using drm_alloc/drm_free in the core and instead use plain malloc/free.
2004-11-06Commit first pieces of port to OpenBSD, done by Martin Lexa (martin atEric Anholt
martinlexa dot cz). Now that we've got porting for all three major BSDs (and the fourth being very similar to FreeBSD), move the mostly-duplication drm_os_* files into drmP.h. Remove some cruft from linux heritage and from pieces of the DRM that have since been removed. Note that things are still not quite working for even FreeBSD, but these are first steps at cleanup, and just a WIP checkpoint.
2004-11-06Commit WIP of BSD conversion to core model. Compiles for r128, radeon, butEric Anholt
doesn't run yet. Moves the ioctl definitions for these two drivers back to the shared code -- they aren't OS-specific.
2004-08-17Merged drmfntbl-0-0-1Dave Airlie
2004-05-09Warning fixes.Eric Anholt
2004-01-06Fix some misuse of NULL where 0 is intended.Eric Anholt
Obtained from: FreeBSD CVS
2003-12-16Don't ioremap the framebuffer area. The ioremapped area wasn't used byEric Anholt
anything, and took up valuable KVA. While I'm in the area, clean up BSD MTRR stuff some more. Suggested by: jonsmirl
2003-11-12Fix a locking nit, and add asserts in some things that should be calledEric Anholt
with locks held.
2003-11-06Return EBUSY when attempting to addmap a DRM_SHM area with a lock in it ifEric Anholt
dev->lock.hw_lock is already set. This fixes the case of two X Servers running on the same head on different VTs with interface 1.1, by making the 2nd head fail to inizialize like before.
2003-10-20Clean up BSD MTRR handling. The NetBSD code is untested, but it's my bestEric Anholt
shot.
2003-10-19- SMPng lock the DRM. This is only partial -- there are a few code pathsEric Anholt
used by root (the X Server) which are not locked. However, it should deal with lost-IRQ issues on -current which I think people have been experiencing but I am unable to reproduce (though I understand why they would occur, because of a bug of mine). Note that most of the locking (DRM_LOCK()/UNLOCK()) is all covered by Giant still, so it doesn't matter yet. - Remove locking on FreeBSD-stable and NetBSD. These are covered by the fact that there is no reentrancy of the kernel except by interrupts, which are locked using spldrm()/splx() instead.
2003-10-03Some code cleanups done while working on locking. Reduces always-trueEric Anholt
tests, excessive indenation, convoluted handling of errors, or code duplication.
2003-10-03Stylistic preparation for SMPng locking work: DRM_LOCK/DRM_UNLOCK have sideEric Anholt
effects, so make them look like functions (add parenthesis).
2003-10-02Mostly whitespace cleanups and style(9) fixes focused on "if(" -> "if ("Eric Anholt
Change some nearby memset()s to bzero()s or to calloc allocations to take advantage of M_ZERO). Reverse some error tests to reduce high levels of indentation. Move the sg_cleanup() call out of the maplist loop in DRM(takedown)-- I can't see any need for it to be inside.
2003-08-19Fix the debug build.Eric Anholt
2003-08-19- Remove $FreeBSD$ tags as they weren't too useful and merges are now beingEric Anholt
done through perforce. - Add copyright headers to drm_os_*bsd.h, still need to research the other copyright-less files better.
2003-05-16Support AGP bridges where the AGP aperture can't be accessed directly byMichel Daenzer
the CPU (David Mosberger, Benjamin Herrenschmidt, myself, Paul Mackerras, Jeff Wiedemeier)
2003-04-26Add PCI DMA memory functions and make addbufs_pci and associated code useEric Anholt
it. To do this we need to save the bus address along with the virtual address in the seglist. Also fix some error handling and a few bits of whitespace.
2003-04-25Merge from FreeBSD-current.Eric Anholt
2003-04-25Clean up the DRM_COPY_TO_USER()ing of DRM(infobufs), making it moreEric Anholt
legible.
2003-04-24Remove more gamma DMA infrastructure. Most of this code was copied straightEric Anholt
from linux, so it could be added back if some driver needed it in the future.
2003-04-24Move some common code from addbufs_<type> to addbufs. Make buf_alloc beEric Anholt
protected by the count_lock and make it non-atomic.
2003-04-24Remove DRM_DMA_HISTOGRAM and associated code.Eric Anholt
2003-03-29Add DRMFILE definitions and supply filp for BSD in theEric Anholt
post-drm-filp-0-1-branch world. The filp is a void * cast from the current pid. This is a temporary solution which maintains the status quo until a proper solution is implemented. What is really needed is a unique pointer per open, hopefully with a device private area. This can be done in FreeBSD for all entry points except mmap, but is difficult (sys/dev/streams/streams.c is an example). I have partially completed code for this but have not had time to debug, so this is a temporary fix.
2003-03-11Merge back from FreeBSD-current, adding FreeBSD ID tags to aid futureEric Anholt
merging. Also includes an update to radeon PCI IDs.
2003-02-21Merge from bsd-4-0-0-branch.Eric Anholt
2002-09-29Remove some noise being spit to the console.Eric Anholt
2002-07-05merged bsd-3-0-0-branchAlan Hourihane
2002-01-27Initial revisionDavid Dawes
generated by cgit v1.2.3 (git 2.25.1) at 2025-01-12 11:45:31 +0000
kwb">struct { unsigned int head; /* Position of head pointer */ unsigned int wrap; /* Primary DMA wrap count */ } drm_mga_age_t; typedef struct _drm_mga_sarea { /* The channel for communication of state information to the kernel * on firing a vertex dma buffer. */ drm_mga_context_regs_t context_state; drm_mga_server_regs_t server_state; drm_mga_texture_regs_t tex_state[2]; unsigned int warp_pipe; unsigned int dirty; unsigned int vertsize; /* The current cliprects, or a subset thereof. */ struct drm_clip_rect boxes[MGA_NR_SAREA_CLIPRECTS]; unsigned int nbox; /* Information about the most recently used 3d drawable. The * client fills in the req_* fields, the server fills in the * exported_ fields and puts the cliprects into boxes, above. * * The client clears the exported_drawable field before * clobbering the boxes data. */ unsigned int req_drawable; /* the X drawable id */ unsigned int req_draw_buffer; /* MGA_FRONT or MGA_BACK */ unsigned int exported_drawable; unsigned int exported_index; unsigned int exported_stamp; unsigned int exported_buffers; unsigned int exported_nfront; unsigned int exported_nback; int exported_back_x, exported_front_x, exported_w; int exported_back_y, exported_front_y, exported_h; struct drm_clip_rect exported_boxes[MGA_NR_SAREA_CLIPRECTS]; /* Counters for aging textures and for client-side throttling. */ unsigned int status[4]; unsigned int last_wrap; drm_mga_age_t last_frame; unsigned int last_enqueue; /* last time a buffer was enqueued */ unsigned int last_dispatch; /* age of the most recently dispatched buffer */ unsigned int last_quiescent; /* */ /* LRU lists for texture memory in agp space and on the card. */ struct drm_tex_region texList[MGA_NR_TEX_HEAPS][MGA_NR_TEX_REGIONS + 1]; unsigned int texAge[MGA_NR_TEX_HEAPS]; /* Mechanism to validate card state. */ int ctxOwner; } drm_mga_sarea_t; /* MGA specific ioctls * The device specific ioctl range is 0x40 to 0x79. */ #define DRM_MGA_INIT 0x00 #define DRM_MGA_FLUSH 0x01 #define DRM_MGA_RESET 0x02 #define DRM_MGA_SWAP 0x03 #define DRM_MGA_CLEAR 0x04 #define DRM_MGA_VERTEX 0x05 #define DRM_MGA_INDICES 0x06 #define DRM_MGA_ILOAD 0x07 #define DRM_MGA_BLIT 0x08 #define DRM_MGA_GETPARAM 0x09 /* 3.2: * ioctls for operating on fences. */ #define DRM_MGA_SET_FENCE 0x0a #define DRM_MGA_WAIT_FENCE 0x0b #define DRM_MGA_DMA_BOOTSTRAP 0x0c #define DRM_IOCTL_MGA_INIT DRM_IOW( DRM_COMMAND_BASE + DRM_MGA_INIT, drm_mga_init_t) #define DRM_IOCTL_MGA_FLUSH DRM_IOW( DRM_COMMAND_BASE + DRM_MGA_FLUSH, drm_lock_t) #define DRM_IOCTL_MGA_RESET DRM_IO( DRM_COMMAND_BASE + DRM_MGA_RESET) #define DRM_IOCTL_MGA_SWAP DRM_IO( DRM_COMMAND_BASE + DRM_MGA_SWAP) #define DRM_IOCTL_MGA_CLEAR DRM_IOW( DRM_COMMAND_BASE + DRM_MGA_CLEAR, drm_mga_clear_t) #define DRM_IOCTL_MGA_VERTEX DRM_IOW( DRM_COMMAND_BASE + DRM_MGA_VERTEX, drm_mga_vertex_t) #define DRM_IOCTL_MGA_INDICES DRM_IOW( DRM_COMMAND_BASE + DRM_MGA_INDICES, drm_mga_indices_t) #define DRM_IOCTL_MGA_ILOAD DRM_IOW( DRM_COMMAND_BASE + DRM_MGA_ILOAD, drm_mga_iload_t) #define DRM_IOCTL_MGA_BLIT DRM_IOW( DRM_COMMAND_BASE + DRM_MGA_BLIT, drm_mga_blit_t) #define DRM_IOCTL_MGA_GETPARAM DRM_IOWR(DRM_COMMAND_BASE + DRM_MGA_GETPARAM, drm_mga_getparam_t) #define DRM_IOCTL_MGA_SET_FENCE DRM_IOW( DRM_COMMAND_BASE + DRM_MGA_SET_FENCE, uint32_t) #define DRM_IOCTL_MGA_WAIT_FENCE DRM_IOWR(DRM_COMMAND_BASE + DRM_MGA_WAIT_FENCE, uint32_t) #define DRM_IOCTL_MGA_DMA_BOOTSTRAP DRM_IOWR(DRM_COMMAND_BASE + DRM_MGA_DMA_BOOTSTRAP, drm_mga_dma_bootstrap_t) typedef struct _drm_mga_warp_index { int installed; unsigned long phys_addr; int size; } drm_mga_warp_index_t; typedef struct drm_mga_init { enum { MGA_INIT_DMA = 0x01, MGA_CLEANUP_DMA = 0x02 } func; unsigned long sarea_priv_offset; int chipset; int sgram; unsigned int maccess; unsigned int fb_cpp; unsigned int front_offset, front_pitch; unsigned int back_offset, back_pitch; unsigned int depth_cpp; unsigned int depth_offset, depth_pitch; unsigned int texture_offset[MGA_NR_TEX_HEAPS]; unsigned int texture_size[MGA_NR_TEX_HEAPS]; unsigned long fb_offset; unsigned long mmio_offset; unsigned long status_offset; unsigned long warp_offset; unsigned long primary_offset; unsigned long buffers_offset; } drm_mga_init_t; typedef struct drm_mga_dma_bootstrap { /** * \name AGP texture region * * On return from the DRM_MGA_DMA_BOOTSTRAP ioctl, these fields will * be filled in with the actual AGP texture settings. * * \warning * If these fields are non-zero, but dma_mga_dma_bootstrap::agp_mode * is zero, it means that PCI memory (most likely through the use of * an IOMMU) is being used for "AGP" textures. */ /*@{*/ unsigned long texture_handle; /**< Handle used to map AGP textures. */ uint32_t texture_size; /**< Size of the AGP texture region. */ /*@}*/ /** * Requested size of the primary DMA region. * * On return from the DRM_MGA_DMA_BOOTSTRAP ioctl, this field will be * filled in with the actual AGP mode. If AGP was not available */ uint32_t primary_size; /** * Requested number of secondary DMA buffers. * * On return from the DRM_MGA_DMA_BOOTSTRAP ioctl, this field will be * filled in with the actual number of secondary DMA buffers * allocated. Particularly when PCI DMA is used, this may be * (subtantially) less than the number requested. */ uint32_t secondary_bin_count; /** * Requested size of each secondary DMA buffer. * * While the kernel \b is free to reduce * dma_mga_dma_bootstrap::secondary_bin_count, it is \b not allowed * to reduce dma_mga_dma_bootstrap::secondary_bin_size. */ uint32_t secondary_bin_size; /** * Bit-wise mask of AGPSTAT2_* values. Currently only \c AGPSTAT2_1X, * \c AGPSTAT2_2X, and \c AGPSTAT2_4X are supported. If this value is * zero, it means that PCI DMA should be used, even if AGP is * possible. * * On return from the DRM_MGA_DMA_BOOTSTRAP ioctl, this field will be * filled in with the actual AGP mode. If AGP was not available * (i.e., PCI DMA was used), this value will be zero. */ uint32_t agp_mode; /** * Desired AGP GART size, measured in megabytes. */ uint8_t agp_size; } drm_mga_dma_bootstrap_t; typedef struct drm_mga_clear { unsigned int flags; unsigned int clear_color; unsigned int clear_depth; unsigned int color_mask; unsigned int depth_mask; } drm_mga_clear_t; typedef struct drm_mga_vertex { int idx; /* buffer to queue */ int used; /* bytes in use */ int discard; /* client finished with buffer? */ } drm_mga_vertex_t; typedef struct drm_mga_indices { int idx; /* buffer to queue */ unsigned int start; unsigned int end; int discard; /* client finished with buffer? */ } drm_mga_indices_t; typedef struct drm_mga_iload { int idx; unsigned int dstorg; unsigned int length; } drm_mga_iload_t; typedef struct _drm_mga_blit { unsigned int planemask; unsigned int srcorg; unsigned int dstorg; int src_pitch, dst_pitch; int delta_sx, delta_sy; int delta_dx, delta_dy; int height, ydir; /* flip image vertically */ int source_pitch, dest_pitch; } drm_mga_blit_t; /* 3.1: An ioctl to get parameters that aren't available to the 3d * client any other way. */ #define MGA_PARAM_IRQ_NR 1 /* 3.2: Query the actual card type. The DDX only distinguishes between * G200 chips and non-G200 chips, which it calls G400. It turns out that * there are some very sublte differences between the G4x0 chips and the G550 * chips. Using this parameter query, a client-side driver can detect the * difference between a G4x0 and a G550. */ #define MGA_PARAM_CARD_TYPE 2 typedef struct drm_mga_getparam { int param; void __user *value; } drm_mga_getparam_t; #endif
generated by cgit v1.2.3 (git 2.25.1) at 2025-01-12 11:45:31 +0000