/* -*- mode: C; c-file-style: "k&r"; tab-width 4; indent-tabs-mode: t; -*- */ /* * Copyright (C) 2011 Texas Instruments, Inc * * Permission is hereby granted, free of charge, to any person obtaining a * copy of this software and associated documentation files (the "Software"), * to deal in the Software without restriction, including without limitation * the rights to use, copy, modify, merge, publish, distribute, sublicense, * and/or sell copies of the Software, and to permit persons to whom the * Software is furnished to do so, subject to the following conditions: * * The above copyright notice and this permission notice (including the next * paragraph) shall be included in all copies or substantial portions of the * Software. * * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE * SOFTWARE. * * Authors: * Rob Clark <rob@ti.com> */ #ifdef HAVE_CONFIG_H #include "config.h" #endif #include <stdlib.h> #include <linux/stddef.h> #include <linux/types.h> #include <errno.h> #include <sys/mman.h> #include <fcntl.h> #include <unistd.h> #include <pthread.h> #include <libdrm_macros.h> #include <xf86drm.h> #include <xf86atomic.h> #include "omap_drm.h" #include "omap_drmif.h" #define __round_mask(x, y) ((__typeof__(x))((y)-1)) #define round_up(x, y) ((((x)-1) | __round_mask(x, y))+1) #define PAGE_SIZE 4096 static pthread_mutex_t table_lock = PTHREAD_MUTEX_INITIALIZER; static void * dev_table; struct omap_device { int fd; atomic_t refcnt; /* The handle_table is used to track GEM bo handles associated w/ * this fd. This is needed, in particular, when importing * dmabuf's because we don't want multiple 'struct omap_bo's * floating around with the same handle. Otherwise, when the * first one is omap_bo_del()'d the handle becomes no longer * valid, and the remaining 'struct omap_bo's are left pointing * to an invalid handle (and possible a GEM bo that is already * free'd). */ void *handle_table; }; /* a GEM buffer object allocated from the DRM device */ struct omap_bo { struct omap_device *dev; void *map; /* userspace mmap'ing (if there is one) */ uint32_t size; uint32_t handle; uint32_t name; /* flink global handle (DRI2 name) */ uint64_t offset; /* offset to mmap() */ int fd; /* dmabuf handle */ atomic_t refcnt; }; static struct omap_device * omap_device_new_impl(int fd) { struct omap_device *dev = calloc(sizeof(*dev), 1); if (!dev) return NULL; dev->fd = fd; atomic_set(&dev->refcnt, 1); dev->handle_table = drmHashCreate(); return dev; } struct omap_device * omap_device_new(int fd) { struct omap_device *dev = NULL; pthread_mutex_lock(&table_lock); if (!dev_table) dev_table = drmHashCreate(); if (drmHashLookup(dev_table, fd, (void **)&dev)) { /* not found, create new device */ dev = omap_device_new_impl(fd); drmHashInsert(dev_table, fd, dev); } else { /* found, just incr refcnt */ dev = omap_device_ref(dev); } pthread_mutex_unlock(&table_lock); return dev; } struct omap_device * omap_device_ref(struct omap_device *dev) { atomic_inc(&dev->refcnt); return dev; } void omap_device_del(struct omap_device *dev) { if (!atomic_dec_and_test(&dev->refcnt)) return; pthread_mutex_lock(&table_lock); drmHashDestroy(dev->handle_table); drmHashDelete(dev_table, dev->fd); pthread_mutex_unlock(&table_lock); free(dev); } int omap_get_param(struct omap_device *dev, uint64_t param, uint64_t *value) { struct drm_omap_param req = { .param = param, }; int ret; ret = drmCommandWriteRead(dev->fd, DRM_OMAP_GET_PARAM, &req, sizeof(req)); if (ret) { return ret; } *value = req.value; return 0; } int omap_set_param(struct omap_device *dev, uint64_t param, uint64_t value) { struct drm_omap_param req = { .param = param, .value = value, }; return drmCommandWrite(dev->fd, DRM_OMAP_SET_PARAM, &req, sizeof(req)); } /* lookup a buffer from it's handle, call w/ table_lock held: */ static struct omap_bo * lookup_bo(struct omap_device *dev, uint32_t handle) { struct omap_bo *bo = NULL; if (!drmHashLookup(dev->handle_table, handle, (void **)&bo)) { /* found, incr refcnt and return: */ bo = omap_bo_ref(bo); } return bo; } /* allocate a new buffer object, call w/ table_lock held */ static struct omap_bo * bo_from_handle(struct omap_device *dev, uint32_t handle) { struct omap_bo *bo = calloc(sizeof(*bo), 1); if (!bo) { struct drm_gem_close req = { .handle = handle, }; drmIoctl(dev->fd, DRM_IOCTL_GEM_CLOSE, &req); return NULL; } bo->dev = omap_device_ref(dev); bo->handle = handle; bo->fd = -1; atomic_set(&bo->refcnt, 1); /* add ourselves to the handle table: */ drmHashInsert(dev->handle_table, handle, bo); return bo; } /* allocate a new buffer object */ static struct omap_bo * omap_bo_new_impl(struct omap_device *dev, union omap_gem_size size, uint32_t flags) { struct omap_bo *bo = NULL; struct drm_omap_gem_new req = { .size = size, .flags = flags, }; if (size.bytes == 0) { goto fail; } if (drmCommandWriteRead(dev->fd, DRM_OMAP_GEM_NEW, &req, sizeof(req))) { goto fail; } pthread_mutex_lock(&table_lock); bo = bo_from_handle(dev, req.handle); pthread_mutex_unlock(&table_lock); if (flags & OMAP_BO_TILED) { bo->size = round_up(size.tiled.width, PAGE_SIZE) * size.tiled.height; } else { bo->size = size.bytes; } return bo; fail: free(bo); return NULL; } /* allocate a new (un-tiled) buffer object */ struct omap_bo * omap_bo_new(struct omap_device *dev, uint32_t size, uint32_t flags) { union omap_gem_size gsize = { .bytes = size, }; if (flags & OMAP_BO_TILED) { return NULL; } return omap_bo_new_impl(dev, gsize, flags); } /* allocate a new buffer object */ struct omap_bo * omap_bo_new_tiled(struct omap_device *dev, uint32_t width, uint32_t height, uint32_t flags) { union omap_gem_size gsize = { .tiled = { .width = width, .height = height, }, }; if (!(flags & OMAP_BO_TILED)) { return NULL; } return omap_bo_new_impl(dev, gsize, flags); } struct omap_bo *omap_bo_ref(struct omap_bo *bo) { atomic_inc(&bo->refcnt); return bo; } /* get buffer info */ static int get_buffer_info(struct omap_bo *bo) { struct drm_omap_gem_info req = { .handle = bo->handle, }; int ret = drmCommandWriteRead(bo->dev->fd, DRM_OMAP_GEM_INFO, &req, sizeof(req)); if (ret) { return ret; } /* really all we need for now is mmap offset */ bo->offset = req.offset; bo->size = req.size; return 0; } /* import a buffer object from DRI2 name */ struct omap_bo * omap_bo_from_name(struct omap_device *dev, uint32_t name) { struct omap_bo *bo = NULL; struct drm_gem_open req = { .name = name, }; pthread_mutex_lock(&table_lock); if (drmIoctl(dev->fd, DRM_IOCTL_GEM_OPEN, &req)) { goto fail; } bo = lookup_bo(dev, req.handle); if (!bo) { bo = bo_from_handle(dev, req.handle); bo->name = name; } pthread_mutex_unlock(&table_lock); return bo; fail: pthread_mutex_unlock(&table_lock); free(bo); return NULL; } /* import a buffer from dmabuf fd, does not take ownership of the * fd so caller should close() the fd when it is otherwise done * with it (even if it is still using the 'struct omap_bo *') */ struct omap_bo * omap_bo_from_dmabuf(struct omap_device *dev, int fd) { struct omap_bo *bo = NULL; struct drm_prime_handle req = { .fd = fd, }; int ret; pthread_mutex_lock(&table_lock); ret = drmIoctl(dev->fd, DRM_IOCTL_PRIME_FD_TO_HANDLE, &req); if (ret) { goto fail; } bo = lookup_bo(dev, req.handle); if (!bo) { bo = bo_from_handle(dev, req.handle); } pthread_mutex_unlock(&table_lock); return bo; fail: pthread_mutex_unlock(&table_lock); free(bo); return NULL; } /* destroy a buffer object */ void omap_bo_del(struct omap_bo *bo) { if (!bo) { return; } if (!atomic_dec_and_test(&bo->refcnt)) return; if (bo->map) { munmap(bo->map, bo->size); } if (bo->fd >= 0) { close(bo->fd); } if (bo->handle) { struct drm_gem_close req = { .handle = bo->handle, }; pthread_mutex_lock(&table_lock); drmHashDelete(bo->dev->handle_table, bo->handle); drmIoctl(bo->dev->fd, DRM_IOCTL_GEM_CLOSE, &req); pthread_mutex_unlock(&table_lock); } omap_device_del(bo->dev); free(bo); } /* get the global flink/DRI2 buffer name */ int omap_bo_get_name(struct omap_bo *bo, uint32_t *name) { if (!bo->name) { struct drm_gem_flink req = { .handle = bo->handle, }; int ret; ret = drmIoctl(bo->dev->fd, DRM_IOCTL_GEM_FLINK, &req); if (ret) { return ret; } bo->name = req.name; } *name = bo->name; return 0; } uint32_t omap_bo_handle(struct omap_bo *bo) { return bo->handle; } /* caller owns the dmabuf fd that is returned and is responsible * to close() it when done */ int omap_bo_dmabuf(struct omap_bo *bo) { if (bo->fd < 0) { struct drm_prime_handle req = { .handle = bo->handle, .flags = DRM_CLOEXEC, }; int ret; ret = drmIoctl(bo->dev->fd, DRM_IOCTL_PRIME_HANDLE_TO_FD, &req); if (ret) { return ret; } bo->fd = req.fd; } return dup(bo->fd); } uint32_t omap_bo_size(struct omap_bo *bo) { if (!bo->size) { get_buffer_info(bo); } return bo->size; } void *omap_bo_map(struct omap_bo *bo) { if (!bo->map) { if (!bo->offset) { get_buffer_info(bo); } bo->map = mmap(0, bo->size, PROT_READ | PROT_WRITE, MAP_SHARED, bo->dev->fd, bo->offset); if (bo->map == MAP_FAILED) { bo->map = NULL; } } return bo->map; } int omap_bo_cpu_prep(struct omap_bo *bo, enum omap_gem_op op) { struct drm_omap_gem_cpu_prep req = { .handle = bo->handle, .op = op, }; return drmCommandWrite(bo->dev->fd, DRM_OMAP_GEM_CPU_PREP, &req, sizeof(req)); } int omap_bo_cpu_fini(struct omap_bo *bo, enum omap_gem_op op) { struct drm_omap_gem_cpu_fini req = { .handle = bo->handle, .op = op, .nregions = 0, }; return drmCommandWrite(bo->dev->fd, DRM_OMAP_GEM_CPU_FINI, &req, sizeof(req)); }