diff options
author | Eric Anholt <[email protected]> | 2017-02-02 16:24:13 -0800 |
---|---|---|
committer | Eric Anholt <[email protected]> | 2017-10-10 11:42:04 -0700 |
commit | 45bb8f2957108056de10721d50259e2f8c5a4ddc (patch) | |
tree | 37abe7d19c4e9d20990e1e42b2eeb61f00eab09e /src/gallium/drivers/vc5/vc5_bufmgr.c | |
parent | ade416d02369cc0942d53ad3cce601d66344f9c3 (diff) |
broadcom: Add V3D 3.3 gallium driver called "vc5", for BCM7268.
V3D 3.3 is a continuation of the 3D implementation in VC4 (v2.1 and v2.6).
V3D 3.3 introduces an MMU (no more CMA allocations) and support for
GLES3.1. This driver is not currently conformant, though that will be a
target as soon as possible.
V3D 3.x parts use a new texture tiling layout common across many Broadcom
graphics parts including and the HVS scanout engine. It also massively
changes the QPU instructions, introducing a common physical register file
(no more A/B split) and half-float instructions, while removing the 4x8
unorm instructions in favor of half-float for talking to fixed function
interfaces. Because so much has changed, vc5 is implemented in a separate
gallium driver, using only the XML code-generation support from vc4.
v2: Fix tile layout for 64bpp textures. Fix texture swizzling for 32-bit
returns. Fix up a bit of MRT setup. Sync the simulator to kernel
behavior a bit more. Improve uniform debugging code. Rebase on
QIR->VIR rename. Move texture state mostly to the CSOs. Improve
cache flushing on the simulator. Fix program deletion
use-after-frees.
Acked-by: Dave Airlie <[email protected]> (uabi plan)
Acked-by: Daniel Vetter <[email protected]> (uabi plan)
Diffstat (limited to 'src/gallium/drivers/vc5/vc5_bufmgr.c')
-rw-r--r-- | src/gallium/drivers/vc5/vc5_bufmgr.c | 580 |
1 files changed, 580 insertions, 0 deletions
diff --git a/src/gallium/drivers/vc5/vc5_bufmgr.c b/src/gallium/drivers/vc5/vc5_bufmgr.c new file mode 100644 index 00000000000..c6c06dcfda7 --- /dev/null +++ b/src/gallium/drivers/vc5/vc5_bufmgr.c @@ -0,0 +1,580 @@ +/* + * Copyright © 2014-2017 Broadcom + * + * 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. + */ + +#include <errno.h> +#include <err.h> +#include <sys/mman.h> +#include <fcntl.h> +#include <xf86drm.h> +#include <xf86drmMode.h> + +#include "util/u_hash_table.h" +#include "util/u_memory.h" +#include "util/ralloc.h" + +#include "vc5_context.h" +#include "vc5_screen.h" + +#ifdef HAVE_VALGRIND +#include <valgrind.h> +#include <memcheck.h> +#define VG(x) x +#else +#define VG(x) +#endif + +static bool dump_stats = false; + +static void +vc5_bo_cache_free_all(struct vc5_bo_cache *cache); + +static void +vc5_bo_dump_stats(struct vc5_screen *screen) +{ + struct vc5_bo_cache *cache = &screen->bo_cache; + + fprintf(stderr, " BOs allocated: %d\n", screen->bo_count); + fprintf(stderr, " BOs size: %dkb\n", screen->bo_size / 1024); + fprintf(stderr, " BOs cached: %d\n", cache->bo_count); + fprintf(stderr, " BOs cached size: %dkb\n", cache->bo_size / 1024); + + if (!list_empty(&cache->time_list)) { + struct vc5_bo *first = LIST_ENTRY(struct vc5_bo, + cache->time_list.next, + time_list); + struct vc5_bo *last = LIST_ENTRY(struct vc5_bo, + cache->time_list.prev, + time_list); + + fprintf(stderr, " oldest cache time: %ld\n", + (long)first->free_time); + fprintf(stderr, " newest cache time: %ld\n", + (long)last->free_time); + + struct timespec time; + clock_gettime(CLOCK_MONOTONIC, &time); + fprintf(stderr, " now: %ld\n", + time.tv_sec); + } +} + +static void +vc5_bo_remove_from_cache(struct vc5_bo_cache *cache, struct vc5_bo *bo) +{ + list_del(&bo->time_list); + list_del(&bo->size_list); + cache->bo_count--; + cache->bo_size -= bo->size; +} + +static struct vc5_bo * +vc5_bo_from_cache(struct vc5_screen *screen, uint32_t size, const char *name) +{ + struct vc5_bo_cache *cache = &screen->bo_cache; + uint32_t page_index = size / 4096 - 1; + + if (cache->size_list_size <= page_index) + return NULL; + + struct vc5_bo *bo = NULL; + mtx_lock(&cache->lock); + if (!list_empty(&cache->size_list[page_index])) { + bo = LIST_ENTRY(struct vc5_bo, cache->size_list[page_index].next, + size_list); + + /* Check that the BO has gone idle. If not, then we want to + * allocate something new instead, since we assume that the + * user will proceed to CPU map it and fill it with stuff. + */ + if (!vc5_bo_wait(bo, 0, NULL)) { + mtx_unlock(&cache->lock); + return NULL; + } + + pipe_reference_init(&bo->reference, 1); + vc5_bo_remove_from_cache(cache, bo); + + bo->name = name; + } + mtx_unlock(&cache->lock); + return bo; +} + +struct vc5_bo * +vc5_bo_alloc(struct vc5_screen *screen, uint32_t size, const char *name) +{ + struct vc5_bo *bo; + int ret; + + size = align(size, 4096); + + bo = vc5_bo_from_cache(screen, size, name); + if (bo) { + if (dump_stats) { + fprintf(stderr, "Allocated %s %dkb from cache:\n", + name, size / 1024); + vc5_bo_dump_stats(screen); + } + return bo; + } + + bo = CALLOC_STRUCT(vc5_bo); + if (!bo) + return NULL; + + pipe_reference_init(&bo->reference, 1); + bo->screen = screen; + bo->size = size; + bo->name = name; + bo->private = true; + + retry: + ; + + bool cleared_and_retried = false; + struct drm_vc5_create_bo create = { + .size = size + }; + + ret = vc5_ioctl(screen->fd, DRM_IOCTL_VC5_CREATE_BO, &create); + bo->handle = create.handle; + bo->offset = create.offset; + + if (ret != 0) { + if (!list_empty(&screen->bo_cache.time_list) && + !cleared_and_retried) { + cleared_and_retried = true; + vc5_bo_cache_free_all(&screen->bo_cache); + goto retry; + } + + free(bo); + return NULL; + } + + screen->bo_count++; + screen->bo_size += bo->size; + if (dump_stats) { + fprintf(stderr, "Allocated %s %dkb:\n", name, size / 1024); + vc5_bo_dump_stats(screen); + } + + return bo; +} + +void +vc5_bo_last_unreference(struct vc5_bo *bo) +{ + struct vc5_screen *screen = bo->screen; + + struct timespec time; + clock_gettime(CLOCK_MONOTONIC, &time); + mtx_lock(&screen->bo_cache.lock); + vc5_bo_last_unreference_locked_timed(bo, time.tv_sec); + mtx_unlock(&screen->bo_cache.lock); +} + +static void +vc5_bo_free(struct vc5_bo *bo) +{ + struct vc5_screen *screen = bo->screen; + + if (bo->map) { + if (using_vc5_simulator && bo->name && + strcmp(bo->name, "winsys") == 0) { + free(bo->map); + } else { + munmap(bo->map, bo->size); + VG(VALGRIND_FREELIKE_BLOCK(bo->map, 0)); + } + } + + struct drm_gem_close c; + memset(&c, 0, sizeof(c)); + c.handle = bo->handle; + int ret = vc5_ioctl(screen->fd, DRM_IOCTL_GEM_CLOSE, &c); + if (ret != 0) + fprintf(stderr, "close object %d: %s\n", bo->handle, strerror(errno)); + + screen->bo_count--; + screen->bo_size -= bo->size; + + if (dump_stats) { + fprintf(stderr, "Freed %s%s%dkb:\n", + bo->name ? bo->name : "", + bo->name ? " " : "", + bo->size / 1024); + vc5_bo_dump_stats(screen); + } + + free(bo); +} + +static void +free_stale_bos(struct vc5_screen *screen, time_t time) +{ + struct vc5_bo_cache *cache = &screen->bo_cache; + bool freed_any = false; + + list_for_each_entry_safe(struct vc5_bo, bo, &cache->time_list, + time_list) { + if (dump_stats && !freed_any) { + fprintf(stderr, "Freeing stale BOs:\n"); + vc5_bo_dump_stats(screen); + freed_any = true; + } + + /* If it's more than a second old, free it. */ + if (time - bo->free_time > 2) { + vc5_bo_remove_from_cache(cache, bo); + vc5_bo_free(bo); + } else { + break; + } + } + + if (dump_stats && freed_any) { + fprintf(stderr, "Freed stale BOs:\n"); + vc5_bo_dump_stats(screen); + } +} + +static void +vc5_bo_cache_free_all(struct vc5_bo_cache *cache) +{ + mtx_lock(&cache->lock); + list_for_each_entry_safe(struct vc5_bo, bo, &cache->time_list, + time_list) { + vc5_bo_remove_from_cache(cache, bo); + vc5_bo_free(bo); + } + mtx_unlock(&cache->lock); +} + +void +vc5_bo_last_unreference_locked_timed(struct vc5_bo *bo, time_t time) +{ + struct vc5_screen *screen = bo->screen; + struct vc5_bo_cache *cache = &screen->bo_cache; + uint32_t page_index = bo->size / 4096 - 1; + + if (!bo->private) { + vc5_bo_free(bo); + return; + } + + if (cache->size_list_size <= page_index) { + struct list_head *new_list = + ralloc_array(screen, struct list_head, page_index + 1); + + /* Move old list contents over (since the array has moved, and + * therefore the pointers to the list heads have to change). + */ + for (int i = 0; i < cache->size_list_size; i++) { + struct list_head *old_head = &cache->size_list[i]; + if (list_empty(old_head)) + list_inithead(&new_list[i]); + else { + new_list[i].next = old_head->next; + new_list[i].prev = old_head->prev; + new_list[i].next->prev = &new_list[i]; + new_list[i].prev->next = &new_list[i]; + } + } + for (int i = cache->size_list_size; i < page_index + 1; i++) + list_inithead(&new_list[i]); + + cache->size_list = new_list; + cache->size_list_size = page_index + 1; + } + + bo->free_time = time; + list_addtail(&bo->size_list, &cache->size_list[page_index]); + list_addtail(&bo->time_list, &cache->time_list); + cache->bo_count++; + cache->bo_size += bo->size; + if (dump_stats) { + fprintf(stderr, "Freed %s %dkb to cache:\n", + bo->name, bo->size / 1024); + vc5_bo_dump_stats(screen); + } + bo->name = NULL; + + free_stale_bos(screen, time); +} + +static struct vc5_bo * +vc5_bo_open_handle(struct vc5_screen *screen, + uint32_t winsys_stride, + uint32_t handle, uint32_t size) +{ + struct vc5_bo *bo; + + assert(size); + + mtx_lock(&screen->bo_handles_mutex); + + bo = util_hash_table_get(screen->bo_handles, (void*)(uintptr_t)handle); + if (bo) { + pipe_reference(NULL, &bo->reference); + goto done; + } + + bo = CALLOC_STRUCT(vc5_bo); + pipe_reference_init(&bo->reference, 1); + bo->screen = screen; + bo->handle = handle; + bo->size = size; + bo->name = "winsys"; + bo->private = false; + +#ifdef USE_VC5_SIMULATOR + vc5_simulator_open_from_handle(screen->fd, winsys_stride, + bo->handle, bo->size); + bo->map = malloc(bo->size); +#endif + + util_hash_table_set(screen->bo_handles, (void *)(uintptr_t)handle, bo); + +done: + mtx_unlock(&screen->bo_handles_mutex); + return bo; +} + +struct vc5_bo * +vc5_bo_open_name(struct vc5_screen *screen, uint32_t name, + uint32_t winsys_stride) +{ + struct drm_gem_open o = { + .name = name + }; + int ret = vc5_ioctl(screen->fd, DRM_IOCTL_GEM_OPEN, &o); + if (ret) { + fprintf(stderr, "Failed to open bo %d: %s\n", + name, strerror(errno)); + return NULL; + } + + return vc5_bo_open_handle(screen, winsys_stride, o.handle, o.size); +} + +struct vc5_bo * +vc5_bo_open_dmabuf(struct vc5_screen *screen, int fd, uint32_t winsys_stride) +{ + uint32_t handle; + int ret = drmPrimeFDToHandle(screen->fd, fd, &handle); + int size; + if (ret) { + fprintf(stderr, "Failed to get vc5 handle for dmabuf %d\n", fd); + return NULL; + } + + /* Determine the size of the bo we were handed. */ + size = lseek(fd, 0, SEEK_END); + if (size == -1) { + fprintf(stderr, "Couldn't get size of dmabuf fd %d.\n", fd); + return NULL; + } + + return vc5_bo_open_handle(screen, winsys_stride, handle, size); +} + +int +vc5_bo_get_dmabuf(struct vc5_bo *bo) +{ + int fd; + int ret = drmPrimeHandleToFD(bo->screen->fd, bo->handle, + O_CLOEXEC, &fd); + if (ret != 0) { + fprintf(stderr, "Failed to export gem bo %d to dmabuf\n", + bo->handle); + return -1; + } + + mtx_lock(&bo->screen->bo_handles_mutex); + bo->private = false; + util_hash_table_set(bo->screen->bo_handles, (void *)(uintptr_t)bo->handle, bo); + mtx_unlock(&bo->screen->bo_handles_mutex); + + return fd; +} + +bool +vc5_bo_flink(struct vc5_bo *bo, uint32_t *name) +{ + struct drm_gem_flink flink = { + .handle = bo->handle, + }; + int ret = vc5_ioctl(bo->screen->fd, DRM_IOCTL_GEM_FLINK, &flink); + if (ret) { + fprintf(stderr, "Failed to flink bo %d: %s\n", + bo->handle, strerror(errno)); + free(bo); + return false; + } + + bo->private = false; + *name = flink.name; + + return true; +} + +static int vc5_wait_seqno_ioctl(int fd, uint64_t seqno, uint64_t timeout_ns) +{ + struct drm_vc5_wait_seqno wait = { + .seqno = seqno, + .timeout_ns = timeout_ns, + }; + int ret = vc5_ioctl(fd, DRM_IOCTL_VC5_WAIT_SEQNO, &wait); + if (ret == -1) + return -errno; + else + return 0; + +} + +bool +vc5_wait_seqno(struct vc5_screen *screen, uint64_t seqno, uint64_t timeout_ns, + const char *reason) +{ + if (screen->finished_seqno >= seqno) + return true; + + if (unlikely(V3D_DEBUG & V3D_DEBUG_PERF) && timeout_ns && reason) { + if (vc5_wait_seqno_ioctl(screen->fd, seqno, 0) == -ETIME) { + fprintf(stderr, "Blocking on seqno %lld for %s\n", + (long long)seqno, reason); + } + } + + int ret = vc5_wait_seqno_ioctl(screen->fd, seqno, timeout_ns); + if (ret) { + if (ret != -ETIME) { + fprintf(stderr, "wait failed: %d\n", ret); + abort(); + } + + return false; + } + + screen->finished_seqno = seqno; + return true; +} + +static int vc5_wait_bo_ioctl(int fd, uint32_t handle, uint64_t timeout_ns) +{ + struct drm_vc5_wait_bo wait = { + .handle = handle, + .timeout_ns = timeout_ns, + }; + int ret = vc5_ioctl(fd, DRM_IOCTL_VC5_WAIT_BO, &wait); + if (ret == -1) + return -errno; + else + return 0; + +} + +bool +vc5_bo_wait(struct vc5_bo *bo, uint64_t timeout_ns, const char *reason) +{ + struct vc5_screen *screen = bo->screen; + + if (unlikely(V3D_DEBUG & V3D_DEBUG_PERF) && timeout_ns && reason) { + if (vc5_wait_bo_ioctl(screen->fd, bo->handle, 0) == -ETIME) { + fprintf(stderr, "Blocking on %s BO for %s\n", + bo->name, reason); + } + } + + int ret = vc5_wait_bo_ioctl(screen->fd, bo->handle, timeout_ns); + if (ret) { + if (ret != -ETIME) { + fprintf(stderr, "wait failed: %d\n", ret); + abort(); + } + + return false; + } + + return true; +} + +void * +vc5_bo_map_unsynchronized(struct vc5_bo *bo) +{ + uint64_t offset; + int ret; + + if (bo->map) + return bo->map; + + struct drm_vc5_mmap_bo map; + memset(&map, 0, sizeof(map)); + map.handle = bo->handle; + ret = vc5_ioctl(bo->screen->fd, DRM_IOCTL_VC5_MMAP_BO, &map); + offset = map.offset; + if (ret != 0) { + fprintf(stderr, "map ioctl failure\n"); + abort(); + } + + bo->map = mmap(NULL, bo->size, PROT_READ | PROT_WRITE, MAP_SHARED, + bo->screen->fd, offset); + if (bo->map == MAP_FAILED) { + fprintf(stderr, "mmap of bo %d (offset 0x%016llx, size %d) failed\n", + bo->handle, (long long)offset, bo->size); + abort(); + } + VG(VALGRIND_MALLOCLIKE_BLOCK(bo->map, bo->size, 0, false)); + + return bo->map; +} + +void * +vc5_bo_map(struct vc5_bo *bo) +{ + void *map = vc5_bo_map_unsynchronized(bo); + + bool ok = vc5_bo_wait(bo, PIPE_TIMEOUT_INFINITE, "bo map"); + if (!ok) { + fprintf(stderr, "BO wait for map failed\n"); + abort(); + } + + return map; +} + +void +vc5_bufmgr_destroy(struct pipe_screen *pscreen) +{ + struct vc5_screen *screen = vc5_screen(pscreen); + struct vc5_bo_cache *cache = &screen->bo_cache; + + vc5_bo_cache_free_all(cache); + + if (dump_stats) { + fprintf(stderr, "BO stats after screen destroy:\n"); + vc5_bo_dump_stats(screen); + } +} |