From 965cf68e8797932e9cd49238a6dd39423ac9b256 Mon Sep 17 00:00:00 2001 From: Francisco Jerez Date: Sat, 6 Mar 2010 13:42:45 +0100 Subject: drm/nouveau: Never evict VRAM buffers to system. VRAM->system is a synchronous operation: it involves scheduling a VRAM->TT DMA transfer and stalling the CPU until it's finished so that we can unbind the new memory from the translation tables. VRAM->TT can always be performed asynchronously, even if TT is already full and we have to move something out of it. Additionally, allowing VRAM->system behaves badly under heavy memory pressure because once we run out of TT, stuff starts to be moved back and forth between VRAM and system, and the TT contents are hardly renewed. Signed-off-by: Francisco Jerez Signed-off-by: Ben Skeggs --- drivers/gpu/drm/nouveau/nouveau_bo.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'drivers/gpu/drm/nouveau/nouveau_bo.c') diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c index 028719fddf7..026612471c9 100644 --- a/drivers/gpu/drm/nouveau/nouveau_bo.c +++ b/drivers/gpu/drm/nouveau/nouveau_bo.c @@ -439,8 +439,7 @@ nouveau_bo_evict_flags(struct ttm_buffer_object *bo, struct ttm_placement *pl) switch (bo->mem.mem_type) { case TTM_PL_VRAM: - nouveau_bo_placement_set(nvbo, TTM_PL_FLAG_TT | - TTM_PL_FLAG_SYSTEM); + nouveau_bo_placement_set(nvbo, TTM_PL_FLAG_TT); break; default: nouveau_bo_placement_set(nvbo, TTM_PL_FLAG_SYSTEM); -- cgit From 5a0e3ad6af8660be21ca98a971cd00f331318c05 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Wed, 24 Mar 2010 17:04:11 +0900 Subject: include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h percpu.h is included by sched.h and module.h and thus ends up being included when building most .c files. percpu.h includes slab.h which in turn includes gfp.h making everything defined by the two files universally available and complicating inclusion dependencies. percpu.h -> slab.h dependency is about to be removed. Prepare for this change by updating users of gfp and slab facilities include those headers directly instead of assuming availability. As this conversion needs to touch large number of source files, the following script is used as the basis of conversion. http://userweb.kernel.org/~tj/misc/slabh-sweep.py The script does the followings. * Scan files for gfp and slab usages and update includes such that only the necessary includes are there. ie. if only gfp is used, gfp.h, if slab is used, slab.h. * When the script inserts a new include, it looks at the include blocks and try to put the new include such that its order conforms to its surrounding. It's put in the include block which contains core kernel includes, in the same order that the rest are ordered - alphabetical, Christmas tree, rev-Xmas-tree or at the end if there doesn't seem to be any matching order. * If the script can't find a place to put a new include (mostly because the file doesn't have fitting include block), it prints out an error message indicating which .h file needs to be added to the file. The conversion was done in the following steps. 1. The initial automatic conversion of all .c files updated slightly over 4000 files, deleting around 700 includes and adding ~480 gfp.h and ~3000 slab.h inclusions. The script emitted errors for ~400 files. 2. Each error was manually checked. Some didn't need the inclusion, some needed manual addition while adding it to implementation .h or embedding .c file was more appropriate for others. This step added inclusions to around 150 files. 3. The script was run again and the output was compared to the edits from #2 to make sure no file was left behind. 4. Several build tests were done and a couple of problems were fixed. e.g. lib/decompress_*.c used malloc/free() wrappers around slab APIs requiring slab.h to be added manually. 5. The script was run on all .h files but without automatically editing them as sprinkling gfp.h and slab.h inclusions around .h files could easily lead to inclusion dependency hell. Most gfp.h inclusion directives were ignored as stuff from gfp.h was usually wildly available and often used in preprocessor macros. Each slab.h inclusion directive was examined and added manually as necessary. 6. percpu.h was updated not to include slab.h. 7. Build test were done on the following configurations and failures were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my distributed build env didn't work with gcov compiles) and a few more options had to be turned off depending on archs to make things build (like ipr on powerpc/64 which failed due to missing writeq). * x86 and x86_64 UP and SMP allmodconfig and a custom test config. * powerpc and powerpc64 SMP allmodconfig * sparc and sparc64 SMP allmodconfig * ia64 SMP allmodconfig * s390 SMP allmodconfig * alpha SMP allmodconfig * um on x86_64 SMP allmodconfig 8. percpu.h modifications were reverted so that it could be applied as a separate patch and serve as bisection point. Given the fact that I had only a couple of failures from tests on step 6, I'm fairly confident about the coverage of this conversion patch. If there is a breakage, it's likely to be something in one of the arch headers which should be easily discoverable easily on most builds of the specific arch. Signed-off-by: Tejun Heo Guess-its-ok-by: Christoph Lameter Cc: Ingo Molnar Cc: Lee Schermerhorn --- drivers/gpu/drm/nouveau/nouveau_bo.c | 1 + 1 file changed, 1 insertion(+) (limited to 'drivers/gpu/drm/nouveau/nouveau_bo.c') diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c index 028719fddf7..69c575dff0a 100644 --- a/drivers/gpu/drm/nouveau/nouveau_bo.c +++ b/drivers/gpu/drm/nouveau/nouveau_bo.c @@ -34,6 +34,7 @@ #include "nouveau_dma.h" #include +#include static void nouveau_bo_del_ttm(struct ttm_buffer_object *bo) -- cgit From 78ad0f7bf2bb667729581f099781fc0b7ae58fcc Mon Sep 17 00:00:00 2001 From: Francisco Jerez Date: Thu, 18 Mar 2010 13:07:47 +0100 Subject: drm/nouveau: Make use of TTM busy_placements. Previously we were filling it the same as "placements", but in some cases there're valid alternatives that we were ignoring completely. Keeping a back-up memory type helps on several low-mem situations. Signed-off-by: Francisco Jerez Signed-off-by: Ben Skeggs --- drivers/gpu/drm/nouveau/nouveau_bo.c | 61 +++++++++++++++++++----------------- 1 file changed, 33 insertions(+), 28 deletions(-) (limited to 'drivers/gpu/drm/nouveau/nouveau_bo.c') diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c index 026612471c9..418d881050a 100644 --- a/drivers/gpu/drm/nouveau/nouveau_bo.c +++ b/drivers/gpu/drm/nouveau/nouveau_bo.c @@ -153,7 +153,7 @@ nouveau_bo_new(struct drm_device *dev, struct nouveau_channel *chan, nvbo->placement.fpfn = 0; nvbo->placement.lpfn = mappable ? dev_priv->fb_mappable_pages : 0; - nouveau_bo_placement_set(nvbo, flags); + nouveau_bo_placement_set(nvbo, flags, 0); nvbo->channel = chan; ret = ttm_bo_init(&dev_priv->ttm.bdev, &nvbo->bo, size, @@ -172,26 +172,33 @@ nouveau_bo_new(struct drm_device *dev, struct nouveau_channel *chan, return 0; } +static void +set_placement_list(uint32_t *pl, unsigned *n, uint32_t type, uint32_t flags) +{ + *n = 0; + + if (type & TTM_PL_FLAG_VRAM) + pl[(*n)++] = TTM_PL_FLAG_VRAM | flags; + if (type & TTM_PL_FLAG_TT) + pl[(*n)++] = TTM_PL_FLAG_TT | flags; + if (type & TTM_PL_FLAG_SYSTEM) + pl[(*n)++] = TTM_PL_FLAG_SYSTEM | flags; +} + void -nouveau_bo_placement_set(struct nouveau_bo *nvbo, uint32_t memtype) +nouveau_bo_placement_set(struct nouveau_bo *nvbo, uint32_t type, uint32_t busy) { - int n = 0; - - if (memtype & TTM_PL_FLAG_VRAM) - nvbo->placements[n++] = TTM_PL_FLAG_VRAM | TTM_PL_MASK_CACHING; - if (memtype & TTM_PL_FLAG_TT) - nvbo->placements[n++] = TTM_PL_FLAG_TT | TTM_PL_MASK_CACHING; - if (memtype & TTM_PL_FLAG_SYSTEM) - nvbo->placements[n++] = TTM_PL_FLAG_SYSTEM | TTM_PL_MASK_CACHING; - nvbo->placement.placement = nvbo->placements; - nvbo->placement.busy_placement = nvbo->placements; - nvbo->placement.num_placement = n; - nvbo->placement.num_busy_placement = n; - - if (nvbo->pin_refcnt) { - while (n--) - nvbo->placements[n] |= TTM_PL_FLAG_NO_EVICT; - } + struct ttm_placement *pl = &nvbo->placement; + uint32_t flags = TTM_PL_MASK_CACHING | + (nvbo->pin_refcnt ? TTM_PL_FLAG_NO_EVICT : 0); + + pl->placement = nvbo->placements; + set_placement_list(nvbo->placements, &pl->num_placement, + type, flags); + + pl->busy_placement = nvbo->busy_placements; + set_placement_list(nvbo->busy_placements, &pl->num_busy_placement, + type | busy, flags); } int @@ -199,7 +206,7 @@ nouveau_bo_pin(struct nouveau_bo *nvbo, uint32_t memtype) { struct drm_nouveau_private *dev_priv = nouveau_bdev(nvbo->bo.bdev); struct ttm_buffer_object *bo = &nvbo->bo; - int ret, i; + int ret; if (nvbo->pin_refcnt && !(memtype & (1 << bo->mem.mem_type))) { NV_ERROR(nouveau_bdev(bo->bdev)->dev, @@ -215,9 +222,7 @@ nouveau_bo_pin(struct nouveau_bo *nvbo, uint32_t memtype) if (ret) goto out; - nouveau_bo_placement_set(nvbo, memtype); - for (i = 0; i < nvbo->placement.num_placement; i++) - nvbo->placements[i] |= TTM_PL_FLAG_NO_EVICT; + nouveau_bo_placement_set(nvbo, memtype, 0); ret = ttm_bo_validate(bo, &nvbo->placement, false, false); if (ret == 0) { @@ -244,7 +249,7 @@ nouveau_bo_unpin(struct nouveau_bo *nvbo) { struct drm_nouveau_private *dev_priv = nouveau_bdev(nvbo->bo.bdev); struct ttm_buffer_object *bo = &nvbo->bo; - int ret, i; + int ret; if (--nvbo->pin_refcnt) return 0; @@ -253,8 +258,7 @@ nouveau_bo_unpin(struct nouveau_bo *nvbo) if (ret) return ret; - for (i = 0; i < nvbo->placement.num_placement; i++) - nvbo->placements[i] &= ~TTM_PL_FLAG_NO_EVICT; + nouveau_bo_placement_set(nvbo, bo->mem.placement, 0); ret = ttm_bo_validate(bo, &nvbo->placement, false, false); if (ret == 0) { @@ -439,10 +443,11 @@ nouveau_bo_evict_flags(struct ttm_buffer_object *bo, struct ttm_placement *pl) switch (bo->mem.mem_type) { case TTM_PL_VRAM: - nouveau_bo_placement_set(nvbo, TTM_PL_FLAG_TT); + nouveau_bo_placement_set(nvbo, TTM_PL_FLAG_TT, + TTM_PL_FLAG_SYSTEM); break; default: - nouveau_bo_placement_set(nvbo, TTM_PL_FLAG_SYSTEM); + nouveau_bo_placement_set(nvbo, TTM_PL_FLAG_SYSTEM, 0); break; } -- cgit From a76fb4e8ffe42144529e21fe1e609b762e8eb5cc Mon Sep 17 00:00:00 2001 From: Ben Skeggs Date: Thu, 18 Mar 2010 09:45:20 +1000 Subject: drm/nouveau: detect vram amount once, and save the value As opposed to repeatedly reading the amount back from the GPU every time we need to know the VRAM size. We should now fail to load gracefully on detecting no VRAM, rather than something potentially messy happening. Signed-off-by: Ben Skeggs --- drivers/gpu/drm/nouveau/nouveau_bo.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'drivers/gpu/drm/nouveau/nouveau_bo.c') diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c index 418d881050a..eeab3fb2708 100644 --- a/drivers/gpu/drm/nouveau/nouveau_bo.c +++ b/drivers/gpu/drm/nouveau/nouveau_bo.c @@ -71,7 +71,7 @@ nouveau_bo_fixup_align(struct drm_device *dev, * many small buffers. */ if (dev_priv->card_type == NV_50) { - uint32_t block_size = nouveau_mem_fb_amount(dev) >> 15; + uint32_t block_size = dev_priv->vram_size >> 15; int i; switch (tile_flags) { @@ -399,8 +399,8 @@ nouveau_bo_init_mem_type(struct ttm_bo_device *bdev, uint32_t type, man->io_addr = NULL; man->io_offset = drm_get_resource_start(dev, 1); man->io_size = drm_get_resource_len(dev, 1); - if (man->io_size > nouveau_mem_fb_amount(dev)) - man->io_size = nouveau_mem_fb_amount(dev); + if (man->io_size > dev_priv->vram_size) + man->io_size = dev_priv->vram_size; man->gpu_offset = dev_priv->vm_vram_base; break; -- cgit