All of lore.kernel.org
 help / color / mirror / Atom feed
From: Chris Wilson <chris@chris-wilson.co.uk>
To: intel-gfx@lists.freedesktop.org
Subject: [PATCH 181/190] drm/i915: Introduce an internal allocator for disposable private objects
Date: Mon, 11 Jan 2016 11:01:22 +0000	[thread overview]
Message-ID: <1452510091-6833-40-git-send-email-chris@chris-wilson.co.uk> (raw)
In-Reply-To: <1452510091-6833-1-git-send-email-chris@chris-wilson.co.uk>

Quite a few of our objects used for internal hardware programming do not
benefit from being swappable or from being zero initialised. As such
they do not benefit from using a shmemfs backing storage and since they
are internal and never directly exposed to the user, we do not need to
worry about providing a filp. For these we can use an
drm_i915_gem_object wrapper around a sg_table of plain struct page. They
are not swapped backed and not automatically pinned. If they are reaped
by the shrinker, the pages are released and the contents discarded. For
the internal use case, this is fine as for example, ringbuffers are
pinned from being written by a request to be read by the hardware. Once
they are idle, they can be discarded entirely. As such they are a good
match for execlist ringbuffers and a small varierty of other internal
objects.

In the first iteration, this is limited to the scratch batch buffers we
use (for command parsing and state initialisation).

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
---
 drivers/gpu/drm/i915/Makefile                |   1 +
 drivers/gpu/drm/i915/i915_drv.h              |   8 ++
 drivers/gpu/drm/i915/i915_gem.c              |   9 +-
 drivers/gpu/drm/i915/i915_gem_batch_pool.c   |  27 ++---
 drivers/gpu/drm/i915/i915_gem_internal.c     | 157 +++++++++++++++++++++++++++
 drivers/gpu/drm/i915/i915_gem_render_state.c |   2 +-
 6 files changed, 180 insertions(+), 24 deletions(-)
 create mode 100644 drivers/gpu/drm/i915/i915_gem_internal.c

diff --git a/drivers/gpu/drm/i915/Makefile b/drivers/gpu/drm/i915/Makefile
index a362425ef862..8d0fae65a5bd 100644
--- a/drivers/gpu/drm/i915/Makefile
+++ b/drivers/gpu/drm/i915/Makefile
@@ -28,6 +28,7 @@ i915-y += i915_cmd_parser.o \
 	  i915_gem_execbuffer.o \
 	  i915_gem_fence.o \
 	  i915_gem_gtt.o \
+	  i915_gem_internal.o \
 	  i915_gem.o \
 	  i915_gem_render_state.o \
 	  i915_gem_request.o \
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 2f8b5e7f9320..e3c77d245a6b 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -1989,6 +1989,9 @@ enum hdmi_force_audio {
 #define I915_GTT_OFFSET_NONE ((u32)-1)
 
 struct drm_i915_gem_object_ops {
+	unsigned int flags;
+#define I915_GEM_OBJECT_HAS_STRUCT_PAGE 0x1
+
 	/* Interface between the GEM object and its backing storage.
 	 * get_pages() is called once prior to the use of the associated set
 	 * of pages before to binding them into the GTT, and put_pages() is
@@ -3117,6 +3120,11 @@ i915_gem_object_create_stolen_for_preallocated(struct drm_device *dev,
 					       u32 gtt_offset,
 					       u32 size);
 
+/* i915_gem_internal.c */
+struct drm_i915_gem_object *
+i915_gem_object_create_internal(struct drm_device *dev,
+				unsigned size);
+
 /* i915_gem_shrinker.c */
 unsigned long i915_gem_shrink(struct drm_i915_private *dev_priv,
 			      unsigned long target,
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 9cd161645041..cca45f60d0bd 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -497,7 +497,7 @@ int i915_gem_obj_prepare_shmem_read(struct drm_i915_gem_object *obj,
 	int ret;
 
 	*needs_clflush = 0;
-	if (!obj->base.filp)
+	if ((obj->ops->flags & I915_GEM_OBJECT_HAS_STRUCT_PAGE) == 0)
 		return -EINVAL;
 
 	ret = i915_gem_object_get_pages(obj);
@@ -547,7 +547,7 @@ int i915_gem_obj_prepare_shmem_write(struct drm_i915_gem_object *obj,
 	int ret;
 
 	*needs_clflush = 0;
-	if (!obj->base.filp)
+	if ((obj->ops->flags & I915_GEM_OBJECT_HAS_STRUCT_PAGE) == 0)
 		return -EINVAL;
 
 	ret = i915_gem_object_get_pages(obj);
@@ -800,7 +800,7 @@ i915_gem_pread_ioctl(struct drm_device *dev, void *data,
 	/* prime objects have no backing filp to GEM pread/pwrite
 	 * pages from.
 	 */
-	if (!obj->base.filp) {
+	if ((obj->ops->flags & I915_GEM_OBJECT_HAS_STRUCT_PAGE) == 0) {
 		ret = -EINVAL;
 		goto out;
 	}
@@ -1131,7 +1131,7 @@ i915_gem_pwrite_ioctl(struct drm_device *dev, void *data,
 	/* prime objects have no backing filp to GEM pread/pwrite
 	 * pages from.
 	 */
-	if (!obj->base.filp) {
+	if ((obj->ops->flags & I915_GEM_OBJECT_HAS_STRUCT_PAGE) == 0) {
 		ret = -EINVAL;
 		goto out;
 	}
@@ -3750,6 +3750,7 @@ void i915_gem_object_init(struct drm_i915_gem_object *obj,
 }
 
 static const struct drm_i915_gem_object_ops i915_gem_object_ops = {
+	.flags = I915_GEM_OBJECT_HAS_STRUCT_PAGE,
 	.get_pages = i915_gem_object_get_pages_gtt,
 	.put_pages = i915_gem_object_put_pages_gtt,
 };
diff --git a/drivers/gpu/drm/i915/i915_gem_batch_pool.c b/drivers/gpu/drm/i915/i915_gem_batch_pool.c
index d46012234db1..b876e334da58 100644
--- a/drivers/gpu/drm/i915/i915_gem_batch_pool.c
+++ b/drivers/gpu/drm/i915/i915_gem_batch_pool.c
@@ -98,9 +98,9 @@ i915_gem_batch_pool_get(struct i915_gem_batch_pool *pool,
 			size_t size)
 {
 	struct drm_i915_gem_object *obj = NULL;
-	struct drm_i915_gem_object *tmp, *next;
+	struct drm_i915_gem_object *tmp;
 	struct list_head *list;
-	int n;
+	int n, ret;
 
 	lockdep_assert_held(&pool->engine->dev->struct_mutex);
 
@@ -113,7 +113,7 @@ i915_gem_batch_pool_get(struct i915_gem_batch_pool *pool,
 		n = ARRAY_SIZE(pool->cache_list) - 1;
 	list = &pool->cache_list[n];
 
-	list_for_each_entry_safe(tmp, next, list, batch_pool_link) {
+	list_for_each_entry(tmp, list, batch_pool_link) {
 		/* The batches are strictly LRU ordered */
 		if (i915_gem_object_is_active(tmp)) {
 			struct drm_i915_gem_request *rq;
@@ -126,13 +126,6 @@ i915_gem_batch_pool_get(struct i915_gem_batch_pool *pool,
 			GEM_BUG_ON(tmp->last_write.request);
 		}
 
-		/* While we're looping, do some clean up */
-		if (tmp->madv == __I915_MADV_PURGED) {
-			list_del(&tmp->batch_pool_link);
-			drm_gem_object_unreference(&tmp->base);
-			continue;
-		}
-
 		if (tmp->base.size >= size) {
 			obj = tmp;
 			break;
@@ -140,19 +133,15 @@ i915_gem_batch_pool_get(struct i915_gem_batch_pool *pool,
 	}
 
 	if (obj == NULL) {
-		int ret;
-
-		obj = i915_gem_alloc_object(pool->engine->dev, size);
+		obj = i915_gem_object_create_internal(pool->engine->dev, size);
 		if (obj == NULL)
 			return ERR_PTR(-ENOMEM);
-
-		ret = i915_gem_object_get_pages(obj);
-		if (ret)
-			return ERR_PTR(ret);
-
-		obj->madv = I915_MADV_DONTNEED;
 	}
 
+	ret = i915_gem_object_get_pages(obj);
+	if (ret)
+		return ERR_PTR(ret);
+
 	list_move_tail(&obj->batch_pool_link, list);
 	i915_gem_object_pin_pages(obj);
 	return obj;
diff --git a/drivers/gpu/drm/i915/i915_gem_internal.c b/drivers/gpu/drm/i915/i915_gem_internal.c
new file mode 100644
index 000000000000..0685a304f3ab
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_gem_internal.c
@@ -0,0 +1,157 @@
+/*
+ * Copyright © 2015 Intel Corporation
+ *
+ * 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 <drm/drmP.h>
+#include <drm/i915_drm.h>
+#include "i915_drv.h"
+
+static void __i915_gem_object_free_pages(struct sg_table *st)
+{
+	struct sg_page_iter sg_iter;
+
+	for_each_sg_page(st->sgl, &sg_iter, st->nents, 0)
+		page_cache_release(sg_page_iter_page(&sg_iter));
+
+	sg_free_table(st);
+	kfree(st);
+}
+
+static int i915_gem_object_get_pages_internal(struct drm_i915_gem_object *obj)
+{
+	const unsigned npages = obj->base.size / PAGE_SIZE;
+	struct sg_table *st;
+	struct scatterlist *sg;
+	unsigned long last_pfn = 0;	/* suppress gcc warning */
+	gfp_t gfp;
+	int i;
+
+	st = kmalloc(sizeof(*st), GFP_KERNEL);
+	if (st == NULL)
+		return -ENOMEM;
+
+	if (sg_alloc_table(st, npages, GFP_KERNEL)) {
+		kfree(st);
+		return -ENOMEM;
+	}
+
+	sg = st->sgl;
+	st->nents = 0;
+
+	gfp = GFP_KERNEL | __GFP_HIGHMEM;
+	gfp |= __GFP_NORETRY | __GFP_NOWARN;
+	gfp &= ~(__GFP_IO | __GFP_RECLAIM);
+	for (i = 0; i < npages; i++) {
+		struct page *page;
+
+		page = alloc_page(gfp);
+		if (page == NULL) {
+			i915_gem_shrink_all(to_i915(obj->base.dev));
+			page = alloc_page(GFP_KERNEL | __GFP_HIGHMEM);
+			if (page == NULL)
+				goto err;
+		}
+
+#ifdef CONFIG_SWIOTLB
+		if (swiotlb_nr_tbl()) {
+			st->nents++;
+			sg_set_page(sg, page, PAGE_SIZE, 0);
+			sg = sg_next(sg);
+			continue;
+		}
+#endif
+		if (!i || page_to_pfn(page) != last_pfn + 1) {
+			if (i)
+				sg = sg_next(sg);
+			st->nents++;
+			sg_set_page(sg, page, PAGE_SIZE, 0);
+		} else {
+			sg->length += PAGE_SIZE;
+		}
+		last_pfn = page_to_pfn(page);
+	}
+#ifdef CONFIG_SWIOTLB
+	if (!swiotlb_nr_tbl())
+#endif
+		sg_mark_end(sg);
+	obj->pages = st;
+
+	if (i915_gem_gtt_prepare_object(obj)) {
+		obj->pages = NULL;
+		goto err;
+	}
+
+	obj->madv = I915_MADV_DONTNEED;
+	i915_gem_object_set_dirty(obj);
+	return 0;
+
+err:
+	sg_mark_end(sg);
+	__i915_gem_object_free_pages(st);
+	return -ENOMEM;
+}
+
+static void i915_gem_object_put_pages_internal(struct drm_i915_gem_object *obj)
+{
+	__i915_gem_object_free_pages(obj->pages);
+
+	i915_gem_object_unset_dirty(obj);
+	obj->madv = I915_MADV_WILLNEED;
+}
+
+static const struct drm_i915_gem_object_ops i915_gem_object_internal_ops = {
+	.flags = I915_GEM_OBJECT_HAS_STRUCT_PAGE,
+	.get_pages = i915_gem_object_get_pages_internal,
+	.put_pages = i915_gem_object_put_pages_internal,
+};
+
+/**
+ * Creates a new object that wraps some internal memory for private use.
+ * This object is not backed by swappable storage, and as such its contents
+ * are volatile and only valid whilst pinned. If the object is reaped by the
+ * shrinker, its pages and data will be discarded. Equally, it is not a full
+ * GEM object and so not valid for access from userspace. This makes it useful
+ * for hardware interfaces like ringbuffers (which are pinned from the time
+ * the request is written to the time the hardware stops accessing it), but
+ * not for contexts (which need to be preserved when not active for later
+ * reuse).
+ */
+struct drm_i915_gem_object *
+i915_gem_object_create_internal(struct drm_device *dev,
+				unsigned size)
+{
+	struct drm_i915_gem_object *obj;
+
+	obj = i915_gem_object_alloc(dev);
+	if (obj == NULL)
+		return NULL;
+
+	drm_gem_private_object_init(dev, &obj->base, size);
+	i915_gem_object_init(obj, &i915_gem_object_internal_ops);
+
+	obj->base.write_domain = I915_GEM_DOMAIN_CPU;
+	obj->base.read_domains = I915_GEM_DOMAIN_CPU;
+	obj->cache_level = HAS_LLC(dev) ? I915_CACHE_LLC: I915_CACHE_NONE;
+
+	return obj;
+}
diff --git a/drivers/gpu/drm/i915/i915_gem_render_state.c b/drivers/gpu/drm/i915/i915_gem_render_state.c
index 2fac95b0ba44..436f097d1907 100644
--- a/drivers/gpu/drm/i915/i915_gem_render_state.c
+++ b/drivers/gpu/drm/i915/i915_gem_render_state.c
@@ -66,7 +66,7 @@ static int render_state_init(struct render_state *so, struct drm_device *dev)
 	if (so->rodata->batch_items * 4 > 4096)
 		return -EINVAL;
 
-	so->obj = i915_gem_alloc_object(dev, 4096);
+	so->obj = i915_gem_object_create_internal(dev, 4096);
 	if (so->obj == NULL)
 		return -ENOMEM;
 
-- 
2.7.0.rc3

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

  parent reply	other threads:[~2016-01-11 11:02 UTC|newest]

Thread overview: 263+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-01-11  9:16 [PATCH 001/190] drm: Release driver references to handle before making it available again Chris Wilson
2016-01-11  9:16 ` [PATCH 002/190] drm/i915: Move the mb() following release-mmap into release-mmap Chris Wilson
2016-01-11  9:16 ` [PATCH 003/190] drm/i915: Add an optional selection from i915 of CONFIG_MMU_NOTIFIER Chris Wilson
2016-02-17 12:59   ` Daniel Vetter
2016-01-11  9:16 ` [PATCH 004/190] drm/i915: Fix some invalid requests cancellations Chris Wilson
2016-01-12 18:16   ` [Intel-gfx] " Dave Gordon
2016-01-12 18:16     ` Dave Gordon
2016-01-13 20:06     ` [Intel-gfx] " Chris Wilson
2016-01-11  9:16 ` [PATCH 005/190] drm/i915: Force clean compilation with -Werror Chris Wilson
2016-01-11  9:16 ` [PATCH 006/190] drm/i915: Add GEM debugging Kconfig option Chris Wilson
2016-01-12 17:44   ` Dave Gordon
2016-01-11  9:16 ` [PATCH 007/190] drm/i915: Hide the atomic_read(reset_counter) behind a helper Chris Wilson
2016-01-11  9:16 ` [PATCH 008/190] drm/i915: Simplify checking of GPU reset_counter in display pageflips Chris Wilson
2016-01-11  9:16 ` [PATCH 009/190] drm/i915: Tighten reset_counter for reset status Chris Wilson
2016-01-11  9:16 ` [PATCH 010/190] drm/i915: Store the reset counter when constructing a request Chris Wilson
2016-01-11  9:16 ` [PATCH 011/190] drm/i915: Simplify reset_counter handling during atomic modesetting Chris Wilson
2016-01-11  9:16 ` [PATCH 012/190] drm/i915: Prevent leaking of -EIO from i915_wait_request() Chris Wilson
2016-01-11  9:16 ` [PATCH 013/190] drm/i915: Suppress error message when GPU resets are disabled Chris Wilson
2016-01-11  9:16 ` [PATCH 014/190] drm/i915: Delay queuing hangcheck to wait-request Chris Wilson
2016-01-11  9:16 ` [PATCH 015/190] drm/i915: Remove the dedicated hangcheck workqueue Chris Wilson
2016-01-11  9:16 ` [PATCH 016/190] drm/i915: Make queueing the hangcheck work inline Chris Wilson
2016-01-11  9:16 ` [PATCH 017/190] drm/i915: Remove forcewake dance from seqno/irq barrier on legacy gen6+ Chris Wilson
2016-01-11 14:02   ` Dave Gordon
2016-01-21 16:27     ` Mika Kuoppala
2016-03-24  6:39   ` David Weinehall
2016-01-11  9:16 ` [PATCH 018/190] drm/i915: Slaughter the thundering i915_wait_request herd Chris Wilson
2016-01-11  9:16 ` [PATCH 019/190] drm/i915: Separate out the seqno-barrier from engine->get_seqno Chris Wilson
2016-01-11 15:43   ` Dave Gordon
2016-01-11  9:16 ` [PATCH 020/190] drm/i915: Remove the lazy_coherency parameter from request-completed? Chris Wilson
2016-01-11 15:45   ` Dave Gordon
2016-01-11 16:24     ` Chris Wilson
2016-01-12 10:27   ` Mika Kuoppala
2016-01-12 10:51     ` Chris Wilson
2016-01-11  9:16 ` [PATCH 021/190] drm/i915: Use HWS for seqno tracking everywhere Chris Wilson
2016-01-11 20:03   ` Dave Gordon
2016-01-12 10:05   ` Mika Kuoppala
2016-01-12 11:03     ` Chris Wilson
2016-01-12 14:30       ` Mika Kuoppala
2016-01-12 14:46         ` Chris Wilson
2016-01-11  9:16 ` [PATCH 022/190] drm/i915: Check the CPU cached value of seqno after waking the waiter Chris Wilson
2016-01-11  9:16 ` [PATCH 023/190] drm/i915: Only apply one barrier after a breadcrumb interrupt is posted Chris Wilson
2016-01-11  9:16 ` [PATCH 024/190] drm/i915: Replace manual barrier() with READ_ONCE() in HWS accessor Chris Wilson
2016-01-12 14:17   ` Mika Kuoppala
2016-01-11  9:16 ` [PATCH 025/190] drm/i915: Broadwell execlists needs exactly the same seqno w/a as legacy Chris Wilson
2016-01-11  9:16 ` [PATCH 026/190] drm/i915: Stop setting wraparound seqno on initialisation Chris Wilson
2016-01-11  9:16 ` [PATCH 027/190] drm/i915: Only query timestamp when measuring elapsed time Chris Wilson
2016-01-11  9:16 ` [PATCH 028/190] drm/i915: On GPU reset, set the HWS breadcrumb to the last seqno Chris Wilson
2016-01-11  9:16 ` [PATCH 029/190] drm/i915: Convert trace-irq to the breadcrumb waiter Chris Wilson
2016-01-11  9:16 ` [PATCH 030/190] drm/i915: Move the get/put irq locking into the caller Chris Wilson
2016-01-11  9:16 ` [PATCH 031/190] drm/i915: Harden detection of missed interrupts Chris Wilson
2016-01-11  9:16 ` [PATCH 032/190] drm/i915: Remove debug noise on detecting fault-injection " Chris Wilson
2016-01-11  9:16 ` [PATCH 033/190] drm/i915: Only start retire worker when idle Chris Wilson
2016-01-11  9:16 ` [PATCH 034/190] drm/i915: Do not keep postponing the idle-work Chris Wilson
2016-01-11  9:16 ` [PATCH 035/190] drm/i915: Remove redundant queue_delayed_work() from throttle ioctl Chris Wilson
2016-01-11  9:16 ` [PATCH 036/190] drm/i915: Restore waitboost credit to the synchronous waiter Chris Wilson
2016-01-11 16:10   ` Jesse Barnes
2016-01-11  9:16 ` [PATCH 037/190] drm/i915: Add background commentary to "waitboosting" Chris Wilson
2016-01-11  9:16 ` [PATCH 038/190] drm/i915: Flush the RPS bottom-half when the GPU idles Chris Wilson
2016-01-11  9:16 ` [PATCH 039/190] drm/i915: Remove stop-rings debugfs interface Chris Wilson
2016-02-25 17:30   ` Arun Siluvery
2016-01-11  9:16 ` [PATCH 040/190] drm/i915: Record the ringbuffer associated with the request Chris Wilson
2016-01-11  9:16 ` [PATCH 041/190] drm/i915: Allow userspace to request no-error-capture upon GPU hangs Chris Wilson
2016-01-11  9:16 ` [PATCH 042/190] drm/i915: Clean up GPU hang message Chris Wilson
2016-02-25 17:40   ` Arun Siluvery
2016-01-11  9:16 ` [PATCH 043/190] drm/i915: Skip capturing an error state if we already have one Chris Wilson
2016-01-11  9:16 ` [PATCH 044/190] drm/i915: Move GEM request routines to i915_gem_request.c Chris Wilson
2016-02-25 17:52   ` Arun Siluvery
2016-03-08 12:58     ` Tvrtko Ursulin
2016-03-08 13:35       ` Arun Siluvery
2016-01-11  9:16 ` [PATCH 045/190] drm/i915: Move releasing of the GEM request from free to retire/cancel Chris Wilson
2016-03-08 13:15   ` Tvrtko Ursulin
2016-04-05 13:42     ` Tvrtko Ursulin
2016-04-05 14:09       ` Chris Wilson
2016-04-05 14:17         ` Tvrtko Ursulin
2016-04-05 14:27           ` Chris Wilson
2016-04-05 14:45             ` Chris Wilson
2016-04-05 14:10       ` Chris Wilson
2016-01-11  9:16 ` [PATCH 046/190] drm/i915: Derive GEM requests from dma-fence Chris Wilson
2016-01-11  9:16 ` [PATCH 047/190] drm/i915: Rename request reference/unreference to get/put Chris Wilson
2016-01-11  9:16 ` [PATCH 048/190] drm/i915: Disable waitboosting for fence_wait() Chris Wilson
2016-01-11  9:17 ` [PATCH 049/190] drm/i915: Disable waitboosting for mmioflips/semaphores Chris Wilson
2016-01-11  9:17 ` [PATCH 050/190] drm/i915: Refactor duplicate object vmap functions Chris Wilson
2016-01-11  9:17 ` [PATCH 051/190] drm,i915: Introduce drm_malloc_gfp() Chris Wilson
2016-01-11  9:17 ` [PATCH 052/190] drm/i915: Treat ringbuffer writes as write to normal memory Chris Wilson
2016-01-11  9:17 ` [PATCH 053/190] drm/i915: Convert i915_semaphores_is_enabled over to early sanitize Chris Wilson
2016-01-12 19:07   ` Dave Gordon
2016-01-11  9:17 ` [PATCH 054/190] drm/i915: Use the new rq->i915 field where appropriate Chris Wilson
2016-01-11  9:17 ` [PATCH 055/190] drm/i915: Unify intel_logical_ring_emit and intel_ring_emit Chris Wilson
2016-01-12 17:29   ` Dave Gordon
2016-01-11  9:17 ` [PATCH 056/190] drm/i915: Unify intel_ring_begin() Chris Wilson
2016-01-11  9:17 ` [PATCH 057/190] drm/i915: Remove the identical implementations of request space reservation Chris Wilson
2016-01-11  9:17 ` [PATCH 058/190] drm/i915: Rename request->ring to request->engine Chris Wilson
2016-01-28 11:45   ` Tvrtko Ursulin
2016-01-11  9:17 ` [PATCH 059/190] drm/i915: Rename request->ringbuf to request->ring Chris Wilson
2016-01-28 11:48   ` Tvrtko Ursulin
2016-01-11  9:17 ` [PATCH 060/190] drm/i915: Rename backpointer from intel_ringbuffer to intel_engine_cs Chris Wilson
2016-01-28 11:49   ` Tvrtko Ursulin
2016-01-11  9:17 ` [PATCH 061/190] drm/i915: Rename intel_context[engine].ringbuf Chris Wilson
2016-01-11  9:17 ` [PATCH 062/190] drm/i915: Rename extern functions operating on intel_engine_cs Chris Wilson
2016-01-11  9:17 ` [PATCH 063/190] drm/i915: Rename struct intel_ringbuffer to intel_ring Chris Wilson
2016-01-28 11:54   ` Tvrtko Ursulin
2016-01-11  9:17 ` [PATCH 064/190] drm/i915: Rename intel_pin_and_map_ring() Chris Wilson
2016-01-11  9:17 ` [PATCH 065/190] drm/i915: Remove obsolete engine->gpu_caches_dirty Chris Wilson
2016-01-11  9:17 ` [PATCH 066/190] drm/i915: Simplify request_alloc by returning the allocated request Chris Wilson
2016-01-12 17:11   ` Dave Gordon
2016-01-11  9:17 ` [PATCH 067/190] drm/i915: Unify legacy/execlists emission of MI_BATCHBUFFER_START Chris Wilson
2016-01-11  9:17 ` [PATCH 068/190] drm/i915: Unify adding requests between ringbuffer and execlists Chris Wilson
2016-01-11  9:17 ` [PATCH 069/190] drm/i915: Remove duplicate golden render state init from execlists Chris Wilson
2016-01-11  9:17 ` [PATCH 070/190] drm/i915: Unify legacy/execlists submit_execbuf callbacks Chris Wilson
2016-01-11  9:17 ` [PATCH 071/190] drm/i915: Simplify calling engine->sync_to Chris Wilson
2016-01-11  9:17 ` [PATCH 072/190] drm/i915: Execlists cannot pin a context without the object Chris Wilson
2016-01-11 15:24   ` Tvrtko Ursulin
2016-01-11  9:17 ` [PATCH 073/190] drm/i915: Introduce i915_gem_active for request tracking Chris Wilson
2016-01-11 17:32   ` Tvrtko Ursulin
2016-01-11 22:49     ` Chris Wilson
2016-01-12 10:04       ` Tvrtko Ursulin
2016-01-12 11:01         ` Chris Wilson
2016-01-12 13:42           ` Tvrtko Ursulin
2016-01-12 13:44           ` Tvrtko Ursulin
2016-01-12 14:08             ` Chris Wilson
2016-01-11  9:17 ` [PATCH 074/190] drm/i915: Rename request->list to link for consistency Chris Wilson
2016-01-12 13:47   ` Tvrtko Ursulin
2016-01-11  9:17 ` [PATCH 075/190] drm/i915: Refactor activity tracking for requests Chris Wilson
2016-01-28 11:41   ` Tvrtko Ursulin
2016-01-28 11:46     ` Chris Wilson
2016-01-28 11:56       ` Tvrtko Ursulin
2016-01-11  9:17 ` [PATCH 076/190] drm/i915: Rename vma->*_list to *_link for consistency Chris Wilson
2016-01-12 13:49   ` Tvrtko Ursulin
2016-01-11  9:17 ` [PATCH 077/190] drm/i915: Amalgamate GGTT/ppGTT vma debug list walkers Chris Wilson
2016-01-11  9:17 ` [PATCH 078/190] drm/i915: Split early global GTT initialisation Chris Wilson
2016-01-11  9:17 ` [PATCH 079/190] drm/i915: Reduce the pointer dance of i915_is_ggtt() Chris Wilson
2016-01-15 12:12   ` Dave Gordon
2016-01-15 12:24     ` Chris Wilson
2016-01-11  9:17 ` [PATCH 080/190] drm/i915: Store owning file on the i915_address_space Chris Wilson
2016-01-11  9:17 ` [PATCH 081/190] drm/i915: i915_vma_move_to_active prep patch Chris Wilson
2016-01-11  9:17 ` [PATCH 082/190] drm/i915: Count how many VMA are bound for an object Chris Wilson
2016-01-11  9:17 ` [PATCH 083/190] drm/i915: Be more careful when unbinding vma Chris Wilson
2016-01-11  9:17 ` [PATCH 084/190] drm/i915: Track active vma requests Chris Wilson
2016-01-11  9:17 ` [PATCH 085/190] drm/i915: Release vma when the handle is closed Chris Wilson
2016-01-11  9:17 ` [PATCH 086/190] drm/i915: Mark the context and address space as closed Chris Wilson
2016-01-11 10:44 ` [PATCH 087/190] Revert "drm/i915: Clean up associated VMAs on context destruction" Chris Wilson
2016-01-11 10:44   ` [PATCH 088/190] drm/i915: Move execlists interrupt based submission to a bottom-half Chris Wilson
2016-02-19 12:08     ` Tvrtko Ursulin
2016-02-19 12:29       ` Chris Wilson
2016-02-19 14:10         ` Tvrtko Ursulin
2016-02-19 14:34           ` Chris Wilson
2016-02-19 14:52             ` Tvrtko Ursulin
2016-02-19 15:02               ` Chris Wilson
2016-02-19 14:41           ` Chris Wilson
2016-01-11 10:44   ` [PATCH 089/190] drm/i915: Tidy execlists submission and tracking Chris Wilson
2016-01-11 10:44   ` [PATCH 090/190] drm/i915: Refactor execlists default context pinning Chris Wilson
2016-01-11 10:44   ` [PATCH 091/190] drm/i915: Move context initialisation to first-use Chris Wilson
2016-01-11 10:44   ` [PATCH 092/190] drm/i915: Move the magical deferred context allocation into the request Chris Wilson
2016-01-11 10:44   ` [PATCH 093/190] drm/i915: Move the forced switch back to the kernel context into eviction Chris Wilson
2016-01-11 10:44   ` [PATCH 094/190] drm/i915: Remove early l3-remap Chris Wilson
2016-01-11 10:44   ` [PATCH 095/190] drm/i915: Rearrange switch_context to load the aliasing ppgtt on first use Chris Wilson
2016-01-11 10:44   ` [PATCH 096/190] drm/i915: Eliminate early submission of context enabling request Chris Wilson
2016-01-11 10:44   ` [PATCH 097/190] drm/i915/shrinker: Flush active on objects before counting Chris Wilson
2016-01-11 10:44   ` [PATCH 098/190] drm/i915: Double check the active status on the batch pool Chris Wilson
2016-01-11 10:44   ` [PATCH 099/190] drm/i915: Check for request completion before choosing CS flips Chris Wilson
2016-01-11 10:44   ` [PATCH 100/190] drm/i915: Remove request retirement before each batch Chris Wilson
2016-01-11 10:44   ` [PATCH 101/190] drm/i915: Only retire if necessary when creating a userptr Chris Wilson
2016-01-11 10:44   ` [PATCH 102/190] drm/i915: Move the "per-ring" default_context to the device Chris Wilson
2016-01-11 14:40     ` Dave Gordon
2016-01-11 10:44   ` [PATCH 103/190] drm/i915: Move pinning of dev_priv->kernel_context into its creator Chris Wilson
2016-01-11 10:44   ` [PATCH 104/190] drm/i915: Remove i915_gem_execbuffer_retire_commands() Chris Wilson
2016-01-11 10:44   ` [PATCH 105/190] drm/i915: Pad GTT views of exec objects up to user specified size Chris Wilson
2016-03-22 14:32     ` David Weinehall
2016-01-11 10:44   ` [PATCH 106/190] drm/i915: Split insertion/binding of an object into the VM Chris Wilson
2016-01-11 10:44   ` [PATCH 107/190] drm/i915: Record allocated vma size Chris Wilson
2016-01-11 10:44   ` [PATCH 108/190] drm/i915: Start passing around i915_vma from execbuffer Chris Wilson
2016-01-11 10:44   ` [PATCH 109/190] drm/i915: Remove highly confusing i915_gem_obj_ggtt_pin() Chris Wilson
2016-01-11 10:44   ` [PATCH 110/190] drm/i915: Move vma->pin_count:4 to vma->flags Chris Wilson
2016-01-11 10:44   ` [PATCH 111/190] drm/i915: Make fb_tracking.lock a spinlock Chris Wilson
2016-01-11 10:44   ` [PATCH 112/190] drm/i915: Move obj->active:5 to obj->flags Chris Wilson
2016-03-24 12:00     ` David Weinehall
2016-01-11 10:44   ` [PATCH 113/190] drm/i915: Enable lockless lookup of request tracking via RCU Chris Wilson
2016-01-11 10:44   ` [PATCH 114/190] drm/i915: Remove (struct_mutex) locking for wait-ioctl Chris Wilson
2016-01-11 10:44   ` [PATCH 115/190] drm/i915: Remove (struct_mutex) locking for busy-ioctl Chris Wilson
2016-01-11 10:45   ` [PATCH 116/190] drm/i915: Reduce locking inside swfinish ioctl Chris Wilson
2016-01-11 10:45   ` [PATCH 117/190] drm/i915: Remove pinned check from madvise ioctl Chris Wilson
2016-01-11 10:45   ` [PATCH 118/190] drm/i915: Remove locking for get_tiling Chris Wilson
2016-01-11 10:45   ` [PATCH 119/190] drm/i915: Reduce amount of duplicate buffer information captured on error Chris Wilson
2016-01-11 10:45   ` [PATCH 120/190] drm/i915: Stop the machine whilst capturing the GPU crash dump Chris Wilson
2016-01-11 10:45   ` [PATCH 121/190] drm/i915: Scan GGTT active list for context object Chris Wilson
2016-01-11 10:45   ` [PATCH 122/190] drm/i915: Move setting of request->batch into its single callsite Chris Wilson
2016-01-11 10:45   ` [PATCH 123/190] drm/i915: Mark unmappable GGTT entries as PIN_HIGH Chris Wilson
2016-01-11 10:45   ` [PATCH 124/190] drm/i915: Track pinned vma inside guc Chris Wilson
2016-01-11 10:45   ` [PATCH 125/190] drm/i915: Track pinned VMA Chris Wilson
2016-01-11 10:45   ` [PATCH 126/190] drm/i915: Print the batchbuffer offset next to BBADDR in error state Chris Wilson
2016-01-11 10:45   ` [PATCH 127/190] drm/i915: Cache kmap between relocations Chris Wilson
2016-01-11 10:45   ` [PATCH 128/190] drm/i915: Extract i915_gem_obj_prepare_shmem_write() Chris Wilson
2016-01-11 10:45   ` [PATCH 129/190] drm/i915: Before accessing an object via the cpu, flush GTT writes Chris Wilson
2016-01-11 10:45   ` [PATCH 130/190] drm/i915: Wait for writes through the GTT to land before reading back Chris Wilson
2016-01-11 10:45   ` [PATCH 131/190] drm/i915: Pin the pages first in shmem prepare read/write Chris Wilson
2016-01-11 10:45   ` [PATCH 132/190] drm/i915: Tidy up flush cpu/gtt write domains Chris Wilson
2016-01-11 10:45   ` [PATCH 133/190] drm/i915: Convert known clflush paths over to clflush_cache_range() Chris Wilson
2016-01-11 10:45   ` [PATCH 134/190] drm/i915: Refactor execbuffer relocation writing Chris Wilson
2016-01-11 10:45   ` [PATCH 135/190] drm/i915: Move map-and-fenceable tracking to the VMA Chris Wilson
2016-01-11 10:45   ` [PATCH 136/190] drm/i915: Move ioremap_wc tracking onto VMA Chris Wilson
2016-02-11 13:20     ` Tvrtko Ursulin
2016-02-11 13:29       ` Chris Wilson
2016-02-11 14:10         ` Tvrtko Ursulin
2016-02-19 15:11           ` Chris Wilson
2016-02-22 15:29             ` Tvrtko Ursulin
2016-02-23 10:21               ` Chris Wilson
2016-01-11 10:45   ` [PATCH 137/190] drm/i915: Shrink pages around failure to dma map Chris Wilson
2016-01-11 10:45   ` [PATCH 138/190] drm/i915/userptr: Make gup errors stickier Chris Wilson
2016-01-11 10:45   ` [PATCH 139/190] drm/i915: Move fence tracking from object to vma Chris Wilson
2016-01-11 10:45   ` [PATCH 140/190] drm/i915: Fix partial GGTT faulting Chris Wilson
2016-01-11 10:45   ` [PATCH 141/190] drm/i915: Choose not to evict faultable objects from the GGTT Chris Wilson
2016-01-11 11:00 ` [PATCH 142/190] drm/i915: Fallback to using unmappable memory for scanout Chris Wilson
2016-01-11 11:00   ` [PATCH 143/190] drm/i915: Track display alignment on VMA Chris Wilson
2016-01-11 11:00   ` [PATCH 144/190] drm/i915: Bump the inactive MRU tracking for all VMA accessed Chris Wilson
2016-01-11 11:00   ` [PATCH 145/190] drm/i915: Stop discarding GTT cache-domain on unbind vma Chris Wilson
2016-01-12 13:22     ` Joonas Lahtinen
2016-01-11 11:00   ` [PATCH 146/190] io-mapping: Always create a struct to hold metadata about the io-mapping Chris Wilson
2016-01-11 11:00   ` [PATCH 147/190] drm/i915: Use remap_io_mapping() to prefault all PTE in a single pass Chris Wilson
2016-01-11 11:00   ` [PATCH 148/190] drm/i915: Stop marking the unaccessible scratch page as UC Chris Wilson
2016-01-11 11:00   ` [PATCH 149/190] drm/i915: Use i915_vm_to_ppgtt() Chris Wilson
2016-01-11 11:00   ` [PATCH 150/190] drm/i915: Embed the scratch page struct into each VM Chris Wilson
2016-01-11 11:00   ` [PATCH 151/190] drm/i915: Allow DMA pagetables to use highmem Chris Wilson
2016-01-11 11:00   ` [PATCH 152/190] drm/i915: Replace request->postfix with ->head for space searching Chris Wilson
2016-01-11 11:00   ` [PATCH 153/190] drm/i915: Record the position of the start of the request Chris Wilson
2016-01-11 11:00   ` [PATCH 154/190] drm/i915: Move per-request pid from request to ctx Chris Wilson
2016-01-11 11:00   ` [PATCH 155/190] drm/i915: Merge legacy+execlists context structs Chris Wilson
2016-01-11 11:00   ` [PATCH 156/190] drm/i915: Store the active context object on all engines upon error Chris Wilson
2016-01-11 11:00   ` [PATCH 157/190] drm/i915: Tidy execlists by using intel_context_engine locals Chris Wilson
2016-01-11 11:00   ` [PATCH 158/190] drm/i915: Skip holding an object reference for execbuf preparation Chris Wilson
2016-01-11 11:01   ` [PATCH 159/190] drm/i915: Defer active reference until required Chris Wilson
2016-01-11 11:01   ` [PATCH 160/190] drm: Track drm_mm nodes with an interval tree Chris Wilson
2016-01-11 11:01   ` [PATCH 161/190] drm: Convert drm_vma_manager to embedded interval-tree in drm_mm Chris Wilson
2016-01-11 11:01   ` [PATCH 162/190] drm/i915: Allow the user to pass a context to any ring Chris Wilson
2016-01-11 11:01   ` [PATCH 163/190] drm/i915: Fix i915_gem_evict_for_vma (soft-pinning) Chris Wilson
2016-01-11 11:01   ` [PATCH 164/190] drm/i915: Move obj->dirty:1 to obj->flags Chris Wilson
2016-03-24  8:17     ` David Weinehall
2016-01-11 11:01   ` [PATCH 165/190] drm/i915: Use the precomputed value for whether to enable command parsing Chris Wilson
2016-01-11 11:01   ` [PATCH 166/190] drm/i915: Drop spinlocks around adding to the client request list Chris Wilson
2016-01-11 11:01   ` [PATCH 167/190] drm/i915: Amalgamate execbuffer parameter structures Chris Wilson
2016-01-11 11:01   ` [PATCH 168/190] drm/i915: Skip holding context reference for duration of execbuffer call Chris Wilson
2016-01-11 11:01   ` [PATCH 169/190] drm/i915: Use vma->exec_entry as our double-entry placeholder Chris Wilson
2016-01-11 11:01   ` [PATCH 170/190] drm/i915: Store a direct lookup from object handle to vma Chris Wilson
2016-01-11 11:01   ` [PATCH 171/190] drm/i915: Pass vma to relocate entry Chris Wilson
2016-01-11 11:01   ` [PATCH 172/190] drm/i915: Eliminate lots of iterations over the execobjects array Chris Wilson
2016-01-11 11:01   ` [PATCH 173/190] drm/i915: Wait upon userptr get-user-pages within execbuffer Chris Wilson
2016-01-11 11:01   ` [PATCH 174/190] drm/i915: Show context objects in debugfs/i915_gem_objects Chris Wilson
2016-03-24  7:58     ` David Weinehall
2016-01-11 11:01   ` [PATCH 175/190] drm/i915: Remove superfluous i915_add_request_no_flush() helper Chris Wilson
2016-01-11 11:01   ` [PATCH 176/190] drm/i915: Use the MRU stack search after evicting Chris Wilson
2016-01-11 11:01   ` [PATCH 177/190] drm/i915: Use VMA as the primary object for context state Chris Wilson
2016-01-11 11:01   ` [PATCH 178/190] drm/i915: Do an inline flush-active before dropping the mutex when waiting Chris Wilson
2016-01-11 11:01   ` [PATCH 179/190] drm/i915: Skip MI_SET_CONTEXT for the same context Chris Wilson
2016-01-11 11:01   ` [PATCH 180/190] drm/i915: Micro-optimise i915_gem_object_get_dirty_page() Chris Wilson
2016-01-11 11:01   ` Chris Wilson [this message]
2016-01-11 11:01   ` [PATCH 182/190] drm/i915: Avoid allocating a vmap arena for a single page Chris Wilson
2016-01-11 11:01   ` [PATCH 183/190] drm/i915/cmdparser: Use cached vmappings Chris Wilson
2016-01-11 11:01   ` [PATCH 184/190] drm/i915/cmdparser: Only cache the dst vmap Chris Wilson
2016-01-11 11:01   ` [PATCH 185/190] drm/i915/cmdparser: Improve hash function Chris Wilson
2016-01-11 11:01   ` [PATCH 186/190] drm/i915/cmdparser: Compare against the previous command descriptor Chris Wilson
2016-01-11 11:01   ` [PATCH 187/190] drm/i915: Allow execbuffer to use the first object as the batch Chris Wilson
2016-01-11 11:01   ` [PATCH 188/190] drm/i915: Use VMA for ringbuffer tracking Chris Wilson
2016-01-11 11:01   ` [PATCH 189/190] drm/i915: Skip clearing the GGTT on full-ppgtt systems Chris Wilson
2016-01-11 11:01   ` [PATCH 190/190] drm/i915: Do a nonblocking wait first in pread/pwrite Chris Wilson

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1452510091-6833-40-git-send-email-chris@chris-wilson.co.uk \
    --to=chris@chris-wilson.co.uk \
    --cc=intel-gfx@lists.freedesktop.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.