Merge tag 'perf-core-for-mingo-5.4-20190814' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux into perf/core

Pull perf/core improvements and fixes from Arnaldo:

Intel PT:

  Adrian Hunter:

  - Add PEBS via Intel PT support, the kernel bits went via PeterZ.

perf record:

  Alexander Shishkin:

  - Add an option to take an AUX snapshot on exit.

  Tan Xiaojun:

  - Support aarch64 random socket_id assignment, just like was fixed for S/390.

tools:

  Andy Shevchenko:

  - Keep list of tools in alphabetical order on 'make -C tools help'.

perf session:

  Arnaldo Carvalho de Melo:

  - Avoid infinite loop when seeing invalid header.size, reported by
    Vince Weaver using a perf.data fuzzer.

Documentation:

  Vince Weaver:

  - Clarify HEADER_SAMPLE_TOPOLOGY format in the perf.data spec.

perf config:

  Arnaldo Carvalho de Melo:

  - Honour $PERF_CONFIG env var to specify alternate .perfconfig.

perf test:

  Arnaldo Carvalho de Melo:

  - Disable ~/.perfconfig to get default output in 'perf trace' tests.

perf top:

  Arnaldo Carvalho de Melo:

  - Set display thread COMM to help with debugging.

  - Collapse and resort evsels in a group, so that we have output
    similar to 'perf report' when using event groups, i.e.

      perf top -e '{cycles,instructions}'

    Will have two columns, and the instructions one will work.

core:

  Igor Lubashev:

  - Detect if libcap development files are available so that we
    can use capabilities to match the checks made by the kernel instead
    of using plain (geteuid() == 0).

Intel:

  Haiyan Song:

  - Add Icelake V1.00 event file.

perf trace:

  Leo Yan:

  - Fix segmentation fault when access syscall info on arm64.

Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

1 files changed, 42 insertions, 5 deletions

diff --git a/drivers/gpu/drm/msm/msm_gem.c b/drivers/gpu/drm/msm/msm_gem.c
index c2114c748c2f..8cf6362e64bf 100644
--- a/drivers/gpu/drm/msm/msm_gem.c
+++ b/drivers/gpu/drm/msm/msm_gem.c
@@ -32,6 +32,46 @@ static bool use_pages(struct drm_gem_object *obj)
 	return !msm_obj->vram_node;
 }
 
+/*
+ * Cache sync.. this is a bit over-complicated, to fit dma-mapping
+ * API.  Really GPU cache is out of scope here (handled on cmdstream)
+ * and all we need to do is invalidate newly allocated pages before
+ * mapping to CPU as uncached/writecombine.
+ *
+ * On top of this, we have the added headache, that depending on
+ * display generation, the display's iommu may be wired up to either
+ * the toplevel drm device (mdss), or to the mdp sub-node, meaning
+ * that here we either have dma-direct or iommu ops.
+ *
+ * Let this be a cautionary tail of abstraction gone wrong.
+ */
+
+static void sync_for_device(struct msm_gem_object *msm_obj)
+{
+	struct device *dev = msm_obj->base.dev->dev;
+
+	if (get_dma_ops(dev)) {
+		dma_sync_sg_for_device(dev, msm_obj->sgt->sgl,
+			msm_obj->sgt->nents, DMA_BIDIRECTIONAL);
+	} else {
+		dma_map_sg(dev, msm_obj->sgt->sgl,
+			msm_obj->sgt->nents, DMA_BIDIRECTIONAL);
+	}
+}
+
+static void sync_for_cpu(struct msm_gem_object *msm_obj)
+{
+	struct device *dev = msm_obj->base.dev->dev;
+
+	if (get_dma_ops(dev)) {
+		dma_sync_sg_for_cpu(dev, msm_obj->sgt->sgl,
+			msm_obj->sgt->nents, DMA_BIDIRECTIONAL);
+	} else {
+		dma_unmap_sg(dev, msm_obj->sgt->sgl,
+			msm_obj->sgt->nents, DMA_BIDIRECTIONAL);
+	}
+}
+
 /* allocate pages from VRAM carveout, used when no IOMMU: */
 static struct page **get_pages_vram(struct drm_gem_object *obj, int npages)
 {
@@ -97,8 +137,7 @@ static struct page **get_pages(struct drm_gem_object *obj)
 		 * because display controller, GPU, etc. are not coherent:
 		 */
 		if (msm_obj->flags & (MSM_BO_WC|MSM_BO_UNCACHED))
-			dma_sync_sg_for_device(dev->dev, msm_obj->sgt->sgl,
-					msm_obj->sgt->nents, DMA_BIDIRECTIONAL);
+			sync_for_device(msm_obj);
 	}
 
 	return msm_obj->pages;
@@ -127,9 +166,7 @@ static void put_pages(struct drm_gem_object *obj)
 			 * GPU, etc. are not coherent:
 			 */
 			if (msm_obj->flags & (MSM_BO_WC|MSM_BO_UNCACHED))
-				dma_sync_sg_for_cpu(obj->dev->dev, msm_obj->sgt->sgl,
-					     msm_obj->sgt->nents,
-					     DMA_BIDIRECTIONAL);
+				sync_for_cpu(msm_obj);
 
 			sg_free_table(msm_obj->sgt);
 			kfree(msm_obj->sgt);