[Intel-gfx] [PATCH 01/39] drm/i915/gem: Avoid implicit vmap for highmem on x86-32

[Chris Wilson]

On 32b, highmem uses a finite set of indirect PTE (i.e. vmap) to provide
virtual mappings of the high pages. As these are finite, map_new_virtual()
must wait for some other kmap() to finish when it runs out. If we map a
large number of objects, there is no method for it to tell us to release
the mappings, and we deadlock.

However, if we make an explicit vmap of the page, that uses a larger
vmalloc arena, and also has the ability to tell us to release unwanted
mappings. Most importantly, it will fail and propagate an error instead
of waiting forever.

Fixes: fb8621d3bee8 ("drm/i915: Avoid allocating a vmap arena for a single page") #x86-32
References: e87666b52f00 ("drm/i915/shrinker: Hook up vmap allocation failure notifier")
Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
Cc: Harald Arnesen <harald at skogtun.org>
Cc: <stable at vger.kernel.org> # v4.7+
---
 drivers/gpu/drm/i915/gem/i915_gem_pages.c | 26 +++++++++++++++++++++--
 1 file changed, 24 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/i915/gem/i915_gem_pages.c b/drivers/gpu/drm/i915/gem/i915_gem_pages.c
index 7050519c87a4..51b63e05dbe4 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_pages.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_pages.c
@@ -255,8 +255,30 @@ static void *i915_gem_object_map(struct drm_i915_gem_object *obj,
 		return NULL;
 
 	/* A single page can always be kmapped */
-	if (n_pte == 1 && type == I915_MAP_WB)
-		return kmap(sg_page(sgt->sgl));
+	if (n_pte == 1 && type == I915_MAP_WB) {
+		struct page *page = sg_page(sgt->sgl);
+
+		/*
+		 * On 32b, highmem uses a finite set of indirect PTE (i.e.
+		 * vmap) to provide virtual mappings of the high pages.
+		 * As these are finite, map_new_virtual() must wait for some
+		 * other kmap() to finish when it runs out. If we map a large
+		 * number of objects, there is no method for it to tell us
+		 * to release the mappings, and we deadlock.
+		 *
+		 * However, if we make an explicit vmap of the page, that
+		 * uses a larger vmalloc arena, and also has the ability
+		 * to tell us to release unwanted mappings. Most importantly,
+		 * it will fail and propagate an error instead of waiting
+		 * forever.
+		 *
+		 * So if the page is beyond the 32b boundary, make an explicit
+		 * vmap. On 64b, this check will be optimised away as we can
+		 * directly kmap any page on the system.
+		 */
+		if (!PageHighMem(page))
+			return kmap(page);
+	}
 
 	mem = stack;
 	if (n_pte > ARRAY_SIZE(stack)) {
-- 
2.20.1