drm/i915: Exclude low pages (128KiB) of stolen from use

The GPU is trashing the low pages of its reserved memory upon reset. If
we are using this memory for ringbuffers, then we will dutiful resubmit
the trashed rings after the reset causing further resets, and worse. We
must exclude this range from our own use. The value of 128KiB was found
by empirical measurement (and verified now with a selftest) on gen9.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: stable@vger.kernel.org
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201019165005.18128-2-chris@chris-wilson.co.uk
(cherry picked from commit d3606757e611fbd48bb239e8c2fe9779b3f50035)
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

1 files changed, 196 insertions, 0 deletions

diff --git a/drivers/gpu/drm/i915/gt/selftest_reset.c b/drivers/gpu/drm/i915/gt/selftest_reset.c
index 35406ecdf0b2..ef5aeebbeeb0 100644
--- a/drivers/gpu/drm/i915/gt/selftest_reset.c
+++ b/drivers/gpu/drm/i915/gt/selftest_reset.c
@@ -3,9 +3,203 @@
  * Copyright © 2018 Intel Corporation
  */
 
+#include <linux/crc32.h>
+
+#include "gem/i915_gem_stolen.h"
+
+#include "i915_memcpy.h"
 #include "i915_selftest.h"
 #include "selftests/igt_reset.h"
 #include "selftests/igt_atomic.h"
+#include "selftests/igt_spinner.h"
+
+static int
+__igt_reset_stolen(struct intel_gt *gt,
+		   intel_engine_mask_t mask,
+		   const char *msg)
+{
+	struct i915_ggtt *ggtt = &gt->i915->ggtt;
+	const struct resource *dsm = &gt->i915->dsm;
+	resource_size_t num_pages, page;
+	struct intel_engine_cs *engine;
+	intel_wakeref_t wakeref;
+	enum intel_engine_id id;
+	struct igt_spinner spin;
+	long max, count;
+	void *tmp;
+	u32 *crc;
+	int err;
+
+	if (!drm_mm_node_allocated(&ggtt->error_capture))
+		return 0;
+
+	num_pages = resource_size(dsm) >> PAGE_SHIFT;
+	if (!num_pages)
+		return 0;
+
+	crc = kmalloc_array(num_pages, sizeof(u32), GFP_KERNEL);
+	if (!crc)
+		return -ENOMEM;
+
+	tmp = kmalloc(PAGE_SIZE, GFP_KERNEL);
+	if (!tmp) {
+		err = -ENOMEM;
+		goto err_crc;
+	}
+
+	igt_global_reset_lock(gt);
+	wakeref = intel_runtime_pm_get(gt->uncore->rpm);
+
+	err = igt_spinner_init(&spin, gt);
+	if (err)
+		goto err_lock;
+
+	for_each_engine(engine, gt, id) {
+		struct intel_context *ce;
+		struct i915_request *rq;
+
+		if (!(mask & engine->mask))
+			continue;
+
+		if (!intel_engine_can_store_dword(engine))
+			continue;
+
+		ce = intel_context_create(engine);
+		if (IS_ERR(ce)) {
+			err = PTR_ERR(ce);
+			goto err_spin;
+		}
+		rq = igt_spinner_create_request(&spin, ce, MI_ARB_CHECK);
+		intel_context_put(ce);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto err_spin;
+		}
+		i915_request_add(rq);
+	}
+
+	for (page = 0; page < num_pages; page++) {
+		dma_addr_t dma = (dma_addr_t)dsm->start + (page << PAGE_SHIFT);
+		void __iomem *s;
+		void *in;
+
+		ggtt->vm.insert_page(&ggtt->vm, dma,
+				     ggtt->error_capture.start,
+				     I915_CACHE_NONE, 0);
+		mb();
+
+		s = io_mapping_map_wc(&ggtt->iomap,
+				      ggtt->error_capture.start,
+				      PAGE_SIZE);
+
+		if (!__drm_mm_interval_first(&gt->i915->mm.stolen,
+					     page << PAGE_SHIFT,
+					     ((page + 1) << PAGE_SHIFT) - 1))
+			memset32(s, STACK_MAGIC, PAGE_SIZE / sizeof(u32));
+
+		in = s;
+		if (i915_memcpy_from_wc(tmp, s, PAGE_SIZE))
+			in = tmp;
+		crc[page] = crc32_le(0, in, PAGE_SIZE);
+
+		io_mapping_unmap(s);
+	}
+	mb();
+	ggtt->vm.clear_range(&ggtt->vm, ggtt->error_capture.start, PAGE_SIZE);
+
+	if (mask == ALL_ENGINES) {
+		intel_gt_reset(gt, mask, NULL);
+	} else {
+		for_each_engine(engine, gt, id) {
+			if (mask & engine->mask)
+				intel_engine_reset(engine, NULL);
+		}
+	}
+
+	max = -1;
+	count = 0;
+	for (page = 0; page < num_pages; page++) {
+		dma_addr_t dma = (dma_addr_t)dsm->start + (page << PAGE_SHIFT);
+		void __iomem *s;
+		void *in;
+		u32 x;
+
+		ggtt->vm.insert_page(&ggtt->vm, dma,
+				     ggtt->error_capture.start,
+				     I915_CACHE_NONE, 0);
+		mb();
+
+		s = io_mapping_map_wc(&ggtt->iomap,
+				      ggtt->error_capture.start,
+				      PAGE_SIZE);
+
+		in = s;
+		if (i915_memcpy_from_wc(tmp, s, PAGE_SIZE))
+			in = tmp;
+		x = crc32_le(0, in, PAGE_SIZE);
+
+		if (x != crc[page] &&
+		    !__drm_mm_interval_first(&gt->i915->mm.stolen,
+					     page << PAGE_SHIFT,
+					     ((page + 1) << PAGE_SHIFT) - 1)) {
+			pr_debug("unused stolen page %pa modified by GPU reset\n",
+				 &page);
+			if (count++ == 0)
+				igt_hexdump(in, PAGE_SIZE);
+			max = page;
+		}
+
+		io_mapping_unmap(s);
+	}
+	mb();
+	ggtt->vm.clear_range(&ggtt->vm, ggtt->error_capture.start, PAGE_SIZE);
+
+	if (count > 0) {
+		pr_info("%s reset clobbered %ld pages of stolen, last clobber at page %ld\n",
+			msg, count, max);
+	}
+	if (max >= I915_GEM_STOLEN_BIAS >> PAGE_SHIFT) {
+		pr_err("%s reset clobbered unreserved area [above %x] of stolen; may cause severe faults\n",
+		       msg, I915_GEM_STOLEN_BIAS);
+		err = -EINVAL;
+	}
+
+err_spin:
+	igt_spinner_fini(&spin);
+
+err_lock:
+	intel_runtime_pm_put(gt->uncore->rpm, wakeref);
+	igt_global_reset_unlock(gt);
+
+	kfree(tmp);
+err_crc:
+	kfree(crc);
+	return err;
+}
+
+static int igt_reset_device_stolen(void *arg)
+{
+	return __igt_reset_stolen(arg, ALL_ENGINES, "device");
+}
+
+static int igt_reset_engines_stolen(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int err;
+
+	if (!intel_has_reset_engine(gt))
+		return 0;
+
+	for_each_engine(engine, gt, id) {
+		err = __igt_reset_stolen(gt, engine->mask, engine->name);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
 
 static int igt_global_reset(void *arg)
 {
@@ -164,6 +358,8 @@ int intel_reset_live_selftests(struct drm_i915_private *i915)
 {
 	static const struct i915_subtest tests[] = {
 		SUBTEST(igt_global_reset), /* attempt to recover GPU first */
+		SUBTEST(igt_reset_device_stolen),
+		SUBTEST(igt_reset_engines_stolen),
 		SUBTEST(igt_wedged_reset),
 		SUBTEST(igt_atomic_reset),
 		SUBTEST(igt_atomic_engine_reset),