diff --git a/gm/blurrect.cpp b/gm/blurrect.cpp
index 1e245712fe..54f3c216c0 100644
--- a/gm/blurrect.cpp
+++ b/gm/blurrect.cpp
@@ -58,7 +58,7 @@ static void draw_donut_skewed(SkCanvas* canvas, const SkRect& r, const SkPaint&
 /*
  * Spits out a dummy gradient to test blur with shader on paint
  */
-static sk_sp<SkShader> MakeRadial() {
+static sk_sp<SkShader> make_radial() {
     SkPoint pts[2] = {
         { 0, 0 },
         { SkIntToScalar(100), SkIntToScalar(100) }
@@ -123,7 +123,7 @@ protected:
                 paint.setAlpha(fAlpha);
 
                 SkPaint paintWithRadial = paint;
-                paintWithRadial.setShader(MakeRadial());
+                paintWithRadial.setShader(make_radial());
 
                 constexpr Proc procs[] = {
                     fill_rect, draw_donut, draw_donut_skewed
diff --git a/gm/yuvtorgbeffect.cpp b/gm/yuvtorgbeffect.cpp
index 001c53d701..f957566d9f 100644
--- a/gm/yuvtorgbeffect.cpp
+++ b/gm/yuvtorgbeffect.cpp
@@ -253,8 +253,9 @@ protected:
                 SkMatrix viewMatrix;
                 viewMatrix.setTranslate(x, y);
                 grPaint.addColorFragmentProcessor(std::move(fp));
-                renderTargetContext->priv().testingOnly_addDrawOp(GrRectOpFactory::MakeNonAAFill(
-                        std::move(grPaint), viewMatrix, renderRect, GrAAType::kNone));
+                std::unique_ptr<GrDrawOp> op(GrRectOpFactory::MakeNonAAFill(
+                                std::move(grPaint), viewMatrix, renderRect, GrAAType::kNone));
+                renderTargetContext->priv().testingOnly_addDrawOp(std::move(op));
             }
         }
     }
diff --git a/src/gpu/GrContext.cpp b/src/gpu/GrContext.cpp
index d05f1976f1..ba26cc844e 100644
--- a/src/gpu/GrContext.cpp
+++ b/src/gpu/GrContext.cpp
@@ -646,12 +646,6 @@ bool GrContextPriv::readSurfacePixels(GrSurfaceContext* src,
         return false;
     }
 
-    if (!proxyToRead->instantiate(fContext->resourceProvider())) {
-        return false;
-    }
-
-    GrSurface* surfaceToRead = proxyToRead->priv().peekSurface();
-
     if (GrGpu::kRequireDraw_DrawPreference == drawPreference && !didTempDraw) {
         return false;
     }
@@ -660,6 +654,13 @@ bool GrContextPriv::readSurfacePixels(GrSurfaceContext* src,
         this->flushSurfaceWrites(proxyToRead.get());
         configToRead = tempDrawInfo.fReadConfig;
     }
+
+    if (!proxyToRead->instantiate(fContext->resourceProvider())) {
+        return false;
+    }
+
+    GrSurface* surfaceToRead = proxyToRead->priv().peekSurface();
+
     if (!fContext->fGpu->readPixels(surfaceToRead, proxyToRead->origin(),
                                     left, top, width, height, configToRead, buffer, rowBytes)) {
         return false;
diff --git a/src/gpu/GrOpList.cpp b/src/gpu/GrOpList.cpp
index 521beea743..17b0f10ebf 100644
--- a/src/gpu/GrOpList.cpp
+++ b/src/gpu/GrOpList.cpp
@@ -31,6 +31,7 @@ GrOpList::GrOpList(GrResourceProvider* resourceProvider,
     fTarget.setProxy(sk_ref_sp(surfaceProxy), kWrite_GrIOType);
     fTarget.get()->setLastOpList(this);
 
+#ifndef MDB_ALLOC_RESOURCES
     // MDB TODO: remove this! We are currently moving to having all the ops that target
     // the RT as a dest (e.g., clear, etc.) rely on the opList's 'fTarget' pointer
     // for the IO Ref. This works well but until they are all swapped over (and none
@@ -39,6 +40,7 @@ GrOpList::GrOpList(GrResourceProvider* resourceProvider,
     // re-use assumptions.
     fTarget.get()->instantiate(resourceProvider);
     fTarget.markPendingIO();
+#endif
 }
 
 GrOpList::~GrOpList() {
@@ -101,6 +103,10 @@ void GrOpList::addDependency(GrSurfaceProxy* dependedOn, const GrCaps& caps) {
 }
 
 #ifdef SK_DEBUG
+bool GrOpList::isInstantiated() const {
+    return fTarget.get()->priv().isInstantiated();
+}
+
 void GrOpList::dump() const {
     SkDebugf("--------------------------------------------------------------\n");
     SkDebugf("node: %d -> RT: %d\n", fUniqueID, fTarget.get() ? fTarget.get()->uniqueID().asUInt()
diff --git a/src/gpu/GrOpList.h b/src/gpu/GrOpList.h
index ae6b9a1906..ef21a24f46 100644
--- a/src/gpu/GrOpList.h
+++ b/src/gpu/GrOpList.h
@@ -98,6 +98,8 @@ public:
     void setStencilLoadOp(GrLoadOp loadOp) { fStencilLoadOp = loadOp; }
 
 protected:
+    SkDEBUGCODE(bool isInstantiated() const;)
+
     GrSurfaceProxyRef fTarget;
     GrAuditTrail*     fAuditTrail;
 
diff --git a/src/gpu/GrRenderTargetContext.cpp b/src/gpu/GrRenderTargetContext.cpp
index 561eaef340..4f1ad1279e 100644
--- a/src/gpu/GrRenderTargetContext.cpp
+++ b/src/gpu/GrRenderTargetContext.cpp
@@ -124,10 +124,12 @@ GrRenderTargetContext::GrRenderTargetContext(GrContext* context,
         fColorXformFromSRGB = GrColorSpaceXform::Make(srgbColorSpace.get(), fColorSpace.get());
     }
 
+#ifndef MDB_ALLOC_RESOURCES
     // MDB TODO: to ensure all resources still get allocated in the correct order in the hybrid
     // world we need to get the correct opList here so that it, in turn, can grab and hold
     // its rendertarget.
     this->getRTOpList();
+#endif
     SkDEBUGCODE(this->validate();)
 }
 
diff --git a/src/gpu/GrResourceAllocator.cpp b/src/gpu/GrResourceAllocator.cpp
index 76abe1855c..e5d6dbf360 100644
--- a/src/gpu/GrResourceAllocator.cpp
+++ b/src/gpu/GrResourceAllocator.cpp
@@ -134,3 +134,35 @@ void GrResourceAllocator::assign() {
         fActiveIntvls.insertByIncreasingEnd(cur);
     }
 }
+
+#ifdef SK_DEBUG
+void GrResourceAllocator::dump() {
+    unsigned int min = fNumOps+1;
+    unsigned int max = 0;
+    for(const Interval* cur = fIntvlList.peekHead(); cur; cur = cur->fNext) {
+        SkDebugf("{ %d,%d }: [%d, %d]\n",
+                 cur->fProxy->uniqueID().asUInt(), cur->fProxy->underlyingUniqueID().asUInt(),
+                 cur->fStart, cur->fEnd);
+        if (min > cur->fStart) {
+            min = cur->fStart;
+        }
+        if (max < cur->fEnd) {
+            max = cur->fEnd;
+        }
+    }
+
+    for(const Interval* cur = fIntvlList.peekHead(); cur; cur = cur->fNext) {
+        SkDebugf("{ %3d,%3d }: ",
+                 cur->fProxy->uniqueID().asUInt(), cur->fProxy->underlyingUniqueID().asUInt());
+        for (unsigned int i = min; i <= max; ++i) {
+            if (i >= cur->fStart && i <= cur->fEnd) {
+                SkDebugf("x");
+            } else {
+                SkDebugf(" ");
+            }
+        }
+        SkDebugf("\n");
+    }
+}
+#endif
+
diff --git a/src/gpu/GrResourceAllocator.h b/src/gpu/GrResourceAllocator.h
index fcaa3d5c9a..bd0c690a58 100644
--- a/src/gpu/GrResourceAllocator.h
+++ b/src/gpu/GrResourceAllocator.h
@@ -52,6 +52,7 @@ public:
     }
 
     void assign();
+    SkDEBUGCODE(void dump();)
 
 private:
     class Interval;
@@ -92,7 +93,7 @@ private:
         static uint32_t Hash(const uint32_t& key) { return key; }
 
         GrSurfaceProxy* fProxy;
-        uint32_t        fProxyID; // This is here b.c. DynamicHash requires a ref to the key 
+        uint32_t        fProxyID; // This is here b.c. DynamicHash requires a ref to the key
         unsigned int    fStart;
         unsigned int    fEnd;
         Interval*       fNext;
diff --git a/src/gpu/text/GrAtlasTextBlob.cpp b/src/gpu/text/GrAtlasTextBlob.cpp
index 823444c972..b88ed9329e 100644
--- a/src/gpu/text/GrAtlasTextBlob.cpp
+++ b/src/gpu/text/GrAtlasTextBlob.cpp
@@ -298,10 +298,6 @@ inline void GrAtlasTextBlob::flushRun(GrRenderTargetContext* rtc, const GrClip&
     int lastRun = fRuns[run].fSubRunInfo.count() - 1;
     for (int subRun = 0; subRun <= lastRun; subRun++) {
         const Run::SubRunInfo& info = fRuns[run].fSubRunInfo[subRun];
-        GrPaint grPaint;
-        if (!paint.toGrPaint(info.maskFormat(), rtc, viewMatrix, &grPaint)) {
-            continue;
-        }
         int glyphCount = info.glyphCount();
         if (0 == glyphCount) {
             continue;
diff --git a/tests/CanvasTest.cpp b/tests/CanvasTest.cpp
index e64f5e95f7..7aaf554015 100644
--- a/tests/CanvasTest.cpp
+++ b/tests/CanvasTest.cpp
@@ -105,8 +105,7 @@ DEF_TEST(canvas_clipbounds, reporter) {
     }
 }
 
-// Will call proc with multiple styles of canse (recording, raster, pdf)
-//
+// Will call proc with multiple styles of canvas (recording, raster, pdf)
 template <typename F> static void multi_canvas_driver(int w, int h, F proc) {
     proc(SkPictureRecorder().beginRecording(SkRect::MakeIWH(w, h)));
 
@@ -116,7 +115,6 @@ template <typename F> static void multi_canvas_driver(int w, int h, F proc) {
     proc(SkSurface::MakeRasterN32Premul(w, h, nullptr)->getCanvas());
 }
 
-
 const SkIRect gBaseRestrictedR = { 0, 0, 10, 10 };
 
 static void test_restriction(skiatest::Reporter* reporter, SkCanvas* canvas) {
diff --git a/tests/GpuSampleLocationsTest.cpp b/tests/GpuSampleLocationsTest.cpp
index 399f4310d6..fa99a9ae9e 100644
--- a/tests/GpuSampleLocationsTest.cpp
+++ b/tests/GpuSampleLocationsTest.cpp
@@ -117,11 +117,11 @@ void test_sampleLocations(skiatest::Reporter* reporter, TestSampleLocationsInter
     for (int i = 0; i < numTestPatterns; ++i) {
         int numSamples = (int)kTestPatterns[i].size();
         GrAlwaysAssert(numSamples > 1 && SkIsPow2(numSamples));
-        bottomUps[i] = ctx->makeDeferredRenderTargetContextWithFallback(
+        bottomUps[i] = ctx->makeDeferredRenderTargetContext(
                            SkBackingFit::kExact, 100, 100, kRGBA_8888_GrPixelConfig, nullptr,
                            rand.nextRangeU(1 + numSamples / 2, numSamples),
                            kBottomLeft_GrSurfaceOrigin);
-        topDowns[i] = ctx->makeDeferredRenderTargetContextWithFallback(
+        topDowns[i] = ctx->makeDeferredRenderTargetContext(
                           SkBackingFit::kExact, 100, 100, kRGBA_8888_GrPixelConfig, nullptr,
                           rand.nextRangeU(1 + numSamples / 2, numSamples),
                           kTopLeft_GrSurfaceOrigin);