Date: Sat, 10 Feb 2024 20:17:14 GMT From: "Jason E. Hale" <jhale@FreeBSD.org> To: ports-committers@FreeBSD.org, dev-commits-ports-all@FreeBSD.org, dev-commits-ports-branches@FreeBSD.org Subject: git: 2acd494ee955 - 2024Q1 - www/qt6-webengine: Address security vulnerabilities Message-ID: <202402102017.41AKHEp0019595@gitrepo.freebsd.org>
next in thread | raw e-mail | index | archive | help
The branch 2024Q1 has been updated by jhale: URL: https://cgit.FreeBSD.org/ports/commit/?id=2acd494ee955806e9af81d1476510dc89eb82786 commit 2acd494ee955806e9af81d1476510dc89eb82786 Author: Jason E. Hale <jhale@FreeBSD.org> AuthorDate: 2024-02-10 18:43:37 +0000 Commit: Jason E. Hale <jhale@FreeBSD.org> CommitDate: 2024-02-10 20:16:31 +0000 www/qt6-webengine: Address security vulnerabilities Patched with security patches up to Chromium version: 121.0.6167.160 MFH: 2024Q1 Security: dc9e5237-c197-11ee-86bb-a8a1599412c6, 19047673-c680-11ee-86bb-a8a1599412c6 (cherry picked from commit 5104585a99d85d1fa23516f95917e8f10beb5bef) --- www/qt6-webengine/Makefile | 2 +- www/qt6-webengine/files/patch-security-rollup | 540 ++++++++++++++++++++++++++ 2 files changed, 541 insertions(+), 1 deletion(-) diff --git a/www/qt6-webengine/Makefile b/www/qt6-webengine/Makefile index b62f3f3a255b..9cf2441c7458 100644 --- a/www/qt6-webengine/Makefile +++ b/www/qt6-webengine/Makefile @@ -12,7 +12,7 @@ PORTNAME?= webengine DISTVERSION= ${QT6_VERSION} -PORTREVISION?= 4 # Master port for print/qt6-pdf. Please keep this line. +PORTREVISION?= 5 # Master port for print/qt6-pdf. Please keep this line. CATEGORIES?= www PKGNAMEPREFIX= qt6- diff --git a/www/qt6-webengine/files/patch-security-rollup b/www/qt6-webengine/files/patch-security-rollup index 3f67e42ad06b..2f8615470498 100644 --- a/www/qt6-webengine/files/patch-security-rollup +++ b/www/qt6-webengine/files/patch-security-rollup @@ -30,6 +30,12 @@ Addresses the following security issues: - Security bug 1511389 - CVE-2024-0810 - Security bug 1407197 +- Security bug 1519980 +- CVE-2024-1060 +- CVE-2024-1077 +- CVE-2024-1059 +- CVE-2024-1283 +- CVE-2024-1284 From 669506a53474e3d7637666d3c53f6101fb94d96f Mon Sep 17 00:00:00 2001 From: Nidhi Jaju <nidhijaju@chromium.org> @@ -4437,3 +4443,537 @@ index d8b3ad83bbb..3087f9c3e0b 100644 } } +From f2480155fcf5f753d60b818986d136fcd2309edc Mon Sep 17 00:00:00 2001 +From: =?UTF-8?q?Peter=20Bostr=C3=B6m?= <pbos@chromium.org> +Date: Tue, 23 Jan 2024 01:06:06 +0000 +Subject: [PATCH] [Backport] Security bug 1519980 +MIME-Version: 1.0 +Content-Type: text/plain; charset=UTF-8 +Content-Transfer-Encoding: 8bit + +Manual cherry-pick of patch originally reviewed on +https://chromium-review.googlesource.com/c/chromium/src/+/5226127: +Speculatively fix race in mojo ShutDownOnIOThread + +This acquires `write_lock_` before resetting handles used by WriteNoLock +(which is called under the same lock in another thread). We also set +`reject_writes_` to prevent future write attempts after shutdown. That +seems strictly more correct. + +We also acquire `fds_to_close_lock_` before clearing the FDs. + +I was unable to repro locally as content_browsertests just times out +in my local setup without reporting anything interesting. This seems +strictly more correct though. + +Bug: 1519980 +Change-Id: I96279936ca908ecb98eddd381df20d61597cba43 +Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5226127 +Auto-Submit: Peter Boström <pbos@chromium.org> +Reviewed-by: Ken Rockot <rockot@google.com> +Commit-Queue: Ken Rockot <rockot@google.com> +Commit-Queue: Peter Boström <pbos@chromium.org> +Cr-Commit-Position: refs/heads/main@{#1250580} +Reviewed-on: https://codereview.qt-project.org/c/qt/qtwebengine-chromium/+/537138 +Reviewed-by: Michal Klocek <michal.klocek@qt.io> +--- + chromium/mojo/core/channel_posix.cc | 25 +++++++++++++++---------- + 1 file changed, 15 insertions(+), 10 deletions(-) + +diff --git a/chromium/mojo/core/channel_posix.cc b/chromium/mojo/core/channel_posix.cc +index f57c9b3cb5f..faf728fdd3d 100644 +--- src/3rdparty/chromium/mojo/core/channel_posix.cc ++++ src/3rdparty/chromium/mojo/core/channel_posix.cc +@@ -264,18 +264,23 @@ void ChannelPosix::WaitForWriteOnIOThreadNoLock() { + void ChannelPosix::ShutDownOnIOThread() { + base::CurrentThread::Get()->RemoveDestructionObserver(this); + +- read_watcher_.reset(); +- write_watcher_.reset(); +- if (leak_handle_) { +- std::ignore = socket_.release(); +- server_.TakePlatformHandle().release(); +- } else { +- socket_.reset(); +- std::ignore = server_.TakePlatformHandle(); +- } ++ { ++ base::AutoLock lock(write_lock_); ++ reject_writes_ = true; ++ read_watcher_.reset(); ++ write_watcher_.reset(); ++ if (leak_handle_) { ++ std::ignore = socket_.release(); ++ server_.TakePlatformHandle().release(); ++ } else { ++ socket_.reset(); ++ std::ignore = server_.TakePlatformHandle(); ++ } + #if BUILDFLAG(IS_IOS) +- fds_to_close_.clear(); ++ base::AutoLock fd_lock(fds_to_close_lock_); ++ fds_to_close_.clear(); + #endif ++ } + + // May destroy the |this| if it was the last reference. + self_ = nullptr; +From d9b4b11c104ec5112900dad72af8ff058c3f069b Mon Sep 17 00:00:00 2001 +From: Jean-Philippe Gravel <jpgravel@chromium.org> +Date: Wed, 17 Jan 2024 17:45:45 +0000 +Subject: [PATCH] [Backport] CVE-2024-1060: Use after free in Canvas + +Manual backport of patch originally reviewed on +https://chromium-review.googlesource.com/c/chromium/src/+/5198419: +Fix use-after-free in DrawTextInternal + +DrawTextInternal was calling GetOrCreatePaintCanvas multiple times, +once at the start of the function, once inside of the +BaseRenderingContext2DAutoRestoreSkCanvas helper class and once in the +Draw call. GetOrCreatePaintCanvas destroys the canvas resource provider +if the GPU context is lost. If this happens on the second call to +GetOrCreatePaintCanvas, destroying the resource provider will +invalidate the cc::PaintCanvas returned by the first call to +GetOrCreatePaintCanvas. + +The GPU process can technically crash at any point during the renderer +process execution (perhaps because of something another renderer +process did). We therefore have to assume that any call to +GetOrCreatePaintCanvas can invalidate previously returned +cc::PaintCanvas. + +Change-Id: Ifa77735ab1b2b55b3d494f886b8566299937f6fe +Fixed: 1511567 +Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5198419 +Reviewed-by: Fernando Serboncini <fserb@chromium.org> +Commit-Queue: Jean-Philippe Gravel <jpgravel@chromium.org> +Cr-Commit-Position: refs/heads/main@{#1248204} +Reviewed-on: https://codereview.qt-project.org/c/qt/qtwebengine-chromium/+/537140 +Reviewed-by: Michal Klocek <michal.klocek@qt.io> +--- + .../canvas2d/canvas_rendering_context_2d.cc | 50 ++++++------------- + .../canvas2d/canvas_rendering_context_2d.h | 2 - + 2 files changed, 16 insertions(+), 36 deletions(-) + +diff --git a/chromium/third_party/blink/renderer/modules/canvas/canvas2d/canvas_rendering_context_2d.cc b/chromium/third_party/blink/renderer/modules/canvas/canvas2d/canvas_rendering_context_2d.cc +index 01720502d6a..adab5144f93 100644 +--- src/3rdparty/chromium/third_party/blink/renderer/modules/canvas/canvas2d/canvas_rendering_context_2d.cc ++++ src/3rdparty/chromium/third_party/blink/renderer/modules/canvas/canvas2d/canvas_rendering_context_2d.cc +@@ -102,35 +102,6 @@ static mojom::blink::ColorScheme GetColorSchemeFromCanvas( + return mojom::blink::ColorScheme::kLight; + } + +-// Drawing methods need to use this instead of SkAutoCanvasRestore in case +-// overdraw detection substitutes the recording canvas (to discard overdrawn +-// draw calls). +-class CanvasRenderingContext2DAutoRestoreSkCanvas { +- STACK_ALLOCATED(); +- +- public: +- explicit CanvasRenderingContext2DAutoRestoreSkCanvas( +- CanvasRenderingContext2D* context) +- : context_(context) { +- DCHECK(context_); +- cc::PaintCanvas* c = context_->GetOrCreatePaintCanvas(); +- if (c) { +- save_count_ = c->getSaveCount(); +- } +- } +- +- ~CanvasRenderingContext2DAutoRestoreSkCanvas() { +- cc::PaintCanvas* c = context_->GetOrCreatePaintCanvas(); +- if (c) +- c->restoreToCount(save_count_); +- context_->ValidateStateStack(); +- } +- +- private: +- CanvasRenderingContext2D* context_; +- int save_count_ = 0; +-}; +- + CanvasRenderingContext* CanvasRenderingContext2D::Factory::Create( + CanvasRenderingContextHost* host, + const CanvasContextCreationAttributesCore& attrs) { +@@ -999,9 +970,11 @@ void CanvasRenderingContext2D::DrawTextInternal( + // to 0, for example), so update style before grabbing the PaintCanvas. + canvas()->GetDocument().UpdateStyleAndLayoutTreeForNode(canvas()); + +- cc::PaintCanvas* c = GetOrCreatePaintCanvas(); +- if (!c) ++ // Abort if we don't have a paint canvas (e.g. the context was lost). ++ cc::PaintCanvas* paint_canvas = GetOrCreatePaintCanvas(); ++ if (!paint_canvas) { + return; ++ } + + if (!std::isfinite(x) || !std::isfinite(y)) + return; +@@ -1066,14 +1039,13 @@ void CanvasRenderingContext2D::DrawTextInternal( + if (paint_type == CanvasRenderingContext2DState::kStrokePaintType) + InflateStrokeRect(bounds); + +- CanvasRenderingContext2DAutoRestoreSkCanvas state_restorer(this); + if (use_max_width) { +- c->save(); ++ paint_canvas->save(); + // We draw when fontWidth is 0 so compositing operations (eg, a "copy" op) + // still work. As the width of canvas is scaled, so text can be scaled to + // match the given maxwidth, update text location so it appears on desired + // place. +- c->scale(ClampTo<float>(width / font_width), 1); ++ paint_canvas->scale(ClampTo<float>(width / font_width), 1); + location.set_x(location.x() / ClampTo<float>(width / font_width)); + } + +@@ -1093,6 +1065,16 @@ void CanvasRenderingContext2D::DrawTextInternal( + { return false; }, + bounds, paint_type, CanvasRenderingContext2DState::kNoImage, + CanvasPerformanceMonitor::DrawType::kText); ++ ++ if (use_max_width) { ++ // Cannot use `paint_canvas` in case recording canvas was substituted or ++ // destroyed during draw call. ++ cc::PaintCanvas* c = GetPaintCanvas(); ++ if (c) { ++ c->restore(); ++ } ++ } ++ ValidateStateStack(); + } + + const Font& CanvasRenderingContext2D::AccessFont() { +diff --git a/chromium/third_party/blink/renderer/modules/canvas/canvas2d/canvas_rendering_context_2d.h b/chromium/third_party/blink/renderer/modules/canvas/canvas2d/canvas_rendering_context_2d.h +index 508af63e75a..59566cb117c 100644 +--- src/3rdparty/chromium/third_party/blink/renderer/modules/canvas/canvas2d/canvas_rendering_context_2d.h ++++ src/3rdparty/chromium/third_party/blink/renderer/modules/canvas/canvas2d/canvas_rendering_context_2d.h +@@ -245,8 +245,6 @@ class MODULES_EXPORT CanvasRenderingContext2D final + void TryRestoreContextEvent(TimerBase*) override; + + private: +- friend class CanvasRenderingContext2DAutoRestoreSkCanvas; +- + void PruneLocalFontCache(size_t target_size); + + void ScrollPathIntoViewInternal(const Path&); +From 5f7b5772910e721f0cbdfd97925e84afa94aeec8 Mon Sep 17 00:00:00 2001 +From: Tsuyoshi Horo <horo@chromium.org> +Date: Tue, 9 Jan 2024 08:40:00 +0000 +Subject: [PATCH] [Backport] CVE-2024-1077: Use after free in Network + +Cherry-pick of patch originally reviewed on +https://chromium-review.googlesource.com/c/chromium/src/+/5179746: +Fix UAF in SourceStreamToDataPipe + +SourceStreamToDataPipe::ReadMore() is passing a callback with +Unretained(this) to net::SourceStream::Read(). But this callback may be +called even after the SourceStream is destructed. This is causing UAF +issue (crbug.com/1511085). + +To solve this problem, this CL changes ReadMore() method to pass a +callback with a weak ptr of this. + +Bug: 1511085 +Change-Id: Idd4e34ff300ff5db2de1de7b303841c7db3a964a +Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5179746 +Reviewed-by: Adam Rice <ricea@chromium.org> +Commit-Queue: Tsuyoshi Horo <horo@chromium.org> +Cr-Commit-Position: refs/heads/main@{#1244526} +Reviewed-on: https://codereview.qt-project.org/c/qt/qtwebengine-chromium/+/537141 +Reviewed-by: Michal Klocek <michal.klocek@qt.io> +--- + .../network/public/cpp/source_stream_to_data_pipe.cc | 6 +++--- + 1 file changed, 3 insertions(+), 3 deletions(-) + +diff --git a/chromium/services/network/public/cpp/source_stream_to_data_pipe.cc b/chromium/services/network/public/cpp/source_stream_to_data_pipe.cc +index bfd85b1a00b..07afd58a40f 100644 +--- src/3rdparty/chromium/services/network/public/cpp/source_stream_to_data_pipe.cc ++++ src/3rdparty/chromium/services/network/public/cpp/source_stream_to_data_pipe.cc +@@ -55,9 +55,9 @@ void SourceStreamToDataPipe::ReadMore() { + + scoped_refptr<net::IOBuffer> buffer( + new network::NetToMojoIOBuffer(pending_write_.get())); +- int result = source_->Read( +- buffer.get(), base::checked_cast<int>(num_bytes), +- base::BindOnce(&SourceStreamToDataPipe::DidRead, base::Unretained(this))); ++ int result = source_->Read(buffer.get(), base::checked_cast<int>(num_bytes), ++ base::BindOnce(&SourceStreamToDataPipe::DidRead, ++ weak_factory_.GetWeakPtr())); + + if (result != net::ERR_IO_PENDING) + DidRead(result); +From 9bcf4d966b8315c3801721222c937f6c4fbc00b2 Mon Sep 17 00:00:00 2001 +From: =?UTF-8?q?Michael=20Br=C3=BCning?= <michael.bruning@qt.io> +Date: Wed, 7 Feb 2024 12:07:44 +0100 +Subject: [PATCH] Fixup: [Backport] Security bug 1407197 + +It was missing setting one of the debug locations in code that we +may potentially compile. + +Change-Id: Ia47c270eb042d131621babaef3927b0745c36014 +Reviewed-on: https://codereview.qt-project.org/c/qt/qtwebengine-chromium/+/537953 +Reviewed-by: Michal Klocek <michal.klocek@qt.io> +--- + chromium/chrome/browser/devtools/devtools_window.cc | 1 + + 1 file changed, 1 insertion(+) + +diff --git a/chromium/chrome/browser/devtools/devtools_window.cc b/chromium/chrome/browser/devtools/devtools_window.cc +index de1b8b019fc..94343b63153 100644 +--- src/3rdparty/chromium/chrome/browser/devtools/devtools_window.cc ++++ src/3rdparty/chromium/chrome/browser/devtools/devtools_window.cc +@@ -1301,6 +1301,7 @@ void DevToolsWindow::AddNewContents( + bool* was_blocked) { + if (new_contents.get() == toolbox_web_contents_) { + owned_toolbox_web_contents_ = std::move(new_contents); ++ owned_toolbox_web_contents_->SetOwnerLocationForDebug(FROM_HERE); + + toolbox_web_contents_->SetDelegate(new DevToolsToolboxDelegate( + toolbox_web_contents_, inspected_web_contents_)); +From beb4a95a8040535701840e84338998b711cf86ff Mon Sep 17 00:00:00 2001 +From: Guido Urdaneta <guidou@chromium.org> +Date: Thu, 18 Jan 2024 16:47:18 +0000 +Subject: [PATCH] [Backport] CVE-2024-1059: Use after free in WebRTC + +Manual backport of patch originally reviewed on +https://chromium-review.googlesource.com/c/chromium/src/+/5210359: +[RTCPeerConnection] Exit early from RTCPeerConnectionHandler + +For certain operations that require a live client +(i.e., RTCPeerConnection, which is garbage collected), +PeerConnectionHandler keeps a pointer to the client on the stack +to prevent garbage collection. + +In some cases, the client may have already been garbage collected +(the client is null). In that case, there is no point in doing the +operation and it should exit early to avoid UAF/crashes. + +This CL adds early exit to the cases that do not already have it. + +Bug: 1514777 +Change-Id: I27e9541cfaa74d978799c03e2832a0980f9e5710 +Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5210359 +Reviewed-by: Tomas Gunnarsson <tommi@chromium.org> +Commit-Queue: Guido Urdaneta <guidou@chromium.org> +Cr-Commit-Position: refs/heads/main@{#1248826} +Reviewed-on: https://codereview.qt-project.org/c/qt/qtwebengine-chromium/+/537139 +Reviewed-by: Michal Klocek <michal.klocek@qt.io> +--- + .../rtc_peer_connection_handler.cc | 16 ++++++++++++---- + 1 file changed, 12 insertions(+), 4 deletions(-) + +diff --git a/chromium/third_party/blink/renderer/modules/peerconnection/rtc_peer_connection_handler.cc b/chromium/third_party/blink/renderer/modules/peerconnection/rtc_peer_connection_handler.cc +index 83853f003c7..fc2336dbb88 100644 +--- src/3rdparty/chromium/third_party/blink/renderer/modules/peerconnection/rtc_peer_connection_handler.cc ++++ src/3rdparty/chromium/third_party/blink/renderer/modules/peerconnection/rtc_peer_connection_handler.cc +@@ -1058,15 +1058,19 @@ bool RTCPeerConnectionHandler::Initialize( + WebLocalFrame* frame, + ExceptionState& exception_state) { + DCHECK(task_runner_->RunsTasksInCurrentSequence()); +- DCHECK(frame); + DCHECK(dependency_factory_); +- frame_ = frame; + + CHECK(!initialize_called_); + initialize_called_ = true; + + // Prevent garbage collection of client_ during processing. + auto* client_on_stack = client_; ++ if (!client_on_stack) { ++ return false; ++ } ++ ++ DCHECK(frame); ++ frame_ = frame; + peer_connection_tracker_ = PeerConnectionTracker::From(*frame); + + configuration_ = server_configuration; +@@ -2268,10 +2272,13 @@ void RTCPeerConnectionHandler::OnIceCandidate(const String& sdp, + int sdp_mline_index, + int component, + int address_family) { ++ DCHECK(task_runner_->RunsTasksInCurrentSequence()); + // In order to ensure that the RTCPeerConnection is not garbage collected + // from under the function, we keep a pointer to it on the stack. + auto* client_on_stack = client_; +- DCHECK(task_runner_->RunsTasksInCurrentSequence()); ++ if (!client_on_stack) { ++ return; ++ } + TRACE_EVENT0("webrtc", "RTCPeerConnectionHandler::OnIceCandidateImpl"); + // This line can cause garbage collection. + auto* platform_candidate = MakeGarbageCollected<RTCIceCandidatePlatform>( +@@ -2281,7 +2288,8 @@ void RTCPeerConnectionHandler::OnIceCandidate(const String& sdp, + this, platform_candidate, PeerConnectionTracker::kSourceLocal, true); + } + +- if (!is_closed_) ++ client_on_stack = client_; ++ if (!is_closed_ && client_on_stack) + client_on_stack->DidGenerateICECandidate(platform_candidate); + } + +From 149e8c185ff1ea7ee0a7037153311b026e142ac3 Mon Sep 17 00:00:00 2001 +From: John Stiles <johnstiles@google.com> +Date: Mon, 29 Jan 2024 23:50:14 +0000 +Subject: [PATCH] [Backport] CVE-2024-1283: Heap buffer overflow in Skia + +Manual cherry-pick of patch originally reviewed on +https://chromium-review.googlesource.com/c/chromium/src/+/5241305: +Fix a crash when a BMP image contains an unnecessary EOF code. + +Previously, this would try to perform color correction on a row +one past the end of the image data. + +Bug: 1521893 +Change-Id: I425437005b9ef400138556705616095857d2cf0d +Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5241305 +Auto-Submit: John Stiles <johnstiles@google.com> +Commit-Queue: John Stiles <johnstiles@google.com> +Reviewed-by: Peter Kasting <pkasting@chromium.org> +Cr-Commit-Position: refs/heads/main@{#1253633} +Reviewed-on: https://codereview.qt-project.org/c/qt/qtwebengine-chromium/+/538110 +Reviewed-by: Michal Klocek <michal.klocek@qt.io> +--- + .../image-decoders/bmp/bmp_image_reader.cc | 17 ++++++++++++++--- + 1 file changed, 14 insertions(+), 3 deletions(-) + +diff --git a/chromium/third_party/blink/renderer/platform/image-decoders/bmp/bmp_image_reader.cc b/chromium/third_party/blink/renderer/platform/image-decoders/bmp/bmp_image_reader.cc +index 063e5385d7f6..b40c8aa5c1fe 100644 +--- src/3rdparty/chromium/third_party/blink/renderer/platform/image-decoders/bmp/bmp_image_reader.cc ++++ src/3rdparty/chromium/third_party/blink/renderer/platform/image-decoders/bmp/bmp_image_reader.cc +@@ -827,8 +827,10 @@ BMPImageReader::ProcessingResult BMPImageReader::ProcessRLEData() { + // the image. + const uint8_t count = ReadUint8(0); + const uint8_t code = ReadUint8(1); +- if ((count || (code != 1)) && PastEndOfImage(0)) ++ const bool is_past_end_of_image = PastEndOfImage(0); ++ if ((count || (code != 1)) && is_past_end_of_image) { + return kFailure; ++ } + + // Decode. + if (!count) { +@@ -849,7 +851,9 @@ BMPImageReader::ProcessingResult BMPImageReader::ProcessRLEData() { + (is_top_down_ ? (coord_.y() < (parent_->Size().height() - 1)) + : (coord_.y() > 0))) + buffer_->SetHasAlpha(true); +- ColorCorrectCurrentRow(); ++ if (!is_past_end_of_image) { ++ ColorCorrectCurrentRow(); ++ } + // There's no need to move |coord_| here to trigger the caller + // to call SetPixelsChanged(). If the only thing that's changed + // is the alpha state, that will be properly written into the +@@ -1061,6 +1065,13 @@ void BMPImageReader::ColorCorrectCurrentRow() { + const ColorProfileTransform* const transform = parent_->ColorTransform(); + if (!transform) + return; ++ int decoder_width = parent_->Size().width(); ++ // Enforce 0 ≤ current row < bitmap height. ++ CHECK_GE(coord_.y(), 0); ++ CHECK_LT(coord_.y(), buffer_->Bitmap().height()); ++ // Enforce decoder width == bitmap width exactly. (The bitmap rowbytes might ++ // add a bit of padding, but we are only converting one row at a time.) ++ CHECK_EQ(decoder_width, buffer_->Bitmap().width()); + ImageFrame::PixelData* const row = buffer_->GetAddr(0, coord_.y()); + const skcms_PixelFormat fmt = XformColorFormat(); + const skcms_AlphaFormat alpha = +@@ -1069,7 +1080,7 @@ void BMPImageReader::ColorCorrectCurrentRow() { + : skcms_AlphaFormat_Unpremul; + const bool success = + skcms_Transform(row, fmt, alpha, transform->SrcProfile(), row, fmt, alpha, +- transform->DstProfile(), parent_->Size().width()); ++ transform->DstProfile(), decoder_width); + DCHECK(success); + buffer_->SetPixelsChanged(true); + } +From 707f4e7c0110c33df3d36a1942ad1b0ea2cb997b Mon Sep 17 00:00:00 2001 +From: Ken Rockot <rockot@google.com> +Date: Fri, 26 Jan 2024 21:53:06 +0000 +Subject: [PATCH] [Backport] CVE-2024-1284: Use after free in Mojo + +Cherry-pick of patch originally reviewed on +https://chromium-review.googlesource.com/c/chromium/src/+/5240312: +ipcz: Fix a few weak asserts + +DriverMemory cloning should not weakly assert success, as it can fail in +real production scenarios. Now Clone() will return an invalid +DriverMemory object if it fails to duplicate the internal handle. +Existing callers of Clone() are already durable to an invalid output, so +this change results in graceful failures instead of undefined behavior. + +This also replaces some weak asserts in DriverTransport creation with +hardening asserts. We may want to fail more gracefully if these end +up crashing a lot, but it seems unlikely. + +Fixed: 1521571 +Change-Id: Id764b33ead8bbba58e61b3270920c839479eaa4a +Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5240312 +Commit-Queue: Ken Rockot <rockot@google.com> +Reviewed-by: Alex Gough <ajgo@chromium.org> +Cr-Commit-Position: refs/heads/main@{#1252882} +Reviewed-on: https://codereview.qt-project.org/c/qt/qtwebengine-chromium/+/538111 +Reviewed-by: Michal Klocek <michal.klocek@qt.io> +--- + .../third_party/ipcz/src/ipcz/driver_memory.cc | 15 +++++++++------ + .../third_party/ipcz/src/ipcz/driver_transport.cc | 4 ++-- + 2 files changed, 11 insertions(+), 8 deletions(-) + +diff --git a/chromium/third_party/ipcz/src/ipcz/driver_memory.cc b/chromium/third_party/ipcz/src/ipcz/driver_memory.cc +index 612eca89d52..b92c04bf521 100644 +--- src/3rdparty/chromium/third_party/ipcz/src/ipcz/driver_memory.cc ++++ src/3rdparty/chromium/third_party/ipcz/src/ipcz/driver_memory.cc +@@ -30,10 +30,11 @@ DriverMemory::DriverMemory(const IpczDriver& driver, size_t num_bytes) + : size_(num_bytes) { + ABSL_ASSERT(num_bytes > 0); + IpczDriverHandle handle; +- IpczResult result = ++ const IpczResult result = + driver.AllocateSharedMemory(num_bytes, IPCZ_NO_FLAGS, nullptr, &handle); +- ABSL_ASSERT(result == IPCZ_RESULT_OK); +- memory_ = DriverObject(driver, handle); ++ if (result == IPCZ_RESULT_OK) { ++ memory_ = DriverObject(driver, handle); ++ } + } + + DriverMemory::DriverMemory(DriverMemory&& other) = default; +@@ -43,12 +44,14 @@ DriverMemory& DriverMemory::operator=(DriverMemory&& other) = default; + DriverMemory::~DriverMemory() = default; + + DriverMemory DriverMemory::Clone() { +- ABSL_ASSERT(is_valid()); ++ ABSL_HARDENING_ASSERT(is_valid()); + + IpczDriverHandle handle; +- IpczResult result = memory_.driver()->DuplicateSharedMemory( ++ const IpczResult result = memory_.driver()->DuplicateSharedMemory( + memory_.handle(), 0, nullptr, &handle); +- ABSL_ASSERT(result == IPCZ_RESULT_OK); ++ if (result != IPCZ_RESULT_OK) { ++ return DriverMemory(); ++ } + + return DriverMemory(DriverObject(*memory_.driver(), handle)); + } +diff --git a/chromium/third_party/ipcz/src/ipcz/driver_transport.cc b/chromium/third_party/ipcz/src/ipcz/driver_transport.cc +index a8cb7a1251f..2550c2891fd 100644 +--- src/3rdparty/chromium/third_party/ipcz/src/ipcz/driver_transport.cc ++++ src/3rdparty/chromium/third_party/ipcz/src/ipcz/driver_transport.cc +@@ -68,14 +68,14 @@ DriverTransport::Pair DriverTransport::CreatePair( + IpczDriverHandle target_transport0 = IPCZ_INVALID_DRIVER_HANDLE; + IpczDriverHandle target_transport1 = IPCZ_INVALID_DRIVER_HANDLE; + if (transport0) { +- ABSL_ASSERT(transport1); ++ ABSL_HARDENING_ASSERT(transport1); + target_transport0 = transport0->driver_object().handle(); + target_transport1 = transport1->driver_object().handle(); + } + IpczResult result = driver.CreateTransports( + target_transport0, target_transport1, IPCZ_NO_FLAGS, nullptr, + &new_transport0, &new_transport1); +- ABSL_ASSERT(result == IPCZ_RESULT_OK); ++ ABSL_HARDENING_ASSERT(result == IPCZ_RESULT_OK); + auto first = + MakeRefCounted<DriverTransport>(DriverObject(driver, new_transport0)); + auto second =
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?202402102017.41AKHEp0019595>