This patch teaches pbtombstone to use llvm-symbolizer to symbolize
stack traces and augment the protobuf tombstones with the symbol
information, before printing tombstones with the symbolized stack
traces included.
The main advantage of adding this information to the tombstone
as opposed to having developers use the stack tool is that stack
does not print all of the information in the original tombstone,
which means that both reports may be required to understand a crash.
Furthermore, stack traces printed by stack are not correlated with
the stack traces in the tombstone, making the report harder to read,
especially with GWP-ASan and MTE which may produce multiple stack
traces for the crashing thread.
Although we could teach stack to print more information, this would
continue to be fragile because stack relies on parsing textual
tombstones. Switching stack to read proto tombstones would be
tantamount to a full rewrite and would require duplicating the C++
proto-to-text logic that we already have in Python. It seems better
to reuse the C++ code for the proto-based symbolization tool.
llvm-symbolizer will look up the symbol files by build ID using a
.build-id directory following the standard here:
https://fedoraproject.org/wiki/RolandMcGrath/BuildID
It will look for .build-id directories under paths specified with
--debug-file-directory, which pbtombstone will pass through to
llvm-symbolizer using its own --debug-file-directory flag. The
intent is that tools for platform developers will pass the flag
--debug-file-directory $ANDROID_PRODUCT_OUT/symbols to pbtombstone.
Soong will start creating .build-id under symbols after a corresponding
Soong CL lands.
Bug: 328531087
Change-Id: Ia4676821cf980c69487cf11aefa2a02dc0c1626f
This is preparation for the next patch, which adds host-side
symbolization capabilities to pbtombstone.
Bug: 328531087
Change-Id: Id5813ae6b121af784643b1ed76084e49fdca118b
When adding a dependencies, if the variants don't match, but the
dependency only has 1 variant anyways, soong will always use that
variant. This makes it hard to add new variants to soong, because the
1-variant fallback stops being used and you start getting missing
variant errors. Make changes to bp files such that all dependencies
correctly specify the variant to use.
Bug: 372091092
Flag: EXEMPT refactor
Test: m nothing
Change-Id: I35a5ac0e6b63080f692be6597edf5f213c4e7acb
When dumping, the code checks if a thread is running as a guest.
If it's not, the code always logs an error message if the header
data structure cannot be read. Don't print an error message in
this case.
Test: Crash and no longer see "failed to get the guest state header"
Test: message.
Change-Id: I6971b57b218a4f653682cfdc82118dc576d9e331
We have fairly fancy detection of stack pointer issues, but nothing for the program counter...
Bug: https://issuetracker.google.com/365602121
Change-Id: I7e7061a1b473089a0f31438df5ebc65b283475bc
There is more than one function that can allocate in the fallback
path. Therefore, make sure that all functions that can allocate have
switched to the fallback linker allocator before allocating. This
is mostly a problem for the trace_handler function call, which iterates
over all known threads and gets backtraces for them.
Add a ScopedUseFallbackAllocator class to do the switching to make it
easier to have the same code everywhere.
Add two tests to verify that no allocations are occuring during
a tombstone or backtrace on the fallback path. These tests are not
comprehensive since they can't verify that the linker allocator is
using the fallback allocator, but they are better than nothing.
Remove the debuggerd_fallback_tombstone() function since it only
enables the linker callback and then calls engrave_tombstone_ucontext().
The enabling is now done with the ScopedUseLinkerAllocator.
Restructure the case where sending the signal to a thread fails so
that the fallback allocator can be enabled properly.
Bug: 359692763
Test: Ran debuggerd -b <PID> and debuggerd <PID> on a process that
Test: goes throught the fallback path.
Test: Unit tests pass along with two new tests.
Test: Forced a fallback process to crash and verified tombstone generated.
Test: Instrumented the linker allocator and verified that the trace_handler
Test: function never calls the normal linker allocator as it runs.
Change-Id: I2710921076634eac97f41bec8c3a29c1d75ae5ec
The kTagGranuleSize is only used in aarch64, so you can get
unused variable warnings for other targets. Move this definition
closer to where it's used and in aarch64 code.
Test: Build with unused checking on.
Change-Id: I9baf660430508ee3ca1efb62585ee1b038c6b0f3
We don't need this any more because the glibc headers have been updated, but it looks like it wasn't being used anyway.
Change-Id: Ie98ae473af504c49a8cae07fa06ec7f1568f662d
We will change the symbolizer to use this information to output
something like:
Potentially referenced stack object:
0 bytes inside a stack variable "variableName" in stack frame of function functionName
at source.cc:1234
Bug: 309446520
Change-Id: I1163ac81ac6b5e184387eb9e058d97a7227e3671
The timeout has to be determined experimentally. Generally, it must be
high enough to at least be the next instruction, and can be otherwise as
low as performance reasons allow.
This feature is for debugging only.
Test: atest PermissiveMteTest
Bug: 309604766
Change-Id: I54eff23374ebb239fd75b3b59ae72a7c33654454
For the debuggerd_test.cpp tests, don't include any libdebuggerd
internal headers.
Test: Compiles and unit tests pass.
Change-Id: I6335dc9e09992f5725a24fc45a8356efab7158e6
... so that we can confirm that it's actually the poll() call that's
timing out.
Bug: 332593241
Test: Treehugger
Change-Id: I529be76a268d7ba1f7f26a953eb84945f3ac4924
Guest thread information will print out follow host thread.
Revert submission 3081452-revert-3062926-CJGHTRPCBP
Reason for revert: Will make the change base on the original CLs for a reland.
Bug: b/321799516
Test: riscv64, checked tombstone file has wanted block.
https://paste.googleplex.com/6282302317658112
Added arm64 support and tested arm64 unwinding in internal repo.
https://paste.googleplex.com/6545612887818240
Change-Id: Ie54ad6f359d60283442adfcd9ee95f5a116e4b72
Revert submission 3081452-revert-3062926-CJGHTRPCBP
Reason for revert: Will make the change base on the original CLs for a reland.
(Original CL commit message)
This CL is to get guest registers information.
Bug: b/321799516
Test: m
Testing for TLS Slot:
Manual testing by: 1. crash the jni tests to produce tombstones file 2.
get the signature field of guest state header 3. verified it is the same
value as NATIVE_BRIDGE_GUEST_STATE_SIGNATURE
Manual test the arm64 by: 1. flash build to pixel phone and verify
retrieving TLS_SLOT_THREAD_ID's tid field is the same as current thread
id.
Testing for register values:
Test and print out registers values for riscv64, looks make sense that
has null zero value slots.
Change-Id: Ieebf845bff517380ee07fac77f24b48efeb53521
Revert submission 3062926
Reason for revert: We want guest state to be present in all threads - revert to be able to fix the proto field type.
Reverted changes: /q/submissionid:3062926
Change-Id: I32b745cca95a619b78bdce0a7d948ac479d42f21
Revert submission 3062926
Reason for revert: We want guest state to be present in all threads - revert to be able to fix the proto field type.
Reverted changes: /q/submissionid:3062926
Change-Id: I87b282a0d9caebe4eae2e7d8eca8ec8ebaa3eca6
This CL is to get guest registers information.
Bug: b/321799516
Test: m
Testing for TLS Slot:
Manual testing by: 1. crash the jni tests to produce tombstones file 2.
get the signature field of guest state header 3. verified it is the same
value as NATIVE_BRIDGE_GUEST_STATE_SIGNATURE
Manual test the arm64 by: 1. flash build to pixel phone and verify
retrieving TLS_SLOT_THREAD_ID's tid field is the same as current thread
id.
Testing for register values:
Test and print out registers values for riscv64, looks make sense that
has null zero value slots.
Change-Id: Iff44ac5c2b202e44f3fb4e6909fbea141e54ae6b
Rather than have to create a number of #if defines for the memory
dumping parts of the tombstone, create a single function to generate
these strings for the memory tests.
Make CrasherTest.smoke use a regex that passes on 32 bit and 64 bit.
Make the tests page size agnostic.
Bug: 339017792
Test: Treehugger.
Test: Ran 32 bit and 64 bit versions of tests on a real device.
Test: Ran on the aosp_cf_x86_64_phone_pgagnostic-trunk_staging-userdebug
Change-Id: If9365061b85de23b00a1bf947d85923cde06c068
Now that Android devices can use 16k page size, it's important that we
know what configuration the device is in when see issues.
1) If the device is in 4k mode, we see nothing new.
2) If the device is in 16k mode, we see this line in the tombstones:
Page size: 16
3) If the device is in 4k mode, but was previous in 16k mode we see:
Has been in 16kb mode: yes
Test: atest debuggerd_test
Test: atest debuggerd_test with ro.misctrl.16kb_before="1"
Test: adb shell cat /data/tombstones/tombstone_00
Bug: 335247092
Change-Id: If7ca3b0954a01070ff413758296460ca1d023ca5
Add signo for target thread in fallback path;
Update test for seccomp tombstone thread abort.
Bug: 336946834
Test: debuggerd_test
Test: Send fatal signal to process with NO_NEW_PRIVS
Change-Id: Ie9d77a93da9cd89ab7093b8949f311e03d96ec50
The malloc_not_svelte variable name is confusing and makes the
low memory config the default. Change this so that the default is
the regular allocator, and that Malloc_low_memory is used to enable
the low memory allocator.
Update blueprint rules so that scudo is the default action.
Test: Verified scudo config is used by default.
Test: Verfified Android GO config uses the jemalloc low memory config.
Change-Id: Ie7b4b005a6377e2a031bbae979d66b50c8b3bcdb
Tests starting failing on aarch64, so force presubmit to run the
debuggerd tests on device to close this coverage hole.
Test: Treehugger and verify tests are run on device.
Change-Id: Ie40eab35e46cefd3cebc132d9c0a3d1ebdb0d766
