llvm's JSON parser supports 64 bit integers, but other tools like the
ones written in JS don't support numbers that big, so we need to
represent these possibly big numbers as a string. This diff uses that to
represent addresses and tsc zero. The former is printed in hex for and
the latter in decimal string form. The schema was updated mentioning
that.
Besides that, I fixed some remaining issues and now all test pass. Before I wasn't running all tests because for some reason my computer reverted perf_paranoid to 1.
Differential Revision: https://reviews.llvm.org/D127819
As discusses offline with @jj10305, we are updating some naming used throughout the code, specially in the json schema
- traceBuffer -> iptTrace
- core -> cpu
Differential Revision: https://reviews.llvm.org/D127817
This applies the changes requested for diff 12.
- use DenseMap<ConstString, _> instead of std::unordered_map<ConstString, _>, which is more idiomatic and possibly performant.
- deduplicate some code in Trace.cpp by using helper functions for fetching in maps
- stop using size and offset when fetching binary data, because we in fact read the entire buffers all the time. If we ever need streaming, we can implement it then. Now, the size is used only to check that we are getting the correct amount of data. This is useful because in some cases determining the size doesn't involve fetching the actual data.
- added back the x86_64 macro to the perf tests
- added more documentation
- simplified some file handling
- fixed some comments
Differential Revision: https://reviews.llvm.org/D127752
This improves several things and addresses comments up to the diff [11] in this stack.
- Simplify many functions to receive less parameters that they can identify easily
- Create Storage classes for Trace and TraceIntelPT that can make it easier to reason about what can change with live process refreshes and what cannot.
- Don't cache the perf zero conversion numbers in lldb-server to make sure we get the most up-to-date numbers.
- Move the thread identifaction from context switches to the bundle parser, to leave TraceIntelPT simpler. This also makes sure that the constructor of TraceIntelPT is invoked when the entire data has been checked to be correct.
- Normalize all bundle paths before the Processes, Threads and Modules are created, so that they can assume that all paths are correct and absolute
- Fix some issues in the tests. Now they all pass.
- return the specific instance when constructing PerThread and MultiCore processor tracers.
- Properly implement IntelPTMultiCoreTrace::TraceStart.
- Improve some comments.
- Use the typedef ContextSwitchTrace more often for clarity.
- Move CreateContextSwitchTracePerfEvent to Perf.h as a utility function.
- Synchronize better the state of the context switch and the intel pt
perf events.
- Use a booblean instead of an enum for the PerfEvent state.
Differential Revision: https://reviews.llvm.org/D127456
- Add the logic that parses all cpu context switch traces and produces blocks of continuous executions, which will be later used to assign intel pt subtraces to threads and to identify gaps. This logic can also identify if the context switch trace is malformed.
- The continuous executions blocks are able to indicate when there were some contention issues when producing the context switch trace. See the inline comments for more information.
- Update the 'dump info' command to show information and stats related to the multicore decoding flow, including timing about context switch decoding.
- Add the logic to conver nanoseconds to TSCs.
- Fix a bug when returning the context switches. Now they data returned makes sense and even empty traces can be returned from lldb-server.
- Finish the necessary bits for loading and saving a multi-core trace bundle from disk.
- Change some size_t to uint64_t for compatibility with 32 bit systems.
Tested by saving a trace session of a program that sleeps 100 times, it was able to produce the following 'dump info' text:
```
(lldb) trace load /tmp/trace3/trace.json (lldb) thread trace dump info Trace technology: intel-pt
thread #1: tid = 4192415
Total number of instructions: 1
Memory usage:
Total approximate memory usage (excluding raw trace): 2.51 KiB
Average memory usage per instruction (excluding raw trace): 2573.00 bytes
Timing for this thread:
Timing for global tasks:
Context switch trace decoding: 0.00s
Events:
Number of instructions with events: 0
Number of individual events: 0
Multi-core decoding:
Total number of continuous executions found: 2499
Number of continuous executions for this thread: 102
Errors:
Number of TSC decoding errors: 0
```
Differential Revision: https://reviews.llvm.org/D126267
:q!
This diff is massive, but it's because it connects the client with lldb-server
and also ensures that the postmortem case works.
- Flatten the postmortem trace schema. The reason is that the schema has become quite complex due to the new multicore case, which defeats the original purpose of having a schema that could work for every trace plug-in. At this point, it's better that each trace plug-in defines it's own full schema. This means that the only common field is "type".
-- Because of this new approach, I merged the "common" trace load and saving functionalities into the IntelPT one. This simplified the code quite a bit. If we eventually implement another trace plug-in, we can see then what we could reuse.
-- The new schema, which is flattened, has now better comments and is parsed better. A change I did was to disallow hex addresses, because they are a bit error prone. I'm asking now to print the address in decimal.
-- Renamed "intel" to "GenuineIntel" in the schema because that's what you see in /proc/cpuinfo.
- Implemented reading the context switch trace data buffer. I had to do
some refactors to do that cleanly.
-- A major change that I did here was to simplify the perf_event circular buffer reading logic. It was too complex. Maybe the original Intel author had something different in mind.
- Implemented all the necessary bits to read trace.json files with per-core data.
- Implemented all the necessary bits to save to disk per-core trace session.
- Added a test that ensures that parsing and saving to disk works.
Differential Revision: https://reviews.llvm.org/D126015
- Add a warnings field in the jLLDBGetState response, for warnings to be delivered to the client for troubleshooting. This removes the need to silently log lldb-server's llvm::Errors and not expose them easily to the user
- Simplify the tscPerfZeroConversion struct and schema. It used to extend a base abstract class, but I'm doubting that we'll ever add other conversion mechanisms because all modern kernels support perf zero. It is also the one who is supposed to work with the timestamps produced by the context switch trace, so expecting it is imperative.
- Force tsc collection for cpu tracing.
- Add a test checking that tscPerfZeroConversion is returned by the GetState request
- Add a pre-check for cpu tracing that makes sure that perf zero values are available.
Differential Revision: https://reviews.llvm.org/D125932
- Add collection of context switches per cpu grouped with the per-cpu intel pt traces.
- Move the state handling from the interl pt trace class to the PerfEvent one.
- Add support for stopping and enabling perf event groups.
- Return context switch entries as part of the jLLDBTraceGetState response.
- Move the triggers of whenever the process stopped or resumed. Now the will-resume notification is in a better location, which will ensure that we'll capture the instructions that will be executed.
- Remove IntelPTSingleBufferTraceUP. The unique pointer was useless.
- Add unit tests
Differential Revision: https://reviews.llvm.org/D125897
We were setting some events to be written in the data buffer of the
perf_event, but we don't need that.
Besides that, we don't need the data buffer to be larger than 1, so we
can reduce its size.
Differential Revision: https://reviews.llvm.org/D125850
We have two different "process trace" implementations: per thread and per core. As a way to simplify the collector who uses both, I'm creating a base abstract class that is used by these implementations. This effectively simplify a good chunk of code.
Differential Revision: https://reviews.llvm.org/D125503
On Ubuntu 18.04 with GCC 7.5 Intel trace code fails to build due to
failure to convert from
lldb_private::process_linux::IntelPTPerThreadProcessTraceUP to
Expected<lldb_private::process_linux::IntelPTPerThreadProcessTraceUP>.
This commit explicitely marks those unique_ptr values as being moved
which fixes the conversion error.
Reviewed By: wallace
Differential Revision: https://reviews.llvm.org/D126402
IntelPTCollector is very big and has 3 classes in it. It's actually cleaner if each one has its own file. This also gives more visibility to the developer about the different kinds of "tracers" that we have.
Besides that, I'm now restricting the creation of the BinaryData chunks to GetState() instead of having it in different places, which is not very clean, because the gdb-remote protocol should be as restricted as possible.
Differential Revision: https://reviews.llvm.org/D125047
When tracing on per-core mode, we are tracing all processes, which means
that after hitting a breakpoint, our process will stop running (thus
producing no more tracing data) but other processes will continue
writing to our trace buffers. This causes a big data loss for our trace.
As a way to remediate this, I'm adding some logic to pause and unpause
tracing based on the target's state. The earlier we do it the better,
however, I'm not adding the trigger at the earliest possible point for
simplicity of this diff. Later we can improve that part.
Differential Revision: https://reviews.llvm.org/D124962
This diffs implements per-core tracing on lldb-server. It also includes tests that ensure that tracing can be initiated from the client and that the jLLDBGetState ppacket returns the list of trace buffers per core.
This doesn't include any decoder changes.
Finally, this makes some little changes here and there improving the existing code.
A specific piece of code that can't reliably be tested is when tracing
per core fails due to permissions. In this case we add a
troubleshooting message and this is the manual test:
```
/proc/sys/kernel/perf_event_paranoid set to 1
(lldb) process trace start --per-core-tracing error: perf event syscall failed: Permission denied
You might need that /proc/sys/kernel/perf_event_paranoid has a value of 0 or -1.
``
Differential Revision: https://reviews.llvm.org/D124858
llvm's json parser supports uint64_t, so let's better use it for the
packets being sent between lldb and lldb-server instead of using int64_t
as an intermediate type, which might be error-prone.
llvm's json parser supports uint64_t, so let's better use it for the
packets being sent between lldb and lldb-server instead of using int64_t
as an intermediate type, which might be error-prone.
I'm refactoring IntelPTThreadTrace into IntelPTSingleBufferTrace so that it can
both single threads or single cores. In this diff I'm basically renaming the
class, moving it to its own file, and removing all the pieces that are not used
along with some basic cleanup.
Differential Revision: https://reviews.llvm.org/D124648
This updates the documentation of the gdb-remote protocol, as well as the help messages, to include the new --per-core-tracing option.
Differential Revision: https://reviews.llvm.org/D124640
In order to open perf events per core, we need to first get the list of
core ids available in the system. So I'm adding a function that does
that by parsing /proc/cpuinfo. That seems to be the simplest and most
portable way to do that.
Besides that, I made a few refactors and renames to reflect better that
the cpu info that we use in lldb-server comes from procfs.
Differential Revision: https://reviews.llvm.org/D124573
This silences warnings like this:
lldb/source/Core/DebuggerEvents.cpp: In member function ‘llvm::StringRef lldb_private::DiagnosticEventData::GetPrefix() const’:
lldb/source/Core/DebuggerEvents.cpp:55:1: warning: control reaches end of non-void function [-Wreturn-type]
55 | }
Differential Revision: https://reviews.llvm.org/D123203
This patch handles the situation where the main thread exits (through
the SYS_exit syscall). In this case, the process as a whole continues
running until all of the other threads exit, or one of them issues an
exit_group syscall.
The patch consists of two changes:
- a moderate redesign of the handling of thread exit (WIFEXITED) events.
Previously, we were removing (forgetting) a thread once we received
the WIFEXITED (or WIFSIGNALED) event. This was problematic for the
main thread, since the main thread WIFEXITED event (which is better thought
of as a process-wide event) gets reported only after the entire process
exits. This resulted in deadlocks, where we were waiting for the
process to stop (because we still considered the main thread "live").
This patch changes the logic such that the main thread is removed as
soon as its PTRACE_EVENT_EXIT (the pre-exit) event is received. At
this point we can consider the thread gone (for most purposes). As a
corrolary, I needed to add special logic to catch process-wide exit
events in the cases where we don't have the main thread around.
- The second part of the patch is the removal of the assumptions that
the main thread is always available. This generally meant replacing
the uses of GetThreadByID(process_id) with GetCurrentThread() in
various process-wide operations (such as memory reads).
Differential Revision: https://reviews.llvm.org/D122716
Update the response schema of the TraceGetState packet and add
Intel PT specific response structure that contains the TSC conversion,
if it exists. The IntelPTCollector loads the TSC conversion and caches
it to prevent unnecessary calls to perf_event_open. Move the TSC conversion
calculation from Perf.h to TraceIntelPTGDBRemotePackets.h to remove
dependency on Linux specific headers.
Differential Revision: https://reviews.llvm.org/D122246
- Add PerfEvent class to handle creating ring buffers and handle the resources associated with a perf_event
- Refactor IntelPT collection code to use this new API
- Add TSC to timestamp conversion logic with unittest
Differential Revision: https://reviews.llvm.org/D121734
- Rename IntelPTManager class and files to IntelPTCollector
- Change GetTimestampCounter API to general trace counter API,
GetCounter
Differential Revision: https://reviews.llvm.org/D121711
Most of our code was including Log.h even though that is not where the
"lldb" log channel is defined (Log.h defines the generic logging
infrastructure). This worked because Log.h included Logging.h, even
though it should.
After the recent refactor, it became impossible the two files include
each other in this direction (the opposite inclusion is needed), so this
patch removes the workaround that was put in place and cleans up all
files to include the right thing. It also renames the file to LLDBLog to
better reflect its purpose.
This better describes the intent of the method. Which for AArch64
is removing the top byte which includes the memory tags.
It does not include pointer signatures, for those we need to use
the ABI plugin. The rename makes this a little more clear.
It's a bit awkward that the memory tag manager is removing the whole
top byte not just the memory tags but it's an improvement for now.
Reviewed By: omjavaid
Differential Revision: https://reviews.llvm.org/D117671
Implement the qXfer:siginfo:read that is used to read the siginfo_t
(extended signal information) for the current thread. This is currently
implemented on FreeBSD and Linux.
Differential Revision: https://reviews.llvm.org/D117113
D116372, while fixing one kind of a race, ended up creating a new one.
The new issue could occur when one inferior thread exits while another
thread initiates termination of the entire process (exit_group(2)).
With some bad luck, we could start processing the exit notification
(PTRACE_EVENT_EXIT) only to have the become unresponsive (ESRCH) in the
middle of the MonitorCallback function. This function would then delete
the thread from our list even though it wasn't completely dead (it stays
zombified until we read the WIFEXITED event). The linux kernel will not
deliver the exited event for the entire process until we process
individual thread exits.
In a pre-D116372 world, this wouldn't be a problem because we would read
this event (even though we would not know what to do with it) with
waitpid(-1). Now, when we issue invididual waitpids, this event will
never be picked up, and we end up hanging.
The fix for this is actually quite simple -- don't delete the thread in
this situation. The thread will be deleted when the WIFEXITED event
comes.
This situation was kind of already tested by
TestCreateDuringInstructionStep (which is how I found this problem), but
it was mostly accidental, so I am also creating a dedicated test which
reproduces this situation.
The lldb-server code is currently set up in a way that each
NativeProcess instance does its own waitpid handling. This works fine
for BSDs, where the code can do a waitpid(process_id), and get
information for all threads in that process.
The situation is trickier on linux, because waitpid(pid) will only
return information for the main thread of the process (one whose tid ==
pid). For this reason the linux code does a waitpid(-1), to get
information for all threads. This was fine while we were supporting just
a single process, but becomes a problem when we have multiple processes
as they end up stealing each others events.
There are two possible solutions to this problem:
- call waitpid(-1) centrally, and then dispatch the events to the
appropriate process
- have each process call waitpid(tid) for all the threads it manages
This patch implements the second approach. Besides fitting better into
the existing design, it also has the added benefit of ensuring
predictable ordering for thread/process creation events (which come in
pairs -- one for the parent and one for the child). The first approach
OTOH, would make this ordering even more complicated since we would
have to keep the half-threads hanging in mid-air until we find the
process we should attach them to.
The downside to this approach is an increased number of syscalls (one
waitpid for each thread), but I think we're pretty far from optimizing
things like this, and so the cleanliness of the design is worth it.
The included test reproduces the circumstances which should demonstrate
the bug (which manifests as a hung test), but I have not been able to
get it to fail. The only place I've seen this failure modes are very
rare hangs in the thread sanitizer tests (tsan forks an addr2line
process to produce its error messages).
Differential Revision: https://reviews.llvm.org/D116372
The MonitorCallback function was assuming that the "exited" argument is
set whenever a thread exits, but the caller was only setting that flag
for the main thread.
This patch deletes the argument altogether, and lets MonitorCallback
compute what it needs itself.
This is almost NFC, since previously we would end up in the
"GetSignalInfo failed for unknown reasons" branch, which was doing the
same thing -- forgetting about the thread.
Support arbitrarily-sized FPR writes on ARM in order to fix writing qN
registers directly. Currently, writing them works only by accident
due to value_regs splitting them into smaller writes via dN and sN
registers.
Differential Revision: https://reviews.llvm.org/D112131
.. and reduce the scope of others. They don't follow llvm coding
standards (which say they should be used only when the same effect
cannot be achieved with the static keyword), and they set a bad example.
Usage of aux_size is guarded against elsewhere in this file, but is missing here.
Reviewed By: wallace
Differential Revision: https://reviews.llvm.org/D110269
Original Author: calebzulawski
This patch fixes register save/restore on expression call to also include SVE registers.
This will fix expression calls like:
re re p1
<Register Value P1 before expression>
p <var-name or function call>
re re p1
<Register Value P1 after expression>
In above example register P1 should remain the same before and after the expression evaluation.
Reviewed By: DavidSpickett
Differential Revision: https://reviews.llvm.org/D108739
This patch removed some typos from NativeRegisterContextLinux_arm and
NativeRegisterContextLinux_arm64. Some of the log/error messages were
being reported as x86_64.
