[llvm-dev] RFC: Sanitizer-based Heap Profiler

Hal Finkel via llvm-dev llvm-dev at lists.llvm.org
Wed Jun 24 18:26:41 PDT 2020


Hi, Teresa,

This looks really useful.

It seems like this general infrastructure could be useful for diagnosing 
places where we have a lot of false sharing too. This could be between 
cores, sockets, or devices. Looking forward, especially with HMM and 
support under Linux for transparent unified memory between CPUs and 
accelerators, I anticipate we'll end up looking for places where some 
data accesses are causing undesired page migrations. Between cores and 
sockets, the interesting granularity is probably a cache line (and it 
looks like this is what you have anyway). This seems similar to me, the 
only difference being that we need to increment the count, not on every 
access, but only when the relevant entity id is different from the last 
one that accessed the relevant memory region. I'm not saying that we 
need to support this, but I'm curious what you think about this 
potential use case.

  -Hal

On 6/24/20 6:57 PM, Teresa Johnson via llvm-dev wrote:
> Hi all,
>
> I've included an RFC for a heap profiler design I've been working on 
> in conjunction with David Li. Please send any questions or feedback. 
> For sanitizer folks, one area of feedback is on refactoring some of 
> the *ASAN shadow setup code (see the Shadow Memory section).
>
> Thanks,
> Teresa
>
> RFC: Sanitizer-based Heap Profiler
>
>
>   Summary
>
> This document provides an overview of an LLVM Sanitizer-based heap 
> profiler design.
>
>
>   Motivation
>
> The objective of heap memory profiling is to collect critical runtime 
> information associated with heap memory references and information on 
> heap memory allocations. The profile information will be used first 
> for tooling, and subsequently to guide the compiler optimizer and 
> allocation runtime to layout heap objects with improved spatial 
> locality. As a  result, DTLB and cache utilization will be improved, 
> and program IPC (performance) will be increased due to reduced TLB and 
> cache misses. More details on the heap profile guided optimizations 
> will be shared in the future.
>
>
>   Overview
>
> The profiler is based on compiler inserted instrumentation of load and 
> store accesses, and utilizes runtime support to monitor heap 
> allocations and profile data. The target consumer of the heap memory 
> profile information is initially tooling and ultimately automatic data 
> layout optimizations performed by the compiler and/or allocation 
> runtime (with the support of new allocation runtime APIs).
>
>
> Each memory address is mapped to Shadow Memory 
> <https://en.wikipedia.org/wiki/Shadow_memory>, similar to the approach 
> used by the Address Sanitizer 
> <https://github.com/google/sanitizers/wiki/AddressSanitizer>(ASAN). 
> Unlike ASAN, which maps each 8 bytes of memory to 1 byte of shadow, 
> the heap profiler maps 64 bytes of memory to 8 bytes of shadow. The 
> shadow location implements the profile counter (incremented on 
> accesses to the corresponding memory). This granularity was chosen to 
> help avoid counter overflow, but it may be possible to consider 
> mapping 32-bytes to 4 bytes. To avoid aliasing of shadow memory for 
> different allocations, we must choose a minimum alignment carefully. 
> As discussed further below, we can attain a 32-byte minimum alignment, 
> instead of a 64-byte alignment, by storing necessary heap information 
> for each allocation in a 32-byte header block.
>
>
> The compiler instruments each load and store to increment the 
> associated shadow memory counter, in order to determine hotness.
>
>
> The heap profiler runtime is responsible for tracking allocations and 
> deallocations, including the stack at each allocation, and information 
> such as the allocation size and other statistics. I have implemented a 
> prototype built using a stripped down and modified version of ASAN, 
> however this will be a separate library utilizing sanitizer_common 
> components.
>
>
>     Compiler
>
> A simple HeapProfiler instrumentation pass instruments interesting 
> memory accesses (loads, stores, atomics), with a simple load, 
> increment, store of the associated shadow memory location (computed 
> via a mask and shift to do the mapping of 64 bytes to 8 byte shadow, 
> and add of the shadow offset). The handling is very similar to and 
> based off of the ASAN instrumentation pass, with slightly different 
> instrumentation code.
>
>
> Various techniques can be used to reduce the overhead, by aggressively 
> coalescing counter updates (e.g. given the 32-byte alignment, accesses 
> known to be in the same 32-byte block, or across possible aliases 
> since we don’t care about the dereferenced values).
>
>
> Additionally, the Clang driver needs to set up to link with the 
> runtime library, much as it does with the sanitizers.
>
>
> A -fheapprofoption is added to enable the instrumentation pass and 
> runtime library linking. Similar to -fprofile-generate, -fheapprofwill 
> accept an argument specifying the directory in which to write the profile.
>
>
>     Runtime
>
> The heap profiler runtime is responsible for tracking and reporting 
> information about heap allocations and accesses, aggregated by 
> allocation calling context. For example, the hotness, lifetime, and 
> cpu affinity.
>
>
> A new heapprof library will be created within compiler-rt. It will 
> leverage support within sanitizer_common, which already contains 
> facilities like stack context tracking, needed by the heap profiler.
>
>
>       Shadow Memory
>
> There are some basic facilities in sanitizer_common for mmap’ing the 
> shadow memory, but most of the existing setup lives in the ASAN and 
> HWASAN libraries. In the case of ASAN, there is support for both 
> statically assigned shadow offsets (the default on most platforms), 
> and for dynamically assigned shadow memory (implemented for Windows 
> and currently also used for Android and iOS). According to kcc, recent 
> experiments show that the performance with a dynamic shadow is close 
> to that with a static mapping. In fact, that is the only approach 
> currently used by HWASAN. Given the simplicity, the heap profiler will 
> be implemented with a dynamic shadow as well.
>
>
> There are a number of functions in ASAN and HWASAN related to setup of 
> the shadow that are duplicated but very nearly identical, at least for 
> linux (which seems to be the only OS flavor currently supported for 
> HWASAN). E.g. ReserveShadowMemoryRange, ProtectGap, and 
> FindDynamicShadowStart (in ASAN there is another nearly identical copy 
> in PremapShadow, used by Android, whereas in HW ASAN the premap 
> handling is already commoned with the non-premap handling). Rather 
> than make yet another copy of these mechanisms, I propose refactoring 
> them into sanitizer_common versions. Like HWASAN, the initial version 
> of the heap profiler will be supported for linux only, but other OSes 
> can be added as needed similar to ASAN.
>
>
>       StackTrace and StackDepot
>
> The sanitizer already contains support for obtaining and representing 
> a stack trace in a StackTrace object, and storing it in the StackDepot 
> which “efficiently stores huge amounts of stack traces”. This is in 
> the sanitizer_common subdirectory and the support is shared by ASAN 
> and ThreadSanitizer. The StackDepot is essentially an unbounded hash 
> table, where each StackTrace is assigned a unique id. ASAN stores this 
> id in the alloc_context_id field in each ChunkHeader (in the redzone 
> preceding each allocation). Additionally, there is support for 
> symbolizing and printing StackTrace objects.
>
>
>       ChunkHeader
>
> The heap profiler needs to track several pieces of information for 
> each allocation. Given the mapping of 64-bytes to 8-bytes shadow, we 
> can achieve a minimum of 32-byte alignment by holding this information 
> in a 32-byte header block preceding each allocation.
>
>
> In ASAN, each allocation is preceded by a 16-byte ChunkHeader. It 
> contains information about the current allocation state, user 
> requested size, allocation and free thread ids, the allocation context 
> id (representing the call stack at allocation, assigned by the 
> StackDepot as described above), and misc other bookkeeping. For heap 
> profiling, this will be converted to a 32-byte header block.
>
>
> Note that we could instead use the metadata section, similar to other 
> sanitizers, which is stored in a separate location. However, as 
> described above, storing the header block with each allocation enables 
> 32-byte alignment without aliasing shadow counters for the same 64 
> bytes of memory.
>
>
> In the prototype heap profiler implementation, the header contains the 
> following fields:
>
>
> // Should be 32 bytes
>
> struct ChunkHeader {
>
>   // 1-st 4 bytes
>
>   // Carry over from ASAN (available, allocated, quarantined). Will be
>
>   // reduced to 1 bit (available or allocated).
>
>   u32 chunk_state       : 8;
>
>   // Carry over from ASAN. Used to determine the start of user allocation.
>
>   u32 from_memalign     : 1;
>
>   // 23 bits available
>
>
>   // 2-nd 4 bytes
>
>   // Carry over from ASAN (comment copied verbatim).
>
>   // This field is used for small sizes. For large sizes it is equal to
>
>   // SizeClassMap::kMaxSize and the actual size is stored in the
>
>   // SecondaryAllocator's metadata.
>
>   u32 user_requested_size : 29;
>
>
>   // 3-rd 4 bytes
>
>   u32 cpu_id; // Allocation cpu id
>
>
>   // 4-th 4 bytes
>
>   // Allocation timestamp in ms from a baseline timestamp computed at
>
>   // the start of profiling (to keep this within 32 bits).
>
>   u32 timestamp_ms;
>
>
>   // 5-th and 6-th 4 bytes
>
>   // Carry over from ASAN. Used to identify allocation stack trace.
>
>   u64 alloc_context_id;
>
>
>   // 7-th and 8-th 4 bytes
>
>   // UNIMPLEMENTED in prototype - needs instrumentation and IR support.
>
>   u64 data_type_id; // hash of type name
>
> };
>
> As noted, the chunk state can be reduced to a single bit (no need for 
> quarantined memory in the heap profiler). The header contains a 
> placeholder for the data type hash, which is not yet implemented as it 
> needs instrumentation and IR support.
>
>
>       Heap Info Block (HIB)
>
> On a deallocation, information from the corresponding shadow block(s) 
> and header are recorded in a Heap Info Block (HIB) object. The access 
> count is computed from the shadow memory locations for the allocation, 
> as well as the percentage of accessed 64-byte blocks (i.e. the 
> percentage of non-zero 8-byte shadow locations for the whole 
> allocation). Other information such as the deallocation timestamp (for 
> lifetime computation) and deallocation cpu id (to determine 
> migrations) are recorded along with the information in the chunk 
> header recorded on allocation.
>
>
> The prototyped HIB object tracks the following:
>
>
> struct HeapInfoBlock {
>
>   // Total allocations at this stack context
>
>   u32 alloc_count;
>
>   // Access count computed from all allocated 64-byte blocks (track total
>
>   // across all allocations, and the min and max).
>
>   u64 total_access_count, min_access_count, max_access_count;
>
>   // Allocated size (track total across all allocations, and the min 
> and max).
>
>   u64 total_size;
>
>   u32 min_size, max_size;
>
>   // Lifetime (track total across all allocations, and the min and max).
>
>   u64 total_lifetime;
>
>   u32 min_lifetime, max_lifetime;
>
>   // Percent utilization of allocated 64-byte blocks (track total
>
>   // across all allocations, and the min and max). The utilization is
>
>   // defined as the percentage of 8-byte shadow counters corresponding to
>
>   // the full allocation that are non-zero.
>
>   u64 total_percent_utilized;
>
>   u32 min_percent_utilized, max_percent_utilized;
>
>   // Allocation and deallocation timestamps from the most recent merge 
> into
>
>   // the table with this stack context.
>
>   u32 alloc_timestamp, dealloc_timestamp;
>
>   // Allocation and deallocation cpu ids from the most recent merge into
>
>   // the table with this stack context.
>
>   u32 alloc_cpu_id, dealloc_cpu_id;
>
>   // Count of allocations at this stack context that had a different
>
>   // allocation and deallocation cpu id.
>
>   u32 num_migrated_cpu;
>
>   // Number of times the lifetime of the entry being merged had its 
> lifetime
>
>   // overlap with the previous entry merged with this stack context (by
>
>   // comparing the new alloc/dealloc timestamp with the one last 
> recorded in
>
>   // the entry in the table.
>
>   u32 num_lifetime_overlaps;
>
>   // Number of times the alloc/dealloc cpu of the entry being merged 
> was the
>
>   // same as that of the previous entry merged with this stack context
>
>   u32 num_same_alloc_cpu;
>
>   u32 num_same_dealloc_cpu;
>
>   // Hash of type name (UNIMPLEMENTED). This needs instrumentation 
> support and
>
>   // possibly IR changes.
>
>   u64 data_type_id;
>
> }
>
>
>       HIB Table
>
> The Heap Info Block Table, which is a multi-way associative cache, 
> holds HIB objects from deallocated objects. It is indexed by the stack 
> allocation context id from the chunk header, and currently utilizes a 
> simple mod with a prime number close to a power of two as the hash 
> (because of the way the stack context ids are assigned, a mod of a 
> power of two performs very poorly). Thus far, only 4-way associativity 
> has been evaluated.
>
>
> HIB entries are added or merged into the HIB Table on each 
> deallocation. If an entry with a matching stack alloc context id is 
> found in the Table, the newly deallocated information is merged into 
> the existing entry. Each HIB Table entry currently tracks the min, max 
> and total value of the various fields for use in computing and 
> reporting the min, max and average when the Table is ultimately dumped.
>
>
> If no entry with a matching stack alloc context id is found, a new 
> entry is created. If this causes an eviction, the evicted entry is 
> dumped immediately (by default to stderr, otherwise to a specified 
> report file). Later post processing can merge dumped entries with the 
> same stack alloc context id.
>
>
>       Initialization
>
>
> For ASAN, an __asan_init function initializes the memory allocation 
> tracking support, and the ASAN instrumentation pass in LLVM creates a 
> global constructor to invoke it. The heap profiler prototype adds a 
> new __heapprof_init function, which performs heap profile specific 
> initialization, and the heap profile instrumentation pass calls this 
> new init function instead by a generated global constructor. It 
> currently additionally invokes __asan_init since we are leveraging a 
> modified ASAN runtime. Eventually, this should be changed to 
> initialize refactored common support.
>
>
> Note that __asan init is also placed in the .preinit_array when it is 
> available, so it is invoked even earlier than global constructors. 
> Currently, it is not possible to do this for __heapprof_init, as it 
> calls timespec_get in order to get a baseline timestamp (as described 
> in the ChunkHeader comments the timestamps (ms) are actually offsets 
> from the baseline timestamp, in order to fit into 32 bits), and system 
> calls cannot be made that early (dl_init is not complete). Since the 
> constructor priority is 1, it should be executed early enough that 
> there are very few allocations before it runs, and likely the best 
> solution is to simply ignore any allocations before initialization.
>
>
>       Dumping
>
> For the prototype, the profile is dumped as text with a compact raw 
> format to limit its size. Ultimately it should be dumped in a more 
> compact binary format (i.e. into a different section of the raw 
> instrumentation based profile, with llvm-profdata performing 
> post-processing) which is TBD.
>
>
>         HIB Dumping
>
> As noted earlier, HIB Table entries are created as memory is 
> deallocated. At the end of the run (or whenever dumping is requested, 
> discussed later), HIB entries need to be created for allocations that 
> are still live. Conveniently, the sanitizer allocator already contains 
> a mechanism to walk through all chunks of memory it is tracking 
> (ForEachChunk). The heap profiler simply looks for all chunks with a 
> chunk state of allocated, and creates a HIB the same as would be done 
> on deallocation, adding each to the table.
>
>
> A HIB Table mechanism for printing each entry is then invoked.
>
>
> By default, the dumping occurs:
>
>  *
>
>     on evictions
>
>  *
>
>     full table at exit (when the static Allocator object is destructed)
>
>
> For running in a load testing scenario, we will want to add a 
> mechanism to provoke finalization (merging currently live allocations) 
> and dumping of the HIB Table before exit. This would be similar to the 
> __llvm_profile_dump facility used for normal PGO counter dumping.
>
>
>         Stack Trace Dumping
>
> There is existing support for dumping symbolized StackTrace objects. A 
> wrapper to dump all StackTrace objects in the StackDepot will be 
> added. This new interface is invoked just after the HIB Table is 
> dumped (on exit or via dumping interface).
>
>
>         Memory Map Dumping
>
> In cases where we may want to symbolize as a post processing step, we 
> may need the memory map (from /proc/self/smaps). Specifically, this is 
> needed to symbolize binaries using ASLR (Address Space Layout 
> Randomization). There is already support for reading this file and 
> dumping it to the specified report output file (DumpProcessMap()). 
> This is invoked when the profile output file is initialized (HIB Table 
> construction), so that the memory map is available at the top of the 
> raw profile.
>
>
>   Current Status and Next Steps
>
>
> As mentioned earlier, I have a working prototype based on a simplified 
> stripped down version of ASAN. My current plan is to do the following:
>
> 1.
>
>     Refactor out some of the shadow setup code common between ASAN and
>     HWASAN into sanitizer_common.
>
> 2.
>
>     Rework my prototype into a separate heapprof library in
>     compiler-rt, using sanitizer_common support where possible, and
>     send patches for review.
>
> 3.
>
>     Send patches for the heap profiler instrumentation pass and
>     related clang options.
>
> 4.
>
>     Design/implement binary profile format
>
>
> -- 
> Teresa Johnson | 	 Software Engineer | 	tejohnson at google.com 
> <mailto:tejohnson at google.com> | 	
>
>
> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org
> https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev

-- 
Hal Finkel
Lead, Compiler Technology and Programming Languages
Leadership Computing Facility
Argonne National Laboratory

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20200624/1dab1801/attachment-0001.html>


More information about the llvm-dev mailing list