llvm-6502/lib/Support/StringMap.cpp
Chandler Carruth 2a79116940 Fix PR13148, an inf-loop in StringMap.
StringMap suffered from the same bug as DenseMap: when you explicitly
construct it with a small number of buckets, you can arrange for the
tombstone-based growth path to be followed when the number of buckets
was less than '8'. In that case, even with a full map, it would compare
'0' as not less than '0', and refuse to grow the table, leading to
inf-loops trying to find an empty bucket on the next insertion. The fix
is very simple: use '<=' as the comparison. The same fix was applied to
DenseMap as well during its recent refactoring.

Thanks to Alex Bolz for the great report and test case. =]

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158725 91177308-0d34-0410-b5e6-96231b3b80d8
2012-06-19 17:40:35 +00:00

238 lines
8.2 KiB
C++

//===--- StringMap.cpp - String Hash table map implementation -------------===//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// This file implements the StringMap class.
//
//===----------------------------------------------------------------------===//
#include "llvm/ADT/StringMap.h"
#include "llvm/ADT/StringExtras.h"
#include <cassert>
using namespace llvm;
StringMapImpl::StringMapImpl(unsigned InitSize, unsigned itemSize) {
ItemSize = itemSize;
// If a size is specified, initialize the table with that many buckets.
if (InitSize) {
init(InitSize);
return;
}
// Otherwise, initialize it with zero buckets to avoid the allocation.
TheTable = 0;
NumBuckets = 0;
NumItems = 0;
NumTombstones = 0;
}
void StringMapImpl::init(unsigned InitSize) {
assert((InitSize & (InitSize-1)) == 0 &&
"Init Size must be a power of 2 or zero!");
NumBuckets = InitSize ? InitSize : 16;
NumItems = 0;
NumTombstones = 0;
TheTable = (StringMapEntryBase **)calloc(NumBuckets+1,
sizeof(StringMapEntryBase **) +
sizeof(unsigned));
// Allocate one extra bucket, set it to look filled so the iterators stop at
// end.
TheTable[NumBuckets] = (StringMapEntryBase*)2;
}
/// LookupBucketFor - Look up the bucket that the specified string should end
/// up in. If it already exists as a key in the map, the Item pointer for the
/// specified bucket will be non-null. Otherwise, it will be null. In either
/// case, the FullHashValue field of the bucket will be set to the hash value
/// of the string.
unsigned StringMapImpl::LookupBucketFor(StringRef Name) {
unsigned HTSize = NumBuckets;
if (HTSize == 0) { // Hash table unallocated so far?
init(16);
HTSize = NumBuckets;
}
unsigned FullHashValue = HashString(Name);
unsigned BucketNo = FullHashValue & (HTSize-1);
unsigned *HashTable = (unsigned *)(TheTable + NumBuckets + 1);
unsigned ProbeAmt = 1;
int FirstTombstone = -1;
while (1) {
StringMapEntryBase *BucketItem = TheTable[BucketNo];
// If we found an empty bucket, this key isn't in the table yet, return it.
if (BucketItem == 0) {
// If we found a tombstone, we want to reuse the tombstone instead of an
// empty bucket. This reduces probing.
if (FirstTombstone != -1) {
HashTable[FirstTombstone] = FullHashValue;
return FirstTombstone;
}
HashTable[BucketNo] = FullHashValue;
return BucketNo;
}
if (BucketItem == getTombstoneVal()) {
// Skip over tombstones. However, remember the first one we see.
if (FirstTombstone == -1) FirstTombstone = BucketNo;
} else if (HashTable[BucketNo] == FullHashValue) {
// If the full hash value matches, check deeply for a match. The common
// case here is that we are only looking at the buckets (for item info
// being non-null and for the full hash value) not at the items. This
// is important for cache locality.
// Do the comparison like this because Name isn't necessarily
// null-terminated!
char *ItemStr = (char*)BucketItem+ItemSize;
if (Name == StringRef(ItemStr, BucketItem->getKeyLength())) {
// We found a match!
return BucketNo;
}
}
// Okay, we didn't find the item. Probe to the next bucket.
BucketNo = (BucketNo+ProbeAmt) & (HTSize-1);
// Use quadratic probing, it has fewer clumping artifacts than linear
// probing and has good cache behavior in the common case.
++ProbeAmt;
}
}
/// FindKey - Look up the bucket that contains the specified key. If it exists
/// in the map, return the bucket number of the key. Otherwise return -1.
/// This does not modify the map.
int StringMapImpl::FindKey(StringRef Key) const {
unsigned HTSize = NumBuckets;
if (HTSize == 0) return -1; // Really empty table?
unsigned FullHashValue = HashString(Key);
unsigned BucketNo = FullHashValue & (HTSize-1);
unsigned *HashTable = (unsigned *)(TheTable + NumBuckets + 1);
unsigned ProbeAmt = 1;
while (1) {
StringMapEntryBase *BucketItem = TheTable[BucketNo];
// If we found an empty bucket, this key isn't in the table yet, return.
if (BucketItem == 0)
return -1;
if (BucketItem == getTombstoneVal()) {
// Ignore tombstones.
} else if (HashTable[BucketNo] == FullHashValue) {
// If the full hash value matches, check deeply for a match. The common
// case here is that we are only looking at the buckets (for item info
// being non-null and for the full hash value) not at the items. This
// is important for cache locality.
// Do the comparison like this because NameStart isn't necessarily
// null-terminated!
char *ItemStr = (char*)BucketItem+ItemSize;
if (Key == StringRef(ItemStr, BucketItem->getKeyLength())) {
// We found a match!
return BucketNo;
}
}
// Okay, we didn't find the item. Probe to the next bucket.
BucketNo = (BucketNo+ProbeAmt) & (HTSize-1);
// Use quadratic probing, it has fewer clumping artifacts than linear
// probing and has good cache behavior in the common case.
++ProbeAmt;
}
}
/// RemoveKey - Remove the specified StringMapEntry from the table, but do not
/// delete it. This aborts if the value isn't in the table.
void StringMapImpl::RemoveKey(StringMapEntryBase *V) {
const char *VStr = (char*)V + ItemSize;
StringMapEntryBase *V2 = RemoveKey(StringRef(VStr, V->getKeyLength()));
(void)V2;
assert(V == V2 && "Didn't find key?");
}
/// RemoveKey - Remove the StringMapEntry for the specified key from the
/// table, returning it. If the key is not in the table, this returns null.
StringMapEntryBase *StringMapImpl::RemoveKey(StringRef Key) {
int Bucket = FindKey(Key);
if (Bucket == -1) return 0;
StringMapEntryBase *Result = TheTable[Bucket];
TheTable[Bucket] = getTombstoneVal();
--NumItems;
++NumTombstones;
assert(NumItems + NumTombstones <= NumBuckets);
return Result;
}
/// RehashTable - Grow the table, redistributing values into the buckets with
/// the appropriate mod-of-hashtable-size.
void StringMapImpl::RehashTable() {
unsigned NewSize;
unsigned *HashTable = (unsigned *)(TheTable + NumBuckets + 1);
// If the hash table is now more than 3/4 full, or if fewer than 1/8 of
// the buckets are empty (meaning that many are filled with tombstones),
// grow/rehash the table.
if (NumItems*4 > NumBuckets*3) {
NewSize = NumBuckets*2;
} else if (NumBuckets-(NumItems+NumTombstones) <= NumBuckets/8) {
NewSize = NumBuckets;
} else {
return;
}
// Allocate one extra bucket which will always be non-empty. This allows the
// iterators to stop at end.
StringMapEntryBase **NewTableArray =
(StringMapEntryBase **)calloc(NewSize+1, sizeof(StringMapEntryBase *) +
sizeof(unsigned));
unsigned *NewHashArray = (unsigned *)(NewTableArray + NewSize + 1);
NewTableArray[NewSize] = (StringMapEntryBase*)2;
// Rehash all the items into their new buckets. Luckily :) we already have
// the hash values available, so we don't have to rehash any strings.
for (unsigned I = 0, E = NumBuckets; I != E; ++I) {
StringMapEntryBase *Bucket = TheTable[I];
if (Bucket && Bucket != getTombstoneVal()) {
// Fast case, bucket available.
unsigned FullHash = HashTable[I];
unsigned NewBucket = FullHash & (NewSize-1);
if (NewTableArray[NewBucket] == 0) {
NewTableArray[FullHash & (NewSize-1)] = Bucket;
NewHashArray[FullHash & (NewSize-1)] = FullHash;
continue;
}
// Otherwise probe for a spot.
unsigned ProbeSize = 1;
do {
NewBucket = (NewBucket + ProbeSize++) & (NewSize-1);
} while (NewTableArray[NewBucket]);
// Finally found a slot. Fill it in.
NewTableArray[NewBucket] = Bucket;
NewHashArray[NewBucket] = FullHash;
}
}
free(TheTable);
TheTable = NewTableArray;
NumBuckets = NewSize;
NumTombstones = 0;
}