/* ----------
* toast_fetch_datum_slice -
*
* Reconstruct a segment of a Datum from the chunks saved
* in the toast relation
* ----------
*/
static struct varlena *
toast_fetch_datum_slice(struct varlena * attr, int32 sliceoffset, int32 length)
{
Relation toastrel;
Relation toastidx;
ScanKeyData toastkey[3];
int nscankeys;
SysScanDesc toastscan;
HeapTuple ttup;
TupleDesc toasttupDesc;
struct varlena *result;
struct varatt_external toast_pointer;
int32 attrsize;
int32 residx;
int32 nextidx;
int numchunks;
int startchunk;
int endchunk;
int32 startoffset;
int32 endoffset;
int totalchunks;
Pointer chunk;
bool isnull;
char *chunkdata;
int32 chunksize;
int32 chcpystrt;
int32 chcpyend;
Assert(VARATT_IS_EXTERNAL(attr));
/* Must copy to access aligned fields */
VARATT_EXTERNAL_GET_POINTER(toast_pointer, attr);
/*
* It's nonsense to fetch slices of a compressed datum -- this isn't lo_*
* we can't return a compressed datum which is meaningful to toast later
*/
Assert(!VARATT_EXTERNAL_IS_COMPRESSED(toast_pointer));
attrsize = toast_pointer.va_extsize;
totalchunks = ((attrsize - 1) / TOAST_MAX_CHUNK_SIZE) + 1;
if (sliceoffset >= attrsize)
{
sliceoffset = 0;
length = 0;
}
if (((sliceoffset + length) > attrsize) || length < 0)
length = attrsize - sliceoffset;
result = (struct varlena *) palloc(length + VARHDRSZ);
if (VARATT_EXTERNAL_IS_COMPRESSED(toast_pointer))
SET_VARSIZE_COMPRESSED(result, length + VARHDRSZ);
else
SET_VARSIZE(result, length + VARHDRSZ);
if (length == 0)
return result; /* Can save a lot of work at this point! */
startchunk = sliceoffset / TOAST_MAX_CHUNK_SIZE;
endchunk = (sliceoffset + length - 1) / TOAST_MAX_CHUNK_SIZE;
numchunks = (endchunk - startchunk) + 1;
startoffset = sliceoffset % TOAST_MAX_CHUNK_SIZE;
endoffset = (sliceoffset + length - 1) % TOAST_MAX_CHUNK_SIZE;
/*
* Open the toast relation and its index
*/
toastrel = heap_open(toast_pointer.va_toastrelid, AccessShareLock);
toasttupDesc = toastrel->rd_att;
toastidx = index_open(toastrel->rd_rel->reltoastidxid, AccessShareLock);
/*
* Setup a scan key to fetch from the index. This is either two keys or
* three depending on the number of chunks.
*/
ScanKeyInit(&toastkey[0],
(AttrNumber) 1,
BTEqualStrategyNumber, F_OIDEQ,
ObjectIdGetDatum(toast_pointer.va_valueid));
/*
* Use equality condition for one chunk, a range condition otherwise:
*/
if (numchunks == 1)
{
ScanKeyInit(&toastkey[1],
(AttrNumber) 2,
BTEqualStrategyNumber, F_INT4EQ,
Int32GetDatum(startchunk));
nscankeys = 2;
}
else
{
ScanKeyInit(&toastkey[1],
(AttrNumber) 2,
BTGreaterEqualStrategyNumber, F_INT4GE,
Int32GetDatum(startchunk));
ScanKeyInit(&toastkey[2],
(AttrNumber) 2,
BTLessEqualStrategyNumber, F_INT4LE,
Int32GetDatum(endchunk));
nscankeys = 3;
}
/*
* Read the chunks by index
*
* The index is on (valueid, chunkidx) so they will come in order
*/
nextidx = startchunk;
toastscan = systable_beginscan_ordered(toastrel, toastidx,
SnapshotToast, nscankeys, toastkey);
while ((ttup = systable_getnext_ordered(toastscan, ForwardScanDirection)) != NULL)
{
/*
* Have a chunk, extract the sequence number and the data
*/
residx = DatumGetInt32(fastgetattr(ttup, 2, toasttupDesc, &isnull));
Assert(!isnull);
chunk = DatumGetPointer(fastgetattr(ttup, 3, toasttupDesc, &isnull));
Assert(!isnull);
if (!VARATT_IS_EXTENDED(chunk))
{
chunksize = VARSIZE(chunk) - VARHDRSZ;
chunkdata = VARDATA(chunk);
}
else if (VARATT_IS_SHORT(chunk))
{
/* could happen due to heap_form_tuple doing its thing */
chunksize = VARSIZE_SHORT(chunk) - VARHDRSZ_SHORT;
chunkdata = VARDATA_SHORT(chunk);
}
else
{
/* should never happen */
elog(ERROR, "found toasted toast chunk for toast value %u in %s",
toast_pointer.va_valueid,
RelationGetRelationName(toastrel));
chunksize = 0; /* keep compiler quiet */
chunkdata = NULL;
}
/*
* Some checks on the data we've found
*/
if ((residx != nextidx) || (residx > endchunk) || (residx < startchunk))
elog(ERROR, "unexpected chunk number %d (expected %d) for toast value %u in %s",
residx, nextidx,
toast_pointer.va_valueid,
RelationGetRelationName(toastrel));
if (residx < totalchunks - 1)
{
if (chunksize != TOAST_MAX_CHUNK_SIZE)
elog(ERROR, "unexpected chunk size %d (expected %d) in chunk %d of %d for toast value %u in %s when fetching slice",
chunksize, (int) TOAST_MAX_CHUNK_SIZE,
residx, totalchunks,
toast_pointer.va_valueid,
RelationGetRelationName(toastrel));
}
else if (residx == totalchunks - 1)
{
if ((residx * TOAST_MAX_CHUNK_SIZE + chunksize) != attrsize)
elog(ERROR, "unexpected chunk size %d (expected %d) in final chunk %d for toast value %u in %s when fetching slice",
chunksize,
(int) (attrsize - residx * TOAST_MAX_CHUNK_SIZE),
residx,
toast_pointer.va_valueid,
RelationGetRelationName(toastrel));
}
else
elog(ERROR, "unexpected chunk number %d (out of range %d..%d) for toast value %u in %s",
residx,
0, totalchunks - 1,
toast_pointer.va_valueid,
RelationGetRelationName(toastrel));
/*
* Copy the data into proper place in our result
*/
chcpystrt = 0;
chcpyend = chunksize - 1;
if (residx == startchunk)
chcpystrt = startoffset;
if (residx == endchunk)
chcpyend = endoffset;
memcpy(VARDATA(result) +
(residx * TOAST_MAX_CHUNK_SIZE - sliceoffset) + chcpystrt,
chunkdata + chcpystrt,
(chcpyend - chcpystrt) + 1);
nextidx++;
}
/*
* Final checks that we successfully fetched the datum
*/
if (nextidx != (endchunk + 1))
elog(ERROR, "missing chunk number %d for toast value %u in %s",
nextidx,
toast_pointer.va_valueid,
RelationGetRelationName(toastrel));
/*
* End scan and close relations
*/
systable_endscan_ordered(toastscan);
index_close(toastidx, AccessShareLock);
heap_close(toastrel, AccessShareLock);
return result;
}
/* ----------
* pglz_compress -
*
* Compresses source into dest using strategy.
* ----------
*/
bool
pglz_compress(const char *source, int32 slen, PGLZ_Header *dest,
const PGLZ_Strategy *strategy)
{
unsigned char *bp = ((unsigned char *) dest) + sizeof(PGLZ_Header);
unsigned char *bstart = bp;
int hist_next = 1;
bool hist_recycle = false;
const char *dp = source;
const char *dend = source + slen;
unsigned char ctrl_dummy = 0;
unsigned char *ctrlp = &ctrl_dummy;
unsigned char ctrlb = 0;
unsigned char ctrl = 0;
bool found_match = false;
int32 match_len;
int32 match_off;
int32 good_match;
int32 good_drop;
int32 result_size;
int32 result_max;
int32 need_rate;
int hashsz;
int mask;
/*
* Our fallback strategy is the default.
*/
if (strategy == NULL)
strategy = PGLZ_strategy_default;
/*
* If the strategy forbids compression (at all or if source chunk size out
* of range), fail.
*/
if (strategy->match_size_good <= 0 ||
slen < strategy->min_input_size ||
slen > strategy->max_input_size)
return false;
/*
* Save the original source size in the header.
*/
dest->rawsize = slen;
/*
* Limit the match parameters to the supported range.
*/
good_match = strategy->match_size_good;
if (good_match > PGLZ_MAX_MATCH)
good_match = PGLZ_MAX_MATCH;
else if (good_match < 17)
good_match = 17;
good_drop = strategy->match_size_drop;
if (good_drop < 0)
good_drop = 0;
else if (good_drop > 100)
good_drop = 100;
need_rate = strategy->min_comp_rate;
if (need_rate < 0)
need_rate = 0;
else if (need_rate > 99)
need_rate = 99;
/*
* Compute the maximum result size allowed by the strategy, namely the
* input size minus the minimum wanted compression rate. This had better
* be <= slen, else we might overrun the provided output buffer.
*/
if (slen > (INT_MAX / 100))
{
/* Approximate to avoid overflow */
result_max = (slen / 100) * (100 - need_rate);
}
else
result_max = (slen * (100 - need_rate)) / 100;
/*
* Experiments suggest that these hash sizes work pretty well. A large
* hash table minimizes collision, but has a higher startup cost. For
* a small input, the startup cost dominates. The table size must be
* a power of two.
*/
if (slen < 128)
hashsz = 512;
else if (slen < 256)
hashsz = 1024;
else if (slen < 512)
hashsz = 2048;
else if (slen < 1024)
hashsz = 4096;
else
hashsz = 8192;
mask = hashsz - 1;
/*
* Initialize the history lists to empty. We do not need to zero the
* hist_entries[] array; its entries are initialized as they are used.
*/
memset(hist_start, 0, hashsz * sizeof(int16));
/*
* Compress the source directly into the output buffer.
*/
while (dp < dend)
{
/*
* If we already exceeded the maximum result size, fail.
*
* We check once per loop; since the loop body could emit as many as 4
* bytes (a control byte and 3-byte tag), PGLZ_MAX_OUTPUT() had better
* allow 4 slop bytes.
*/
if (bp - bstart >= result_max)
return false;
/*
* If we've emitted more than first_success_by bytes without finding
* anything compressible at all, fail. This lets us fall out
* reasonably quickly when looking at incompressible input (such as
* pre-compressed data).
*/
if (!found_match && bp - bstart >= strategy->first_success_by)
return false;
/*
* Try to find a match in the history
*/
if (pglz_find_match(hist_start, dp, dend, &match_len,
&match_off, good_match, good_drop, mask))
{
/*
* Create the tag and add history entries for all matched
* characters.
*/
pglz_out_tag(ctrlp, ctrlb, ctrl, bp, match_len, match_off);
while (match_len--)
{
pglz_hist_add(hist_start, hist_entries,
hist_next, hist_recycle,
dp, dend, mask);
dp++; /* Do not do this ++ in the line above! */
/* The macro would do it four times - Jan. */
}
found_match = true;
}
else
{
/*
* No match found. Copy one literal byte.
*/
pglz_out_literal(ctrlp, ctrlb, ctrl, bp, *dp);
pglz_hist_add(hist_start, hist_entries,
hist_next, hist_recycle,
dp, dend, mask);
dp++; /* Do not do this ++ in the line above! */
/* The macro would do it four times - Jan. */
}
}
/*
* Write out the last control byte and check that we haven't overrun the
* output size allowed by the strategy.
*/
*ctrlp = ctrlb;
result_size = bp - bstart;
if (result_size >= result_max)
return false;
/*
* Success - need only fill in the actual length of the compressed datum.
*/
SET_VARSIZE_COMPRESSED(dest, result_size + sizeof(PGLZ_Header));
return true;
}
/* ----------
* toast_fetch_datum -
*
* Reconstruct an in memory Datum from the chunks saved
* in the toast relation
* ----------
*/
static struct varlena *
toast_fetch_datum(struct varlena * attr)
{
Relation toastrel;
Relation toastidx;
ScanKeyData toastkey;
SysScanDesc toastscan;
HeapTuple ttup;
TupleDesc toasttupDesc;
struct varlena *result;
struct varatt_external toast_pointer;
int32 ressize;
int32 residx,
nextidx;
int32 numchunks;
Pointer chunk;
bool isnull;
char *chunkdata;
int32 chunksize;
/* Must copy to access aligned fields */
VARATT_EXTERNAL_GET_POINTER(toast_pointer, attr);
ressize = toast_pointer.va_extsize;
numchunks = ((ressize - 1) / TOAST_MAX_CHUNK_SIZE) + 1;
result = (struct varlena *) palloc(ressize + VARHDRSZ);
if (VARATT_EXTERNAL_IS_COMPRESSED(toast_pointer))
SET_VARSIZE_COMPRESSED(result, ressize + VARHDRSZ);
else
SET_VARSIZE(result, ressize + VARHDRSZ);
/*
* Open the toast relation and its index
*/
toastrel = heap_open(toast_pointer.va_toastrelid, AccessShareLock);
toasttupDesc = toastrel->rd_att;
toastidx = index_open(toastrel->rd_rel->reltoastidxid, AccessShareLock);
/*
* Setup a scan key to fetch from the index by va_valueid
*/
ScanKeyInit(&toastkey,
(AttrNumber) 1,
BTEqualStrategyNumber, F_OIDEQ,
ObjectIdGetDatum(toast_pointer.va_valueid));
/*
* Read the chunks by index
*
* Note that because the index is actually on (valueid, chunkidx) we will
* see the chunks in chunkidx order, even though we didn't explicitly ask
* for it.
*/
nextidx = 0;
toastscan = systable_beginscan_ordered(toastrel, toastidx,
SnapshotToast, 1, &toastkey);
while ((ttup = systable_getnext_ordered(toastscan, ForwardScanDirection)) != NULL)
{
/*
* Have a chunk, extract the sequence number and the data
*/
residx = DatumGetInt32(fastgetattr(ttup, 2, toasttupDesc, &isnull));
Assert(!isnull);
chunk = DatumGetPointer(fastgetattr(ttup, 3, toasttupDesc, &isnull));
Assert(!isnull);
if (!VARATT_IS_EXTENDED(chunk))
{
chunksize = VARSIZE(chunk) - VARHDRSZ;
chunkdata = VARDATA(chunk);
}
else if (VARATT_IS_SHORT(chunk))
{
/* could happen due to heap_form_tuple doing its thing */
chunksize = VARSIZE_SHORT(chunk) - VARHDRSZ_SHORT;
chunkdata = VARDATA_SHORT(chunk);
}
else
{
/* should never happen */
elog(ERROR, "found toasted toast chunk for toast value %u in %s",
toast_pointer.va_valueid,
RelationGetRelationName(toastrel));
chunksize = 0; /* keep compiler quiet */
chunkdata = NULL;
}
/*
* Some checks on the data we've found
*/
if (residx != nextidx)
elog(ERROR, "unexpected chunk number %d (expected %d) for toast value %u in %s",
residx, nextidx,
toast_pointer.va_valueid,
RelationGetRelationName(toastrel));
if (residx < numchunks - 1)
{
if (chunksize != TOAST_MAX_CHUNK_SIZE)
elog(ERROR, "unexpected chunk size %d (expected %d) in chunk %d of %d for toast value %u in %s",
chunksize, (int) TOAST_MAX_CHUNK_SIZE,
residx, numchunks,
toast_pointer.va_valueid,
RelationGetRelationName(toastrel));
}
else if (residx == numchunks - 1)
{
if ((residx * TOAST_MAX_CHUNK_SIZE + chunksize) != ressize)
elog(ERROR, "unexpected chunk size %d (expected %d) in final chunk %d for toast value %u in %s",
chunksize,
(int) (ressize - residx * TOAST_MAX_CHUNK_SIZE),
residx,
toast_pointer.va_valueid,
RelationGetRelationName(toastrel));
}
else
elog(ERROR, "unexpected chunk number %d (out of range %d..%d) for toast value %u in %s",
residx,
0, numchunks - 1,
toast_pointer.va_valueid,
RelationGetRelationName(toastrel));
/*
* Copy the data into proper place in our result
*/
memcpy(VARDATA(result) + residx * TOAST_MAX_CHUNK_SIZE,
chunkdata,
chunksize);
nextidx++;
}
/*
* Final checks that we successfully fetched the datum
*/
if (nextidx != numchunks)
elog(ERROR, "missing chunk number %d for toast value %u in %s",
nextidx,
toast_pointer.va_valueid,
RelationGetRelationName(toastrel));
/*
* End scan and close relations
*/
systable_endscan_ordered(toastscan);
index_close(toastidx, AccessShareLock);
heap_close(toastrel, AccessShareLock);
return result;
}
/* ----------
* pglz_compress -
*
* Compresses source into dest using strategy.
* ----------
*/
bool
pglz_compress(const char *source, int32 slen, PGLZ_Header *dest,
const PGLZ_Strategy *strategy)
{
unsigned char *bp = ((unsigned char *) dest) + sizeof(PGLZ_Header);
unsigned char *bstart = bp;
int hist_next = 0;
bool hist_recycle = false;
const char *dp = source;
const char *dend = source + slen;
unsigned char ctrl_dummy = 0;
unsigned char *ctrlp = &ctrl_dummy;
unsigned char ctrlb = 0;
unsigned char ctrl = 0;
int32 match_len;
int32 match_off;
int32 good_match;
int32 good_drop;
int32 result_size;
int32 result_max;
int32 need_rate;
/*
* Our fallback strategy is the default.
*/
if (strategy == NULL)
strategy = PGLZ_strategy_default;
/*
* If the strategy forbids compression (at all or if source chunk too
* small), fail.
*/
if (strategy->match_size_good <= 0 ||
slen < strategy->min_input_size)
return false;
/*
* Save the original source size in the header.
*/
dest->rawsize = slen;
/*
* Limit the match size to the maximum implementation allowed value
*/
if ((good_match = strategy->match_size_good) > PGLZ_MAX_MATCH)
good_match = PGLZ_MAX_MATCH;
if (good_match < 17)
good_match = 17;
if ((good_drop = strategy->match_size_drop) < 0)
good_drop = 0;
if (good_drop > 100)
good_drop = 100;
/*
* Initialize the history lists to empty. We do not need to zero the
* hist_entries[] array; its entries are initialized as they are used.
*/
memset((void *) hist_start, 0, sizeof(hist_start));
/*
* Compute the maximum result size allowed by the strategy. If the input
* size exceeds force_input_size, the max result size is the input size
* itself. Otherwise, it is the input size minus the minimum wanted
* compression rate.
*/
if (slen >= strategy->force_input_size)
result_max = slen;
else
{
need_rate = strategy->min_comp_rate;
if (need_rate < 0)
need_rate = 0;
else if (need_rate > 99)
need_rate = 99;
result_max = slen - ((slen * need_rate) / 100);
}
/*
* Compress the source directly into the output buffer.
*/
while (dp < dend)
{
/*
* If we already exceeded the maximum result size, fail.
*
* We check once per loop; since the loop body could emit as many as 4
* bytes (a control byte and 3-byte tag), PGLZ_MAX_OUTPUT() had better
* allow 4 slop bytes.
*/
if (bp - bstart >= result_max)
return false;
/*
* Try to find a match in the history
*/
if (pglz_find_match(hist_start, dp, dend, &match_len,
&match_off, good_match, good_drop))
{
/*
* Create the tag and add history entries for all matched
* characters.
*/
pglz_out_tag(ctrlp, ctrlb, ctrl, bp, match_len, match_off);
while (match_len--)
{
pglz_hist_add(hist_start, hist_entries,
hist_next, hist_recycle,
dp, dend);
dp++; /* Do not do this ++ in the line above! */
/* The macro would do it four times - Jan. */
}
}
else
{
/*
* No match found. Copy one literal byte.
*/
pglz_out_literal(ctrlp, ctrlb, ctrl, bp, *dp);
pglz_hist_add(hist_start, hist_entries,
hist_next, hist_recycle,
dp, dend);
dp++; /* Do not do this ++ in the line above! */
/* The macro would do it four times - Jan. */
}
}
/*
* Write out the last control byte and check that we haven't overrun the
* output size allowed by the strategy.
*/
*ctrlp = ctrlb;
result_size = bp - bstart;
if (result_size >= result_max)
return false;
/*
* Success - need only fill in the actual length of the compressed datum.
*/
SET_VARSIZE_COMPRESSED(dest, result_size + sizeof(PGLZ_Header));
return true;
}
