/* ----------
* heap_tuple_untoast_attr -
*
* Public entry point to get back a toasted value from compression
* or external storage.
* ----------
*/
varattrib *
heap_tuple_untoast_attr(varattrib *attr)
{
varattrib *result;
if (VARATT_IS_EXTERNAL(attr))
{
if (VARATT_IS_COMPRESSED(attr))
{
/* ----------
* This is an external stored compressed value
* Fetch it from the toast heap and decompress.
* ----------
*/
varattrib *tmp;
tmp = toast_fetch_datum(attr);
result = (varattrib *) palloc(attr->va_content.va_external.va_rawsize
+ VARHDRSZ);
VARATT_SIZEP(result) = attr->va_content.va_external.va_rawsize
+ VARHDRSZ;
pglz_decompress((PGLZ_Header *) tmp, VARATT_DATA(result));
pfree(tmp);
}
else
{
/*
* This is an external stored plain value
*/
result = toast_fetch_datum(attr);
}
}
else if (VARATT_IS_COMPRESSED(attr))
{
/*
* This is a compressed value inside of the main tuple
*/
result = (varattrib *) palloc(attr->va_content.va_compressed.va_rawsize
+ VARHDRSZ);
VARATT_SIZEP(result) = attr->va_content.va_compressed.va_rawsize
+ VARHDRSZ;
pglz_decompress((PGLZ_Header *) attr, VARATT_DATA(result));
}
else
/*
* This is a plain value inside of the main tuple - why am I
* called?
*/
return attr;
return result;
}
/* ----------
* heap_tuple_untoast_attr -
*
* Public entry point to get back a toasted value from compression
* or external storage.
* ----------
*/
struct varlena *
heap_tuple_untoast_attr(struct varlena * attr)
{
if (VARATT_IS_EXTERNAL(attr))
{
/*
* This is an externally stored datum --- fetch it back from there
*/
attr = toast_fetch_datum(attr);
/* If it's compressed, decompress it */
if (VARATT_IS_COMPRESSED(attr))
{
PGLZ_Header *tmp = (PGLZ_Header *) attr;
attr = (struct varlena *) palloc(PGLZ_RAW_SIZE(tmp) + VARHDRSZ);
SET_VARSIZE(attr, PGLZ_RAW_SIZE(tmp) + VARHDRSZ);
pglz_decompress(tmp, VARDATA(attr));
pfree(tmp);
}
}
else if (VARATT_IS_COMPRESSED(attr))
{
/*
* This is a compressed value inside of the main tuple
*/
PGLZ_Header *tmp = (PGLZ_Header *) attr;
attr = (struct varlena *) palloc(PGLZ_RAW_SIZE(tmp) + VARHDRSZ);
SET_VARSIZE(attr, PGLZ_RAW_SIZE(tmp) + VARHDRSZ);
pglz_decompress(tmp, VARDATA(attr));
}
else if (VARATT_IS_SHORT(attr))
{
/*
* This is a short-header varlena --- convert to 4-byte header format
*/
Size data_size = VARSIZE_SHORT(attr) - VARHDRSZ_SHORT;
Size new_size = data_size + VARHDRSZ;
struct varlena *new_attr;
new_attr = (struct varlena *) palloc(new_size);
SET_VARSIZE(new_attr, new_size);
memcpy(VARDATA(new_attr), VARDATA_SHORT(attr), data_size);
attr = new_attr;
}
return attr;
}
/* ----------
* toast_compress_datum -
*
* Create a compressed version of a varlena datum
*
* If we fail (ie, compressed result is actually bigger than original)
* then return NULL. We must not use compressed data if it'd expand
* the tuple!
*
* We use VAR{SIZE,DATA}_ANY so we can handle short varlenas here without
* copying them. But we can't handle external or compressed datums.
* ----------
*/
Datum
toast_compress_datum(Datum value)
{
varattrib *tmp;
int32 valsize = VARSIZE_ANY_EXHDR_D(value);
Assert(!VARATT_IS_EXTERNAL(DatumGetPointer(value)));
Assert(!VARATT_IS_COMPRESSED(DatumGetPointer(value)));
/*
* No point in wasting a palloc cycle if value size is out of the allowed
* range for compression
*/
if (valsize < PGLZ_strategy_default->min_input_size ||
valsize > PGLZ_strategy_default->max_input_size)
return PointerGetDatum(NULL);
tmp = (varattrib *) palloc(PGLZ_MAX_OUTPUT(valsize));
if (pglz_compress(VARDATA_ANY_D(value), valsize,
(PGLZ_Header *) tmp, PGLZ_strategy_default) &&
VARSIZE(tmp) < VARSIZE_ANY_D(value))
{
/* successful compression */
VARATT_SET_COMPRESSED(tmp);
return PointerGetDatum(tmp);
}
else
{
/* incompressible data */
pfree(tmp);
return PointerGetDatum(NULL);
}
}
/* ----------
* toast_raw_datum_size -
*
* Return the raw (detoasted) size of a varlena datum
* (including the VARHDRSZ header)
* ----------
*/
Size
toast_raw_datum_size(Datum value)
{
struct varlena *attr = (struct varlena *) DatumGetPointer(value);
Size result;
if (VARATT_IS_EXTERNAL(attr))
{
/* va_rawsize is the size of the original datum -- including header */
struct varatt_external toast_pointer;
VARATT_EXTERNAL_GET_POINTER(toast_pointer, attr);
result = toast_pointer.va_rawsize;
}
else if (VARATT_IS_COMPRESSED(attr))
{
/* here, va_rawsize is just the payload size */
result = VARRAWSIZE_4B_C(attr) + VARHDRSZ;
}
else if (VARATT_IS_SHORT(attr))
{
/*
* we have to normalize the header length to VARHDRSZ or else the
* callers of this function will be confused.
*/
result = VARSIZE_SHORT(attr) - VARHDRSZ_SHORT + VARHDRSZ;
}
else
{
/* plain untoasted datum */
result = VARSIZE(attr);
}
return result;
}
/* ----------
* toast_raw_datum_size -
*
* Return the raw (detoasted) size of a varlena datum
* ----------
*/
Size
toast_raw_datum_size(Datum value)
{
varattrib *attr = (varattrib *) DatumGetPointer(value);
Size result;
if (VARATT_IS_COMPRESSED(attr))
{
/*
* va_rawsize shows the original data size, whether the datum is
* external or not.
*/
result = attr->va_content.va_compressed.va_rawsize + VARHDRSZ;
}
else if (VARATT_IS_EXTERNAL(attr))
{
/*
* an uncompressed external attribute has rawsize including the
* header (not too consistent!)
*/
result = attr->va_content.va_external.va_rawsize;
}
else
{
/* plain untoasted datum */
result = VARSIZE(attr);
}
return result;
}
/* ----------
* toast_raw_datum_size -
*
* Return the raw (detoasted) size of a varlena datum
* (including the VARHDRSZ header)
* ----------
*/
Size
toast_raw_datum_size(Datum value)
{
varattrib *attr = (varattrib *) DatumGetPointer(value);
Size result;
if (VARATT_IS_EXTERNAL(attr))
{
/* va_rawsize is the size of the original datum -- including header */
result = attr->va_external.va_rawsize;
}
else if (VARATT_IS_COMPRESSED(attr))
{
/* here, va_rawsize is just the payload size */
result = attr->va_compressed.va_rawsize + VARHDRSZ;
}
else if (VARATT_IS_SHORT(attr))
{
/*
* we have to normalize the header length to VARHDRSZ or else the
* callers of this function will be confused.
*/
result = VARSIZE_SHORT(attr) - VARHDRSZ_SHORT + VARHDRSZ;
}
else
{
/* plain untoasted datum */
result = VARSIZE(attr);
}
return result;
}
/* ----------
* heap_tuple_untoast_attr_slice -
*
* Public entry point to get back part of a toasted value
* from compression or external storage.
* ----------
*/
varattrib *
heap_tuple_untoast_attr_slice(varattrib *attr, int32 sliceoffset, int32 slicelength)
{
varattrib *preslice;
varattrib *result;
int32 attrsize;
if (VARATT_IS_COMPRESSED(attr))
{
varattrib *tmp;
if (VARATT_IS_EXTERNAL(attr))
tmp = toast_fetch_datum(attr);
else
{
tmp = attr; /* compressed in main tuple */
}
preslice = (varattrib *) palloc(attr->va_content.va_external.va_rawsize
+ VARHDRSZ);
VARATT_SIZEP(preslice) = attr->va_content.va_external.va_rawsize + VARHDRSZ;
pglz_decompress((PGLZ_Header *) tmp, VARATT_DATA(preslice));
if (tmp != attr)
pfree(tmp);
}
else
{
/* Plain value */
if (VARATT_IS_EXTERNAL(attr))
{
/* fast path */
return (toast_fetch_datum_slice(attr, sliceoffset, slicelength));
}
else
preslice = attr;
}
/* slicing of datum for compressed cases and plain value */
attrsize = VARSIZE(preslice) - VARHDRSZ;
if (sliceoffset >= attrsize)
{
sliceoffset = 0;
slicelength = 0;
}
if (((sliceoffset + slicelength) > attrsize) || slicelength < 0)
slicelength = attrsize - sliceoffset;
result = (varattrib *) palloc(slicelength + VARHDRSZ);
VARATT_SIZEP(result) = slicelength + VARHDRSZ;
memcpy(VARDATA(result), VARDATA(preslice) + sliceoffset, slicelength);
if (preslice != attr)
pfree(preslice);
return result;
}
/**
* If this function is changed then update varattrib_untoast_len as well
*/
void varattrib_untoast_ptr_len(Datum d, char **datastart, int *len, void **tofree)
{
if (DatumGetPointer(d) == NULL)
{
ereport(ERROR,
(errcode(ERRCODE_INTERNAL_ERROR),
errmsg(" Unable to detoast datum "),
errprintstack(true)));
}
struct varlena *va = (struct varlena *) DatumGetPointer(d);
varattrib *attr = (varattrib *) va;
*len = -1;
*tofree = NULL;
if(VARATT_IS_EXTENDED(attr))
{
if(VARATT_IS_EXTERNAL(attr))
{
attr = (varattrib *)toast_fetch_datum((struct varlena *)attr);
/* toast_fetch_datum will palloc, so set it up for free */
*tofree = attr;
}
if(VARATT_IS_COMPRESSED(attr))
{
PGLZ_Header *tmp = (PGLZ_Header *) attr;
attr = (varattrib *) palloc(PGLZ_RAW_SIZE(tmp) + VARHDRSZ);
SET_VARSIZE(attr, PGLZ_RAW_SIZE(tmp) + VARHDRSZ);
pglz_decompress(tmp, VARDATA(attr));
/* If tofree is set, that is, we get it from toast_fetch_datum.
* We need to free it here
*/
if(*tofree)
pfree(*tofree);
*tofree = attr;
}
else if(VARATT_IS_SHORT(attr))
{
/* Warning! Return unaligned pointer! */
*len = VARSIZE_SHORT(attr) - VARHDRSZ_SHORT;
*datastart = VARDATA_SHORT(attr);
attr = NULL;
}
}
if(*len == -1)
{
*datastart = VARDATA(attr);
*len = VARSIZE(attr) - VARHDRSZ;
}
Assert(*len >= 0);
}
/*
* Extract data from the pg_statistic arrays into useful format.
*/
static Selectivity
mcelem_tsquery_selec(TSQuery query, Datum *mcelem, int nmcelem,
float4 *numbers, int nnumbers)
{
float4 minfreq;
TextFreq *lookup;
Selectivity selec;
int i;
/*
* There should be two more Numbers than Values, because the last two
* cells are taken for minimal and maximal frequency. Punt if not.
*
* (Note: the MCELEM statistics slot definition allows for a third extra
* number containing the frequency of nulls, but we're not expecting that
* to appear for a tsvector column.)
*/
if (nnumbers != nmcelem + 2)
return tsquery_opr_selec_no_stats(query);
/*
* Transpose the data into a single array so we can use bsearch().
*/
lookup = (TextFreq *) palloc(sizeof(TextFreq) * nmcelem);
for (i = 0; i < nmcelem; i++)
{
/*
* The text Datums came from an array, so it cannot be compressed or
* stored out-of-line -- it's safe to use VARSIZE_ANY*.
*/
Assert(!VARATT_IS_COMPRESSED(mcelem[i]) && !VARATT_IS_EXTERNAL(mcelem[i]));
lookup[i].element = (text *) DatumGetPointer(mcelem[i]);
lookup[i].frequency = numbers[i];
}
/*
* Grab the lowest frequency. compute_tsvector_stats() stored it for us in
* the one before the last cell of the Numbers array. See ts_typanalyze.c
*/
minfreq = numbers[nnumbers - 2];
selec = tsquery_opr_selec(GETQUERY(query), GETOPERAND(query), lookup,
nmcelem, minfreq);
pfree(lookup);
return selec;
}
/**
* If this function is changed then update varattrib_untoast_ptr_len as well
*/
int varattrib_untoast_len(Datum d)
{
if (DatumGetPointer(d) == NULL)
{
ereport(ERROR,
(errcode(ERRCODE_INTERNAL_ERROR),
errmsg(" Unable to detoast datum "),
errprintstack(true)));
}
struct varlena *va = (struct varlena *) DatumGetPointer(d);
varattrib *attr = (varattrib *) va;
int len = -1;
void *toFree = NULL;
if(VARATT_IS_EXTENDED(attr))
{
if(VARATT_IS_EXTERNAL(attr))
{
attr = (varattrib *)toast_fetch_datum((struct varlena *)attr);
/* toast_fetch_datum will palloc, so set it up for free */
toFree = attr;
}
if(VARATT_IS_COMPRESSED(attr))
{
PGLZ_Header *tmp = (PGLZ_Header *) attr;
len = PGLZ_RAW_SIZE(tmp);
}
else if(VARATT_IS_SHORT(attr))
{
len = VARSIZE_SHORT(attr) - VARHDRSZ_SHORT;
}
}
if(len == -1)
{
len = VARSIZE(attr) - VARHDRSZ;
}
if ( toFree)
pfree(toFree);
Assert(len >= 0);
return len;
}
/* ----------
* toast_compress_datum -
*
* Create a compressed version of a varlena datum
*
* If we fail (ie, compressed result is actually bigger than original)
* then return NULL. We must not use compressed data if it'd expand
* the tuple!
*
* We use VAR{SIZE,DATA}_ANY so we can handle short varlenas here without
* copying them. But we can't handle external or compressed datums.
* ----------
*/
Datum
toast_compress_datum(Datum value)
{
struct varlena *tmp;
int32 valsize = VARSIZE_ANY_EXHDR(DatumGetPointer(value));
Assert(!VARATT_IS_EXTERNAL(DatumGetPointer(value)));
Assert(!VARATT_IS_COMPRESSED(DatumGetPointer(value)));
/*
* No point in wasting a palloc cycle if value size is out of the allowed
* range for compression
*/
if (valsize < PGLZ_strategy_default->min_input_size ||
valsize > PGLZ_strategy_default->max_input_size)
return PointerGetDatum(NULL);
tmp = (struct varlena *) palloc(PGLZ_MAX_OUTPUT(valsize));
/*
* We recheck the actual size even if pglz_compress() reports success,
* because it might be satisfied with having saved as little as one byte
* in the compressed data --- which could turn into a net loss once you
* consider header and alignment padding. Worst case, the compressed
* format might require three padding bytes (plus header, which is
* included in VARSIZE(tmp)), whereas the uncompressed format would take
* only one header byte and no padding if the value is short enough. So
* we insist on a savings of more than 2 bytes to ensure we have a gain.
*/
if (pglz_compress(VARDATA_ANY(DatumGetPointer(value)), valsize,
(PGLZ_Header *) tmp, PGLZ_strategy_default) &&
VARSIZE(tmp) < valsize - 2)
{
/* successful compression */
return PointerGetDatum(tmp);
}
else
{
/* incompressible data */
pfree(tmp);
return PointerGetDatum(NULL);
}
}
/* ----------
* heap_tuple_untoast_attr -
*
* Public entry point to get back a toasted value from compression
* or external storage.
* ----------
*/
struct varlena *
heap_tuple_untoast_attr(struct varlena *attr)
{
if (VARATT_IS_EXTERNAL(attr))
{
/*
* This is an externally stored datum --- fetch it back from there
*/
attr = toast_fetch_datum(attr);
/* fall through to IS_COMPRESSED if it's a compressed external datum */
}
if (VARATT_IS_COMPRESSED(attr))
{
/*
* This is a compressed value inside of the main tuple
*/
PGLZ_Header *tmp = (PGLZ_Header *) attr;
attr = (struct varlena *) palloc(PGLZ_RAW_SIZE(tmp) + VARHDRSZ);
SET_VARSIZE(attr, PGLZ_RAW_SIZE(tmp) + VARHDRSZ);
pglz_decompress(tmp, VARDATA(attr));
}
else if (VARATT_IS_SHORT(attr))
{
/*
* This is a short-header varlena --- convert to 4-byte header format
*/
Size data_size = VARSIZE_SHORT(attr);
Size new_size = data_size - VARHDRSZ_SHORT + VARHDRSZ;
varattrib *tmp = (varattrib *)attr;
/* This is a "short" varlena header but is otherwise a normal varlena */
attr = (struct varlena *) palloc(new_size);
SET_VARSIZE(attr, new_size);
memcpy(VARDATA(attr), VARDATA_SHORT(tmp), data_size - VARHDRSZ_SHORT);
}
return attr;
}
/* ----------
* toast_fetch_datum_slice -
*
* Reconstruct a segment of a varattrib from the chunks saved
* in the toast relation
* ----------
*/
static varattrib *
toast_fetch_datum_slice(varattrib *attr, int32 sliceoffset, int32 length)
{
Relation toastrel;
Relation toastidx;
ScanKeyData toastkey[3];
int nscankeys;
IndexScanDesc toastscan;
HeapTuple ttup;
TupleDesc toasttupDesc;
varattrib *result;
int32 attrsize;
int32 residx;
int32 nextidx;
int numchunks;
int startchunk;
int endchunk;
int32 startoffset;
int32 endoffset;
int totalchunks;
Pointer chunk;
bool isnull;
int32 chunksize;
int32 chcpystrt;
int32 chcpyend;
attrsize = attr->va_content.va_external.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 = (varattrib *) palloc(length + VARHDRSZ);
VARATT_SIZEP(result) = length + VARHDRSZ;
if (VARATT_IS_COMPRESSED(attr))
VARATT_SIZEP(result) |= VARATT_FLAG_COMPRESSED;
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 it's index
*/
toastrel = heap_open(attr->va_content.va_external.va_toastrelid,
AccessShareLock);
toasttupDesc = toastrel->rd_att;
toastidx = index_open(toastrel->rd_rel->reltoastidxid);
/*
* Setup a scan key to fetch from the index. This is either two keys
* or three depending on the number of chunks.
*/
ScanKeyEntryInitialize(&toastkey[0],
(bits16) 0,
(AttrNumber) 1,
(RegProcedure) F_OIDEQ,
ObjectIdGetDatum(attr->va_content.va_external.va_valueid));
/*
* Now dependent on number of chunks:
*/
if (numchunks == 1)
{
ScanKeyEntryInitialize(&toastkey[1],
(bits16) 0,
(AttrNumber) 2,
(RegProcedure) F_INT4EQ,
Int32GetDatum(startchunk));
nscankeys = 2;
}
else
{
ScanKeyEntryInitialize(&toastkey[1],
(bits16) 0,
(AttrNumber) 2,
(RegProcedure) F_INT4GE,
Int32GetDatum(startchunk));
ScanKeyEntryInitialize(&toastkey[2],
(bits16) 0,
(AttrNumber) 2,
(RegProcedure) 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 = index_beginscan(toastrel, toastidx, SnapshotToast,
nscankeys, toastkey);
while ((ttup = index_getnext(toastscan, ForwardScanDirection)) != NULL)
{
/*
* Have a chunk, extract the sequence number and the data
*/
residx = DatumGetInt32(heap_getattr(ttup, 2, toasttupDesc, &isnull));
Assert(!isnull);
chunk = DatumGetPointer(heap_getattr(ttup, 3, toasttupDesc, &isnull));
Assert(!isnull);
chunksize = VARATT_SIZE(chunk) - VARHDRSZ;
/*
* 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",
residx, nextidx,
attr->va_content.va_external.va_valueid);
if (residx < totalchunks - 1)
{
if (chunksize != TOAST_MAX_CHUNK_SIZE)
elog(ERROR, "unexpected chunk size %d in chunk %d for toast value %u",
chunksize, residx,
attr->va_content.va_external.va_valueid);
}
else
{
if ((residx * TOAST_MAX_CHUNK_SIZE + chunksize) != attrsize)
elog(ERROR, "unexpected chunk size %d in chunk %d for toast value %u",
chunksize, residx,
attr->va_content.va_external.va_valueid);
}
/*
* 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(((char *) VARATT_DATA(result)) +
(residx * TOAST_MAX_CHUNK_SIZE - sliceoffset) + chcpystrt,
VARATT_DATA(chunk) + 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",
nextidx,
attr->va_content.va_external.va_valueid);
/*
* End scan and close relations
*/
index_endscan(toastscan);
index_close(toastidx);
heap_close(toastrel, AccessShareLock);
return result;
}
/* ----------
* heap_tuple_untoast_attr_slice -
*
* Public entry point to get back part of a toasted value
* from compression or external storage.
* ----------
*/
struct varlena *
heap_tuple_untoast_attr_slice(struct varlena * attr,
int32 sliceoffset, int32 slicelength)
{
struct varlena *preslice;
struct varlena *result;
char *attrdata;
int32 attrsize;
if (VARATT_IS_EXTERNAL(attr))
{
struct varatt_external toast_pointer;
VARATT_EXTERNAL_GET_POINTER(toast_pointer, attr);
/* fast path for non-compressed external datums */
if (!VARATT_EXTERNAL_IS_COMPRESSED(toast_pointer))
return toast_fetch_datum_slice(attr, sliceoffset, slicelength);
/* fetch it back (compressed marker will get set automatically) */
preslice = toast_fetch_datum(attr);
}
else
preslice = attr;
if (VARATT_IS_COMPRESSED(preslice))
{
PGLZ_Header *tmp = (PGLZ_Header *) preslice;
Size size = PGLZ_RAW_SIZE(tmp) + VARHDRSZ;
preslice = (struct varlena *) palloc(size);
SET_VARSIZE(preslice, size);
pglz_decompress(tmp, VARDATA(preslice));
if (tmp != (PGLZ_Header *) attr)
pfree(tmp);
}
if (VARATT_IS_SHORT(preslice))
{
attrdata = VARDATA_SHORT(preslice);
attrsize = VARSIZE_SHORT(preslice) - VARHDRSZ_SHORT;
}
else
{
attrdata = VARDATA(preslice);
attrsize = VARSIZE(preslice) - VARHDRSZ;
}
/* slicing of datum for compressed cases and plain value */
if (sliceoffset >= attrsize)
{
sliceoffset = 0;
slicelength = 0;
}
if (((sliceoffset + slicelength) > attrsize) || slicelength < 0)
slicelength = attrsize - sliceoffset;
result = (struct varlena *) palloc(slicelength + VARHDRSZ);
SET_VARSIZE(result, slicelength + VARHDRSZ);
memcpy(VARDATA(result), attrdata + sliceoffset, slicelength);
if (preslice != attr)
pfree(preslice);
return result;
}
/* ----------
* toast_save_datum -
*
* Save one single datum into the secondary relation and return
* a Datum reference for it.
* ----------
*/
static Datum
toast_save_datum(Relation rel, Datum value, int options)
{
Relation toastrel;
Relation toastidx;
HeapTuple toasttup;
TupleDesc toasttupDesc;
Datum t_values[3];
bool t_isnull[3];
CommandId mycid = GetCurrentCommandId(true);
struct varlena *result;
struct varatt_external toast_pointer;
struct
{
struct varlena hdr;
char data[TOAST_MAX_CHUNK_SIZE]; /* make struct big enough */
int32 align_it; /* ensure struct is aligned well enough */
} chunk_data;
int32 chunk_size;
int32 chunk_seq = 0;
char *data_p;
int32 data_todo;
Pointer dval = DatumGetPointer(value);
/*
* Open the toast relation and its index. We can use the index to check
* uniqueness of the OID we assign to the toasted item, even though it has
* additional columns besides OID.
*/
toastrel = heap_open(rel->rd_rel->reltoastrelid, RowExclusiveLock);
toasttupDesc = toastrel->rd_att;
toastidx = index_open(toastrel->rd_rel->reltoastidxid, RowExclusiveLock);
/*
* Get the data pointer and length, and compute va_rawsize and va_extsize.
*
* va_rawsize is the size of the equivalent fully uncompressed datum, so
* we have to adjust for short headers.
*
* va_extsize is the actual size of the data payload in the toast records.
*/
if (VARATT_IS_SHORT(dval))
{
data_p = VARDATA_SHORT(dval);
data_todo = VARSIZE_SHORT(dval) - VARHDRSZ_SHORT;
toast_pointer.va_rawsize = data_todo + VARHDRSZ; /* as if not short */
toast_pointer.va_extsize = data_todo;
}
else if (VARATT_IS_COMPRESSED(dval))
{
data_p = VARDATA(dval);
data_todo = VARSIZE(dval) - VARHDRSZ;
/* rawsize in a compressed datum is just the size of the payload */
toast_pointer.va_rawsize = VARRAWSIZE_4B_C(dval) + VARHDRSZ;
toast_pointer.va_extsize = data_todo;
/* Assert that the numbers look like it's compressed */
Assert(VARATT_EXTERNAL_IS_COMPRESSED(toast_pointer));
}
else
{
data_p = VARDATA(dval);
data_todo = VARSIZE(dval) - VARHDRSZ;
toast_pointer.va_rawsize = VARSIZE(dval);
toast_pointer.va_extsize = data_todo;
}
/*
* Insert the correct table OID into the result TOAST pointer.
*
* Normally this is the actual OID of the target toast table, but during
* table-rewriting operations such as CLUSTER, we have to insert the OID
* of the table's real permanent toast table instead. rd_toastoid is set
* if we have to substitute such an OID.
*/
if (OidIsValid(rel->rd_toastoid))
toast_pointer.va_toastrelid = rel->rd_toastoid;
else
toast_pointer.va_toastrelid = RelationGetRelid(toastrel);
/*
* Choose an unused OID within the toast table for this toast value.
*/
toast_pointer.va_valueid = GetNewOidWithIndex(toastrel,
RelationGetRelid(toastidx),
(AttrNumber) 1);
/*
* Initialize constant parts of the tuple data
*/
t_values[0] = ObjectIdGetDatum(toast_pointer.va_valueid);
t_values[2] = PointerGetDatum(&chunk_data);
t_isnull[0] = false;
t_isnull[1] = false;
t_isnull[2] = false;
/*
* Split up the item into chunks
*/
while (data_todo > 0)
{
/*
* Calculate the size of this chunk
*/
chunk_size = Min(TOAST_MAX_CHUNK_SIZE, data_todo);
/*
* Build a tuple and store it
*/
t_values[1] = Int32GetDatum(chunk_seq++);
SET_VARSIZE(&chunk_data, chunk_size + VARHDRSZ);
memcpy(VARDATA(&chunk_data), data_p, chunk_size);
toasttup = heap_form_tuple(toasttupDesc, t_values, t_isnull);
heap_insert(toastrel, toasttup, mycid, options, NULL);
/*
* Create the index entry. We cheat a little here by not using
* FormIndexDatum: this relies on the knowledge that the index columns
* are the same as the initial columns of the table.
*
* Note also that there had better not be any user-created index on
* the TOAST table, since we don't bother to update anything else.
*/
index_insert(toastidx, t_values, t_isnull,
&(toasttup->t_self),
toastrel,
toastidx->rd_index->indisunique ?
UNIQUE_CHECK_YES : UNIQUE_CHECK_NO);
/*
* Free memory
*/
heap_freetuple(toasttup);
/*
* Move on to next chunk
*/
data_todo -= chunk_size;
data_p += chunk_size;
}
/*
* Done - close toast relation
*/
index_close(toastidx, RowExclusiveLock);
heap_close(toastrel, RowExclusiveLock);
/*
* Create the TOAST pointer value that we'll return
*/
result = (struct varlena *) palloc(TOAST_POINTER_SIZE);
SET_VARSIZE_EXTERNAL(result, TOAST_POINTER_SIZE);
memcpy(VARDATA_EXTERNAL(result), &toast_pointer, sizeof(toast_pointer));
return PointerGetDatum(result);
}
/* ----------
* toast_save_datum -
*
* Save one single datum into the secondary relation and return
* a varattrib reference for it.
* ----------
*/
static Datum
toast_save_datum(Relation rel, Datum value)
{
Relation toastrel;
Relation toastidx;
HeapTuple toasttup;
InsertIndexResult idxres;
TupleDesc toasttupDesc;
Datum t_values[3];
char t_nulls[3];
varattrib *result;
struct
{
struct varlena hdr;
char data[TOAST_MAX_CHUNK_SIZE];
} chunk_data;
int32 chunk_size;
int32 chunk_seq = 0;
char *data_p;
int32 data_todo;
/*
* Create the varattrib reference
*/
result = (varattrib *) palloc(sizeof(varattrib));
result->va_header = sizeof(varattrib) | VARATT_FLAG_EXTERNAL;
if (VARATT_IS_COMPRESSED(value))
{
result->va_header |= VARATT_FLAG_COMPRESSED;
result->va_content.va_external.va_rawsize =
((varattrib *) value)->va_content.va_compressed.va_rawsize;
}
else
result->va_content.va_external.va_rawsize = VARATT_SIZE(value);
result->va_content.va_external.va_extsize =
VARATT_SIZE(value) - VARHDRSZ;
result->va_content.va_external.va_valueid = newoid();
result->va_content.va_external.va_toastrelid =
rel->rd_rel->reltoastrelid;
/*
* Initialize constant parts of the tuple data
*/
t_values[0] = ObjectIdGetDatum(result->va_content.va_external.va_valueid);
t_values[2] = PointerGetDatum(&chunk_data);
t_nulls[0] = ' ';
t_nulls[1] = ' ';
t_nulls[2] = ' ';
/*
* Get the data to process
*/
data_p = VARATT_DATA(value);
data_todo = VARATT_SIZE(value) - VARHDRSZ;
/*
* Open the toast relation
*/
toastrel = heap_open(rel->rd_rel->reltoastrelid, RowExclusiveLock);
toasttupDesc = toastrel->rd_att;
toastidx = index_open(toastrel->rd_rel->reltoastidxid);
/*
* Split up the item into chunks
*/
while (data_todo > 0)
{
/*
* Calculate the size of this chunk
*/
chunk_size = Min(TOAST_MAX_CHUNK_SIZE, data_todo);
/*
* Build a tuple and store it
*/
t_values[1] = Int32GetDatum(chunk_seq++);
VARATT_SIZEP(&chunk_data) = chunk_size + VARHDRSZ;
memcpy(VARATT_DATA(&chunk_data), data_p, chunk_size);
toasttup = heap_formtuple(toasttupDesc, t_values, t_nulls);
if (!HeapTupleIsValid(toasttup))
elog(ERROR, "failed to build TOAST tuple");
simple_heap_insert(toastrel, toasttup);
/*
* Create the index entry. We cheat a little here by not using
* FormIndexDatum: this relies on the knowledge that the index
* columns are the same as the initial columns of the table.
*
* Note also that there had better not be any user-created index on
* the TOAST table, since we don't bother to update anything else.
*/
idxres = index_insert(toastidx, t_values, t_nulls,
&(toasttup->t_self),
toastrel, toastidx->rd_index->indisunique);
if (idxres == NULL)
elog(ERROR, "failed to insert index entry for TOAST tuple");
/*
* Free memory
*/
pfree(idxres);
heap_freetuple(toasttup);
/*
* Move on to next chunk
*/
data_todo -= chunk_size;
data_p += chunk_size;
}
/*
* Done - close toast relation and return the reference
*/
index_close(toastidx);
heap_close(toastrel, RowExclusiveLock);
return PointerGetDatum(result);
}
static void
convert_heaptuple(HeapTupleHeader htup)
{
/* HEAP_HASEXTENDED and HEAP_HASCOMPRESSED flags were removed. Clear them out */
htup->t_infomask &= ~(VERSION4_HEAP_HASEXTENDED | VERSION4_HEAP_HASCOMPRESSED);
/* HEAP_HASOID flag was moved */
if (htup->t_infomask & VERSION4_HEAP_HASOID)
{
htup->t_infomask &= ~(VERSION4_HEAP_HASOID);
htup->t_infomask |= HEAP_HASOID;
}
/* Any numeric columns? */
if (curr_hasnumerics)
{
/* This is like heap_deform_tuple() */
bool hasnulls = (htup->t_infomask & HEAP_HASNULL) != 0;
int attnum;
char *tp;
long off;
bits8 *bp = htup->t_bits; /* ptr to null bitmap in tuple */
tp = (char *) htup + htup->t_hoff;
off = 0;
for (attnum = 0; attnum < curr_natts; attnum++)
{
AttInfo *thisatt = &curr_atts[attnum];
if (hasnulls && att_isnull(attnum, bp))
continue;
if (thisatt->attlen == -1)
{
off = att_align_pointer(off, thisatt->attalign, -1,
tp + off);
}
else
{
/* not varlena, so safe to use att_align_nominal */
off = att_align_nominal(off, thisatt->attalign);
}
if (thisatt->is_numeric)
{
/*
* Before PostgreSQL 8.3, the n_weight and n_sign_dscale fields
* were the other way 'round. Swap them.
*
* FIXME: need to handle toasted datums here.
*/
Datum datum = PointerGetDatum(tp + off);
if (VARATT_IS_COMPRESSED(datum))
pg_log(PG_FATAL, "converting compressed numeric datums is not implemented\n");
else if (VARATT_IS_EXTERNAL(datum))
pg_log(PG_FATAL, "converting toasted numeric datums is not implemented\n");
else
{
char *numericdata = VARDATA_ANY(DatumGetPointer(datum));
int sz = VARSIZE_ANY_EXHDR(DatumGetPointer(datum));
uint16 tmp;
if (sz < 4)
pg_log(PG_FATAL, "unexpected size for numeric datum: %d\n", sz);
memcpy(&tmp, &numericdata[0], 2);
memcpy(&numericdata[0], &numericdata[2], 2);
memcpy(&numericdata[2], &tmp, 2);
}
}
off = att_addlength_pointer(off, thisatt->attlen, tp + off);
}
}
}
/* ----------
* toast_insert_or_update -
*
* Delete no-longer-used toast-entries and create new ones to
* make the new tuple fit on INSERT or UPDATE
*
* Inputs:
* newtup: the candidate new tuple to be inserted
* oldtup: the old row version for UPDATE, or NULL for INSERT
* options: options to be passed to heap_insert() for toast rows
* Result:
* either newtup if no toasting is needed, or a palloc'd modified tuple
* that is what should actually get stored
*
* NOTE: neither newtup nor oldtup will be modified. This is a change
* from the pre-8.1 API of this routine.
* ----------
*/
HeapTuple
toast_insert_or_update(Relation rel, HeapTuple newtup, HeapTuple oldtup,
int options)
{
HeapTuple result_tuple;
TupleDesc tupleDesc;
Form_pg_attribute *att;
int numAttrs;
int i;
bool need_change = false;
bool need_free = false;
bool need_delold = false;
bool has_nulls = false;
Size maxDataLen;
Size hoff;
char toast_action[MaxHeapAttributeNumber];
bool toast_isnull[MaxHeapAttributeNumber];
bool toast_oldisnull[MaxHeapAttributeNumber];
Datum toast_values[MaxHeapAttributeNumber];
Datum toast_oldvalues[MaxHeapAttributeNumber];
int32 toast_sizes[MaxHeapAttributeNumber];
bool toast_free[MaxHeapAttributeNumber];
bool toast_delold[MaxHeapAttributeNumber];
/*
* We should only ever be called for tuples of plain relations ---
* recursing on a toast rel is bad news.
*/
Assert(rel->rd_rel->relkind == RELKIND_RELATION);
/*
* Get the tuple descriptor and break down the tuple(s) into fields.
*/
tupleDesc = rel->rd_att;
att = tupleDesc->attrs;
numAttrs = tupleDesc->natts;
Assert(numAttrs <= MaxHeapAttributeNumber);
heap_deform_tuple(newtup, tupleDesc, toast_values, toast_isnull);
if (oldtup != NULL)
heap_deform_tuple(oldtup, tupleDesc, toast_oldvalues, toast_oldisnull);
/* ----------
* Then collect information about the values given
*
* NOTE: toast_action[i] can have these values:
* ' ' default handling
* 'p' already processed --- don't touch it
* 'x' incompressible, but OK to move off
*
* NOTE: toast_sizes[i] is only made valid for varlena attributes with
* toast_action[i] different from 'p'.
* ----------
*/
memset(toast_action, ' ', numAttrs * sizeof(char));
memset(toast_free, 0, numAttrs * sizeof(bool));
memset(toast_delold, 0, numAttrs * sizeof(bool));
for (i = 0; i < numAttrs; i++)
{
struct varlena *old_value;
struct varlena *new_value;
if (oldtup != NULL)
{
/*
* For UPDATE get the old and new values of this attribute
*/
old_value = (struct varlena *) DatumGetPointer(toast_oldvalues[i]);
new_value = (struct varlena *) DatumGetPointer(toast_values[i]);
/*
* If the old value is an external stored one, check if it has
* changed so we have to delete it later.
*/
if (att[i]->attlen == -1 && !toast_oldisnull[i] &&
VARATT_IS_EXTERNAL(old_value))
{
if (toast_isnull[i] || !VARATT_IS_EXTERNAL(new_value) ||
memcmp((char *) old_value, (char *) new_value,
VARSIZE_EXTERNAL(old_value)) != 0)
{
/*
* The old external stored value isn't needed any more
* after the update
*/
toast_delold[i] = true;
need_delold = true;
}
else
{
/*
* This attribute isn't changed by this update so we reuse
* the original reference to the old value in the new
* tuple.
*/
toast_action[i] = 'p';
continue;
}
}
}
else
{
/*
* For INSERT simply get the new value
*/
new_value = (struct varlena *) DatumGetPointer(toast_values[i]);
}
/*
* Handle NULL attributes
*/
if (toast_isnull[i])
{
toast_action[i] = 'p';
has_nulls = true;
continue;
}
/*
* Now look at varlena attributes
*/
if (att[i]->attlen == -1)
{
/*
* If the table's attribute says PLAIN always, force it so.
*/
if (att[i]->attstorage == 'p')
toast_action[i] = 'p';
/*
* We took care of UPDATE above, so any external value we find
* still in the tuple must be someone else's we cannot reuse.
* Fetch it back (without decompression, unless we are forcing
* PLAIN storage). If necessary, we'll push it out as a new
* external value below.
*/
if (VARATT_IS_EXTERNAL(new_value))
{
if (att[i]->attstorage == 'p')
new_value = heap_tuple_untoast_attr(new_value);
else
new_value = heap_tuple_fetch_attr(new_value);
toast_values[i] = PointerGetDatum(new_value);
toast_free[i] = true;
need_change = true;
need_free = true;
}
/*
* Remember the size of this attribute
*/
toast_sizes[i] = VARSIZE_ANY(new_value);
}
else
{
/*
* Not a varlena attribute, plain storage always
*/
toast_action[i] = 'p';
}
}
/* ----------
* Compress and/or save external until data fits into target length
*
* 1: Inline compress attributes with attstorage 'x', and store very
* large attributes with attstorage 'x' or 'e' external immediately
* 2: Store attributes with attstorage 'x' or 'e' external
* 3: Inline compress attributes with attstorage 'm'
* 4: Store attributes with attstorage 'm' external
* ----------
*/
/* compute header overhead --- this should match heap_form_tuple() */
hoff = offsetof(HeapTupleHeaderData, t_bits);
if (has_nulls)
hoff += BITMAPLEN(numAttrs);
if (newtup->t_data->t_infomask & HEAP_HASOID)
hoff += sizeof(Oid);
hoff = MAXALIGN(hoff);
Assert(hoff == newtup->t_data->t_hoff);
/* now convert to a limit on the tuple data size */
maxDataLen = TOAST_TUPLE_TARGET - hoff;
/*
* Look for attributes with attstorage 'x' to compress. Also find large
* attributes with attstorage 'x' or 'e', and store them external.
*/
while (heap_compute_data_size(tupleDesc,
toast_values, toast_isnull) > maxDataLen)
{
int biggest_attno = -1;
int32 biggest_size = MAXALIGN(TOAST_POINTER_SIZE);
Datum old_value;
Datum new_value;
/*
* Search for the biggest yet unprocessed internal attribute
*/
for (i = 0; i < numAttrs; i++)
{
if (toast_action[i] != ' ')
continue;
if (VARATT_IS_EXTERNAL(DatumGetPointer(toast_values[i])))
continue; /* can't happen, toast_action would be 'p' */
if (VARATT_IS_COMPRESSED(DatumGetPointer(toast_values[i])))
continue;
if (att[i]->attstorage != 'x' && att[i]->attstorage != 'e')
continue;
if (toast_sizes[i] > biggest_size)
{
biggest_attno = i;
biggest_size = toast_sizes[i];
}
}
if (biggest_attno < 0)
break;
/*
* Attempt to compress it inline, if it has attstorage 'x'
*/
i = biggest_attno;
if (att[i]->attstorage == 'x')
{
old_value = toast_values[i];
new_value = toast_compress_datum(old_value);
if (DatumGetPointer(new_value) != NULL)
{
/* successful compression */
if (toast_free[i])
pfree(DatumGetPointer(old_value));
toast_values[i] = new_value;
toast_free[i] = true;
toast_sizes[i] = VARSIZE(DatumGetPointer(toast_values[i]));
need_change = true;
need_free = true;
}
else
{
/* incompressible, ignore on subsequent compression passes */
toast_action[i] = 'x';
}
}
else
{
/* has attstorage 'e', ignore on subsequent compression passes */
toast_action[i] = 'x';
}
/*
* If this value is by itself more than maxDataLen (after compression
* if any), push it out to the toast table immediately, if possible.
* This avoids uselessly compressing other fields in the common case
* where we have one long field and several short ones.
*
* XXX maybe the threshold should be less than maxDataLen?
*/
if (toast_sizes[i] > maxDataLen &&
rel->rd_rel->reltoastrelid != InvalidOid)
{
old_value = toast_values[i];
toast_action[i] = 'p';
toast_values[i] = toast_save_datum(rel, toast_values[i], options);
if (toast_free[i])
pfree(DatumGetPointer(old_value));
toast_free[i] = true;
need_change = true;
need_free = true;
}
}
/*
* Second we look for attributes of attstorage 'x' or 'e' that are still
* inline. But skip this if there's no toast table to push them to.
*/
while (heap_compute_data_size(tupleDesc,
toast_values, toast_isnull) > maxDataLen &&
rel->rd_rel->reltoastrelid != InvalidOid)
{
int biggest_attno = -1;
int32 biggest_size = MAXALIGN(TOAST_POINTER_SIZE);
Datum old_value;
/*------
* Search for the biggest yet inlined attribute with
* attstorage equals 'x' or 'e'
*------
*/
for (i = 0; i < numAttrs; i++)
{
if (toast_action[i] == 'p')
continue;
if (VARATT_IS_EXTERNAL(DatumGetPointer(toast_values[i])))
continue; /* can't happen, toast_action would be 'p' */
if (att[i]->attstorage != 'x' && att[i]->attstorage != 'e')
continue;
if (toast_sizes[i] > biggest_size)
{
biggest_attno = i;
biggest_size = toast_sizes[i];
}
}
if (biggest_attno < 0)
break;
/*
* Store this external
*/
i = biggest_attno;
old_value = toast_values[i];
toast_action[i] = 'p';
toast_values[i] = toast_save_datum(rel, toast_values[i], options);
if (toast_free[i])
pfree(DatumGetPointer(old_value));
toast_free[i] = true;
need_change = true;
need_free = true;
}
/*
* Round 3 - this time we take attributes with storage 'm' into
* compression
*/
while (heap_compute_data_size(tupleDesc,
toast_values, toast_isnull) > maxDataLen)
{
int biggest_attno = -1;
int32 biggest_size = MAXALIGN(TOAST_POINTER_SIZE);
Datum old_value;
Datum new_value;
/*
* Search for the biggest yet uncompressed internal attribute
*/
for (i = 0; i < numAttrs; i++)
{
if (toast_action[i] != ' ')
continue;
if (VARATT_IS_EXTERNAL(DatumGetPointer(toast_values[i])))
continue; /* can't happen, toast_action would be 'p' */
if (VARATT_IS_COMPRESSED(DatumGetPointer(toast_values[i])))
continue;
if (att[i]->attstorage != 'm')
continue;
if (toast_sizes[i] > biggest_size)
{
biggest_attno = i;
biggest_size = toast_sizes[i];
}
}
if (biggest_attno < 0)
break;
/*
* Attempt to compress it inline
*/
i = biggest_attno;
old_value = toast_values[i];
new_value = toast_compress_datum(old_value);
if (DatumGetPointer(new_value) != NULL)
{
/* successful compression */
if (toast_free[i])
pfree(DatumGetPointer(old_value));
toast_values[i] = new_value;
toast_free[i] = true;
toast_sizes[i] = VARSIZE(DatumGetPointer(toast_values[i]));
need_change = true;
need_free = true;
}
else
{
/* incompressible, ignore on subsequent compression passes */
toast_action[i] = 'x';
}
}
/*
* Finally we store attributes of type 'm' externally. At this point we
* increase the target tuple size, so that 'm' attributes aren't stored
* externally unless really necessary.
*/
maxDataLen = TOAST_TUPLE_TARGET_MAIN - hoff;
while (heap_compute_data_size(tupleDesc,
toast_values, toast_isnull) > maxDataLen &&
rel->rd_rel->reltoastrelid != InvalidOid)
{
int biggest_attno = -1;
int32 biggest_size = MAXALIGN(TOAST_POINTER_SIZE);
Datum old_value;
/*--------
* Search for the biggest yet inlined attribute with
* attstorage = 'm'
*--------
*/
for (i = 0; i < numAttrs; i++)
{
if (toast_action[i] == 'p')
continue;
if (VARATT_IS_EXTERNAL(DatumGetPointer(toast_values[i])))
continue; /* can't happen, toast_action would be 'p' */
if (att[i]->attstorage != 'm')
continue;
if (toast_sizes[i] > biggest_size)
{
biggest_attno = i;
biggest_size = toast_sizes[i];
}
}
if (biggest_attno < 0)
break;
/*
* Store this external
*/
i = biggest_attno;
old_value = toast_values[i];
toast_action[i] = 'p';
toast_values[i] = toast_save_datum(rel, toast_values[i], options);
if (toast_free[i])
pfree(DatumGetPointer(old_value));
toast_free[i] = true;
need_change = true;
need_free = true;
}
/*
* In the case we toasted any values, we need to build a new heap tuple
* with the changed values.
*/
if (need_change)
{
HeapTupleHeader olddata = newtup->t_data;
HeapTupleHeader new_data;
int32 new_len;
int32 new_data_len;
/*
* Calculate the new size of the tuple. Header size should not
* change, but data size might.
*/
new_len = offsetof(HeapTupleHeaderData, t_bits);
if (has_nulls)
new_len += BITMAPLEN(numAttrs);
if (olddata->t_infomask & HEAP_HASOID)
new_len += sizeof(Oid);
new_len = MAXALIGN(new_len);
Assert(new_len == olddata->t_hoff);
new_data_len = heap_compute_data_size(tupleDesc,
toast_values, toast_isnull);
new_len += new_data_len;
/*
* Allocate and zero the space needed, and fill HeapTupleData fields.
*/
result_tuple = (HeapTuple) palloc0(HEAPTUPLESIZE + new_len);
result_tuple->t_len = new_len;
result_tuple->t_self = newtup->t_self;
result_tuple->t_tableOid = newtup->t_tableOid;
new_data = (HeapTupleHeader) ((char *) result_tuple + HEAPTUPLESIZE);
result_tuple->t_data = new_data;
/*
* Put the existing tuple header and the changed values into place
*/
memcpy(new_data, olddata, olddata->t_hoff);
heap_fill_tuple(tupleDesc,
toast_values,
toast_isnull,
(char *) new_data + olddata->t_hoff,
new_data_len,
&(new_data->t_infomask),
has_nulls ? new_data->t_bits : NULL);
}
else
result_tuple = newtup;
/*
* Free allocated temp values
*/
if (need_free)
for (i = 0; i < numAttrs; i++)
if (toast_free[i])
pfree(DatumGetPointer(toast_values[i]));
/*
* Delete external values from the old tuple
*/
if (need_delold)
for (i = 0; i < numAttrs; i++)
if (toast_delold[i])
toast_delete_datum(rel, toast_oldvalues[i]);
return result_tuple;
}
/* ----------
* toast_fetch_datum -
*
* Reconstruct an in memory varattrib from the chunks saved
* in the toast relation
* ----------
*/
static varattrib *
toast_fetch_datum(varattrib *attr)
{
Relation toastrel;
Relation toastidx;
ScanKeyData toastkey;
IndexScanDesc toastscan;
HeapTuple ttup;
TupleDesc toasttupDesc;
varattrib *result;
int32 ressize;
int32 residx,
nextidx;
int32 numchunks;
Pointer chunk;
bool isnull;
int32 chunksize;
ressize = attr->va_content.va_external.va_extsize;
numchunks = ((ressize - 1) / TOAST_MAX_CHUNK_SIZE) + 1;
result = (varattrib *) palloc(ressize + VARHDRSZ);
VARATT_SIZEP(result) = ressize + VARHDRSZ;
if (VARATT_IS_COMPRESSED(attr))
VARATT_SIZEP(result) |= VARATT_FLAG_COMPRESSED;
/*
* Open the toast relation and its index
*/
toastrel = heap_open(attr->va_content.va_external.va_toastrelid,
AccessShareLock);
toasttupDesc = toastrel->rd_att;
toastidx = index_open(toastrel->rd_rel->reltoastidxid);
/*
* Setup a scan key to fetch from the index by va_valueid
*/
ScanKeyEntryInitialize(&toastkey,
(bits16) 0,
(AttrNumber) 1,
(RegProcedure) F_OIDEQ,
ObjectIdGetDatum(attr->va_content.va_external.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 = index_beginscan(toastrel, toastidx, SnapshotToast,
1, &toastkey);
while ((ttup = index_getnext(toastscan, ForwardScanDirection)) != NULL)
{
/*
* Have a chunk, extract the sequence number and the data
*/
residx = DatumGetInt32(heap_getattr(ttup, 2, toasttupDesc, &isnull));
Assert(!isnull);
chunk = DatumGetPointer(heap_getattr(ttup, 3, toasttupDesc, &isnull));
Assert(!isnull);
chunksize = VARATT_SIZE(chunk) - VARHDRSZ;
/*
* Some checks on the data we've found
*/
if (residx != nextidx)
elog(ERROR, "unexpected chunk number %d (expected %d) for toast value %u",
residx, nextidx,
attr->va_content.va_external.va_valueid);
if (residx < numchunks - 1)
{
if (chunksize != TOAST_MAX_CHUNK_SIZE)
elog(ERROR, "unexpected chunk size %d in chunk %d for toast value %u",
chunksize, residx,
attr->va_content.va_external.va_valueid);
}
else if (residx < numchunks)
{
if ((residx * TOAST_MAX_CHUNK_SIZE + chunksize) != ressize)
elog(ERROR, "unexpected chunk size %d in chunk %d for toast value %u",
chunksize, residx,
attr->va_content.va_external.va_valueid);
}
else
elog(ERROR, "unexpected chunk number %d for toast value %u",
residx,
attr->va_content.va_external.va_valueid);
/*
* Copy the data into proper place in our result
*/
memcpy(((char *) VARATT_DATA(result)) + residx * TOAST_MAX_CHUNK_SIZE,
VARATT_DATA(chunk),
chunksize);
nextidx++;
}
/*
* Final checks that we successfully fetched the datum
*/
if (nextidx != numchunks)
elog(ERROR, "missing chunk number %d for toast value %u",
nextidx,
attr->va_content.va_external.va_valueid);
/*
* End scan and close relations
*/
index_endscan(toastscan);
index_close(toastidx);
heap_close(toastrel, AccessShareLock);
return result;
}
/* ----------
* heap_tuple_untoast_attr_slice -
*
* Public entry point to get back part of a toasted value
* from compression or external storage.
* ----------
*/
struct varlena *
heap_tuple_untoast_attr_slice(struct varlena * attr,
int32 sliceoffset, int32 slicelength)
{
varattrib *preslice;
varattrib *result;
char *attrdata;
int32 attrsize;
if (VARATT_IS_EXTERNAL(attr))
{
/* fast path for non-compressed external datums */
if (!VARATT_EXTERNAL_IS_COMPRESSED(attr))
return toast_fetch_datum_slice(attr, sliceoffset, slicelength);
/* this automatically sets the compressed flag if appropriate */
preslice = (varattrib *)toast_fetch_datum(attr);
}
else
preslice = (varattrib *)attr;
if (VARATT_IS_COMPRESSED(preslice))
{
unsigned size;
PGLZ_Header *tmp;
tmp = (PGLZ_Header *) preslice;
size = PGLZ_RAW_SIZE(tmp) + VARHDRSZ;
preslice = (varattrib *) palloc(size);
SET_VARSIZE(preslice, size);
pglz_decompress(tmp, VARDATA(preslice));
if (tmp != (PGLZ_Header *) attr)
pfree(tmp);
}
if (VARATT_IS_SHORT(preslice))
{
attrdata = VARDATA_SHORT(preslice);
attrsize = VARSIZE_SHORT(preslice) - VARHDRSZ_SHORT;
}
else
{
attrdata = VARDATA(preslice);
attrsize = VARSIZE(preslice) - VARHDRSZ;
}
/* slicing of datum for compressed cases and plain value */
if (sliceoffset >= attrsize)
{
sliceoffset = 0;
slicelength = 0;
}
if (((sliceoffset + slicelength) > attrsize) || slicelength < 0)
slicelength = attrsize - sliceoffset;
result = (varattrib *) palloc(slicelength + VARHDRSZ);
SET_VARSIZE(result, slicelength + VARHDRSZ);
memcpy(VARDATA(result), attrdata + sliceoffset, slicelength);
if ((struct varlena *)preslice != (struct varlena *)attr)
pfree(preslice);
return (struct varlena *)result;
}
/* ----------
* toast_flatten_tuple_attribute -
*
* If a Datum is of composite type, "flatten" it to contain no toasted fields.
* This must be invoked on any potentially-composite field that is to be
* inserted into a tuple. Doing this preserves the invariant that toasting
* goes only one level deep in a tuple.
*
* Note that flattening does not mean expansion of short-header varlenas,
* so in one sense toasting is allowed within composite datums.
* ----------
*/
Datum
toast_flatten_tuple_attribute(Datum value,
Oid typeId, int32 typeMod)
{
TupleDesc tupleDesc;
HeapTupleHeader olddata;
HeapTupleHeader new_data;
int32 new_len;
int32 new_data_len;
HeapTupleData tmptup;
Form_pg_attribute *att;
int numAttrs;
int i;
bool need_change = false;
bool has_nulls = false;
Datum toast_values[MaxTupleAttributeNumber];
bool toast_isnull[MaxTupleAttributeNumber];
bool toast_free[MaxTupleAttributeNumber];
/*
* See if it's a composite type, and get the tupdesc if so.
*/
tupleDesc = lookup_rowtype_tupdesc_noerror(typeId, typeMod, true);
if (tupleDesc == NULL)
return value; /* not a composite type */
att = tupleDesc->attrs;
numAttrs = tupleDesc->natts;
/*
* Break down the tuple into fields.
*/
olddata = DatumGetHeapTupleHeader(value);
Assert(typeId == HeapTupleHeaderGetTypeId(olddata));
Assert(typeMod == HeapTupleHeaderGetTypMod(olddata));
/* Build a temporary HeapTuple control structure */
tmptup.t_len = HeapTupleHeaderGetDatumLength(olddata);
ItemPointerSetInvalid(&(tmptup.t_self));
tmptup.t_tableOid = InvalidOid;
tmptup.t_data = olddata;
Assert(numAttrs <= MaxTupleAttributeNumber);
heap_deform_tuple(&tmptup, tupleDesc, toast_values, toast_isnull);
memset(toast_free, 0, numAttrs * sizeof(bool));
for (i = 0; i < numAttrs; i++)
{
/*
* Look at non-null varlena attributes
*/
if (toast_isnull[i])
has_nulls = true;
else if (att[i]->attlen == -1)
{
struct varlena *new_value;
new_value = (struct varlena *) DatumGetPointer(toast_values[i]);
if (VARATT_IS_EXTERNAL(new_value) ||
VARATT_IS_COMPRESSED(new_value))
{
new_value = heap_tuple_untoast_attr(new_value);
toast_values[i] = PointerGetDatum(new_value);
toast_free[i] = true;
need_change = true;
}
}
}
/*
* If nothing to untoast, just return the original tuple.
*/
if (!need_change)
{
ReleaseTupleDesc(tupleDesc);
return value;
}
/*
* Calculate the new size of the tuple. Header size should not change,
* but data size might.
*/
new_len = offsetof(HeapTupleHeaderData, t_bits);
if (has_nulls)
new_len += BITMAPLEN(numAttrs);
if (olddata->t_infomask & HEAP_HASOID)
new_len += sizeof(Oid);
new_len = MAXALIGN(new_len);
Assert(new_len == olddata->t_hoff);
new_data_len = heap_compute_data_size(tupleDesc,
toast_values, toast_isnull);
new_len += new_data_len;
new_data = (HeapTupleHeader) palloc0(new_len);
/*
* Put the tuple header and the changed values into place
*/
memcpy(new_data, olddata, olddata->t_hoff);
HeapTupleHeaderSetDatumLength(new_data, new_len);
heap_fill_tuple(tupleDesc,
toast_values,
toast_isnull,
(char *) new_data + olddata->t_hoff,
new_data_len,
&(new_data->t_infomask),
has_nulls ? new_data->t_bits : NULL);
/*
* Free allocated temp values
*/
for (i = 0; i < numAttrs; i++)
if (toast_free[i])
pfree(DatumGetPointer(toast_values[i]));
ReleaseTupleDesc(tupleDesc);
return PointerGetDatum(new_data);
}
/* checks the individual attributes of the tuple */
uint32 check_index_tuple_attributes(Relation rel, PageHeader header, int block, int i, char *buffer) {
IndexTuple tuple;
uint32 nerrs = 0;
int j, off;
bits8 * bitmap;
BTPageOpaque opaque;
ereport(DEBUG2,(errmsg("[%d:%d] checking attributes for the tuple", block, i)));
/* get the index tuple and info about the page */
tuple = (IndexTuple)(buffer + header->pd_linp[i].lp_off);
opaque = (BTPageOpaque)(buffer + header->pd_special);
/* current attribute offset - always starts at (buffer + off) */
off = header->pd_linp[i].lp_off + IndexInfoFindDataOffset(tuple->t_info);
ereport(DEBUG3,(errmsg("[%d:%d] tuple has %d attributes", block, (i+1),
RelationGetNumberOfAttributes(rel))));
bitmap = (bits8*)(buffer + header->pd_linp[i].lp_off + sizeof(IndexTupleData));
/* TODO This is mostly copy'n'paste from check_heap_tuple_attributes,
so maybe it could be refactored to share the code. */
/* For left-most tuples on non-leaf pages, there are no data actually
(see src/backend/access/nbtree/README, last paragraph in section "Notes
About Data Representation")
Use P_LEFTMOST/P_ISLEAF to identify such cases (for the leftmost item only)
and set len = 0.
*/
if (P_LEFTMOST(opaque) && (! P_ISLEAF(opaque)) && (i == 0)) {
ereport(DEBUG3, (errmsg("[%d:%d] leftmost tuple on non-leaf block => no data, skipping", block, i)));
return nerrs;
}
/* check all the index attributes */
for (j = 0; j < rel->rd_att->natts; j++) {
/* default length of the attribute */
int len = rel->rd_att->attrs[j]->attlen;
/* copy from src/backend/commands/analyze.c */
bool is_varlena = (!rel->rd_att->attrs[j]->attbyval && len == -1);
bool is_varwidth = (!rel->rd_att->attrs[j]->attbyval && len < 0); /* thus it's "len = -2" */
/* if the attribute is marked as NULL (in the tuple header), skip to the next attribute */
if (IndexTupleHasNulls(tuple) && att_isnull(j, bitmap)) {
ereport(DEBUG3, (errmsg("[%d:%d] attribute '%s' is NULL (skipping)", block, (i+1), rel->rd_att->attrs[j]->attname.data)));
continue;
}
/* fix the alignment (see src/include/access/tupmacs.h) */
off = att_align_pointer(off, rel->rd_att->attrs[j]->attalign, rel->rd_att->attrs[j]->attlen, buffer+off);
if (is_varlena) {
/*
other interesting macros (see postgres.h) - should do something about those ...
VARATT_IS_COMPRESSED(PTR) VARATT_IS_4B_C(PTR)
VARATT_IS_EXTERNAL(PTR) VARATT_IS_1B_E(PTR)
VARATT_IS_SHORT(PTR) VARATT_IS_1B(PTR)
VARATT_IS_EXTENDED(PTR) (!VARATT_IS_4B_U(PTR))
*/
len = VARSIZE_ANY(buffer + off);
if (len < 0) {
ereport(WARNING, (errmsg("[%d:%d] attribute '%s' has negative length < 0 (%d)", block, (i+1), rel->rd_att->attrs[j]->attname.data, len)));
++nerrs;
break;
}
if (VARATT_IS_COMPRESSED(buffer + off)) {
/* the raw length should be less than 1G (and positive) */
if ((VARRAWSIZE_4B_C(buffer + off) < 0) || (VARRAWSIZE_4B_C(buffer + off) > 1024*1024)) {
ereport(WARNING, (errmsg("[%d:%d] attribute '%s' has invalid length %d (should be between 0 and 1G)", block, (i+1), rel->rd_att->attrs[j]->attname.data, VARRAWSIZE_4B_C(buffer + off))));
++nerrs;
/* no break here, this does not break the page structure - we may check the other attributes */
}
}
/* FIXME Check if the varlena value may be detoasted. */
} else if (is_varwidth) {
/* get the C-string length (at most to the end of tuple), +1 as it does not include '\0' at the end */
/* if the string is not properly terminated, then this returns 'remaining space + 1' so it's detected */
len = strnlen(buffer + off, header->pd_linp[i].lp_off + len + header->pd_linp[i].lp_len - off) + 1;
}
/* Check if the length makes sense (is not negative and does not overflow
* the tuple end, stop validating the other rows (we don't know where to
* continue anyway). */
if (off + len > (header->pd_linp[i].lp_off + header->pd_linp[i].lp_len)) {
ereport(WARNING, (errmsg("[%d:%d] attribute '%s' (off=%d len=%d) overflows tuple end (off=%d, len=%d)", block, (i+1), rel->rd_att->attrs[j]->attname.data, off, len, header->pd_linp[i].lp_off, header->pd_linp[i].lp_len)));
++nerrs;
break;
}
/* skip to the next attribute */
off += len;
ereport(DEBUG3,(errmsg("[%d:%d] attribute '%s' len=%d", block, (i+1), rel->rd_att->attrs[j]->attname.data, len)));
}
ereport(DEBUG3,(errmsg("[%d:%d] last attribute ends at %d, tuple ends at %d", block, (i+1), off, header->pd_linp[i].lp_off + header->pd_linp[i].lp_len)));
/* after the last attribute, the offset should be exactly the same as the end of the tuple */
if (MAXALIGN(off) != header->pd_linp[i].lp_off + header->pd_linp[i].lp_len) {
ereport(WARNING, (errmsg("[%d:%d] the last attribute ends at %d but the tuple ends at %d", block, (i+1), off, header->pd_linp[i].lp_off + header->pd_linp[i].lp_len)));
++nerrs;
}
return nerrs;
}