/*
* mdopen() -- Open the specified relation. ereport's on failure.
* (Optionally, can return NULL instead of ereport for ENOENT.)
*
* Note we only open the first segment, when there are multiple segments.
*/
static MdfdVec *
mdopen(SMgrRelation reln, bool allowNotFound)
{
MdfdVec *mdfd;
char *path;
File fd;
/* No work if already open */
if (reln->md_fd)
return reln->md_fd;
path = relpath(reln->smgr_rnode);
fd = PathNameOpenFile(path, O_RDWR | PG_BINARY, 0600);
if (fd < 0)
{
/*
* During bootstrap, there are cases where a system relation will be
* accessed (by internal backend processes) before the bootstrap
* script nominally creates it. Therefore, accept mdopen() as a
* substitute for mdcreate() in bootstrap mode only. (See mdcreate)
*/
if (IsBootstrapProcessingMode())
fd = PathNameOpenFile(path, O_RDWR | O_CREAT | O_EXCL | PG_BINARY, 0600);
if (fd < 0)
{
pfree(path);
if (allowNotFound && errno == ENOENT)
return NULL;
ereport(ERROR,
(errcode_for_file_access(),
errmsg("could not open relation %u/%u/%u: %m",
reln->smgr_rnode.spcNode,
reln->smgr_rnode.dbNode,
reln->smgr_rnode.relNode)));
}
}
pfree(path);
reln->md_fd = mdfd = _fdvec_alloc();
mdfd->mdfd_vfd = fd;
mdfd->mdfd_segno = 0;
#ifndef LET_OS_MANAGE_FILESIZE
mdfd->mdfd_chain = NULL;
Assert(_mdnblocks(fd, BLCKSZ) <= ((BlockNumber) RELSEG_SIZE));
#endif
return mdfd;
}
/*
* mdopen() -- Open the specified relation.
*
* Note we only open the first segment, when there are multiple segments.
*
* If first segment is not present, either ereport or return NULL according
* to "behavior". We treat EXTENSION_CREATE the same as EXTENSION_FAIL;
* EXTENSION_CREATE means it's OK to extend an existing relation, not to
* invent one out of whole cloth.
*/
static MdfdVec *
mdopen(SMgrRelation reln, ForkNumber forknum, ExtensionBehavior behavior)
{
MdfdVec *mdfd;
char *path;
File fd;
/* No work if already open */
if (reln->md_fd[forknum])
return reln->md_fd[forknum];
path = relpath(reln->smgr_rnode, forknum);
fd = PathNameOpenFile(path, O_RDWR | PG_BINARY, 0600);
if (fd < 0)
{
/*
* During bootstrap, there are cases where a system relation will be
* accessed (by internal backend processes) before the bootstrap
* script nominally creates it. Therefore, accept mdopen() as a
* substitute for mdcreate() in bootstrap mode only. (See mdcreate)
*/
if (IsBootstrapProcessingMode())
fd = PathNameOpenFile(path, O_RDWR | O_CREAT | O_EXCL | PG_BINARY, 0600);
if (fd < 0)
{
if (behavior == EXTENSION_RETURN_NULL &&
FILE_POSSIBLY_DELETED(errno))
{
pfree(path);
return NULL;
}
ereport(ERROR,
(errcode_for_file_access(),
errmsg("could not open file \"%s\": %m", path)));
}
}
pfree(path);
reln->md_fd[forknum] = mdfd = _fdvec_alloc();
mdfd->mdfd_vfd = fd;
mdfd->mdfd_segno = 0;
mdfd->mdfd_chain = NULL;
Assert(_mdnblocks(reln, forknum, mdfd) <= ((BlockNumber) RELSEG_SIZE));
return mdfd;
}
/*
* mdcreate() -- Create a new relation on magnetic disk.
*
* If isRedo is true, it's okay for the relation to exist already.
*/
bool
mdcreate(SMgrRelation reln, bool isRedo)
{
char *path;
File fd;
if (isRedo && reln->md_fd != NULL)
return true; /* created and opened already... */
Assert(reln->md_fd == NULL);
path = relpath(reln->smgr_rnode);
fd = PathNameOpenFile(path, O_RDWR | O_CREAT | O_EXCL | PG_BINARY, 0600);
if (fd < 0)
{
int save_errno = errno;
/*
* During bootstrap, there are cases where a system relation will be
* accessed (by internal backend processes) before the bootstrap
* script nominally creates it. Therefore, allow the file to exist
* already, even if isRedo is not set. (See also mdopen)
*/
if (isRedo || IsBootstrapProcessingMode())
fd = PathNameOpenFile(path, O_RDWR | PG_BINARY, 0600);
if (fd < 0)
{
pfree(path);
/* be sure to return the error reported by create, not open */
errno = save_errno;
return false;
}
errno = 0;
}
pfree(path);
reln->md_fd = _fdvec_alloc();
reln->md_fd->mdfd_vfd = fd;
reln->md_fd->mdfd_segno = 0;
#ifndef LET_OS_MANAGE_FILESIZE
reln->md_fd->mdfd_chain = NULL;
#endif
return true;
}
/*
* Do open the next segment file to read, but don't do error processing.
*
*/
static File ParquetStorageRead_DoOpenFile(ParquetStorageRead *storageRead,
char *filePathName) {
int fileFlags = O_RDONLY | PG_BINARY;
int fileMode = 0400; /* File mode is S_IRUSR 00400 user has read permission */
File file;
Assert(storageRead != NULL);
Assert(storageRead->isActive);
Assert(filePathName != NULL);
if (Debug_appendonly_print_read_block) {
elog(
LOG,
"Parquet storage read: opening table '%s', segment file '%s', fileFlags 0x%x, fileMode 0x%x",
storageRead->relationName,
storageRead->segmentFileName,
fileFlags,
fileMode);
}
/*
* Open the file for read.
*/
file = PathNameOpenFile(filePathName, fileFlags, fileMode);
return file;
}
/*
* Do open the next segment file to read, but don't do error processing.
*
* This routine is responsible for seeking to the proper
* read location given the logical EOF.
*/
static File AppendOnlyStorageRead_DoOpenFile(
AppendOnlyStorageRead *storageRead,
char *filePathName)
/* The name of the segment file to open. */
{
int fileFlags = O_RDONLY | PG_BINARY;
int fileMode = 0400;
/* File mode is S_IRUSR 00400 user has read permission */
File file;
Assert(storageRead != NULL);
Assert(storageRead->isActive);
Assert(filePathName != NULL);
elogif(Debug_appendonly_print_read_block, LOG,
"Append-Only storage read: opening table '%s', segment file '%s', fileFlags 0x%x, fileMode 0x%x",
storageRead->relationName,
storageRead->segmentFileName,
fileFlags,
fileMode);
/*
* Open the file for read.
*/
file = PathNameOpenFile(filePathName, fileFlags, fileMode);
return file;
}
static Oid
lo_import_internal(text *filename, Oid lobjOid)
{
File fd;
int nbytes,
tmp;
char buf[BUFSIZE];
char fnamebuf[MAXPGPATH];
LargeObjectDesc *lobj;
Oid oid;
#ifndef ALLOW_DANGEROUS_LO_FUNCTIONS
if (!superuser())
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
errmsg("must be superuser to use server-side lo_import()"),
errhint("Anyone can use the client-side lo_import() provided by libpq.")));
#endif
CreateFSContext();
/*
* open the file to be read in
*/
text_to_cstring_buffer(filename, fnamebuf, sizeof(fnamebuf));
fd = PathNameOpenFile(fnamebuf, O_RDONLY | PG_BINARY, S_IRWXU);
if (fd < 0)
ereport(ERROR,
(errcode_for_file_access(),
errmsg("could not open server file \"%s\": %m",
fnamebuf)));
/*
* create an inversion object
*/
oid = inv_create(lobjOid);
/*
* read in from the filesystem and write to the inversion object
*/
lobj = inv_open(oid, INV_WRITE, fscxt);
while ((nbytes = FileRead(fd, buf, BUFSIZE)) > 0)
{
tmp = inv_write(lobj, buf, nbytes);
Assert(tmp == nbytes);
}
if (nbytes < 0)
ereport(ERROR,
(errcode_for_file_access(),
errmsg("could not read server file \"%s\": %m",
fnamebuf)));
inv_close(lobj);
FileClose(fd);
return oid;
}
/*
* mdcreate() -- Create a new relation on magnetic disk.
*
* If isRedo is true, it's okay for the relation to exist already.
*/
void
mdcreate(SMgrRelation reln, ForkNumber forkNum, bool isRedo)
{
char *path;
File fd;
if (isRedo && reln->md_fd[forkNum] != NULL)
return; /* created and opened already... */
Assert(reln->md_fd[forkNum] == NULL);
path = relpath(reln->smgr_rnode, forkNum);
fd = PathNameOpenFile(path, O_RDWR | O_CREAT | O_EXCL | PG_BINARY, 0600);
if (fd < 0)
{
int save_errno = errno;
/*
* During bootstrap, there are cases where a system relation will be
* accessed (by internal backend processes) before the bootstrap
* script nominally creates it. Therefore, allow the file to exist
* already, even if isRedo is not set. (See also mdopen)
*/
if (isRedo || IsBootstrapProcessingMode())
fd = PathNameOpenFile(path, O_RDWR | PG_BINARY, 0600);
if (fd < 0)
{
/* be sure to report the error reported by create, not open */
errno = save_errno;
ereport(ERROR,
(errcode_for_file_access(),
errmsg("could not create file \"%s\": %m", path)));
}
}
pfree(path);
reln->md_fd[forkNum] = _fdvec_alloc();
reln->md_fd[forkNum]->mdfd_vfd = fd;
reln->md_fd[forkNum]->mdfd_segno = 0;
reln->md_fd[forkNum]->mdfd_chain = NULL;
}
static void DatabaseInfo_AddFile(
DatabaseInfo *info,
HTAB *dbInfoRelHashTable,
Oid tablespace,
char *dbDirPath,
char *name)
{
int64 eof;
int itemCount;
Oid relfilenode;
uint32 segmentFileNum;
char path[MAXPGPATH];
int fileFlags = O_RDONLY | PG_BINARY;
int fileMode = 0400;
/* File mode is S_IRUSR 00400 user has read permission */
File file;
sprintf(path, "%s/%s", dbDirPath, name);
/*
* Open the file for read.
*/
file = PathNameOpenFile(path, fileFlags, fileMode);
if(file < 0)
{
ereport(ERROR,
(errcode_for_file_access(),
errmsg("Could not open segment file '%s'",
path)));
}
eof = FileSeek(file, 0L, SEEK_END);
if (eof < 0) {
ereport(ERROR,
(errcode_for_file_access(),
errmsg("Could not seek to end of file \"%s\" : %m",
path)));
}
FileClose(file);
itemCount = sscanf(name, "%u.%u", &relfilenode, &segmentFileNum);
// UNDONE: sscanf is a rather poor scanner.
// UNDONE: For right now, just assume properly named files....
if (itemCount == 0)
{
DatabaseInfo_AddMiscEntry(info, tablespace, false, name);
return;
}
else if (itemCount == 1)
segmentFileNum = 0;
else
Assert(itemCount == 2);
DatabaseInfo_AddRelSegFile(info, dbInfoRelHashTable, tablespace, relfilenode, segmentFileNum, eof);
}
/*
* open a file in the database directory ($PGDATA/base/DIROID/)
*
* if we are using the system default filespace. Otherwise open
* the file in the filespace configured for temporary files.
* The passed name MUST be a relative path. Effectively, this
* prepends DatabasePath or path of the filespace to it and then
* acts like PathNameOpenFile.
*/
File
FileNameOpenFile(FileName fileName, int fileFlags, int fileMode)
{
File fd;
char *fname;
Assert(!is_absolute_path(fileName));
fname = (char*)palloc(PATH_MAX);
if (snprintf(fname, PATH_MAX, "%s/%s", getCurrentTempFilePath, fileName) > PATH_MAX)
{
ereport(ERROR, (errmsg("cannot generate path %s/%s", getCurrentTempFilePath,
fileName)));
}
fd = PathNameOpenFile(fname, fileFlags, fileMode);
pfree(fname);
return fd;
}
/*
* Open the specified segment of the relation,
* and make a MdfdVec object for it. Returns NULL on failure.
*/
static MdfdVec *
_mdfd_openseg(SMgrRelation reln, BlockNumber segno, int oflags)
{
MdfdVec *v;
int fd;
char *path,
*fullpath;
path = relpath(reln->smgr_rnode);
if (segno > 0)
{
/* be sure we have enough space for the '.segno' */
fullpath = (char *) palloc(strlen(path) + 12);
sprintf(fullpath, "%s.%u", path, segno);
pfree(path);
}
else
fullpath = path;
/* open the file */
fd = PathNameOpenFile(fullpath, O_RDWR | PG_BINARY | oflags, 0600);
pfree(fullpath);
if (fd < 0)
return NULL;
/* allocate an mdfdvec entry for it */
v = _fdvec_alloc();
/* fill the entry */
v->mdfd_vfd = fd;
v->mdfd_segno = segno;
v->mdfd_chain = NULL;
Assert(_mdnblocks(fd, BLCKSZ) <= ((BlockNumber) RELSEG_SIZE));
/* all done */
return v;
}
/*
* lo_export -
* exports an (inversion) large object.
*/
Datum
lo_export(PG_FUNCTION_ARGS)
{
Oid lobjId = PG_GETARG_OID(0);
text *filename = PG_GETARG_TEXT_PP(1);
File fd;
int nbytes,
tmp;
char buf[BUFSIZE];
char fnamebuf[MAXPGPATH];
LargeObjectDesc *lobj;
mode_t oumask;
#ifndef ALLOW_DANGEROUS_LO_FUNCTIONS
if (!superuser())
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
errmsg("must be superuser to use server-side lo_export()"),
errhint("Anyone can use the client-side lo_export() provided by libpq.")));
#endif
CreateFSContext();
/*
* open the inversion object (no need to test for failure)
*/
lobj = inv_open(lobjId, INV_READ, fscxt);
/*
* open the file to be written to
*
* Note: we reduce backend's normal 077 umask to the slightly friendlier
* 022. This code used to drop it all the way to 0, but creating
* world-writable export files doesn't seem wise.
*/
text_to_cstring_buffer(filename, fnamebuf, sizeof(fnamebuf));
oumask = umask(S_IWGRP | S_IWOTH);
fd = PathNameOpenFile(fnamebuf, O_CREAT | O_WRONLY | O_TRUNC | PG_BINARY,
S_IRUSR | S_IWUSR | S_IRGRP | S_IROTH);
umask(oumask);
if (fd < 0)
ereport(ERROR,
(errcode_for_file_access(),
errmsg("could not create server file \"%s\": %m",
fnamebuf)));
/*
* read in from the inversion file and write to the filesystem
*/
while ((nbytes = inv_read(lobj, buf, BUFSIZE)) > 0)
{
tmp = FileWrite(fd, buf, nbytes);
if (tmp != nbytes)
ereport(ERROR,
(errcode_for_file_access(),
errmsg("could not write server file \"%s\": %m",
fnamebuf)));
}
FileClose(fd);
inv_close(lobj);
PG_RETURN_INT32(1);
}
/*
* Open a temporary file in a specific tablespace.
* Subroutine for OpenTemporaryFile, which see for details.
*/
static File
OpenTemporaryFileInTablespace(Oid tblspcOid, bool rejectError)
{
char tempdirpath[MAXPGPATH];
char tempfilepath[MAXPGPATH];
File file;
/*
* Identify the tempfile directory for this tablespace.
*
* If someone tries to specify pg_global, use pg_default instead.
*/
if (tblspcOid == DEFAULTTABLESPACE_OID ||
tblspcOid == GLOBALTABLESPACE_OID)
{
/* The default tablespace is {datadir}/base */
snprintf(tempdirpath, sizeof(tempdirpath), "base/%s",
PG_TEMP_FILES_DIR);
}
else
{
/* All other tablespaces are accessed via symlinks */
snprintf(tempdirpath, sizeof(tempdirpath), "pg_tblspc/%u/%s",
tblspcOid, PG_TEMP_FILES_DIR);
}
/*
* Generate a tempfile name that should be unique within the current
* database instance.
*/
snprintf(tempfilepath, sizeof(tempfilepath), "%s/%s%d.%ld",
tempdirpath, PG_TEMP_FILE_PREFIX, MyProcPid, tempFileCounter++);
/*
* Open the file. Note: we don't use O_EXCL, in case there is an orphaned
* temp file that can be reused.
*/
file = PathNameOpenFile(tempfilepath,
O_RDWR | O_CREAT | O_TRUNC | PG_BINARY,
0600);
if (file <= 0)
{
/*
* We might need to create the tablespace's tempfile directory, if no
* one has yet done so.
*
* Don't check for error from mkdir; it could fail if someone else
* just did the same thing. If it doesn't work then we'll bomb out on
* the second create attempt, instead.
*/
mkdir(tempdirpath, S_IRWXU);
file = PathNameOpenFile(tempfilepath,
O_RDWR | O_CREAT | O_TRUNC | PG_BINARY,
0600);
if (file <= 0 && rejectError)
elog(ERROR, "could not create temporary file \"%s\": %m",
tempfilepath);
}
return file;
}
/*
* lo_import -
* imports a file as an (inversion) large object.
*/
Datum
lo_import(PG_FUNCTION_ARGS)
{
text *filename = PG_GETARG_TEXT_P(0);
File fd;
int nbytes,
tmp;
char buf[BUFSIZE];
char fnamebuf[MAXPGPATH];
LargeObjectDesc *lobj;
Oid lobjOid;
#ifndef ALLOW_DANGEROUS_LO_FUNCTIONS
if (!superuser())
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
errmsg("must be superuser to use server-side lo_import()"),
errhint("Anyone can use the client-side lo_import() provided by libpq.")));
#endif
CreateFSContext();
/*
* open the file to be read in
*/
nbytes = VARSIZE(filename) - VARHDRSZ;
if (nbytes >= MAXPGPATH)
nbytes = MAXPGPATH - 1;
memcpy(fnamebuf, VARDATA(filename), nbytes);
fnamebuf[nbytes] = '\0';
fd = PathNameOpenFile(fnamebuf, O_RDONLY | PG_BINARY, 0666);
if (fd < 0)
ereport(ERROR,
(errcode_for_file_access(),
errmsg("could not open server file \"%s\": %m",
fnamebuf)));
/*
* create an inversion object
*/
lobjOid = inv_create(InvalidOid);
/*
* read in from the filesystem and write to the inversion object
*/
lobj = inv_open(lobjOid, INV_WRITE, fscxt);
while ((nbytes = FileRead(fd, buf, BUFSIZE)) > 0)
{
tmp = inv_write(lobj, buf, nbytes);
Assert(tmp == nbytes);
}
if (nbytes < 0)
ereport(ERROR,
(errcode_for_file_access(),
errmsg("could not read server file \"%s\": %m",
fnamebuf)));
inv_close(lobj);
FileClose(fd);
PG_RETURN_OID(lobjOid);
}
/*
* Open a relation for mirrored write.
*/
static void MirroredBufferPool_DoOpen(
MirroredBufferPoolOpen *open,
/* The resulting open struct. */
RelFileNode *relFileNode,
/* The tablespace, database, and relation OIDs for the open. */
uint32 segmentFileNum,
/* Which segment file. */
char *relationName,
/* For tracing only. Can be NULL in some execution paths. */
MirrorDataLossTrackingState mirrorDataLossTrackingState,
int64 mirrorDataLossTrackingSessionNum,
bool create,
bool mirrorOnly,
bool copyToMirror,
int *primaryError,
bool *mirrorDataLossOccurred)
{
int fileFlags = O_RDWR | PG_BINARY;
int fileMode = 0600;
/*
* File mode is S_IRUSR 00400 user has read permission
* + S_IWUSR 00200 user has write permission
*/
char *primaryFilespaceLocation = NULL;
char *mirrorFilespaceLocation = NULL;
Assert(open != NULL);
*primaryError = 0;
*mirrorDataLossOccurred = false;
if (create)
fileFlags = O_CREAT | O_RDWR | PG_BINARY;
PersistentTablespace_GetPrimaryAndMirrorFilespaces(
relFileNode->spcNode,
&primaryFilespaceLocation,
&mirrorFilespaceLocation);
if (Debug_persistent_print
&&
(create || mirrorOnly || copyToMirror))
{
SUPPRESS_ERRCONTEXT_DECLARE;
SUPPRESS_ERRCONTEXT_PUSH();
elog(Persistent_DebugPrintLevel(),
"MirroredBufferPool_DoOpen: Special open %u/%u/%u --> create %s, mirrorOnly %s, copyToMirror %s, "
"primary filespace location %s "
"mirror filespace location %s ",
relFileNode->spcNode,
relFileNode->dbNode,
relFileNode->relNode,
(create ? "true" : "false"),
(mirrorOnly ? "true" : "false"),
(copyToMirror ? "true" : "false"),
(primaryFilespaceLocation == NULL) ? "<null>" : primaryFilespaceLocation,
(mirrorFilespaceLocation == NULL) ? "<null>" : mirrorFilespaceLocation);
SUPPRESS_ERRCONTEXT_POP();
}
MemSet(open, 0, sizeof(MirroredBufferPoolOpen));
open->primaryFile = -1;
if (mirrorFilespaceLocation == NULL)
sprintf(open->mirrorFilespaceLocation, "%s", "");
else
sprintf(open->mirrorFilespaceLocation, "%s", mirrorFilespaceLocation);
open->relFileNode = *relFileNode;
open->segmentFileNum = segmentFileNum;
open->create = create;
open->mirrorOnly = mirrorOnly;
open->copyToMirror = copyToMirror;
MirroredBufferPool_SetUpMirrorAccess(
relFileNode,
segmentFileNum,
relationName,
mirrorDataLossTrackingState,
mirrorDataLossTrackingSessionNum,
/* primaryOnly */ false,
mirrorOnly,
&open->mirrorMode,
&open->mirrorDataLossOccurred);
if (StorageManagerMirrorMode_DoPrimaryWork(open->mirrorMode))
{
char *dbPath;
char *path;
dbPath = (char*)palloc(MAXPGPATH + 1);
path = (char*)palloc(MAXPGPATH + 1);
/*
* Do the primary work first so we don't leave files on the mirror or have an
* open to clean up.
*/
FormDatabasePath(
dbPath,
primaryFilespaceLocation,
relFileNode->spcNode,
relFileNode->dbNode);
if (segmentFileNum == 0)
sprintf(path, "%s/%u", dbPath, relFileNode->relNode);
else
sprintf(path, "%s/%u.%u", dbPath, relFileNode->relNode, segmentFileNum);
errno = 0;
open->primaryFile = PathNameOpenFile(path, fileFlags, fileMode);
if (open->primaryFile < 0)
{
*primaryError = errno;
}
pfree(dbPath);
pfree(path);
}
if (StorageManagerMirrorMode_SendToMirror(open->mirrorMode) &&
*primaryError == 0 &&
!open->mirrorDataLossOccurred)
{
if (FileRepPrimary_MirrorOpen(
FileRep_GetRelationIdentifier(
open->mirrorFilespaceLocation,
open->relFileNode,
open->segmentFileNum),
FileRepRelationTypeBufferPool,
FILEREP_OFFSET_UNDEFINED,
fileFlags,
fileMode,
TRUE /* supressError */) != 0)
{
if (Debug_filerep_print)
ereport(LOG,
(errmsg("could not sent file open request to mirror "),
FileRep_ReportRelationPath(
open->mirrorFilespaceLocation,
open->relFileNode,
open->segmentFileNum)));
}
open->mirrorDataLossOccurred = FileRepPrimary_IsMirrorDataLossOccurred();
}
if (*primaryError != 0)
{
open->isActive = false;
}
else if (StorageManagerMirrorMode_DoPrimaryWork(open->mirrorMode))
{
open->isActive = true;
}
else if (StorageManagerMirrorMode_SendToMirror(open->mirrorMode) &&
!open->mirrorDataLossOccurred)
{
open->isActive = true;
}
*mirrorDataLossOccurred = open->mirrorDataLossOccurred;
if (primaryFilespaceLocation != NULL)
pfree(primaryFilespaceLocation);
if (mirrorFilespaceLocation != NULL)
pfree(mirrorFilespaceLocation);
}
