/*@
MPI_Info_dup - Returns a duplicate of the info object
Input Parameters:
. info - info object (handle)
Output Parameters:
. newinfo - duplicate of info object (handle)
.N fortran
@*/
int MPI_Info_dup(MPI_Info info, MPI_Info *newinfo)
{
MPI_Info curr_old, curr_new;
if ((info <= (MPI_Info) 0) || (info->cookie != MPIR_INFO_COOKIE)) {
FPRINTF(stderr, "MPI_Info_dup: Invalid info object\n");
MPI_Abort(MPI_COMM_WORLD, 1);
}
*newinfo = (MPI_Info) ADIOI_Malloc(sizeof(struct MPIR_Info));
curr_new = *newinfo;
curr_new->cookie = MPIR_INFO_COOKIE;
curr_new->key = 0;
curr_new->value = 0;
curr_new->next = 0;
curr_old = info->next;
while (curr_old) {
curr_new->next = (MPI_Info) ADIOI_Malloc(sizeof(struct MPIR_Info));
curr_new = curr_new->next;
curr_new->cookie = 0; /* cookie not set on purpose */
curr_new->key = ADIOI_Strdup(curr_old->key);
curr_new->value = ADIOI_Strdup(curr_old->value);
curr_new->next = 0;
curr_old = curr_old->next;
}
return MPI_SUCCESS;
}
/* ADIO_FileSysType_parentdir
*
* Returns pointer to string in dirnamep; that string is allocated with
* strdup and must be free()'d.
*/
static void ADIO_FileSysType_parentdir(const char *filename, char **dirnamep)
{
int err;
char *dir = NULL, *slash;
struct stat statbuf;
err = lstat(filename, &statbuf);
if (err || (!S_ISLNK(statbuf.st_mode))) {
/* no such file, or file is not a link; these are the "normal"
* cases where we can just return the parent directory.
*/
dir = ADIOI_Strdup(filename);
}
else {
/* filename is a symlink. we've presumably already tried
* to stat it and found it to be missing (dangling link),
* but this code doesn't care if the target is really there
* or not.
*/
ssize_t namelen;
char *linkbuf;
linkbuf = ADIOI_Malloc(PATH_MAX+1);
namelen = readlink(filename, linkbuf, PATH_MAX+1);
if (namelen == -1) {
/* something strange has happened between the time that
* we determined that this was a link and the time that
* we attempted to read it; punt and use the old name.
*/
dir = ADIOI_Strdup(filename);
}
else {
/* successfully read the link */
linkbuf[namelen] = '\0'; /* readlink doesn't null terminate */
dir = ADIOI_Strdup(linkbuf);
}
ADIOI_Free(linkbuf);
}
slash = strrchr(dir, '/');
if (!slash) ADIOI_Strncpy(dir, ".", 2);
else {
if (slash == dir) *(dir + 1) = '\0';
else *slash = '\0';
}
*dirnamep = dir;
return;
}
void ADIOI_BEEGFS_SetInfo( ADIO_File fd, MPI_Info users_info, int *error_code )
{
char *value, *pathname, *dname, *slash;
int flag, stripe_val[2], numtargets = 0, chunksize = 0;
struct BeegfsIoctl_MkFileWithStripeHints_Arg createFileArg;
int err, myrank, fd_pdir, perm, old_mask;
static char myname[] = "ADIOI_BEEGFS_SETINFO";
/* set error code to success */
*error_code = MPI_SUCCESS;
value = ( char * )ADIOI_Malloc( ( MPI_MAX_INFO_VAL + 1 ) * sizeof( char ) );
MPI_Comm_rank( fd->comm, &myrank );
/* set hints */
if( ( fd->info ) == MPI_INFO_NULL ) {
MPI_Info_create( &( fd->info ) );
ADIOI_Info_set( fd->info, "striping_unit", "0" );
ADIOI_Info_set( fd->info, "striping_factor", "0" );
/* set users infos */
if( users_info != MPI_INFO_NULL ) {
/* striping information */
ADIOI_Info_get( users_info, "striping_unit", MPI_MAX_INFO_VAL, value, &flag );
if( flag )
chunksize = atoi( value );
ADIOI_Info_get( users_info, "striping_factor", MPI_MAX_INFO_VAL, value, &flag );
if( flag )
numtargets = atoi( value );
/* check stripe info consistency */
if( myrank == 0 ) {
stripe_val[0] = numtargets;
stripe_val[1] = chunksize;
}
MPI_Bcast( stripe_val, 2, MPI_INT, 0, fd->comm );
if( stripe_val[0] != numtargets || stripe_val[1] != chunksize ) {
FPRINTF( stderr, "ADIOI_BEEGFS_SetInfo: All keys"
"-striping_factor:striping_unit "
"need to be identical across all processes\n" );
MPI_Abort( MPI_COMM_WORLD, 1 );
}
/* if user has specified striping info, process 0 tries to set it */
if( myrank == 0 && ( fd->access_mode & ADIO_CREATE ) && numtargets && chunksize ) {
/* open the parent dir to get/set striping info */
pathname = ADIOI_Strdup( fd->filename );
dname = strrchr( pathname, '/' );
if( dname != NULL ) {
*dname = '\0'; // replace / with nul-character
fd_pdir = open( pathname, O_RDONLY );
if( fd_pdir == -1 ) {
FPRINTF( stderr, "Error opening %s: %s\n", pathname, strerror( errno ) );
}
}
else {
/* current dir relative path */
fd_pdir = open( ".", O_RDONLY );
if( fd_pdir == -1 ) {
FPRINTF( stderr, "Error opening .: %s\n", strerror( errno ) );
}
}
ADIOI_Free( pathname );
if( fd->perm == ADIO_PERM_NULL ) {
old_mask = umask( 022 );
umask( old_mask );
perm = old_mask ^ 0666;
}
else perm = fd->perm;
/* set create hints depending on e10 hints previously set */
slash = strrchr( fd->filename, '/' );
if( slash != NULL )
slash += 1;
else
slash = fd->filename;
createFileArg.filename = slash;
createFileArg.mode = perm;
createFileArg.numtargets = numtargets;
createFileArg.chunksize = chunksize;
/* create the hint file */
err = ioctl( fd_pdir, BEEGFS_IOC_MKFILE_STRIPEHINTS, &createFileArg );
if( err ) {
FPRINTF( stderr, "BEEGFS_IOC_MKFILE_STRIPEHINTS: %s. ", strerror( errno ) );
if( errno == EEXIST ) {
/* ignore user striping and use current file info */
FPRINTF( stderr, "[rank:%d] Failure to set stripe info for %s!\n", myrank, fd->filename );
}
}
/* close the parent dir file descriptor */
close( fd_pdir );
} /* End of striping parameters validation */
}
MPI_Barrier( fd->comm );
}
/* set rest of the MPI hints (including E10 hints) */
ADIOI_GEN_SetInfo( fd, users_info, error_code );
ADIOI_Free( value );
}
/*@
MPI_Register_datarep - Register functions for user-defined data
representations
Input Parameters:
+ datarep - data representation name (string)
. read_conversion_fn - function invoked to convert from file representation to
native representation (function)
. write_conversion_fn - function invoked to convert from native representation to
file representation (function)
. dtype_file_extent_fn - function invoked to get the exted of a datatype as represented
in the file (function)
- extra_state - pointer to extra state that is passed to each of the
three functions
Notes:
This function allows the user to provide routines to convert data from
an external representation, used within a file, and the native representation,
used within the CPU. There is one predefined data representation,
'external32'. Please consult the MPI-2 standard for details on this
function.
.N fortran
@*/
int MPI_Register_datarep(ROMIO_CONST char *datarep,
MPI_Datarep_conversion_function *read_conversion_fn,
MPI_Datarep_conversion_function *write_conversion_fn,
MPI_Datarep_extent_function *dtype_file_extent_fn,
void *extra_state)
{
int error_code;
ADIOI_Datarep *adio_datarep;
static char myname[] = "MPI_REGISTER_DATAREP";
ROMIO_THREAD_CS_ENTER();
/* --BEGIN ERROR HANDLING-- */
/* check datarep name (use strlen instead of strnlen because
strnlen is not portable) */
if (datarep == NULL ||
strlen(datarep) < 1 ||
strlen(datarep) > MPI_MAX_DATAREP_STRING)
{
error_code = MPIO_Err_create_code(MPI_SUCCESS,
MPIR_ERR_RECOVERABLE,
myname, __LINE__,
MPI_ERR_ARG,
"**datarepname", 0);
error_code = MPIO_Err_return_file(MPI_FILE_NULL, error_code);
goto fn_exit;
}
/* --END ERROR HANDLING-- */
MPIR_MPIOInit(&error_code);
if (error_code != MPI_SUCCESS) goto fn_exit;
/* --BEGIN ERROR HANDLING-- */
/* check datarep isn't already registered */
for (adio_datarep = ADIOI_Datarep_head; adio_datarep; adio_datarep = adio_datarep->next) {
if (!strncmp(datarep, adio_datarep->name, MPI_MAX_DATAREP_STRING)) {
error_code = MPIO_Err_create_code(MPI_SUCCESS,
MPIR_ERR_RECOVERABLE,
myname, __LINE__,
MPI_ERR_DUP_DATAREP,
"**datarepused",
"**datarepused %s",
datarep);
error_code = MPIO_Err_return_file(MPI_FILE_NULL, error_code);
goto fn_exit;
}
}
/* Check Non-NULL Read and Write conversion function pointer */
/* Read and Write conversions are currently not supported. */
if ( (read_conversion_fn != NULL) || (write_conversion_fn != NULL) )
{
error_code = MPIO_Err_create_code(MPI_SUCCESS, MPIR_ERR_RECOVERABLE,
myname, __LINE__,
MPI_ERR_CONVERSION,
"**drconvnotsupported", 0);
error_code = MPIO_Err_return_file(MPI_FILE_NULL, error_code);
goto fn_exit;
}
/* check extent function pointer */
if (dtype_file_extent_fn == NULL)
{
error_code = MPIO_Err_create_code(MPI_SUCCESS,
MPIR_ERR_RECOVERABLE,
myname, __LINE__,
MPI_ERR_ARG,
"**datarepextent", 0);
error_code = MPIO_Err_return_file(MPI_FILE_NULL, error_code);
goto fn_exit;
}
/* --END ERROR HANDLING-- */
adio_datarep = ADIOI_Malloc(sizeof(ADIOI_Datarep));
adio_datarep->name = ADIOI_Strdup(datarep);
adio_datarep->state = extra_state;
adio_datarep->read_conv_fn = read_conversion_fn;
adio_datarep->write_conv_fn = write_conversion_fn;
adio_datarep->extent_fn = dtype_file_extent_fn;
adio_datarep->next = ADIOI_Datarep_head;
ADIOI_Datarep_head = adio_datarep;
error_code = MPI_SUCCESS;
fn_exit:
ROMIO_THREAD_CS_EXIT();
return error_code;
}
MPI_File ADIO_Open(MPI_Comm orig_comm,
MPI_Comm comm, const char *filename, int file_system,
ADIOI_Fns *ops,
int access_mode, ADIO_Offset disp, MPI_Datatype etype,
MPI_Datatype filetype,
MPI_Info info, int perm, int *error_code)
{
MPI_File mpi_fh;
ADIO_File fd;
int err, rank, procs;
static char myname[] = "ADIO_OPEN";
int max_error_code;
MPI_Info dupinfo;
int syshints_processed, can_skip;
char *p;
*error_code = MPI_SUCCESS;
/* obtain MPI_File handle */
mpi_fh = MPIO_File_create(sizeof(struct ADIOI_FileD));
if (mpi_fh == MPI_FILE_NULL) {
fd = MPI_FILE_NULL;
*error_code = MPIO_Err_create_code(*error_code,
MPIR_ERR_RECOVERABLE,
myname,
__LINE__,
MPI_ERR_OTHER,
"**nomem2",0);
goto fn_exit;
}
fd = MPIO_File_resolve(mpi_fh);
fd->cookie = ADIOI_FILE_COOKIE;
fd->fp_ind = disp;
fd->fp_sys_posn = 0;
fd->comm = comm; /* dup'ed in MPI_File_open */
fd->filename = ADIOI_Strdup(filename);
fd->file_system = file_system;
fd->fs_ptr = NULL;
fd->fns = ops;
fd->disp = disp;
fd->split_coll_count = 0;
fd->shared_fp_fd = ADIO_FILE_NULL;
fd->atomicity = 0;
fd->etype = etype; /* MPI_BYTE by default */
fd->filetype = filetype; /* MPI_BYTE by default */
fd->etype_size = 1; /* default etype is MPI_BYTE */
fd->file_realm_st_offs = NULL;
fd->file_realm_types = NULL;
fd->perm = perm;
fd->async_count = 0;
fd->fortran_handle = -1;
fd->err_handler = ADIOI_DFLT_ERR_HANDLER;
fd->io_buf_window = MPI_WIN_NULL;
fd->io_buf_put_amounts_window = MPI_WIN_NULL;
MPI_Comm_rank(comm, &rank);
MPI_Comm_size(comm, &procs);
/* create and initialize info object */
fd->hints = (ADIOI_Hints *)ADIOI_Calloc(1, sizeof(struct ADIOI_Hints_struct));
if (fd->hints == NULL) {
*error_code = MPIO_Err_create_code(*error_code,
MPIR_ERR_RECOVERABLE,
myname,
__LINE__,
MPI_ERR_OTHER,
"**nomem2",0);
goto fn_exit;
}
fd->hints->cb_config_list = NULL;
fd->hints->ranklist = NULL;
fd->hints->initialized = 0;
fd->info = MPI_INFO_NULL;
/* move system-wide hint processing *back* into open, but this time the
* hintfile reader will do a scalable read-and-broadcast. The global
* ADIOI_syshints will get initialized at first open. subsequent open
* calls will just use result from first open.
*
* We have two goals here:
* 1: avoid processing the hintfile multiple times
* 2: have all processes participate in hintfile processing (so we can read-and-broadcast)
*
* a code might do an "initialize from 0", so we can only skip hint
* processing once everyone has particpiated in hint processing */
if (ADIOI_syshints == MPI_INFO_NULL)
syshints_processed = 0;
else
syshints_processed = 1;
MPI_Allreduce(&syshints_processed, &can_skip, 1, MPI_INT, MPI_MIN, fd->comm);
if (!can_skip) {
if (ADIOI_syshints == MPI_INFO_NULL)
MPI_Info_create(&ADIOI_syshints);
ADIOI_process_system_hints(fd, ADIOI_syshints);
}
ADIOI_incorporate_system_hints(info, ADIOI_syshints, &dupinfo);
ADIO_SetInfo(fd, dupinfo, &err);
if (dupinfo != MPI_INFO_NULL) {
*error_code = MPI_Info_free(&dupinfo);
if (*error_code != MPI_SUCCESS)
goto fn_exit;
}
ADIOI_Info_set(fd->info, "romio_filesystem_type", fd->fns->fsname);
/* Instead of repeatedly allocating this buffer in collective read/write,
* allocating up-front might make memory management on small platforms
* (e.g. Blue Gene) more efficent */
fd->io_buf = ADIOI_Malloc(fd->hints->cb_buffer_size);
/* deferred open:
* we can only do this optimization if 'fd->hints->deferred_open' is set
* (which means the user hinted 'no_indep_rw' and collective buffering).
* Furthermore, we only do this if our collective read/write routines use
* our generic function, and not an fs-specific routine (we can defer opens
* only if we use our aggreagation code). */
if (fd->hints->deferred_open &&
!(uses_generic_read(fd) \
&& uses_generic_write(fd))) {
fd->hints->deferred_open = 0;
}
if (ADIO_Feature(fd, ADIO_SCALABLE_OPEN))
/* disable deferred open on these fs so that scalable broadcast
* will always use the propper communicator */
fd->hints->deferred_open = 0;
/* on BlueGene, the cb_config_list is built when hints are processed. No
* one else does that right now */
if (fd->hints->ranklist == NULL) {
build_cb_config_list(fd, orig_comm, comm, rank, procs, error_code);
if (*error_code != MPI_SUCCESS)
goto fn_exit;
}
fd->is_open = 0;
fd->my_cb_nodes_index = -2;
fd->is_agg = is_aggregator(rank, fd);
/* deferred open used to split the communicator to create an "aggregator
* communicator", but we only used it as a way to indicate that deferred
* open happened. fd->is_open and fd->is_agg are sufficient */
/* actual opens start here */
/* generic open: one process opens to create the file, all others open */
/* nfs open: everybody opens or else you'll end up with "file not found"
* due to stupid nfs consistency semantics */
/* scalable open: one process opens and broadcasts results to everyone */
ADIOI_OpenColl(fd, rank, access_mode, error_code);
/* for debugging, it can be helpful to see the hints selected. Some file
* systes set up the hints in the open call (e.g. lustre) */
p = getenv("ROMIO_PRINT_HINTS");
if (rank == 0 && p != NULL ) {
ADIOI_Info_print_keyvals(fd->info);
}
fn_exit:
MPI_Allreduce(error_code, &max_error_code, 1, MPI_INT, MPI_MAX, comm);
if (max_error_code != MPI_SUCCESS) {
/* If the file was successfully opened, close it */
if (*error_code == MPI_SUCCESS) {
/* in the deferred open case, only those who have actually
opened the file should close it */
if (fd->hints->deferred_open) {
if (fd->is_agg) {
(*(fd->fns->ADIOI_xxx_Close))(fd, error_code);
}
}
else {
(*(fd->fns->ADIOI_xxx_Close))(fd, error_code);
}
}
ADIOI_Free(fd->filename);
ADIOI_Free(fd->hints->ranklist);
ADIOI_Free(fd->hints->cb_config_list);
ADIOI_Free(fd->hints);
if (fd->info != MPI_INFO_NULL) MPI_Info_free(&(fd->info));
ADIOI_Free(fd->io_buf);
ADIOI_Free(fd);
fd = ADIO_FILE_NULL;
if (*error_code == MPI_SUCCESS)
{
*error_code = MPIO_Err_create_code(MPI_SUCCESS,
MPIR_ERR_RECOVERABLE, myname,
__LINE__, MPI_ERR_IO,
"**oremote_fail", 0);
}
}
return fd;
}
void ADIOI_PANFS_Open(ADIO_File fd, int *error_code)
{
char* value;
int perm, old_mask, amode, flag;
static char myname[] = "ADIOI_PANFS_OPEN";
if (fd->perm == ADIO_PERM_NULL) {
old_mask = umask(022);
umask(old_mask);
perm = ~old_mask & 0666;
}
else perm = fd->perm;
amode = 0;
if (fd->access_mode & ADIO_CREATE)
{
pan_fs_client_layout_agg_type_t layout_type = PAN_FS_CLIENT_LAYOUT_TYPE__DEFAULT;
unsigned long int layout_stripe_unit = 0;
unsigned long int layout_parity_stripe_width = 0;
unsigned long int layout_parity_stripe_depth = 0;
unsigned long int layout_total_num_comps = 0;
pan_fs_client_layout_visit_t layout_visit_policy = PAN_FS_CLIENT_LAYOUT_VISIT__ROUND_ROBIN;
int myrank;
MPI_Comm_rank(fd->comm, &myrank);
*error_code = MPI_SUCCESS;
value = (char *) ADIOI_Malloc((MPI_MAX_INFO_VAL+1)*sizeof(char));
ADIOI_Info_get(fd->info, "panfs_layout_type", MPI_MAX_INFO_VAL,
value, &flag);
if (flag) {
layout_type = strtoul(value,NULL,10);
}
ADIOI_Info_get(fd->info, "panfs_layout_stripe_unit", MPI_MAX_INFO_VAL,
value, &flag);
if (flag) {
layout_stripe_unit = strtoul(value,NULL,10);
}
ADIOI_Info_get(fd->info, "panfs_layout_total_num_comps", MPI_MAX_INFO_VAL,
value, &flag);
if (flag) {
layout_total_num_comps = strtoul(value,NULL,10);
}
ADIOI_Info_get(fd->info, "panfs_layout_parity_stripe_width", MPI_MAX_INFO_VAL,
value, &flag);
if (flag) {
layout_parity_stripe_width = strtoul(value,NULL,10);
}
ADIOI_Info_get(fd->info, "panfs_layout_parity_stripe_depth", MPI_MAX_INFO_VAL,
value, &flag);
if (flag) {
layout_parity_stripe_depth = strtoul(value,NULL,10);
}
ADIOI_Info_get(fd->info, "panfs_layout_visit_policy", MPI_MAX_INFO_VAL,
value, &flag);
if (flag) {
layout_visit_policy = strtoul(value,NULL,10);
}
ADIOI_Free(value);
amode = amode | O_CREAT;
/* Check for valid set of hints */
if ((layout_type < PAN_FS_CLIENT_LAYOUT_TYPE__DEFAULT) ||
(layout_type > PAN_FS_CLIENT_LAYOUT_TYPE__RAID10))
{
FPRINTF(stderr, "%s: panfs_layout_type is not a valid value: %u.\n", myname, layout_type);
MPI_Abort(MPI_COMM_WORLD, 1);
}
if ((layout_type == PAN_FS_CLIENT_LAYOUT_TYPE__RAID0) &&
((layout_stripe_unit == 0) || (layout_total_num_comps == 0)))
{
if(layout_stripe_unit == 0)
{
FPRINTF(stderr, "%s: MPI_Info does not contain the panfs_layout_stripe_unit hint which is necessary to specify a valid RAID0 layout to the PAN_FS_CLIENT_LAYOUT_CREATE_FILE ioctl.\n", myname);
}
if(layout_total_num_comps == 0)
{
FPRINTF(stderr, "%s: MPI_Info does not contain the panfs_layout_total_num_comps hint which is necessary to specify a valid RAID0 layout to the PAN_FS_CLIENT_LAYOUT_CREATE_FILE ioctl.\n", myname);
}
MPI_Abort(MPI_COMM_WORLD, 1);
}
if (layout_type == PAN_FS_CLIENT_LAYOUT_TYPE__RAID1_5_PARITY_STRIPE)
{
if ((layout_stripe_unit == 0) ||
(layout_parity_stripe_width == 0) ||
(layout_parity_stripe_depth == 0) ||
(layout_total_num_comps == 0))
{
if(layout_stripe_unit == 0)
{
FPRINTF(stderr, "%s: MPI_Info does not contain the panfs_layout_stripe_unit hint which is necessary to specify a valid RAID5 parity stripe layout to the PAN_FS_CLIENT_LAYOUT_CREATE_FILE ioctl.\n", myname);
}
if(layout_total_num_comps == 0)
{
FPRINTF(stderr, "%s: MPI_Info does not contain the panfs_layout_total_num_comps hint which is necessary to specify a valid RAID5 parity stripe layout to the PAN_FS_CLIENT_LAYOUT_CREATE_FILE ioctl.\n", myname);
}
if(layout_parity_stripe_width == 0)
{
FPRINTF(stderr, "%s: MPI_Info does not contain the panfs_layout_parity_stripe_width hint which is necessary to specify a valid RAID5 parity stripe layout to the PAN_FS_CLIENT_LAYOUT_CREATE_FILE ioctl.\n", myname);
}
if(layout_parity_stripe_depth == 0)
{
FPRINTF(stderr, "%s: MPI_Info does not contain the panfs_layout_parity_stripe_depth hint which is necessary to specify a valid RAID5 parity stripe layout to the PAN_FS_CLIENT_LAYOUT_CREATE_FILE ioctl.\n", myname);
}
MPI_Abort(MPI_COMM_WORLD, 1);
}
if ((layout_visit_policy < PAN_FS_CLIENT_LAYOUT_VISIT__ROUND_ROBIN) ||
(layout_visit_policy > PAN_FS_CLIENT_LAYOUT_VISIT__ROUND_ROBIN_WITH_HASHED_OFFSET))
{
FPRINTF(stderr, "%s: panfs_layout_visit_policy is not a valid value: %u.\n", myname, layout_visit_policy);
MPI_Abort(MPI_COMM_WORLD, 1);
}
}
if (layout_type == PAN_FS_CLIENT_LAYOUT_TYPE__RAID10)
{
if ((layout_stripe_unit == 0) || (layout_total_num_comps == 0))
{
if(layout_stripe_unit == 0)
{
FPRINTF(stderr, "%s: MPI_Info does not contain the panfs_layout_stripe_unit hint which is necessary to specify a valid RAID10 layout to the PAN_FS_CLIENT_LAYOUT_CREATE_FILE ioctl.\n", myname);
}
if(layout_total_num_comps == 0)
{
FPRINTF(stderr, "%s: MPI_Info does not contain the panfs_layout_total_num_comps hint which is necessary to specify a valid RAID10 layout to the PAN_FS_CLIENT_LAYOUT_CREATE_FILE ioctl.\n", myname);
}
MPI_Abort(MPI_COMM_WORLD, 1);
}
if ((layout_visit_policy < PAN_FS_CLIENT_LAYOUT_VISIT__ROUND_ROBIN) ||
(layout_visit_policy > PAN_FS_CLIENT_LAYOUT_VISIT__ROUND_ROBIN_WITH_HASHED_OFFSET))
{
FPRINTF(stderr, "%s: panfs_layout_visit_policy is not a valid value: %u.\n", myname, layout_visit_policy);
MPI_Abort(MPI_COMM_WORLD, 1);
}
}
/* Create the file via ioctl() or open(). ADIOI_PANFS_Open's caller
* already optimizes performance by only calling this function with
* ADIO_CREATE on rank 0. Therefore, we don't need to worry about
* implementing that optimization here. */
if((layout_type == PAN_FS_CLIENT_LAYOUT_TYPE__RAID0) || (layout_type == PAN_FS_CLIENT_LAYOUT_TYPE__RAID1_5_PARITY_STRIPE)
|| (layout_type == PAN_FS_CLIENT_LAYOUT_TYPE__RAID10)) {
pan_fs_client_layout_create_args_t file_create_args;
int fd_dir;
char* slash;
struct stat stat_buf;
int err;
char *path;
/* Check that the file does not exist before
* trying to create it. The ioctl itself should
* be able to handle this condition. Currently,
* the ioctl will return successfully if the file
* has been previously created. Filed bug 33862
* to track the problem.
*/
err = stat(fd->filename,&stat_buf);
if((err == -1) && (errno != ENOENT))
{
FPRINTF(stderr,"%s: Unexpected I/O Error calling stat() on PanFS file: %s.\n", myname, strerror(errno));
MPI_Abort(MPI_COMM_WORLD, 1);
}
else if (err == 0)
{
FPRINTF(stderr,"%s: Cannot create PanFS file with ioctl when file already exists.\n", myname);
MPI_Abort(MPI_COMM_WORLD, 1);
}
else
{
/* (err == -1) && (errno == ENOENT) */
/* File does not exist */
path = ADIOI_Strdup(fd->filename);
slash = strrchr(path, '/');
if (!slash)
ADIOI_Strncpy(path, ".", 2);
else {
if (slash == path)
*(path + 1) = '\0';
else *slash = '\0';
}
/* create PanFS object */
bzero(&file_create_args,sizeof(pan_fs_client_layout_create_args_t));
/* open directory */
fd_dir = open(path, O_RDONLY);
if (fd_dir < 0) {
FPRINTF(stderr, "%s: I/O Error opening parent directory to create PanFS file using ioctl: %s.\n", myname, strerror(errno));
MPI_Abort(MPI_COMM_WORLD, 1);
}
else
{
char *file_name_ptr = fd->filename;
slash = strrchr(fd->filename, '/');
if (slash)
{
file_name_ptr = slash + 1;
}
/* create file in the directory */
file_create_args.mode = perm;
file_create_args.version = PAN_FS_CLIENT_LAYOUT_VERSION;
file_create_args.flags = PAN_FS_CLIENT_LAYOUT_CREATE_F__NONE;
ADIOI_Strncpy(file_create_args.filename, file_name_ptr, strlen(fd->filename)+1);
file_create_args.layout.agg_type = layout_type;
file_create_args.layout.layout_is_valid = 1;
if(layout_type == PAN_FS_CLIENT_LAYOUT_TYPE__RAID1_5_PARITY_STRIPE)
{
file_create_args.layout.u.raid1_5_parity_stripe.total_num_comps = layout_total_num_comps;
file_create_args.layout.u.raid1_5_parity_stripe.parity_stripe_width = layout_parity_stripe_width;
file_create_args.layout.u.raid1_5_parity_stripe.parity_stripe_depth = layout_parity_stripe_depth;
file_create_args.layout.u.raid1_5_parity_stripe.stripe_unit = layout_stripe_unit;
file_create_args.layout.u.raid1_5_parity_stripe.layout_visit_policy = layout_visit_policy;
}
else if(layout_type == PAN_FS_CLIENT_LAYOUT_TYPE__RAID0)
{
file_create_args.layout.u.raid0.total_num_comps = layout_total_num_comps;
file_create_args.layout.u.raid0.stripe_unit = layout_stripe_unit;
}
else if(layout_type == PAN_FS_CLIENT_LAYOUT_TYPE__RAID10)
{
file_create_args.layout.u.raid10.total_num_comps = layout_total_num_comps;
file_create_args.layout.u.raid10.stripe_unit = layout_stripe_unit;
file_create_args.layout.u.raid10.layout_visit_policy = layout_visit_policy;
}
err = ioctl(fd_dir, PAN_FS_CLIENT_LAYOUT_CREATE_FILE, &file_create_args);
if (err < 0) {
FPRINTF(stderr, "%s: I/O Error doing ioctl on parent directory to create PanFS file using ioctl: %s.\n", myname, strerror(errno));
MPI_Abort(MPI_COMM_WORLD, 1);
}
err = close(fd_dir);
}
ADIOI_Free(path);
}
}
else
{
int create_fd = open(fd->filename,amode,perm);
if(create_fd != -1)
{
close(create_fd);
}
else
{
FPRINTF(stderr, "%s: I/O Error creating PanFS file using open: %s.\n", myname, strerror(errno));
MPI_Abort(MPI_COMM_WORLD, 1);
}
}
}
if (fd->access_mode & ADIO_RDONLY)
amode = amode | O_RDONLY;
if (fd->access_mode & ADIO_WRONLY)
amode = amode | O_WRONLY;
if (fd->access_mode & ADIO_RDWR)
amode = amode | O_RDWR;
if (fd->access_mode & ADIO_EXCL)
amode = amode | O_EXCL;
value = (char *) ADIOI_Malloc((MPI_MAX_INFO_VAL+1)*sizeof(char));
ADIOI_Info_get(fd->info, "panfs_concurrent_write", MPI_MAX_INFO_VAL,
value, &flag);
if (flag) {
unsigned long int concurrent_write = strtoul(value,NULL,10);
if(concurrent_write == 1)
{
amode = amode | O_CONCURRENT_WRITE;
}
}
ADIOI_Free(value);
fd->fd_sys = open(fd->filename, amode, perm);
fd->fd_direct = -1;
if (fd->fd_sys != -1)
{
int rc;
char temp_buffer[TEMP_BUFFER_SIZE];
pan_fs_client_layout_query_args_t file_query_args;
bzero(&file_query_args,sizeof(pan_fs_client_layout_query_args_t));
file_query_args.version = PAN_FS_CLIENT_LAYOUT_VERSION;
rc = ioctl(fd->fd_sys, PAN_FS_CLIENT_LAYOUT_QUERY_FILE, &file_query_args);
if (rc < 0)
{
/* Error - set layout type to unknown */
ADIOI_Info_set(fd->info, "panfs_layout_type", "PAN_FS_CLIENT_LAYOUT_TYPE__INVALID");
}
else
{
ADIOI_Snprintf(temp_buffer,TEMP_BUFFER_SIZE,"%u",file_query_args.layout.agg_type);
ADIOI_Info_set(fd->info, "panfs_layout_type", temp_buffer);
if (file_query_args.layout.layout_is_valid == 1)
{
switch (file_query_args.layout.agg_type)
{
case PAN_FS_CLIENT_LAYOUT_TYPE__RAID0:
ADIOI_Snprintf(temp_buffer,TEMP_BUFFER_SIZE,"%u",file_query_args.layout.u.raid0.stripe_unit);
ADIOI_Info_set(fd->info, "panfs_layout_stripe_unit", temp_buffer);
ADIOI_Snprintf(temp_buffer,TEMP_BUFFER_SIZE,"%u",file_query_args.layout.u.raid0.total_num_comps);
ADIOI_Info_set(fd->info, "panfs_layout_total_num_comps", temp_buffer);
break;
case PAN_FS_CLIENT_LAYOUT_TYPE__RAID1_5_PARITY_STRIPE:
ADIOI_Snprintf(temp_buffer,TEMP_BUFFER_SIZE,"%u",file_query_args.layout.u.raid1_5_parity_stripe.stripe_unit);
ADIOI_Info_set(fd->info, "panfs_layout_stripe_unit", temp_buffer);
ADIOI_Snprintf(temp_buffer,TEMP_BUFFER_SIZE,"%u",file_query_args.layout.u.raid1_5_parity_stripe.parity_stripe_width);
ADIOI_Info_set(fd->info, "panfs_layout_parity_stripe_width", temp_buffer);
ADIOI_Snprintf(temp_buffer,TEMP_BUFFER_SIZE,"%u",file_query_args.layout.u.raid1_5_parity_stripe.parity_stripe_depth);
ADIOI_Info_set(fd->info, "panfs_layout_parity_stripe_depth", temp_buffer);
ADIOI_Snprintf(temp_buffer,TEMP_BUFFER_SIZE,"%u",file_query_args.layout.u.raid1_5_parity_stripe.total_num_comps);
ADIOI_Info_set(fd->info, "panfs_layout_total_num_comps", temp_buffer);
ADIOI_Snprintf(temp_buffer,TEMP_BUFFER_SIZE,"%u",file_query_args.layout.u.raid1_5_parity_stripe.layout_visit_policy);
ADIOI_Info_set(fd->info, "panfs_layout_visit_policy", temp_buffer);
break;
case PAN_FS_CLIENT_LAYOUT_TYPE__RAID10:
ADIOI_Snprintf(temp_buffer,TEMP_BUFFER_SIZE,"%u",file_query_args.layout.u.raid10.stripe_unit);
ADIOI_Info_set(fd->info, "panfs_layout_stripe_unit", temp_buffer);
ADIOI_Snprintf(temp_buffer,TEMP_BUFFER_SIZE,"%u",file_query_args.layout.u.raid10.total_num_comps);
ADIOI_Info_set(fd->info, "panfs_layout_total_num_comps", temp_buffer);
ADIOI_Snprintf(temp_buffer,TEMP_BUFFER_SIZE,"%u",file_query_args.layout.u.raid10.layout_visit_policy);
ADIOI_Info_set(fd->info, "panfs_layout_visit_policy", temp_buffer);
break;
default:
break;
}
}
}
}
if ((fd->fd_sys != -1) && (fd->access_mode & ADIO_APPEND))
fd->fp_ind = fd->fp_sys_posn = lseek(fd->fd_sys, 0, SEEK_END);
if (fd->fd_sys == -1) {
if (errno == ENAMETOOLONG)
*error_code = MPIO_Err_create_code(MPI_SUCCESS,
MPIR_ERR_RECOVERABLE, myname,
__LINE__, MPI_ERR_BAD_FILE,
"**filenamelong",
"**filenamelong %s %d",
fd->filename,
strlen(fd->filename));
else if (errno == ENOENT)
*error_code = MPIO_Err_create_code(MPI_SUCCESS,
MPIR_ERR_RECOVERABLE, myname,
__LINE__, MPI_ERR_NO_SUCH_FILE,
"**filenoexist",
"**filenoexist %s",
fd->filename);
else if (errno == ENOTDIR || errno == ELOOP)
*error_code = MPIO_Err_create_code(MPI_SUCCESS,
MPIR_ERR_RECOVERABLE,
myname, __LINE__,
MPI_ERR_BAD_FILE,
"**filenamedir",
"**filenamedir %s",
fd->filename);
else if (errno == EACCES) {
*error_code = MPIO_Err_create_code(MPI_SUCCESS,
MPIR_ERR_RECOVERABLE, myname,
__LINE__, MPI_ERR_ACCESS,
"**fileaccess",
"**fileaccess %s",
fd->filename );
}
else if (errno == EROFS) {
/* Read only file or file system and write access requested */
*error_code = MPIO_Err_create_code(MPI_SUCCESS,
MPIR_ERR_RECOVERABLE, myname,
__LINE__, MPI_ERR_READ_ONLY,
"**ioneedrd", 0 );
}
else {
*error_code = MPIO_Err_create_code(MPI_SUCCESS,
MPIR_ERR_RECOVERABLE, myname,
__LINE__, MPI_ERR_IO, "**io",
"**io %s", strerror(errno));
}
}
else *error_code = MPI_SUCCESS;
}