diff options
author | Albert Cheng <acheng@hdfgroup.org> | 2013-06-12 16:29:09 (GMT) |
---|---|---|
committer | Albert Cheng <acheng@hdfgroup.org> | 2013-06-12 16:29:09 (GMT) |
commit | d01b40e43956e8494b7b536c8cca6246b81fe887 (patch) | |
tree | d78335668556f63ae231b51a023e513b7cbb5dd1 /test | |
parent | a33955d60f3d6bd85f32b3b21f4f6ffb63ea1121 (diff) | |
download | hdf5-d01b40e43956e8494b7b536c8cca6246b81fe887.zip hdf5-d01b40e43956e8494b7b536c8cca6246b81fe887.tar.gz hdf5-d01b40e43956e8494b7b536c8cca6246b81fe887.tar.bz2 |
[svn-r23760] Implement SWMR-5:
Implemented use case 1.8 program--SWRM read/write multiple chunks at a time.
test/use_append_mchunks.c:
The program.
test/Makefile.am:
test/Makefile.in:
Added program use_append_mchunks.
test/test_usecases.sh.in:
Added simple tests for use_append_mchunks.
MANIFEST:
Updated for new file.
Tested: h5committest, except cmake, passed. Hand tested in Koala for various options.
Diffstat (limited to 'test')
-rw-r--r-- | test/Makefile.am | 5 | ||||
-rw-r--r-- | test/Makefile.in | 22 | ||||
-rw-r--r-- | test/test_usecases.sh.in | 6 | ||||
-rw-r--r-- | test/use_append_mchunks.c | 576 |
4 files changed, 601 insertions, 8 deletions
diff --git a/test/Makefile.am b/test/Makefile.am index 03241c1..397d20f 100644 --- a/test/Makefile.am +++ b/test/Makefile.am @@ -51,7 +51,7 @@ TEST_PROG= testhdf5 lheap ohdr stab gheap cache cache_api cache_tagging \ bin_PROGRAMS=swmr_generator swmr_reader swmr_writer swmr_remove_reader \ swmr_remove_writer swmr_addrem_writer swmr_sparse_reader swmr_sparse_writer \ - use_append_chunk + use_append_chunk use_append_mchunks # List programs to be built when testing here. error_test and err_compat are # built at the same time as the other tests, but executed by testerror.sh. @@ -159,7 +159,7 @@ CHECK_CLEANFILES+=accum.h5 cmpd_dset.h5 compact_dataset.h5 dataset.h5 \ new_multi_file_v16-r.h5 new_multi_file_v16-s.h5 \ split_get_file_image_test-m.h5 split_get_file_image_test-r.h5 \ file_image_core_test.h5.copy \ - swmr_data.h5 use_append_chunk.h5 \ + swmr_data.h5 use_append_chunk.h5 use_append_mchunks.h5 \ flushrefresh.h5 flushrefresh_VERIFICATION_START \ flushrefresh_VERIFICATION_CHECKPOINT1 flushrefresh_VERIFICATION_CHECKPOINT2 \ flushrefresh_VERIFICATION_DONE @@ -172,6 +172,7 @@ testhdf5_SOURCES=testhdf5.c tarray.c tattr.c tchecksum.c tconfig.c tfile.c \ # Sources for Use Cases use_append_chunk_SOURCES=use_append_chunk.c use_common.c +use_append_mchunks_SOURCES=use_append_mchunks.c use_common.c # Temporary files. DISTCLEANFILES=testerror.sh testlibinfo.sh testcheck_version.sh testlinks_env.sh test_plugin.sh \ diff --git a/test/Makefile.in b/test/Makefile.in index b3c9653..b45746f 100644 --- a/test/Makefile.in +++ b/test/Makefile.in @@ -83,7 +83,7 @@ bin_PROGRAMS = swmr_generator$(EXEEXT) swmr_reader$(EXEEXT) \ swmr_writer$(EXEEXT) swmr_remove_reader$(EXEEXT) \ swmr_remove_writer$(EXEEXT) swmr_addrem_writer$(EXEEXT) \ swmr_sparse_reader$(EXEEXT) swmr_sparse_writer$(EXEEXT) \ - use_append_chunk$(EXEEXT) + use_append_chunk$(EXEEXT) use_append_mchunks$(EXEEXT) check_PROGRAMS = $(am__EXEEXT_1) error_test$(EXEEXT) \ err_compat$(EXEEXT) tcheck_version$(EXEEXT) testmeta$(EXEEXT) \ links_env$(EXEEXT) flushrefresh$(EXEEXT) $(am__EXEEXT_2) @@ -538,6 +538,11 @@ am_use_append_chunk_OBJECTS = use_append_chunk.$(OBJEXT) \ use_append_chunk_OBJECTS = $(am_use_append_chunk_OBJECTS) use_append_chunk_LDADD = $(LDADD) use_append_chunk_DEPENDENCIES = libh5test.la $(LIBHDF5) +am_use_append_mchunks_OBJECTS = use_append_mchunks.$(OBJEXT) \ + use_common.$(OBJEXT) +use_append_mchunks_OBJECTS = $(am_use_append_mchunks_OBJECTS) +use_append_mchunks_LDADD = $(LDADD) +use_append_mchunks_DEPENDENCIES = libh5test.la $(LIBHDF5) vfd_SOURCES = vfd.c vfd_OBJECTS = vfd.$(OBJEXT) vfd_LDADD = $(LDADD) @@ -596,7 +601,8 @@ SOURCES = $(libdynlib1_la_SOURCES) $(libdynlib2_la_SOURCES) \ swmr_generator.c swmr_reader.c swmr_remove_reader.c \ swmr_remove_writer.c swmr_sparse_reader.c swmr_sparse_writer.c \ swmr_writer.c tcheck_version.c $(testhdf5_SOURCES) testmeta.c \ - $(ttsafe_SOURCES) unlink.c $(use_append_chunk_SOURCES) vfd.c + $(ttsafe_SOURCES) unlink.c $(use_append_chunk_SOURCES) \ + $(use_append_mchunks_SOURCES) vfd.c DIST_SOURCES = $(am__libdynlib1_la_SOURCES_DIST) \ $(am__libdynlib2_la_SOURCES_DIST) \ $(am__libdynlib3_la_SOURCES_DIST) $(libh5test_la_SOURCES) \ @@ -618,7 +624,8 @@ DIST_SOURCES = $(am__libdynlib1_la_SOURCES_DIST) \ swmr_generator.c swmr_reader.c swmr_remove_reader.c \ swmr_remove_writer.c swmr_sparse_reader.c swmr_sparse_writer.c \ swmr_writer.c tcheck_version.c $(testhdf5_SOURCES) testmeta.c \ - $(ttsafe_SOURCES) unlink.c $(use_append_chunk_SOURCES) vfd.c + $(ttsafe_SOURCES) unlink.c $(use_append_chunk_SOURCES) \ + $(use_append_mchunks_SOURCES) vfd.c am__can_run_installinfo = \ case $$AM_UPDATE_INFO_DIR in \ n|no|NO) false;; \ @@ -940,8 +947,8 @@ CHECK_CLEANFILES = *.chkexe *.chklog *.clog accum.h5 cmpd_dset.h5 \ earray.h5 efc[0-5].h5 log_vfd_out.log new_multi_file_v16-r.h5 \ new_multi_file_v16-s.h5 split_get_file_image_test-m.h5 \ split_get_file_image_test-r.h5 file_image_core_test.h5.copy \ - swmr_data.h5 use_append_chunk.h5 flushrefresh.h5 \ - flushrefresh_VERIFICATION_START \ + swmr_data.h5 use_append_chunk.h5 use_append_mchunks.h5 \ + flushrefresh.h5 flushrefresh_VERIFICATION_START \ flushrefresh_VERIFICATION_CHECKPOINT1 \ flushrefresh_VERIFICATION_CHECKPOINT2 \ flushrefresh_VERIFICATION_DONE @@ -1009,6 +1016,7 @@ testhdf5_SOURCES = testhdf5.c tarray.c tattr.c tchecksum.c tconfig.c tfile.c \ # Sources for Use Cases use_append_chunk_SOURCES = use_append_chunk.c use_common.c +use_append_mchunks_SOURCES = use_append_mchunks.c use_common.c # Temporary files. DISTCLEANFILES = testerror.sh testlibinfo.sh testcheck_version.sh testlinks_env.sh test_plugin.sh \ @@ -1457,6 +1465,9 @@ unlink$(EXEEXT): $(unlink_OBJECTS) $(unlink_DEPENDENCIES) $(EXTRA_unlink_DEPENDE use_append_chunk$(EXEEXT): $(use_append_chunk_OBJECTS) $(use_append_chunk_DEPENDENCIES) $(EXTRA_use_append_chunk_DEPENDENCIES) @rm -f use_append_chunk$(EXEEXT) $(AM_V_CCLD)$(LINK) $(use_append_chunk_OBJECTS) $(use_append_chunk_LDADD) $(LIBS) +use_append_mchunks$(EXEEXT): $(use_append_mchunks_OBJECTS) $(use_append_mchunks_DEPENDENCIES) $(EXTRA_use_append_mchunks_DEPENDENCIES) + @rm -f use_append_mchunks$(EXEEXT) + $(AM_V_CCLD)$(LINK) $(use_append_mchunks_OBJECTS) $(use_append_mchunks_LDADD) $(LIBS) vfd$(EXEEXT): $(vfd_OBJECTS) $(vfd_DEPENDENCIES) $(EXTRA_vfd_DEPENDENCIES) @rm -f vfd$(EXEEXT) $(AM_V_CCLD)$(LINK) $(vfd_OBJECTS) $(vfd_LDADD) $(LIBS) @@ -1587,6 +1598,7 @@ distclean-compile: @AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/tvltypes.Po@am__quote@ @AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/unlink.Po@am__quote@ @AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/use_append_chunk.Po@am__quote@ +@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/use_append_mchunks.Po@am__quote@ @AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/use_common.Po@am__quote@ @AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/vfd.Po@am__quote@ diff --git a/test/test_usecases.sh.in b/test/test_usecases.sh.in index bd98b60..771be71 100644 --- a/test/test_usecases.sh.in +++ b/test/test_usecases.sh.in @@ -31,7 +31,7 @@ EXIT_FAILURE=1 RESULT_PASSED=" PASSED" RESULT_FAILED="*FAILED*" RESULT_SKIP="-SKIP-" -USECASES_PROGRAMS=use_append_chunk +USECASES_PROGRAMS="use_append_chunk use_append_mchunks" TESTNAME="Use Case" # Define variables @@ -106,6 +106,10 @@ for p in $USECASES_PROGRAMS; do TOOLTEST $p -f $tmpfile; rm -f $tmpfile TOOLTEST $p -l w TOOLTEST $p -l r + # cleanup temp datafile + if test -z "$HDF5_NOCLEANUP"; then + rm -f $p.h5 + fi done diff --git a/test/use_append_mchunks.c b/test/use_append_mchunks.c new file mode 100644 index 0000000..85abb74 --- /dev/null +++ b/test/use_append_mchunks.c @@ -0,0 +1,576 @@ +/* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * + * Copyright by The HDF Group. * + * All rights reserved. * + * * + * This file is part of HDF5. The full HDF5 copyright notice, including * + * terms governing use, modification, and redistribution, is contained in * + * the files COPYING and Copyright.html. COPYING can be found at the root * + * of the source code distribution tree; Copyright.html can be found at the * + * root level of an installed copy of the electronic HDF5 document set and * + * is linked from the top-level documents page. It can also be found at * + * http://hdfgroup.org/HDF5/doc/Copyright.html. If you do not have * + * access to either file, you may request a copy from help@hdfgroup.org. * + * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */ + +/* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * + * Use Case 1.8 Appending a hyperslab of multiple chunks. + * Description: + * Appending a hyperslab that spans several chunks of a dataset with + * unlimited dimensions within a pre-created file and reading the new + * data back. + * Goal: + * Read data appended by the Writer to a pre-existing dataset in a + * file. The dataset has one or more unlimited dimensions. The data + * is appended by a hyperslab that is contained in several chunks (for + * example, appending 2-dim planes along the slowest changing dimension + * in the 3-dim dataset and each plane is covered by 4 chunks). + * Level: + * User Level + * Guarantees: + * o Readers will see the modified dimension sizes after the Writer + * finishes HDF5 metadata updates and issues H5Fflush or H5Oflush calls. + * o Readers will see newly appended data after the Writer finishes + * the flush operation. + * + * Preconditions: + * o Readers are not allowed to modify the file. + * o All datasets that are modified by the Writer exist when the + * Writer opens the file. + * o All datasets that are modified by the Writer exist when a Reader + * opens the file. + * + * Main Success Scenario: + * 1. An application creates a file with required objects (groups, + * datasets, and attributes). + * 2. The Writer opens the file and datasets in the file and starts + * adding data using H5Dwrite call with a hyperslab selection that + * spans several chunks. + * 3. A Reader opens the file and a dataset in a file; if the size of + * the unlimited dimension has changed, reads the appended data back. + * + * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */ + +/* Created: Albert Cheng, 2013/6/1. + * Modified: + */ + +#include "use.h" + +/* Global Variable definitions */ +options_t UC_opts; /* Use Case Options */ +const char *progname_g="use_append_mchunks"; /* program name */ + +/* Create the skeleton use case file for testing. + * It has one 3d dataset using chunked storage. + * The dataset is (unlimited, 2*chunksize, 2*chunksize). + * Dataset type is 2 bytes integer. + * It starts out "empty", i.e., first dimension is 0. + * + * Return: 0 succeed; -1 fail. + */ +int create_uc_file(void) +{ + hsize_t dims[3]; /* Dataset starting dimensions */ + hsize_t max_dims[3]; /* Dataset maximum dimensions */ + hsize_t chunk_dims[3]; /* Chunk dimensions */ + hid_t fid; /* File ID for new HDF5 file */ + hid_t dcpl; /* Dataset creation property list */ + hid_t sid; /* Dataspace ID */ + hid_t dsid; /* Dataset ID */ + + /* Create the file */ + if((fid = H5Fcreate(UC_opts.filename, H5F_ACC_TRUNC, H5P_DEFAULT, H5P_DEFAULT)) < 0) + return -1; + + /* Set up dimension sizes */ + chunk_dims[0] = 1; + dims[0] = 0; + max_dims[0] = H5S_UNLIMITED; + chunk_dims[1] = chunk_dims[2] = UC_opts.chunksize; + max_dims[1] = max_dims[2] = dims[1] = dims[2] = 2*UC_opts.chunksize; + + /* Create dataspace for creating datasets */ + if((sid = H5Screate_simple(3, dims, max_dims)) < 0) + return -1; + + /* Create dataset creation property list */ + if((dcpl = H5Pcreate(H5P_DATASET_CREATE)) < 0) + return -1; + if(H5Pset_chunk(dcpl, 3, chunk_dims) < 0) + return -1; + + /* create dataset of progname */ + if((dsid = H5Dcreate2(fid, progname_g, UC_DATATYPE, sid, H5P_DEFAULT, dcpl, H5P_DEFAULT)) < 0) + return -1; + + /* Close everythign */ + if(H5Dclose(dsid) < 0) + return -1; + + if(H5Sclose(sid) < 0) + return -1; + if(H5Fclose(fid) < 0) + return -1; + + return 0; +} + + +/* Append planes, each of (1,2*chunksize,2*chunksize) to the dataset. + * In other words, 4 chunks are appended to the dataset at a time. + * Fill each plan with the plane number and then write it at the nth plane. + * Increase the plane number and repeat till the end of dataset, when it + * reaches chunksize long. End product is a (2*chunksize)^3 cube. + * + * Return: 0 succeed; -1 fail. + */ +int write_uc_file(void) +{ + hid_t fid; /* File ID for new HDF5 file */ + hid_t dsid; /* dataset ID */ + hid_t dcpl; /* Dataset creation property list */ + char *name; + UC_CTYPE *buffer, *bufptr; /* data buffer */ + int cz=UC_opts.chunksize; /* Chunk size */ + hid_t f_sid; /* dataset file space id */ + hid_t m_sid; /* memory space id */ + int rank; /* rank */ + hsize_t chunk_dims[3]; /* Chunk dimensions */ + hsize_t dims[3]; /* Dataspace dimensions */ + hsize_t memdims[3]; /* Memory space dimensions */ + hsize_t start[3] = {0,0,0}, count[3]; /* Hyperslab selection values */ + int i, j, k; + + name = UC_opts.filename; + + /* Open the file */ + if((fid = H5Fopen(name, H5F_ACC_RDWR | (UC_opts.use_swmr ? H5F_ACC_SWMR_WRITE : 0), H5P_DEFAULT)) < 0){ + fprintf(stderr, "H5Fopen failed\n"); + return -1; + } + + /* Open the dataset of the program name */ + if((dsid = H5Dopen2(fid, progname_g, H5P_DEFAULT)) < 0){ + fprintf(stderr, "H5Dopen2 failed\n"); + return -1; + } + + /* Find chunksize used */ + if ((dcpl = H5Dget_create_plist(dsid)) < 0){ + fprintf(stderr, "H5Dget_create_plist failed\n"); + return -1; + } + if (H5D_CHUNKED != H5Pget_layout(dcpl)){ + fprintf(stderr, "storage layout is not chunked\n"); + return -1; + } + if ((rank = H5Pget_chunk(dcpl, 3, chunk_dims)) != 3){ + fprintf(stderr, "storage rank is not 3\n"); + return -1; + } + + /* verify chunk_dims is (1, chunkszie, chunksize) */ + if (chunk_dims[0]!=1 || chunk_dims[1] != cz || chunk_dims[2] != cz){ + fprintf(stderr, "chunk size is not as expected. Got dims=(%ld,%ld,%ld)\n", + (long)chunk_dims[0], (long)chunk_dims[1], (long)chunk_dims[2]); + return -1; + } + + /* allocate space for data buffer 1 X 2*chunksize X 2*chunksize of UC_CTYPE */ + memdims[0]=1; + memdims[1] = memdims[2] = 2*cz; + if ((buffer=(UC_CTYPE*)HDmalloc(4*cz*cz*sizeof(UC_CTYPE)))==NULL) { + fprintf(stderr, "malloc: failed\n"); + return -1; + }; + + /* + * Get dataset rank and dimension. + */ + f_sid = H5Dget_space(dsid); /* Get filespace handle first. */ + rank = H5Sget_simple_extent_ndims(f_sid); + if (rank != 3){ + fprintf(stderr, "rank(%d) of dataset does not match\n", rank); + return -1; + } + if (H5Sget_simple_extent_dims(f_sid, dims, NULL) < 0){ + fprintf(stderr, "H5Sget_simple_extent_dims got error\n"); + return -1; + } + printf("dataset rank %d, dimensions %lu x %lu x %lu\n", + rank, (unsigned long)(dims[0]), (unsigned long)(dims[1]), (unsigned long)(dims[2])); + /* verify that file space dims are as expected and are consistent with memory space dims */ + if (dims[0] != 0 || dims[1] != memdims[1] || dims[2] != memdims[2]){ + fprintf(stderr, "dataset is not empty. Got dims=(%ld,%ld,%ld)\n", + (long)dims[0], (long)dims[1], (long)dims[2]); + return -1; + } + + /* setup mem-space for buffer */ + if ((m_sid=H5Screate_simple(rank, memdims, NULL))<0){ + fprintf(stderr, "H5Screate_simple for memory failed\n"); + return -1; + }; + + /* write planes */ + count[0]=1; + count[1]=count[2]=2*cz; + for (i=0; i<UC_opts.nplanes; i++){ + /* fill buffer with value i+1 */ + bufptr = buffer; + for (j=0; j<2*cz; j++) + for (k=0; k<2*cz; k++) + *bufptr++ = i; + + /* extend the dataset by one for new plane */ + dims[0]=i+1; + if(H5Dset_extent(dsid, dims) < 0){ + fprintf(stderr, "H5Dset_extent failed\n"); + return -1; + } + + /* Get the dataset's dataspace */ + if((f_sid = H5Dget_space(dsid)) < 0){ + fprintf(stderr, "H5Dset_extent failed\n"); + return -1; + } + + start[0]=i; + /* Choose the next plane to write */ + if(H5Sselect_hyperslab(f_sid, H5S_SELECT_SET, start, NULL, count, NULL) < 0){ + fprintf(stderr, "Failed H5Sselect_hyperslab\n"); + return -1; + } + + /* Write plane to the dataset */ + if(H5Dwrite(dsid, UC_DATATYPE, m_sid, f_sid, H5P_DEFAULT, buffer) < 0){ + fprintf(stderr, "Failed H5Dwrite\n"); + return -1; + } + /* flush file to make the just written plane available. */ + if(H5Fflush(fid, H5F_SCOPE_GLOBAL) < 0){ + fprintf(stderr, "Failed to H5Fflush file\n"); + return -1; + } + } + + /* Done writing. Free/Close all resources including data file */ + HDfree(buffer); + if (H5Dclose(dsid) < 0){ + fprintf(stderr, "Failed to close datasete\n"); + return -1; + } + if (H5Sclose(m_sid) < 0){ + fprintf(stderr, "Failed to close memory space\n"); + return -1; + } + if (H5Sclose(f_sid) < 0){ + fprintf(stderr, "Failed to close file space\n"); + return -1; + } + if (H5Fclose(fid) < 0){ + fprintf(stderr, "Failed to close file id\n"); + return -1; + } + + return 0; +} + + +/* Read planes from the dataset. + * It expects the dataset is being changed (growing). + * It checks the unlimited dimension (1st one). When it increases, + * it will read in the new planes, one by one, and verify the data correctness. + * (The nth plan should contain all "n".) + * When the unlimited dimension grows to the chunksize (it becomes a cube), + * that is the expected end of data, the reader exits. + * + * Return: 0 succeed; -1 fail. + */ +int read_uc_file(void) +{ + hid_t fid; /* File ID for new HDF5 file */ + hid_t dsid; /* dataset ID */ + char *name; + UC_CTYPE *buffer, *bufptr; /* read data buffer */ + int cz=UC_opts.chunksize; /* Chunk size */ + hid_t f_sid; /* dataset file space id */ + hid_t m_sid; /* memory space id */ + int rank; /* rank */ + hsize_t dims[3]; /* Dataspace dimensions */ + hsize_t memdims[3]; /* Memory space dimensions */ + int nplane=0, nplane_old=0; /* nth plane, last nth plane */ + hsize_t start[3] = {0,0,0}, count[3]; /* Hyperslab selection values */ + int j, k; + int nreadererr=0; + int nerrs; + int nonewplane; + + name = UC_opts.filename; + + /* Open the file */ + if((fid = H5Fopen(name, H5F_ACC_RDONLY | (UC_opts.use_swmr ? H5F_ACC_SWMR_READ : 0), H5P_DEFAULT)) < 0){ + fprintf(stderr, "H5Fopen failed\n"); + return -1; + } + + /* Open the dataset of the program name */ + if((dsid = H5Dopen2(fid, progname_g, H5P_DEFAULT)) < 0){ + fprintf(stderr, "H5Dopen2 failed\n"); + return -1; + } + + /* allocate space for data buffer 1 X 2*chunksize X 2*chunksize of UC_CTYPE */ + memdims[0]=1; + memdims[1] = memdims[2] = 2*cz; + if ((buffer=(UC_CTYPE*)HDmalloc(4*cz*cz*sizeof(UC_CTYPE)))==NULL) { + fprintf(stderr, "malloc: failed\n"); + return -1; + }; + + /* + * Get dataset rank and dimension. + * Verify dimension is as expected (unlimited,2*chunksize,2*chunksize). + */ + f_sid = H5Dget_space(dsid); /* Get filespace handle first. */ + rank = H5Sget_simple_extent_ndims(f_sid); + if (rank != 3){ + fprintf(stderr, "rank(%d) of dataset does not match\n", rank); + return -1; + } + if (H5Sget_simple_extent_dims(f_sid, dims, NULL) < 0){ + fprintf(stderr, "H5Sget_simple_extent_dims got error\n"); + return -1; + } + printf("dataset rank %d, dimensions %lu x %lu x %lu\n", + rank, (unsigned long)(dims[0]), (unsigned long)(dims[1]), (unsigned long)(dims[2])); + /* verify that file space dims are as expected and are consistent with memory space dims */ + if (dims[1] != memdims[1] || dims[2] != memdims[2]){ + fprintf(stderr, "dataset dimension is not as expected. Got dims=(%ld,%ld,%ld)\n", + (long)dims[0], (long)dims[1], (long)dims[2]); + fprintf(stderr, "But memdims=(%ld,%ld,%ld)\n", + (long)memdims[0], (long)memdims[1], (long)memdims[2]); + return -1; + } + + /* setup mem-space for buffer */ + if ((m_sid=H5Screate_simple(rank, memdims, NULL))<0){ + fprintf(stderr, "H5Screate_simple for memory failed\n"); + return -1; + }; + + /* Read 1 plane at a time whenever the dataset grows larger + * (along dim[0]) */ + count[0]=1; + count[1]=count[2]=2*cz; + /* quit when all nplanes, default cz, have been read */ + nonewplane=0; + while (nplane_old < UC_opts.nplanes ){ + /* print progress message according to if new planes are availalbe */ + if (nplane_old < dims[0]) { + if (nonewplane){ + /* end the previous message */ + printf("\n"); + nonewplane=0; + } + printf("reading planes %d to %d\n", nplane_old, (int)dims[0]); + }else{ + if (nonewplane){ + printf("."); + if (nonewplane>=30){ + fprintf(stderr, "waited too long for new plane, quit.\n"); + return -1; + } + }else{ + /* print mesg only the first time; dots still no new plane */ + printf("no new planes to read "); + } + nonewplane++; + /* pause for a second */ + sleep(1); + } + for (nplane=nplane_old; nplane < dims[0]; nplane++){ + /* read planes between last old nplanes and current extent */ + /* Get the dataset's dataspace */ + if((f_sid = H5Dget_space(dsid)) < 0){ + fprintf(stderr, "H5Dget_space failed\n"); + return -1; + } + + start[0]=nplane; + /* Choose the next plane to read */ + if(H5Sselect_hyperslab(f_sid, H5S_SELECT_SET, start, NULL, count, NULL) < 0){ + fprintf(stderr, "H5Sselect_hyperslab failed\n"); + return -1; + } + + /* Read the plane from the dataset */ + if(H5Dread(dsid, UC_DATATYPE, m_sid, f_sid, H5P_DEFAULT, buffer) < 0){ + fprintf(stderr, "H5Dread failed\n"); + return -1; + } + + /* compare read data with expected data value which is nplane */ + bufptr = buffer; + nerrs=0; + for (j=0; j<2*cz; j++){ + for (k=0; k<2*cz; k++){ + if (*bufptr++ != nplane){ + if (++nerrs < ErrorReportMax){ + fprintf(stderr, + "found error %d plane(%d,%d), expected %d, got %d\n", + nplane, j, k, nplane, (int)*(bufptr-1)); + } + } + } + } + if (nerrs){ + nreadererr++; + fprintf(stderr, "found %d unexpected values in plane %d\n", nerrs, nplane); + } + } + /* Have read all current planes */ + nplane_old=dims[0]; + + /* check if dataset has grown since last time */ + /* close dsid and file, then reopen them */ + if (H5Dclose(dsid) < 0){ + fprintf(stderr, "H5Dclose failed\n"); + return -1; + } + if (H5Fclose(fid) < 0){ + fprintf(stderr, "H5Fclose failed\n"); + return -1; + } + if((fid = H5Fopen(name, H5F_ACC_RDONLY | (UC_opts.use_swmr ? H5F_ACC_SWMR_READ : 0), H5P_DEFAULT)) < 0){ + fprintf(stderr, "H5Fopen failed\n"); + return -1; + } + if((dsid = H5Dopen2(fid, progname_g, H5P_DEFAULT)) < 0){ + fprintf(stderr, "H5Dopen2 failed\n"); + return -1; + } + f_sid = H5Dget_space(dsid); /* Get filespace handle first. */ + if (H5Sget_simple_extent_dims(f_sid, dims, NULL) < 0){ + fprintf(stderr, "H5Sget_simple_extent_dims got error\n"); + return -1; + } + } + + if (nreadererr) + return -1; + else + return 0; +} + + +/* Overall Algorithm: + * Parse options from user; + * Generate/pre-created test files needed and close it; + * fork: child process becomes the reader process; + * while parent process continues as the writer process; + * both run till ending conditions are met. + */ +int +main(int argc, char *argv[]) +{ + pid_t childpid=0; + pid_t mypid, tmppid; + int child_status; + int child_wait_option=0; + int ret_value = 0; + int child_ret_value; + + /* initialization */ + /* use case defaults */ + HDmemset(&UC_opts, 0, sizeof(options_t)); + UC_opts.h5_use_chunks = 1; /* use chunked datasets */ + UC_opts.chunksize = Chunksize_DFT; + UC_opts.use_swmr = 1; /* use swmr open */ + + /* parse options */ + if (parse_option(argc, argv) < 0){ + Hgoto_error(1); + } + + /* ==============================================================*/ + /* UC_READWRITE: create datafile, launch both reader and writer. */ + /* UC_WRITER: create datafile, skip reader, launch writer. */ + /* UC_READER: skip create, launch reader, exit. */ + /* ==============================================================*/ + /* ============*/ + /* Create file */ + /* ============*/ + if (UC_opts.launch != UC_READER){ + printf("Creating skeleton data file for test...\n"); + if (create_uc_file() < 0){ + fprintf(stderr, "***encounter error\n"); + Hgoto_error(1); + }else + printf("File created.\n"); + } + + if (UC_opts.launch==UC_READWRITE){ + /* fork process */ + if((childpid = fork()) < 0) { + perror("fork"); + Hgoto_error(1); + }; + }; + mypid = getpid(); + + /* ============= */ + /* launch reader */ + /* ============= */ + if (UC_opts.launch != UC_WRITER){ + /* child process launch the reader */ + if(0 == childpid) { + printf("%d: launch reader process\n", mypid); + if (read_uc_file() < 0){ + fprintf(stderr, "read_uc_file encountered error\n"); + exit(1); + } + exit(0); + } + } + + /* ============= */ + /* launch writer */ + /* ============= */ + /* this process continues to launch the writer */ + printf("%d: continue as the writer process\n", mypid); + if (write_uc_file() < 0){ + fprintf(stderr, "write_uc_file encountered error\n"); + Hgoto_error(1); + } + + /* ================================================ */ + /* If readwrite, collect exit code of child process */ + /* ================================================ */ + if (UC_opts.launch == UC_READWRITE){ + if ((tmppid = waitpid(childpid, &child_status, child_wait_option)) < 0){ + perror("waitpid"); + Hgoto_error(1); + } + if (WIFEXITED(child_status)){ + if ((child_ret_value=WEXITSTATUS(child_status)) != 0){ + printf("%d: child process exited with non-zero code (%d)\n", + mypid, child_ret_value); + Hgoto_error(2); + } + } else { + printf("%d: child process terminated abnormally\n", mypid); + Hgoto_error(2); + } + } + +done: + /* Print result and exit */ + if (ret_value != 0){ + printf("Error(s) encountered\n"); + }else{ + printf("All passed\n"); + } + + return(ret_value); +} |