diff options
Diffstat (limited to 'src/H5Dchunk.c')
| -rw-r--r-- | src/H5Dchunk.c | 1129 |
1 files changed, 948 insertions, 181 deletions
diff --git a/src/H5Dchunk.c b/src/H5Dchunk.c index ab621ca..c3a3d9d 100644 --- a/src/H5Dchunk.c +++ b/src/H5Dchunk.c @@ -56,6 +56,7 @@ #endif /* H5_HAVE_PARALLEL */ #include "H5Dpkg.h" /* Dataset functions */ #include "H5Eprivate.h" /* Error handling */ +#include "H5Fprivate.h" /* File functions */ #include "H5FLprivate.h" /* Free Lists */ #include "H5Iprivate.h" /* IDs */ #include "H5MMprivate.h" /* Memory management */ @@ -72,6 +73,13 @@ #define H5D_CHUNK_GET_NODE_INFO(map, node) (map->use_single ? map->single_chunk_info : (H5D_chunk_info_t *)H5SL_item(node)) #define H5D_CHUNK_GET_NEXT_NODE(map, node) (map->use_single ? (H5SL_node_t *)NULL : H5SL_next(node)) +/* Sanity check on chunk index types: commonly used by a lot of routines in this file */ +#define H5D_CHUNK_STORAGE_INDEX_CHK(storage) \ + HDassert((H5D_CHUNK_IDX_EARRAY == storage->idx_type && H5D_COPS_EARRAY == storage->ops) || \ + (H5D_CHUNK_IDX_FARRAY == storage->idx_type && H5D_COPS_FARRAY == storage->ops) || \ + (H5D_CHUNK_IDX_BT2 == storage->idx_type && H5D_COPS_BT2 == storage->ops) || \ + (H5D_CHUNK_IDX_BTREE == storage->idx_type && H5D_COPS_BTREE == storage->ops)); + /* * Feature: If this constant is defined then every cache preemption and load * causes a character to be printed on the standard error stream: @@ -98,11 +106,36 @@ /*#define H5D_CHUNK_DEBUG */ +/* Flags for the "edge_chunk_state" field below */ +#define H5D_RDCC_DISABLE_FILTERS 0x01u /* Disable filters on this chunk */ +#define H5D_RDCC_NEWLY_DISABLED_FILTERS 0x02u /* Filters have been disabled since + * the last flush */ + /******************/ /* Local Typedefs */ /******************/ +/* Raw data chunks are cached. Each entry in the cache is: */ +typedef struct H5D_rdcc_ent_t { + hbool_t locked; /*entry is locked in cache */ + hbool_t dirty; /*needs to be written to disk? */ + hbool_t deleted; /*chunk about to be deleted */ + unsigned edge_chunk_state; /*states related to edge chunks (see above) */ + hsize_t scaled[H5O_LAYOUT_NDIMS]; /*scaled chunk 'name' (coordinates) */ + uint32_t rd_count; /*bytes remaining to be read */ + uint32_t wr_count; /*bytes remaining to be written */ + H5F_block_t chunk_block; /*offset/length of chunk in file */ + hsize_t chunk_idx; /*index of chunk in dataset */ + uint8_t *chunk; /*the unfiltered chunk data */ + unsigned idx; /*index in hash table */ + struct H5D_rdcc_ent_t *next;/*next item in doubly-linked list */ + struct H5D_rdcc_ent_t *prev;/*previous item in doubly-linked list */ + struct H5D_rdcc_ent_t *tmp_next;/*next item in temporary doubly-linked list */ + struct H5D_rdcc_ent_t *tmp_prev;/*previous item in temporary doubly-linked list */ +} H5D_rdcc_ent_t; +typedef H5D_rdcc_ent_t *H5D_rdcc_ent_ptr_t; /* For free lists */ + /* Callback info for iteration to prune chunks */ typedef struct H5D_chunk_it_ud1_t { H5D_chunk_common_ud_t common; /* Common info for B-tree user data (must be first) */ @@ -150,6 +183,8 @@ typedef struct H5D_chunk_it_ud3_t { /* needed for compressed variable-length data */ const H5O_pline_t *pline; /* Filter pipeline */ + unsigned dset_ndims; /* Number of dimensions in dataset */ + const hsize_t *dset_dims; /* Dataset dimensions */ /* needed for copy object pointed by refs */ H5O_copy_t *cpy_info; /* Copy options */ @@ -215,7 +250,7 @@ H5D__nonexistent_readvv(const H5D_io_info_t *io_info, /* Helper routines */ static herr_t H5D__chunk_set_info_real(H5O_layout_chunk_t *layout, unsigned ndims, - const hsize_t *curr_dims); + const hsize_t *curr_dims, const hsize_t *max_dims); static void *H5D__chunk_mem_alloc(size_t size, const H5O_pline_t *pline); static void *H5D__chunk_mem_xfree(void *chk, const H5O_pline_t *pline); static void *H5D__chunk_mem_realloc(void *chk, size_t size, @@ -240,9 +275,16 @@ static herr_t H5D__chunk_flush_entry(const H5D_t *dset, hid_t dxpl_id, const H5D_dxpl_cache_t *dxpl_cache, H5D_rdcc_ent_t *ent, hbool_t reset); static herr_t H5D__chunk_cache_evict(const H5D_t *dset, hid_t dxpl_id, const H5D_dxpl_cache_t *dxpl_cache, H5D_rdcc_ent_t *ent, hbool_t flush); +static hbool_t H5D__chunk_is_partial_edge_chunk(unsigned dset_ndims, + const uint32_t *chunk_dims, const hsize_t *chunk_scaled, const hsize_t *dset_dims); +static void *H5D__chunk_lock(const H5D_io_info_t *io_info, + H5D_chunk_ud_t *udata, hbool_t relax, hbool_t prev_unfilt_chunk); +static herr_t H5D__chunk_unlock(const H5D_io_info_t *io_info, + const H5D_chunk_ud_t *udata, hbool_t dirty, void *chunk, + uint32_t naccessed); static herr_t H5D__chunk_cache_prune(const H5D_t *dset, hid_t dxpl_id, const H5D_dxpl_cache_t *dxpl_cache, size_t size); -static herr_t H5D__chunk_prune_fill(H5D_chunk_it_ud1_t *udata); +static herr_t H5D__chunk_prune_fill(H5D_chunk_it_ud1_t *udata, hbool_t new_unfilt_chunk); static herr_t H5D__chunk_file_alloc(const H5D_chk_idx_info_t *idx_info, const H5F_block_t *old_chunk, H5F_block_t *new_chunk, hbool_t *need_insert); #ifdef H5_HAVE_PARALLEL @@ -333,22 +375,35 @@ H5D__chunk_direct_write(const H5D_t *dset, hid_t dxpl_id, uint32_t filters, H5D_chk_idx_info_t idx_info; /* Chunked index info */ hsize_t scaled[H5S_MAX_RANK]; /* Scaled coordinates for this chunk */ hbool_t need_insert = FALSE; /* Whether the chunk needs to be inserted into the index */ + H5D_io_info_t io_info; /* to hold the dset and two dxpls (meta and raw data) */ + hbool_t md_dxpl_generated = FALSE; /* bool to indicate whether we should free the md_dxpl_id at exit */ herr_t ret_value = SUCCEED; /* Return value */ FUNC_ENTER_STATIC_TAG(dxpl_id, dset->oloc.addr, FAIL) + io_info.dset = dset; + io_info.raw_dxpl_id = dxpl_id; + io_info.md_dxpl_id = dxpl_id; + + /* set the dxpl IO type for sanity checking at the FD layer */ +#ifdef H5_DEBUG_BUILD + if(H5D_set_io_info_dxpls(&io_info, dxpl_id) < 0) + HGOTO_ERROR(H5E_DATASET, H5E_CANTCOPY, FAIL, "can't set metadata and raw data dxpls") + md_dxpl_generated = TRUE; +#endif /* H5_DEBUG_BUILD */ + /* Allocate dataspace and initialize it if it hasn't been. */ - if(!(*layout->ops->is_space_alloc)(&layout->storage)) + if(!(*layout->ops->is_space_alloc)(&layout->storage)) { /* Allocate storage */ - if(H5D__alloc_storage(dset, dxpl_id, H5D_ALLOC_WRITE, FALSE, NULL) < 0) + if(H5D__alloc_storage(&io_info, H5D_ALLOC_WRITE, FALSE, NULL) < 0) HGOTO_ERROR(H5E_DATASET, H5E_CANTINIT, FAIL, "unable to initialize storage") - + } /* Calculate the index of this chunk */ H5VM_chunk_scaled(dset->shared->ndims, offset, layout->u.chunk.dim, scaled); scaled[dset->shared->ndims] = 0; /* Find out the file address of the chunk (if any) */ - if(H5D__chunk_lookup(dset, dxpl_id, scaled, &udata) < 0) + if(H5D__chunk_lookup(dset, io_info.md_dxpl_id, scaled, &udata) < 0) HGOTO_ERROR(H5E_DATASET, H5E_CANTGET, FAIL, "error looking up chunk address") /* Sanity check */ @@ -366,7 +421,7 @@ H5D__chunk_direct_write(const H5D_t *dset, hid_t dxpl_id, uint32_t filters, /* Compose chunked index info struct */ idx_info.f = dset->oloc.file; - idx_info.dxpl_id = dxpl_id; + idx_info.dxpl_id = io_info.md_dxpl_id; idx_info.pline = &(dset->shared->dcpl_cache.pline); idx_info.layout = &(dset->shared->layout.u.chunk); idx_info.storage = &(dset->shared->layout.storage.u.chunk); @@ -392,15 +447,15 @@ H5D__chunk_direct_write(const H5D_t *dset, hid_t dxpl_id, uint32_t filters, const H5D_rdcc_t *rdcc = &(dset->shared->cache.chunk); /*raw data chunk cache */ /* Fill the DXPL cache values for later use */ - if(H5D__get_dxpl_cache(dxpl_id, &dxpl_cache) < 0) + if(H5D__get_dxpl_cache(io_info.raw_dxpl_id, &dxpl_cache) < 0) HGOTO_ERROR(H5E_DATASET, H5E_CANTGET, FAIL, "can't fill dxpl cache") - if(H5D__chunk_cache_evict(dset, dxpl_id, dxpl_cache, rdcc->slot[udata.idx_hint], FALSE) < 0) + if(H5D__chunk_cache_evict(dset, io_info.md_dxpl_id, dxpl_cache, rdcc->slot[udata.idx_hint], FALSE) < 0) HGOTO_ERROR(H5E_DATASET, H5E_CANTREMOVE, FAIL, "unable to evict chunk") } /* end if */ /* Write the data to the file */ - if(H5F_block_write(dset->oloc.file, H5FD_MEM_DRAW, udata.chunk_block.offset, data_size, dxpl_id, buf) < 0) + if(H5F_block_write(dset->oloc.file, H5FD_MEM_DRAW, udata.chunk_block.offset, data_size, io_info.raw_dxpl_id, buf) < 0) HGOTO_ERROR(H5E_DATASET, H5E_WRITEERROR, FAIL, "unable to write raw data to file") /* Insert the chunk record into the index */ @@ -413,6 +468,10 @@ H5D__chunk_direct_write(const H5D_t *dset, hid_t dxpl_id, uint32_t filters, } /* end if */ done: +#ifdef H5_DEBUG_BUILD + if(md_dxpl_generated && H5I_dec_ref(io_info.md_dxpl_id) < 0) + HDONE_ERROR(H5E_DATASET, H5E_CANTDEC, FAIL, "can't close metadata dxpl") +#endif /* H5_DEBUG_BUILD */ FUNC_LEAVE_NOAPI_TAG(ret_value, FAIL) } /* end H5D__chunk_direct_write() */ @@ -430,7 +489,8 @@ done: *------------------------------------------------------------------------- */ static herr_t -H5D__chunk_set_info_real(H5O_layout_chunk_t *layout, unsigned ndims, const hsize_t *curr_dims) +H5D__chunk_set_info_real(H5O_layout_chunk_t *layout, unsigned ndims, + const hsize_t *curr_dims, const hsize_t *max_dims) { unsigned u; /* Local index variable */ herr_t ret_value = SUCCEED; /* Return value */ @@ -443,17 +503,21 @@ H5D__chunk_set_info_real(H5O_layout_chunk_t *layout, unsigned ndims, const hsize HDassert(curr_dims); /* Compute the # of chunks in dataset dimensions */ - for(u = 0, layout->nchunks = 1; u < ndims; u++) { + for(u = 0, layout->nchunks = 1, layout->max_nchunks = 1; u < ndims; u++) { /* Round up to the next integer # of chunks, to accomodate partial chunks */ layout->chunks[u] = ((curr_dims[u] + layout->dim[u]) - 1) / layout->dim[u]; + layout->max_chunks[u] = ((max_dims[u] + layout->dim[u]) - 1) / layout->dim[u]; /* Accumulate the # of chunks */ layout->nchunks *= layout->chunks[u]; + layout->max_nchunks *= layout->max_chunks[u]; } /* end for */ /* Get the "down" sizes for each dimension */ if(H5VM_array_down(ndims, layout->chunks, layout->down_chunks) < 0) HGOTO_ERROR(H5E_DATASET, H5E_CANTSET, FAIL, "can't compute 'down' chunk size value") + if(H5VM_array_down(ndims, layout->max_chunks, layout->max_down_chunks) < 0) + HGOTO_ERROR(H5E_DATASET, H5E_CANTSET, FAIL, "can't compute 'down' chunk size value") done: FUNC_LEAVE_NOAPI(ret_value) @@ -483,7 +547,7 @@ H5D__chunk_set_info(const H5D_t *dset) HDassert(dset); /* Set the base layout information */ - if(H5D__chunk_set_info_real(&dset->shared->layout.u.chunk, dset->shared->ndims, dset->shared->curr_dims) < 0) + if(H5D__chunk_set_info_real(&dset->shared->layout.u.chunk, dset->shared->ndims, dset->shared->curr_dims, dset->shared->max_dims) < 0) HGOTO_ERROR(H5E_DATASET, H5E_CANTSET, FAIL, "can't set layout's chunk info") /* Call the index's "resize" callback */ @@ -496,6 +560,69 @@ done: /*------------------------------------------------------------------------- + * Function: H5D__chunk_set_sizes + * + * Purpose: Sets chunk and type sizes. + * + * Return: SUCCEED/FAIL + * + * Programmer: Dana Robinson + * December 2015 + * + *------------------------------------------------------------------------- + */ +herr_t +H5D__chunk_set_sizes(H5D_t *dset) +{ + uint64_t chunk_size; /* Size of chunk in bytes */ + unsigned max_enc_bytes_per_dim; /* Max. number of bytes required to encode this dimension */ + unsigned u; /* Iterator */ + herr_t ret_value = SUCCEED; /* Return value */ + + FUNC_ENTER_NOAPI(FAIL) + + /* Sanity checks */ + HDassert(dset); + + /* Increment # of chunk dimensions, to account for datatype size as last element */ + dset->shared->layout.u.chunk.ndims++; + + /* Set the last dimension of the chunk size to the size of the datatype */ + dset->shared->layout.u.chunk.dim[dset->shared->layout.u.chunk.ndims - 1] = (uint32_t)H5T_GET_SIZE(dset->shared->type); + + /* Compute number of bytes to use for encoding chunk dimensions */ + max_enc_bytes_per_dim = 0; + for(u = 0; u < (unsigned)dset->shared->layout.u.chunk.ndims; u++) { + unsigned enc_bytes_per_dim; /* Number of bytes required to encode this dimension */ + + /* Get encoded size of dim, in bytes */ + enc_bytes_per_dim = (H5VM_log2_gen(dset->shared->layout.u.chunk.dim[u]) + 8) / 8; + + /* Check if this is the largest value so far */ + if(enc_bytes_per_dim > max_enc_bytes_per_dim) + max_enc_bytes_per_dim = enc_bytes_per_dim; + } /* end for */ + HDassert(max_enc_bytes_per_dim > 0 && max_enc_bytes_per_dim <= 8); + dset->shared->layout.u.chunk.enc_bytes_per_dim = max_enc_bytes_per_dim; + + /* Compute and store the total size of a chunk */ + /* (Use 64-bit value to ensure that we can detect >4GB chunks) */ + for(u = 1, chunk_size = (uint64_t)dset->shared->layout.u.chunk.dim[0]; u < dset->shared->layout.u.chunk.ndims; u++) + chunk_size *= (uint64_t)dset->shared->layout.u.chunk.dim[u]; + + /* Check for chunk larger than can be represented in 32-bits */ + /* (Chunk size is encoded in 32-bit value in v1 B-tree records) */ + if(chunk_size > (uint64_t)0xffffffff) + HGOTO_ERROR(H5E_DATASET, H5E_CANTINIT, FAIL, "chunk size must be < 4GB") + + H5_CHECKED_ASSIGN(dset->shared->layout.u.chunk.size, uint32_t, chunk_size, uint64_t); + +done: + FUNC_LEAVE_NOAPI(ret_value) +} /* end H5D__chunk_set_sizes */ + + +/*------------------------------------------------------------------------- * Function: H5D__chunk_construct * * Purpose: Constructs new chunked layout information for dataset @@ -510,8 +637,6 @@ done: static herr_t H5D__chunk_construct(H5F_t H5_ATTR_UNUSED *f, H5D_t *dset) { - const H5T_t *type = dset->shared->type; /* Convenience pointer to dataset's datatype */ - uint64_t chunk_size; /* Size of chunk in bytes */ unsigned u; /* Local index variable */ herr_t ret_value = SUCCEED; /* Return value */ @@ -524,22 +649,18 @@ H5D__chunk_construct(H5F_t H5_ATTR_UNUSED *f, H5D_t *dset) /* Check for invalid chunk dimension rank */ if(0 == dset->shared->layout.u.chunk.ndims) HGOTO_ERROR(H5E_DATASET, H5E_BADVALUE, FAIL, "no chunk information set?") - - /* Set up layout information */ if(dset->shared->layout.u.chunk.ndims != dset->shared->ndims) HGOTO_ERROR(H5E_DATASET, H5E_BADVALUE, FAIL, "dimensionality of chunks doesn't match the dataspace") - /* Increment # of chunk dimensions, to account for datatype size as last element */ - dset->shared->layout.u.chunk.ndims++; + /* Set chunk sizes */ + if(H5D__chunk_set_sizes(dset) < 0) + HGOTO_ERROR(H5E_DATASET, H5E_BADVALUE, FAIL, "unable to set chunk sizes") HDassert((unsigned)(dset->shared->layout.u.chunk.ndims) <= NELMTS(dset->shared->layout.u.chunk.dim)); /* Chunked storage is not compatible with external storage (currently) */ if(dset->shared->dcpl_cache.efl.nused > 0) HGOTO_ERROR(H5E_DATASET, H5E_BADVALUE, FAIL, "external storage not supported with chunked layout") - /* Set the last dimension of the chunk size to the size of the datatype */ - dset->shared->layout.u.chunk.dim[dset->shared->layout.u.chunk.ndims - 1] = (uint32_t)H5T_GET_SIZE(type); - /* Sanity check dimensions */ for(u = 0; u < dset->shared->layout.u.chunk.ndims - 1; u++) { /* Don't allow zero-sized chunk dimensions */ @@ -555,19 +676,6 @@ H5D__chunk_construct(H5F_t H5_ATTR_UNUSED *f, H5D_t *dset) HGOTO_ERROR(H5E_DATASET, H5E_CANTINIT, FAIL, "chunk size must be <= maximum dimension size for fixed-sized dimensions") } /* end for */ - /* Compute the total size of a chunk */ - /* (Use 64-bit value to ensure that we can detect >4GB chunks) */ - for(u = 1, chunk_size = (uint64_t)dset->shared->layout.u.chunk.dim[0]; u < dset->shared->layout.u.chunk.ndims; u++) - chunk_size *= (uint64_t)dset->shared->layout.u.chunk.dim[u]; - - /* Check for chunk larger than can be represented in 32-bits */ - /* (Chunk size is encoded in 32-bit value in v1 B-tree records) */ - if(chunk_size > (uint64_t)0xffffffff) - HGOTO_ERROR(H5E_DATASET, H5E_CANTINIT, FAIL, "chunk size must be < 4GB") - - /* Retain computed chunk size */ - H5_CHECKED_ASSIGN(dset->shared->layout.u.chunk.size, uint32_t, chunk_size, uint64_t); - /* Reset address and pointer of the array struct for the chunked storage index */ if(H5D_chunk_idx_reset(&dset->shared->layout.storage.u.chunk, TRUE) < 0) HGOTO_ERROR(H5E_DATASET, H5E_CANTINIT, FAIL, "unable to reset chunked storage index") @@ -596,6 +704,7 @@ H5D__chunk_init(H5F_t *f, hid_t dxpl_id, const H5D_t *dset, hid_t dapl_id) H5D_chk_idx_info_t idx_info; /* Chunked index info */ H5D_rdcc_t *rdcc = &(dset->shared->cache.chunk); /* Convenience pointer to dataset's chunk cache */ H5P_genplist_t *dapl; /* Data access property list object pointer */ + H5O_storage_chunk_t *sc = &(dset->shared->layout.storage.u.chunk); herr_t ret_value = SUCCEED; /* Return value */ FUNC_ENTER_STATIC @@ -603,6 +712,7 @@ H5D__chunk_init(H5F_t *f, hid_t dxpl_id, const H5D_t *dset, hid_t dapl_id) /* Sanity check */ HDassert(f); HDassert(dset); + H5D_CHUNK_STORAGE_INDEX_CHK(sc); if(NULL == (dapl = (H5P_genplist_t *)H5I_object(dapl_id))) HGOTO_ERROR(H5E_ATOM, H5E_BADATOM, FAIL, "can't find object for fapl ID") @@ -686,12 +796,14 @@ done: hbool_t H5D__chunk_is_space_alloc(const H5O_storage_t *storage) { + const H5O_storage_chunk_t *sc = &(storage->u.chunk); hbool_t ret_value = FALSE; /* Return value */ FUNC_ENTER_PACKAGE_NOERR /* Sanity checks */ HDassert(storage); + H5D_CHUNK_STORAGE_INDEX_CHK(sc); /* Query index layer */ ret_value = (storage->u.chunk.ops->is_space_alloc)(&storage->u.chunk); @@ -1020,9 +1132,8 @@ H5D__chunk_mem_alloc(size_t size, const H5O_pline_t *pline) FUNC_ENTER_STATIC_NOERR HDassert(size); - HDassert(pline); - if(pline->nused > 0) + if(pline && pline->nused) ret_value = H5MM_malloc(size); else ret_value = H5FL_BLK_MALLOC(chunk, size); @@ -1050,10 +1161,8 @@ H5D__chunk_mem_xfree(void *chk, const H5O_pline_t *pline) { FUNC_ENTER_STATIC_NOERR - HDassert(pline); - if(chk) { - if(pline->nused > 0) + if(pline && pline->nused) H5MM_xfree(chk); else chk = H5FL_BLK_FREE(chunk, chk); @@ -1343,6 +1452,9 @@ H5D__create_chunk_file_map_hyper(H5D_chunk_map_t *fm, const H5D_io_info_t HDmemcpy(new_chunk_info->scaled, scaled, sizeof(hsize_t) * fm->f_ndims); new_chunk_info->scaled[fm->f_ndims] = 0; + /* Copy the chunk's scaled coordinates */ + HDmemcpy(new_chunk_info->scaled, scaled, sizeof(hsize_t) * fm->f_ndims); + /* Insert the new chunk into the skip list */ if(H5SL_insert(fm->sel_chunks, new_chunk_info, &new_chunk_info->index) < 0) { H5D__free_chunk_info(new_chunk_info, NULL, NULL); @@ -1600,6 +1712,7 @@ H5D__chunk_file_cb(void H5_ATTR_UNUSED *elem, const H5T_t H5_ATTR_UNUSED *type, /* Set the chunk's scaled coordinates */ HDmemcpy(chunk_info->scaled, scaled, sizeof(hsize_t) * fm->f_ndims); chunk_info->scaled[fm->f_ndims] = 0; + HDmemcpy(chunk_info->scaled, scaled, sizeof(hsize_t) * fm->f_ndims); /* Insert the new chunk into the skip list */ if(H5SL_insert(fm->sel_chunks,chunk_info,&chunk_info->index) < 0) { @@ -1729,15 +1842,32 @@ htri_t H5D__chunk_cacheable(const H5D_io_info_t *io_info, haddr_t caddr, hbool_t write_op) { const H5D_t *dataset = io_info->dset; /* Local pointer to dataset info */ + hbool_t has_filters = FALSE; /* Whether there are filters on the chunk or not */ htri_t ret_value = FAIL; /* Return value */ FUNC_ENTER_PACKAGE + /* Sanity check */ HDassert(io_info); HDassert(dataset); - /* Must bring the whole chunk in if there are any filters */ - if(dataset->shared->dcpl_cache.pline.nused > 0) + /* Must bring the whole chunk in if there are any filters on the chunk. + * Make sure to check if filters are on the dataset but disabled for the + * chunk because it is a partial edge chunk. */ + if(dataset->shared->dcpl_cache.pline.nused > 0) { + if(dataset->shared->layout.u.chunk.flags + & H5O_LAYOUT_CHUNK_DONT_FILTER_PARTIAL_BOUND_CHUNKS) { + has_filters = !H5D__chunk_is_partial_edge_chunk( + io_info->dset->shared->ndims, + io_info->dset->shared->layout.u.chunk.dim, + io_info->store->chunk.scaled, + io_info->dset->shared->curr_dims); + } /* end if */ + else + has_filters = TRUE; + } /* end if */ + + if(has_filters) ret_value = TRUE; else { #ifdef H5_HAVE_PARALLEL @@ -1871,7 +2001,7 @@ H5D__chunk_read(H5D_io_info_t *io_info, const H5D_type_info_t *type_info, chunk_info = H5D_CHUNK_GET_NODE_INFO(fm, chunk_node); /* Get the info for the chunk in the file */ - if(H5D__chunk_lookup(io_info->dset, io_info->dxpl_id, chunk_info->scaled, &udata) < 0) + if(H5D__chunk_lookup(io_info->dset, io_info->md_dxpl_id, chunk_info->scaled, &udata) < 0) HGOTO_ERROR(H5E_DATASET, H5E_CANTGET, FAIL, "error looking up chunk address") /* Sanity check */ @@ -1885,6 +2015,9 @@ H5D__chunk_read(H5D_io_info_t *io_info, const H5D_type_info_t *type_info, void *chunk = NULL; /* Pointer to locked chunk buffer */ htri_t cacheable; /* Whether the chunk is cacheable */ + /* Set chunk's [scaled] coordinates */ + io_info->store->chunk.scaled = chunk_info->scaled; + /* Determine if we should use the chunk cache */ if((cacheable = H5D__chunk_cacheable(io_info, udata.chunk_block.offset, FALSE)) < 0) HGOTO_ERROR(H5E_DATASET, H5E_CANTGET, FAIL, "can't tell if chunk is cacheable") @@ -1895,11 +2028,8 @@ H5D__chunk_read(H5D_io_info_t *io_info, const H5D_type_info_t *type_info, H5_CHECK_OVERFLOW(type_info->src_type_size, /*From:*/ size_t, /*To:*/ uint32_t); src_accessed_bytes = chunk_info->chunk_points * (uint32_t)type_info->src_type_size; - /* Set chunk's [scaled] coordinates */ - io_info->store->chunk.scaled = chunk_info->scaled; - /* Lock the chunk into the cache */ - if(NULL == (chunk = H5D__chunk_lock(io_info, &udata, FALSE))) + if(NULL == (chunk = H5D__chunk_lock(io_info, &udata, FALSE, FALSE))) HGOTO_ERROR(H5E_IO, H5E_READERROR, FAIL, "unable to read raw data chunk") /* Set up the storage buffer information for this chunk */ @@ -2004,13 +2134,16 @@ H5D__chunk_write(H5D_io_info_t *io_info, const H5D_type_info_t *type_info, chunk_info = H5D_CHUNK_GET_NODE_INFO(fm, chunk_node); /* Look up the chunk */ - if(H5D__chunk_lookup(io_info->dset, io_info->dxpl_id, chunk_info->scaled, &udata) < 0) + if(H5D__chunk_lookup(io_info->dset, io_info->md_dxpl_id, chunk_info->scaled, &udata) < 0) HGOTO_ERROR(H5E_DATASET, H5E_CANTGET, FAIL, "error looking up chunk address") /* Sanity check */ HDassert((H5F_addr_defined(udata.chunk_block.offset) && udata.chunk_block.length > 0) || (!H5F_addr_defined(udata.chunk_block.offset) && udata.chunk_block.length == 0)); + /* Set chunk's [scaled] coordinates */ + io_info->store->chunk.scaled = chunk_info->scaled; + /* Determine if we should use the chunk cache */ if((cacheable = H5D__chunk_cacheable(io_info, udata.chunk_block.offset, TRUE)) < 0) HGOTO_ERROR(H5E_DATASET, H5E_CANTGET, FAIL, "can't tell if chunk is cacheable") @@ -2029,11 +2162,8 @@ H5D__chunk_write(H5D_io_info_t *io_info, const H5D_type_info_t *type_info, fm->fsel_type == H5S_SEL_POINTS) entire_chunk = FALSE; - /* Set chunk's [scaled] coordinates */ - io_info->store->chunk.scaled = chunk_info->scaled; - /* Lock the chunk into the cache */ - if(NULL == (chunk = H5D__chunk_lock(io_info, &udata, entire_chunk))) + if(NULL == (chunk = H5D__chunk_lock(io_info, &udata, entire_chunk, FALSE))) HGOTO_ERROR(H5E_IO, H5E_READERROR, FAIL, "unable to read raw data chunk") /* Set up the storage buffer information for this chunk */ @@ -2047,7 +2177,7 @@ H5D__chunk_write(H5D_io_info_t *io_info, const H5D_type_info_t *type_info, if(!H5F_addr_defined(udata.chunk_block.offset)) { /* Compose chunked index info struct */ idx_info.f = io_info->dset->oloc.file; - idx_info.dxpl_id = io_info->dxpl_id; + idx_info.dxpl_id = io_info->md_dxpl_id; idx_info.pline = &(io_info->dset->shared->dcpl_cache.pline); idx_info.layout = &(io_info->dset->shared->layout.u.chunk); idx_info.storage = &(io_info->dset->shared->layout.storage.u.chunk); @@ -2221,12 +2351,14 @@ H5D__chunk_dest(H5D_t *dset, hid_t dxpl_id) H5D_rdcc_t *rdcc = &(dset->shared->cache.chunk); /* Dataset's chunk cache */ H5D_rdcc_ent_t *ent = NULL, *next = NULL; /* Pointer to current & next cache entries */ int nerrors = 0; /* Accumulated count of errors */ + H5O_storage_chunk_t *sc = &(dset->shared->layout.storage.u.chunk); herr_t ret_value = SUCCEED; /* Return value */ FUNC_ENTER_STATIC_TAG(dxpl_id, dset->oloc.addr, FAIL) /* Sanity checks */ HDassert(dset); + H5D_CHUNK_STORAGE_INDEX_CHK(sc); /* Fill the DXPL cache values for later use */ if(H5D__get_dxpl_cache(dxpl_id, &dxpl_cache) < 0) @@ -2287,6 +2419,7 @@ H5D_chunk_idx_reset(H5O_storage_chunk_t *storage, hbool_t reset_addr) /* Sanity checks */ HDassert(storage); HDassert(storage->ops); + H5D_CHUNK_STORAGE_INDEX_CHK(storage); /* Reset index structures */ if((storage->ops->reset)(storage, reset_addr) < 0) @@ -2351,6 +2484,7 @@ H5D__chunk_cinfo_cache_update(H5D_chunk_cached_t *last, const H5D_chunk_ud_t *ud HDmemcpy(last->scaled, udata->common.scaled, sizeof(hsize_t) * udata->common.layout->ndims); last->addr = udata->chunk_block.offset; H5_CHECKED_ASSIGN(last->nbytes, uint32_t, udata->chunk_block.length, hsize_t); + last->chunk_idx = udata->chunk_idx; last->filter_mask = udata->filter_mask; /* Indicate that the cached info is valid */ @@ -2397,6 +2531,7 @@ H5D__chunk_cinfo_cache_found(const H5D_chunk_cached_t *last, H5D_chunk_ud_t *uda /* Retrieve the information from the cache */ udata->chunk_block.offset = last->addr; udata->chunk_block.length = last->nbytes; + udata->chunk_idx = last->chunk_idx; udata->filter_mask = last->filter_mask; /* Indicate that the data was found */ @@ -2427,6 +2562,7 @@ herr_t H5D__chunk_create(const H5D_t *dset /*in,out*/, hid_t dxpl_id) { H5D_chk_idx_info_t idx_info; /* Chunked index info */ + H5O_storage_chunk_t *sc = &(dset->shared->layout.storage.u.chunk); herr_t ret_value = SUCCEED; /* Return value */ FUNC_ENTER_PACKAGE @@ -2435,6 +2571,8 @@ H5D__chunk_create(const H5D_t *dset /*in,out*/, hid_t dxpl_id) HDassert(dset); HDassert(H5D_CHUNKED == dset->shared->layout.type); HDassert(dset->shared->layout.u.chunk.ndims > 0 && dset->shared->layout.u.chunk.ndims <= H5O_LAYOUT_NDIMS); + H5D_CHUNK_STORAGE_INDEX_CHK(sc); + #ifndef NDEBUG { unsigned u; /* Local index variable */ @@ -2525,44 +2663,59 @@ H5D__chunk_lookup(const H5D_t *dset, hid_t dxpl_id, const hsize_t *scaled, H5D_chunk_ud_t *udata) { H5D_rdcc_ent_t *ent = NULL; /* Cache entry */ - hbool_t found = FALSE; /* In cache? */ - unsigned u; /* Counter */ - herr_t ret_value = SUCCEED; /* Return value */ + H5O_storage_chunk_t *sc = &(dset->shared->layout.storage.u.chunk); + unsigned idx; /* Index of chunk in cache, if present */ + hbool_t found = FALSE; /* In cache? */ + herr_t ret_value = SUCCEED; /* Return value */ FUNC_ENTER_PACKAGE + /* Sanity checks */ HDassert(dset); HDassert(dset->shared->layout.u.chunk.ndims > 0); + H5D_CHUNK_STORAGE_INDEX_CHK(sc); HDassert(scaled); HDassert(udata); /* Initialize the query information about the chunk we are looking for */ udata->common.layout = &(dset->shared->layout.u.chunk); udata->common.storage = &(dset->shared->layout.storage.u.chunk); - udata->common.scaled = scaled; + udata->common.scaled = scaled; /* Reset information about the chunk we are looking for */ udata->chunk_block.offset = HADDR_UNDEF; udata->chunk_block.length = 0; udata->filter_mask = 0; + udata->new_unfilt_chunk = FALSE; /* Check for chunk in cache */ if(dset->shared->cache.chunk.nslots > 0) { - udata->idx_hint = H5D__chunk_hash_val(dset->shared, scaled); - ent = dset->shared->cache.chunk.slot[udata->idx_hint]; + /* Determine the chunk's location in the hash table */ + idx = H5D__chunk_hash_val(dset->shared, scaled); + + /* Get the chunk cache entry for that location */ + ent = dset->shared->cache.chunk.slot[idx]; + if(ent) { + unsigned u; /* Counter */ - if(ent) - for(u = 0, found = TRUE; u < dset->shared->ndims; u++) + /* Speculatively set the 'found' flag */ + found = TRUE; + + /* Verify that the cache entry is the correct chunk */ + for(u = 0; u < dset->shared->ndims; u++) if(scaled[u] != ent->scaled[u]) { found = FALSE; break; } /* end if */ + } /* end if */ } /* end if */ - /* Find chunk addr */ + /* Retrieve chunk addr */ if(found) { + udata->idx_hint = idx; udata->chunk_block.offset = ent->chunk_block.offset; udata->chunk_block.length = ent->chunk_block.length;; + udata->chunk_idx = ent->chunk_idx; } /* end if */ else { /* Invalidate idx_hint, to signal that the chunk is not in cache */ @@ -2571,6 +2724,9 @@ H5D__chunk_lookup(const H5D_t *dset, hid_t dxpl_id, const hsize_t *scaled, /* Check for cached information */ if(!H5D__chunk_cinfo_cache_found(&dset->shared->cache.chunk.last, udata)) { H5D_chk_idx_info_t idx_info; /* Chunked index info */ +#ifdef H5_HAVE_PARALLEL + H5P_coll_md_read_flag_t temp_cmr; /* Temp value to hold the coll metadata read setting */ +#endif /* H5_HAVE_PARALLEL */ /* Compose chunked index info struct */ idx_info.f = dset->oloc.file; @@ -2579,10 +2735,26 @@ H5D__chunk_lookup(const H5D_t *dset, hid_t dxpl_id, const hsize_t *scaled, idx_info.layout = &dset->shared->layout.u.chunk; idx_info.storage = &dset->shared->layout.storage.u.chunk; +#ifdef H5_HAVE_PARALLEL + if(H5F_HAS_FEATURE(idx_info.f, H5FD_FEAT_HAS_MPI)) { + /* disable collective metadata read for chunk indexes + as it is highly unlikely that users would read the + same chunks from all processes. MSC - might turn on + for root node? */ + temp_cmr = H5F_COLL_MD_READ(idx_info.f); + H5F_set_coll_md_read(idx_info.f, H5P_FORCE_FALSE); + } /* end if */ +#endif /* H5_HAVE_PARALLEL */ + /* Go get the chunk information */ if((dset->shared->layout.storage.u.chunk.ops->get_addr)(&idx_info, udata) < 0) HGOTO_ERROR(H5E_DATASET, H5E_CANTGET, FAIL, "can't query chunk address") +#ifdef H5_HAVE_PARALLEL + if(H5F_HAS_FEATURE(idx_info.f, H5FD_FEAT_HAS_MPI)) + H5F_set_coll_md_read(idx_info.f, temp_cmr); +#endif /* H5_HAVE_PARALLEL */ + /* Cache the information retrieved */ H5D__chunk_cinfo_cache_update(&dset->shared->cache.chunk.last, udata); } /* end if */ @@ -2614,12 +2786,14 @@ H5D__chunk_flush_entry(const H5D_t *dset, hid_t dxpl_id, const H5D_dxpl_cache_t { void *buf = NULL; /* Temporary buffer */ hbool_t point_of_no_return = FALSE; + H5O_storage_chunk_t *sc = &(dset->shared->layout.storage.u.chunk); herr_t ret_value = SUCCEED; /* Return value */ FUNC_ENTER_STATIC_TAG(dxpl_id, dset->oloc.addr, FAIL) HDassert(dset); HDassert(dset->shared); + H5D_CHUNK_STORAGE_INDEX_CHK(sc); HDassert(dxpl_cache); HDassert(ent); HDassert(!ent->locked); @@ -2638,9 +2812,11 @@ H5D__chunk_flush_entry(const H5D_t *dset, hid_t dxpl_id, const H5D_dxpl_cache_t udata.chunk_block.offset = ent->chunk_block.offset; udata.chunk_block.length = dset->shared->layout.u.chunk.size; udata.filter_mask = 0; + udata.chunk_idx = ent->chunk_idx; /* Should the chunk be filtered before writing it to disk? */ - if(dset->shared->dcpl_cache.pline.nused) { + if(dset->shared->dcpl_cache.pline.nused + && !(ent->edge_chunk_state & H5D_RDCC_DISABLE_FILTERS)) { size_t alloc = udata.chunk_block.length; /* Bytes allocated for BUF */ size_t nbytes; /* Chunk size (in bytes) */ @@ -2679,10 +2855,27 @@ H5D__chunk_flush_entry(const H5D_t *dset, hid_t dxpl_id, const H5D_dxpl_cache_t /* Indicate that the chunk must be allocated */ must_alloc = TRUE; } /* end if */ - else if(!H5F_addr_defined(udata.chunk_block.offset)) + else if(!H5F_addr_defined(udata.chunk_block.offset)) { /* Indicate that the chunk must be allocated */ must_alloc = TRUE; + /* This flag could be set for this chunk, just remove and ignore it + */ + ent->edge_chunk_state &= ~H5D_RDCC_NEWLY_DISABLED_FILTERS; + } /* end else */ + else if(ent->edge_chunk_state & H5D_RDCC_NEWLY_DISABLED_FILTERS) { + /* Chunk on disk is still filtered, must insert to allocate correct + * size */ + must_alloc = TRUE; + + /* Set the disable filters field back to the standard disable + * filters setting, as it no longer needs to be inserted with every + * flush */ + ent->edge_chunk_state &= ~H5D_RDCC_NEWLY_DISABLED_FILTERS; + } /* end else */ + + HDassert(!(ent->edge_chunk_state & H5D_RDCC_NEWLY_DISABLED_FILTERS)); + /* Check if the chunk needs to be allocated (it also could exist already * and the chunk alloc operation could resize it) */ @@ -2708,7 +2901,7 @@ H5D__chunk_flush_entry(const H5D_t *dset, hid_t dxpl_id, const H5D_dxpl_cache_t /* Write the data to the file */ HDassert(H5F_addr_defined(udata.chunk_block.offset)); H5_CHECK_OVERFLOW(udata.chunk_block.length, hsize_t, size_t); - if(H5F_block_write(dset->oloc.file, H5FD_MEM_DRAW, udata.chunk_block.offset, (size_t)udata.chunk_block.length, dxpl_id, buf) < 0) + if(H5F_block_write(dset->oloc.file, H5FD_MEM_DRAW, udata.chunk_block.offset, (size_t)udata.chunk_block.length, H5AC_rawdata_dxpl_id, buf) < 0) HGOTO_ERROR(H5E_DATASET, H5E_WRITEERROR, FAIL, "unable to write raw data to file") /* Insert the chunk record into the index */ @@ -2732,7 +2925,9 @@ H5D__chunk_flush_entry(const H5D_t *dset, hid_t dxpl_id, const H5D_dxpl_cache_t if(buf == ent->chunk) buf = NULL; if(ent->chunk != NULL) - ent->chunk = (uint8_t *)H5D__chunk_mem_xfree(ent->chunk, &(dset->shared->dcpl_cache.pline)); + ent->chunk = (uint8_t *)H5D__chunk_mem_xfree(ent->chunk, + ((ent->edge_chunk_state & H5D_RDCC_DISABLE_FILTERS) ? NULL + : &(dset->shared->dcpl_cache.pline))); } /* end if */ done: @@ -2748,7 +2943,9 @@ done: */ if(ret_value < 0 && point_of_no_return) if(ent->chunk) - ent->chunk = (uint8_t *)H5D__chunk_mem_xfree(ent->chunk, &(dset->shared->dcpl_cache.pline)); + ent->chunk = (uint8_t *)H5D__chunk_mem_xfree(ent->chunk, + ((ent->edge_chunk_state & H5D_RDCC_DISABLE_FILTERS) ? NULL + : &(dset->shared->dcpl_cache.pline))); FUNC_LEAVE_NOAPI_TAG(ret_value, FAIL) } /* end H5D__chunk_flush_entry() */ @@ -2790,7 +2987,9 @@ H5D__chunk_cache_evict(const H5D_t *dset, hid_t dxpl_id, const H5D_dxpl_cache_t else { /* Don't flush, just free chunk */ if(ent->chunk != NULL) - ent->chunk = (uint8_t *)H5D__chunk_mem_xfree(ent->chunk, &(dset->shared->dcpl_cache.pline)); + ent->chunk = (uint8_t *)H5D__chunk_mem_xfree(ent->chunk, + ((ent->edge_chunk_state & H5D_RDCC_DISABLE_FILTERS) ? NULL + : &(dset->shared->dcpl_cache.pline))); } /* end else */ /* Unlink from list */ @@ -2804,8 +3003,19 @@ H5D__chunk_cache_evict(const H5D_t *dset, hid_t dxpl_id, const H5D_dxpl_cache_t rdcc->tail = ent->prev; ent->prev = ent->next = NULL; - /* Only clear hash table slot if chunk was not marked as deleted already */ - if(!ent->deleted) + /* Unlink from temporary list */ + if(ent->tmp_prev) { + HDassert(rdcc->tmp_head->tmp_next); + ent->tmp_prev->tmp_next = ent->tmp_next; + if(ent->tmp_next) { + ent->tmp_next->tmp_prev = ent->tmp_prev; + ent->tmp_next = NULL; + } /* end if */ + ent->tmp_prev = NULL; + } /* end if */ + else + /* Only clear hash table slot if the chunk was not on the temporary list + */ rdcc->slot[ent->idx] = NULL; /* Remove from cache */ @@ -2956,12 +3166,13 @@ done: * *------------------------------------------------------------------------- */ -void * +static void * H5D__chunk_lock(const H5D_io_info_t *io_info, H5D_chunk_ud_t *udata, - hbool_t relax) + hbool_t relax, hbool_t prev_unfilt_chunk) { const H5D_t *dset = io_info->dset; /* Local pointer to the dataset info */ const H5O_pline_t *pline = &(dset->shared->dcpl_cache.pline); /* I/O pipeline info - always equal to the pline passed to H5D__chunk_mem_alloc */ + const H5O_pline_t *old_pline = pline; /* Old pipeline, i.e. pipeline used to read the chunk */ const H5O_layout_t *layout = &(dset->shared->layout); /* Dataset layout */ const H5O_fill_t *fill = &(dset->shared->dcpl_cache.fill); /* Fill value info */ H5D_fill_buf_info_t fb_info; /* Dataset's fill buffer info */ @@ -2969,17 +3180,21 @@ H5D__chunk_lock(const H5D_io_info_t *io_info, H5D_chunk_ud_t *udata, H5D_rdcc_t *rdcc = &(dset->shared->cache.chunk); /*raw data chunk cache*/ H5D_rdcc_ent_t *ent; /*cache entry */ size_t chunk_size; /*size of a chunk */ + hbool_t disable_filters = FALSE; /* Whether to disable filters (when adding to cache) */ void *chunk = NULL; /*the file chunk */ void *ret_value = NULL; /* Return value */ - FUNC_ENTER_PACKAGE + FUNC_ENTER_STATIC HDassert(io_info); HDassert(io_info->dxpl_cache); HDassert(io_info->store); HDassert(udata); HDassert(dset); - HDassert(TRUE == H5P_isa_class(io_info->dxpl_id, H5P_DATASET_XFER)); + HDassert(TRUE == H5P_isa_class(io_info->md_dxpl_id, H5P_DATASET_XFER)); + HDassert(TRUE == H5P_isa_class(io_info->raw_dxpl_id, H5P_DATASET_XFER)); + HDassert(!(udata->new_unfilt_chunk && prev_unfilt_chunk)); + HDassert(!rdcc->tmp_head); /* Get the chunk's size */ HDassert(layout->u.chunk.size > 0); @@ -2999,7 +3214,7 @@ H5D__chunk_lock(const H5D_io_info_t *io_info, H5D_chunk_ud_t *udata, unsigned u; /*counters */ /* Make sure this is the right chunk */ - for(u = 0; u < layout->u.chunk.ndims; u++) + for(u = 0; u < layout->u.chunk.ndims - 1; u++) HDassert(io_info->store->chunk.scaled[u] == ent->scaled[u]); } #endif /* NDEBUG */ @@ -3009,6 +3224,66 @@ H5D__chunk_lock(const H5D_io_info_t *io_info, H5D_chunk_ud_t *udata, */ rdcc->stats.nhits++; + /* Make adjustments if the edge chunk status changed recently */ + if(pline->nused) { + /* If the chunk recently became an unfiltered partial edge chunk + * while in cache, we must make some changes to the entry */ + if(udata->new_unfilt_chunk) { + /* If this flag is set then partial chunk filters must be + * disabled, and the chunk must not have previously been a + * partial chunk (with disabled filters) */ + HDassert(layout->u.chunk.flags + & H5O_LAYOUT_CHUNK_DONT_FILTER_PARTIAL_BOUND_CHUNKS); + HDassert(!(ent->edge_chunk_state & H5D_RDCC_DISABLE_FILTERS)); + HDassert(old_pline->nused); + + /* Disable filters. Set pline to NULL instead of just the + * default pipeline to make a quick failure more likely if the + * code is changed in an inappropriate/incomplete way. */ + pline = NULL; + + /* Reallocate the chunk so H5D__chunk_mem_xfree doesn't get confused + */ + if(NULL == (chunk = H5D__chunk_mem_alloc(chunk_size, pline))) + HGOTO_ERROR(H5E_RESOURCE, H5E_NOSPACE, NULL, "memory allocation failed for raw data chunk") + HDmemcpy(chunk, ent->chunk, chunk_size); + ent->chunk = (uint8_t *)H5D__chunk_mem_xfree(ent->chunk, old_pline); + ent->chunk = (uint8_t *)chunk; + chunk = NULL; + + /* Mark the chunk as having filters disabled as well as "newly + * disabled" so it is inserted on flush */ + ent->edge_chunk_state |= H5D_RDCC_DISABLE_FILTERS; + ent->edge_chunk_state |= H5D_RDCC_NEWLY_DISABLED_FILTERS; + } /* end if */ + else if(prev_unfilt_chunk) { + /* If this flag is set then partial chunk filters must be + * disabled, and the chunk must have previously been a partial + * chunk (with disabled filters) */ + HDassert(layout->u.chunk.flags + & H5O_LAYOUT_CHUNK_DONT_FILTER_PARTIAL_BOUND_CHUNKS); + HDassert((ent->edge_chunk_state & H5D_RDCC_DISABLE_FILTERS)); + HDassert(pline->nused); + + /* Mark the old pipeline as having been disabled */ + old_pline = NULL; + + /* Reallocate the chunk so H5D__chunk_mem_xfree doesn't get confused + */ + if(NULL == (chunk = H5D__chunk_mem_alloc(chunk_size, pline))) + HGOTO_ERROR(H5E_RESOURCE, H5E_NOSPACE, NULL, "memory allocation failed for raw data chunk") + HDmemcpy(chunk, ent->chunk, chunk_size); + + ent->chunk = (uint8_t *)H5D__chunk_mem_xfree(ent->chunk, old_pline); + ent->chunk = (uint8_t *)chunk; + chunk = NULL; + + /* Mark the chunk as having filters enabled */ + ent->edge_chunk_state &= ~(H5D_RDCC_DISABLE_FILTERS + | H5D_RDCC_NEWLY_DISABLED_FILTERS); + } /* end else */ + } /* end if */ + /* * If the chunk is not at the beginning of the cache; move it backward * by one slot. This is how we implement the LRU preemption @@ -3037,6 +3312,39 @@ H5D__chunk_lock(const H5D_io_info_t *io_info, H5D_chunk_ud_t *udata, chunk_addr = udata->chunk_block.offset; chunk_alloc = udata->chunk_block.length; + /* Check if we should disable filters on this chunk */ + if(pline->nused) { + if(udata->new_unfilt_chunk) { + HDassert(layout->u.chunk.flags + & H5O_LAYOUT_CHUNK_DONT_FILTER_PARTIAL_BOUND_CHUNKS); + + /* Disable the filters for writing */ + disable_filters = TRUE; + pline = NULL; + } /* end if */ + else if(prev_unfilt_chunk) { + HDassert(layout->u.chunk.flags + & H5O_LAYOUT_CHUNK_DONT_FILTER_PARTIAL_BOUND_CHUNKS); + + /* Mark the filters as having been previously disabled (for the + * chunk as currently on disk) - disable the filters for reading + */ + old_pline = NULL; + } /* end if */ + else if(layout->u.chunk.flags + & H5O_LAYOUT_CHUNK_DONT_FILTER_PARTIAL_BOUND_CHUNKS) { + /* Check if this is an edge chunk */ + if(H5D__chunk_is_partial_edge_chunk(io_info->dset->shared->ndims, + layout->u.chunk.dim, io_info->store->chunk.scaled, + io_info->dset->shared->curr_dims)) { + /* Disable the filters for both writing and reading */ + disable_filters = TRUE; + old_pline = NULL; + pline = NULL; + } /* end if */ + } /* end if */ + } /* end if */ + if(relax) { /* * Not in the cache, but we're about to overwrite the whole thing @@ -3066,16 +3374,32 @@ H5D__chunk_lock(const H5D_io_info_t *io_info, H5D_chunk_ud_t *udata, /* Chunk size on disk isn't [likely] the same size as the final chunk * size in memory, so allocate memory big enough. */ - if(NULL == (chunk = H5D__chunk_mem_alloc(my_chunk_alloc, pline))) + if(NULL == (chunk = H5D__chunk_mem_alloc(my_chunk_alloc, (udata->new_unfilt_chunk ? old_pline : pline)))) HGOTO_ERROR(H5E_RESOURCE, H5E_NOSPACE, NULL, "memory allocation failed for raw data chunk") - if(H5F_block_read(dset->oloc.file, H5FD_MEM_DRAW, chunk_addr, my_chunk_alloc, io_info->dxpl_id, chunk) < 0) + if(H5F_block_read(dset->oloc.file, H5FD_MEM_DRAW, chunk_addr, my_chunk_alloc, io_info->raw_dxpl_id, chunk) < 0) HGOTO_ERROR(H5E_IO, H5E_READERROR, NULL, "unable to read raw data chunk") - if(pline->nused) - if(H5Z_pipeline(pline, H5Z_FLAG_REVERSE, &(udata->filter_mask), io_info->dxpl_cache->err_detect, - io_info->dxpl_cache->filter_cb, &my_chunk_alloc, &buf_alloc, &chunk) < 0) + if(old_pline && old_pline->nused) { + if(H5Z_pipeline(old_pline, H5Z_FLAG_REVERSE, + &(udata->filter_mask), + io_info->dxpl_cache->err_detect, + io_info->dxpl_cache->filter_cb, + &my_chunk_alloc, &buf_alloc, &chunk) < 0) HGOTO_ERROR(H5E_PLINE, H5E_CANTFILTER, NULL, "data pipeline read failed") + /* Reallocate chunk if necessary */ + if(udata->new_unfilt_chunk) { + void *tmp_chunk = chunk; + + if(NULL == (chunk = H5D__chunk_mem_alloc(my_chunk_alloc, pline))) { + (void)H5D__chunk_mem_xfree(tmp_chunk, old_pline); + HGOTO_ERROR(H5E_RESOURCE, H5E_NOSPACE, NULL, "memory allocation failed for raw data chunk") + } /* end if */ + HDmemcpy(chunk, tmp_chunk, chunk_size); + (void)H5D__chunk_mem_xfree(tmp_chunk, old_pline); + } /* end if */ + } /* end if */ + /* Increment # of cache misses */ rdcc->stats.nmisses++; } /* end if */ @@ -3106,14 +3430,14 @@ H5D__chunk_lock(const H5D_io_info_t *io_info, H5D_chunk_ud_t *udata, /* (use the compact dataset storage buffer as the fill value buffer) */ if(H5D__fill_init(&fb_info, chunk, NULL, NULL, NULL, NULL, &dset->shared->dcpl_cache.fill, dset->shared->type, - dset->shared->type_id, (size_t)0, chunk_size, io_info->dxpl_id) < 0) + dset->shared->type_id, (size_t)0, chunk_size, io_info->md_dxpl_id) < 0) HGOTO_ERROR(H5E_DATASET, H5E_CANTINIT, NULL, "can't initialize fill buffer info") fb_info_init = TRUE; /* Check for VL datatype & non-default fill value */ if(fb_info.has_vlen_fill_type) /* Fill the buffer with VL datatype fill values */ - if(H5D__fill_refill_vl(&fb_info, fb_info.elmts_per_buf, io_info->dxpl_id) < 0) + if(H5D__fill_refill_vl(&fb_info, fb_info.elmts_per_buf, io_info->md_dxpl_id) < 0) HGOTO_ERROR(H5E_DATASET, H5E_CANTCONVERT, NULL, "can't refill fill value buffer") } /* end if */ else @@ -3134,19 +3458,24 @@ H5D__chunk_lock(const H5D_io_info_t *io_info, H5D_chunk_ud_t *udata, if(!ent || !ent->locked) { /* Preempt enough things from the cache to make room */ if(ent) { - if(H5D__chunk_cache_evict(io_info->dset, io_info->dxpl_id, io_info->dxpl_cache, ent, TRUE) < 0) + if(H5D__chunk_cache_evict(io_info->dset, io_info->md_dxpl_id, io_info->dxpl_cache, ent, TRUE) < 0) HGOTO_ERROR(H5E_IO, H5E_CANTINIT, NULL, "unable to preempt chunk from cache") } /* end if */ - if(H5D__chunk_cache_prune(io_info->dset, io_info->dxpl_id, io_info->dxpl_cache, chunk_size) < 0) + if(H5D__chunk_cache_prune(io_info->dset, io_info->md_dxpl_id, io_info->dxpl_cache, chunk_size) < 0) HGOTO_ERROR(H5E_IO, H5E_CANTINIT, NULL, "unable to preempt chunk(s) from cache") /* Create a new entry */ if(NULL == (ent = H5FL_CALLOC(H5D_rdcc_ent_t))) HGOTO_ERROR(H5E_DATASET, H5E_CANTALLOC, NULL, "can't allocate raw data chunk entry") + ent->edge_chunk_state = disable_filters ? H5D_RDCC_DISABLE_FILTERS : 0; + if(udata->new_unfilt_chunk) + ent->edge_chunk_state |= H5D_RDCC_NEWLY_DISABLED_FILTERS; + /* Initialize the new entry */ ent->chunk_block.offset = chunk_addr; ent->chunk_block.length = chunk_alloc; + ent->chunk_idx = udata->chunk_idx; HDmemcpy(ent->scaled, udata->common.scaled, sizeof(hsize_t) * layout->u.chunk.ndims); H5_CHECKED_ASSIGN(ent->rd_count, uint32_t, chunk_size, size_t); H5_CHECKED_ASSIGN(ent->wr_count, uint32_t, chunk_size, size_t); @@ -3167,6 +3496,9 @@ H5D__chunk_lock(const H5D_io_info_t *io_info, H5D_chunk_ud_t *udata, } /* end if */ else rdcc->head = rdcc->tail = ent; + ent->tmp_next = NULL; + ent->tmp_prev = NULL; + } /* end if */ else /* We did not add the chunk to cache */ @@ -3229,7 +3561,7 @@ done: * *------------------------------------------------------------------------- */ -herr_t +static herr_t H5D__chunk_unlock(const H5D_io_info_t *io_info, const H5D_chunk_ud_t *udata, hbool_t dirty, void *chunk, uint32_t naccessed) { @@ -3237,8 +3569,9 @@ H5D__chunk_unlock(const H5D_io_info_t *io_info, const H5D_chunk_ud_t *udata, const H5D_rdcc_t *rdcc = &(io_info->dset->shared->cache.chunk); herr_t ret_value = SUCCEED; /* Return value */ - FUNC_ENTER_PACKAGE + FUNC_ENTER_STATIC + /* Sanity check */ HDassert(io_info); HDassert(udata); @@ -3246,26 +3579,45 @@ H5D__chunk_unlock(const H5D_io_info_t *io_info, const H5D_chunk_ud_t *udata, /* * It's not in the cache, probably because it's too big. If it's * dirty then flush it to disk. In any case, free the chunk. - * Note: we have to copy the layout and filter messages so we - * don't discard the `const' qualifier. */ + hbool_t is_unfiltered_edge_chunk = FALSE; /* Whether the chunk is an unfiltered edge chunk */ + + /* Check if we should disable filters on this chunk */ + if(udata->new_unfilt_chunk) { + HDassert(layout->u.chunk.flags & H5O_LAYOUT_CHUNK_DONT_FILTER_PARTIAL_BOUND_CHUNKS); + + is_unfiltered_edge_chunk = TRUE; + } /* end if */ + else if(layout->u.chunk.flags & H5O_LAYOUT_CHUNK_DONT_FILTER_PARTIAL_BOUND_CHUNKS) { + /* Check if the chunk is an edge chunk, and disable filters if so */ + is_unfiltered_edge_chunk = H5D__chunk_is_partial_edge_chunk( + io_info->dset->shared->ndims, layout->u.chunk.dim, + io_info->store->chunk.scaled, io_info->dset->shared->curr_dims); + } /* end if */ + if(dirty) { H5D_rdcc_ent_t fake_ent; /* "fake" chunk cache entry */ HDmemset(&fake_ent, 0, sizeof(fake_ent)); fake_ent.dirty = TRUE; + if(is_unfiltered_edge_chunk) + fake_ent.edge_chunk_state = H5D_RDCC_DISABLE_FILTERS; + if(udata->new_unfilt_chunk) + fake_ent.edge_chunk_state |= H5D_RDCC_NEWLY_DISABLED_FILTERS; HDmemcpy(fake_ent.scaled, udata->common.scaled, sizeof(hsize_t) * layout->u.chunk.ndims); HDassert(layout->u.chunk.size > 0); + fake_ent.chunk_idx = udata->chunk_idx; fake_ent.chunk_block.offset = udata->chunk_block.offset; fake_ent.chunk_block.length = udata->chunk_block.length; fake_ent.chunk = (uint8_t *)chunk; - if(H5D__chunk_flush_entry(io_info->dset, io_info->dxpl_id, io_info->dxpl_cache, &fake_ent, TRUE) < 0) + if(H5D__chunk_flush_entry(io_info->dset, io_info->md_dxpl_id, io_info->dxpl_cache, &fake_ent, TRUE) < 0) HGOTO_ERROR(H5E_IO, H5E_WRITEERROR, FAIL, "cannot flush indexed storage buffer") } /* end if */ else { if(chunk) - chunk = H5D__chunk_mem_xfree(chunk, &(io_info->dset->shared->dcpl_cache.pline)); + chunk = H5D__chunk_mem_xfree(chunk, (is_unfiltered_edge_chunk ? NULL + : &(io_info->dset->shared->dcpl_cache.pline))); } /* end else */ } /* end if */ else { @@ -3344,12 +3696,14 @@ H5D__chunk_allocated(H5D_t *dset, hid_t dxpl_id, hsize_t *nbytes) H5D_dxpl_cache_t _dxpl_cache; /* Data transfer property cache buffer */ H5D_dxpl_cache_t *dxpl_cache = &_dxpl_cache; /* Data transfer property cache */ hsize_t chunk_bytes = 0; /* Number of bytes allocated for chunks */ + H5O_storage_chunk_t *sc = &(dset->shared->layout.storage.u.chunk); herr_t ret_value = SUCCEED; /* Return value */ FUNC_ENTER_PACKAGE HDassert(dset); HDassert(dset->shared); + H5D_CHUNK_STORAGE_INDEX_CHK(sc); /* Fill the DXPL cache values for later use */ if(H5D__get_dxpl_cache(dxpl_id, &dxpl_cache) < 0) @@ -3396,9 +3750,9 @@ done: *------------------------------------------------------------------------- */ herr_t -H5D__chunk_allocate(const H5D_t *dset, hid_t dxpl_id, hbool_t full_overwrite, - hsize_t old_dim[]) +H5D__chunk_allocate(const H5D_io_info_t *io_info, hbool_t full_overwrite, hsize_t old_dim[]) { + const H5D_t *dset = io_info->dset; /* the dataset pointer */ H5D_chk_idx_info_t idx_info; /* Chunked index info */ const H5D_chunk_ops_t *ops = dset->shared->layout.storage.u.chunk.ops; /* Chunk operations */ hsize_t min_unalloc[H5O_LAYOUT_NDIMS]; /* First chunk in each dimension that is unallocated (in scaled coordinates) */ @@ -3409,9 +3763,12 @@ H5D__chunk_allocate(const H5D_t *dset, hid_t dxpl_id, hbool_t full_overwrite, unsigned filter_mask = 0; /* Filter mask for chunks that have them */ const H5O_layout_t *layout = &(dset->shared->layout); /* Dataset layout */ const H5O_pline_t *pline = &(dset->shared->dcpl_cache.pline); /* I/O pipeline info */ + const H5O_pline_t def_pline = H5O_CRT_PIPELINE_DEF; /* Default pipeline */ const H5O_fill_t *fill = &(dset->shared->dcpl_cache.fill); /* Fill value info */ H5D_fill_value_t fill_status; /* The fill value status */ hbool_t should_fill = FALSE; /* Whether fill values should be written */ + void *unfilt_fill_buf = NULL; /* Unfiltered fill value buffer */ + void **fill_buf = NULL; /* Pointer to the fill buffer to use for a chunk */ H5D_dxpl_cache_t _dxpl_cache; /* Data transfer property cache buffer */ H5D_dxpl_cache_t *dxpl_cache = &_dxpl_cache; /* Data transfer property cache */ #ifdef H5_HAVE_PARALLEL @@ -3419,6 +3776,8 @@ H5D__chunk_allocate(const H5D_t *dset, hid_t dxpl_id, hbool_t full_overwrite, hbool_t using_mpi = FALSE; /* Flag to indicate that the file is being accessed with an MPI-capable file driver */ H5D_chunk_coll_info_t chunk_info; /* chunk address information for doing I/O */ #endif /* H5_HAVE_PARALLEL */ + hid_t md_dxpl_id = io_info->md_dxpl_id; + hid_t raw_dxpl_id = io_info->raw_dxpl_id; hbool_t carry; /* Flag to indicate that chunk increment carrys to higher dimension (sorta) */ unsigned space_ndims; /* Dataset's space rank */ const hsize_t *space_dim; /* Dataset's dataspace dimensions */ @@ -3426,14 +3785,21 @@ H5D__chunk_allocate(const H5D_t *dset, hid_t dxpl_id, hbool_t full_overwrite, unsigned op_dim; /* Current operating dimension */ H5D_fill_buf_info_t fb_info; /* Dataset's fill buffer info */ hbool_t fb_info_init = FALSE; /* Whether the fill value buffer has been initialized */ + hbool_t has_unfilt_edge_chunks = FALSE; /* Whether there are partial edge chunks with disabled filters */ + hbool_t unfilt_edge_chunk_dim[H5O_LAYOUT_NDIMS]; /* Whether there are unfiltered edge chunks at the edge of each dimension */ + hsize_t edge_chunk_scaled[H5O_LAYOUT_NDIMS]; /* Offset of the unfiltered edge chunks at the edge of each dimension */ + unsigned nunfilt_edge_chunk_dims = 0; /* Number of dimensions on an edge */ + const H5O_storage_chunk_t *sc = &(layout->storage.u.chunk); herr_t ret_value = SUCCEED; /* Return value */ - FUNC_ENTER_PACKAGE_TAG(dxpl_id, dset->oloc.addr, FAIL) + FUNC_ENTER_PACKAGE_TAG(md_dxpl_id, dset->oloc.addr, FAIL) /* Check args */ HDassert(dset && H5D_CHUNKED == layout->type); HDassert(layout->u.chunk.ndims > 0 && layout->u.chunk.ndims <= H5O_LAYOUT_NDIMS); - HDassert(TRUE == H5P_isa_class(dxpl_id, H5P_DATASET_XFER)); + H5D_CHUNK_STORAGE_INDEX_CHK(sc); + HDassert(TRUE == H5P_isa_class(md_dxpl_id, H5P_DATASET_XFER)); + HDassert(TRUE == H5P_isa_class(raw_dxpl_id, H5P_DATASET_XFER)); /* Retrieve the dataset dimensions */ space_dim = dset->shared->curr_dims; @@ -3464,9 +3830,32 @@ H5D__chunk_allocate(const H5D_t *dset, hid_t dxpl_id, hbool_t full_overwrite, #endif /* H5_HAVE_PARALLEL */ /* Fill the DXPL cache values for later use */ - if(H5D__get_dxpl_cache(dxpl_id, &dxpl_cache) < 0) + if(H5D__get_dxpl_cache(raw_dxpl_id, &dxpl_cache) < 0) HGOTO_ERROR(H5E_DATASET, H5E_CANTGET, FAIL, "can't fill dxpl cache") + /* Calculate the minimum and maximum chunk offsets in each dimension, and + * determine if there are any unfiltered partial edge chunks. Note that we + * assume here that all elements of space_dim are > 0. This is checked at + * the top of this function. */ + for(op_dim=0; op_dim<space_ndims; op_dim++) { + min_unalloc[op_dim] = (old_dim[op_dim] + chunk_dim[op_dim] - 1) / chunk_dim[op_dim]; + max_unalloc[op_dim] = (space_dim[op_dim] - 1) / chunk_dim[op_dim]; + + /* Calculate if there are unfiltered edge chunks at the edge of this + * dimension. Note the edge_chunk_scaled is uninitialized for + * dimensions where unfilt_edge_chunk_dim is FALSE. Also */ + if((layout->u.chunk.flags + & H5O_LAYOUT_CHUNK_DONT_FILTER_PARTIAL_BOUND_CHUNKS) + && pline->nused > 0 + && space_dim[op_dim] % chunk_dim[op_dim] != 0) { + has_unfilt_edge_chunks = TRUE; + unfilt_edge_chunk_dim[op_dim] = TRUE; + edge_chunk_scaled[op_dim] = max_unalloc[op_dim]; + } /* end if */ + else + unfilt_edge_chunk_dim[op_dim] = FALSE; + } /* end for */ + /* Get original chunk size */ H5_CHECKED_ASSIGN(orig_chunk_size, size_t, layout->u.chunk.size, uint32_t); @@ -3494,10 +3883,15 @@ H5D__chunk_allocate(const H5D_t *dset, hid_t dxpl_id, hbool_t full_overwrite, if(H5D__fill_init(&fb_info, NULL, (H5MM_allocate_t)H5D__chunk_mem_alloc, (void *)pline, (H5MM_free_t)H5D__chunk_mem_xfree, (void *)pline, &dset->shared->dcpl_cache.fill, dset->shared->type, - dset->shared->type_id, (size_t)0, orig_chunk_size, dxpl_id) < 0) + dset->shared->type_id, (size_t)0, orig_chunk_size, md_dxpl_id) < 0) HGOTO_ERROR(H5E_DATASET, H5E_CANTINIT, FAIL, "can't initialize fill buffer info") fb_info_init = TRUE; + /* Initialize the fill_buf pointer to the buffer in fb_info. If edge + * chunk filters are disabled, we will switch the buffer as appropriate + * for each chunk. */ + fill_buf = &fb_info.fill_buf; + /* Check if there are filters which need to be applied to the chunk */ /* (only do this in advance when the chunk info can be re-used (i.e. * it doesn't contain any non-default VL datatype fill values) @@ -3505,6 +3899,14 @@ H5D__chunk_allocate(const H5D_t *dset, hid_t dxpl_id, hbool_t full_overwrite, if(!fb_info.has_vlen_fill_type && pline->nused > 0) { size_t buf_size = orig_chunk_size; + /* If the dataset has disabled partial chunk filters, create a copy + * of the unfiltered fill_buf to use for partial chunks */ + if(has_unfilt_edge_chunks) { + if(NULL == (unfilt_fill_buf = H5D__chunk_mem_alloc(orig_chunk_size, &def_pline))) + HGOTO_ERROR(H5E_RESOURCE, H5E_NOSPACE, FAIL, "memory allocation failed for raw data chunk") + HDmemcpy(unfilt_fill_buf, fb_info.fill_buf, orig_chunk_size); + } /* end if */ + /* Push the chunk through the filters */ if(H5Z_pipeline(pline, 0, &filter_mask, dxpl_cache->err_detect, dxpl_cache->filter_cb, &orig_chunk_size, &buf_size, &fb_info.fill_buf) < 0) HGOTO_ERROR(H5E_PLINE, H5E_WRITEERROR, FAIL, "output pipeline failed") @@ -3518,19 +3920,11 @@ H5D__chunk_allocate(const H5D_t *dset, hid_t dxpl_id, hbool_t full_overwrite, /* Compose chunked index info struct */ idx_info.f = dset->oloc.file; - idx_info.dxpl_id = dxpl_id; + idx_info.dxpl_id = md_dxpl_id; idx_info.pline = &dset->shared->dcpl_cache.pline; idx_info.layout = &dset->shared->layout.u.chunk; idx_info.storage = &dset->shared->layout.storage.u.chunk; - /* Calculate the minimum and maximum chunk offsets in each dimension. Note - * that we assume here that all elements of space_dim are > 0. This is - * checked at the top of this function. */ - for(op_dim = 0; op_dim < (unsigned)space_ndims; op_dim++) { - min_unalloc[op_dim] = (old_dim[op_dim] + chunk_dim[op_dim] - 1) / chunk_dim[op_dim]; - max_unalloc[op_dim] = (space_dim[op_dim] - 1) / chunk_dim[op_dim]; - } /* end for */ - /* Loop over all chunks */ /* The algorithm is: * For each dimension: @@ -3552,8 +3946,10 @@ H5D__chunk_allocate(const H5D_t *dset, hid_t dxpl_id, hbool_t full_overwrite, * Note that min_unalloc & max_unalloc are in scaled coordinates. * */ + chunk_size = orig_chunk_size; for(op_dim = 0; op_dim < space_ndims; op_dim++) { H5D_chunk_ud_t udata; /* User data for querying chunk info */ + unsigned u; /* Local index variable */ int i; /* Local index variable */ /* Check if allocation along this dimension is really necessary */ @@ -3564,34 +3960,50 @@ H5D__chunk_allocate(const H5D_t *dset, hid_t dxpl_id, hbool_t full_overwrite, HDmemset(scaled, 0, (space_ndims * sizeof(scaled[0]))); scaled[op_dim] = min_unalloc[op_dim]; + if(has_unfilt_edge_chunks) { + /* Initialize nunfilt_edge_chunk_dims */ + nunfilt_edge_chunk_dims = 0; + for(u = 0; u < space_ndims; u++) + if(unfilt_edge_chunk_dim[u] && scaled[u] + == edge_chunk_scaled[u]) + nunfilt_edge_chunk_dims++; + + /* Initialize chunk_size and fill_buf */ + if(should_fill && !fb_info.has_vlen_fill_type) { + HDassert(fb_info_init); + HDassert(unfilt_fill_buf); + if(nunfilt_edge_chunk_dims) { + fill_buf = &unfilt_fill_buf; + chunk_size = layout->u.chunk.size; + } /* end if */ + else { + fill_buf = &fb_info.fill_buf; + chunk_size = orig_chunk_size; + } /* end else */ + } /* end if */ + } /* end if */ + carry = FALSE; } /* end else */ while(!carry) { hbool_t need_insert = FALSE; /* Whether the chunk needs to be inserted into the index */ - /* Reset size of chunk in bytes, in case filtered size changes */ - chunk_size = orig_chunk_size; - + /* Look up this chunk */ + if(H5D__chunk_lookup(dset, md_dxpl_id, scaled, &udata) < 0) + HGOTO_ERROR(H5E_DATASET, H5E_CANTGET, FAIL, "error looking up chunk address") #ifndef NDEBUG /* None of the chunks should be allocated */ - { - /* Look up this chunk */ - if(H5D__chunk_lookup(dset, dxpl_id, scaled, &udata) < 0) - HGOTO_ERROR(H5E_DATASET, H5E_CANTGET, FAIL, "error looking up chunk address") - - HDassert(!H5F_addr_defined(udata.chunk_block.offset)); - } /* end block */ /* Make sure the chunk is really in the dataset and outside the * original dimensions */ { - unsigned u; /* Local index variable */ + unsigned v; /* Local index variable */ hbool_t outside_orig = FALSE; - for(u = 0; u < space_ndims; u++) { - HDassert((scaled[u] * chunk_dim[u]) < space_dim[u]); - if((scaled[u] * chunk_dim[u]) >= old_dim[u]) + for(v = 0; v < space_ndims; v++) { + HDassert((scaled[v] * chunk_dim[v]) < space_dim[v]); + if((scaled[v] * chunk_dim[v]) >= old_dim[v]) outside_orig = TRUE; } /* end for */ HDassert(outside_orig); @@ -3602,6 +4014,7 @@ H5D__chunk_allocate(const H5D_t *dset, hid_t dxpl_id, hbool_t full_overwrite, if(fb_info_init && fb_info.has_vlen_fill_type) { /* Sanity check */ HDassert(should_fill); + HDassert(!unfilt_fill_buf); #ifdef H5_HAVE_PARALLEL HDassert(!using_mpi); /* Can't write VL datatypes in parallel currently */ #endif @@ -3616,11 +4029,11 @@ H5D__chunk_allocate(const H5D_t *dset, hid_t dxpl_id, hbool_t full_overwrite, } /* end if */ /* Fill the buffer with VL datatype fill values */ - if(H5D__fill_refill_vl(&fb_info, fb_info.elmts_per_buf, dxpl_id) < 0) + if(H5D__fill_refill_vl(&fb_info, fb_info.elmts_per_buf, md_dxpl_id) < 0) HGOTO_ERROR(H5E_DATASET, H5E_CANTCONVERT, FAIL, "can't refill fill value buffer") /* Check if there are filters which need to be applied to the chunk */ - if(pline->nused > 0) { + if((pline->nused > 0) && !nunfilt_edge_chunk_dims) { size_t nbytes = orig_chunk_size; /* Push the chunk through the filters */ @@ -3636,6 +4049,10 @@ H5D__chunk_allocate(const H5D_t *dset, hid_t dxpl_id, hbool_t full_overwrite, /* Keep the number of bytes the chunk turned in to */ chunk_size = nbytes; } /* end if */ + else + chunk_size = layout->u.chunk.size; + + HDassert(*fill_buf == fb_info.fill_buf); } /* end if */ /* Initialize the chunk information */ @@ -3663,8 +4080,11 @@ H5D__chunk_allocate(const H5D_t *dset, hid_t dxpl_id, hbool_t full_overwrite, /* collect all chunk addresses to be written to write collectively at the end */ /* allocate/resize address array if no more space left */ + /* Note that if we add support for parallel filters we must + * also store an array of chunk sizes and pass it to the + * apporpriate collective write function */ if(0 == chunk_info.num_io % 1024) - if(NULL == (chunk_info.addr = (haddr_t *)HDrealloc(chunk_info.addr, (chunk_info.num_io + 1024) * sizeof(haddr_t)))) + if(NULL == (chunk_info.addr = (haddr_t *)H5MM_realloc(chunk_info.addr, (chunk_info.num_io + 1024) * sizeof(haddr_t)))) HGOTO_ERROR(H5E_DATASET, H5E_CANTALLOC, FAIL, "memory allocation failed for chunk addresses") /* Store the chunk's address for later */ @@ -3676,7 +4096,7 @@ H5D__chunk_allocate(const H5D_t *dset, hid_t dxpl_id, hbool_t full_overwrite, } /* end if */ else { #endif /* H5_HAVE_PARALLEL */ - if(H5F_block_write(dset->oloc.file, H5FD_MEM_DRAW, udata.chunk_block.offset, chunk_size, dxpl_id, fb_info.fill_buf) < 0) + if(H5F_block_write(dset->oloc.file, H5FD_MEM_DRAW, udata.chunk_block.offset, chunk_size, raw_dxpl_id, *fill_buf) < 0) HGOTO_ERROR(H5E_IO, H5E_WRITEERROR, FAIL, "unable to write raw data to file") #ifdef H5_HAVE_PARALLEL } /* end else */ @@ -3684,9 +4104,6 @@ H5D__chunk_allocate(const H5D_t *dset, hid_t dxpl_id, hbool_t full_overwrite, } /* end if */ /* Insert the chunk record into the index */ - /* (Note that this isn't safe, from a SWMR perspective, unlike - * serial operation. -QAK - */ if(need_insert && ops->insert) if((ops->insert)(&idx_info, &udata, dset) < 0) HGOTO_ERROR(H5E_DATASET, H5E_CANTINSERT, FAIL, "unable to insert chunk addr into index") @@ -3700,8 +4117,31 @@ H5D__chunk_allocate(const H5D_t *dset, hid_t dxpl_id, hbool_t full_overwrite, scaled[i] = min_unalloc[i]; else scaled[i] = 0; - } /* end if */ + + /* Check if we just left the edge in this dimension */ + if(unfilt_edge_chunk_dim[i] + && edge_chunk_scaled[i] == max_unalloc[i] + && scaled[i] < edge_chunk_scaled[i]) { + nunfilt_edge_chunk_dims--; + if(should_fill && nunfilt_edge_chunk_dims == 0 && !fb_info.has_vlen_fill_type) { + HDassert(!H5D__chunk_is_partial_edge_chunk(space_ndims, chunk_dim, scaled, space_dim)); + fill_buf = &fb_info.fill_buf; + chunk_size = orig_chunk_size; + } /* end if */ + } /* end if */ + } /* end if */ else { + /* Check if we just entered the edge in this dimension */ + if(unfilt_edge_chunk_dim[i] && scaled[i] == edge_chunk_scaled[i]) { + HDassert(edge_chunk_scaled[i] == max_unalloc[i]); + nunfilt_edge_chunk_dims++; + if(should_fill && nunfilt_edge_chunk_dims == 1 && !fb_info.has_vlen_fill_type) { + HDassert(H5D__chunk_is_partial_edge_chunk(space_ndims, chunk_dim, scaled, space_dim)); + fill_buf = &unfilt_fill_buf; + chunk_size = layout->u.chunk.size; + } /* end if */ + } /* end if */ + carry = FALSE; break; } /* end else */ @@ -3720,7 +4160,7 @@ H5D__chunk_allocate(const H5D_t *dset, hid_t dxpl_id, hbool_t full_overwrite, #ifdef H5_HAVE_PARALLEL /* do final collective I/O */ if(using_mpi && blocks_written) - if(H5D__chunk_collective_fill(dset, dxpl_id, &chunk_info, chunk_size, fb_info.fill_buf) < 0) + if(H5D__chunk_collective_fill(dset, raw_dxpl_id, &chunk_info, chunk_size, fb_info.fill_buf) < 0) HGOTO_ERROR(H5E_IO, H5E_WRITEERROR, FAIL, "unable to write raw data to file") #endif /* H5_HAVE_PARALLEL */ @@ -3732,14 +4172,196 @@ done: if(fb_info_init && H5D__fill_term(&fb_info) < 0) HDONE_ERROR(H5E_DATASET, H5E_CANTFREE, FAIL, "Can't release fill buffer info") + /* Free the unfiltered fill value buffer */ + unfilt_fill_buf = H5D__chunk_mem_xfree(unfilt_fill_buf, &def_pline); + #ifdef H5_HAVE_PARALLEL if(using_mpi && chunk_info.addr) - HDfree(chunk_info.addr); + H5MM_free(chunk_info.addr); #endif FUNC_LEAVE_NOAPI_TAG(ret_value, FAIL) } /* end H5D__chunk_allocate() */ + +/*------------------------------------------------------------------------- + * Function: H5D__chunk_update_old_edge_chunks + * + * Purpose: Update all chunks which were previously partial edge + * chunks and are now complete. Determines exactly which + * chunks need to be updated and locks each into cache using + * the 'prev_unfilt_chunk' flag, then unlocks it, causing + * filters to be applied as necessary. + * + * Return: Non-negative on success/Negative on failure + * + * Programmer: Neil Fortner + * April 14, 2010 + * + *------------------------------------------------------------------------- + */ +herr_t +H5D__chunk_update_old_edge_chunks(H5D_t *dset, hid_t dxpl_id, hsize_t old_dim[]) +{ + hsize_t old_edge_chunk_sc[H5O_LAYOUT_NDIMS]; /* Offset of first previously incomplete chunk in each dimension */ + hsize_t max_edge_chunk_sc[H5O_LAYOUT_NDIMS]; /* largest offset of chunks that might need to be modified in each dimension */ + hbool_t new_full_dim[H5O_LAYOUT_NDIMS]; /* Whether the plane of chunks in this dimension needs to be modified */ + const H5O_layout_t *layout = &(dset->shared->layout); /* Dataset layout */ + const H5O_pline_t *pline = &(dset->shared->dcpl_cache.pline); /* I/O pipeline info */ + hsize_t chunk_sc[H5O_LAYOUT_NDIMS]; /* Offset of current chunk */ + const uint32_t *chunk_dim = layout->u.chunk.dim; /* Convenience pointer to chunk dimensions */ + unsigned space_ndims; /* Dataset's space rank */ + const hsize_t *space_dim; /* Dataset's dataspace dimensions */ + unsigned op_dim; /* Current operationg dimension */ + H5D_io_info_t chk_io_info; /* Chunked I/O info object */ + H5D_chunk_ud_t chk_udata; /* User data for locking chunk */ + H5D_storage_t chk_store; /* Chunk storage information */ + H5D_dxpl_cache_t _dxpl_cache; /* Data transfer property cache buffer */ + H5D_dxpl_cache_t *dxpl_cache = &_dxpl_cache; /* Data transfer property cache */ + void *chunk; /* The file chunk */ + hbool_t carry; /* Flag to indicate that chunk increment carrys to higher dimension (sorta) */ + const H5O_storage_chunk_t *sc = &(layout->storage.u.chunk); + herr_t ret_value = SUCCEED; /* Return value */ + + FUNC_ENTER_PACKAGE + + /* Check args */ + HDassert(dset && H5D_CHUNKED == layout->type); + HDassert(layout->u.chunk.ndims > 0 && layout->u.chunk.ndims <= H5O_LAYOUT_NDIMS); + H5D_CHUNK_STORAGE_INDEX_CHK(sc); + HDassert(TRUE == H5P_isa_class(dxpl_id, H5P_DATASET_XFER)); + HDassert(pline->nused > 0); + HDassert(layout->u.chunk.flags + & H5O_LAYOUT_CHUNK_DONT_FILTER_PARTIAL_BOUND_CHUNKS); + + /* Retrieve the dataset dimensions */ + space_dim = dset->shared->curr_dims; + space_ndims = dset->shared->ndims; + + /* The last dimension in chunk_offset is always 0 */ + chunk_sc[space_ndims] = (hsize_t)0; + + /* Check if any current dimensions are smaller than the chunk size, or if + * any old dimensions are 0. If so we do not have to do anything. */ + for(op_dim=0; op_dim<space_ndims; op_dim++) + if((space_dim[op_dim] < chunk_dim[op_dim]) || old_dim[op_dim] == 0) { + /* Reset any cached chunk info for this dataset */ + H5D__chunk_cinfo_cache_reset(&dset->shared->cache.chunk.last); + HGOTO_DONE(SUCCEED) + } /* end if */ + + /* + * Initialize structures needed to lock chunks into cache + */ + /* Fill the DXPL cache values for later use */ + if(H5D__get_dxpl_cache(dxpl_id, &dxpl_cache) < 0) + HGOTO_ERROR(H5E_DATASET, H5E_CANTGET, FAIL, "can't fill dxpl cache") + + /* Set up chunked I/O info object, for operations on chunks (in callback). + * Note that we only need to set chunk_offset once, as the array's address + * will never change. */ + chk_store.chunk.scaled = chunk_sc; + H5D_BUILD_IO_INFO_RD(&chk_io_info, dset, dxpl_cache, dxpl_id, H5AC_rawdata_dxpl_id, &chk_store, NULL); + + /* + * Determine the edges of the dataset which need to be modified + */ + for(op_dim=0; op_dim<space_ndims; op_dim++) { + /* Start off with this dimension marked as not needing to be modified */ + new_full_dim[op_dim] = FALSE; + + /* Calulate offset of first previously incomplete chunk in this + * dimension */ + old_edge_chunk_sc[op_dim] = (old_dim[op_dim] / chunk_dim[op_dim]); + + /* Calculate the largest offset of chunks that might need to be + * modified in this dimension */ + max_edge_chunk_sc[op_dim] = MIN((old_dim[op_dim] - 1) / chunk_dim[op_dim], + MAX((space_dim[op_dim] / chunk_dim[op_dim]), 1) - 1); + + /* Check for old_dim aligned with chunk boundary in this dimension, if + * so we do not need to modify chunks along the edge in this dimension + */ + if(old_dim[op_dim] % chunk_dim[op_dim] == 0) + continue; + + /* Check if the dataspace expanded enough to cause the old edge chunks + * in this dimension to become full */ + if((space_dim[op_dim]/chunk_dim[op_dim]) >= (old_edge_chunk_sc[op_dim] + 1)) + new_full_dim[op_dim] = TRUE; + } /* end for */ + + /* Main loop: fix old edge chunks */ + for(op_dim=0; op_dim<space_ndims; op_dim++) { + /* Check if allocation along this dimension is really necessary */ + if(!new_full_dim[op_dim]) + continue; + else { + HDassert(max_edge_chunk_sc[op_dim] == old_edge_chunk_sc[op_dim]); + + /* Reset the chunk offset indices */ + HDmemset(chunk_sc, 0, (space_ndims * sizeof(chunk_sc[0]))); + chunk_sc[op_dim] = old_edge_chunk_sc[op_dim]; + + carry = FALSE; + } /* end if */ + + while(!carry) { + int i; /* Local index variable */ + + /* Make sure the chunk is really a former edge chunk */ + HDassert(H5D__chunk_is_partial_edge_chunk(space_ndims, chunk_dim, chunk_sc, old_dim) + && !H5D__chunk_is_partial_edge_chunk(space_ndims, chunk_dim, chunk_sc, space_dim)); + + /* Lookup the chunk */ + if(H5D__chunk_lookup(dset, dxpl_id, chunk_sc, &chk_udata) < 0) + HGOTO_ERROR(H5E_DATASET, H5E_CANTGET, FAIL, "error looking up chunk address") + + /* If this chunk does not exist in cache or on disk, no need to do + * anything */ + if(H5F_addr_defined(chk_udata.chunk_block.offset) + || (UINT_MAX != chk_udata.idx_hint)) { + /* Lock the chunk into cache. H5D__chunk_lock will take care of + * updating the chunk to no longer be an edge chunk. */ + if(NULL == (chunk = (void *)H5D__chunk_lock(&chk_io_info, &chk_udata, FALSE, TRUE))) + HGOTO_ERROR(H5E_DATASET, H5E_READERROR, FAIL, "unable to lock raw data chunk") + + /* Unlock the chunk */ + if(H5D__chunk_unlock(&chk_io_info, &chk_udata, TRUE, chunk, (uint32_t)0) < 0) + HGOTO_ERROR(H5E_IO, H5E_WRITEERROR, FAIL, "unable to unlock raw data chunk") + } /* end if */ + + /* Increment indices */ + carry = TRUE; + for(i = ((int)space_ndims - 1); i >= 0; --i) { + if((unsigned)i != op_dim) { + ++chunk_sc[i]; + if(chunk_sc[i] > (hsize_t) max_edge_chunk_sc[i]) + chunk_sc[i] = 0; + else { + carry = FALSE; + break; + } /* end else */ + } /* end if */ + } /* end for */ + } /* end while(!carry) */ + + /* Adjust max_edge_chunk_sc so we don't modify the same chunk twice. + * Also check if this dimension started from 0 (and hence modified all + * of the old edge chunks. */ + if(old_edge_chunk_sc[op_dim] == 0) + break; + else + --max_edge_chunk_sc[op_dim]; + } /* end for(op_dim=0...) */ + + /* Reset any cached chunk info for this dataset */ + H5D__chunk_cinfo_cache_reset(&dset->shared->cache.chunk.last); + +done: + FUNC_LEAVE_NOAPI(ret_value) +} /* end H5D__chunk_update_old_edge_chunks() */ + #ifdef H5_HAVE_PARALLEL /*------------------------------------------------------------------------- @@ -3910,7 +4532,7 @@ done: *------------------------------------------------------------------------- */ static herr_t -H5D__chunk_prune_fill(H5D_chunk_it_ud1_t *udata) +H5D__chunk_prune_fill(H5D_chunk_it_ud1_t *udata, hbool_t new_unfilt_chunk) { const H5D_io_info_t *io_info = udata->io_info; /* Local pointer to I/O info */ const H5D_t *dset = io_info->dset; /* Local pointer to the dataset info */ @@ -3935,8 +4557,9 @@ H5D__chunk_prune_fill(H5D_chunk_it_ud1_t *udata) H5_CHECKED_ASSIGN(chunk_size, size_t, layout->u.chunk.size, uint32_t); /* Get the info for the chunk in the file */ - if(H5D__chunk_lookup(dset, io_info->dxpl_id, scaled, &chk_udata) < 0) + if(H5D__chunk_lookup(dset, io_info->md_dxpl_id, scaled, &chk_udata) < 0) HGOTO_ERROR(H5E_DATASET, H5E_CANTGET, FAIL, "error looking up chunk address") + chk_udata.new_unfilt_chunk = new_unfilt_chunk; /* If this chunk does not exist in cache or on disk, no need to do anything */ if(!H5F_addr_defined(chk_udata.chunk_block.offset) && UINT_MAX == chk_udata.idx_hint) @@ -3948,7 +4571,7 @@ H5D__chunk_prune_fill(H5D_chunk_it_ud1_t *udata) if(H5D__fill_init(&udata->fb_info, NULL, NULL, NULL, NULL, NULL, &dset->shared->dcpl_cache.fill, dset->shared->type, dset->shared->type_id, (size_t)udata->elmts_per_chunk, - chunk_size, io_info->dxpl_id) < 0) + chunk_size, io_info->md_dxpl_id) < 0) HGOTO_ERROR(H5E_DATASET, H5E_CANTINIT, FAIL, "can't initialize fill buffer info") udata->fb_info_init = TRUE; } /* end if */ @@ -3968,7 +4591,7 @@ H5D__chunk_prune_fill(H5D_chunk_it_ud1_t *udata) HGOTO_ERROR(H5E_DATASET, H5E_CANTSELECT, FAIL, "unable to select hyperslab") /* Lock the chunk into the cache, to get a pointer to the chunk buffer */ - if(NULL == (chunk = (void *)H5D__chunk_lock(io_info, &chk_udata, FALSE))) + if(NULL == (chunk = (void *)H5D__chunk_lock(io_info, &chk_udata, FALSE, FALSE))) HGOTO_ERROR(H5E_DATASET, H5E_READERROR, FAIL, "unable to lock raw data chunk") @@ -3985,7 +4608,7 @@ H5D__chunk_prune_fill(H5D_chunk_it_ud1_t *udata) /* Check for VL datatype & non-default fill value */ if(udata->fb_info.has_vlen_fill_type) /* Re-fill the buffer to use for this I/O operation */ - if(H5D__fill_refill_vl(&udata->fb_info, (size_t)sel_nelmts, io_info->dxpl_id) < 0) + if(H5D__fill_refill_vl(&udata->fb_info, (size_t)sel_nelmts, io_info->md_dxpl_id) < 0) HGOTO_ERROR(H5E_DATASET, H5E_CANTCONVERT, FAIL, "can't refill fill value buffer") /* Create a selection iterator for scattering the elements to memory buffer */ @@ -4122,6 +4745,8 @@ H5D__chunk_prune_by_extent(H5D_t *dset, hid_t dxpl_id, const hsize_t *old_dim) hsize_t max_mod_chunk_sc[H5O_LAYOUT_NDIMS]; /* Scaled offset of last chunk to modify in each dimension */ hssize_t max_fill_chunk_sc[H5O_LAYOUT_NDIMS]; /* Scaled offset of last chunk that might be filled in each dimension */ hbool_t fill_dim[H5O_LAYOUT_NDIMS]; /* Whether the plane of edge chunks in this dimension needs to be filled */ + hsize_t min_partial_chunk_sc[H5O_LAYOUT_NDIMS]; /* Offset of first partial (or empty) chunk in each dimension */ + hbool_t new_unfilt_dim[H5O_LAYOUT_NDIMS]; /* Whether the plane of edge chunks in this dimension are newly unfiltered */ H5D_chk_idx_info_t idx_info; /* Chunked index info */ H5D_io_info_t chk_io_info; /* Chunked I/O info object */ H5D_storage_t chk_store; /* Chunk storage information */ @@ -4141,7 +4766,10 @@ H5D__chunk_prune_by_extent(H5D_t *dset, hid_t dxpl_id, const hsize_t *old_dim) hsize_t scaled[H5O_LAYOUT_NDIMS]; /* Scaled offset of current chunk */ hsize_t hyper_start[H5O_LAYOUT_NDIMS]; /* Starting location of hyperslab */ uint32_t elmts_per_chunk; /* Elements in chunk */ + hbool_t disable_edge_filters = FALSE; /* Whether to disable filters on partial edge chunks */ + hbool_t new_unfilt_chunk = FALSE; /* Whether the chunk is newly unfiltered */ unsigned u; /* Local index variable */ + const H5O_storage_chunk_t *sc = &(layout->storage.u.chunk); herr_t ret_value = SUCCEED; /* Return value */ FUNC_ENTER_PACKAGE @@ -4149,6 +4777,7 @@ H5D__chunk_prune_by_extent(H5D_t *dset, hid_t dxpl_id, const hsize_t *old_dim) /* Check args */ HDassert(dset && H5D_CHUNKED == layout->type); HDassert(layout->u.chunk.ndims > 0 && layout->u.chunk.ndims <= H5O_LAYOUT_NDIMS); + H5D_CHUNK_STORAGE_INDEX_CHK(sc); HDassert(dxpl_cache); /* Fill the DXPL cache values for later use */ @@ -4194,7 +4823,8 @@ H5D__chunk_prune_by_extent(H5D_t *dset, hid_t dxpl_id, const hsize_t *old_dim) * Note that we only need to set scaled once, as the array's address * will never change. */ chk_store.chunk.scaled = scaled; - H5D_BUILD_IO_INFO_RD(&chk_io_info, dset, dxpl_cache, dxpl_id, &chk_store, NULL); + H5D_BUILD_IO_INFO_RD(&chk_io_info, dset, dxpl_cache, dxpl_id, H5AC_rawdata_dxpl_id, &chk_store, NULL); + chk_io_info.raw_dxpl_id = H5AC_rawdata_dxpl_id; /* Compose chunked index info struct */ idx_info.f = dset->oloc.file; @@ -4221,6 +4851,11 @@ H5D__chunk_prune_by_extent(H5D_t *dset, hid_t dxpl_id, const hsize_t *old_dim) idx_udata.layout = &layout->u.chunk; idx_udata.storage = &layout->storage.u.chunk; + /* Determine if partial edge chunk filters are disabled */ + disable_edge_filters = (layout->u.chunk.flags + & H5O_LAYOUT_CHUNK_DONT_FILTER_PARTIAL_BOUND_CHUNKS) + && (idx_info.pline->nused > 0); + /* * Determine the chunks which need to be filled or removed */ @@ -4247,13 +4882,31 @@ H5D__chunk_prune_by_extent(H5D_t *dset, hid_t dxpl_id, const hsize_t *old_dim) min_mod_chunk_sc[op_dim] = space_dim[op_dim] / chunk_dim[op_dim]; /* Determine if we need to fill chunks in this dimension */ - if((hssize_t)min_mod_chunk_sc[op_dim] == max_fill_chunk_sc[op_dim]) + if((hssize_t)min_mod_chunk_sc[op_dim] == max_fill_chunk_sc[op_dim]) { fill_dim[op_dim] = TRUE; - else + + /* If necessary, check if chunks in this dimension that need to + * be filled are new partial edge chunks */ + if(disable_edge_filters && old_dim[op_dim] >= (min_mod_chunk_sc[op_dim] + 1)) + new_unfilt_dim[op_dim] = TRUE; + else + new_unfilt_dim[op_dim] = FALSE; + } /* end if */ + else { fill_dim[op_dim] = FALSE; + new_unfilt_dim[op_dim] = FALSE; + } /* end else */ } /* end if */ - else + else { fill_dim[op_dim] = FALSE; + new_unfilt_dim[op_dim] = FALSE; + } /* end else */ + + /* If necessary, calculate the smallest offset of non-previously full + * chunks in this dimension, so we know these chunks were previously + * unfiltered */ + if(disable_edge_filters) + min_partial_chunk_sc[op_dim] = old_dim[op_dim] / chunk_dim[op_dim]; } /* end for */ /* Main loop: fill or remove chunks */ @@ -4287,12 +4940,33 @@ H5D__chunk_prune_by_extent(H5D_t *dset, hid_t dxpl_id, const hsize_t *old_dim) while(!carry) { int i; /* Local index variable */ + udata.common.scaled = scaled; + if(0 == ndims_outside_fill) { HDassert(fill_dim[op_dim]); HDassert(scaled[op_dim] == min_mod_chunk_sc[op_dim]); + /* Make sure this is an edge chunk */ + HDassert(H5D__chunk_is_partial_edge_chunk(space_ndims, layout->u.chunk.dim, scaled, space_dim)); + + /* Determine if the chunk just became an unfiltered chunk */ + if(new_unfilt_dim[op_dim]) { + new_unfilt_chunk = TRUE; + for(u = 0; u < space_ndims; u++) + if(scaled[u] == min_partial_chunk_sc[u]) { + new_unfilt_chunk = FALSE; + break; + } /* end if */ + } /* end if */ + + /* Make sure that, if we think this is a new unfiltered chunk, + * it was previously not an edge chunk */ + HDassert(!new_unfilt_dim[op_dim] || (!new_unfilt_chunk != + !H5D__chunk_is_partial_edge_chunk(space_ndims, layout->u.chunk.dim, scaled, old_dim))); + HDassert(!new_unfilt_chunk || new_unfilt_dim[op_dim]); + /* Fill the unused parts of the chunk */ - if(H5D__chunk_prune_fill(&udata) < 0) + if(H5D__chunk_prune_fill(&udata, new_unfilt_chunk) < 0) HGOTO_ERROR(H5E_DATASET, H5E_WRITEERROR, FAIL, "unable to write fill value") } /* end if */ else { @@ -4325,7 +4999,7 @@ H5D__chunk_prune_by_extent(H5D_t *dset, hid_t dxpl_id, const hsize_t *old_dim) /* Remove the chunk from disk, if present */ if(H5F_addr_defined(chk_udata.chunk_block.offset)) { /* Update the offset in idx_udata */ - idx_udata.scaled = scaled; + idx_udata.scaled = udata.common.scaled; /* Remove the chunk from disk */ if((layout->storage.u.chunk.ops->remove)(&idx_info, &idx_udata) < 0) @@ -4447,12 +5121,14 @@ H5D__chunk_addrmap(const H5D_io_info_t *io_info, haddr_t chunk_addr[]) H5D_chk_idx_info_t idx_info; /* Chunked index info */ const H5D_t *dset = io_info->dset; /* Local pointer to dataset info */ H5D_chunk_it_ud2_t udata; /* User data for iteration callback */ + H5O_storage_chunk_t *sc = &(dset->shared->layout.storage.u.chunk); herr_t ret_value = SUCCEED; /* Return value */ FUNC_ENTER_PACKAGE HDassert(dset); HDassert(dset->shared); + H5D_CHUNK_STORAGE_INDEX_CHK(sc); HDassert(chunk_addr); /* Set up user data for B-tree callback */ @@ -4463,7 +5139,7 @@ H5D__chunk_addrmap(const H5D_io_info_t *io_info, haddr_t chunk_addr[]) /* Compose chunked index info struct */ idx_info.f = dset->oloc.file; - idx_info.dxpl_id = io_info->dxpl_id; + idx_info.dxpl_id = io_info->md_dxpl_id; idx_info.pline = &dset->shared->dcpl_cache.pline; idx_info.layout = &dset->shared->layout.u.chunk; idx_info.storage = &dset->shared->layout.storage.u.chunk; @@ -4500,6 +5176,7 @@ H5D__chunk_delete(H5F_t *f, hid_t dxpl_id, H5O_t *oh, H5O_storage_t *storage) H5O_pline_t pline; /* I/O pipeline message */ hbool_t pline_read = FALSE; /* Whether the I/O pipeline message was read from the file */ htri_t exists; /* Flag if header message of interest exists */ + H5O_storage_chunk_t *sc = &(storage->u.chunk); herr_t ret_value = SUCCEED; /* Return value */ FUNC_ENTER_PACKAGE @@ -4508,6 +5185,7 @@ H5D__chunk_delete(H5F_t *f, hid_t dxpl_id, H5O_t *oh, H5O_storage_t *storage) HDassert(f); HDassert(oh); HDassert(storage); + H5D_CHUNK_STORAGE_INDEX_CHK(sc); /* Check for I/O pipeline message */ if((exists = H5O_msg_exists_oh(oh, H5O_PLINE_ID)) < 0) @@ -4574,6 +5252,8 @@ H5D__chunk_update_cache(H5D_t *dset, hid_t dxpl_id) { H5D_rdcc_t *rdcc = &(dset->shared->cache.chunk); /*raw data chunk cache */ H5D_rdcc_ent_t *ent, *next; /*cache entry */ + H5D_rdcc_ent_t tmp_head; /* Sentinel entry for temporary entry list */ + H5D_rdcc_ent_t *tmp_tail; /* Tail pointer for temporary entry list */ H5D_dxpl_cache_t _dxpl_cache; /* Data transfer property cache buffer */ H5D_dxpl_cache_t *dxpl_cache = &_dxpl_cache; /* Data transfer property cache */ herr_t ret_value = SUCCEED; /* Return value */ @@ -4591,6 +5271,11 @@ H5D__chunk_update_cache(H5D_t *dset, hid_t dxpl_id) if(H5D__get_dxpl_cache(dxpl_id, &dxpl_cache) < 0) HGOTO_ERROR(H5E_DATASET, H5E_CANTGET, FAIL, "can't fill dxpl cache") + /* Add temporary entry list to rdcc */ + (void)HDmemset(&tmp_head, 0, sizeof(tmp_head)); + rdcc->tmp_head = &tmp_head; + tmp_tail = &tmp_head; + /* Recompute the index for each cached chunk that is in a dataset */ for(ent = rdcc->head; ent; ent = next) { unsigned old_idx; /* Previous index number */ @@ -4611,37 +5296,58 @@ H5D__chunk_update_cache(H5D_t *dset, hid_t dxpl_id) HDassert(old_ent->locked == FALSE); HDassert(old_ent->deleted == FALSE); - /* Mark the old entry as deleted, but do not evict (yet). - * Make sure we do not make any calls to the index + /* Insert the old entry into the temporary list, but do not + * evict (yet). Make sure we do not make any calls to the index * until all chunks have updated indices! */ - old_ent->deleted = TRUE; + HDassert(!old_ent->tmp_next); + HDassert(!old_ent->tmp_prev); + tmp_tail->tmp_next = old_ent; + old_ent->tmp_prev = tmp_tail; + tmp_tail = old_ent; } /* end if */ /* Insert this chunk into correct location in hash table */ rdcc->slot[ent->idx] = ent; - /* If this chunk was previously marked as deleted and therefore - * not in the hash table, reset the deleted flag. + /* If this chunk was previously on the temporary list and therefore + * not in the hash table, remove it from the temporary list. * Otherwise clear the old hash table slot. */ - if(ent->deleted) - ent->deleted = FALSE; + if(ent->tmp_prev) { + HDassert(tmp_head.tmp_next); + HDassert(tmp_tail != &tmp_head); + ent->tmp_prev->tmp_next = ent->tmp_next; + if(ent->tmp_next) { + ent->tmp_next->tmp_prev = ent->tmp_prev; + ent->tmp_next = NULL; + } /* end if */ + else { + HDassert(tmp_tail == ent); + tmp_tail = ent->tmp_prev; + } /* end else */ + ent->tmp_prev = NULL; + } /* end if */ else rdcc->slot[old_idx] = NULL; } /* end if */ } /* end for */ - /* Evict chunks that are still marked as deleted */ - for(ent = rdcc->head; ent; ent = next) { - /* Get the pointer to the next cache entry */ - next = ent->next; + /* tmp_tail is no longer needed, and will be invalidated by + * H5D_chunk_cache_evict anyways. */ + tmp_tail = NULL; + + /* Evict chunks that are still on the temporary list */ + while(tmp_head.tmp_next) { + ent = tmp_head.tmp_next; /* Remove the old entry from the cache */ - if(ent->deleted) - if(H5D__chunk_cache_evict(dset, dxpl_id, dxpl_cache, ent, TRUE) < 0) - HGOTO_ERROR(H5E_IO, H5E_CANTFLUSH, FAIL, "unable to flush one or more raw data chunks") - } /* end for */ + if(H5D__chunk_cache_evict(dset, dxpl_id, dxpl_cache, ent, TRUE) < 0) + HGOTO_ERROR(H5E_IO, H5E_CANTFLUSH, FAIL, "unable to flush one or more raw data chunks") + } /* end while */ done: + /* Remove temporary list from rdcc */ + rdcc->tmp_head = NULL; + FUNC_LEAVE_NOAPI(ret_value) } /* end H5D__chunk_update_cache() */ @@ -4678,7 +5384,6 @@ H5D__chunk_copy_cb(const H5D_chunk_rec_t *chunk_rec, void *_udata) hbool_t must_filter = FALSE; /* Whether chunk must be filtered during copy */ size_t nbytes; /* Size of chunk in file (in bytes) */ H5Z_cb_t cb_struct; /* Filter failure callback struct */ - int ret_value = H5_ITER_CONT; /* Return value */ FUNC_ENTER_STATIC @@ -4698,8 +5403,15 @@ H5D__chunk_copy_cb(const H5D_chunk_rec_t *chunk_rec, void *_udata) /* Check for filtered chunks */ if((is_vlen || fix_ref) && pline && pline->nused) { - must_filter = TRUE; - cb_struct.func = NULL; /* no callback function when failed */ + /* Check if we should disable filters on this chunk */ + if(udata->common.layout->flags + & H5O_LAYOUT_CHUNK_DONT_FILTER_PARTIAL_BOUND_CHUNKS) { + /* Check if the chunk is an edge chunk, and disable filters if so */ + if(!H5D__chunk_is_partial_edge_chunk(udata->dset_ndims, udata->common.layout->dim, chunk_rec->scaled, udata->dset_dims)) + must_filter = TRUE; + } /* end if */ + else + must_filter = TRUE; } /* end if */ /* Resize the buf if it is too small to hold the data */ @@ -4725,13 +5437,14 @@ H5D__chunk_copy_cb(const H5D_chunk_rec_t *chunk_rec, void *_udata) } /* end if */ /* read chunk data from the source file */ - if(H5F_block_read(udata->file_src, H5FD_MEM_DRAW, chunk_rec->chunk_addr, nbytes, udata->idx_info_dst->dxpl_id, buf) < 0) + if(H5F_block_read(udata->file_src, H5FD_MEM_DRAW, chunk_rec->chunk_addr, nbytes, H5AC_rawdata_dxpl_id, buf) < 0) HGOTO_ERROR(H5E_IO, H5E_READERROR, H5_ITER_ERROR, "unable to read raw data chunk") /* Need to uncompress variable-length & reference data elements */ if(must_filter) { unsigned filter_mask = chunk_rec->filter_mask; + cb_struct.func = NULL; /* no callback function when failed */ if(H5Z_pipeline(pline, H5Z_FLAG_REVERSE, &filter_mask, H5Z_NO_EDC, cb_struct, &nbytes, &buf_size, &buf) < 0) HGOTO_ERROR(H5E_PLINE, H5E_CANTFILTER, H5_ITER_ERROR, "data pipeline read failed") } /* end if */ @@ -4763,7 +5476,7 @@ H5D__chunk_copy_cb(const H5D_chunk_rec_t *chunk_rec, void *_udata) HGOTO_ERROR(H5E_DATATYPE, H5E_CANTINIT, H5_ITER_ERROR, "datatype conversion failed") /* Reclaim space from variable length data */ - if(H5D_vlen_reclaim(tid_mem, buf_space, H5P_DATASET_XFER_DEFAULT, reclaim_buf) < 0) + if(H5D_vlen_reclaim(tid_mem, buf_space, udata->idx_info_dst->dxpl_id, reclaim_buf) < 0) HGOTO_ERROR(H5E_DATASET, H5E_BADITER, H5_ITER_ERROR, "unable to reclaim variable-length data") } /* end if */ else if(fix_ref) { @@ -4806,13 +5519,16 @@ H5D__chunk_copy_cb(const H5D_chunk_rec_t *chunk_rec, void *_udata) udata->buf_size = buf_size; } /* end if */ + udata_dst.chunk_idx = H5VM_array_offset_pre(udata_dst.common.layout->ndims - 1, + udata_dst.common.layout->down_chunks, udata_dst.common.scaled); + /* Allocate chunk in the file */ if(H5D__chunk_file_alloc(udata->idx_info_dst, NULL, &udata_dst.chunk_block, &need_insert) < 0) HGOTO_ERROR(H5E_DATASET, H5E_CANTINSERT, FAIL, "unable to insert/resize chunk on chunk level") /* Write chunk data to destination file */ HDassert(H5F_addr_defined(udata_dst.chunk_block.offset)); - if(H5F_block_write(udata->idx_info_dst->f, H5FD_MEM_DRAW, udata_dst.chunk_block.offset, nbytes, udata->idx_info_dst->dxpl_id, buf) < 0) + if(H5F_block_write(udata->idx_info_dst->f, H5FD_MEM_DRAW, udata_dst.chunk_block.offset, nbytes, H5AC_rawdata_dxpl_id, buf) < 0) HGOTO_ERROR(H5E_DATASET, H5E_WRITEERROR, H5_ITER_ERROR, "unable to write raw data to file") /* Set metadata tag in dxpl_id */ @@ -4853,6 +5569,9 @@ H5D__chunk_copy(H5F_t *f_src, H5O_storage_chunk_t *storage_src, H5D_chunk_it_ud3_t udata; /* User data for iteration callback */ H5D_chk_idx_info_t idx_info_dst; /* Dest. chunked index info */ H5D_chk_idx_info_t idx_info_src; /* Source chunked index info */ + int sndims; /* Rank of dataspace */ + hsize_t curr_dims[H5O_LAYOUT_NDIMS]; /* Curr. size of dataset dimensions */ + hsize_t max_dims[H5O_LAYOUT_NDIMS]; /* Curr. size of dataset dimensions */ H5O_pline_t _pline; /* Temporary pipeline info */ const H5O_pline_t *pline; /* Pointer to pipeline info to use */ H5T_path_t *tpath_src_mem = NULL, *tpath_mem_dst = NULL; /* Datatype conversion paths */ @@ -4876,9 +5595,11 @@ H5D__chunk_copy(H5F_t *f_src, H5O_storage_chunk_t *storage_src, /* Check args */ HDassert(f_src); HDassert(storage_src); + H5D_CHUNK_STORAGE_INDEX_CHK(storage_src); HDassert(layout_src); HDassert(f_dst); HDassert(storage_dst); + H5D_CHUNK_STORAGE_INDEX_CHK(storage_dst); HDassert(ds_extent_src); HDassert(dt_src); @@ -4896,17 +5617,15 @@ H5D__chunk_copy(H5F_t *f_src, H5O_storage_chunk_t *storage_src, /* Initialize layout information */ { - hsize_t curr_dims[H5O_LAYOUT_NDIMS]; /* Curr. size of dataset dimensions */ - int sndims; /* Rank of dataspace */ unsigned ndims; /* Rank of dataspace */ /* Get the dim info for dataset */ - if((sndims = H5S_extent_get_dims(ds_extent_src, curr_dims, NULL)) < 0) + if((sndims = H5S_extent_get_dims(ds_extent_src, curr_dims, max_dims)) < 0) HGOTO_ERROR(H5E_DATASET, H5E_CANTGET, FAIL, "can't get dataspace dimensions") H5_CHECKED_ASSIGN(ndims, unsigned, sndims, int); /* Set the source layout chunk information */ - if(H5D__chunk_set_info_real(layout_src, ndims, curr_dims) < 0) + if(H5D__chunk_set_info_real(layout_src, ndims, curr_dims, max_dims) < 0) HGOTO_ERROR(H5E_DATASET, H5E_CANTSET, FAIL, "can't set layout's chunk info") } /* end block */ @@ -5108,6 +5827,7 @@ H5D__chunk_bh_info(const H5O_loc_t *loc, hid_t dxpl_id, H5O_t *oh, H5O_layout_t H5D_chk_idx_info_t idx_info; /* Chunked index info */ H5S_t *space = NULL; /* Dataset's dataspace */ H5O_pline_t pline; /* I/O pipeline message */ + H5O_storage_chunk_t *sc = &(layout->storage.u.chunk); htri_t exists; /* Flag if header message of interest exists */ hbool_t idx_info_init = FALSE; /* Whether the chunk index info has been initialized */ hbool_t pline_read = FALSE; /* Whether the I/O pipeline message was read */ @@ -5120,6 +5840,7 @@ H5D__chunk_bh_info(const H5O_loc_t *loc, hid_t dxpl_id, H5O_t *oh, H5O_layout_t HDassert(loc->file); HDassert(H5F_addr_defined(loc->addr)); HDassert(layout); + H5D_CHUNK_STORAGE_INDEX_CHK(sc); HDassert(index_size); /* Check for I/O pipeline message */ @@ -5229,12 +5950,14 @@ H5D__chunk_dump_index_cb(const H5D_chunk_rec_t *chunk_rec, void *_udata) herr_t H5D__chunk_dump_index(H5D_t *dset, hid_t dxpl_id, FILE *stream) { + H5O_storage_chunk_t *sc = &(dset->shared->layout.storage.u.chunk); herr_t ret_value = SUCCEED; /* Return value */ FUNC_ENTER_PACKAGE /* Sanity check */ HDassert(dset); + H5D_CHUNK_STORAGE_INDEX_CHK(sc); /* Only display info if stream is defined */ if(stream) { @@ -5423,7 +6146,7 @@ H5D__nonexistent_readvv(const H5D_io_info_t *io_info, /* Set up user data for H5VM_opvv() */ udata.rbuf = (unsigned char *)io_info->u.rbuf; udata.dset = io_info->dset; - udata.dxpl_id = io_info->dxpl_id; + udata.dxpl_id = io_info->md_dxpl_id; /* Call generic sequence operation routine */ if((ret_value = H5VM_opvv(chunk_max_nseq, chunk_curr_seq, chunk_len_arr, chunk_off_arr, @@ -5437,6 +6160,45 @@ done: /*------------------------------------------------------------------------- + * Function: H5D__chunk_is_partial_edge_chunk + * + * Purpose: Checks to see if the chunk is a partial edge chunk. + * Either dset or (dset_dims and dset_ndims) must be + * provided. + * + * Return: Non-negative on success/Negative on failure + * + * Programmer: Neil Fortner + * 19 Nov 2009 + * + *------------------------------------------------------------------------- + */ +static hbool_t +H5D__chunk_is_partial_edge_chunk(unsigned dset_ndims, const uint32_t *chunk_dims, + const hsize_t scaled[], const hsize_t *dset_dims) +{ + unsigned u; /* Local index variable */ + hbool_t ret_value = FALSE; /* Return value */ + + FUNC_ENTER_STATIC_NOERR + + /* Check args */ + HDassert(scaled); + HDassert(dset_ndims > 0); + HDassert(dset_dims); + HDassert(chunk_dims); + + /* check if this is a partial edge chunk */ + for(u = 0; u < dset_ndims; u++) + if(((scaled[u] + 1) * chunk_dims[u]) > dset_dims[u]) + HGOTO_DONE(TRUE); + +done: + FUNC_LEAVE_NOAPI(ret_value) +} /* H5D__chunk_is_partial_edge_chunk() */ + + +/*------------------------------------------------------------------------- * Function: H5D__chunk_file_alloc() * * Purpose: Chunk allocation: @@ -5469,6 +6231,8 @@ H5D__chunk_file_alloc(const H5D_chk_idx_info_t *idx_info, const H5F_block_t *old HDassert(new_chunk); HDassert(need_insert); + *need_insert = FALSE; + /* Check for filters on chunks */ if(idx_info->pline->nused > 0) { /* Sanity/error checking block */ @@ -5523,21 +6287,24 @@ H5D__chunk_file_alloc(const H5D_chk_idx_info_t *idx_info, const H5F_block_t *old /* Actually allocate space for the chunk in the file */ if(alloc_chunk) { - switch(idx_info->storage->idx_type) { - case H5D_CHUNK_IDX_BTREE: + switch(idx_info->storage->idx_type) { + case H5D_CHUNK_IDX_EARRAY: + case H5D_CHUNK_IDX_FARRAY: + case H5D_CHUNK_IDX_BT2: + case H5D_CHUNK_IDX_BTREE: HDassert(new_chunk->length > 0); - H5_CHECK_OVERFLOW(new_chunk->length, /*From: */uint32_t, /*To: */hsize_t); - new_chunk->offset = H5MF_alloc(idx_info->f, H5FD_MEM_DRAW, idx_info->dxpl_id, (hsize_t)new_chunk->length); - if(!H5F_addr_defined(new_chunk->offset)) - HGOTO_ERROR(H5E_DATASET, H5E_CANTALLOC, FAIL, "file allocation failed") - *need_insert = TRUE; - break; - - case H5D_CHUNK_IDX_NTYPES: - default: - HDassert(0 && "This should never be executed!"); - break; - } /* end switch */ + H5_CHECK_OVERFLOW(new_chunk->length, /*From: */uint32_t, /*To: */hsize_t); + new_chunk->offset = H5MF_alloc(idx_info->f, H5FD_MEM_DRAW, idx_info->dxpl_id, (hsize_t)new_chunk->length); + if(!H5F_addr_defined(new_chunk->offset)) + HGOTO_ERROR(H5E_DATASET, H5E_CANTALLOC, FAIL, "file allocation failed") + *need_insert = TRUE; + break; + + case H5D_CHUNK_IDX_NTYPES: + default: + HDassert(0 && "This should never be executed!"); + break; + } /* end switch */ } /* end if */ HDassert(H5F_addr_defined(new_chunk->offset)); |