Add new internal data structure
Description:
Add an implementation of skip lists to the library (see comment in
src/H5SL.c for references to the papers describing them) as a potential
replacement for our current threaded, balanced binary tree container.
Skip lists are much simpler to implement and should be faster to use.
Also, added new error codes to release branch, so bump the minor version
number to indicate that the library is no longer perfectly compatible with
the 1.6.3 release.
Platforms tested:
FreeBSD 4.10 (sleipnir) w/parallel
Solaris 2.7 (arabica)
Too minor to require further testing (the skip lists aren't actually
used by any library code yet)
Cleanup & add test
Description:
Clean up test makefile a bit and add metadata cache test to 1.6 branch.
Platforms tested:
FreeBSD 4.10 (sleipnir)
Too minor to require h5committest
Bug fix
Description:
Fix situation where deleting a chunked datasets with B-tree nodes that
weren't in the metadata cache would die with a core dump.
Platforms tested:
FreeBSD 4.10 (sleipnir)
Linux 2.4 (heping)
Solaris 2.7 (arabica)
Code cleanup
Description:
Various minor tweaks to clean code up and bring it into closer
syncronization with the release branch.
Platforms tested:
FreeBSD 4.10 (sleipnir) w/parallel
h5committested
IRIX64 6.5 (modi4)
Purpose:
Bug Fix
Description:
If an HDF5 file grows larger than its address space, it dies and is unable to
write any data. This is more likely to happen since users are able to change
the number of bytes used to store addresses in the file.
Solution:
HDF5 now throws an error instead of dying. In addition, it "reserves" address
space for the local heap and for object headers (which do not allocate space
immediately). This ensures that after the error occurs, there is enough address
space left to flush the entire file to disk, so no data is lost.
A more complete explanation is at /doc/html/TechNotes/ReservedFileSpace.html
Platforms tested:
sleipnir, copper (parallel), verbena, arabica, Windows (Visual Studio 7)
Solution:
Platforms tested:
Misc. update:
Description:
Replaced the old metadata cache with a cache with a modified LRU
replacement policy. This should improve the hit rate.
Solution:
Since we want to flush cache entries in increasing address order, I
used the threaded binary B-tree code to store the cache entries.
There is a fair bit of overhead here, so we may want to consider
other options.
While the code is designed to allow the support of other replacement
algorithms, at present, only a modified version of LRU is supported.
The modified LRU algorithm requires that a user selectable portion
of the cache entries be clean. The clean entries are evicted first
when writes are not permitted. If the pool of clean entries is used
up, the cache grows beyond its user specified maximum size. The
cache can also exceed its maximum size if the combined size of the
protected (or locked) entries exceeds the maximum size of the cache.
Platforms tested:
eirene (serial, parallel, fp), h5committested
Misc. update:
Description: Added new API H5Fget_name and new test program called filename.c. This function
returns the name of the file by object ID(file, group, dataset, named datatype, and attribute)
which belongs to the file.
Platforms tested: h5committest and fuss.
Misc. update: MANIFEST and RELEASE.txt
Code cleanup
Description:
Add more files that should be removed to the clean & distclean targets.
Platforms tested:
FreeBSD 4.10 (sleipnir)
Too minor to require h5committest
New Feature
Description:
Add the data transform function, H5Pset_transform().
Platforms tested:
"h5committested".
Copper was down. Ran parallel tests in sol instead.
Misc. update:
Code cleanup.
Description:
The routines residing in testframe.c was defined in testhdf5.h due to
historical reason. It really belongs to h5test.h because those routines
reside in libh55test.a.
Solution:
Moved them to the right place. Also removed the duplicated occurance
of testframe.c in the TEST_SRC.
Platforms tested:
Tested in o2 (SGI) parallel.
Bug fix
Description:
Correct two problems with variable-length datatypes in datasets:
- When overwriting an entire dataset, writing the fill value to the
file would be skipped, causing problems for VL datatypes when
objects in the file had been unlinked (and thus the space in the
file was not all zeros)
- When an application has set a fill-value for a dataset and the
dataset's datatype contained a VL datatype, the library was filling
space on disk with the memory form of the VL information, instead
of the disk form.
Platforms tested:
FreeBSD 4.9 (sleipnir)
h5committest
Code cleanup & reorganization
Description:
Move further in the testing framework cleanup, eliminating all the
global variables (moving them into testframe.c as static variables) from the
testing framework code and moving it into the libh5test.a.
Platforms tested:
FreeBSD 4.9 (sleipnir) w & w/o thread-safety, c++ & parallel
h5committested
Code cleanup
Description:
Refactor library testing framework (used for the testhdf5 & ttsafe tests)
to remove almost all of the duplicated code, moving the common code into a
new 'testframe.c' source file.
Platforms tested:
FreeBSD 4.9 (sleipnir) w & w/o thread-safety
h5committest
Code cleanup & add a feature
Description:
Added support for querying the file ID of named datatypes to new
H5Iget_file_id function.
Platforms tested:
FreeBSD 4.9 (sleipnir)
too minor to require h5committest
Description: The error test script compares the error messages to the standard
output. The file names and line numbers in the messages can be different
for users.
Solution: Temporarily disable the test until find a good solution.
Platforms tested: None. Disabling a test.
Bug fix
Description:
When datasets are deleted from a file, they are removed from the sieve
buffer, but instead of invalidating only the part of the sieve buffer affected,
the sieve buffer code would throw away the entire sieve buffer, potentially
including other raw data in the buffer that hadn't been written to disk yet.
Solution:
Improve the sieve buffer clearing code to handle partial invalidations.
Platforms tested:
FreeBSD 4.8 (sleipnir)
h5committest
Bug fix
Description:
Allow a user block to be "inserted" in front of a file (probably by
writing a validly-sized userblock to a new file and then appending another
HDF5 file to the new file).
Platforms tested:
FreeBSD 4.8 (sleipnir)
h5committested
New test
Description:
Added test to check for dangling objects being handled correctly by the
library.
Platforms tested:
FreeBSD 4.8 (sleipnir)
h5committested
New test feature
Description:
Added the VFD_LIST definition to support the "make check-vfd" tests.
Platforms tested:
h5committested.
Since check-vfd is not invoked by default, I went to each machine
to invoke them by hand. The target worked fine but there were
errors reported from those driver tests. (Those errors were beyond
the scope of this make target test.)
Misc. update:
New internal feature
Description:
Add internal API for building and working with heaps (H5HP). This will be
used for the LRU algorithm in the new metadata cache code.
Platforms tested:
Tested h5committest {arabica (fortran), eirene (fortran, C++)
modi4 (parallel, fortran)}
FreeBSD 4.7 (sleipnir)
Bug Fix
Description:
When calling H5Fopen with the core VFL driver, but without the
H5F_ACC_CREAT flag goes ahead and creates a memory file.
Solution:
Check for the H5F_ACC_CREAT flag before allowing the memory file to be
created.
Platforms tested:
FreeBSD 4.7 (sleipnir)
Lots of performance improvements & a couple new internal API interfaces.
Description:
Performance Improvements:
- Cached file offset & length sizes in shared file struct, to avoid
constantly looking them up in the FCPL.
- Generic property improvements:
- Added "revision" number to generic property classes to speed
up comparisons.
- Changed method of storing properties from using a hash-table
to the TBBT routines in the library.
- Share the propery names between classes and the lists derived
from them.
- Removed redundant 'def_value' buffer from each property.
- Switching code to use a "copy on write" strategy for
properties in each list, where the properties in each list
are shared with the properties in the class, until a
property's value is changed in a list.
- Fixed error in layout code which was allocating too many buffers.
- Redefined public macros of the form (H5open()/H5check, <variable>)
internally to only be (<variable>), avoiding innumerable useless
calls to H5open() and H5check_version().
- Reuse already zeroed buffers in H5F_contig_fill instead of
constantly re-zeroing them.
- Don't write fill values if writing entire dataset.
- Use gettimeofday() system call instead of time() system when
checking the modification time of a dataset.
- Added reference counted string API and use it for tracking the
names of objects opening in a file (for the ID->name code).
- Removed redundant H5P_get() calls in B-tree routines.
- Redefine H5T datatype macros internally to the library, to avoid
calling H5check redundantly.
- Keep dataspace information for dataset locally instead of reading
from disk each time. Added new module to track open objects
in a file, to allow this (which will be useful eventually for
some FPH5 metadata caching issues).
- Remove H5AC_find macro which was inlining metadata cache lookups,
and call function instead.
- Remove redundant memset() calls from H5G_namei() routine.
- Remove redundant checking of object type when locating objects
in metadata cache and rely on the address only.
- Create default dataset object to use when default dataset creation
property list is used to create datasets, bypassing querying
for all the property list values.
- Use default I/O vector size when performing raw data with the
default dataset transfer property list, instead of querying for
I/O vector size.
- Remove H5P_DEFAULT internally to the library, replacing it with
more specific default property list based on the type of
property list needed.
- Remove redundant memset() calls in object header message (H5O*)
routines.
- Remove redunant memset() calls in data I/O routines.
- Split free-list allocation routines into malloc() and calloc()-
like routines, instead of one combined routine.
- Remove lots of indirection in H5O*() routines.
- Simplify metadata cache entry comparison routine (used when
flushing entire cache out).
- Only enable metadata cache statistics when H5AC_DEBUG is turned
on, instead of always tracking them.
- Simplify address comparison macro (H5F_addr_eq).
- Remove redundant metadata cache entry protections during dataset
creation by protecting the object header once and making all
the modifications necessary for the dataset creation before
unprotecting it.
- Reduce # of "number of element in extent" computations performed
by computing and storing the value during dataspace creation.
- Simplify checking for group location's file information, when file
has not been involving in file-mounting operations.
- Use binary encoding for modification time, instead of ASCII.
- Hoist H5HL_peek calls (to get information in a local heap)
out of loops in many group routine.
- Use static variable for iterators of selections, instead of
dynamically allocation them each time.
- Lookup & insert new entries in one step, avoiding traversing
group's B-tree twice.
- Fixed memory leak in H5Gget_objname_idx() routine (tangential to
performance improvements, but fixed along the way).
- Use free-list for reference counted strings.
- Don't bother copying object names into cached group entries,
since they are re-created when an object is opened.
The benchmark I used to measure these results created several thousand
small (2K) datasets in a file and wrote out the data for them. This is
Elena's "regular.c" benchmark.
These changes resulted in approximately ~4.3x speedup of the
development branch when compared to the previous code in the
development branch and ~1.4x speedup compared to the release
branch.
Additionally, these changes reduce the total memory used (code and
data) by the development branch by ~800KB, bringing the development
branch back into the same ballpark as the release branch.
I'll send out a more detailed description of the benchmark results
as a followup note.
New internal API routines:
Added "reference counted strings" API for tracking strings that get
used by multiple owners without duplicating the strings.
Added "ternary search tree" API for text->object mappings.
Platforms tested:
Tested h5committest {arabica (fortran), eirene (fortran, C++)
modi4 (parallel, fortran)}
Other platforms/configurations tested?
FreeBSD 4.7 (sleipnir) serial & parallel
Solaris 2.6 (baldric) serial
Code cleanup
Description:
Added more test datafiles to clean directive.
Platforms tested:
FreeBSD 4.6 (sleipnir) w and w/o parallel
Linux 2.2.x (eirene) w/FORTRAN & C++
Solaris 2.7 (arabica) w/FORTRAN
IRIX64 6.5 (modi4) w/FORTRAN & parallel
Purpose:
New API functions
Description:
Added API functions to return pointer to low-level file handle
(H5Fget_vfd_handle and H5FDget_vfd_handle) and related property list
setting functions(H5Pset_family_offset and H5Pset_multi_type).
Platforms tested:
Linux 2.2(eirene), Solaris 2.7(arabica), IRIX64 6.5(modi4)
Purpose:
Added 'ID to name' support
Description:
There is a new API function H5Iget_name
Most of the changes are on H5G.c , regarding the symbol table entry struct H5G_entry_t
which has 2 new fields 'name' and 'old_name'
A new private function was introduced H5G_ent_copy, that does a deep copy
between 2 symbol table entries
The test file is getname.c
Platforms tested:
windows 2000, Linux, Solaris
Code cleanup/More tests
Description:
Cleaned up some compiler warnings and wrote additional tests for space
allocation and storage size routines.
Platforms tested:
FreeBSD 4.6 (sleipnir) w/serial & parallel. Will be testing on IRIX64
6.5 (modi4) in serial & parallel shortly.
Additional regression tests & bug fixes
Description:
There was no testing for the H5Dget_storage_size function and it seemed to
be having problems with compressed, chunked datasets, so write some tests
to verify that its working correctly.
Also, fix case for allocating storage early for chunked datasets
Platforms tested:
FreeBSD 4.6 (sleipnir) serial & parallel
Bug Fix
Description:
H5Dcreate and H5Tcommit allow "empty" compound and enumerated types (i.e.
ones with no members) to be stored in the file, but this causes an assertion
failure and is somewhat vapid.
Solution:
Check the datatype "makes sense" before using it for H5Dcreate and
H5Tcommit.
Platforms tested:
FreeBSD 4.5 (sleipnir)
Test Bug Fix
Description:
Under certain [obscure] circumstances, an object header would get paged out
of the metadata cache, and when it was accessed again and brought back into
the cache, and immediately had additional metadata added to it (an
attribute, usually, or perhaps adding an object to a group), and needed to
be extended with a continuation message, but there was no room in any
existing object header chunks for the continuation message and an existing
object header message needed to be moved to the new object header chunk (I
told you it was obscure :-), the object header message moved to the new
chunk (not the new metadata being added) would get corrupted. *whew* :-)
Solution:
Actually copy the "raw" object header message information of the object
header message being moved to the new chunk, instead of relying on the
"native" object header message information being re-encoded when the object
header is flushed. This is because when an object header is paged out of
the metadata cache and subsequently brought back in, the "native"
information pointer in memory is reset to NULL and only the "raw"
information exists.
[Actually, this additional testing doesn't trigger the bug, which needs
_lots_ of objects to be created and accessed, but it does execise the
object header continuation code more than other tests in the library.]
Platforms tested:
Solaris 2.7 (arabica) & FreeBSD 4.5 (sleipnir)
Bug fix
Description:
$(LT_RUN) was used to execute ./H5detect in src/Makefile.in.
$RUNTEST was set to $(LT_RUN) which was used to execute sequential
executable. $(LT_RUN) is "../libtool --mode=execute". But libtool
invokes some commands that are not supported in the Tflops machine.
That caused failures during the build and check processes.
Solution:
Upon investigation, there does not seem to be need to use $(LT_RUN)
any more. The "libtool --mode=link" now generates a "fake" executable
that is actually a command-script file that can regenerate the real
executable with dynamic libraries hooks setup properly.
Undo all those $LT_RUN substitute and let $RUNSERIAL execute those
sequential executables.
Platforms tested:
Eirene(serial), modi4(parallel)
Bug fix
Description:
When several level deep nested compound & VL datatypes are used, the data
in the nested compound datatypes is incorrectly sharing the same "background
buffer", causing data corruption when the data is written to the file.
Solution:
Allocate a separate background buffer for each level of the nested types
to convert. (Also allocate temporary background buffers for array
datatypes, where this sort of problem could occur also)
Added more regression tests to check for these errors.
Platforms tested:
FreeBSD 4.5 (sleipnir) & Solaris 2.6 (baldric)
Bug Fix
Description:
Regression test for following bug:
The H5Gget_objinfo() function was not setting the 'fileno' field in the
H5G_stat_t struct passed in.
Solution:
Added a "file serial number" to each file currently open in the library
and put that in the 'fileno' field. If a file is opened twice (with
H5Fopen) and the VFL driver detects that it is the same file (i.e. the
two file structures have the same "shared file info" in the library's
memory structures), they will have the same serial number.
This serial number has two drawbacks:
- If a VFL driver doesn't/can't detect that two calls to H5Fopen with
the same file actually _are_ the same file, each will get a
different serial number
- If the same file is closed and re-opened, the serial number will be
different.
It is be possible to fix the second drawback for many VFL drivers, but it
would be a lot of effort and probably isn't worth it until we've got a
good reason to do it. Dunno if we'll ever be able to fix the first
drawback...
Platforms tested:
FreeBSD 4.5 (sleipnir)
VS: ----------------------------------------------------------------------
Cleanup
Description:
Added set_extend.h5 file to the CLEANUP macro in the Makefile so that
it'll be removed when the other temp files are removed.
Platforms tested:
Linux
Bug Fix
Description:
When file space was returned to the file space free-list for reuse,
occasionally raw data allocations which used space from the free-list
would overlap with the metadata accumulator and get over-written with
the cached information in the accumulator, corrupting the data.
Solution:
Check if the space about to be recycled on the free-list is going to be
used for raw data and also overlaps with the metadata accumulator cache,
avoiding using space that fits those criteria.
This fixes bug #701
Platforms tested:
FreeBSD 4.5 (sleipnir)
