Bring some of Neil's changes to help support SWMR access in the v1
B-tree code from the revise_chunks branch back to the trunk.
Tested on:
FreeBSD/32 6.3 (duty) in debug mode
FreeBSD/64 6.3 (liberty) w/C++ & FORTRAN, in debug mode
Linux/32 2.6 (jam) w/PGI compilers, w/default API=1.8.x,
w/C++ & FORTRAN, w/threadsafe, in debug mode
Linux/64-amd64 2.6 (amani) w/Intel compilers, w/default API=1.6.x,
w/C++ & FORTRAN, in production mode
Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN,
w/szip filter, w/threadsafe, in production mode
Linux/PPC 2.6 (heiwa) w/C++ & FORTRAN, w/threadsafe, in debug mode
Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN,
in production mode
Linux/64-amd64 2.6 (abe) w/parallel, w/FORTRAN, in debug mode
Mac OS X/32 10.6.4 (amazon) in debug mode
Mac OS X/32 10.6.4 (amazon) w/C++ & FORTRAN, w/threadsafe,
in production mode
Mac OS X/32 10.6.4 (amazon) w/parallel, in debug mode
- Fix a bug in each of the metadata accumulator source and test code
Description:
- In accum.c test file, switch dxpl_id used in H5F_accum_*
function calls to H5P_DATASET_XFER_DEFAULT (instead of
H5AC_dxpl_id), to fix compilation on windows.
- Changed boundary checking from <= to < when checking
if a read from disk with overlapping dirty metadata in the
accumulator has the read ending such that it aligns exactly
with the dirty accumulator (line 234 of H5Faccum.c).
Tested:
- h5committested
of the H5Ocache.c code to update its image of the on disk representation
of the object header on a call to the clear callback.
This wasn't an issue as long as all flushes of the object header were
made from the same process, but if an object header is modified, and
then flushed on one process and cleared on the rest, the changes were
not be reflected in the images of the on disk representation on all
processes where the object header was cleared rather than flushed.
If one of these processes did the next flush, the changes were lost in
the on disk representation.
Fixed this by causing all dirty messages and to be written to the copy
of the on disk image maintained by the object header code on both flush
and clear.
Also added associated test code in t_mdset.c.
Also checking in some cache debug code developed while chasing this bug.
Commit tested and tested (parallel) on phoenix.
Merge accum_tests branch back to the trunk.
Description:
Changes consist of addition of tests for H5Faccum.c source code,
as well as some fixes that address some discovered bugs in the
metadata accumulator on several corner cases.
Tested:
h5committested
FAILED. Rejected by the xlc compiler.
Solution: changed it to reutrn NULL.
Tested: UP make passed. serial passed but some parallel tests failed. The
parallel tests failed were not related to this fix.
Address issue with object headers being created getting evicted from
the metadata cache cache before they are completely initialized. This is
done by pinning the object header in the cache until it is completely
initialized and attached to a group.
Tested on:
FreeBSD/32 6.3 (duty) in debug mode
FreeBSD/64 6.3 (liberty) w/C++ & FORTRAN, in debug mode
Linux/32 2.6 (jam) w/PGI compilers, w/default API=1.8.x,
w/C++ & FORTRAN, w/threadsafe, in debug mode
Linux/64-amd64 2.6 (amani) w/Intel compilers, w/default API=1.6.x,
w/C++ & FORTRAN, in production mode
Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN,
w/szip filter, w/threadsafe, in production mode
Linux/PPC 2.6 (heiwa) w/C++ & FORTRAN, w/threadsafe, in debug mode
Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN,
in production mode
Linux/64-amd64 2.6 (abe) w/parallel, w/FORTRAN, in debug mode
Mac OS X/32 10.6.4 (amazon) in debug mode
Mac OS X/32 10.6.4 (amazon) w/C++ & FORTRAN, w/threadsafe,
in production mode
Mac OS X/32 10.6.4 (amazon) w/parallel, in debug mode
with the 1.8 library to compile. The cause is from the mismatch of H5Eprint1 and H5Eprint2
set through H5Eset_auto. I changed the structure H5E_auto_t. I added a IS_DEDAULT flag in
it. Both H5Eprint1/2 are the default now. If the user sets his/her own printing function.
Then a call to H5Eget_auto1/2 will have to match H5Eset_auto1/2.
Tested on heiwa, jam, and amani.
The property change in configure.in, config, and Makefile.am came from the merge of the 1.8
library change.
Patch for metadata accumulator bug
Description:
Linew failed one of the randomly seeded fheap tests
due to a corner case bug in the metadata accumulator code.
This patch fixes the corner case.
Tested:
used same random seed in fheap tests to reproduce and
verify on linew; plus full make check on jam.
unit tests for accumulator code coming soon.
flag IS_DEDAULT in the H5E_auto_t structure. Both H5Eprint1/2 are the default now. If the user
sets his/her own printing function. Then a call to H5Eget_auto1/2 will have to match
H5Eset_auto1/2.
Tested on jam, heiwa, and amani.
Optimize the vector-vector memcpy() routine even further, for a total
of ~2x speedup. :-)
Make a generic vector-vector operation routine and convert other
vector-vector read & write routines to use generic routine instead of multiple
copies of the basic algorithm.
Tested on:
FreeBSD/32 6.3 (duty) in debug mode
FreeBSD/64 6.3 (liberty) w/C++ & FORTRAN, in debug mode
Linux/32 2.6 (jam) w/PGI compilers, w/default API=1.8.x,
w/C++ & FORTRAN, w/threadsafe, in debug mode
Linux/64-amd64 2.6 (amani) w/Intel compilers, w/default API=1.6.x,
w/C++ & FORTRAN, in production mode
Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN,
w/szip filter, w/threadsafe, in production mode
Linux/PPC 2.6 (heiwa) w/C++ & FORTRAN, w/threadsafe, in debug mode
Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN,
in production mode
Linux/64-amd64 2.6 (abe) w/parallel, w/FORTRAN, in debug mode
Mac OS X/32 10.6.4 (amazon) in debug mode
Mac OS X/32 10.6.4 (amazon) w/C++ & FORTRAN, w/threadsafe,
in production mode
Mac OS X/32 10.6.4 (amazon) w/parallel, in debug mode
Speed up hyperslab sequence generation routine some more, giving
about a 10% improvement in Ger Van Diepen's LOFAR benchmark.
Tested on:
FreeBSD/32 6.3 (duty) in debug mode
FreeBSD/64 6.3 (liberty) w/C++ & FORTRAN, in debug mode
Linux/32 2.6 (jam) w/PGI compilers, w/default API=1.8.x,
w/C++ & FORTRAN, w/threadsafe, in debug mode
Linux/64-amd64 2.6 (amani) w/Intel compilers, w/default API=1.6.x,
w/C++ & FORTRAN, in production mode
Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN,
w/szip filter, w/threadsafe, in production mode
Linux/PPC 2.6 (heiwa) w/C++ & FORTRAN, w/threadsafe, in debug mode
Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN,
in production mode
Linux/64-amd64 2.6 (abe) w/parallel, w/FORTRAN, in debug mode
Mac OS X/32 10.6.4 (amazon) in debug mode
Mac OS X/32 10.6.4 (amazon) w/C++ & FORTRAN, w/threadsafe,
in production mode
Mac OS X/32 10.6.4 (amazon) w/parallel, in debug mode
Bug 1707 is that H5Eset_auto causes a seg fault when an application uses -DH5_USE_16_API with
the 1.8 library to compile.
I created a branch off the trunk called set_auto to continue working on the problem.
Tested on jam - backing out, simple.
Polish the vectorized memory <-> memory and memory <-> file I/O
routines to speed them up a little bit (~10%)
Tested on:
FreeBSD/32 6.3 (duty) in debug mode
FreeBSD/64 6.3 (liberty) w/C++ & FORTRAN, in debug mode
Linux/32 2.6 (jam) w/PGI compilers, w/default API=1.8.x,
w/C++ & FORTRAN, w/threadsafe, in debug mode
Linux/64-amd64 2.6 (amani) w/Intel compilers, w/default API=1.6.x,
w/C++ & FORTRAN, in production mode
Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN,
w/szip filter, w/threadsafe, in production mode
Linux/PPC 2.6 (heiwa) w/C++ & FORTRAN, w/threadsafe, in debug mode
Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN,
in production mode
Linux/64-amd64 2.6 (abe) w/parallel, w/FORTRAN, in debug mode
Mac OS X/32 10.6.4 (amazon) in debug mode
Mac OS X/32 10.6.4 (amazon) w/C++ & FORTRAN, w/threadsafe,
in production mode
Mac OS X/32 10.6.4 (amazon) w/parallel, in debug mode
Create a new, more streamlined method of generating I/O sequences when
there is only a single block defined by a hyperslab. This improves the
performance of Ger Van Diepen's (LOFAR team) benchmark by about 20%.
Tested on:
FreeBSD/32 6.3 (duty) in debug mode
FreeBSD/64 6.3 (liberty) w/C++ & FORTRAN, in debug mode
Linux/32 2.6 (jam) w/PGI compilers, w/default API=1.8.x,
w/C++ & FORTRAN, w/threadsafe, in debug mode
Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN,
w/szip filter, w/threadsafe, in production mode
Linux/PPC 2.6 (heiwa) w/C++ & FORTRAN, w/threadsafe, in debug mode
Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN,
in production mode
Linux/64-amd64 2.6 (abe) w/parallel, w/FORTRAN, in debug mode
Mac OS X/32 10.6.4 (amazon) in debug mode
Mac OS X/32 10.6.4 (amazon) w/C++ & FORTRAN, w/threadsafe,
in production mode
Mac OS X/32 10.6.4 (amazon) w/parallel, in debug mode
deprecated symbols were disabled. I made a different definition for H5E_auto_op_t when the
deprecated symbols were disabled.
Tested on jam - a simple change.
Description:
Library versions 1.6.3 and earlier contain a bug which causes them to be unable
to perform certain operations on a group if that group's symbol table
information is not cached in the parent group's symbol table. Versions 1.8.0
to 1.8.5 did not cache this information. Modified library to cache this
information.
Tested: jam, amani, heiwa (h5committest)
through H5Eset_auto. I changed the union in the structure H5E_auto_t. Another change is to
make H5Eget_auto fail if H5Eset_auto is called to set the printing function. I'll write a
document for it.
Tested on heiwa, jam, and amani.
The property change in configure.in, config, and Makefile.am came from the merge of the 1.8
library change.
Correct Bz#1968 (Problem when opening new format file with
H5F_LIBVER_EARLIEST) by initializing symbol table leaf information in a
more universal way. Also cleaned up other superblock initialization a
little bit while I'm here.
Tested on:
FreeBSD/32 6.3 (duty) in debug mode
FreeBSD/64 6.3 (liberty) w/C++ & FORTRAN, in debug mode
Linux/32 2.6 (jam) w/PGI compilers, w/default API=1.8.x,
w/C++ & FORTRAN, w/threadsafe, in debug mode
Linux/64-amd64 2.6 (amani) w/Intel compilers, w/default API=1.6.x,
w/C++ & FORTRAN, in production mode
Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN,
w/szip filter, w/threadsafe, in production mode
Linux/PPC 2.6 (heiwa) w/C++ & FORTRAN, w/threadsafe, in debug mode
Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN,
in production mode
Linux/64-amd64 2.6 (abe) w/parallel, w/FORTRAN, in debug mode
Mac OS X/32 10.6.4 (amazon) in debug mode
Mac OS X/32 10.6.4 (amazon) w/C++ & FORTRAN, w/threadsafe,
in production mode
Mac OS X/32 10.6.4 (amazon) w/parallel, in debug mode
Clean up the internal usage H5I_dec_ref() to eliminate the sequence of
flags that has been creeping in.
Tested on:
FreeBSD/32 6.3 (duty) in debug mode
FreeBSD/64 6.3 (liberty) w/C++ & FORTRAN, in debug mode
Linux/32 2.6 (jam) w/PGI compilers, w/default API=1.8.x,
w/C++ & FORTRAN, w/threadsafe, in debug mode
Linux/64-amd64 2.6 (amani) w/Intel compilers, w/default API=1.6.x,
w/C++ & FORTRAN, in production mode
Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN,
w/szip filter, w/threadsafe, in production mode
Linux/PPC 2.6 (heiwa) w/C++ & FORTRAN, w/threadsafe, in debug mode
Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN,
in production mode
Linux/64-amd64 2.6 (abe) w/parallel, w/FORTRAN, in debug mode
Mac OS X/32 10.6.4 (amazon) in debug mode
Mac OS X/32 10.6.4 (amazon) w/C++ & FORTRAN, w/threadsafe,
in production mode
Mac OS X/32 10.6.4 (amazon) w/parallel, in debug mode
Bring back various modifications to the v2 B-trees that were introduced
in the revise_chunks branch but aren't dependent on any file format or API
changes.
Tested on:
FreeBSD/32 6.3 (duty) in debug mode
FreeBSD/64 6.3 (liberty) w/C++ & FORTRAN, in debug mode
Linux/32 2.6 (jam) w/PGI compilers, w/default API=1.8.x,
w/C++ & FORTRAN, w/threadsafe, in debug mode
Linux/64-amd64 2.6 (amani) w/Intel compilers, w/default API=1.6.x,
w/C++ & FORTRAN, in production mode
Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN,
w/szip filter, w/threadsafe, in production mode
Linux/PPC 2.6 (heiwa) w/C++ & FORTRAN, w/threadsafe, in debug mode
Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN,
in production mode
Linux/64-amd64 2.6 (abe) w/parallel, w/FORTRAN, in debug mode
Mac OS X/32 10.6.4 (amazon) in debug mode
Mac OS X/32 10.6.4 (amazon) w/C++ & FORTRAN, w/threadsafe,
in production mode
Mac OS X/32 10.6.4 (amazon) w/parallel, in debug mode
Description:
The fractal heap header structure keeps a pointer to the file associated with
it. However, it is possible for that file pointer to be closed while the
header is still in cache (through the shared file pointer). Previously, the
header's file pointer was not updated and subsequently pointed to an invalid
file structure. Modified fractal heap code to update the file pointer every
time the header is accessed.
Tested: jam, linew, amani (h5committest)
Correct an error I convinced Mike to introduce with the phread_once()
code, :-( along with another one that wasn't my fault. Also, clean up
warnings and restructure code in the thread-safe code a bit.
Tested on:
Mac OS X/32 10.6.4 (amazon) w/production + thread-safe
(not a configuration that h5committest tests)
Add windows threads support to HDF5.
Description:
Added calls to the windows threads library to the H5TS layer, and wrapped
most calls to either pthreads or windows threads library with portable
H5TS-style defines. Modified tests to use portable function definitions
as well.
This can be configured via CMake with the HDF5_ENABLE_THREADSAFE
option, and should work on windows vista and later operating systems.
Tested:
h5committest, plus threadsafe with pthreads on jam and amani,
and tested on a Windows Vista VM with threadsafe using windows threads.
Add some extra range checking for invalid chunk sizes.
Tweak default chunk sizes to be zero, so that it's more likely that
application developers will be forced to set them correctly.
Add a few tests to verify these checks.
Tested on:
Mac OS X/32 10.6.4 (amazon) w/debug & production
(too minor to require h5committest)
Bring recent tweaks to extensible and fixed array file formats from
revise_chunks branch back to trunk.
Tested on:
FreeBSD/32 6.3 (duty) in debug mode
FreeBSD/64 6.3 (liberty) w/C++ & FORTRAN, in debug mode
Linux/32 2.6 (jam) w/PGI compilers, w/default API=1.8.x,
w/C++ & FORTRAN, w/threadsafe, in debug mode
Linux/64-amd64 2.6 (amani) w/Intel compilers, w/default API=1.6.x,
w/C++ & FORTRAN, in production mode
Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN,
w/szip filter, w/threadsafe, in production mode
Linux/PPC 2.6 (heiwa) w/C++ & FORTRAN, w/threadsafe, in debug mode
Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN,
in production mode
Linux/64-amd64 2.6 (abe) w/parallel, w/FORTRAN, in debug mode
Mac OS X/32 10.6.4 (amazon) in debug mode
Mac OS X/32 10.6.4 (amazon) w/C++ & FORTRAN, w/threadsafe,
in production mode
Mac OS X/32 10.6.4 (amazon) w/parallel, in debug mode
Clean up a few compiler warnings and style issues in/near recent changes.
Tested on:
Mac OS X/32 10.6.4 (amazon) w/debug & production
(too minor to require h5committest)
Bring revisions from Coverity branch back to trunk:
r19191:
Fix coverity items 104 and 105. Added calls to H5HF_sect_indirect_free to
H5HF_sect_indirect_reduce and H5HF_sect_indirect_reduce_row if there is an errorbefore "peer_sect" is linked into the main free space structure via its direct
sections. Also delayed call to H5HF_sect_indirect_first to prevent peer_sect
from being left in an inconsistent state.
r19268:
Added fix to disallow extendible compact dataset.
This was the same check as in H5D_contig_construct() in H5Dcontig.c.
Added test to verify the creation of extendible dataset with various layouts.
Tested on:
Mac OS X/32 10.6.4 (amazon) w/debug & production
(h5committested on branch)
Whack a few more memory errors exposed by valgrind.
Make the debugging dump output a little bit prettier.
Tested on:
Mac OS X/32 10.6.4 (amazon) w/debug
(h5committest forthcoming)
Merge r19290 & r19291 from 1.8 branch to trunk:
r19290:
Correct another error in metadata accumulator dirty region calculations
(this time with a corner case when freeing data in the file).
r19291:
Avoid getting object information for soft/external links if we aren't
going to traverse across the link itself.
Tested on:
FreeBSD/32 6.3 (duty) in debug mode
FreeBSD/64 6.3 (liberty) w/C++ & FORTRAN, in debug mode
Linux/32 2.6 (jam) w/PGI compilers, w/default API=1.8.x,
w/C++ & FORTRAN, w/threadsafe, in debug mode
Linux/64-amd64 2.6 (amani) w/Intel compilers, w/default API=1.6.x,
w/C++ & FORTRAN, in production mode
Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN,
w/szip filter, w/threadsafe, in production mode
Linux/PPC 2.6 (heiwa) w/C++ & FORTRAN, w/threadsafe, in debug mode
Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN,
in production mode
Linux/64-amd64 2.6 (abe) w/parallel, w/FORTRAN, in debug mode
Mac OS X/32 10.6.4 (amazon) in debug mode
Mac OS X/32 10.6.4 (amazon) w/C++ & FORTRAN, w/threadsafe,
in production mode
Mac OS X/32 10.6.4 (amazon) w/parallel, in debug mode
H5T_ORDER_MIXED specifically for compound types and its derived types. They report mixed
orders if they have members of different orders.
There is no change to configure.in, config, and Makefile.am in the top directory. They have
some property changes when I merged the code from 1.8.
Tested on jam. I tested the same change for 1.8 with h5committest.
More fixes to memory allocation, etc. exposed by valgrind.
Tested on:
FreeBSD/32 6.3 (duty) in debug mode
FreeBSD/64 6.3 (liberty) w/C++ & FORTRAN, in debug mode
Linux/32 2.6 (jam) w/PGI compilers, w/default API=1.8.x,
w/C++ & FORTRAN, w/threadsafe, in debug mode
Linux/64-amd64 2.6 (amani) w/Intel compilers, w/default API=1.6.x,
w/C++ & FORTRAN, in production mode
Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN,
w/szip filter, w/threadsafe, in production mode
Linux/PPC 2.6 (heiwa) w/C++ & FORTRAN, w/threadsafe, in debug mode
Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN,
in production mode
Linux/64-amd64 2.6 (abe) w/parallel, w/FORTRAN, in debug mode
Mac OS X/32 10.6.4 (amazon) in debug mode
Mac OS X/32 10.6.4 (amazon) w/C++ & FORTRAN, w/threadsafe,
in production mode
Mac OS X/32 10.6.4 (amazon) w/parallel, in debug mode
Correct calculation of end of dirty information in metadata accumulator
Tested on:
FreeBSD/32 6.3 (duty) in debug mode
FreeBSD/64 6.3 (liberty) w/C++ & FORTRAN, in debug mode
Linux/32 2.6 (jam) w/PGI compilers, w/default API=1.8.x,
w/C++ & FORTRAN, w/threadsafe, in debug mode
Linux/64-amd64 2.6 (amani) w/Intel compilers, w/default API=1.6.x,
w/C++ & FORTRAN, in production mode
Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN,
w/szip filter, w/threadsafe, in production mode
Linux/PPC 2.6 (heiwa) w/C++ & FORTRAN, w/threadsafe, in debug mode
Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN,
in production mode
Linux/64-amd64 2.6 (abe) w/parallel, w/FORTRAN, in debug mode
Mac OS X/32 10.6.4 (amazon) in debug mode
Mac OS X/32 10.6.4 (amazon) w/C++ & FORTRAN, w/threadsafe,
in production mode
Mac OS X/32 10.6.4 (amazon) w/parallel, in debug mode
Close out various resource leaks and usages of uninitialized memory that
were flagged by valgrind. [There's still some more valgrind warnings, but
it's better now... :-/ ]
Also clean up warnings and code formatting.
Tested on:
Mac OS X/32 10.6.4 (amazon) w/debug & valgrind
(h5committest forthcoming)
