| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
Verified with valgrind -v --tool=memcheck --leak-check=full h5dump POV-GH-2603
The several invalid reads shown originally are now gone.
|
| |
|
|
|
|
| |
(#2679) (#2730)
|
|
|
|
| |
(#2727)
|
|
|
|
|
|
|
|
|
| |
Malformed hdf5 files may have trunkated content which does not match
the expected size. When this function attempts to decode these it may
read past the end of the allocated space leading to heap overflows
as bounds checking is incomplete.
Make sure each element is within bounds before reading.
This fixes CVE-2019-8396 / HDFFV-10712 / github bug #2209.
|
|
|
|
|
|
|
|
|
|
|
|
| |
As indicated in the description, memory leak is detected when running "./h5dump pov".
The problem is: when calling H5O__add_cont_msg() from H5O__chunk_deserialize(),
memory is allocated for cont_msg_info->msgs. Eventually, when the library tries to load
the continuation message via H5AC_protect() in H5O_protect(), error is
encountered due to illegal info in the continuation message.
Due to the error, H5O_protect() exits but the memory allocated for cont_msg_info->msgs is not freed.
When we figure out how to handle fuzzed files that we didn't generate,
a test needs to be added to run h5dump with the provided "pov" file.
|
|
|
|
|
|
| |
There is no need to maintain the release schedule in each branch, so
it's being removed from all branches aside from develop.
Also removes the badges that don't pertain to 1.12
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Convert asserts to error handling in efl decode
The function that decodes external data files object header messages
would call assert() when parsing malformed files, causing applications
to crash when linked against the debug library.
This change converts these assert() calls to HDF5 error checks, so
the messages are sanity checked in both release and debug mode and
debug mode no longer crashes applications.
Also cleaned up some error handling usage and debug checks.
* Free memory on H5O efl decode errors
* Add buffer size checks to efl msg decode
* Add parentheses to math expressions
Fixes GitHub #2605
|
|
|
|
| |
(#2632) (#2668)
|
| |
|
|
|
|
|
|
|
|
| |
* Update HDF5 version after 1.10.10 release.
Update so numbers to match 1.10.10 release so numbers.
Update HISTORY-1_10.txt.
Clean RELEASE.txt entries.
* Missed one version change.
|
| |
|
|
|
|
|
|
|
|
|
| |
* Revert the removal of HDF5GroupInfo class and deprecate.
* revert H5_QUARTER_HADDR_MAX removal
---------
Co-authored-by: github-actions <41898282+github-actions[bot]@users.noreply.github.com>
|
|
|
|
|
|
| |
* Revert "1 10 revert 2615 (#2629)"
This reverts commit 43e4e64d886e9072a6075c6369e84c0e273fa44f.
|
|
|
| |
Co-authored-by: github-actions <41898282+github-actions[bot]@users.noreply.github.com>
|
| |
|
|
|
|
|
| |
* Merge CMake cleanup process from develop
* Update test script files
|
|
|
| |
Unused in the library
|
|
|
|
|
|
|
|
| |
The small perf tool conflicts with the standard perf tool and is being
removed from all HDF5 releases.
Building h5perf in "standalone" mode (i.e., independently of the
library) has been broken for some time, so the build code for
that has also been removed.
|
|
|
|
|
|
|
| |
* Add Option to not build gif tools and remove javadoc option
* Remove old option
* Fix CMake command format
|
|
|
|
|
|
|
|
|
| |
* Add missing items for hpc/scripts and release versions of extra
* Update examples project version
* Add VS2022 options to build scripts
* And VS2022 added to windefs comment
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Add Developer build mode to CMake (#1659)
* Add Developer build mode to CMake
* Set a few CMake variables for Developer build modes
* Refactor enabling of debug and developer-level compile definitions
* Convert cache debugging macros to normal ifdef style
Normal ifdef-style instead of if-style allows build system to define macros
without warning about redefining macros with different values (0 vs. 1)
* Add HDF5 Developer compile definitions to testing files
* Temporarily disable -fanalyzer flag for GCC 12+ Developer builds
* Fix Java tests for Developer build modes (#2079)
* Minor adjustment of Developer build mode changes for 1.10 branch
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Normalize platform-independence code w/ develop
* Use C99 types and functions in lieu of C89 work-arounds
* Align key files with develop
- H5public.h
- H5private.h
- H5system.c
- H5win32defs.h
* Minor fixes elsewhere to support changes in above files
* Incidentally brings Fortran mod directory settings file change over
This does NOT change the configure/build files. Those will still do the
checks needed for the C89 work-around cruft until the perf and perform
code gets cleaned up.
* Add C++98 fixes
* Explicitly set -std=c++98 in Autotools
* Do not include cstdlib in H5public.h (requires C++11)
* Remove redundant stdbool.h include
* Fix alarm issues on Windows
* Bring parallel alarm() changes from develop
|
|
|
|
| |
Add HDF5_SRC_INCLUDE_DIRS variable
cleanup cmake comments and documentation
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
| |
As in develop
|
|
|
| |
As in develop
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
| |
This has not been used to debug the library in a very long time. Most
developers use valgrind, -fsanitize=address, or some other memory checker
instead of this library.
This removes:
* dmalloc.h include from H5private.h
* --with-dmalloc= Autotools configure option
* HDF5_ENABLE_USING_DMALLOC CMake option
|
| |
|
|
|
|
|
|
| |
Brings many compiler flag changes over from develop. Does NOT bring over
-Werror changes.
Also turns on Fortran silent rules, as in develop.
|
|
|
|
|
| |
* Add README.md and NEWSLETTER.txt files
* Remove HISTORY-1_8.txt
* Rename HISTORY-1_0-1_8_0_rc3.txt to HISTORY-1_0-1_8_0.txt
|
|
|
|
|
|
|
|
| |
* Allow parallel + thread-safe builds
No longer require the "enable unsupported" option to build this
configuration in CMake and the Autotools.
* Add a release note for parallel thread-safety
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Elaborate how cd_values get stored (#2522)
* Enclose MESG in do...while loop (#2576)
Enclose MSG macro in a do...while loop
* Add a clang-format comment about permissions (#2577)
* Check for overflow when calculating on-disk attribute data size (#2459)
* Remove duplicate code
Signed-off-by: Egbert Eich <eich@suse.com>
* Add test case for CVE-2021-37501
Bogus sizes in this test case causes the on-disk data size
calculation in H5O__attr_decode() to overflow so that the
calculated size becomes 0. This causes the read to overflow
and h5dump to segfault.
This test case was crafted, the test file was not directly
generated by HDF5.
Test case from:
https://github.com/ST4RF4LL/Something_Found/blob/main/HDF5_v1.13.0_h5dump_heap_overflow.md
---------
Co-authored-by: Mark (he/his) C. Miller <miller86@llnl.gov>
Co-authored-by: glennsong09 <43005495+glennsong09@users.noreply.github.com>
Co-authored-by: Dana Robinson <43805+derobins@users.noreply.github.com>
Co-authored-by: Egbert Eich <eich@suse.com>
|
|
|
|
|
|
| |
* Comment & whitespace differences
* Minor tweak to the "find szip" code
* Yank a STDC_HEADERS block
* Remove ST from the list of debug packages
|
|
|
| |
PR #1925 from develop
|
| |
|
| |
|