Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Fixed HDFFV-10480 (CVE-2018-11206) and HDFFV-11159 (CVE-2018-14033) #405

Merged
merged 22 commits into from
Mar 19, 2021
Merged

Fixed HDFFV-10480 (CVE-2018-11206) and HDFFV-11159 (CVE-2018-14033) #405

merged 22 commits into from
Mar 19, 2021

Conversation

bmribler
Copy link
Contributor

@bmribler bmribler commented Mar 1, 2021

Description
Checked against buffer size to prevent segfault, in case of data corruption.

+ HDFFV-11159 CVE-2018-14033 (same issue as CVE-2020-10811) Buffer over-read in H5O_layout_decode
+ HDFFV-10480 CVE-2018-11206 Buffer over-read in H5O_fill_new[/old]_decode

Platforms tested:
Linux/64 (jelly)

Description
    Checked against buffer size to prevent segfault, in case of data corruption.

    + HDFFV-11159 CVE-2018-14033 Buffer over-read in H5O_layout_decode
    + HDFFV-10480 CVE-2018-11206 Buffer over-read in H5O_fill_new[/old]_decode
Platforms tested:
    Linux/64 (jelly)
   after a more correct fix was applied, that is the check now accounted
   for the previous advance of the buffer pointer.  Removed it.
@bmribler bmribler requested a review from byrnHDF as a code owner March 15, 2021 20:36
@bmribler bmribler requested a review from lrknox as a code owner March 16, 2021 05:23
@@ -1172,6 +1174,10 @@
# test to verify HDFFV-9407: long double full precision
ADD_H5_GREP_TEST (t128bit_float 1 "1.123456789012345" -m %.35Lf t128bit_float.h5)

# test to verify HDFFV-10480: out of bounds read in H5O_fill_new[old]_decode
ADD_H5_TEST (tCVE_2018_11206_fill_old 1 tCVE_2018_11206_fill_old.h5)
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I wonder if ADD_H5_TEST should be ADD_H5_GREP_TEST and search that the test failure that is expected and doesn't fail for some other reason.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'm not sure, but TOOLTEST_FAIL took care of that, it would detect the segfault. I wish the author would put some comments to these macros... :-/

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Oh it segfaults - okay then this is fine.

@bmribler bmribler requested a review from bljhdf as a code owner March 16, 2021 17:21
# Segfault occurred
if [ $RET == 139 ] ; then
nerrors="`expr $nerrors + 1`"
echo " FAILED"
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think it might be worthwhile to differentiate the two "FAILED" messages to make the log more meaningful and also to make clear in the second case that an expected h5dump failure didn't happen. Also note that all other FAILED messages in the script are surrounded by *s: FAILED. If you follow that convention it makes it easier to distinguish failed tests in the logs from incidental occurrences of "FAILED" that don't indicate test failures.

I suggest:
line 891 "FAILED - [h5dump] test failed with segmentation fault"
line 895 "FAILED - [h5dump] test did not fail as expected"

[ ] - optional - could also be test name if available.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Ah, thank you! I knew about '' but must have used the existing PASSED line and changed to FAIL and forgot about the ''.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I've made the changes and resolved the conflict.

@lrknox lrknox merged commit dafc728 into HDFGroup:develop Mar 19, 2021
lrknox added a commit that referenced this pull request Apr 1, 2021
* Snapshot version 1.12 release 1-3.  Update  version to 1.12.1-4.

* First cut of the H5 public API documentation. (#80)

* First cut of the H5 public API documentation.

* Added H5Z "bonus track."

* Applied Quincey's patch.

* Added the missing patches from Quincey's original patch.

* H5PL (complete) and basic H5VL API documentation.

* Added H5I API docs.

* Added H5L API docs.

* First installment from Elena's H5T batch.

* Second installment of Elena's H5T batch.

* Final installment of Elena's H5T batch.

* Full set of current H5F documentation. (#105)

* First cut of the H5 public API documentation.

* Added H5Z "bonus track."

* Applied Quincey's patch.

* Added the missing patches from Quincey's original patch.

* H5PL (complete) and basic H5VL API documentation.

* Added H5I API docs.

* Added H5L API docs.

* First installment from Elena's H5T batch.

* Second installment of Elena's H5T batch.

* Final installment of Elena's H5T batch.

* Migrated documentation for SWMR functions.

* Catching up on MDC functions.

* Integrated the H5F MDC function documentation.

* Added MDC and parallel H5F functions.

* Slightly updated main page.

* Added doxygen/dox/H5AC_cache_config_t.dox to MANIFEST.

* Doxygen - added (mostly) beginner functions (#112)

* Doxygen - added (mostly) beginner functions

* Removed duplicate H5Pset_szip function

* Add src/H5module.h to MANIFEST.

* close #195. (#196)

* Update HDF5PluginMacros.cmake

* Update HDF5PluginMacros.cmake

* Avoid aligned access for references by decoding into temporary buffer and then copying the result into the actual buffer.   Update test to be more thorough with using compound datatype fields everywhere. (#206)

* Modify temporary rpath for testing in java example scripts. (#230)

* Fix undefined left shifting of negative numbers (#338)

Undefined Bahavior Sanitizer errored here about left shifting negative numbers.

* Fixes various warnings noticed on Windows (#425)

* Fixes various warnings noticed on Windows

- Adds a prototype for our implementation of vasprintf
- Return type of H5_get_utf16_str() is now non-const
- Fixes possible uninitialized return type in Wremove_utf8
- Better isolation of fork() code in accum.c:test_swmr_write_big()
- Better isolation of non-zlib code in dsets.c:test_filter_delete()
- Removed unused variable in trefer.c:test_reference_cmpnd_obj()

* Fixes clang-format issues

* Applied clang-tidy readability-non-const-parameter warning fixes auto… (#429)

* Automatically applied clang-tidy readability-avoid-const-params-in-decls fixes

Removes useless const declarations.

* Fixed most readability-non-const-parameter warnings

These changes were made automatically by clang-tidy, but I manually reverted the changes related to the H5Z_func_t signature.

* Reformat source with clang v10.0.1.

Co-authored-by: Larry Knox <lrknox@hdfgroup.org>

* Added C++11 override keyword where appropriate (#433)

Added H5_OVERRIDE macro for compatibility with both C++11 and older.

* Various clang tidy warning fixes (#448)

* Fixed clang-tidy bugprone-reserved-identifier warnings

* Fixed clang-tidy bugprone-assert-side-effect warnings

* Fixed clang-tidy bugprone-copy-constructor-init warning

* Fixed clang-tidy readability-redundant-preprocessor warning

For error_test.c the removed code was already dead, because it was in the else of an `#if H5_USE_16_API` block.

Based on H5Location.h, I think p_get_ref_obj_type was meant to be in `#ifndef DOXYGEN_SHOULD_SKIP_THIS` and an `#endif` was missing.  Similarly, in the header, getObjTypeByIdx is only in H5_NO_DEPRECATED_SYMBOLS, not DOXYGEN_SHOULD_SKIP_THIS.

* Fixed clang-tidy readability-redundant-string-init warnings

* Fixed some clang-tidy performance-type-promotion-in-math-fn warnings

* Fixed clang-tidy performance-unnecessary-value-param warnings

* Reformat source with clang v10.0.1.

Co-authored-by: Larry Knox <lrknox@hdfgroup.org>

* Removed checks/workarounds for pre-C++89 compatibility (#449)

After 30+ years, just assume that the following exist:
- extension-less includes
- namespaces
- std::
- static_cast
- bool

* Fixed all clang-tidy bugprone-suspicious-string-compare warnings (#451)

* Fixed all clang-tidy bugprone-suspicious-string-compare warnings

This change was generated entirely by clang-tidy itself.

* Reformat code with clang v10.0.1.

Co-authored-by: Larry Knox <lrknox@hdfgroup.org>

* Remove 2 functions incorrectly merged from develop in a cherry-pick merge of PR #451.

* Purge the buffer used in type conversion. (#263)

Some of the uniniitialized bits in the buffer may get carried through
all the way to disk, creating a risk for leaks.

We observed an msan error during the floating point output conversion.
Due to the encoding certain bits could remain untouched during the conversion.

In this draft we zero initialize the dbuf used by every convertor.

* Fixed HDFFV-10480 (CVE-2018-11206) and HDFFV-11159 (CVE-2018-14033) (#405)

* Fixed HDFFV-10480 (CVE-2018-11206) and HDFFV-11159 (CVE-2018-14033)
Description
    Checked against buffer size to prevent segfault, in case of data corruption.

    + HDFFV-11159 CVE-2018-14033 Buffer over-read in H5O_layout_decode
    + HDFFV-10480 CVE-2018-11206 Buffer over-read in H5O_fill_new[/old]_decode
Platforms tested:
    Linux/64 (jelly)

* Accidentally left in another occurrence of the previous patch from user
   after a more correct fix was applied, that is the check now accounted
   for the previous advance of the buffer pointer.  Removed it.

* Typo

* Fixed format issues.

* Added test.

* Changed arguments to ADD_H5_TEST

* Fixing arguments to ADD_H5_TEST again.

* Fixing arguments again.

* Took out the CMake changes until Allen can help.

* Added files:

tCVE_2018_11206_fill_old.h5
tCVE_2018_11206_fill_new.h5

* Revert "Took out the CMake changes until Allen can help."

This reverts commit c21324d.

* Revert "Fixing arguments again."

This reverts commit 5832a70.

* Revert "Fixing arguments to ADD_H5_TEST again."

This reverts commit b45de82.

* Revert "Changed arguments to ADD_H5_TEST"

This reverts commit 1671982.

* Added first argument to ADD_H5_TEST for HDFFV-10480 fix.

* Changed argument 0 to 1

* Revert "Changed argument 0 to 1"

This reverts commit b343d66.

* Revert "Added first argument to ADD_H5_TEST for HDFFV-10480 fix."

This reverts commit b8a0f9a.

* Added first argument and corrected the second.

* Updated fixes for HDFFV-10480 and HDFFV-11159/HDFFV-11049

* Improved error messages.

* Added description of the current HDF5 branches; added a draft of contribution policy. (#445)

* Added description of the current HDF5 branches.

* Removed capitalization in from Develop, Release, Feature to reflect the real naming schema

* Added a draft of contributions guidance document.

* Fixed typos.

* Fixed section title.

* Fixed typo.

* Fixed typos and formatting.

* Fixed many typos and simplified the text (e.g., removed testing instructions and left pointers to where to find them, etc.)

* Fixed a typo.
 Please enter the commit message for your changes. Lines starting

* Added contributing.md file and rearranged doc entries in alphabetical order.

* ddressed Gerd's review comments; found and fixed more typos.

* Addressed comments from Larry and Scot.

* Fix CMake error message location. (#478)

Print error message if Perl is not found.

* Committing clang-format changes

Co-authored-by: Gerd Heber <gheber@hdfgroup.org>
Co-authored-by: bljhdf <58825073+bljhdf@users.noreply.github.com>
Co-authored-by: H. Joe Lee <hyoklee@hdfgroup.org>
Co-authored-by: Quincey Koziol <quincey@koziol.cc>
Co-authored-by: Sean McBride <sean@rogue-research.com>
Co-authored-by: Dana Robinson <43805+derobins@users.noreply.github.com>
Co-authored-by: Yu Feng <rainwoodman@gmail.com>
Co-authored-by: bmribler <39579120+bmribler@users.noreply.github.com>
Co-authored-by: epourmal <epourmal@hdfgroup.org>
Co-authored-by: github-actions <41898282+github-actions[bot]@users.noreply.github.com>
lrknox added a commit that referenced this pull request Apr 2, 2021
* close #195. (#196)

* Update HDF5PluginMacros.cmake

* Update HDF5PluginMacros.cmake

* Modify temporary rpath for testing in java example scripts. (#230)

* Fix undefined left shifting of negative numbers (#338)

Undefined Bahavior Sanitizer errored here about left shifting negative numbers.

* Update license url (#332)

* Modify temporary rpath for testing in java example scripts.

* Update URL in source file Copyright headers for web copy of COPYING
file - src and test directories.

* Cleans up a couple of MSVC warnings in testhdf5 (#475)
* Fixes a few testhdf5 warnings raised in Visual Studio

Visual Studio is grumpier about treating pointers like integers than
gcc.

* Committing clang-format changes

Co-authored-by: github-actions <41898282+github-actions[bot]@users.noreply.github.com>

* Fix typos and grammar errors. (#476)

* Suppresses the tcheck_version test's abort dialog on Windows (#477)

* Suppresses the tcheck_version test's abort dialog on Windows

Windows raises a modal abort/retry/ignore dialog box when CRT
calls abort(). This change installs a report hook that suppresses
the dialog so that the CMake tests don't time out waiting for a
nonexistent user to click a dialog box.

* Committing clang-format changes

* Removes __cdecl from callback

Co-authored-by: github-actions <41898282+github-actions[bot]@users.noreply.github.com>

* Purge the buffer used in type conversion. (#263)

Some of the uniniitialized bits in the buffer may get carried through
all the way to disk, creating a risk for leaks.

We observed an msan error during the floating point output conversion.
Due to the encoding certain bits could remain untouched during the conversion.

In this draft we zero initialize the dbuf used by every convertor.

* Fixed HDFFV-10480 (CVE-2018-11206) and HDFFV-11159 (CVE-2018-14033) (#405)

* Fixed HDFFV-10480 (CVE-2018-11206) and HDFFV-11159 (CVE-2018-14033)
Description
    Checked against buffer size to prevent segfault, in case of data corruption.

    + HDFFV-11159 CVE-2018-14033 Buffer over-read in H5O_layout_decode
    + HDFFV-10480 CVE-2018-11206 Buffer over-read in H5O_fill_new[/old]_decode
Platforms tested:
    Linux/64 (jelly)

* Accidentally left in another occurrence of the previous patch from user
   after a more correct fix was applied, that is the check now accounted
   for the previous advance of the buffer pointer.  Removed it.

* Typo

* Fixed format issues.

* Added test.

* Changed arguments to ADD_H5_TEST

* Fixing arguments to ADD_H5_TEST again.

* Fixing arguments again.

* Took out the CMake changes until Allen can help.

* Added files:

tCVE_2018_11206_fill_old.h5
tCVE_2018_11206_fill_new.h5

* Revert "Took out the CMake changes until Allen can help."

This reverts commit c21324d.

* Revert "Fixing arguments again."

This reverts commit 5832a70.

* Revert "Fixing arguments to ADD_H5_TEST again."

This reverts commit b45de82.

* Revert "Changed arguments to ADD_H5_TEST"

This reverts commit 1671982.

* Added first argument to ADD_H5_TEST for HDFFV-10480 fix.

* Changed argument 0 to 1

* Revert "Changed argument 0 to 1"

This reverts commit b343d66.

* Revert "Added first argument to ADD_H5_TEST for HDFFV-10480 fix."

This reverts commit b8a0f9a.

* Added first argument and corrected the second.

* Updated fixes for HDFFV-10480 and HDFFV-11159/HDFFV-11049

* Improved error messages.

* Added description of the current HDF5 branches; added a draft of contribution policy. (#445)

* Added description of the current HDF5 branches.

* Removed capitalization in from Develop, Release, Feature to reflect the real naming schema

* Added a draft of contributions guidance document.

* Fixed typos.

* Fixed section title.

* Fixed typo.

* Fixed typos and formatting.

* Fixed many typos and simplified the text (e.g., removed testing instructions and left pointers to where to find them, etc.)

* Fixed a typo.
 Please enter the commit message for your changes. Lines starting

* Added contributing.md file and rearranged doc entries in alphabetical order.

* ddressed Gerd's review comments; found and fixed more typos.

* Addressed comments from Larry and Scot.

* Fix CMake error message location. (#478)

Print error message if Perl is not found.

* Committing clang-format changes

* Update MANIFEST.

Co-authored-by: H. Joe Lee <hyoklee@hdfgroup.org>
Co-authored-by: Sean McBride <sean@rogue-research.com>
Co-authored-by: Dana Robinson <43805+derobins@users.noreply.github.com>
Co-authored-by: github-actions <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: Yu Feng <rainwoodman@gmail.com>
Co-authored-by: bmribler <39579120+bmribler@users.noreply.github.com>
Co-authored-by: epourmal <epourmal@hdfgroup.org>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

5 participants