* wip
* Merged PR 28837: Fix CICD
Fix CICD build failures.
* wip
* wip
* Revert "wip"
This reverts commit e7311fef62.
* sync
* sync
* sync
* feedback
* draft
* sync
* sync
* draft
* wip
* nit
* wip
* wip
* wip
* wip
* wip
* wip
* nit
* add AKS evn detection, add TBD check eBPF usage on uninstall
* wip
* Merged PR 30605: Bump bpf_conformance to 0.6.0 to unblock conformance testing
Bump bpf_conformance to 0.6.0 to unblock conformance testing
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Merged PR 30598: Add new onefuzz task to perform continual fuzzing of eBPF for Windows
Add new onefuzz task to perform continual fuzzing of eBPF for Windows
* Merged PR 30633: Add missing bug filing info for netebpfext
Add missing bug filing info for netebpfext
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Merged PR 30634: Add codecoverage for fuzzing pipeline
Add codecoverage for fuzzing pipeline
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* wip
* Merged PR 30808: Apply correct fuzzer settings
Apply correct fuzzer settings
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Merged PR 30804: Fix division by zero in batch delete.
Fix division by zero in batch delete:
`input_count = key_length / map_definition->key_size;`
Related work items: #159139
* Merged PR 30814: Fix OneFuzzConfig.json to fuzz feature/security_fix and correctly assign bugs
Fix OneFuzzConfig.json to fuzz feature/security_fix and correctly assign bugs
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Merged PR 30825: Reformat IOCTL fuzzing seed to include reply length
Reformat IOCTL fuzzing seed to include reply length and regenerate corpus.
Issue:
The fuzzer was always setting an invalid reply length, even for "good" cases. This results in many paths being blocked because reply length != expected reply length in ebpf_core_invoke_protocol_handler.
By including the correct reply length in the seed, we permit the fuzzer to explore more code spaces.
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Merged PR 30880: Add ASAN_OPTIONS=allocator_may_return_null=1
Add ASAN_OPTIONS=allocator_may_return_null=1
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Merged PR 30890: Run all libfuzzers in OneFuzz on the feature/security_fix branch
Run all libfuzzers in OneFuzz on the feature/security_fix branch
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Merged PR 30897: _ebpf_program_load_byte_code should reject zero length eBPF programs
_ebpf_program_load_byte_code should reject zero length eBPF programs
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Merged PR 30922: Improve verifier coverage by permitting more program types to run.
Improve verifier coverage by permitting more program types to run.
Due to context create for sample program type, all calls to prog run for a sample program fail with minimal coverage.
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Merged PR 30905: Prevent fuzzer from triggering long running bpf_prog_test_run
The execution_context_fuzzer can get hung on executing a BPF program if it fuzzes the iteration count to a huge value, resulting in the test timing out after several minutes. This slows the rate of useful fuzzing.
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Fix buffer overflow in ebpf_program_get_info
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Workaround for verifier bug
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Fix division by zero in _ebpf_core_protocol_map_update_element_batch
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Merged PR 30956: Improve code coverage of execution context fuzzer
During creation of the various map types, creation of several map types was failing, preventing fuzzing of those maps.
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Merged PR 30819: Check LPM key length
Check LPM key length
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
Related work items: #159141
* Merged PR 30971: Remove PATs from onefuzz task.
As per guidance from the OneFuzz team, a PAT is no longer required, so switch to the new option.
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Merged PR 30991: Prevent ebpf_program_t from switching between JIT and interpreter
Prevent ebpf_program_t from switching between JIT and interpreter.
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Merged PR 30992: Improve execution context fuzzer - add link object and speed up fuzzing
Improve execution context fuzzer - add link object and speed up fuzzing
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Merged PR 30993: Make helper fuzzer more generic
1. Remove dependency on ebpf_program_t.
2. Bind directly to program info provider.
3. Move ebpf_core_initiate/terminate to global.
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Merged PR 31238: Fix assigned to in .onefuzz\OneFuzzConfig.json
Fix assigned to in .onefuzz\OneFuzzConfig.json
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Merged PR 31250: Merge main into feature/security_fix
* Merged PR 31358: Add HLK playlist link to the HLK documentation.
Add playlist link, whitespace optimizations.
* Merged PR 31516: Disable onefuzz on main
Disable onefuzz on main
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Merged PR 30823: Quick-guide: how to debug an eBPF program
This is a quick guide on how to use WinDbg to debug kernel code on a target machine.
* add checks for code_type
* Merged PR 31674: Update netebpfext bind context app_id truncation logic for safety
_net_ebpf_ext_resource_truncate_appid made some assumptions about the input data it received. In theory, improper data could be passed into ebpf, which could potentially cause a buffer overrun/crash. Additional checks were added to ensure safety of the truncation logic, even if improper input was passed.
Additional test cases were added to ensure the truncation logic was safe and validating improper input.
Related work items: #159406
* Merged PR 30986: Add LPM map unit tests, fix lookup and input checking bugs
## Description
This improves and adds ebpf map unit testing and fixes some bugs.
Changes:
- `_create_lpm_map`: fixes length calculation for key size bitmap
- `max_prefix` should be the passed value, bitmap size gets `+1` to account for max length keys
- `_find_lpm_map_entry`: initialize search from search key bit length instead of maximum stored key length
- reduces number of hashes to try during lookup (for short keys) and fixes bug of looking for map key longer than passed search key
- adds `ebpf_bitmap_start_reverse_search_at` to initialize set bit search to a specific bit index
- `_create_array_map_with_map_struct_size`: argument check for value size/capacity combination that allocates >128GB
- `_next_lpm_map_key_and_value`/`_next_hash_map_key_and_value` - fix SAL annotation for `next_key` (should be `_Inout_`)
- Refactor lpm key handling to use new `ebpf_core_lpm_key_t` struct instead of pointer to prefix length
- Add, fix, and extend LPM map unit tests
## Testing
Extends the `map_crud_operations_lpm_trie_32` test case and adds a second `[negative]` tagged version which tests operations that shouldn't work (including some edge cases that were previously missed).
`map_create_invalid` test case added to try creating various maps with invalid parameters.
Extends and refactors `map_crud_operations_lpm_trie_128`
## Documentation
Comments in new unit testing functions, and more documentation added in LPM and bitmap code.
## Installation
N/A
Related work items: #159137, #159144, #159270, #159273
* Merged PR 31697: Add map type validation in debug mode
This PR adds map type validation in debug mode.
The map type was already checked on creation and is never user-writeable, so these checks are just against code bugs or outside code modifying the type field.
In ebpf_maps.c lookup into `ebpf_map_metadata_tables` is now factored out into calls to `ebpf_map_get_table()`, which validates the map type with a debug assert.
Related work items: #159405
* Merged PR 31809: Update VM extension script to invoke `export_program_info`
This PR contains changes to the VM extension scripts to invoke `export_program_info.exe`. The invocation is kept optional based on whether the file is present or not. This is for 2 reasons:
1. ebpf v0.11 released earlier to IMDS does not contain this exe. So in case of cleanup or rollback, it is possible that the exe is not available.
2. We should ideally change the store implementation to move from registry to file based. Keeping the new code optional allows to change store implementation without requiring the script to mandatorily updated at the same time.
* Merged PR 32086: Update ADO pipeline
This PR updates the clang version to be used in ADO pipeline. Since the clang version has been updated in GitHub, and correspnding tests have also been updated, CICD is now failing in ADO since then.
#### PR Summary
This pull request updates the ADO pipeline to use Visual Studio LLVM tools instead of installing LLVM via Chocolatey.
- `.pipelines/reusable-build.yml`: Replaced Chocolatey installation of LLVM with Visual Studio LLVM tools installation using PowerShell tasks.
- `.pipelines/reusable-build.yml`: Added MSVC LLVM Clang to PATH and logged the CLANG version.
- `.pipelines/reusable-build.yml`: Set `VSINSTALLDIR` environment variable.
* Merged PR 32085: VM extension script fixes
This PR contains 2 changes:
1. Fix the source path for upgrade flow. The script was incorrectly using a hardcoded path.
2. Suppress error messages from `Get-ItemProperty` when querying for registry path. This change is only cosmetic and has no functional impact.
----
#### AI description (iteration 1)
#### PR Classification
Bug fix for VM extension scripts.
#### PR Summary
This pull request addresses several issues in the VM extension scripts, primarily focusing on error handling and parameter corrections.
- `/.internal/vm-extension/src/scripts/common.ps1`: Added `-ErrorAction SilentlyContinue` to `Get-ItemProperty` to handle missing registry keys gracefully.
- Corrected parameter usage in `Update-eBPF` and `InstallOrUpdate-eBPF` functions to ensure the correct source path is used during installation.
- Removed trailing whitespace in multiple functions for code cleanliness.
* Merged PR 30804: Fix division by zero in batch delete.
Fix division by zero in batch delete:
`input_count = key_length / map_definition->key_size;`
Related work items: #159139
* Merged PR 30814: Fix OneFuzzConfig.json to fuzz feature/security_fix and correctly assign bugs
Fix OneFuzzConfig.json to fuzz feature/security_fix and correctly assign bugs
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Merged PR 30825: Reformat IOCTL fuzzing seed to include reply length
Reformat IOCTL fuzzing seed to include reply length and regenerate corpus.
Issue:
The fuzzer was always setting an invalid reply length, even for "good" cases. This results in many paths being blocked because reply length != expected reply length in ebpf_core_invoke_protocol_handler.
By including the correct reply length in the seed, we permit the fuzzer to explore more code spaces.
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Merged PR 30890: Run all libfuzzers in OneFuzz on the feature/security_fix branch
Run all libfuzzers in OneFuzz on the feature/security_fix branch
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Merged PR 30897: _ebpf_program_load_byte_code should reject zero length eBPF programs
_ebpf_program_load_byte_code should reject zero length eBPF programs
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Merged PR 30922: Improve verifier coverage by permitting more program types to run.
Improve verifier coverage by permitting more program types to run.
Due to context create for sample program type, all calls to prog run for a sample program fail with minimal coverage.
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Merged PR 30905: Prevent fuzzer from triggering long running bpf_prog_test_run
The execution_context_fuzzer can get hung on executing a BPF program if it fuzzes the iteration count to a huge value, resulting in the test timing out after several minutes. This slows the rate of useful fuzzing.
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Fix division by zero in _ebpf_core_protocol_map_update_element_batch
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Fix buffer overflow in ebpf_program_get_info
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Workaround for verifier bug
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Merged PR 30956: Improve code coverage of execution context fuzzer
During creation of the various map types, creation of several map types was failing, preventing fuzzing of those maps.
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Merged PR 30819: Check LPM key length
Check LPM key length
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
Related work items: #159141
* Merged PR 30991: Prevent ebpf_program_t from switching between JIT and interpreter
Prevent ebpf_program_t from switching between JIT and interpreter.
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Merged PR 30992: Improve execution context fuzzer - add link object and speed up fuzzing
Improve execution context fuzzer - add link object and speed up fuzzing
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Merged PR 30993: Make helper fuzzer more generic
1. Remove dependency on ebpf_program_t.
2. Bind directly to program info provider.
3. Move ebpf_core_initiate/terminate to global.
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Merged PR 31250: Merge main into feature/security_fix
* Merged PR 31674: Update netebpfext bind context app_id truncation logic for safety
_net_ebpf_ext_resource_truncate_appid made some assumptions about the input data it received. In theory, improper data could be passed into ebpf, which could potentially cause a buffer overrun/crash. Additional checks were added to ensure safety of the truncation logic, even if improper input was passed.
Additional test cases were added to ensure the truncation logic was safe and validating improper input.
Related work items: #159406
* Merged PR 30986: Add LPM map unit tests, fix lookup and input checking bugs
## Description
This improves and adds ebpf map unit testing and fixes some bugs.
Changes:
- `_create_lpm_map`: fixes length calculation for key size bitmap
- `max_prefix` should be the passed value, bitmap size gets `+1` to account for max length keys
- `_find_lpm_map_entry`: initialize search from search key bit length instead of maximum stored key length
- reduces number of hashes to try during lookup (for short keys) and fixes bug of looking for map key longer than passed search key
- adds `ebpf_bitmap_start_reverse_search_at` to initialize set bit search to a specific bit index
- `_create_array_map_with_map_struct_size`: argument check for value size/capacity combination that allocates >128GB
- `_next_lpm_map_key_and_value`/`_next_hash_map_key_and_value` - fix SAL annotation for `next_key` (should be `_Inout_`)
- Refactor lpm key handling to use new `ebpf_core_lpm_key_t` struct instead of pointer to prefix length
- Add, fix, and extend LPM map unit tests
## Testing
Extends the `map_crud_operations_lpm_trie_32` test case and adds a second `[negative]` tagged version which tests operations that shouldn't work (including some edge cases that were previously missed).
`map_create_invalid` test case added to try creating various maps with invalid parameters.
Extends and refactors `map_crud_operations_lpm_trie_128`
## Documentation
Comments in new unit testing functions, and more documentation added in LPM and bitmap code.
## Installation
N/A
Related work items: #159137, #159144, #159270, #159273
* Merged PR 31697: Add map type validation in debug mode
This PR adds map type validation in debug mode.
The map type was already checked on creation and is never user-writeable, so these checks are just against code bugs or outside code modifying the type field.
In ebpf_maps.c lookup into `ebpf_map_metadata_tables` is now factored out into calls to `ebpf_map_get_table()`, which validates the map type with a debug assert.
Related work items: #159405
* add checks for code_type
* fix bad merge
* Override `bpf_get_current_pid_tgid` for sock_ops hook. (#3765)
* initial changes.
* sock_addr.
* fixes.
* Update include/ebpf_nethooks.h
Co-authored-by: Dave Thaler <dthaler1968@gmail.com>
---------
Co-authored-by: Dave Thaler <dthaler1968@gmail.com>
* Add ADO pipelines (#3741)
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
Co-authored-by: Alan Jowett <alan.jowett@microsoft.com>
* Bump external/ubpf from `19cd22c` to `762a98d` (#3783)
Bumps [external/ubpf](https://github.com/iovisor/ubpf) from `19cd22c` to `762a98d`.
- [Commits](19cd22c6fb...762a98dd1a)
---
updated-dependencies:
- dependency-name: external/ubpf
dependency-type: direct:production
...
Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
* Bump external/usersim from `46c6b4a` to `ca8f2de` (#3782)
Bumps [external/usersim](https://github.com/microsoft/usersim) from `46c6b4a` to `ca8f2de`.
- [Commits](46c6b4adf1...ca8f2defc6)
---
updated-dependencies:
- dependency-name: external/usersim
dependency-type: direct:production
...
Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
* Bump external/ebpf-verifier from `dc2d6ea` to `cd9344b` (#3784)
Bumps [external/ebpf-verifier](https://github.com/vbpf/ebpf-verifier) from `dc2d6ea` to `cd9344b`.
- [Release notes](https://github.com/vbpf/ebpf-verifier/releases)
- [Commits](dc2d6ea75a...cd9344bdbb)
---
updated-dependencies:
- dependency-name: external/ebpf-verifier
dependency-type: direct:production
...
Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
* Bump github/codeql-action from 3.26.2 to 3.26.5 in the actions group (#3790)
Bumps the actions group with 1 update: [github/codeql-action](https://github.com/github/codeql-action).
Updates `github/codeql-action` from 3.26.2 to 3.26.5
- [Release notes](https://github.com/github/codeql-action/releases)
- [Changelog](https://github.com/github/codeql-action/blob/main/CHANGELOG.md)
- [Commits](429e197704...2c779ab0d0)
---
updated-dependencies:
- dependency-name: github/codeql-action
dependency-type: direct:production
update-type: version-update:semver-patch
dependency-group: actions
...
Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
* Stamp all tools with version information. (#3792)
* Stamp all tools with version information.
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Add commit id to nuget package
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Move versioning to Directory.Build.props
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Set correct project type in sample.vcxproj
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Add version info to fuzz tests
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* PR feedback
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Update tests/socket/resource.h
Co-authored-by: Dave Thaler <dthaler1968@gmail.com>
* Revert editor change
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
---------
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
Co-authored-by: Alan Jowett <alan.jowett@microsoft.com>
Co-authored-by: Dave Thaler <dthaler1968@gmail.com>
* First string functions for ebpf general helpers. (#3780)
* Initial progress on function implementations.
* Moved code.
* Updated kernel project.
* Working on parallel user and kernel mode versions.
* Work in Progress.
* Added tests, added a lot.
* Cleaned up test code some, still doesn't build.
* Re-enabled some tests.
* Got a working set of tests, now to figure out what to do with this in the interim.
* Removing as-yet-unimplemented functions for now.
* Fixed a name.
* Doxygen fixes. Also removed a function header that'll be needed another time.
* Got down to one source file, got a build working in user mode.
* Renamed string_opts to ebpf_strings
* Corrected build issues, removed other extraneous comments.
* Fixed the test selection name, and suppressed an analysis error in a Windows header.
---------
Co-authored-by: Ben Lewis (REDMOND) <Ben.Lewis@microsoft.com>
* Bump external/ubpf from `762a98d` to `f1ecb7a` (#3797)
Bumps [external/ubpf](https://github.com/iovisor/ubpf) from `762a98d` to `f1ecb7a`.
- [Commits](762a98dd1a...f1ecb7a245)
---
updated-dependencies:
- dependency-name: external/ubpf
dependency-type: direct:production
...
Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
* Bump external/usersim from `ca8f2de` to `bded12c` (#3796)
Bumps [external/usersim](https://github.com/microsoft/usersim) from `ca8f2de` to `bded12c`.
- [Commits](ca8f2defc6...bded12ceee)
---
updated-dependencies:
- dependency-name: external/usersim
dependency-type: direct:production
...
Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
* Added ntosext extensions in the GettingStarted.md document. (#3785)
* Added ntosext extension in the doc
* Addressed PR comment
* Update docs/GettingStarted.md
Co-authored-by: Dave Thaler <dthaler1968@gmail.com>
---------
Co-authored-by: Dave Thaler <dthaler1968@gmail.com>
* relocate installer license to avoid "multiple licenses" text on GitHub (#3802)
* Inline _ebpf_adjust_value_pointer to reduce cost of ebpf_map_find_entry (#3804)
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
Co-authored-by: Alan Jowett <alan.jowett@microsoft.com>
* Bump external/ebpf-verifier from `cd9344b` to `559482c` (#3807)
Bumps [external/ebpf-verifier](https://github.com/vbpf/ebpf-verifier) from `cd9344b` to `559482c`.
- [Release notes](https://github.com/vbpf/ebpf-verifier/releases)
- [Commits](cd9344bdbb...559482ccaf)
---
updated-dependencies:
- dependency-name: external/ebpf-verifier
dependency-type: direct:production
...
Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
* Bump the actions group with 2 updates (#3806)
Bumps the actions group with 2 updates: [actions/upload-artifact](https://github.com/actions/upload-artifact) and [github/codeql-action](https://github.com/github/codeql-action).
Updates `actions/upload-artifact` from 4.3.6 to 4.4.0
- [Release notes](https://github.com/actions/upload-artifact/releases)
- [Commits](834a144ee9...50769540e7)
Updates `github/codeql-action` from 3.26.5 to 3.26.6
- [Release notes](https://github.com/github/codeql-action/releases)
- [Changelog](https://github.com/github/codeql-action/blob/main/CHANGELOG.md)
- [Commits](2c779ab0d0...4dd16135b6)
---
updated-dependencies:
- dependency-name: actions/upload-artifact
dependency-type: direct:production
update-type: version-update:semver-minor
dependency-group: actions
- dependency-name: github/codeql-action
dependency-type: direct:production
update-type: version-update:semver-patch
dependency-group: actions
...
Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
* Switch to crc32 if available for hashing function (#3803)
* Switch to crc32 if available for hashing function
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Only do crc32 when keys are directly accessible
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Minimize branches and extra ALU ops
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
---------
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
Co-authored-by: Alan Jowett <alan.jowett@microsoft.com>
* Multi-attach support for SOCK_ADDR programs (#3751)
* initial commit
* second commit
* third commit
* update lock logic
* fix analysis errors
* backup
* Add tests
* fix
* fix
* fix
* code refactor
* update test
* Revert "update test"
This reverts commit 6770269907.
* Revert "code refactor"
This reverts commit fae36a7d0e.
* Revert "Revert "code refactor""
This reverts commit 3b8e762d33.
* Revert "Revert "update test""
This reverts commit 11610f06ce.
* fix capability in xdp
* add tests, remove filter weight
* invoke wildcard programs, update tests
* fix wildcard invocation logic
* fix wildcard invocation logic, remove sleep from concurrency tests
* move invocation out of the lock, fix tests
* code cleanup, update test
* code cleanup
* remove trace
* cleanup
* update wildcard invocation logic
* code cleanup
* code cleanup
* code cleanup, CR comments
* cr comments, code cleanup
* Update netebpfext/net_ebpf_ext_hook_provider.h
Co-authored-by: Dave Thaler <dthaler1968@gmail.com>
* cr comments
* add tests
* cr comments
* code cleanup
* tracing changes
* fix sal
* code cleanup
* code cleanup
* Update netebpfext/net_ebpf_ext.h
Co-authored-by: Dave Thaler <dthaler1968@gmail.com>
---------
Co-authored-by: Dave Thaler <dthaler1968@gmail.com>
fix.
* Updated onebranch.vcxproj
* Updated post-build.ps1
* Stack expansion in netebpfext (#3817)
* expand stack
* update usersim submodule
* fix analysis build
* cr comments
* fix analysis failure
* cr comments
* Update netebpfext/net_ebpf_ext_hook_provider.h
Co-authored-by: Dave Thaler <dthaler1968@gmail.com>
---------
Co-authored-by: Dave Thaler <dthaler1968@gmail.com>
* Bump external/usersim from `bded12c` to `a1ba035` (#3818)
Bumps [external/usersim](https://github.com/microsoft/usersim) from `bded12c` to `a1ba035`.
- [Commits](bded12ceee...a1ba0355fb)
---
updated-dependencies:
- dependency-name: external/usersim
dependency-type: direct:production
...
Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
* change expanded stack size to 16KB.
* Merged PR 11480212: Update store APIs to also update HKLM (#3660) (#3853)
Update store APIs to also update HKLM (#3660) (#3853)
Cherry-picked from commit `462b8e3c`.
----
#### AI description (iteration 1)
#### PR Classification
API change to update store APIs to also update HKLM.
#### PR Summary
This pull request updates the store APIs to handle both HKCU and HKLM registry keys, ensuring that operations are attempted on both keys and errors are suppressed if access to HKLM is denied.
- `ebpf_store_helper.c`: Added functions to update and delete program and section information for both HKCU and HKLM.
- `store_helper_internal.cpp`: Modified functions to handle both HKCU and HKLM registry keys.
- `Product.wxs`: Updated installer scripts to clear and set up eBPF store for both HKCU and HKLM.
- `ebpf_store_helper.h` and `ebpf_registry_helper.cpp`: Introduced separate variables for HKCU and HKLM root keys.
* Merged PR 11486801: Update version to 0.19.1
update version to 0.19.1
* remove .internal
* Miscellaneous fixes.
* PR Feedback.
---------
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: Gianni Trevisiol <gtrevi@users.noreply.github.com>
Co-authored-by: Anurag Saxena <anusa@microsoft.com>
Co-authored-by: Microsoft.VisualStudio.Services.TFS <Microsoft.VisualStudio.Services.TFS@microsoft.com>
Co-authored-by: Gianni Trevisiol <gtrevi@microsoft.com>
Co-authored-by: Alan Jowett <alanjo@microsoft.com>
Co-authored-by: Alan Jowett <Alan Jowett@microsoft.com>
Co-authored-by: Igor Klemenski <igklemen@microsoft.com>
Co-authored-by: Shankar Seal <Shankar Seal@microsoft.com>
Co-authored-by: Alan Jowett <alan.jowett@microsoft.com>
Co-authored-by: Matt Ige <maige@microsoft.com>
Co-authored-by: Michael Agun <danielagun@microsoft.com>
Co-authored-by: Dave Thaler <dthaler1968@gmail.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
Co-authored-by: Ben Lewis <1884340+ben-zen@users.noreply.github.com>
Co-authored-by: Ben Lewis (REDMOND) <Ben.Lewis@microsoft.com>
Co-authored-by: Sharmi <shpalan@microsoft.com>
Co-authored-by: Michael Friesen <3517159+mtfriesen@users.noreply.github.com>
Co-authored-by: Anurag Saxena <43585259+saxena-anurag@users.noreply.github.com>
* Build with WDK as nuget
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Update test collateral
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Doc update
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Add old and new sizes
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
---------
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
Co-authored-by: Alan Jowett <alan.jowett@microsoft.com>
* Switch to cxplat_get_current_processor_number
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Fix runtime failure
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
---------
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
Co-authored-by: Alan Jowett <alan.jowett@microsoft.com>
* Stamp all tools with version information.
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Add commit id to nuget package
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Move versioning to Directory.Build.props
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Set correct project type in sample.vcxproj
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Add version info to fuzz tests
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* PR feedback
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Update tests/socket/resource.h
Co-authored-by: Dave Thaler <dthaler1968@gmail.com>
* Revert editor change
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
---------
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
Co-authored-by: Alan Jowett <alan.jowett@microsoft.com>
Co-authored-by: Dave Thaler <dthaler1968@gmail.com>
* Update copyright to be eBPF for Windows contributors
Fixes#3507
Signed-off-by: Dave Thaler <dthaler1968@gmail.com>
* Undo gratuitous formatting changes in expected files
---------
Signed-off-by: Dave Thaler <dthaler1968@gmail.com>
* Adding a call to ebpf_trace_terminate in case of subsequent APIs faliure
Signed-off-by: Amit Schendel <amithagever@gmail.com>
* Adding ebpf_trace_terminate to the Exit section of the DriverEntry
Signed-off-by: Amit Schendel <amitschendel@gmail.com>
---------
Signed-off-by: Amit Schendel <amithagever@gmail.com>
Signed-off-by: Amit Schendel <amitschendel@gmail.com>
* Update to use latest usersim
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Update to latest usersim
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Use NonPagedPoolNx instead of NonPagedPool for netebpfext state
NonPagedPool is executable, which we don't want
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Update usersim
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
---------
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* build ebpfcore under usersim
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Work around analysis warnings
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Fix installer project
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Temporarily use ubpf branch with memory fix
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* WIP
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* WIP
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Create project for files shared between ebpfapi and the ebpf runtime
Move existing files from libs/platform either to:
* libs/shared, if meant to be common between user-mode and ebpf runtime
* libs/runtime, if meant to be used by the ebpf runtime only (whether in
user or kernel mode)
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Move ebpf_tracelog.c to shared
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Create shared_user and shared_kernel projects
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* WIP
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* WIP
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Remove usersim from ebpfsvc
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Update to cxplat branch of usersim
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Use cxplat_utf8_string
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Cleanup
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Fix build errors
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Fix ebpf_allocate_with_tag
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Fixes
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Bug fixes
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Fix use of ebpf_reallocate to not pass a null pointer
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Fix printf tests
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Fix memory leaks
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Temporary change to help debug double-free in test
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Update to latest usersim
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Fix analysis warnings
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Update to latest usersim
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Update usersim to latest
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Change native module to use latest usersim
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Update nuget package
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Only call cxplat_cleanup if cxplat_initialize succeeds
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
* Revert setup_build.vcxproj change
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Update usersim to latest
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
---------
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
Co-authored-by: Alan Jowett <alan.jowett@microsoft.com>
Other projects can now consume ebpf-for-windows in binary form,
using either nuget or MSI install. We no longer need to build
ebpf-for-windows both ways, so removing the cmake build to free
up build resources and reduce the development cost of maintaining
both msbuild and cmake builds.
Fixes#2743
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Update to use latest usersim
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* build ebpfcore under usersim
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Update usersim to latest
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Fix compilation
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Work around analysis warnings
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Fix installer project
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Fix memory issues in libbpf implementation
Found my latest usersim code
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Update usersim to latest
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Fix mismatched calloc/free function in ubpf_user.c
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Fix mismatched alloc/free in verifier_fuzzer
The verifier_fuzzer used ebpf_allocate() but then called free()
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Temporarily use ubpf branch with memory fix
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Fix memory issues in ebpfsvc
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* PR feedback from Anurag
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
---------
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Use usersim repo
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Remove DISPATCH test code from _preprocess_load_native_module
Per Anurag: the code was added earlier to execute the 2 paths in
native module, and it is not needed anmore IIRC, as the "DISPATCH"
code has been removed now from the native module
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Bug fixes
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Unify some APIs
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Fix warnings and clean up code
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* updated
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
---------
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Remove JIT and interpreter code from NativeOnly builds
Fixes#2030Fixes#2488
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Try conditional project references
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Various test fixes
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Native-only fixes to deploy-ebpf.ps1
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Native-only support for setup-ebpf.ps1
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Native-only support for api_test.exe
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Add connect_redirect_tests to deploy-ebpf.ps1
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* PR feedback from Anurag
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
---------
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Fix ProductVersion string in ebpf_resource.rc
Add git commit hash as "FileVersion" in ebpf_resource.rc
Signed-off-by: Alan Jowett <alanjo@microsoft.com>
* Add version string to drivers
Signed-off-by: Alan Jowett <alanjo@microsoft.com>
* Fix cmake build
Signed-off-by: Alan Jowett <alanjo@microsoft.com>
---------
Signed-off-by: Alan Jowett <alanjo@microsoft.com>
* clean up code
* fixing the build
* Update libs/ebpfnetsh/elf.cpp
Co-authored-by: Dave Thaler <dthaler@microsoft.com>
* fixing the build
* fixing the build
* fixing the build
* fixing the build
* fixing the build
* fixing the build
* revert HRESULT
* change BOOL type
* change BOOL type
* change BOOL type
* change BOOL type
* change BOOL type
* change BOOL type
* revert LPSERVICE_MAIN_FUNCTION
* revert SERVICE_MAIN_FUNCTION*
* SAL annotation for svcmain.cpp
* revert some SAL annotation
* remove some extra spaces
* fix tests
* change to bool type
* Update map_dll.c
* fix according to dll_main prototype
* fix BOOL
* fix according to dll_main prototype
* BOOl type
* BOOl type
* modify doc for self host runner setup
* modify doc for self host runner setup
* update tests with bool
* Update tests/unit/wer_report_test_wrapper.cpp
Co-authored-by: Dave Thaler <dthaler@microsoft.com>
* Update libs/platform/user/ebpf_platform_user.cpp
Co-authored-by: Dave Thaler <dthaler@microsoft.com>
* Update libs/ebpfnetsh/programs.cpp
Co-authored-by: Dave Thaler <dthaler@microsoft.com>
* change _Null_terminated_ to _Field_z_
* address PR comment
---------
Co-authored-by: Dave Thaler <dthaler@microsoft.com>
Co-authored-by: Gianni Trevisiol <gtrevi@users.noreply.github.com>
Co-authored-by: Alan Jowett <alanjo@microsoft.com>
* Clean up order of includes
Fixes#1963
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Update clang-format
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Clean up headers
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Update sort
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* FIx various build errors
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Fix ordering if including a .c file
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Fix cmake build
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* More header cleanup
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Update expected bpf2c output
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* More header fixes
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Update expected bpf2c output for custom bpf.c
The generate script does not handle this file.
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Fix tests
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Fix more expected bpf2c output
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
---------
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
Co-authored-by: Anurag Saxena <43585259+saxena-anurag@users.noreply.github.com>
Co-authored-by: Alan Jowett <alanjo@microsoft.com>
* Fix code analysis annotations for _In_ and _In_opt_ parameters,
add const where appropriate or change to _Inout_ where appropriate.
The following regex expressions were used to find the problems:
Find _In_ or _In_opt_ that are not const and are not followed by _Post_invalid_ or _Frees_ptr_ or _Post_ptr_invalid_:
```_In_[ ]+(?!.*(const|_Post_invalid_|_Frees_ptr_|_Post_ptr_invalid_))```
```_In_opt_[ ]+(?!.*(const|_Post_invalid_|_Frees_ptr_|_Post_ptr_invalid_))```
Some of the _In_ and _In_opt_ parameters are not const due to required compatibility with the Windows API.
Pointers to functions are const by definition, so the const qualifier is not required.
Signed-off-by: Alan Jowett <alanjo@microsoft.com>
* PR feedback
Signed-off-by: Alan Jowett <alanjo@microsoft.com>
* Fix code analysis failure
Signed-off-by: Alan Jowett <alanjo@microsoft.com>
* PR feedback
Signed-off-by: Alan Jowett <alanjo@microsoft.com>
* PR feedback
Signed-off-by: Alan Jowett <alanjo@microsoft.com>
* PR feedback
Signed-off-by: Alan Jowett <alanjo@microsoft.com>
Signed-off-by: Alan Jowett <alanjo@microsoft.com>
* kernel and platform changes
* fix
* user mode changes, other fixes
* fix test failure
* change function name
* add tests
* fix failing test
* cr comments
* CR comments
* fix cmake build
* fix
* Apply suggestions from code review
Co-authored-by: Dave Thaler <dthaler@microsoft.com>
* fix assert
* cr comments
* fix analysis failure
* fix test to avoid double close of fd
* test case fixes
* test fix
* fix unit test, analysis failure
* fix tests
* fix memory leak, possible race condition
Co-authored-by: Dave Thaler <dthaler@microsoft.com>
Co-authored-by: Alan Jowett <alanjo@microsoft.com>
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
Signed-off-by: Alan Jowett <alan.jowett@microsoft.com>
Co-authored-by: Alan Jowett <alan.jowett@microsoft.com>
Co-authored-by: Dave Thaler <dthaler@microsoft.com>
* WIP: Developer nuget package
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Add bpf2c utility
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Build nupkg from within VisualStudio
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Don't use relative path to externals
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Update props file in nuget package
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Fix include and library paths
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Fix debug build
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
Co-authored-by: saxena-anurag <43585259+saxena-anurag@users.noreply.github.com>
* Fix reply id/length on failure in KM
And make UM and KM more closely follow the same logic to catch this sort
of bug in the future. Before this, UM would succeed and KM would fail,
and there were only UM tests for this case.
Fixes#946
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Fix device_helper.hpp
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Fix reply length
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Cleanup
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* PR feedback
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
Co-authored-by: Alan Jowett <alanjo@microsoft.com>
* Add libbpf bpf_create_map_in_map() API
* bpf_create_map() now fails for outer maps. You must use
bpf_create_map_in_map() instead.
* Fix bug where EBPF_INVALID_FD was incorrectly converted to
EBPF_INVALID_ARGUMENT by ioctl handling code (part of issue #595)
One symptom of this bug was that errno was being set to EINVAL
in a number of cases which should have been EBADF.
* Fix bug where a HASH_OF_MAPS (unlike ARRAY_OF_MAPS) wasn't enforcing
that an inner map value had to match the inner map template.
Refactored the code in ebpf_maps.c so the checking is in done in one
place called by both maps, to ensure consistency.
* Fix bug in HASH_OF_MAPS where if an update failed, it would leave
the old entry but incorrectly drop the reference it held. It now
preserves the reference since the entry is unchanged.
* Added test case for ARRAY_OF_MAPS created via libbpf. Previously
only HASH_OF_MAPS creation was tested for that path.
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Fix bug caught by kernel test
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
Addresses part of issue #595
A more comprehensive fix is left for a separate PR.
This PR fixes some bad error messages from bpftool.
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Fix sample project dependencies
Bug was that if you cleaned everything and then tried to build just the
sample_ebpf_ext project, building would fail. A race condition also
existed when trying to build everything in that if
execution_context_kernel wasn't't done building by the
time sample_ebpf_ext was ready to link, the build would also fail.
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Clean up other linker inputs
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
* Fix non-determinism in build due to multiple projects generating header
* Only regenerate git_commit_id.h on change
Signed-off-by: Alan Jowett <alanjo@microsoft.com>
* Set device type to something _open_osfhandle understands
Signed-off-by: Alan Jowett <alanjo@microsoft.com>
* PR feedback
Signed-off-by: Alan Jowett <alanjo@microsoft.com>
Shankar Seal
Alan Jowett
Anurag Saxena
Dave Thaler
Sharmi
Amit Schendel
Dave Thaler
Gianni Trevisiol
Delaram Amiri
Dhiren Vispute
Alessandro Gario