Update TODO.sanitizers

Mark as done:
- upstream local patches, mostly to compiler-rt
  <- all patches submitted to review
- investigate and address the libcxx failing tests on NetBSD
  <- done by Michal Gorny
- FILE and DIR sanitization (needed by at least: ESan, MSan)
  <- done by Michal Gorny
- add missing interceptors (mostly MSan-centric)
  <- all local patches merged upstream, it's too opaque to keep it around
- port or finish the port of: XRay, profile
  <- done
- detect and fix more bugs in basesystem
  <- too opaque to keep it here
- uvm_map.c E2BIG workaround or limit raise (1MB -> 10MB?), needed by libFuzzer
  <- done
This commit is contained in:
kamil 2018-12-27 21:20:40 +00:00
parent 4b140abf1c
commit 988e2611ad
1 changed files with 0 additions and 8 deletions

View File

@ -4,32 +4,25 @@ Short term:
- compiler-rt import and integration with base
- merge TSan, MSan and libFuzzer ATF tests
- prepare MKSANITIZER readme
- upstream local patches, mostly to compiler-rt
Long term:
- handle fts_open(3) compar callback in sanitizers (use TLS to store
the real callback and call a wrapper that will [un]poison appropriate
memory regions)
- investigate and address the libcxx failing tests on NetBSD
- FILE and DIR sanitization (needed by at least: ESan, MSan)
- add missing interceptors (mostly MSan-centric)
- improve the framework for ioctl(2) database and handle special cases when a
part of a struct can be uninitialized and passed to the kernel (MSan)
- port or finish the port of:
* ESan,
* LSan,
* XRay,
* shadowcallstack,
* cfi,
* scudo,
* profile,
* DFSan.
- finish the research of a new syscall for StopTheWorld() operation
ptrace(2) style for self-introspection and debugging (LSan)
- port HWASan -- aarch64 specific, needs hardware or emulator
- make MKSANITIZER usable with any supported by a compiler combination of
sanitizers
- detect and fix more bugs in basesystem
- pkgsrc integration with sanitizers and MKSANITIZER (PKGSANITIZER?)
- port what possible to !amd64
- fixes with the signal code
@ -39,7 +32,6 @@ Long term:
- lld port to NetBSD (GNU ld(1) is too slow for serious LLVM development)
- merge of the libFuzzer integration with the basesystem
- oom-killer ATF tests and fixes
- uvm_map.c E2BIG workaround or limit raise (1MB -> 10MB?), needed by libFuzzer
- utime(3) allow changing actime regardless of noatime mount property
- NetBSD tar: handle character escaping in file names (\\ \), needed in
check-lld (lld::reproduce-backslash.s)