- User Since
- Sep 8 2017, 3:58 PM (124 w, 1 h)
- Merged from master (after enable/disable + fork support by @eugenis).
Reviving this a little. check-gwp_asan in my aarch64 crostool now WAI, with some warnings during compilation from ar and ranlib: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0000000. Looks like this is the PAC/BTI properties from D62609, but doesn't really hurt us here.
AFAIK we do test on android.
LGTM with installAtFork() and cond_wait() comments.
For tracking purposes.
This patch broke the UBSan build bots here.
Tue, Jan 21
This is unfortunately one of those situations where we have to bend to
existing code because we doubt it's going to change any time soon.
Mon, Jan 20
This test broke the UBSan buildbot here. I've rolled it back for now, but to reproduce the issue should be as simple as adding -DLLVM_USE_SANITIZER=Undefined and running ninja check-lld or make check-lld depending on which tool you use :)
Mon, Jan 13
Sun, Jan 12
Oh, also, would you mind please updating the commit message to mention that this change adds enable/disable support, and fork protections? Thanks!
Thanks for doing this.
Tue, Jan 7
Dec 23 2019
Dec 18 2019
This change broke the sanitizer buildbots: http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux-fast/builds/37446
Fyi - slight lint bug introduced here, fixed in 2f4c3c565a34b2c30fb883c85acbf4114b455c26.
Dec 17 2019
Reverted in 2423774cc2a040d34d32aacaf261805cb195ebd2 due to breaking the sanitizer buildbots: http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux-fast/builds/37410
Dec 16 2019
Dec 13 2019
Dec 11 2019
Dec 10 2019
- Enforce alignment to GWP-ASan. If userspace calls posix_memalign, this alignment guarantee should be kept even if the allocation is chosen to be sampled.
- Guard isOwned behind macro.
Dec 9 2019
- Synced to master, added isOwned and /s/size_t/uptr.
LGTM with a couple comments
Dec 6 2019
Dec 4 2019
Dec 2 2019
So the model presented asserts that if the chunk header is truncated, the pointer is not owned by us. Is this WAI? I can forsee that a chunk header was truncated, and then the pointer to the associated allocation is checked for ownership, which the ownership will fail as the chunk header check didn't succeed.
Nov 27 2019
either side of the allocation are tagged with tag 0, which is normally
excluded from the set of tags that may be selected randomly
Nov 26 2019
Nov 25 2019
Nov 21 2019
Other than that, LGTM