- User Since
- May 5 2016, 2:57 PM (88 w, 5 d)
Fri, Jan 12
Thu, Jan 4
Addressing review comments.
Wed, Jan 3
Tue, Jan 2
Mon, Jan 1
Sun, Dec 31
For ASan, an identical check exists as a background thread. We decided not to do that due to some requirements on Android.
It could potentially come back as a background thread check on supported platforms if that would address the concerns you are raising.
How about using calloc instead and benefiting of the "included" zero'd memory feature? That would avoid implementing the touch function.
I won't be able to test anything before Jan. 2nd.
Tue, Dec 19
As discussed: you probably want to enforce mutual exclusion of the sanitizers (eg: ASan is not compatible with Scudo).
I have not test Scudo+Safestack but they should be compatible (or at least with a small amount of work can should be made to be).
Dec 16 2017
Dec 14 2017
Dec 13 2017
Adding a TODO to add tests for the various defines.
Currently the tests we run leverage a static library built with a default
configuration. We should add gtests like the other sanitizers to ensure all
the compile time configuration options work as expected.
Added some words about security in the interface comment.
Added a define to allow the disabling of the function.
Dec 12 2017
I should add here that this is required for programs wanting to set their RSS limit based on computations done at runtime due to hardware constraints or other factors.
From a security perspective, this could have an impact if called with parameters controlled by a malicious user, or if set too low by a program (memory exhaustion exploitation is on the rise).
Correct the last #endif comment of the file while I am at it, since it
didn't include SANITIZER_NETBSD.
As suggested, split the code cleanly between Linux-non-Go and the rest.
Moving the vDSO check under a SANITIZER_LINUX define.
Dec 11 2017
Rebasing one last time.
Dec 8 2017
Dec 7 2017
Dec 6 2017
Adding a MonotonicNanoTime stub for mac.
Replacing some const with constexpr.
Dec 5 2017
@krytarowski: I'd be grateful if you could confirm that this would work on NetBSD. Thanks!
Moving the NetBSD clock_gettime syscall mangling to
sanitizer_syscall_generic.inc as is the case for others.
Use MonotonicNanoTime for the Scudo RSS checks.
Dec 4 2017
Introduce IsRegionExhausted that checks if the region has enough room prior
to calling Mmap for the user & metadata portions of memory.
Fold all the metadata work in a if (kMetadataSize).
Dec 2 2017
Dec 1 2017