http://llvm.org/bugs/show_bug.cgi?id=17059
Kostya Serebryany <[email protected]> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED CC| |[email protected] Resolution|--- |FIXED --- Comment #1 from Kostya Serebryany <[email protected]> --- r189806 implements almost what you've asked for. The reports look like this now: [32, 42) 'AAA' [96, 106) 'BBB' <== Memory access at offset 94 underflows this variable [160, 170) 'CCC' [32, 42) 'AAA' [96, 106) 'BBB' <== Memory access at offset 95 partially underflows this variable [160, 170) 'CCC' [32, 42) 'AAA' [96, 106) 'BBB' <== Memory access at offset 105 partially overflows this variable [160, 170) 'CCC' [32, 42) 'AAA' [96, 106) 'BBB' <== Memory access at offset 106 overflows this variable [160, 170) 'CCC' Sometimes like this (when the access is somewhere in between two vars): [32, 42) 'AAA' [96, 106) 'BBB' <== Memory access at offset 132 overflows this variable [160, 170) 'CCC' <== Memory access at offset 132 underflows this variable The reports do not mention the access sizes. Fixing this is a bit more involved: when memset touches invalid memory, we report the size of the whole memset but the reported address is the fist invalid address. But that's another story. -- You are receiving this mail because: You are on the CC list for the bug.
_______________________________________________ LLVMbugs mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/llvmbugs
