[LLVMdev] UBSan detects misaligned memory accesses in llvm-profdata and llvm-cov

Alexey Samsonov vonosmas at gmail.com
Fri Sep 12 15:48:19 PDT 2014


Thanks!

On Fri, Sep 12, 2014 at 2:42 PM, Justin Bogner <mail at justinbogner.com>
wrote:

> Alexey Samsonov <vonosmas at gmail.com> writes:
> > Hi!
> >
> > UBSan bootstrap bot fails with error report on 5 llvm-cov and
> llvm-profdata
> > lit-tests:
> >
> http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux-bootstrap/builds/4526
> > /steps/check-llvm%20ubsan/logs/stdio
> > Also see http://llvm.org/bugs/show_bug.cgi?id=20815.
>
> Thanks for pointing these out!
>
> > The error reports look vaild: in general it's incorrect to load
> uint64_t, or
> > even structures like "RawHeader" or "CoverageMappingTURecord" from a
> random
> > address. Also, if you parse files stored on disk, it would certainly
> help to
> > know endianness of data in these files and, e.g., use routines from
> Support/
> > Endian.h. Can you take a look at this?
>
> I've fixed the llvm-profdata failure in r217708. The file format has a
> known endianness and is always padded to a proper alignment, so this is
> just a matter of rejecting files that are malformed.
>
> I'm a bit less familiar with the llvm-cov format, so I'll look into that
> one and get it fixed in the next day or two.
>



-- 
Alexey Samsonov
vonosmas at gmail.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20140912/a9377ba7/attachment.html>


More information about the llvm-dev mailing list