[llvm-commits] [PATCH] asan-rt: discover main thread stack limits without pthread

[Kostya Serebryany]

On Wed, Dec 28, 2011 at 3:38 AM, Evgeniy Stepanov <eugeni.stepanov at gmail.com
> wrote:

> In fact, I'd like copy or reimplement ProcMapsIterator as part of ASan
> to escape the ifdef hell.
>

Yea, I've got tired of it myself too. I'll try to re-implement the linux
version next week.


> But this should be good enough, lets land it first.
>
r147321 (fixed style)
Thanks (let's see if this fixed the wine problem)

--kcc


> On Wed, Dec 28, 2011 at 3:24 AM, Kostya Serebryany <kcc at google.com> wrote:
> > I like the idea (it may actually fix the problem reported today on wine),
> > but we need to guard the code that uses sysinfo/sysinfo.h
> > with ASAN_USE_SYSINFO==1 (see asan_stack.cc).
> > Or add some macro machinery to get sysinfo.h from an alternative place
> as in
> > perftools (base/sysinfo.h).
> >
> > --kcc
> >
> >
> > On Wed, Dec 21, 2011 at 4:04 AM, Evgeniy Stepanov
> > <eugeni.stepanov at gmail.com> wrote:
> >>
> >> Hi,
> >>
> >> if __asan_init is called from .preinit_array, pthread_getattr_np may
> >> become unsafe. This patch adds a different way of locating the stack
> >> of the main thread with a combination of getlrimit() and
> >> /proc/self/maps.
> >
> >
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20111228/83fd01e5/attachment.html>