pkgsrc-Users archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: __stack_chk_fail_local pkgsrc with gcc4.8 (was Re: minimum gcc to compile pkgsrc?



On Tue, Aug 01, 2017 at 11:24:42PM -0000, yancm%sdf.org@localhost wrote:
> > On 07/13/17 17:36, Brook Milligan wrote:
> >> is there any sense of what minimum gcc version is
> >> required as the native compiler for all the packages
> >> that are not marked with special GCC_REQD?
> 
> > 4.8.1.
> >
> > It's the first to fully support c++11:
> >
> > https://gcc.gnu.org/projects/cxx-status.html
> >
> > Currently there aren't many packages that will fail with GCC 4.4.7
> > but it's enough to be a nuisance, so I'm setting
> > GCC_REQD=4.8 in mk.conf on my Linux systems until there's an official
> > solution in place.
> 
> I'm on NetBSD 6_Stable and recently started having issues compiling
> various pkgsrc packages... so I individually set them to require
> gcc 4.8...this worked at first...
> 
> then I ran in to a new (to me) issue when trying to compile...
> I receive "__stack_chk_fail_local" which after googling, has been
> around a while, it's apparently new behavior in 4.8 that was
> not in earlier version maybe?
> 
> I found this "stack smash protection which is enabled by default
> in gcc - I had similar issues when trying to compile patched
> kernels. To circumvent this issue you could compile with
> CFLAGS="-fno-stack-protector -fno-stack-protector-all" to
> turn it off. Settin ghti flag did not help...
> 
> Any clues would be helpful?!

This might be a new issue from the addition of PKGSRC_USE_SSP=yes by
default. I suspect it is a package specific issue, not a netbsd6 + old
gcc issue. Which package?

You can also set PKGSRC_USE_SSP=no to work around this for now.


Home | Main Index | Thread Index | Old Index