GCC 6.3.0 build fails on Ubuntu 20.04: make: *** [Makefile:924: all] Error 2 - libitm / libquadmath

lordmund lordmund@protonmail.com
Wed Jan 27 14:34:45 GMT 2021

Hi Paul / Hi All,

Thank you very much. I appreciate your very quick answer.

All your recommendations proved to be very useful.

As you suggested, I redirected the output of both the build and stderr to an external file, and let the build process run in serial mode instead of the parallel one that I had done before.

The serial build took much more time (about 2 hours) but this is normal behavior. I presume.

In fact, I had to let it run twice as after the first run. The same error message about 'sanitizer_platform_limits_posix.cc' occured as explained in the second link of my original post.

This one was straightforward to fix based on that information.

Then after the second run, which was successful, I have investigated again the log file. No error message appeared only a few warning messages. I think they are acceptable.

So after all these, I went further as described at the very first link of my OP.
which is for easier reference here again:

These are the specific steps that I closely followed:

Once the build is finished, we can continue to actually installing GCC with the following command:

sudo make install

With adding the following lines at the end of the .bashrc file, we will permanently add the compilers to the system’s path.

nano /root/.bashrc

Add this to the end of that file:

export export PATH=/usr/local/gcc-8.2/bin:$PATH
export LD_LIBRARY_PATH=/usr/local/gcc-8.2/lib64:$LD_LIBRARY_PATH

With this command, we will reload .bashrc without restarting the machine:

. .bashrc

No error messages or anything similar, therefore my opinion is that everything was successful.

The GCC 6.3.0 was installed in the below directory of my computer:


However, my Ubuntu (terminal) cannot find the path to the recently installed GCC 6.3.0, which I find very interesting why.

Could you please help me what I still miss to properly configure GCC 6.3.0?

In my opinion, there must be only a very minor command that needs to be run for the entirely correct configuration.

Thank you very much for your help in advance again.


‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐
On Tuesday, January 26, 2021 10:23 PM, Paul Smith <psmith@gnu.org> wrote:

> On Tue, 2021-01-26 at 14:04 -0700, lordmund--- via Gcc-help wrote:
> > Unfortunately, the above error message is quite informative. (maybe
> > just for me)
> I don't see any actual error messages here (other than the last line
> which is just make telling you that something earlier failed).
> I suspect that the actual error is much earlier in your build process
> than what you quote here, and since you're using parallel builds the
> build continues on with other things even after the error.
> You should be sure you redirect the output of the build (and stderr!)
> to a file, then search the file for the FIRST instance of a make error
> message (has the "***" in the message) and look at what errors happened
> just before that. Maybe that will help you understand the problem. If
> not reply to the list with those error messages.

More information about the Gcc-help mailing list