[Top][All Lists]

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

[Bug ld/20173] LD: Link issue with grantlee 0.5.1

From: romain.naour at gmail dot com
Subject: [Bug ld/20173] LD: Link issue with grantlee 0.5.1
Date: Fri, 03 Jun 2016 21:31:08 +0000


--- Comment #4 from romain.naour at gmail dot com ---
Hi Nick,

(In reply to Nick Clifton from comment #3)
> Hi Romain,
>   Ha - found it.  You are linking with "--fatal-warnings" specified on the
> linker command line.  Hence the warnings about being unable to create a
> .eh_frame_hdr section are causing the link to fail.

Ha right ! I missed --fatal-warnings flag hardcoded in the grantlee

>   I am not an expert in your build system, but I suspect that this option is
> the result of a configure-time option, or possibly the absence of a
> configure-time option to turn it off.

Actually, Buildroot itself doesn't turn a warning into error, it's the grantlee
build system. I don't think the grantlee developers ever tried to compile with
a microblaze or nios2 toolchain.

>   I can understand that you might be reluctant to turn this option off, as
> it would allow other, possibly important, warnings to fly by without
> necessarily receiving the attention of the programmer.  If that is the case
> then the only thing that I can suggest is that you do not enable PIC for
> these libraries.

Well, you're probably right but it's sometime difficult to keep all warnings
enabled since in the autobuilders try to build every packages with different
kind of toolchains (glibc, uClibc, musl based) and for each CPU architecture
Buildroot support. We try to the best effort by resolving each critical build
error but leaving warnings aside.

Since you said that this warning is not so serious (PR19405#c5), I'll simply
remove this flag from the CMakeLists.txt.

>   The root cause of the problem is that the nios2 target (and microblaze as
> well) does not generate PC-relative addresses in its .eh_frame sections, and
> this makes it impossible to generate the .eh_frame_hdr section.  The warning
> is intended as a hint to the programmer to let them know why the linker is
> not doing something that they might otherwise expect.

Thanks for the explanation, I think nios2 user/programmer should be aware of

In the end, this not a linker issue. I'll clone this bug report.

Thanks for your time,

Best regards,

> Cheers
>   Nick

You are receiving this mail because:
You are on the CC list for the bug.

reply via email to

[Prev in Thread] Current Thread [Next in Thread]