Re: [PATCH] bpf: bpftool: convert ‘const char *’ type into 'char *' in assignment

[Bo YU <tsu.yubo@gmail.com>]

On Thu, Mar 28, 2019 at 10:33 PM Quentin Monnet
<quentin.monnet@netronome.com> wrote:
>
> 2019-03-28 10:16 UTC-0400 ~ Bo YU <tsu.yubo@gmail.com>
> > When compiling with check flag: make -C tools/bpf/bpftool/
> > gcc will warning:
> >
> > jit_disasm.c:119:29: warning: assignment discards ‘const’ qualifier from
> > pointer target type [-Wdiscarded-qualifiers]
> >   info.disassembler_options = disassembler_options;
> >
> > So convert 'const char *' type into 'char *' to fix the warning.
> >
> > Fixes:3ddeac6705aba(tools: bpftool: use 4 context mode for the NFP disasm)
> > Signed-off-by: Bo YU <tsu.yubo@gmail.com>
> > ---
> > tools/bpf/bpftool/jit_disasm.c | 2 +-
> > 1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/tools/bpf/bpftool/jit_disasm.c
> > b/tools/bpf/bpftool/jit_disasm.c
> > index 3ef3093560ba..40b46e46454a 100644
> > --- a/tools/bpf/bpftool/jit_disasm.c
> > +++ b/tools/bpf/bpftool/jit_disasm.c
> > @@ -116,7 +116,7 @@ void disasm_print_insn(unsigned char *image, ssize_t
> > len, int opcodes,
> >     info.arch = bfd_get_arch(bfdf);
> >     info.mach = bfd_get_mach(bfdf);
> >     if (disassembler_options)
> > -        info.disassembler_options = disassembler_options;
> > +        info.disassembler_options = (char *)disassembler_options;
> >     info.buffer = image;
> >     info.buffer_length = len;
> >
>
>
> Hi, the warning does not show with latest versions of binutils, it was
> fixed with this (binutils) commit:
>
>     commit f995bbe8e62fdb5607acb1ee127240cfe50d2b8f
>     Author: Pedro Alves <palves@redhat.com>
>     Date:   Wed Apr 5 19:21:33 2017 +0100
>
>         -Wwrite-strings: Constify struct disassemble_info's
>                 disassembler_options field
>
> Distributions with recent binutils (Ubuntu 18.04, Fedora 29 at least)
> should not have this message. I thought about sending the same patch
> some time ago, but in the end Jakub and I decided that it may not be
> worth casting the const away to silence the warning just for older
> distributions. Daniel also shared the same opinion at the time (but that
> was off the mailing-list).
Ok, good to know it.
Thanks,
>
> Quentin