Re: [PATCH bpf-next v2 2/3] bpf: btf: add btf print functionality
From: Martin KaFai Lau
Date: Tue Jul 03 2018 - 19:34:25 EST
On Tue, Jul 03, 2018 at 03:38:43PM -0700, Jakub Kicinski wrote:
> On Tue, 3 Jul 2018 15:23:31 -0700, Jakub Kicinski wrote:
> > > > > + else
> > > > > + jsonw_printf(jw, "%hhd", *((char *)data));
> > > >
> > > > ... I think you need to always print a string, and express it as
> > > > \u00%02hhx for non-printable.
> > > Okay that makes sense
> >
> > Yeah, IDK, char can be used as a byte as well as a string. In eBPF
> > it may actually be more likely to just be used as a raw byte buffer...
>
> Actually, what is the definition/purpose of BTF_INT_CHAR? There seems
> to be no BTF_INT_SHORT and BTF_INT_SIGNED can simply be of size 8...
> Is normal int only used for bitfields of size 8 and BTF_INT_CHAR for
> char variables?
>
> The kernel seems to be rejecting combinations of those flags, is
> unsigned char going to not be marked as char then?
BTF_INT_ENOCODING (CHAR/SIGNED/BOOL) is for formatting (e.g. pretty
print). It is mainly how CTF is using it also. Hence, BTF_INT_ENCODINGs
is not a 1:1 mapping to C integer types.
The size of an interger is described by BTF_INT_BITS instead.
>
> > Either way I think it may be nice to keep it consistent, at least for
> > the JSON output could we do either always ints or always characters?
>