readelf: simplify Xen string note printing

[tig_freebsdfoundation.org]

contrib/elftoolchain/readelf/readelf.c

3710 ↗

(On Diff #69725)

I noticed that this type==0 case didn't exist before. Is this needed? If not, I think you can get rid of the function pointer field in struct note_desc.

[markj]

contrib/elftoolchain/readelf/readelf.c

3685 ↗

(On Diff #69725)

Why don't we dump data in this case anymore?

3710 ↗

(On Diff #69725)

It's needed to ensure that loop iteration over xen_notes[] terminates. One common trick which eliminates the need for the nul terminator is to iterate using this pattern:

for (i = 0; i < sizeof(xen_notes) / sizeof(xen_notes[0]) /* sometimes written "nitems(xen_notes)" */; i++)

but existing practice in this code seems to be what Ed is doing, so I think it is ok.

I don't have a strong feeling about the function pointer. Indeed, it can be dropped, but if in the future a note is added that requires a custom print function, it'll need to be re-added.

[emaste]

contrib/elftoolchain/readelf/readelf.c

3685 ↗

(On Diff #69725)

returning false now falls through to the unknown: case in dump_notes_data and prints it as hex

3710 ↗

(On Diff #69725)

Yeah, my goal is to extend this to add custom print functions, and move all note printing to be done via tables like this.

[markj]

contrib/elftoolchain/readelf/readelf.c

3685 ↗

(On Diff #69725)

Oh, I mixed up the handling of nd->fp's return value.