I think the CFI support is about ready to merge in. I did a single-commit version on the new branch roland/merge-cfi. This has some recent fixups for reloc support, and I dropped dwfl_addrframe for now since I'm not real sure about it.
Please give it a good eyeball and tell me if you think it is OK to merge.
Thanks, Roland
Roland McGrath wrote:
I did a single-commit version on the new branch roland/merge-cfi.
- please update the copyright years to show 2009. Even if the files are older, today is when they get added
- you should run pahole on libdw. Some of the new data structures contain unnecessary holes
- do you plan to use the new functions in readelf?
It's OK to merge once you fix the two points above.
- please update the copyright years to show 2009. Even if the files are
older, today is when they get added
Done.
- you should run pahole on libdw. Some of the new data structures
contain unnecessary holes
I first did this on the existing libelf and libdw and they had some improvements to be made in their private structs too.
- do you plan to use the new functions in readelf?
On the old branch I had made readelf use dwarf_next_cfi, which is just the minimal harder parser a la dwarf_nextcu. I punted that when you implemented CFI reading in readelf separately. I don't really recall if what I'd done before gave output that was any better.
I don't intend to have readelf use the fancy stuff for -wframe. It's useful for readelf just to be an independent reader/dumper of the low-level format.
It's OK to merge once you fix the two points above.
Merged!
Thanks, Roland
elfutils-devel@lists.fedorahosted.org