This is the mail archive of the dwarf2@corp.sgi.com mailing list for the dwarf2 project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: A DWARF3 call frame issue


Lee Smith wrote:
Mike,

We are busy nailing down the final loose corners of the ABI for the ARM
Architecture.

One such is the use of DW_CFA_same_value in CIE entries in
dot-debug_frame sections.

The DWARF standard provides DW_CFA_same_value as a way of stating
explicitly that the value of a register as the same as in the previous
frame. In general this cannot work in the ARM world because a producer
cannot know what registers will be present in a system using its code
(e.g. both VFP and Wireless MMX might or might not be present, and
future architecture versions might introduce new core registers).

A consumer (debugger) therefore _must_ assume that registers not
mentioned are not perturbed by code associated with the frame unwinding
description.

Having accepted this principle, we feel that it should be adopted
uniformly and, therefore, that the generation of DW_CFA_same_value
entries should be entirely optional.

We propose to write this requirement into "DWARF for the ARM
Architecture" unless consultation with interested parties reveals that
it's a bad thing to do.

Do you have an opinion either way?

Comments?



-- Michael Eager Eager Consulting eager@eagercon.com 1960 Park Blvd., Palo Alto, CA 94306 650-325-8077


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]