This is the mail archive of the
mailing list for the GDB project.
Re: Call stack trace implementation
On Tue, Dec 11, 2007 at 12:59:33PM -0800, Ananth Sowda wrote:
> My question is on implementation of stack backtrace on architures,
> MIPS and PPC. Is the implementation based on libunwind and unwind
> directives in the executable? Or is it based on older code which
> analyzes prologue of a function? Or a mix of both approaches?
Both prologue analysis and call frame information. Libunwind is not
used except on IA64; GDB has its own support for DWARF-2 unwind tables.