[Dwarf-Discuss] DWARF piece questions
paul.robinson at sony.com
Fri Jan 27 11:40:27 PST 2017
> -----Original Message-----
> From: Dwarf-Discuss [mailto:dwarf-discuss-bounces at lists.dwarfstd.org] On
> Behalf Of Andreas Arnez
> Sent: Friday, January 27, 2017 11:16 AM
> To: Michael Eager
> Cc: DWARF Discussion; Ulrich Weigand
> Subject: Re: [Dwarf-Discuss] DWARF piece questions
> On Fri, Jan 27 2017, Michael Eager wrote:
> > On 01/27/2017 06:49 AM, Andreas Arnez wrote:
> >> But if some "even less significant" bits were added (such as with
> >> z/Architecture, where a newer release extended 64-bit FP-registers to
> >> 128-bit vectors), then the numbering scheme has to change. This breaks
> >> compatibility with the debug info in existing programs. That's the
> >> problem I was trying to outline above.
> > You need to emulate the old architecture on the new architecture. You
> > cannot assume that DWARF generated for an old architecture will be
> > usable without interpretation on an arbitrarily different new
> > architecture.
> So, from a DWARF perspective, you'd expect that all libraries shall be
> recompiled when migrating from an older x86-64 CPU to a newer one that
> has AVX-512? Or, as in the z/Architecture case, from a zEC12 to a z13
> system? You don't consider it valid for old and new binaries to coexist
> in the same program?
I'd expect that the DWARF consumer will understand what the target for the
original executable was, and interpret any DWARF expressions appropriately.
My understanding is that AVX and AVX-512 were considered to expand the
vector registers at the most-significant-bit end, not the LSB end.
This allows the same DWARF register numbers to be reused and counting
bits from the LSB end works regardless of the register width.
I'm not able to find equivalent comparison of the z/Architecture models,
but if they do something similar then it should all Just Work.
More information about the Dwarf-Discuss