#1800 closed defect (wontfix)

PowerPC: IEEE 754 conformance of e200/e500 embedded floating point unit

Reported by: Sebastian Huber Owned by: Sebastian Huber
Priority: normal Milestone: 4.11
Component: bsps Version: 4.11
Severity: normal Keywords:
Cc: ralf.corsepius@… Blocked By:
Blocking:

Description (last modified by Sebastian Huber)

For full IEEE 754 conformance of the e200 and e500 embedded floating point unit we need software support. This consists of an exception handler that decodes the offending instruction and provides the results via a software floating point implementation.

The suggested software floating point library is:

http://www.jhauser.us/arithmetic/SoftFloat.html

Change History (9)

comment:1 Changed on 05/19/11 at 05:17:52 by Sebastian Huber

dependson: 1740
Owner: changed from Joel Sherrill to Sebastian Huber

comment:2 Changed on 05/19/11 at 05:18:14 by Sebastian Huber

Status: newassigned

comment:3 Changed on 05/19/11 at 05:52:43 by Ralf Corsepius

Cc: Ralf Corsepius added

comment:4 Changed on 05/19/11 at 07:03:36 by Sebastian Huber

Replying to comment:3:

Replying to comment:2:

Replying to comment:1:

Replying to comment:0:

For full IEEE 754 conformance of the e200 and e500 embedded floating point unit
we need software support. This consists of an exception handler that decodes
the offending instruction and provides the results via a software floating
point implementation.

The suggested software floating point library is:

http://www.jhauser.us/arithmetic/SoftFloat.html

I fail to understand this proposal, once again presented in eat-or-die fashion.

What do you mean with "eat-or-die fashion"?

Let me answer in German: Most of your proposals are presented in a way, you are
telling "I must have, eat it or die"

It's not my fault that this hardware floating point unit is that way.

In German: Die meisten Deiner Vorschläge lassen sich auf einen Satz
zusammenfassen der lautet "friss das oder strib".

RTEMS uses newlib's soft-float - What's wrong with it?

The e200 and e500 embedded floating point unit is a hardware floating point
unit.

=> Integrate it into newlib or gcc.

I absolutely don't understand what this has to do with Newlib or GCC.

The Linux equivalent is here:

http://lxr.free-electrons.com/source/arch/powerpc/math-emu/math_efp.c?a=powerpc

comment:5 Changed on 04/01/14 at 11:28:22 by Sebastian Huber

Milestone: 4.114.12

comment:6 Changed on 11/20/14 at 03:31:07 by Chris Johns

Milestone: 4.124.11

comment:7 Changed on 11/21/14 at 12:33:44 by Sebastian Huber

Description: modified (diff)
Resolution: wontfix
Status: assignedclosed

No demand to fix this from the users in the last four years.

comment:8 Changed on 11/24/14 at 18:58:28 by Gedare Bloom

Version: HEAD4.11

Replace Version=HEAD with Version=4.11 for the tickets with Milestone >= 4.11

comment:9 Changed on 12/12/14 at 07:24:41 by Sebastian Huber

Description: modified (diff)
Milestone: 4.124.11
Resolution: wontfix
Status: assignedclosed
Note: See TracTickets for help on using tickets.