[RTEMS Project] #747: m68k-rtems-gcc-3.4.3 -m528x links against wrong multilib variant

RTEMS trac trac at rtems.org
Fri Nov 21 14:12:09 UTC 2014


#747: m68k-rtems-gcc-3.4.3 -m528x links against wrong multilib variant
--------------------+----------------------------
 Reporter:  norume  |       Owner:  joel.sherrill
     Type:  defect  |      Status:  closed
 Priority:  normal  |   Milestone:  2
Component:  tools   |     Version:  unknown
 Severity:  major   |  Resolution:  wontfix
 Keywords:          |
--------------------+----------------------------
Changes (by joel.sherrill):

 * status:   => closed
 * resolution:   => wontfix


Old description:

> GCC doesn't seem to know that '-m528x' should imply the use of the
> 'm5200' multilib variant when linking.  Instead the link editor ends up
> using the default multilib variant (corresponding to a 68020, I believe)
> -- which has non-ColdFire instructions.  These, of course, cause
> exceptions when the mcf5282 tries to execute them.
>
> My quick-fix attempt was to addm528x=m5200 to the MULTILIB_MATCHES in
> gcc/config/m68k/t-m68kelf but this seems to have had no effect.
>
> Release:
> gcc-3.4.3
>
> Environment:
> Linux/Darwin hosts, MCF5282 target
>
> How-To-Repeat:
> Add -Wl,--trace to a -m5828x link and you'll see that the wrong libraries
> are used.

New description:

 GCC doesn't seem to know that '-m528x' should imply the use of the 'm5200'
 multilib variant when linking.  Instead the link editor ends up using the
 default multilib variant (corresponding to a 68020, I believe) -- which
 has non-ColdFire instructions.  These, of course, cause exceptions when
 the mcf5282 tries to execute them.

 My quick-fix attempt was to addm528x=m5200 to the MULTILIB_MATCHES in
 gcc/config/m68k/t-m68kelf but this seems to have had no effect.

 Release:
 gcc-3.4.3

 Environment:
 Linux/Darwin hosts, MCF5282 target

 How-To-Repeat:
 Add -Wl,--trace to a -m5828x link and you'll see that the wrong libraries
 are used.

--

--
Ticket URL: <http://devel.rtems.org/ticket/747#comment:2>
RTEMS Project <http://www.rtems.org/>
RTEMS Project


More information about the bugs mailing list