[Bug 1950] arm/armeabi: error: unknown type name 'CPU_Interrupt_frame'

bugzilla-daemon at rtems.org bugzilla-daemon at rtems.org
Mon Nov 7 07:38:22 UTC 2011


https://www.rtems.org/bugzilla/show_bug.cgi?id=1950

--- Comment #5 from Sebastian Huber <sebastian.huber at embedded-brains.de> 2011-11-07 01:38:18 CST ---
This break is more or less intentional, because I don't know which multilib is
appropriate for the ARM1136JF-S core.  We also have no real BSP using this
core.

(In reply to comment #4)
> (In reply to comment #1)
> > Another case of missing a multilib variant.
> 
> I now realize why you are talking about "multilib variant"s.
> 
> You are referring to the ARM_MULTILIB_ARCH_* defines cascade from 
> score/cpu/arm/rtems/score/arm.h.
> 
> Two comments on this:
> a) I consider these defines to be mis-named. They reflects groups of
> sub-targets RTEMS support and have nothing to do with multilibs.

They have to do with multilibs.  In the cpukit we can only use architecture
specific defines which relate to a multilib.  Based on the __ARM_ARCH_* define
the appropriate multilib is selected via the ARM_MULTILIB_ARCH_* define.  All
the ARMv4* and ARMv5* cores will use the ARMv4 or ARMv4T multilib.

> 
> b) I consider this cascade to be harmful and to be a massive regression,
> because they are trying to connect unconnected compiler features.

How else would you do this?

> 
> [Several years ago I had spend a considerable amount of time to reduce and
> remove similar cascades, esp. for the powerpc, where they had proven to be
> unmaintainable and to be the cause of the exception handling mess the powerpc
> had suffered from.]

This is completely different to the PowerPC.

-- 
Configure bugmail: https://www.rtems.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are watching all bug changes.



More information about the bugs mailing list