Hitachi SH target is slightly broken

[Tony Bryant <brd at paradise dot net dot nz>]

Just to save somebody else a day of ****ing around:

To do interrupt handlers with the SH target, on gcc 2.96 (and probably 
later), you've got to do both

__attribute__ ((interrupt_handler))	

AND

#pragma interrupt


The attribute causes an RTE to be generated, while the
#pragma stops the final stack pop being put in its delay slot, (which
is an awefully hard problem to track down)

Bizzare behaviour, but one quick look at the machine description and I
can see that "sh.c" now uses the attribute, but "sh.md" still uses the 
#pragma.

Who should I report this to, to get this fixed?


------
Want more information?  See the CrossGCC FAQ, http://www.objsw.com/CrossGCC/
Want to unsubscribe? Send a note to crossgcc-unsubscribe@sourceware.cygnus.com