------- Comment #40 from dave at hiauly1 dot hia dot nrc dot ca  2006-01-14 
00:57 -------
Subject: Re:  [4.1/4.2 Regression] internal compiler error: verify_flow_info
failed

> ------- Comment #30 from steven at gcc dot gnu dot org  2006-01-13 18:51 
> -------
> I wonder if HP-PA should not just make targetm.cannot_modify_jumps_p return
> true after reload...

I assume this question is somewhat independent of the PR.

I didn't think that this was necessary.  There is some somewhat horrible
code to save register r1 into the frame and restore it in the delay slot
of a jump in situations where we need a register for the jump (i.e.,
the jump is too large to be handled using a pc-relative offset).  The
downside is the jump needs several instructions more than would be needed
if we had a register handy.

Dave


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24626

Reply via email to