[Bug target/69030] [6 Regression] ICE on x86_64-linux-gnu at -O2 and above in 32-bit mode (ICE in copy_rtx, at rtl.c:358)

[vmakarov at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69030

--- Comment #5 from Vladimir Makarov <vmakarov at gcc dot gnu.org> ---
(In reply to Vladimir Makarov from comment #4)
> I am working on the
> problem and hope to have a fix by Monday.
> 
> 

The problem is more severe than I thought.  It is not only choice between
removed debug info (removing some debug insns in bery rare cases) or wrong
debug info (assigning memory to pseudo living only in debug insns).  The two
different solutions I have now unfortunately affect fundamental assumption that
code with debug and without debug info should be the same: I have 2 new GCC
test regressions when -fcompare-debug is used.

So fixing it will take more time than I evaluated before.