This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: is subreg with subreg_byte outside source valid?


> >>>>> Jan Hubicka writes:
> 
> Jan> I am just investigating the second failure in powerpc testsuite.
> 
> 	Is this a failure in the GCC trunk?  I am not sure that I can help
Yes, it is. It was introduced by me, sadly. I don't think I will have
problems to fix it soon, as this design issue gets resolved.
> with your question, but I would like to understand the context because it
I've reorganized subreg simplification code and added some sanity checking.
The checking alarms on the subreg, as I believe it is not correct.

If you remove offending abort, the test may pass, but this is of course
not the way to go officially.

If you are familiar with the powerpc port, you may help me with the
last failure. I am getting crash on outputting insn:
(insn:TI 35 8 15 (set (reg:QI 9 r9)
        (mem/s/u:QI (plus:SI (lo_sum:SI (reg/f:SI 9 r9 [84])
                    (symbol_ref:SI ("*.LC0")))
                (const_int 1 [0x1])) 2)) 289 {*rs6000.md:7526} (insn_list 8 (nil))
    (nil))

it looks like the mem expression is wrong. Do you have some idea what
goes wrong and how should look the correct expression? I don't have any
idea how my patch can introduce such a extraordinary crash so late in
the way, as all subregs disappear after reload and compiler should crash
if insn were invalid.

Honza
> affects the PowerPC port.
> 
> Thanks, David


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]