Does ada work on x32 currently (gcc 5.x)?

Does ada work on x32 currently (gcc 5.x)?

[Giuseppe Caizzone]

Hello,

I've been unable to bootstrap gcc 5.[12], starting from gcc 4.9.2, if
I configure it with --with-abi=mx32 and having "ada" among the values
for --enable-languages; searching the web I've found some old mailing
list messages about ada not working on x32 (however gcc 4.9.x at least
did complete the bootstrap under the same conditions). Is it still the
case that ada isn't supported on x32 (and therefore I should drop ada
from the language list for x32) or it's supposed to work instead (and
therefore I can invest some more time finding out what I'm doing
wrong)?

Thanks,
Giuseppe

(In case it's of interest, what *I think* happens on my particular
machine is that the stage2 gnat1 binary fails with EFAULT when ran
during the bootstrap because it passes a wrong pointer to the write
syscall; the pointer has in its lower 32 bits the correct address it's
supposed to have, but other stuff in the upper bits. The stage1
binary, instead, is fine in this respect. Also, gcc 4.9.2 does
bootstrap successfully on the same box with the same configure
parameters. I don't know if this is due to a bug or rather a
misconfiguration of my system.)

Re: Does ada work on x32 currently (gcc 5.x)?

[Ali Abdul Ghani]

its same for me

2015-07-25 0:48 جرينتش-07:00, Giuseppe Caizzone <acaizzo@gmail.com>:
> Hello,
>
> I've been unable to bootstrap gcc 5.[12], starting from gcc 4.9.2, if
> I configure it with --with-abi=mx32 and having "ada" among the values
> for --enable-languages; searching the web I've found some old mailing
> list messages about ada not working on x32 (however gcc 4.9.x at least
> did complete the bootstrap under the same conditions). Is it still the
> case that ada isn't supported on x32 (and therefore I should drop ada
> from the language list for x32) or it's supposed to work instead (and
> therefore I can invest some more time finding out what I'm doing
> wrong)?
>
> Thanks,
> Giuseppe
>
> (In case it's of interest, what *I think* happens on my particular
> machine is that the stage2 gnat1 binary fails with EFAULT when ran
> during the bootstrap because it passes a wrong pointer to the write
> syscall; the pointer has in its lower 32 bits the correct address it's
> supposed to have, but other stuff in the upper bits. The stage1
> binary, instead, is fine in this respect. Also, gcc 4.9.2 does
> bootstrap successfully on the same box with the same configure
> parameters. I don't know if this is due to a bug or rather a
> misconfiguration of my system.)
>


-- 
Think not of them, thou hast thy music too