This is the mail archive of the
mailing list for the GCC project.
Re: [patch] Record REG_POINTER in more cases
- From: Ulrich Weigand <weigand at i1 dot informatik dot uni-erlangen dot de>
- To: law at redhat dot com
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Mon, 22 Mar 2004 23:59:35 +0100 (CET)
- Subject: Re: [patch] Record REG_POINTER in more cases
Jeff Law wrote:
>That works on most architectures -- but the PA's braindamaged notion that
>segment selection should be from the base register only rather than the
>effective address means that such code would be unsafe. One of the mn10x
>targets had this insanity as well IIRC.
In one specific mode of operation on the S/390, it is also the case
that the particular base register used in an address selects the
address space the address refers to. (Note that we are talking full
address spaces, not segments, here -- thus it wouldn't even be possible
to make the selection based on the effective address.)
While Linux currently doesn't support that mode (access register mode),
we would still like to have precise REG_POINTER semantics, because we
can use LOAD ADDRESS as addition instruction only if we know we're
adding 31-bit values (i.e. pointers), not full 32-bit values ...
Dr. Ulrich Weigand