public inbox for gcc@gcc.gnu.org
 help / color / mirror / Atom feed
* Re: aix4.1 alias problem
@ 1997-11-06  1:06 Christian Iseli
  0 siblings, 0 replies; only message in thread
From: Christian Iseli @ 1997-11-06  1:06 UTC (permalink / raw)
  To: law, mengel; +Cc: egcs

> Well, it's the same basic problem as Christian ran into, except
> that the scheduler has moved things around in such a way that
> the copy out of the arg register appears after the function begin
> note.
>
> ...
> 
> The other possibility is to set reg_seen for all the registers
> which might hold values at function entry.  This might be overly
> pessimistic and defeat the entire purpose of tracking argument
> values for alias purposes.  But should be 100% safe.

How about trying to do some clever walking through the reg_base_value
table, detecting cycles after the first loop iteration, and avoiding
to cycle during subsequent iterations of the loop?

					Christian

^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~1997-11-06  1:06 UTC | newest]

Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
1997-11-06  1:06 aix4.1 alias problem Christian Iseli

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).