This is the mail archive of the gcc-patches@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]
Other format:	[Raw text]

Re: PATCH to expand_builtin_setjmp_receiver for middle-end/28493

From: Paolo Bonzini <paolo dot bonzini at lu dot unisi dot ch>
To: Jason Merrill <jason at redhat dot com>, GCC Patches <gcc-patches at gcc dot gnu dot org>
Date: Wed, 13 Sep 2006 11:42:01 +0200
Subject: Re: PATCH to expand_builtin_setjmp_receiver for middle-end/28493
References: <4506F4C8.4080600@redhat.com>

Jason Merrill wrote:

The problem in 28493 turned out to be that we had
  (set virtual-stack-vars hard-frame-pointer)
  ...
  (compare virtual-stack-vars ptr)
devirtualization turned this into
  (set soft-frame-pointer (minus hard-frame-pointer 8))
  ...
  (set tmp (plus soft-frame-pointer 8))
  (compare tmp ptr)
[...]

and things break. Fixed by forcing a clobber of the hard frame pointer so that the soft/hard duality doesn't confuse CSE.

With much handwaving, I wonder if the clobber of the hard frame pointer shouldn't be emitted by devirtualization whenever it sets the soft frame pointer.

Paolo

References:
- PATCH to expand_builtin_setjmp_receiver for middle-end/28493
  - From: Jason Merrill

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