From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 15688 invoked by alias); 9 Apr 2003 16:57:34 -0000 Mailing-List: contact gcc-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Archive: List-Post: List-Help: Sender: gcc-owner@gcc.gnu.org Received: (qmail 15680 invoked from network); 9 Apr 2003 16:57:33 -0000 Received: from unknown (HELO mail.libertysurf.net) (213.36.80.91) by sources.redhat.com with SMTP; 9 Apr 2003 16:57:33 -0000 Received: from localhost.localdomain (212.83.141.69) by mail.libertysurf.net (6.5.026) id 3E8A4CE60015418C; Wed, 9 Apr 2003 18:57:30 +0200 Content-Type: text/plain; charset="iso-8859-1" From: Eric Botcazou To: Jan Hubicka Subject: Re: Reload bug Date: Wed, 09 Apr 2003 18:10:00 -0000 User-Agent: KMail/1.4.3 Cc: gcc@gcc.gnu.org References: <200304081937.16859.ebotcazou@libertysurf.fr> <200304091105.24896.ebotcazou@libertysurf.fr> <20030409112544.GA12755@kam.mff.cuni.cz> In-Reply-To: <20030409112544.GA12755@kam.mff.cuni.cz> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Message-Id: <200304091847.23040.ebotcazou@libertysurf.fr> X-SW-Source: 2003-04/txt/msg00372.txt.bz2 > Concerning this testcase, the reload used to manage to get around this > problem. It didn't get around intentionally and just because the > instruction contrain does not accept SSE register so reload is needed > anyway. The idea of reloading whole SSE register into integer is no-go > as can be see from the other testcases. There are two cases when the > reload is really needed: > 1) we are not able to access the SUBREG_BYTE part because of > the problem above because SUBREG_BYTE is not divisible by size of the > register (there is no way to get the size of the register, but I guess > we can use GET_MODE_SIZE / HARD_REGNO_NREGS) and we will also need to > compensate the lowparts, > > This is the problem we are seeing. It is IMO questionable whether it > is valid to see such scenario and whether local alloc can produce > such a subregs that has no representation. My CANNOT_CHANGE_MODE_P > patch avoids that as it causes the problem described above. > > 2) We are outputing into the register and the HARD_REGNO_NREGS is > decreasing so we may convert subreg rewriting just part of the > register to the subreg rewriting the lower part and clobbering upper > part. This is what I beleive the conditional was originally invented > for. > > Are we in the sync now? I don't fully understand the second point. Could you give an example? > It seems to be that the conditional is still overconservative for 2) as > it is still possible that the subreg is large enought to fit into > HARD_REGNO_NREGS of the destination and it now ignores 1). It is > question whether 1) can legaly happen at all but I see it does for 3.2 > (and does not for 3.3). Are you sure of that for 3.3? Because we will silently miscompile too if there are a few leaks. -- Eric Botcazou