From: Michael Veksler <VEKSLER@il.ibm.com>
To: Sebastian Pop <sebastian.pop@cri.ensmp.fr>
Cc: Robert Dewar <dewar@adacore.com>,
gcc@gcc.gnu.org, Kai Henningsen <kaih@khms.westfalen.de>
Subject: Re: basic VRP min/max range overflow question
Date: Tue, 19 Jul 2005 07:07:00 -0000 [thread overview]
Message-ID: <OF42B07BFE.4BCEC179-ON43257043.00260CD6-43257043.00271359@il.ibm.com> (raw)
In-Reply-To: <20050719064910.GA4849@napoca.cri.ensmp.fr>
Sebastian Pop wrote on 19/07/2005 09:49:11:
> Robert Dewar wrote:
> >
> > and that is called a false positive if in fact the loop does
> > not overrun. this sounds very dubious to me
I concur.
>
> The problem is that the compiler has no other information about the
> number of iterations in the loop, otherwise it wouldn't spend cycles
> on computing such estimations. Because the compiler will use the
> estimation for transforming the code, it should warn the user.
But isn't it only an estimation? In other words, does it affect anything
but performance (in case of a wrong estimation)?
>
> If I understand correctly, you would like a warning that says: "your
> code is wrong: you access the array outside the allocated size".
> Well, this is feasible, but then, the warning will be effective only
> on loops that have statically determinable number of iterations.
>
> In my opinion both these warnings are useful.
>
I am not sure about it. If you have false positives that you have
no reasonable way to turn off (other than -Wno-xxxxx), then
the warning is useless. The only alternative for -Wno-xxxxx
in your example would be:
void foo (int N)
{
int i;
int A[123];
if (N > 123) // Turn the warning off.
abort();
for (i = 0; i < N; i++)
A[i] = ...
}
This is problematic:
1. I am not sure it will turn the warning off.
2. Even if it does, this defeats the whole idea of the original
optimization (gcc could have put the abort - it is allowed
to do so because N>123 triggers undefined behavior).
3. If the line with A[i]=.... contains a function call then it is
possible that the function calls exit, abort, longjmp or throw.
In that case, N>123 is perfectly legal (and my if..abort
is not viable - so again there is no way to shut the warning
off).
I guess that VRP is not allowed to propagate N <= 123 in
this case.
Michael
next prev parent reply other threads:[~2005-07-19 7:07 UTC|newest]
Thread overview: 70+ messages / expand[flat|nested] mbox.gz Atom feed top
2005-06-18 14:15 Paul Schlie
2005-06-18 16:19 ` Joseph S. Myers
2005-06-18 17:44 ` Paul Schlie
2005-06-18 18:05 ` Paul Schlie
2005-06-18 18:21 ` Joseph S. Myers
2005-06-18 18:50 ` Paul Schlie
2005-06-18 19:57 ` Joseph S. Myers
2005-06-18 21:01 ` Florian Weimer
2005-06-19 10:18 ` Kai Henningsen
2005-06-20 22:20 ` Mike Stump
2005-06-18 19:16 ` Paul Schlie
2005-06-18 21:26 ` Robert Dewar
2005-06-19 18:23 ` Paul Schlie
2005-06-20 2:44 ` Robert Dewar
2005-06-20 5:55 ` Paul Schlie
2005-06-20 10:14 ` Michael Veksler
2005-06-20 11:04 ` Paul Schlie
2005-06-20 11:32 ` Robert Dewar
2005-06-20 12:08 ` Paul Schlie
2005-06-20 12:39 ` Robert Dewar
2005-06-20 13:01 ` Paul Schlie
2005-06-20 13:14 ` Robert Dewar
2005-06-20 12:23 ` Michael Veksler
2005-06-20 12:46 ` Robert Dewar
2005-06-20 10:43 ` Robert Dewar
2005-06-20 10:54 ` Robert Dewar
2005-06-20 11:26 ` Paul Schlie
2005-06-20 11:34 ` Robert Dewar
2005-06-20 12:09 ` Paul Schlie
2005-06-20 11:14 ` Paul Schlie
2005-06-20 12:53 ` Michael Veksler
2005-06-20 12:59 ` Robert Dewar
2005-06-20 13:09 ` Paul Schlie
2005-06-20 13:17 ` Robert Dewar
2005-06-20 14:54 ` Michael Veksler
2005-06-20 18:01 ` Paul Schlie
2005-06-18 18:08 ` Joseph S. Myers
2005-06-18 21:08 ` Robert Dewar
2005-06-19 10:18 ` Kai Henningsen
2005-06-19 10:49 ` Robert Dewar
2005-06-20 13:22 ` Sebastian Pop
2005-06-20 18:10 ` DJ Delorie
2005-07-13 7:57 ` Sebastian Pop
2005-07-13 15:18 ` DJ Delorie
2005-06-21 15:21 ` Robert Dewar
2005-07-18 16:34 ` Sebastian Pop
2005-07-18 16:44 ` Robert Dewar
2005-07-19 6:44 ` Sebastian Pop
2005-07-19 7:07 ` Michael Veksler [this message]
2005-07-19 9:55 ` Sebastian Pop
2005-07-19 10:22 ` Michael Veksler
2005-06-20 19:53 ` Kai Henningsen
2005-06-18 20:55 ` Robert Dewar
2005-06-18 22:45 ` Tristan Wibberley
2005-06-20 21:47 ` Mike Stump
-- strict thread matches above, loose matches on Subject: below --
2005-06-17 12:00 Paul Schlie
2005-06-17 12:09 ` Paolo Bonzini
2005-06-17 18:29 ` Paul Schlie
2005-06-17 22:09 ` Paolo Bonzini
2005-06-17 22:48 ` Diego Novillo
2005-06-18 0:20 ` Paul Schlie
2005-06-18 0:23 ` Andrew Pinski
2005-06-18 0:59 ` Paul Schlie
2005-06-18 1:10 ` Dale Johannesen
2005-06-18 4:09 ` Mike Stump
[not found] ` <25364524.1119085038744.JavaMail.root@dtm1eusosrv72.dtm.ops.eu.uu.net>
2005-06-18 11:47 ` Toon Moene
2005-06-18 2:02 ` Robert Dewar
2005-06-17 4:28 Paul Schlie
2005-06-17 7:03 ` Paolo Bonzini
2005-06-17 12:28 ` Diego Novillo
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=OF42B07BFE.4BCEC179-ON43257043.00260CD6-43257043.00271359@il.ibm.com \
--to=veksler@il.ibm.com \
--cc=dewar@adacore.com \
--cc=gcc@gcc.gnu.org \
--cc=kaih@khms.westfalen.de \
--cc=sebastian.pop@cri.ensmp.fr \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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).