From: Thomas Rast <trast@student.ethz.ch>
To: Erick Mattos <erick.mattos@gmail.com>
Cc: "Junio C Hamano" <gitster@pobox.com>,
git@vger.kernel.org, "Ævar Arnfjörð Bjarmason" <avarab@gmail.com>
Subject: Re: [PATCH next v2] log_ref_setup: don't return stack-allocated array
Date: Thu, 10 Jun 2010 19:29:10 +0200 [thread overview]
Message-ID: <201006101929.11034.trast@student.ethz.ch> (raw)
In-Reply-To: <AANLkTillDOCNQrpaEiFsFdq6HpU_LlwWI2ELIrEcrWHc@mail.gmail.com>
Erick Mattos wrote:
> 2010/6/10 Thomas Rast <trast@student.ethz.ch>
> > -int log_ref_setup(const char *ref_name, char **log_file)
> > +int log_ref_setup(const char *ref_name, char *logfile, int bufsize)
> > {
> > int logfd, oflags = O_APPEND | O_WRONLY;
> > - char logfile[PATH_MAX];
> >
> > - git_snpath(logfile, sizeof(logfile), "logs/%s", ref_name);
> > - *log_file = logfile;
> > + git_snpath(logfile, bufsize, "logs/%s", ref_name);
[...]
> I don't see any improvement here. Unless you want to get rid of using
> references on calling functions which is only going to add another
> buffer to the stack, sized PATH_MAX, once that log_file is going to be
> really allocated in the heap after git_snpath(). As folks use to say
> here: "changing six by half a dozen".
What the - side of the hunk above does is returning a local (stack
allocated) variable, in the form of a pointer to logfile. Once those
go out of scope, you have zero guarantees on what happens with them.
Try the following snippet, it should cause a similar problem:
#include <stdio.h>
int* f()
{
int i;
i = 42;
return &i;
}
int main()
{
int *p = f();
if (1) {
char buf[1024];
memset(buf, 0, sizeof(buf));
}
printf("I got: %d\n", *p);
}
Only in this case the issue is so obvious that the compiler will warn
(at least mine does).
> I haven't ever seen this happening so I think you have found some
> particularity of valgrind which could route a patch to it.
Admittedly my experience is somewhat limited since I don't do C coding
outside of git and some teaching. But so far I have not had a single
false alarm with valgrind (when compiled without optimizations;
otherwise the compiler may do some magic).
--
Thomas Rast
trast@{inf,student}.ethz.ch
next prev parent reply other threads:[~2010-06-10 17:29 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-06-10 12:43 [PATCH next] log_ref_setup: don't return stack-allocated array Thomas Rast
2010-06-10 12:54 ` [PATCH next v2] " Thomas Rast
2010-06-10 16:48 ` Erick Mattos
2010-06-10 17:29 ` Thomas Rast [this message]
2010-06-10 23:09 ` Erick Mattos
2010-06-11 5:12 ` Jeff King
2010-06-11 18:54 ` Erick Mattos
[not found] ` <AANLkTinI44rPfeXvWr-7jvAVyw5itX_gUsHimwSL74Lv@mail.gmail.com>
2010-06-10 18:09 ` Ævar Arnfjörð Bjarmason
2010-06-10 18:43 ` [PATCH] check_aliased_update: strcpy() instead of strcat() to copy Thomas Rast
2010-06-10 19:00 ` Ævar Arnfjörð Bjarmason
2010-06-10 19:26 ` Jay Soffian
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
List information: http://vger.kernel.org/majordomo-info.html
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=201006101929.11034.trast@student.ethz.ch \
--to=trast@student.ethz.ch \
--cc=avarab@gmail.com \
--cc=erick.mattos@gmail.com \
--cc=git@vger.kernel.org \
--cc=gitster@pobox.com \
/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.
Code repositories for project(s) associated with this public inbox
https://80x24.org/mirrors/git.git
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).