From: Rich Felker <dalias@libc.org>
To: musl@lists.openwall.com
Subject: Re: [PATCH] drop unused extra char from getnameinfo() local buffer
Date: Fri, 28 Jun 2019 11:13:50 -0400 [thread overview]
Message-ID: <20190628151350.GE1506@brightrain.aerifal.cx> (raw)
In-Reply-To: <1561683273-18265-1-git-send-email-armccurdy@gmail.com>
On Thu, Jun 27, 2019 at 05:54:33PM -0700, Andre McCurdy wrote:
> The num local buffer is only passed to itoa(), which expects a buffer
> size of 3*sizeof(int), not 3*sizeof(int)+1. Also change the data type
> of the port local variable to clarify that itoa() only handles
> unsigned values.
> ---
> src/network/getnameinfo.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/src/network/getnameinfo.c b/src/network/getnameinfo.c
> index f77e73a..02c2c09 100644
> --- a/src/network/getnameinfo.c
> +++ b/src/network/getnameinfo.c
> @@ -124,7 +124,7 @@ int getnameinfo(const struct sockaddr *restrict sa, socklen_t sl,
> int flags)
> {
> char ptr[PTR_MAX];
> - char buf[256], num[3*sizeof(int)+1];
> + char buf[256], num[3*sizeof(int)];
I think the 3*sizeof(int)+1 idiom is a standard one we use throughout
the code, because it's clearly valid for any size int. It's based
ceil(log10(256))==3 and one byte for termination, and it would
actually be sharp in the pathological case sizeof(int)==1 (which of
course we don't support and can't actually be supported on a hosted
implementation due to stdio constaints).
In practice a constant 11 would work for known-32-bit int, but the
desire here is to be obviously-safe, not to be "optimal".
I think what you have found though is that the expectation in the
definition of itoa is inconsistent. I probably didn't notice the
inconsistency because of the *--p instead of *p. It should be either
*p=0 or p+=3*sizeof(int)+1 initially, I think. Does that sound right?
Either way it's harmless on the only value of sizeof(int) that
actually occurs, but I'd like to fix the inconsistency here.
> int af = sa->sa_family;
> unsigned char *a;
> unsigned scopeid;
> @@ -184,7 +184,7 @@ int getnameinfo(const struct sockaddr *restrict sa, socklen_t sl,
>
> if (serv && servlen) {
> char *p = buf;
> - int port = ntohs(((struct sockaddr_in *)sa)->sin_port);
> + unsigned port = ntohs(((struct sockaddr_in *)sa)->sin_port);
> buf[0] = 0;
> if (!(flags & NI_NUMERICSERV))
> reverse_services(buf, port, flags & NI_DGRAM);
This is ok-ish since it's consistent with the signature for itoa, but
the range of value is such that it can never be negative either way.
Rich
next prev parent reply other threads:[~2019-06-28 15:13 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-06-28 0:54 Andre McCurdy
2019-06-28 15:13 ` Rich Felker [this message]
2019-06-28 19:24 ` Andre McCurdy
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=20190628151350.GE1506@brightrain.aerifal.cx \
--to=dalias@libc.org \
--cc=musl@lists.openwall.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://git.vuxu.org/mirror/musl/
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 NNTP newsgroup(s).