mailing list of musl libc
 help / color / mirror / code / Atom feed
From: Petr Skocik <pskocik@gmail.com>
To: Alexander Monakov <amonakov@ispras.ru>, musl@lists.openwall.com
Subject: Re: [musl] Musl's FD_{SET,ISSET,CLR} macros from sys/select.h trigger gcc's -Wsign-conversion warnings
Date: Sun, 2 Aug 2020 21:43:56 +0200	[thread overview]
Message-ID: <9d46d00c-9a67-30eb-d60e-8bd3b0661178@gmail.com> (raw)
In-Reply-To: <alpine.LNX.2.20.13.2008022148160.2454@monopod.intra.ispras.ru>

Yeah, you're right. The signedness does make the codegen worse because of the signed division (glibc still does the (int) cast of sizeof).

A direct size_t cast of (d) could potentially do some type-unsafe things (like convert a pointer), but I guess you could
do something like `(size_t)(int){d}` or use an inline function to silence the warnings without worsening the codegen and without worsening the typesafety of the macro.

I don't super care about this in Musl as it's quite easily silencable with an `unsigned` cast from the user-side but Cygwin had the macros written in such a way that the `unsigned` cast on the descriptor argument wouldn't help, so I reported it there and also sent one bug report to Musl while I was at it, but I will the -Wsystem-headers problem to gcc.

Thanks for the suggestion, Alexander Monakov.

Regards,
Petr Skocik


On 8/2/20 8:53 PM, Alexander Monakov wrote:
> On Sun, 2 Aug 2020, Petr Skocik wrote:
>
>>     #define FD_SET(d, s)   ((s)->fds_bits[(d)/(8*(int)sizeof(long))] |=
>> (1UL<<((d)%(8*(int)sizeof(long)))))
>>     #define FD_CLR(d, s)   ((s)->fds_bits[(d)/(8*(int)sizeof(long))] &=
>> ~(1UL<<((d)%(8*(int)sizeof(long)))))
>>     #define FD_ISSET(d, s) !!((s)->fds_bits[(d)/(8*(int)sizeof(long))] &
>> (1UL<<((d)%(8*(int)sizeof(long)))))
>>
>> You might want to add them.
> (casting 'd' to size_t would have been more appropriate, as there's no need
> to perform signed division and modulus here)
>
> This is one of the cases where the warning should have been suppressed by
> GCC unless -Wsystem-headers is also given: the problem appears when expanding
> a macro defined in a system header, so the user can't cleanly avoid it.
>
> Would you care to open an issue in the GCC Bugzilla about it?
>
> Alexander
>
>
>

  parent reply	other threads:[~2020-08-02 20:06 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-08-02 18:24 Petr Skocik
2020-08-02 18:53 ` Alexander Monakov
2020-08-02 18:56   ` Alexander Monakov
2020-08-02 19:43   ` Petr Skocik [this message]
2020-08-03  5:43     ` Markus Wichmann
2020-08-03 17:47       ` Rich Felker

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=9d46d00c-9a67-30eb-d60e-8bd3b0661178@gmail.com \
    --to=pskocik@gmail.com \
    --cc=amonakov@ispras.ru \
    --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).