linux-kernel - Re: [PATCH] getdents.2: Use 'ssize

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-ID: <1195fda6-4de3-09fc-8653-42655d7e2b76@gmail.com>
Date:   Thu, 29 Oct 2020 12:02:44 +0100
From:   Alejandro Colomar <colomar.6.4.3@...il.com>
To:     "Michael Kerrisk (man-pages)" <mtk.manpages@...il.com>
Cc:     linux-man@...r.kernel.org,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "libc-alpha@...rceware.org" <libc-alpha@...rceware.org>
Subject: Re: [PATCH] getdents.2: Use 'ssize_t' instead of 'int'

[[ CC += linux-kernel, libc-alpha ]]

Hi Michael,

On 2020-10-29 08:13, Michael Kerrisk (man-pages) wrote:
 > Hi Alex,
 >
 > On 10/28/20 11:11 PM, Alejandro Colomar wrote:
 >> The glibc wrapper for getdents64() uses ssize_t.
 >
 > It also changed the types for the arguments, so those need to be
 > fixed too.

I saw that, but I sent an email to glibc asking if it was a bug.
I'll hold this patch, and I'll CC this list in that other thread.

 >
 >> And let's use it also for getdents().
 >
 > I actually think we should *not* change that. So long as their is
 > no wrapper, we should show pretty much what the ABI exposes. (That
 > makes me think that the return type should really be long; see
 > what you think about DEFINE_SYSCALL3 in the kernel sources.)

Agreed; then getdents() (and all syscalls without a wrapper)
should use 'long', as you said.

BTW: My mind explodes when I try to read through SYSCALL_DEFINEx :)

 >
 > And you added an include for <sys/types.h>. I'm not sure
 > whether that's needed, but it should be explained in the
 > commit message.

Agreed.
I should be able to check if it's needed, in ssize_t(3) :p
... Actually it's not needed (<stdio.h> already provides it).

Thanks,

Alex