linux-kernel - RE: [PATCH v3] tools/nolibc: fix up size inflate regression

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

Date:   Mon, 7 Aug 2023 09:53:07 +0000
From:   David Laight <David.Laight@...LAB.COM>
To:     David Laight <David.Laight@...LAB.COM>,
        'Zhangjin Wu' <falcon@...ylab.org>, "w@....eu" <w@....eu>
CC:     "arnd@...db.de" <arnd@...db.de>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "linux-kselftest@...r.kernel.org" <linux-kselftest@...r.kernel.org>,
        "thomas@...ch.de" <thomas@...ch.de>
Subject: RE: [PATCH v3] tools/nolibc: fix up size inflate regression

From: David Laight
> Sent: 07 August 2023 09:40
....
> with (retyped so it may be wrong):
> #define is_constexpr(x) sizeof(*(0 ? (void *)((long)(x) * 0) : (int *)0)) == 1)

Bah, I know why that works and I still got is backwards :-(
Basically the compiler needs to find a type that is 'compatible'
with both the possible results.
Since '(void *)0' is a valid 'int *' value that gives 'int *'.
But '(void *)(anything_else)' requires the 'int *' be converted
to 'void *'.

Also the following seems to compile to sane code
for all types on 32bit and 64bit x86.

int errno;

#define MAXERRNO 0x4000
#define type int *

type sysret(type arg)
{
    if ((unsigned long)arg < 0ul - MAXERRNO)
        return arg;

    errno = -(long)arg;
    return (__typeof(arg))-1;        
}

You do get a comparison but no sign extensions.

	David

-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)