linux-kernel - RE: [PATCH next resend 2/5] minmax: Allow min()/max()/clamp() if the arguments have the same signedness.

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>] [day] [month] [year] [list]

Message-ID: <4d15014d6b4f434786d50bc4f2199f23@AcuMS.aculab.com>
Date:   Wed, 26 Jul 2023 10:25:11 +0000
From:   David Laight <David.Laight@...LAB.COM>
To:     'Marc Zyngier' <maz@...nel.org>
CC:     'kernel test robot' <lkp@...el.com>,
        "'linux-kernel@...r.kernel.org'" <linux-kernel@...r.kernel.org>,
        'Andy Shevchenko' <andriy.shevchenko@...ux.intel.com>,
        'Andrew Morton' <akpm@...ux-foundation.org>,
        "'Matthew Wilcox (Oracle)'" <willy@...radead.org>,
        'Christoph Hellwig' <hch@...radead.org>,
        "'Jason A. Donenfeld'" <Jason@...c4.com>,
        "llvm@...ts.linux.dev" <llvm@...ts.linux.dev>,
        "oe-kbuild-all@...ts.linux.dev" <oe-kbuild-all@...ts.linux.dev>,
        "Linux Memory Management List" <linux-mm@...ck.org>,
        Rob Herring <robh@...nel.org>
Subject: RE: [PATCH next resend 2/5] minmax: Allow min()/max()/clamp() if the
 arguments have the same signedness.

From: Marc Zyngier
> Sent: 26 July 2023 10:51
> 
> On Wed, 26 Jul 2023 10:19:48 +0100,
> David Laight <David.Laight@...LAB.COM> wrote:
> >
> > > From: kernel test robot <lkp@...el.com>
> > > Sent: 25 July 2023 19:33
> > ...
> > >
> > > All errors (new ones prefixed by >>):
> > >
> > > >> drivers/irqchip/irq-mips-cpu.c:288:1: error: call to undeclared function '__typecheck'; ISO C99
> and
> > > later do not support implicit function declarations [-Wimplicit-function-declaration]
> > >      288 | IRQCHIP_DECLARE(cpu_intc, "mti,cpu-interrupt-controller", mips_cpu_irq_of_init);
> > >          | ^
> > >    include/linux/irqchip.h:37:38: note: expanded from macro 'IRQCHIP_DECLARE'
> > >       37 |         OF_DECLARE_2(irqchip, name, compat, typecheck_irq_init_cb(fn))
> > >          |                                             ^
> > >    include/linux/irqchip.h:24:3: note: expanded from macro 'typecheck_irq_init_cb'
> > >       24 |         (__typecheck(typecheck_irq_init_cb, &fn) ? fn : fn)
> > >          |          ^
> > > >> drivers/irqchip/irq-mips-cpu.c:288:1: error: initializer element is not a compile-time constant
> > >      288 | IRQCHIP_DECLARE(cpu_intc, "mti,cpu-interrupt-controller", mips_cpu_irq_of_init);
> > >          | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > >    include/linux/irqchip.h:37:2: note: expanded from macro 'IRQCHIP_DECLARE'
> > >       37 |         OF_DECLARE_2(irqchip, name, compat, typecheck_irq_init_cb(fn))
> > >          |         ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > >    include/linux/of.h:1493:3: note: expanded from macro 'OF_DECLARE_2'
> > >     1493 |                 _OF_DECLARE(table, name, compat, fn, of_init_fn_2)
> > >          |                 ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > >    include/linux/of.h:1481:2: note: expanded from macro '_OF_DECLARE'
> > >     1481 |         _OF_DECLARE_STUB(table, name, compat, fn, fn_type)
> > >          |         ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > >    include/linux/of.h:1470:16: note: expanded from macro '_OF_DECLARE_STUB'
> > >     1470 |                      .data = (fn == (fn_type)NULL) ? fn : fn }
> > >          |                              ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> >
> > It looks like irqchip.h is using __typecheck() which is really
> > an internal part of the implementation of min() and max().
> > The patched version doesn't use it - hence the build fail.
> > I can re-instate it, but this all looks wrong to me.
> 
> Please see f1985002839a ("irqchip: Provide stronger type checking for
> IRQCHIP_MATCH/IRQCHIP_DECLARE") for the rationale.
> 
> Given that this has uncovered a number of bugs, I'm not letting this
> go without an equivalent replacement.
..
> 
> They are used in different contexts. See IRQCHIP_MATCH().

Ah, I was seeing the error in the expansion of IRQCHIP_DECLARE()
which is doing the type check twice.
It can just pass 'fn'.

The cast NULL check does work.
So IRQCHIP_MATCH() can use the simpler 'fn = (fn_type)NULL ? fn : fn' test
that _OF_DECLARE_STUB() uses.

	David

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