linux-kernel - RE: [PATCH] net: via-rhine: fix compiler warning

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]

Message-ID: <063D6719AE5E284EB5DD2968C1650D6D0F702D00@AcuExch.aculab.com>
Date:	Wed, 30 Apr 2014 09:35:28 +0000
From:	David Laight <David.Laight@...LAB.COM>
To:	'Alexey Charkov' <alchark@...il.com>
CC:	Jan Moskyto Matejka <mq@...e.cz>, Roger Luethi <rl@...lgate.ch>,
	"netdev@...r.kernel.org" <netdev@...r.kernel.org>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: RE: [PATCH] net: via-rhine: fix compiler warning

From: Alexey Charkov [mailto:alchark@...il.com]
 
> 2014-04-30 12:49 GMT+04:00 David Laight <David.Laight@...lab.com>:
> > From: Jan Moskyto Matejka
> >> Fixed different size cast warning:
> >>
> >>       drivers/net/ethernet/via/via-rhine.c: In function rhine_init_one_platform:
> >>       drivers/net/ethernet/via/via-rhine.c:1132:13: warning: cast from pointer to integer of different
> >> size [-Wpointer-to-int-cast]
> >>         revision = (u32)match->data;
> >>                    ^
> >>
> >> That code was added in commit 2d283862dc62daead9db0dc89cd0d0351e91f765
> >> ("net: via-rhine: add OF bus binding").
> > ...
> >> diff --git a/drivers/net/ethernet/via/via-rhine.c b/drivers/net/ethernet/via/via-rhine.c
> >> index 4fa9201..76d18e0 100644
> >> --- a/drivers/net/ethernet/via/via-rhine.c
> >> +++ b/drivers/net/ethernet/via/via-rhine.c
> >> @@ -288,7 +288,7 @@ MODULE_DEVICE_TABLE(pci, rhine_pci_tbl);
> >>   * (for quirks etc.)
> >>   */
> >>  static struct of_device_id rhine_of_tbl[] = {
> >> -     { .compatible = "via,vt8500-rhine", .data = (void *)0x84 },
> >> +     { .compatible = "via,vt8500-rhine", .data = (u32 []) { 0x84 } },
> >>       { }     /* terminate list */
> >>  };
> >>  MODULE_DEVICE_TABLE(of, rhine_of_tbl);
> >> @@ -1129,7 +1129,7 @@ static int rhine_init_one_platform(struct platform_device *pdev)
> >>       if (!irq)
> >>               return -EINVAL;
> >>
> >> -     revision = (u32)match->data;
> >> +     revision = *((u32 *) match->data);
> >>       if (!revision)
> >>               return -EINVAL;
> >
> > Both those casts look horrid.
> > I'm not entirely convinced that the first is valid C - It would have to be
> > something specific to C99 initialisers.
> > Casts like *(u32 *)foo are also likely to be bugs (esp. on BE systems)
> > so themselves start ringing alarm bells.
> >
> > So why not just:
> >         revision = (unsigned long)match->data;
> > and add a comment that the 0x84 is the revision - #define ??
> 
> There is no particular reason why it should be u32 now - this is a
> leftover from the previous iteration of code where revision was a
> separate property in DT (sized u32). It actually mirrors the
> respective field in struct pci_dev, which is u8 - don't see any issue
> defining it as unsigned long (and also changing the definition in
> struct rhine_private).

I'd guess that the field in 'struct rhine_private' only needs to
be large enough for the domain of the values that are actually
saved in it.
If these are versions (of something) then a fixed size type
(or just int) is probably more appropriate.

	David