[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <201312052131.42316.arnd@arndb.de>
Date: Thu, 5 Dec 2013 21:31:42 +0100
From: Arnd Bergmann <arnd@...db.de>
To: haver@...ux.vnet.ibm.com
Cc: linux-kernel@...r.kernel.org, gregkh@...uxfoundation.org,
cody@...ux.vnet.ibm.com, schwidefsky@...ibm.com,
utz.bacher@...ibm.com, mmarek@...e.cz, rmallon@...il.com,
jsvogt@...ibm.com, MIJUNG@...ibm.com, cascardo@...ux.vnet.ibm.com,
michael@...ra.de
Subject: Re: [PATCH 1/6] GenWQE PCI support, health monitoring and recovery
On Thursday 05 December 2013, Frank Haverkamp wrote:
> > > Was wrong, as already pointed out before. It is now:
> > >
> > > struct genwqe_mem {
> > > __u64 addr;
> > > __u64 size;
> > > int direction;
> > > };
> > >
> > > I hope the int is ok here.
> >
> > No, it's not. The problem is that sizeof(struct genwqe_mem) is now 24 on
> > most architectures (including x86-64) and 20 on x86-32.
>
> Interesting. So int is like long architecture specific. I changed it to
> be __u64 too, to avoid any problem.
The solution is ok, but the problem is different from what you thought:
On all architectures that Linux runs on, 'int' is 32 bit. The problem is
again the alignment of __u64. On normal architectures, it is naturally
aligned, and gcc adds 4 byte padding so that 'sizeof (struct genwqe_mem)'
is multiple of the required alignment. On x86-32, the required alignment
for the __u64 members is only 4 bytes, so no padding is added.
Arnd
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists