On 23 August 2016 at 17:05, Joe Perches <j...@perches.com> wrote:
> On Tue, 2016-08-23 at 07:21 -0700, Eric Dumazet wrote:
>> On Tue, 2016-08-23 at 14:41 +0100, Luis Henriques wrote:
>> > From: Avijit Kanti Das <avijitn...@codeaurora.org>
>> >
>> > memset() the structure ethtool_wolinfo that has padded bytes
>> > but the padded bytes have not been zeroed out.
> []
>> > diff --git a/net/core/ethtool.c b/net/core/ethtool.c
> []
>> > @@ -1435,11 +1435,13 @@ static int ethtool_reset(struct net_device *dev,
>> > char __user *useraddr)
>> >
>> > static int ethtool_get_wol(struct net_device *dev, char __user *useraddr)
>> > {
>> > - struct ethtool_wolinfo wol = { .cmd = ETHTOOL_GWOL };
>> > + struct ethtool_wolinfo wol;
>> >
>> > if (!dev->ethtool_ops->get_wol)
>> > return -EOPNOTSUPP;
>> >
>> > + memset(&wol, 0, sizeof(struct ethtool_wolinfo));
>> > + wol.cmd = ETHTOOL_GWOL;
>> > dev->ethtool_ops->get_wol(dev, &wol);
>> >
>> > if (copy_to_user(useraddr, &wol, sizeof(wol)))
>> This would suggest a compiler bug to me.
>
> A compiler does not have a standards based requirement to
> initialize arbitrary padding bytes.
>
> I believe gcc always does zero all padding anyway.
>
>> I checked that my compiler does properly put zeros there, even in the
>> padding area.
>>
>> If we can not rely on such constructs, we have hundreds of similar
>> patches to submit.
>
> True.
>
> From a practical point of view, does any compiler used for
> kernel compilation (gcc/icc/llvm/any others?) not always
> perform zero padding of alignment bytes?
>
gcc often does not do it, depends on a few factors though:
https://lkml.org/lkml/2016/5/20/389
Vegard
