On Tue, 11 Feb 2025 20:59:24 +0200 Gal Pressman wrote:
> > Everything else looks very good, though, yes, I would agree with the
> > alignment comments made down-thread. This was "accidentally correct"
> > before in the sense that the end of the struct would be padded for
> > alignment, but isn't gu
On 11/02/2025 19:49, Kees Cook wrote:
>> @@ -659,7 +654,7 @@ static inline void ip_tunnel_info_opts_set(struct
>> ip_tunnel_info *info,
>> {
>> info->options_len = len;
>> if (len > 0) {
>> -memcpy(ip_tunnel_info_opts(info), from, len);
>> +memcpy(info->options,
On Sun, Feb 09, 2025 at 12:18:53PM +0200, Gal Pressman wrote:
> Remove the hidden assumption that options are allocated at the end of
> the struct, and teach the compiler about them using a flexible array.
>
> With this, we can revert the unsafe_memcpy() call we have in
> tun_dst_unclone() [1], an
On 09/02/2025 22:16, Ilya Maximets wrote:
> Ideally we would have a proper union with all the potential option types
> instead of this hacky construct. But if that's not the the way to go, then
> 8 bytes may indeed be the way, as it is the maximum guaranteed alignment
> for allocations and the cur
On 2/9/25 20:37, Gal Pressman wrote:
> Hi Ilya, thanks for the review.
>
> On 09/02/2025 18:21, Ilya Maximets wrote:
>> On 2/9/25 11:18, Gal Pressman via dev wrote:
>>> Remove the hidden assumption that options are allocated at the end of
>>> the struct, and teach the compiler about them using a f
Hi Ilya, thanks for the review.
On 09/02/2025 18:21, Ilya Maximets wrote:
> On 2/9/25 11:18, Gal Pressman via dev wrote:
>> Remove the hidden assumption that options are allocated at the end of
>> the struct, and teach the compiler about them using a flexible array.
>>
>> With this, we can revert
On 2/9/25 11:18, Gal Pressman via dev wrote:
> Remove the hidden assumption that options are allocated at the end of
> the struct, and teach the compiler about them using a flexible array.
>
> With this, we can revert the unsafe_memcpy() call we have in
> tun_dst_unclone() [1], and resolve the fal
Remove the hidden assumption that options are allocated at the end of
the struct, and teach the compiler about them using a flexible array.
With this, we can revert the unsafe_memcpy() call we have in
tun_dst_unclone() [1], and resolve the false field-spanning write
warning caused by the memcpy()
