On Wed, Mar 28, 2018 at 09:32:40AM +0300, Leon Romanovsky wrote:
> From: Leon Romanovsky <leonro@xxxxxxxxxxxx>
>
> Ensure that user provides valid AF family prior to query device address.
>
> Reported-by: <syzbot+2a2c48fc189ed5125b9c@xxxxxxxxxxxxxxxxxxxxxxxxx>
> Fixes: ee7aed4528fb ("RDMA/ucma: Support querying for AF_IB addresses")
> Signed-off-by: Leon Romanovsky <leonro@xxxxxxxxxxxx>
> ---
> Changelog v0 -> v1:
> * Fix wrongly placed memcpy
> ---
> drivers/infiniband/core/ucma.c | 6 ++++++
> 1 file changed, 6 insertions(+)
I don't think we should apply this, it doesn't seem to fix any
syzkaller bug.
I'm scared this will break a legitimate user space flow, since it
converts something that used to return success with a zero sa_family
into a failure.
Particularly I see in rdma-core that this flow is called after doing
listen, and I'm not confident that dst_addr is valid after listen.
There also seems to be nothing wrong with returning a 0 address API
wise, so I prefer to leave it alone.
Jason
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
