On Wed, Jul 19, 2017 at 11:07:19AM +0800, JingPiao Chen wrote:
> Extended ACK reporting introduced by linux kernel commit
> v4.11-rc5-1382-g2d4bc93.
>
> * netlink.h (NLM_F_CAPPED): New macro.
> * netlink.c (decode_payload): Pass
> nlmsghdr->nlmsg_flags & NLM_F_CAPPED to decode_nlmsgerr.
> (decode_nlmsgerr): Adjust the length pass to
> decode_nlmsghdr_with_payload.
> ---
> netlink.c | 14 +++++++++++---
> netlink.h | 4 ++++
> 2 files changed, 15 insertions(+), 3 deletions(-)
>
> diff --git a/netlink.c b/netlink.c
> index d3ad8b0..122422b 100644
> --- a/netlink.c
> +++ b/netlink.c
> @@ -297,7 +297,8 @@ decode_nlmsgerr(struct tcb *const tcp,
> const int fd,
> const int family,
> kernel_ulong_t addr,
> - kernel_ulong_t len)
> + kernel_ulong_t len,
> + const bool capped)
> {
> struct nlmsgerr err;
>
> @@ -320,10 +321,16 @@ decode_nlmsgerr(struct tcb *const tcp,
> len -= offsetof(struct nlmsgerr, msg);
>
> if (len) {
> + unsigned int payload = len;
> +
> tprints(", msg=");
> if (fetch_nlmsghdr(tcp, &err.msg, addr, len)) {
> + payload = capped ? sizeof(err.msg) : err.msg.nlmsg_len;
Subsequent decode_nlmsghdr_with_payload call already does all necessary
checks, why do you bother with err.msg.nlmsg_len here?
> + if (payload > len)
> + payload = len;
Consider this instead:
const unsigned int payload =
(capped && sizeof(err.msg) < len) ? sizeof(err.msg) :
len;
> +
> decode_nlmsghdr_with_payload(tcp, fd, family,
> - &err.msg, addr, len);
> + &err.msg, addr, payload);
--
ldv
signature.asc
Description: PGP signature
------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________ Strace-devel mailing list Strace-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/strace-devel
