On Mon, Apr 30, 2018 at 11:35 PM, kefu chai <tchaikov@xxxxxxxxx> wrote: > On Mon, Apr 30, 2018 at 8:12 PM, Jeff Layton <jlayton@xxxxxxxxxx> wrote: >> On Mon, 2018-04-30 at 07:23 -0400, Jeff Layton wrote: >>> I recently upgraded my main workstation to Fedora 28, and I've been >>> unable to build ceph under it. It looks like the main problem is new >>> compiler warnings that have cropped up, and the fact that DPDK builds >>> with -Werror: >>> >>> Here's one that looks like an actual bug, fwiw: >>> >>> /home/jlayton/git/ceph-build/src/spdk/dpdk/lib/librte_metrics/rte_metrics.c: In function ‘rte_metrics_get_names’: >>> /home/jlayton/git/ceph-build/src/spdk/dpdk/lib/librte_metrics/rte_metrics.c:247:4: error: ‘strncpy’ specified bound 64 equals destination size [-Werror=stringop-truncation] >>> strncpy(names[idx_name].name, >>> ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~ >>> stats->metadata[idx_name].name, >>> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ >>> RTE_METRICS_MAX_NAME_LEN); >>> ~~~~~~~~~~~~~~~~~~~~~~~~~ >>> >>> names[X].name is only RTE_METRICS_MAX_NAME_LEN bytes, so this can >>> overflow. Do we need a newer dpdk submodule? > > yeah, see https://github.com/ceph/ceph/pull/21744 there are pending patches addressing these issues in the dpdk upstream. see https://www.mail-archive.com/dev@xxxxxxxx/msg100134.html i guess we just need to wait for dpdk to merge them, and spdk to incorporate a commit of spdk with these fixes. regarding to the hash function, see http://dpdk.org/dev/patchwork/patch/39431/ . > >>> >> >> s/overflow/end up not being null terminated/ >> >> ...still, it's a bug either way... >> -- Regards Kefu Chai -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
