On Thu, Apr 29, 2021 at 11:05 AM Gioh Kim <gi-oh.kim@xxxxxxxxx> wrote: > > On Thu, Apr 29, 2021 at 10:58 AM Dan Carpenter <dan.carpenter@xxxxxxxxxx> wrote: > > > > tree: https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git master > > head: 593ef1658ecf61d3619885bdbbcfffa3d1417891 > > commit: fa607fcb87f6e3dca38b9984c4ccde3e36a43721 [13667/14048] block/rnbd-clt: Support polling mode for IO latency optimization > > config: x86_64-randconfig-m031-20210420 (attached as .config) > > compiler: gcc-9 (Debian 9.3.0-22) 9.3.0 > > > > If you fix the issue, kindly add following tag as appropriate > > Reported-by: kernel test robot <lkp@xxxxxxxxx> > > Reported-by: Dan Carpenter <dan.carpenter@xxxxxxxxxx> > > > > smatch warnings: > > drivers/infiniband/ulp/rtrs/rtrs-clt.c:2922 rtrs_clt_rdma_cq_direct() error: uninitialized symbol 'cnt'. > > > > vim +/cnt +2922 drivers/infiniband/ulp/rtrs/rtrs-clt.c > > > > fa607fcb87f6e3 Gioh Kim 2021-04-19 2901 int rtrs_clt_rdma_cq_direct(struct rtrs_clt *clt, unsigned int index) > > fa607fcb87f6e3 Gioh Kim 2021-04-19 2902 { > > fa607fcb87f6e3 Gioh Kim 2021-04-19 2903 int cnt; > > fa607fcb87f6e3 Gioh Kim 2021-04-19 2904 struct rtrs_con *con; > > fa607fcb87f6e3 Gioh Kim 2021-04-19 2905 struct rtrs_clt_sess *sess; > > fa607fcb87f6e3 Gioh Kim 2021-04-19 2906 struct path_it it; > > fa607fcb87f6e3 Gioh Kim 2021-04-19 2907 > > fa607fcb87f6e3 Gioh Kim 2021-04-19 2908 rcu_read_lock(); > > fa607fcb87f6e3 Gioh Kim 2021-04-19 2909 for (path_it_init(&it, clt); > > fa607fcb87f6e3 Gioh Kim 2021-04-19 2910 (sess = it.next_path(&it)) && it.i < it.clt->paths_num; it.i++) { > > fa607fcb87f6e3 Gioh Kim 2021-04-19 2911 if (READ_ONCE(sess->state) != RTRS_CLT_CONNECTED) > > fa607fcb87f6e3 Gioh Kim 2021-04-19 2912 continue; > > > > What if no sessions are connected? So we dont enter the loop or we > > hit continue every time? > > Hi Dan, > > I just finished a internal process for the fix of that issue. > In short, I would like to initialize cnt with -1. > So rtrs_clt_rdma_cq_direct will return -1 if there is no session. > And the block layer will not call rtrs_clt_rdma_cq_direct again > because the return value is negative. > > I am going to send the patch today with the tags above. > Thank you very much. Hi Dan, FYI, I just sent a patch including the tags. > > > > > fa607fcb87f6e3 Gioh Kim 2021-04-19 2913 > > fa607fcb87f6e3 Gioh Kim 2021-04-19 2914 con = sess->s.con[index + 1]; > > fa607fcb87f6e3 Gioh Kim 2021-04-19 2915 cnt = ib_process_cq_direct(con->cq, -1); > > fa607fcb87f6e3 Gioh Kim 2021-04-19 2916 if (cnt) > > fa607fcb87f6e3 Gioh Kim 2021-04-19 2917 break; > > fa607fcb87f6e3 Gioh Kim 2021-04-19 2918 } > > fa607fcb87f6e3 Gioh Kim 2021-04-19 2919 path_it_deinit(&it); > > fa607fcb87f6e3 Gioh Kim 2021-04-19 2920 rcu_read_unlock(); > > fa607fcb87f6e3 Gioh Kim 2021-04-19 2921 > > fa607fcb87f6e3 Gioh Kim 2021-04-19 @2922 return cnt; > > fa607fcb87f6e3 Gioh Kim 2021-04-19 2923 } > > > > --- > > 0-DAY CI Kernel Test Service, Intel Corporation > > https://lists.01.org/hyperkitty/list/kbuild-all@xxxxxxxxxxxx > >
