tree: git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband.git for-next head: 2d79d42b11e84e5bf577a6ef0917ad799dfe610e commit: 72278f474221abd5c09af5d3a4dff39f9d10948f [11/36] RDMA/ocrdma: Move PD resource management to driver.
New smatch warnings: drivers/infiniband/hw/ocrdma/ocrdma_verbs.c:683 ocrdma_alloc_pd() error: double free of 'pd' drivers/infiniband/hw/ocrdma/ocrdma_verbs.c:686 ocrdma_alloc_pd() error: passing non negative 255 to ERR_PTR This bug is that _ocrdma_pd_mgr_put_bitmap() truncates its error values to u8. I will push a Smatch check to warn about these sorts of signedness bugs. Also that function would be readable if it just returned directly instead of setting the temporary "status" variable. There are some other similar typing bugs in this file. Old smatch warnings: drivers/infiniband/hw/ocrdma/ocrdma_verbs.c:1766 ocrdma_destroy_qp() error: we previously assumed 'qp->rq_cq' could be null (see line 1760) git remote add infiniband git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband.git git remote update infiniband git checkout 72278f474221abd5c09af5d3a4dff39f9d10948f vim +/pd +683 drivers/infiniband/hw/ocrdma/ocrdma_verbs.c fe2caefc Parav Pandit 2012-03-21 677 fe2caefc Parav Pandit 2012-03-21 678 err: cffce990 Naresh Gottumukkala 2013-08-26 679 if (is_uctx_pd) { cffce990 Naresh Gottumukkala 2013-08-26 680 ocrdma_release_ucontext_pd(uctx); cffce990 Naresh Gottumukkala 2013-08-26 681 } else { 72278f47 Mitesh Ahuja 2014-12-18 682 status = _ocrdma_dealloc_pd(dev, pd); 45e86b33 Naresh Gottumukkala 2013-08-07 @683 kfree(pd); cffce990 Naresh Gottumukkala 2013-08-26 684 } cffce990 Naresh Gottumukkala 2013-08-26 685 exit: fe2caefc Parav Pandit 2012-03-21 @686 return ERR_PTR(status); fe2caefc Parav Pandit 2012-03-21 687 } fe2caefc Parav Pandit 2012-03-21 688 fe2caefc Parav Pandit 2012-03-21 689 int ocrdma_dealloc_pd(struct ib_pd *ibpd) --- 0-DAY kernel test infrastructure Open Source Technology Center http://lists.01.org/mailman/listinfo/kbuild Intel Corporation _______________________________________________ kbuild mailing list kbuild@lists.01.org https://lists.01.org/mailman/listinfo/kbuild