From owner-svn-src-all@freebsd.org Fri May 18 03:38:20 2018 Return-Path: Delivered-To: svn-src-all@mailman.ysv.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2610:1c1:1:606c::19:1]) by mailman.ysv.freebsd.org (Postfix) with ESMTP id 41949EE81A4; Fri, 18 May 2018 03:38:20 +0000 (UTC) (envelope-from mmacy@FreeBSD.org) Received: from mxrelay.nyi.freebsd.org (mxrelay.nyi.freebsd.org [IPv6:2610:1c1:1:606c::19:3]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (Client CN "mxrelay.nyi.freebsd.org", Issuer "Let's Encrypt Authority X3" (verified OK)) by mx1.freebsd.org (Postfix) with ESMTPS id E199082FE9; Fri, 18 May 2018 03:38:19 +0000 (UTC) (envelope-from mmacy@FreeBSD.org) Received: from repo.freebsd.org (repo.freebsd.org [IPv6:2610:1c1:1:6068::e6a:0]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (Client did not present a certificate) by mxrelay.nyi.freebsd.org (Postfix) with ESMTPS id A86A212B1E; Fri, 18 May 2018 03:38:19 +0000 (UTC) (envelope-from mmacy@FreeBSD.org) Received: from repo.freebsd.org ([127.0.1.37]) by repo.freebsd.org (8.15.2/8.15.2) with ESMTP id w4I3cJq7065231; Fri, 18 May 2018 03:38:19 GMT (envelope-from mmacy@FreeBSD.org) Received: (from mmacy@localhost) by repo.freebsd.org (8.15.2/8.15.2/Submit) id w4I3cHln065221; Fri, 18 May 2018 03:38:17 GMT (envelope-from mmacy@FreeBSD.org) Message-Id: <201805180338.w4I3cHln065221@repo.freebsd.org> X-Authentication-Warning: repo.freebsd.org: mmacy set sender to mmacy@FreeBSD.org using -f From: Matt Macy Date: Fri, 18 May 2018 03:38:17 +0000 (UTC) To: src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-head@freebsd.org Subject: svn commit: r333778 - in head/sys: conf dev/netmap X-SVN-Group: head X-SVN-Commit-Author: mmacy X-SVN-Commit-Paths: in head/sys: conf dev/netmap X-SVN-Commit-Revision: 333778 X-SVN-Commit-Repository: base MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit X-BeenThere: svn-src-all@freebsd.org X-Mailman-Version: 2.1.26 Precedence: list List-Id: "SVN commit messages for the entire src tree \(except for " user" and " projects" \)" List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 18 May 2018 03:38:20 -0000 Author: mmacy Date: Fri May 18 03:38:17 2018 New Revision: 333778 URL: https://svnweb.freebsd.org/changeset/base/333778 Log: netmap: pull fix for 32-bit support from upstream Approved by: sbruno Modified: head/sys/conf/makeLINT.mk head/sys/dev/netmap/if_ptnet.c head/sys/dev/netmap/netmap.c head/sys/dev/netmap/netmap_freebsd.c head/sys/dev/netmap/netmap_generic.c head/sys/dev/netmap/netmap_kern.h head/sys/dev/netmap/netmap_legacy.c head/sys/dev/netmap/netmap_mem2.c head/sys/dev/netmap/netmap_monitor.c head/sys/dev/netmap/netmap_pipe.c head/sys/dev/netmap/netmap_pt.c head/sys/dev/netmap/netmap_vale.c Modified: head/sys/conf/makeLINT.mk ============================================================================== --- head/sys/conf/makeLINT.mk Fri May 18 02:58:26 2018 (r333777) +++ head/sys/conf/makeLINT.mk Fri May 18 03:38:17 2018 (r333778) @@ -53,7 +53,6 @@ LINT: ${NOTES} ${MAKELINT_SED} .if ${TARGET} == "powerpc" # cat is available, not sure if cp is? cat ${.TARGET} > ${.TARGET}64 - echo "nodevice netmap" >> ${.TARGET} echo "machine ${TARGET} powerpc" >> ${.TARGET} echo "machine ${TARGET} powerpc64" >> ${.TARGET}64 .endif Modified: head/sys/dev/netmap/if_ptnet.c ============================================================================== --- head/sys/dev/netmap/if_ptnet.c Fri May 18 02:58:26 2018 (r333777) +++ head/sys/dev/netmap/if_ptnet.c Fri May 18 03:38:17 2018 (r333778) @@ -210,7 +210,7 @@ static int ptnet_irqs_init(struct ptnet_softc *sc); static void ptnet_irqs_fini(struct ptnet_softc *sc); static uint32_t ptnet_nm_ptctl(if_t ifp, uint32_t cmd); -static int ptnet_nm_config(struct netmap_adapter *na, +static int ptnet_nm_config(struct netmap_adapter *na, struct nm_config_info *info); static void ptnet_update_vnet_hdr(struct ptnet_softc *sc); static int ptnet_nm_register(struct netmap_adapter *na, int onoff); Modified: head/sys/dev/netmap/netmap.c ============================================================================== --- head/sys/dev/netmap/netmap.c Fri May 18 02:58:26 2018 (r333777) +++ head/sys/dev/netmap/netmap.c Fri May 18 03:38:17 2018 (r333778) @@ -536,35 +536,35 @@ SYSBEGIN(main_init); SYSCTL_DECL(_dev_netmap); SYSCTL_NODE(_dev, OID_AUTO, netmap, CTLFLAG_RW, 0, "Netmap args"); SYSCTL_INT(_dev_netmap, OID_AUTO, verbose, - CTLFLAG_RW, &netmap_verbose, 0, "Verbose mode"); + CTLFLAG_RW, &netmap_verbose, 0, "Verbose mode"); SYSCTL_INT(_dev_netmap, OID_AUTO, no_timestamp, - CTLFLAG_RW, &netmap_no_timestamp, 0, "no_timestamp"); + CTLFLAG_RW, &netmap_no_timestamp, 0, "no_timestamp"); SYSCTL_INT(_dev_netmap, OID_AUTO, no_pendintr, CTLFLAG_RW, &netmap_no_pendintr, - 0, "Always look for new received packets."); + 0, "Always look for new received packets."); SYSCTL_INT(_dev_netmap, OID_AUTO, txsync_retry, CTLFLAG_RW, - &netmap_txsync_retry, 0, "Number of txsync loops in bridge's flush."); + &netmap_txsync_retry, 0, "Number of txsync loops in bridge's flush."); SYSCTL_INT(_dev_netmap, OID_AUTO, fwd, CTLFLAG_RW, &netmap_fwd, 0, - "Force NR_FORWARD mode"); + "Force NR_FORWARD mode"); SYSCTL_INT(_dev_netmap, OID_AUTO, admode, CTLFLAG_RW, &netmap_admode, 0, - "Adapter mode. 0 selects the best option available," - "1 forces native adapter, 2 forces emulated adapter"); + "Adapter mode. 0 selects the best option available," + "1 forces native adapter, 2 forces emulated adapter"); SYSCTL_INT(_dev_netmap, OID_AUTO, generic_mit, CTLFLAG_RW, &netmap_generic_mit, - 0, "RX notification interval in nanoseconds"); + 0, "RX notification interval in nanoseconds"); SYSCTL_INT(_dev_netmap, OID_AUTO, generic_ringsize, CTLFLAG_RW, - &netmap_generic_ringsize, 0, - "Number of per-ring slots for emulated netmap mode"); + &netmap_generic_ringsize, 0, + "Number of per-ring slots for emulated netmap mode"); SYSCTL_INT(_dev_netmap, OID_AUTO, generic_rings, CTLFLAG_RW, - &netmap_generic_rings, 0, - "Number of TX/RX queues for emulated netmap adapters"); + &netmap_generic_rings, 0, + "Number of TX/RX queues for emulated netmap adapters"); #ifdef linux SYSCTL_INT(_dev_netmap, OID_AUTO, generic_txqdisc, CTLFLAG_RW, - &netmap_generic_txqdisc, 0, "Use qdisc for generic adapters"); + &netmap_generic_txqdisc, 0, "Use qdisc for generic adapters"); #endif SYSCTL_INT(_dev_netmap, OID_AUTO, ptnet_vnet_hdr, CTLFLAG_RW, &ptnet_vnet_hdr, - 0, "Allow ptnet devices to use virtio-net headers"); + 0, "Allow ptnet devices to use virtio-net headers"); SYSCTL_INT(_dev_netmap, OID_AUTO, ptnetmap_tx_workers, CTLFLAG_RW, - &ptnetmap_tx_workers, 0, "Use worker threads for pnetmap TX processing"); + &ptnetmap_tx_workers, 0, "Use worker threads for pnetmap TX processing"); SYSEND; @@ -765,15 +765,15 @@ netmap_update_config(struct netmap_adapter *na) na->rx_buf_maxsize == info.rx_buf_maxsize) return 0; /* nothing changed */ if (na->active_fds == 0) { - D("configuration changed for %s: txring %d x %d, " - "rxring %d x %d, rxbufsz %d", - na->name, na->num_tx_rings, na->num_tx_desc, - na->num_rx_rings, na->num_rx_desc, na->rx_buf_maxsize); na->num_tx_rings = info.num_tx_rings; na->num_tx_desc = info.num_tx_descs; na->num_rx_rings = info.num_rx_rings; na->num_rx_desc = info.num_rx_descs; na->rx_buf_maxsize = info.rx_buf_maxsize; + D("configuration changed for %s: txring %d x %d, " + "rxring %d x %d, rxbufsz %d", + na->name, na->num_tx_rings, na->num_tx_desc, + na->num_rx_rings, na->num_rx_desc, na->rx_buf_maxsize); return 0; } D("WARNING: configuration changed for %s while active: " @@ -830,7 +830,7 @@ netmap_krings_create(struct netmap_adapter *na, u_int n[NR_TX] = na->num_tx_rings + 1; n[NR_RX] = na->num_rx_rings + 1; - len = (n[NR_TX] + n[NR_RX]) * + len = (n[NR_TX] + n[NR_RX]) * (sizeof(struct netmap_kring) + sizeof(struct netmap_kring *)) + tailroom; @@ -841,7 +841,7 @@ netmap_krings_create(struct netmap_adapter *na, u_int } na->rx_rings = na->tx_rings + n[NR_TX]; na->tailroom = na->rx_rings + n[NR_RX]; - + /* link the krings in the krings array */ kring = (struct netmap_kring *)((char *)na->tailroom + tailroom); for (i = 0; i < n[NR_TX] + n[NR_RX]; i++) { @@ -1006,9 +1006,9 @@ netmap_do_unregif(struct netmap_priv_d *priv) if (netmap_verbose) D("deleting last instance for %s", na->name); - if (nm_netmap_on(na)) { - D("BUG: netmap on while going to delete the krings"); - } + if (nm_netmap_on(na)) { + D("BUG: netmap on while going to delete the krings"); + } na->nm_krings_delete(na); } @@ -1324,7 +1324,7 @@ netmap_rxsync_from_host(struct netmap_kring *kring, in m_copydata(m, 0, len, NMB(na, slot)); ND("nm %d len %d", nm_i, len); if (netmap_verbose) - D("%s", nm_dump_buf(NMB(na, slot),len, 128, NULL)); + D("%s", nm_dump_buf(NMB(na, slot),len, 128, NULL)); slot->len = len; slot->flags = 0; @@ -1476,7 +1476,7 @@ netmap_get_na(struct nmreq_header *hdr, struct netmap_adapter **na, struct ifnet **ifp, struct netmap_mem_d *nmd, int create) { - struct nmreq_register *req = (struct nmreq_register *)hdr->nr_body; + struct nmreq_register *req = (struct nmreq_register *)(uintptr_t)hdr->nr_body; int error = 0; struct netmap_adapter *ret = NULL; int nmd_ref = 0; @@ -2091,9 +2091,6 @@ netmap_do_regif(struct netmap_priv_d *priv, struct net NMG_LOCK_ASSERT(); priv->np_na = na; /* store the reference */ - error = netmap_set_ringid(priv, nr_mode, nr_ringid, nr_flags); - if (error) - goto err; error = netmap_mem_finalize(na->nm_mem, na); if (error) goto err; @@ -2109,7 +2106,14 @@ netmap_do_regif(struct netmap_priv_d *priv, struct net /* ring configuration may have changed, fetch from the card */ netmap_update_config(na); + } + /* compute the range of tx and rx rings to monitor */ + error = netmap_set_ringid(priv, nr_mode, nr_ringid, nr_flags); + if (error) + goto err_put_lut; + + if (na->active_fds == 0) { /* * If this is the first registration of the adapter, * perform sanity checks and create the in-kernel view @@ -2117,12 +2121,18 @@ netmap_do_regif(struct netmap_priv_d *priv, struct net */ if (na->ifp && nm_priv_rx_enabled(priv)) { /* This netmap adapter is attached to an ifnet. */ - unsigned nbs = netmap_mem_bufsize(na->nm_mem); + unsigned nbs = NETMAP_BUF_SIZE(na); unsigned mtu = nm_os_ifnet_mtu(na->ifp); - ND("mtu %d rx_buf_maxsize %d netmap_buf_size %d", - mtu, na->rx_buf_maxsize, nbs); + ND("%s: mtu %d rx_buf_maxsize %d netmap_buf_size %d", + na->name, mtu, na->rx_buf_maxsize, nbs); + if (na->rx_buf_maxsize == 0) { + D("%s: error: rx_buf_maxsize == 0", na->name); + error = EIO; + goto err_drop_mem; + } + if (mtu <= na->rx_buf_maxsize) { /* The MTU fits a single NIC slot. We only * Need to check that netmap buffers are @@ -2191,7 +2201,7 @@ netmap_do_regif(struct netmap_priv_d *priv, struct net nifp = netmap_mem_if_new(na, priv); if (nifp == NULL) { error = ENOMEM; - goto err_del_rings; + goto err_rel_excl; } if (nm_kring_pending(priv)) { @@ -2217,10 +2227,9 @@ netmap_do_regif(struct netmap_priv_d *priv, struct net err_del_if: netmap_mem_if_delete(na, nifp); -err_del_rings: - netmap_mem_rings_delete(na); err_rel_excl: netmap_krings_put(priv); + netmap_mem_rings_delete(na); err_del_krings: if (na->active_fds == 0) na->nm_krings_delete(na); @@ -2313,8 +2322,8 @@ netmap_ioctl(struct netmap_priv_d *priv, u_long cmd, c * For convenince, the nr_body pointer and the pointers * in the options list will be replaced with their * kernel-space counterparts. The original pointers are - * saved internally and later restored by nmreq_copyout - */ + * saved internally and later restored by nmreq_copyout + */ error = nmreq_copyin(hdr, nr_body_is_user); if (error) { return error; @@ -2326,7 +2335,7 @@ netmap_ioctl(struct netmap_priv_d *priv, u_long cmd, c switch (hdr->nr_reqtype) { case NETMAP_REQ_REGISTER: { struct nmreq_register *req = - (struct nmreq_register *)hdr->nr_body; + (struct nmreq_register *)(uintptr_t)hdr->nr_body; /* Protect access to priv from concurrent requests. */ NMG_LOCK(); do { @@ -2341,7 +2350,7 @@ netmap_ioctl(struct netmap_priv_d *priv, u_long cmd, c } #ifdef WITH_EXTMEM - opt = nmreq_findoption((struct nmreq_option *)hdr->nr_options, + opt = nmreq_findoption((struct nmreq_option *)(uintptr_t)hdr->nr_options, NETMAP_REQ_OPT_EXTMEM); if (opt != NULL) { struct nmreq_opt_extmem *e = @@ -2444,7 +2453,7 @@ netmap_ioctl(struct netmap_priv_d *priv, u_long cmd, c case NETMAP_REQ_PORT_INFO_GET: { struct nmreq_port_info_get *req = - (struct nmreq_port_info_get *)hdr->nr_body; + (struct nmreq_port_info_get *)(uintptr_t)hdr->nr_body; NMG_LOCK(); do { @@ -2463,10 +2472,10 @@ netmap_ioctl(struct netmap_priv_d *priv, u_long cmd, c /* get a refcount */ hdr->nr_reqtype = NETMAP_REQ_REGISTER; - hdr->nr_body = (uint64_t)®req; + hdr->nr_body = (uintptr_t)®req; error = netmap_get_na(hdr, &na, &ifp, NULL, 1 /* create */); hdr->nr_reqtype = NETMAP_REQ_PORT_INFO_GET; /* reset type */ - hdr->nr_body = (uint64_t)req; /* reset nr_body */ + hdr->nr_body = (uintptr_t)req; /* reset nr_body */ if (error) { na = NULL; ifp = NULL; @@ -2517,7 +2526,7 @@ netmap_ioctl(struct netmap_priv_d *priv, u_long cmd, c case NETMAP_REQ_PORT_HDR_SET: { struct nmreq_port_hdr *req = - (struct nmreq_port_hdr *)hdr->nr_body; + (struct nmreq_port_hdr *)(uintptr_t)hdr->nr_body; /* Build a nmreq_register out of the nmreq_port_hdr, * so that we can call netmap_get_bdg_na(). */ struct nmreq_register regreq; @@ -2533,10 +2542,10 @@ netmap_ioctl(struct netmap_priv_d *priv, u_long cmd, c } NMG_LOCK(); hdr->nr_reqtype = NETMAP_REQ_REGISTER; - hdr->nr_body = (uint64_t)®req; + hdr->nr_body = (uintptr_t)®req; error = netmap_get_bdg_na(hdr, &na, NULL, 0); hdr->nr_reqtype = NETMAP_REQ_PORT_HDR_SET; - hdr->nr_body = (uint64_t)req; + hdr->nr_body = (uintptr_t)req; if (na && !error) { struct netmap_vp_adapter *vpna = (struct netmap_vp_adapter *)na; @@ -2556,7 +2565,7 @@ netmap_ioctl(struct netmap_priv_d *priv, u_long cmd, c case NETMAP_REQ_PORT_HDR_GET: { /* Get vnet-header length for this netmap port */ struct nmreq_port_hdr *req = - (struct nmreq_port_hdr *)hdr->nr_body; + (struct nmreq_port_hdr *)(uintptr_t)hdr->nr_body; /* Build a nmreq_register out of the nmreq_port_hdr, * so that we can call netmap_get_bdg_na(). */ struct nmreq_register regreq; @@ -2565,10 +2574,10 @@ netmap_ioctl(struct netmap_priv_d *priv, u_long cmd, c bzero(®req, sizeof(regreq)); NMG_LOCK(); hdr->nr_reqtype = NETMAP_REQ_REGISTER; - hdr->nr_body = (uint64_t)®req; + hdr->nr_body = (uintptr_t)®req; error = netmap_get_na(hdr, &na, &ifp, NULL, 0); hdr->nr_reqtype = NETMAP_REQ_PORT_HDR_GET; - hdr->nr_body = (uint64_t)req; + hdr->nr_body = (uintptr_t)req; if (na && !error) { req->nr_hdr_len = na->virt_hdr_len; } @@ -2595,7 +2604,7 @@ netmap_ioctl(struct netmap_priv_d *priv, u_long cmd, c #endif /* WITH_VALE */ case NETMAP_REQ_POOLS_INFO_GET: { struct nmreq_pools_info *req = - (struct nmreq_pools_info *)hdr->nr_body; + (struct nmreq_pools_info *)(uintptr_t)hdr->nr_body; /* Get information from the memory allocator. This * netmap device must already be bound to a port. * Note that hdr->nr_name is ignored. */ @@ -2774,8 +2783,8 @@ nmreq_copyin(struct nmreq_header *hdr, int nr_body_is_ error = EMSGSIZE; goto out_err; } - if ((rqsz && hdr->nr_body == (uint64_t)NULL) || - (!rqsz && hdr->nr_body != (uint64_t)NULL)) { + if ((rqsz && hdr->nr_body == (uintptr_t)NULL) || + (!rqsz && hdr->nr_body != (uintptr_t)NULL)) { /* Request body expected, but not found; or * request body found but unexpected. */ error = EINVAL; @@ -2784,8 +2793,8 @@ nmreq_copyin(struct nmreq_header *hdr, int nr_body_is_ bufsz = 2 * sizeof(void *) + rqsz; optsz = 0; - for (src = (struct nmreq_option *)hdr->nr_options; src; - src = (struct nmreq_option *)buf.nro_next) + for (src = (struct nmreq_option *)(uintptr_t)hdr->nr_options; src; + src = (struct nmreq_option *)(uintptr_t)buf.nro_next) { error = copyin(src, &buf, sizeof(*src)); if (error) @@ -2813,11 +2822,11 @@ nmreq_copyin(struct nmreq_header *hdr, int nr_body_is_ p = (char *)ptrs; /* copy the body */ - error = copyin((void *)hdr->nr_body, p, rqsz); + error = copyin((void *)(uintptr_t)hdr->nr_body, p, rqsz); if (error) goto out_restore; /* overwrite the user pointer with the in-kernel one */ - hdr->nr_body = (uint64_t)p; + hdr->nr_body = (uintptr_t)p; p += rqsz; /* copy the options */ @@ -2874,7 +2883,7 @@ static int nmreq_copyout(struct nmreq_header *hdr, int rerror) { struct nmreq_option *src, *dst; - void *ker = (void *)hdr->nr_body, *bufstart; + void *ker = (void *)(uintptr_t)hdr->nr_body, *bufstart; uint64_t *ptrs; size_t bodysz; int error; @@ -2886,13 +2895,13 @@ nmreq_copyout(struct nmreq_header *hdr, int rerror) ptrs = (uint64_t *)ker - 2; bufstart = ptrs; hdr->nr_body = *ptrs++; - src = (struct nmreq_option *)hdr->nr_options; + src = (struct nmreq_option *)(uintptr_t)hdr->nr_options; hdr->nr_options = *ptrs; if (!rerror) { /* copy the body */ bodysz = nmreq_size_by_type(hdr->nr_reqtype); - error = copyout(ker, (void *)hdr->nr_body, bodysz); + error = copyout(ker, (void *)(uintptr_t)hdr->nr_body, bodysz); if (error) { rerror = error; goto out; @@ -2900,7 +2909,7 @@ nmreq_copyout(struct nmreq_header *hdr, int rerror) } /* copy the options */ - dst = (struct nmreq_option *)hdr->nr_options; + dst = (struct nmreq_option *)(uintptr_t)hdr->nr_options; while (src) { size_t optsz; uint64_t next; @@ -2916,7 +2925,7 @@ nmreq_copyout(struct nmreq_header *hdr, int rerror) rerror = error; goto out; } - + /* copy the option body only if there was no error */ if (!rerror && !src->nro_status) { optsz = nmreq_opt_size_by_type(src->nro_reqtype); @@ -2928,8 +2937,8 @@ nmreq_copyout(struct nmreq_header *hdr, int rerror) } } } - src = (struct nmreq_option *)next; - dst = (struct nmreq_option *)*ptrs; + src = (struct nmreq_option *)(uintptr_t)next; + dst = (struct nmreq_option *)(uintptr_t)*ptrs; } @@ -2942,7 +2951,7 @@ out: struct nmreq_option * nmreq_findoption(struct nmreq_option *opt, uint16_t reqtype) { - for ( ; opt; opt = (struct nmreq_option *)opt->nro_next) + for ( ; opt; opt = (struct nmreq_option *)(uintptr_t)opt->nro_next) if (opt->nro_reqtype == reqtype) return opt; return NULL; @@ -2953,7 +2962,7 @@ nmreq_checkduplicate(struct nmreq_option *opt) { uint16_t type = opt->nro_reqtype; int dup = 0; - while ((opt = nmreq_findoption((struct nmreq_option *)opt->nro_next, + while ((opt = nmreq_findoption((struct nmreq_option *)(uintptr_t)opt->nro_next, type))) { dup++; opt->nro_status = EINVAL; @@ -2969,8 +2978,8 @@ nmreq_checkoptions(struct nmreq_header *hdr) * marked as not supported */ - for (opt = (struct nmreq_option *)hdr->nr_options; opt; - opt = (struct nmreq_option *)opt->nro_next) + for (opt = (struct nmreq_option *)(uintptr_t)hdr->nr_options; opt; + opt = (struct nmreq_option *)(uintptr_t)opt->nro_next) if (opt->nro_status == EOPNOTSUPP) return EOPNOTSUPP; @@ -3643,9 +3652,9 @@ netmap_transmit(struct ifnet *ifp, struct mbuf *m) */ mbq_lock(q); - busy = kring->nr_hwtail - kring->nr_hwcur; - if (busy < 0) - busy += kring->nkr_num_slots; + busy = kring->nr_hwtail - kring->nr_hwcur; + if (busy < 0) + busy += kring->nkr_num_slots; if (busy + mbq_len(q) >= kring->nkr_num_slots - 1) { RD(2, "%s full hwcur %d hwtail %d qlen %d", na->name, kring->nr_hwcur, kring->nr_hwtail, mbq_len(q)); Modified: head/sys/dev/netmap/netmap_freebsd.c ============================================================================== --- head/sys/dev/netmap/netmap_freebsd.c Fri May 18 02:58:26 2018 (r333777) +++ head/sys/dev/netmap/netmap_freebsd.c Fri May 18 03:38:17 2018 (r333778) @@ -138,13 +138,13 @@ nm_os_put_module(void) static void netmap_ifnet_arrival_handler(void *arg __unused, struct ifnet *ifp) { - netmap_undo_zombie(ifp); + netmap_undo_zombie(ifp); } static void netmap_ifnet_departure_handler(void *arg __unused, struct ifnet *ifp) { - netmap_make_zombie(ifp); + netmap_make_zombie(ifp); } static eventhandler_tag nm_ifnet_ah_tag; @@ -153,33 +153,33 @@ static eventhandler_tag nm_ifnet_dh_tag; int nm_os_ifnet_init(void) { - nm_ifnet_ah_tag = - EVENTHANDLER_REGISTER(ifnet_arrival_event, - netmap_ifnet_arrival_handler, - NULL, EVENTHANDLER_PRI_ANY); - nm_ifnet_dh_tag = - EVENTHANDLER_REGISTER(ifnet_departure_event, - netmap_ifnet_departure_handler, - NULL, EVENTHANDLER_PRI_ANY); - return 0; + nm_ifnet_ah_tag = + EVENTHANDLER_REGISTER(ifnet_arrival_event, + netmap_ifnet_arrival_handler, + NULL, EVENTHANDLER_PRI_ANY); + nm_ifnet_dh_tag = + EVENTHANDLER_REGISTER(ifnet_departure_event, + netmap_ifnet_departure_handler, + NULL, EVENTHANDLER_PRI_ANY); + return 0; } void nm_os_ifnet_fini(void) { - EVENTHANDLER_DEREGISTER(ifnet_arrival_event, - nm_ifnet_ah_tag); - EVENTHANDLER_DEREGISTER(ifnet_departure_event, - nm_ifnet_dh_tag); + EVENTHANDLER_DEREGISTER(ifnet_arrival_event, + nm_ifnet_ah_tag); + EVENTHANDLER_DEREGISTER(ifnet_departure_event, + nm_ifnet_dh_tag); } unsigned nm_os_ifnet_mtu(struct ifnet *ifp) { #if __FreeBSD_version < 1100030 - return ifp->if_data.ifi_mtu; + return ifp->if_data.ifi_mtu; #else /* __FreeBSD_version >= 1100030 */ - return ifp->if_mtu; + return ifp->if_mtu; #endif } @@ -625,14 +625,14 @@ nm_os_vi_detach(struct ifnet *ifp) struct nm_os_extmem { vm_object_t obj; vm_offset_t kva; - vm_offset_t size; - vm_pindex_t scan; + vm_offset_t size; + uintptr_t scan; }; void nm_os_extmem_delete(struct nm_os_extmem *e) { - D("freeing %zx bytes", (size_t)e->size); + D("freeing %jx bytes", (uintmax_t)e->size); vm_map_remove(kernel_map, e->kva, e->kva + e->size); nm_os_free(e); } @@ -701,7 +701,7 @@ nm_os_extmem_create(unsigned long p, struct nmreq_pool VMFS_OPTIMAL_SPACE, VM_PROT_READ | VM_PROT_WRITE, VM_PROT_READ | VM_PROT_WRITE, 0); if (rv != KERN_SUCCESS) { - D("vm_map_find(%zx) failed", (size_t)e->size); + D("vm_map_find(%jx) failed", (uintmax_t)e->size); goto out_rel; } rv = vm_map_wire(kernel_map, e->kva, e->kva + e->size, @@ -942,7 +942,7 @@ struct netmap_vm_handle_t { static int netmap_dev_pager_ctor(void *handle, vm_ooffset_t size, vm_prot_t prot, - vm_ooffset_t foff, struct ucred *cred, u_short *color) + vm_ooffset_t foff, struct ucred *cred, u_short *color) { struct netmap_vm_handle_t *vmh = handle; @@ -1519,7 +1519,7 @@ freebsd_netmap_poll(struct cdev *cdevi __unused, int e static int freebsd_netmap_ioctl(struct cdev *dev __unused, u_long cmd, caddr_t data, - int ffla __unused, struct thread *td) + int ffla __unused, struct thread *td) { int error; struct netmap_priv_d *priv; Modified: head/sys/dev/netmap/netmap_generic.c ============================================================================== --- head/sys/dev/netmap/netmap_generic.c Fri May 18 02:58:26 2018 (r333777) +++ head/sys/dev/netmap/netmap_generic.c Fri May 18 03:38:17 2018 (r333778) @@ -235,14 +235,14 @@ nm_os_get_mbuf(struct ifnet *ifp, int len) for ((_k)=*(_karr), (_i) = 0; (_i) < (_n); (_i)++, (_k) = (_karr)[(_i)]) #define for_each_tx_kring(_i, _k, _na) \ - for_each_kring_n(_i, _k, (_na)->tx_rings, (_na)->num_tx_rings) + for_each_kring_n(_i, _k, (_na)->tx_rings, (_na)->num_tx_rings) #define for_each_tx_kring_h(_i, _k, _na) \ - for_each_kring_n(_i, _k, (_na)->tx_rings, (_na)->num_tx_rings + 1) + for_each_kring_n(_i, _k, (_na)->tx_rings, (_na)->num_tx_rings + 1) #define for_each_rx_kring(_i, _k, _na) \ - for_each_kring_n(_i, _k, (_na)->rx_rings, (_na)->num_rx_rings) + for_each_kring_n(_i, _k, (_na)->rx_rings, (_na)->num_rx_rings) #define for_each_rx_kring_h(_i, _k, _na) \ - for_each_kring_n(_i, _k, (_na)->rx_rings, (_na)->num_rx_rings + 1) + for_each_kring_n(_i, _k, (_na)->rx_rings, (_na)->num_rx_rings + 1) /* ======================== PERFORMANCE STATISTICS =========================== */ @@ -297,12 +297,12 @@ static struct rate_context rate_ctx; void generic_rate(int txp, int txs, int txi, int rxp, int rxs, int rxi) { - if (txp) rate_ctx.new.txpkt++; - if (txs) rate_ctx.new.txsync++; - if (txi) rate_ctx.new.txirq++; - if (rxp) rate_ctx.new.rxpkt++; - if (rxs) rate_ctx.new.rxsync++; - if (rxi) rate_ctx.new.rxirq++; + if (txp) rate_ctx.new.txpkt++; + if (txs) rate_ctx.new.txsync++; + if (txi) rate_ctx.new.txirq++; + if (rxp) rate_ctx.new.rxpkt++; + if (rxs) rate_ctx.new.rxsync++; + if (rxi) rate_ctx.new.rxirq++; } #else /* !RATE */ @@ -586,7 +586,7 @@ generic_mbuf_destructor(struct mbuf *m) * MBUF_TXQ(m) under our feet. If the match is not found * on 'r', we try to see if it belongs to some other ring. */ - for (;;) { + for (;;) { bool match = false; kring = na->tx_rings[r]; Modified: head/sys/dev/netmap/netmap_kern.h ============================================================================== --- head/sys/dev/netmap/netmap_kern.h Fri May 18 02:58:26 2018 (r333777) +++ head/sys/dev/netmap/netmap_kern.h Fri May 18 03:38:17 2018 (r333778) @@ -450,7 +450,7 @@ struct netmap_kring { /* the adapter the owns this kring */ struct netmap_adapter *na; - + /* the adapter that wants to be notified when this kring has * new slots avaialable. This is usually the same as the above, * but wrappers may let it point to themselves @@ -1528,7 +1528,7 @@ int netmap_get_monitor_na(struct nmreq_header *hdr, st void netmap_monitor_stop(struct netmap_adapter *na); #else #define netmap_get_monitor_na(hdr, _2, _3, _4) \ - (((struct nmreq_register *)hdr->nr_body)->nr_flags & (NR_MONITOR_TX | NR_MONITOR_RX) ? EOPNOTSUPP : 0) + (((struct nmreq_register *)(uintptr_t)hdr->nr_body)->nr_flags & (NR_MONITOR_TX | NR_MONITOR_RX) ? EOPNOTSUPP : 0) #endif #ifdef CONFIG_NET_NS @@ -1823,7 +1823,7 @@ struct lut_entry { }; #else /* linux & _WIN32 */ /* dma-mapping in linux can assign a buffer a different address - * depending on the device, so we need to have a separate + * depending on the device, so we need to have a separate * physical-address look-up table for each na. * We can still share the vaddrs, though, therefore we split * the lut_entry structure. @@ -2177,7 +2177,7 @@ nm_ptnetmap_host_on(struct netmap_adapter *na) } #else /* !WITH_PTNETMAP_HOST */ #define netmap_get_pt_host_na(hdr, _2, _3, _4) \ - (((struct nmreq_register *)hdr->nr_body)->nr_flags & (NR_PTNETMAP_HOST) ? EOPNOTSUPP : 0) + (((struct nmreq_register *)(uintptr_t)hdr->nr_body)->nr_flags & (NR_PTNETMAP_HOST) ? EOPNOTSUPP : 0) #define ptnetmap_ctl(_1, _2, _3) EINVAL #define nm_ptnetmap_host_on(_1) EINVAL #endif /* !WITH_PTNETMAP_HOST */ Modified: head/sys/dev/netmap/netmap_legacy.c ============================================================================== --- head/sys/dev/netmap/netmap_legacy.c Fri May 18 02:58:26 2018 (r333777) +++ head/sys/dev/netmap/netmap_legacy.c Fri May 18 03:38:17 2018 (r333778) @@ -132,8 +132,8 @@ nmreq_from_legacy(struct nmreq *nmr, u_long ioctl_cmd) /* First prepare the request header. */ hdr->nr_version = NETMAP_API; /* new API */ strncpy(hdr->nr_name, nmr->nr_name, sizeof(nmr->nr_name)); - hdr->nr_options = (uint64_t)NULL; - hdr->nr_body = (uint64_t)NULL; + hdr->nr_options = (uintptr_t)NULL; + hdr->nr_body = (uintptr_t)NULL; switch (ioctl_cmd) { case NIOCREGIF: { @@ -142,7 +142,7 @@ nmreq_from_legacy(struct nmreq *nmr, u_long ioctl_cmd) /* Regular NIOCREGIF operation. */ struct nmreq_register *req = nm_os_malloc(sizeof(*req)); if (!req) { goto oom; } - hdr->nr_body = (uint64_t)req; + hdr->nr_body = (uintptr_t)req; hdr->nr_reqtype = NETMAP_REQ_REGISTER; if (nmreq_register_from_legacy(nmr, hdr, req)) { goto oom; @@ -152,7 +152,7 @@ nmreq_from_legacy(struct nmreq *nmr, u_long ioctl_cmd) case NETMAP_BDG_ATTACH: { struct nmreq_vale_attach *req = nm_os_malloc(sizeof(*req)); if (!req) { goto oom; } - hdr->nr_body = (uint64_t)req; + hdr->nr_body = (uintptr_t)req; hdr->nr_reqtype = NETMAP_REQ_VALE_ATTACH; if (nmreq_register_from_legacy(nmr, hdr, &req->reg)) { goto oom; @@ -167,14 +167,14 @@ nmreq_from_legacy(struct nmreq *nmr, u_long ioctl_cmd) } case NETMAP_BDG_DETACH: { hdr->nr_reqtype = NETMAP_REQ_VALE_DETACH; - hdr->nr_body = (uint64_t)nm_os_malloc(sizeof(struct nmreq_vale_detach)); + hdr->nr_body = (uintptr_t)nm_os_malloc(sizeof(struct nmreq_vale_detach)); break; } case NETMAP_BDG_VNET_HDR: case NETMAP_VNET_HDR_GET: { struct nmreq_port_hdr *req = nm_os_malloc(sizeof(*req)); if (!req) { goto oom; } - hdr->nr_body = (uint64_t)req; + hdr->nr_body = (uintptr_t)req; hdr->nr_reqtype = (nmr->nr_cmd == NETMAP_BDG_VNET_HDR) ? NETMAP_REQ_PORT_HDR_SET : NETMAP_REQ_PORT_HDR_GET; req->nr_hdr_len = nmr->nr_arg1; @@ -183,7 +183,7 @@ nmreq_from_legacy(struct nmreq *nmr, u_long ioctl_cmd) case NETMAP_BDG_NEWIF : { struct nmreq_vale_newif *req = nm_os_malloc(sizeof(*req)); if (!req) { goto oom; } - hdr->nr_body = (uint64_t)req; + hdr->nr_body = (uintptr_t)req; hdr->nr_reqtype = NETMAP_REQ_VALE_NEWIF; req->nr_tx_slots = nmr->nr_tx_slots; req->nr_rx_slots = nmr->nr_rx_slots; @@ -200,7 +200,7 @@ nmreq_from_legacy(struct nmreq *nmr, u_long ioctl_cmd) case NETMAP_BDG_POLLING_OFF: { struct nmreq_vale_polling *req = nm_os_malloc(sizeof(*req)); if (!req) { goto oom; } - hdr->nr_body = (uint64_t)req; + hdr->nr_body = (uintptr_t)req; hdr->nr_reqtype = (nmr->nr_cmd == NETMAP_BDG_POLLING_ON) ? NETMAP_REQ_VALE_POLLING_ENABLE : NETMAP_REQ_VALE_POLLING_DISABLE; @@ -232,7 +232,7 @@ nmreq_from_legacy(struct nmreq *nmr, u_long ioctl_cmd) if (nmr->nr_cmd == NETMAP_BDG_LIST) { struct nmreq_vale_list *req = nm_os_malloc(sizeof(*req)); if (!req) { goto oom; } - hdr->nr_body = (uint64_t)req; + hdr->nr_body = (uintptr_t)req; hdr->nr_reqtype = NETMAP_REQ_VALE_LIST; req->nr_bridge_idx = nmr->nr_arg1; req->nr_port_idx = nmr->nr_arg2; @@ -240,7 +240,7 @@ nmreq_from_legacy(struct nmreq *nmr, u_long ioctl_cmd) /* Regular NIOCGINFO. */ struct nmreq_port_info_get *req = nm_os_malloc(sizeof(*req)); if (!req) { goto oom; } - hdr->nr_body = (uint64_t)req; + hdr->nr_body = (uintptr_t)req; hdr->nr_reqtype = NETMAP_REQ_PORT_INFO_GET; req->nr_offset = nmr->nr_offset; req->nr_memsize = nmr->nr_memsize; @@ -258,7 +258,7 @@ nmreq_from_legacy(struct nmreq *nmr, u_long ioctl_cmd) oom: if (hdr) { if (hdr->nr_body) { - nm_os_free((void *)hdr->nr_body); + nm_os_free((void *)(uintptr_t)hdr->nr_body); } nm_os_free(hdr); } @@ -293,13 +293,13 @@ nmreq_to_legacy(struct nmreq_header *hdr, struct nmreq switch (hdr->nr_reqtype) { case NETMAP_REQ_REGISTER: { struct nmreq_register *req = - (struct nmreq_register *)hdr->nr_body; + (struct nmreq_register *)(uintptr_t)hdr->nr_body; nmreq_register_to_legacy(req, nmr); break; } case NETMAP_REQ_PORT_INFO_GET: { struct nmreq_port_info_get *req = - (struct nmreq_port_info_get *)hdr->nr_body; + (struct nmreq_port_info_get *)(uintptr_t)hdr->nr_body; nmr->nr_offset = req->nr_offset; nmr->nr_memsize = req->nr_memsize; nmr->nr_tx_slots = req->nr_tx_slots; @@ -311,7 +311,7 @@ nmreq_to_legacy(struct nmreq_header *hdr, struct nmreq } case NETMAP_REQ_VALE_ATTACH: { struct nmreq_vale_attach *req = - (struct nmreq_vale_attach *)hdr->nr_body; + (struct nmreq_vale_attach *)(uintptr_t)hdr->nr_body; nmreq_register_to_legacy(&req->reg, nmr); break; } @@ -320,7 +320,7 @@ nmreq_to_legacy(struct nmreq_header *hdr, struct nmreq } case NETMAP_REQ_VALE_LIST: { struct nmreq_vale_list *req = - (struct nmreq_vale_list *)hdr->nr_body; + (struct nmreq_vale_list *)(uintptr_t)hdr->nr_body; strncpy(nmr->nr_name, hdr->nr_name, sizeof(nmr->nr_name)); nmr->nr_arg1 = req->nr_bridge_idx; nmr->nr_arg2 = req->nr_port_idx; @@ -329,13 +329,13 @@ nmreq_to_legacy(struct nmreq_header *hdr, struct nmreq case NETMAP_REQ_PORT_HDR_SET: case NETMAP_REQ_PORT_HDR_GET: { struct nmreq_port_hdr *req = - (struct nmreq_port_hdr *)hdr->nr_body; + (struct nmreq_port_hdr *)(uintptr_t)hdr->nr_body; nmr->nr_arg1 = req->nr_hdr_len; break; } case NETMAP_REQ_VALE_NEWIF: { struct nmreq_vale_newif *req = - (struct nmreq_vale_newif *)hdr->nr_body; + (struct nmreq_vale_newif *)(uintptr_t)hdr->nr_body; nmr->nr_tx_slots = req->nr_tx_slots; nmr->nr_rx_slots = req->nr_rx_slots; nmr->nr_tx_rings = req->nr_tx_rings; @@ -375,7 +375,7 @@ netmap_ioctl_legacy(struct netmap_priv_d *priv, u_long nmreq_to_legacy(hdr, nmr); } if (hdr->nr_body) { - nm_os_free((void *)hdr->nr_body); + nm_os_free((void *)(uintptr_t)hdr->nr_body); } nm_os_free(hdr); break; Modified: head/sys/dev/netmap/netmap_mem2.c ============================================================================== --- head/sys/dev/netmap/netmap_mem2.c Fri May 18 02:58:26 2018 (r333777) +++ head/sys/dev/netmap/netmap_mem2.c Fri May 18 03:38:17 2018 (r333778) @@ -673,10 +673,10 @@ nm_mem_assign_id_locked(struct netmap_mem_d *nmd) static int nm_mem_assign_id(struct netmap_mem_d *nmd) { - int ret; + int ret; NM_MTX_LOCK(nm_mem_list_lock); - ret = nm_mem_assign_id_locked(nmd); + ret = nm_mem_assign_id_locked(nmd); NM_MTX_UNLOCK(nm_mem_list_lock); return ret; @@ -1143,7 +1143,7 @@ netmap_extra_alloc(struct netmap_adapter *na, uint32_t static void netmap_extra_free(struct netmap_adapter *na, uint32_t head) { - struct lut_entry *lut = na->na_lut.lut; + struct lut_entry *lut = na->na_lut.lut; struct netmap_mem_d *nmd = na->nm_mem; struct netmap_obj_pool *p = &nmd->pools[NETMAP_BUF_POOL]; uint32_t i, cur, *buf; @@ -1516,10 +1516,11 @@ netmap_mem_unmap(struct netmap_obj_pool *p, struct net return 0; #if defined(__FreeBSD__) + /* On FreeBSD mapping and unmapping is performed by the txsync + * and rxsync routine, packet by packet. */ (void)i; (void)lim; (void)lut; - D("unsupported on FreeBSD"); #elif defined(_WIN32) (void)i; (void)lim; @@ -1551,10 +1552,11 @@ netmap_mem_map(struct netmap_obj_pool *p, struct netma return 0; #if defined(__FreeBSD__) + /* On FreeBSD mapping and unmapping is performed by the txsync + * and rxsync routine, packet by packet. */ (void)i; (void)lim; (void)lut; - D("unsupported on FreeBSD"); #elif defined(_WIN32) (void)i; (void)lim; @@ -1572,7 +1574,7 @@ netmap_mem_map(struct netmap_obj_pool *p, struct netma if (lut->plut == NULL) { D("Failed to allocate physical lut for %s", na->name); return ENOMEM; - } + } for (i = 0; i < lim; i += p->_clustentries) { lut->plut[i].paddr = 0; @@ -1644,7 +1646,7 @@ error: * allocator for private memory */ static void * -_netmap_mem_private_new(size_t size, struct netmap_obj_params *p, +_netmap_mem_private_new(size_t size, struct netmap_obj_params *p, struct netmap_mem_ops *ops, int *perr) { struct netmap_mem_d *d = NULL; @@ -1722,16 +1724,16 @@ netmap_mem_private_new(u_int txr, u_int txd, u_int rxr if (p[NETMAP_RING_POOL].size < v) p[NETMAP_RING_POOL].size = v; /* each pipe endpoint needs two tx rings (1 normal + 1 host, fake) - * and two rx rings (again, 1 normal and 1 fake host) - */ + * and two rx rings (again, 1 normal and 1 fake host) + */ v = txr + rxr + 8 * npipes; if (p[NETMAP_RING_POOL].num < v) p[NETMAP_RING_POOL].num = v; /* for each pipe we only need the buffers for the 4 "real" rings. - * On the other end, the pipe ring dimension may be different from - * the parent port ring dimension. As a compromise, we allocate twice the - * space actually needed if the pipe rings were the same size as the parent rings - */ + * On the other end, the pipe ring dimension may be different from + * the parent port ring dimension. As a compromise, we allocate twice the + * space actually needed if the pipe rings were the same size as the parent rings + */ v = (4 * npipes + rxr) * rxd + (4 * npipes + txr) * txd + 2 + extra_bufs; /* the +2 is for the tx and rx fake buffers (indices 0 and 1) */ if (p[NETMAP_BUF_POOL].num < v) @@ -1942,7 +1944,11 @@ netmap_mem2_rings_create(struct netmap_adapter *na) return 0; cleanup: - netmap_free_rings(na); + /* we cannot actually cleanup here, since we don't own kring->users + * and kring->nr_klags & NKR_NEEDRING. The caller must decrement + * the first or zero-out the second, then call netmap_free_rings() + * to do the cleanup + */ return ENOMEM; } @@ -2155,7 +2161,7 @@ netmap_mem_ext_delete(struct netmap_mem_d *d) for (i = 0; i < NETMAP_POOLS_NR; i++) { struct netmap_obj_pool *p = &d->pools[i]; - + if (p->lut) { nm_free_lut(p->lut, p->objtotal); p->lut = NULL; @@ -2215,7 +2221,7 @@ netmap_mem_ext_create(uint64_t usrptr, struct nmreq_po pi->nr_if_pool_objtotal, pi->nr_if_pool_objsize, pi->nr_ring_pool_objtotal, pi->nr_ring_pool_objsize, pi->nr_buf_pool_objtotal, pi->nr_buf_pool_objsize); - + os = nm_os_extmem_create(usrptr, pi, &error); if (os == NULL) { D("os extmem creation failed"); @@ -2238,7 +2244,7 @@ netmap_mem_ext_create(uint64_t usrptr, struct nmreq_po &error); if (nme == NULL) goto out_unmap; - + nr_pages = nm_os_extmem_nr_pages(os); /* from now on pages will be released by nme destructor; @@ -2262,7 +2268,7 @@ netmap_mem_ext_create(uint64_t usrptr, struct nmreq_po error = ENOMEM; goto out_delete; } - + p->bitmap_slots = (o->num + sizeof(uint32_t) - 1) / sizeof(uint32_t); p->invalid_bitmap = nm_os_malloc(sizeof(uint32_t) * p->bitmap_slots); if (p->invalid_bitmap == NULL) { @@ -2515,11 +2521,11 @@ netmap_mem_pt_guest_finalize(struct netmap_mem_d *nmd) if (error) goto out; - /* Initialize the lut using the information contained in the + /* Initialize the lut using the information contained in the * ptnetmap memory device. */ - bufsize = nm_os_pt_memdev_ioread(ptnmd->ptn_dev, + bufsize = nm_os_pt_memdev_ioread(ptnmd->ptn_dev, PTNET_MDEV_IO_BUF_POOL_OBJSZ); - nbuffers = nm_os_pt_memdev_ioread(ptnmd->ptn_dev, + nbuffers = nm_os_pt_memdev_ioread(ptnmd->ptn_dev, PTNET_MDEV_IO_BUF_POOL_OBJNUM); /* allocate the lut */ @@ -2740,7 +2746,7 @@ netmap_mem_pt_guest_create(nm_memid_t mem_id) ptnmd->host_mem_id = mem_id; ptnmd->pt_ifs = NULL; - /* Assign new id in the guest (We have the lock) */ + /* Assign new id in the guest (We have the lock) */ err = nm_mem_assign_id_locked(&ptnmd->up); if (err) goto error; Modified: head/sys/dev/netmap/netmap_monitor.c ============================================================================== --- head/sys/dev/netmap/netmap_monitor.c Fri May 18 02:58:26 2018 (r333777) +++ head/sys/dev/netmap/netmap_monitor.c Fri May 18 03:38:17 2018 (r333778) @@ -139,7 +139,7 @@ nm_is_zmon(struct netmap_adapter *na) static int netmap_monitor_txsync(struct netmap_kring *kring, int flags) { - RD(1, "%s %x", kring->name, flags); + RD(1, "%s %x", kring->name, flags); return EIO; } @@ -152,10 +152,10 @@ netmap_monitor_txsync(struct netmap_kring *kring, int static int netmap_monitor_rxsync(struct netmap_kring *kring, int flags) { - ND("%s %x", kring->name, flags); + ND("%s %x", kring->name, flags); kring->nr_hwcur = kring->rhead; mb(); - return 0; + return 0; } /* nm_krings_create callbacks for monitors. @@ -198,7 +198,7 @@ nm_monitor_alloc(struct netmap_kring *kring, u_int n) return 0; old_len = sizeof(struct netmap_kring *)*kring->max_monitors; - len = sizeof(struct netmap_kring *) * n; + len = sizeof(struct netmap_kring *) * n; nm = nm_os_realloc(kring->monitors, len, old_len); if (nm == NULL) return ENOMEM; @@ -621,14 +621,14 @@ out_rxsync: static int netmap_zmon_parent_txsync(struct netmap_kring *kring, int flags) { - return netmap_zmon_parent_sync(kring, flags, NR_TX); *** DIFF OUTPUT TRUNCATED AT 1000 LINES ***