Date: Tue, 14 Feb 2012 18:00:37 +0000 (UTC) From: "Kenneth D. Merry" <ken@FreeBSD.org> To: src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-stable@freebsd.org, svn-src-stable-9@freebsd.org Subject: svn commit: r231697 - in stable/9: share/man/man4 sys/conf sys/dev/xen/blkback sys/dev/xen/netback sys/i386/include/xen sys/kern sys/xen/interface/io Message-ID: <201202141800.q1EI0bHb011087@svn.freebsd.org>
next in thread | raw e-mail | index | archive | help
Author: ken Date: Tue Feb 14 18:00:37 2012 New Revision: 231697 URL: http://svn.freebsd.org/changeset/base/231697 Log: MFC 230587, 230916 Xen netback driver rewrite. r230587 | ken | 2012-01-26 09:35:09 -0700 (Thu, 26 Jan 2012) | 38 lines Xen netback driver rewrite. share/man/man4/Makefile, share/man/man4/xnb.4, sys/dev/xen/netback/netback.c, sys/dev/xen/netback/netback_unit_tests.c: Rewrote the netback driver for xen to attach properly via newbus and work properly in both HVM and PVM mode (only HVM is tested). Works with the in-tree FreeBSD netfront driver or the Windows netfront driver from SuSE. Has not been extensively tested with a Linux netfront driver. Does not implement LRO, TSO, or polling. Includes unit tests that may be run through sysctl after compiling with XNB_DEBUG defined. sys/dev/xen/blkback/blkback.c, sys/xen/interface/io/netif.h: Comment elaboration. sys/kern/uipc_mbuf.c: Fix page fault in kernel mode when calling m_print() on a null mbuf. Since m_print() is only used for debugging, there are no performance concerns for extra error checking code. sys/kern/subr_scanf.c: Add the "hh" and "ll" width specifiers from C99 to scanf(). A few callers were already using "ll" even though scanf() was handling it as "l". Submitted by: Alan Somers <alans@spectralogic.com> Submitted by: John Suykerbuyk <johns@spectralogic.com> Sponsored by: Spectra Logic Reviewed by: ken r230916 | ken | 2012-02-02 10:54:35 -0700 (Thu, 02 Feb 2012) | 13 lines Fix the netback driver build for i386. netback.c: Add missing VM includes. xen/xenvar.h, xen/xenpmap.h: Move some XENHVM macros from <machine/xen/xenpmap.h> to <machine/xen/xenvar.h> on i386 to match the amd64 headers. conf/files: Add netback to the build. Submitted by: jhb Added: stable/9/share/man/man4/xnb.4 - copied unchanged from r230587, head/share/man/man4/xnb.4 stable/9/sys/dev/xen/netback/netback_unit_tests.c - copied unchanged from r230587, head/sys/dev/xen/netback/netback_unit_tests.c Modified: stable/9/share/man/man4/Makefile stable/9/sys/conf/files stable/9/sys/dev/xen/blkback/blkback.c stable/9/sys/dev/xen/netback/netback.c stable/9/sys/i386/include/xen/xenpmap.h stable/9/sys/i386/include/xen/xenvar.h stable/9/sys/kern/subr_scanf.c stable/9/sys/kern/uipc_mbuf.c stable/9/sys/xen/interface/io/netif.h Directory Properties: stable/9/ (props changed) stable/9/share/ (props changed) stable/9/share/man/ (props changed) stable/9/share/man/man4/ (props changed) stable/9/sys/ (props changed) stable/9/sys/conf/ (props changed) Modified: stable/9/share/man/man4/Makefile ============================================================================== --- stable/9/share/man/man4/Makefile Tue Feb 14 17:35:44 2012 (r231696) +++ stable/9/share/man/man4/Makefile Tue Feb 14 18:00:37 2012 (r231697) @@ -528,6 +528,7 @@ MAN= aac.4 \ ${_xen.4} \ xhci.4 \ xl.4 \ + ${_xnb.4} \ xpt.4 \ zero.4 \ zyd.4 @@ -722,6 +723,7 @@ _urtw.4= urtw.4 _viawd.4= viawd.4 _wpi.4= wpi.4 _xen.4= xen.4 +_xnb.4= xnb.4 MLINKS+=lindev.4 full.4 .endif Copied: stable/9/share/man/man4/xnb.4 (from r230587, head/share/man/man4/xnb.4) ============================================================================== --- /dev/null 00:00:00 1970 (empty, because file is newly added) +++ stable/9/share/man/man4/xnb.4 Tue Feb 14 18:00:37 2012 (r231697, copy of r230587, head/share/man/man4/xnb.4) @@ -0,0 +1,134 @@ +.\" Copyright (c) 2012 Spectra Logic Corporation +.\" All rights reserved. +.\" +.\" +.\" Redistribution and use in source and binary forms, with or without +.\" modification, are permitted provided that the following conditions +.\" are met: +.\" 1. Redistributions of source code must retain the above copyright +.\" notice, this list of conditions, and the following disclaimer, +.\" without modification. +.\" 2. Redistributions in binary form must reproduce at minimum a disclaimer +.\" substantially similar to the "NO WARRANTY" disclaimer below +.\" ("Disclaimer") and any redistribution must be conditioned upon +.\" including a substantially similar Disclaimer requirement for further +.\" binary redistribution. +.\" +.\" NO WARRANTY +.\" THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS +.\" "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT +.\" LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR +.\" A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT +.\" HOLDERS OR CONTRIBUTORS BE LIABLE FOR SPECIAL, EXEMPLARY, OR CONSEQUENTIAL +.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS +.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) +.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, +.\" STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING +.\" IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE +.\" POSSIBILITY OF SUCH DAMAGES. +.\" +.\" Authors: Alan Somers (Spectra Logic Corporation) +.\" +.\" $FreeBSD$ +.\" + +.Dd January 6, 2012 +.Dt XNB 4 +.Os +.Sh NAME +.Nm xnb +.Nd "Xen Paravirtualized Backend Ethernet Driver" +.Sh SYNOPSIS +To compile this driver into the kernel, place the following lines in your +kernel configuration file: +.Bd -ragged -offset indent +.Cd "options XENHVM" +.Cd "device xenpci" +.Ed +.Sh DESCRIPTION +The +.Nm +driver provides the back half of a paravirtualized +.Xr xen 4 +network connection. The netback and netfront drivers appear to their +respective operating systems as Ethernet devices linked by a crossover cable. +Typically, +.Nm +will run on Domain 0 and the netfront driver will run on a guest domain. +However, it is also possible to run +.Nm +on a guest domain. It may be bridged or routed to provide the netfront's +domain access to other guest domains or to a physical network. +.Pp +In most respects, the +.Nm +device appears to the OS as an other Ethernet device. It can be configured at +runtime entirely with +.Xr ifconfig 8 +\&. In particular, it supports MAC changing, arbitrary MTU sizes, checksum +offload for IP, UDP, and TCP for both receive and transmit, and TSO. However, +see +.Sx CAVEATS +before enabling txcsum, rxcsum, or tso. +.Sh SYSCTL VARIABLES +The following read-only variables are available via +.Xr sysctl 8 : +.Bl -tag -width indent +.It Va dev.xnb.%d.dump_rings +Displays information about the ring buffers used to pass requests between the +netfront and netback. Mostly useful for debugging, but can also be used to +get traffic statistics. +.It Va dev.xnb.%d.unit_test_results +Runs a builtin suite of unit tests and displays the results. Does not affect +the operation of the driver in any way. Note that the test suite simulates +error conditions; this will result in error messages being printed to the +system system log. +.Sh CAVEATS +Packets sent through Xennet pass over shared memory, so the protocol includes +no form of link-layer checksum or CRC. Furthermore, Xennet drivers always +report to their hosts that they support receive and transmit checksum +offloading. They "offload" the checksum calculation by simply skipping it. +That works fine for packets that are exchanged between two domains on the same +machine. However, when a Xennet interface is bridged to a physical interface, +a correct checksum must be attached to any packets bound for that physical +interface. Currently, FreeBSD lacks any mechanism for an ethernet device to +inform the OS that newly received packets are valid even though their checksums +are not. So if the netfront driver is configured to offload checksum +calculations, it will pass non-checksumed packets to +.Nm +, which must then calculate the checksum in software before passing the packet +to the OS. +.Pp +For this reason, it is recommended that if +.Nm +is bridged to a physcal interface, then transmit checksum offloading should be +disabled on the netfront. The Xennet protocol does not have any mechanism for +the netback to request the netfront to do this; the operator must do it +manually. +.Sh SEE ALSO +.Xr arp 4 , +.Xr netintro 4 , +.Xr ng_ether 4 , +.Xr ifconfig 8 , +.Xr xen 4 +.Sh HISTORY +The +.Nm +device driver first appeared in +.Fx 10.0 +. +.Sh AUTHORS +The +.Nm +driver was written by +.An Alan Somers +.Aq alans@spectralogic.com +and +.An John Suykerbuyk +.Aq johns@spectralogic.com +.Sh BUGS +The +.Nm +driver does not properly checksum UDP datagrams that span more than one +Ethernet frame. Nor does it correctly checksum IPv6 packets. To workaround +that bug, disable transmit checksum offloading on the netfront driver. Modified: stable/9/sys/conf/files ============================================================================== --- stable/9/sys/conf/files Tue Feb 14 17:35:44 2012 (r231696) +++ stable/9/sys/conf/files Tue Feb 14 18:00:37 2012 (r231697) @@ -3493,6 +3493,7 @@ dev/xen/blkback/blkback.c optional xen | dev/xen/console/console.c optional xen dev/xen/console/xencons_ring.c optional xen dev/xen/control/control.c optional xen | xenhvm +dev/xen/netback/netback.c optional xen | xenhvm dev/xen/netfront/netfront.c optional xen | xenhvm dev/xen/xenpci/xenpci.c optional xenpci dev/xen/xenpci/evtchn.c optional xenpci Modified: stable/9/sys/dev/xen/blkback/blkback.c ============================================================================== --- stable/9/sys/dev/xen/blkback/blkback.c Tue Feb 14 17:35:44 2012 (r231696) +++ stable/9/sys/dev/xen/blkback/blkback.c Tue Feb 14 18:00:37 2012 (r231697) @@ -3434,6 +3434,10 @@ xbb_shutdown(struct xbb_softc *xbb) DPRINTF("\n"); + /* + * Before unlocking mutex, set this flag to prevent other threads from + * getting into this function + */ xbb->flags |= XBBF_IN_SHUTDOWN; mtx_unlock(&xbb->lock); Modified: stable/9/sys/dev/xen/netback/netback.c ============================================================================== --- stable/9/sys/dev/xen/netback/netback.c Tue Feb 14 17:35:44 2012 (r231696) +++ stable/9/sys/dev/xen/netback/netback.c Tue Feb 14 18:00:37 2012 (r231697) @@ -1,1596 +1,2537 @@ -/* - * Copyright (c) 2006, Cisco Systems, Inc. +/*- + * Copyright (c) 2009-2011 Spectra Logic Corporation * All rights reserved. * - * Redistribution and use in source and binary forms, with or without - * modification, are permitted provided that the following conditions + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions, and the following disclaimer, + * without modification. + * 2. Redistributions in binary form must reproduce at minimum a disclaimer + * substantially similar to the "NO WARRANTY" disclaimer below + * ("Disclaimer") and any redistribution must be conditioned upon + * including a substantially similar Disclaimer requirement for further + * binary redistribution. * - * 1. Redistributions of source code must retain the above copyright - * notice, this list of conditions and the following disclaimer. - * 2. Redistributions in binary form must reproduce the above copyright - * notice, this list of conditions and the following disclaimer in the - * documentation and/or other materials provided with the distribution. - * 3. Neither the name of Cisco Systems, Inc. nor the names of its contributors - * may be used to endorse or promote products derived from this software - * without specific prior written permission. - * - * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" - * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE - * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE - * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE - * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR - * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF - * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS - * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN - * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) - * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE - * POSSIBILITY OF SUCH DAMAGE. + * NO WARRANTY + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS + * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT + * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR + * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT + * HOLDERS OR CONTRIBUTORS BE LIABLE FOR SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, + * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING + * IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE + * POSSIBILITY OF SUCH DAMAGES. + * + * Authors: Justin T. Gibbs (Spectra Logic Corporation) + * Alan Somers (Spectra Logic Corporation) + * John Suykerbuyk (Spectra Logic Corporation) */ #include <sys/cdefs.h> __FBSDID("$FreeBSD$"); + +/** + * \file netback.c + * + * \brief Device driver supporting the vending of network access + * from this FreeBSD domain to other domains. + */ +#include "opt_inet.h" +#include "opt_global.h" + #include "opt_sctp.h" #include <sys/param.h> -#include <sys/systm.h> -#include <sys/sockio.h> -#include <sys/mbuf.h> -#include <sys/malloc.h> #include <sys/kernel.h> -#include <sys/socket.h> -#include <sys/queue.h> -#include <sys/taskqueue.h> -#include <sys/module.h> #include <sys/bus.h> +#include <sys/module.h> +#include <sys/rman.h> +#include <sys/socket.h> +#include <sys/sockio.h> #include <sys/sysctl.h> #include <net/if.h> #include <net/if_arp.h> -#include <net/if_types.h> #include <net/ethernet.h> -#include <net/if_bridgevar.h> +#include <net/if_dl.h> +#include <net/if_media.h> +#include <net/if_types.h> -#include <netinet/in_systm.h> #include <netinet/in.h> -#include <netinet/in_var.h> #include <netinet/ip.h> +#include <netinet/if_ether.h> +#if __FreeBSD_version >= 700000 #include <netinet/tcp.h> -#include <netinet/udp.h> -#ifdef SCTP -#include <netinet/sctp.h> -#include <netinet/sctp_crc32.h> #endif +#include <netinet/ip_icmp.h> +#include <netinet/udp.h> +#include <machine/in_cksum.h> +#include <vm/vm.h> +#include <vm/pmap.h> #include <vm/vm_extern.h> #include <vm/vm_kern.h> -#include <machine/in_cksum.h> -#include <machine/xen-os.h> -#include <machine/hypervisor.h> -#include <machine/hypervisor-ifs.h> -#include <machine/xen_intr.h> -#include <machine/evtchn.h> -#include <machine/xenbus.h> -#include <machine/gnttab.h> -#include <machine/xen-public/memory.h> -#include <dev/xen/xenbus/xenbus_comms.h> +#include <machine/_inttypes.h> +#include <machine/xen/xen-os.h> +#include <machine/xen/xenvar.h> + +#include <xen/evtchn.h> +#include <xen/xen_intr.h> +#include <xen/interface/io/netif.h> +#include <xen/xenbus/xenbusvar.h> +/*--------------------------- Compile-time Tunables --------------------------*/ + +/*---------------------------------- Macros ----------------------------------*/ +/** + * Custom malloc type for all driver allocations. + */ +static MALLOC_DEFINE(M_XENNETBACK, "xnb", "Xen Net Back Driver Data"); -#ifdef XEN_NETBACK_DEBUG -#define DPRINTF(fmt, args...) \ - printf("netback (%s:%d): " fmt, __FUNCTION__, __LINE__, ##args) +#define XNB_SG 1 /* netback driver supports feature-sg */ +#define XNB_GSO_TCPV4 1 /* netback driver supports feature-gso-tcpv4 */ +#define XNB_RX_COPY 1 /* netback driver supports feature-rx-copy */ +#define XNB_RX_FLIP 0 /* netback driver does not support feature-rx-flip */ + +#undef XNB_DEBUG +#define XNB_DEBUG /* hardcode on during development */ + +#ifdef XNB_DEBUG +#define DPRINTF(fmt, args...) \ + printf("xnb(%s:%d): " fmt, __FUNCTION__, __LINE__, ##args) #else -#define DPRINTF(fmt, args...) ((void)0) +#define DPRINTF(fmt, args...) do {} while (0) #endif -#ifdef XEN_NETBACK_DEBUG_LOTS -#define DDPRINTF(fmt, args...) \ - printf("netback (%s:%d): " fmt, __FUNCTION__, __LINE__, ##args) -#define DPRINTF_MBUF(_m) print_mbuf(_m, 0) -#define DPRINTF_MBUF_LEN(_m, _len) print_mbuf(_m, _len) -#else -#define DDPRINTF(fmt, args...) ((void)0) -#define DPRINTF_MBUF(_m) ((void)0) -#define DPRINTF_MBUF_LEN(_m, _len) ((void)0) +/* Default length for stack-allocated grant tables */ +#define GNTTAB_LEN (64) + +/* Features supported by all backends. TSO and LRO can be negotiated */ +#define XNB_CSUM_FEATURES (CSUM_TCP | CSUM_UDP) + +#define NET_TX_RING_SIZE __RING_SIZE((netif_tx_sring_t *)0, PAGE_SIZE) +#define NET_RX_RING_SIZE __RING_SIZE((netif_rx_sring_t *)0, PAGE_SIZE) + +/** + * Two argument version of the standard macro. Second argument is a tentative + * value of req_cons + */ +#define RING_HAS_UNCONSUMED_REQUESTS_2(_r, cons) ({ \ + unsigned int req = (_r)->sring->req_prod - cons; \ + unsigned int rsp = RING_SIZE(_r) - \ + (cons - (_r)->rsp_prod_pvt); \ + req < rsp ? req : rsp; \ +}) + +#define virt_to_mfn(x) (vtomach(x) >> PAGE_SHIFT) +#define virt_to_offset(x) ((x) & (PAGE_SIZE - 1)) + +/** + * Predefined array type of grant table copy descriptors. Used to pass around + * statically allocated memory structures. + */ +typedef struct gnttab_copy gnttab_copy_table[GNTTAB_LEN]; + +/*--------------------------- Forward Declarations ---------------------------*/ +struct xnb_softc; +struct xnb_pkt; + +static void xnb_attach_failed(struct xnb_softc *xnb, + int err, const char *fmt, ...) + __printflike(3,4); +static int xnb_shutdown(struct xnb_softc *xnb); +static int create_netdev(device_t dev); +static int xnb_detach(device_t dev); +static int xen_net_read_mac(device_t dev, uint8_t mac[]); +static int xnb_ifmedia_upd(struct ifnet *ifp); +static void xnb_ifmedia_sts(struct ifnet *ifp, struct ifmediareq *ifmr); +static void xnb_intr(void *arg); +static int xnb_send(netif_rx_back_ring_t *rxb, domid_t otherend, + const struct mbuf *mbufc, gnttab_copy_table gnttab); +static int xnb_recv(netif_tx_back_ring_t *txb, domid_t otherend, + struct mbuf **mbufc, struct ifnet *ifnet, + gnttab_copy_table gnttab); +static int xnb_ring2pkt(struct xnb_pkt *pkt, + const netif_tx_back_ring_t *tx_ring, + RING_IDX start); +static void xnb_txpkt2rsp(const struct xnb_pkt *pkt, + netif_tx_back_ring_t *ring, int error); +static struct mbuf *xnb_pkt2mbufc(const struct xnb_pkt *pkt, struct ifnet *ifp); +static int xnb_txpkt2gnttab(const struct xnb_pkt *pkt, + const struct mbuf *mbufc, + gnttab_copy_table gnttab, + const netif_tx_back_ring_t *txb, + domid_t otherend_id); +static void xnb_update_mbufc(struct mbuf *mbufc, + const gnttab_copy_table gnttab, int n_entries); +static int xnb_mbufc2pkt(const struct mbuf *mbufc, + struct xnb_pkt *pkt, + RING_IDX start, int space); +static int xnb_rxpkt2gnttab(const struct xnb_pkt *pkt, + const struct mbuf *mbufc, + gnttab_copy_table gnttab, + const netif_rx_back_ring_t *rxb, + domid_t otherend_id); +static int xnb_rxpkt2rsp(const struct xnb_pkt *pkt, + const gnttab_copy_table gnttab, int n_entries, + netif_rx_back_ring_t *ring); +static void xnb_add_mbuf_cksum(struct mbuf *mbufc); +static void xnb_stop(struct xnb_softc*); +static int xnb_ioctl(struct ifnet*, u_long, caddr_t); +static void xnb_start_locked(struct ifnet*); +static void xnb_start(struct ifnet*); +static void xnb_ifinit_locked(struct xnb_softc*); +static void xnb_ifinit(void*); +#ifdef XNB_DEBUG +static int xnb_unit_test_main(SYSCTL_HANDLER_ARGS); +static int xnb_dump_rings(SYSCTL_HANDLER_ARGS); #endif +/*------------------------------ Data Structures -----------------------------*/ -#define WPRINTF(fmt, args...) \ - printf("netback (%s:%d): " fmt, __FUNCTION__, __LINE__, ##args) -#define ARRAY_SIZE(x) (sizeof(x)/sizeof(x[0])) -#define BUG_ON PANIC_IF +/** + * Representation of a xennet packet. Simplified version of a packet as + * stored in the Xen tx ring. Applicable to both RX and TX packets + */ +struct xnb_pkt{ + /** + * Array index of the first data-bearing (eg, not extra info) entry + * for this packet + */ + RING_IDX car; + + /** + * Array index of the second data-bearing entry for this packet. + * Invalid if the packet has only one data-bearing entry. If the + * packet has more than two data-bearing entries, then the second + * through the last will be sequential modulo the ring size + */ + RING_IDX cdr; + + /** + * Optional extra info. Only valid if flags contains + * NETTXF_extra_info. Note that extra.type will always be + * XEN_NETIF_EXTRA_TYPE_GSO. Currently, no known netfront or netback + * driver will ever set XEN_NETIF_EXTRA_TYPE_MCAST_* + */ + netif_extra_info_t extra; + + /** Size of entire packet in bytes. */ + uint16_t size; -#define IFNAME(_np) (_np)->ifp->if_xname + /** The size of the first entry's data in bytes */ + uint16_t car_size; -#define NET_TX_RING_SIZE __RING_SIZE((netif_tx_sring_t *)0, PAGE_SIZE) -#define NET_RX_RING_SIZE __RING_SIZE((netif_rx_sring_t *)0, PAGE_SIZE) + /** + * Either NETTXF_ or NETRXF_ flags. Note that the flag values are + * not the same for TX and RX packets + */ + uint16_t flags; + + /** + * The number of valid data-bearing entries (either netif_tx_request's + * or netif_rx_response's) in the packet. If this is 0, it means the + * entire packet is invalid. + */ + uint16_t list_len; -struct ring_ref { - vm_offset_t va; - grant_handle_t handle; - uint64_t bus_addr; + /** There was an error processing the packet */ + uint8_t error; }; -typedef struct netback_info { +/** xnb_pkt method: initialize it */ +static inline void +xnb_pkt_initialize(struct xnb_pkt *pxnb) +{ + bzero(pxnb, sizeof(*pxnb)); +} - /* Schedule lists */ - STAILQ_ENTRY(netback_info) next_tx; - STAILQ_ENTRY(netback_info) next_rx; - int on_tx_sched_list; - int on_rx_sched_list; - - struct xenbus_device *xdev; - XenbusState frontend_state; - - domid_t domid; - int handle; - char *bridge; - - int rings_connected; - struct ring_ref tx_ring_ref; - struct ring_ref rx_ring_ref; - netif_tx_back_ring_t tx; - netif_rx_back_ring_t rx; - evtchn_port_t evtchn; - int irq; - void *irq_cookie; +/** xnb_pkt method: mark the packet as valid */ +static inline void +xnb_pkt_validate(struct xnb_pkt *pxnb) +{ + pxnb->error = 0; +}; - struct ifnet *ifp; - int ref_cnt; +/** xnb_pkt method: mark the packet as invalid */ +static inline void +xnb_pkt_invalidate(struct xnb_pkt *pxnb) +{ + pxnb->error = 1; +}; - device_t ndev; - int attached; -} netif_t; - - -#define MAX_PENDING_REQS 256 -#define PKT_PROT_LEN 64 - -static struct { - netif_tx_request_t req; - netif_t *netif; -} pending_tx_info[MAX_PENDING_REQS]; -static uint16_t pending_ring[MAX_PENDING_REQS]; -typedef unsigned int PEND_RING_IDX; -#define MASK_PEND_IDX(_i) ((_i)&(MAX_PENDING_REQS-1)) -static PEND_RING_IDX pending_prod, pending_cons; -#define NR_PENDING_REQS (MAX_PENDING_REQS - pending_prod + pending_cons) - -static unsigned long mmap_vstart; -#define MMAP_VADDR(_req) (mmap_vstart + ((_req) * PAGE_SIZE)) - -/* Freed TX mbufs get batched on this ring before return to pending_ring. */ -static uint16_t dealloc_ring[MAX_PENDING_REQS]; -static PEND_RING_IDX dealloc_prod, dealloc_cons; - -static multicall_entry_t rx_mcl[NET_RX_RING_SIZE+1]; -static mmu_update_t rx_mmu[NET_RX_RING_SIZE]; -static gnttab_transfer_t grant_rx_op[NET_RX_RING_SIZE]; - -static grant_handle_t grant_tx_handle[MAX_PENDING_REQS]; -static gnttab_unmap_grant_ref_t tx_unmap_ops[MAX_PENDING_REQS]; -static gnttab_map_grant_ref_t tx_map_ops[MAX_PENDING_REQS]; - -static struct task net_tx_task, net_rx_task; -static struct callout rx_task_callout; - -static STAILQ_HEAD(netback_tx_sched_list, netback_info) tx_sched_list = - STAILQ_HEAD_INITIALIZER(tx_sched_list); -static STAILQ_HEAD(netback_rx_sched_list, netback_info) rx_sched_list = - STAILQ_HEAD_INITIALIZER(rx_sched_list); -static struct mtx tx_sched_list_lock; -static struct mtx rx_sched_list_lock; - -static int vif_unit_maker = 0; - -/* Protos */ -static void netback_start(struct ifnet *ifp); -static int netback_ioctl(struct ifnet *ifp, u_long cmd, caddr_t data); -static int vif_add_dev(struct xenbus_device *xdev); -static void disconnect_rings(netif_t *netif); +/** xnb_pkt method: Check whether the packet is valid */ +static inline int +xnb_pkt_is_valid(const struct xnb_pkt *pxnb) +{ + return (! pxnb->error); +} + +#ifdef XNB_DEBUG +/** xnb_pkt method: print the packet's contents in human-readable format*/ +static void __unused +xnb_dump_pkt(const struct xnb_pkt *pkt) { + if (pkt == NULL) { + DPRINTF("Was passed a null pointer.\n"); + return; + } + DPRINTF("pkt address= %p\n", pkt); + DPRINTF("pkt->size=%d\n", pkt->size); + DPRINTF("pkt->car_size=%d\n", pkt->car_size); + DPRINTF("pkt->flags=0x%04x\n", pkt->flags); + DPRINTF("pkt->list_len=%d\n", pkt->list_len); + /* DPRINTF("pkt->extra"); TODO */ + DPRINTF("pkt->car=%d\n", pkt->car); + DPRINTF("pkt->cdr=%d\n", pkt->cdr); + DPRINTF("pkt->error=%d\n", pkt->error); +} +#endif /* XNB_DEBUG */ -#ifdef XEN_NETBACK_DEBUG_LOTS -/* Debug code to display the contents of an mbuf */ static void -print_mbuf(struct mbuf *m, int max) +xnb_dump_txreq(RING_IDX idx, const struct netif_tx_request *txreq) { - int i, j=0; - printf("mbuf %08x len = %d", (unsigned int)m, m->m_pkthdr.len); - for (; m; m = m->m_next) { - unsigned char *d = m->m_data; - for (i=0; i < m->m_len; i++) { - if (max && j == max) - break; - if ((j++ % 16) == 0) - printf("\n%04x:", j); - printf(" %02x", d[i]); - } + if (txreq != NULL) { + DPRINTF("netif_tx_request index =%u\n", idx); + DPRINTF("netif_tx_request.gref =%u\n", txreq->gref); + DPRINTF("netif_tx_request.offset=%hu\n", txreq->offset); + DPRINTF("netif_tx_request.flags =%hu\n", txreq->flags); + DPRINTF("netif_tx_request.id =%hu\n", txreq->id); + DPRINTF("netif_tx_request.size =%hu\n", txreq->size); } - printf("\n"); } -#endif -#define MAX_MFN_ALLOC 64 -static unsigned long mfn_list[MAX_MFN_ALLOC]; -static unsigned int alloc_index = 0; +/** + * \brief Configuration data for a shared memory request ring + * used to communicate with the front-end client of this + * this driver. + */ +struct xnb_ring_config { + /** + * Runtime structures for ring access. Unfortunately, TX and RX rings + * use different data structures, and that cannot be changed since it + * is part of the interdomain protocol. + */ + union{ + netif_rx_back_ring_t rx_ring; + netif_tx_back_ring_t tx_ring; + } back_ring; + + /** + * The device bus address returned by the hypervisor when + * mapping the ring and required to unmap it when a connection + * is torn down. + */ + uint64_t bus_addr; + + /** The pseudo-physical address where ring memory is mapped.*/ + uint64_t gnt_addr; -static unsigned long -alloc_mfn(void) -{ - unsigned long mfn = 0; - struct xen_memory_reservation reservation = { - .extent_start = mfn_list, - .nr_extents = MAX_MFN_ALLOC, - .extent_order = 0, - .domid = DOMID_SELF - }; - if ( unlikely(alloc_index == 0) ) - alloc_index = HYPERVISOR_memory_op( - XENMEM_increase_reservation, &reservation); - if ( alloc_index != 0 ) - mfn = mfn_list[--alloc_index]; - return mfn; -} + /** KVA address where ring memory is mapped. */ + vm_offset_t va; -static unsigned long -alloc_empty_page_range(unsigned long nr_pages) + /** + * Grant table handles, one per-ring page, returned by the + * hyperpervisor upon mapping of the ring and required to + * unmap it when a connection is torn down. + */ + grant_handle_t handle; + + /** The number of ring pages mapped for the current connection. */ + unsigned ring_pages; + + /** + * The grant references, one per-ring page, supplied by the + * front-end, allowing us to reference the ring pages in the + * front-end's domain and to map these pages into our own domain. + */ + grant_ref_t ring_ref; +}; + +/** + * Per-instance connection state flags. + */ +typedef enum { - void *pages; - int i = 0, j = 0; - multicall_entry_t mcl[17]; - unsigned long mfn_list[16]; - struct xen_memory_reservation reservation = { - .extent_start = mfn_list, - .nr_extents = 0, - .address_bits = 0, - .extent_order = 0, - .domid = DOMID_SELF - }; + /** Communication with the front-end has been established. */ + XNBF_RING_CONNECTED = 0x01, - pages = malloc(nr_pages*PAGE_SIZE, M_DEVBUF, M_NOWAIT); - if (pages == NULL) - return 0; + /** + * Front-end requests exist in the ring and are waiting for + * xnb_xen_req objects to free up. + */ + XNBF_RESOURCE_SHORTAGE = 0x02, + + /** Connection teardown has started. */ + XNBF_SHUTDOWN = 0x04, - memset(mcl, 0, sizeof(mcl)); + /** A thread is already performing shutdown processing. */ + XNBF_IN_SHUTDOWN = 0x08 +} xnb_flag_t; - while (i < nr_pages) { - unsigned long va = (unsigned long)pages + (i++ * PAGE_SIZE); +/** + * Types of rings. Used for array indices and to identify a ring's control + * data structure type + */ +typedef enum{ + XNB_RING_TYPE_TX = 0, /* ID of TX rings, used for array indices */ + XNB_RING_TYPE_RX = 1, /* ID of RX rings, used for array indices */ + XNB_NUM_RING_TYPES +} xnb_ring_type_t; - mcl[j].op = __HYPERVISOR_update_va_mapping; - mcl[j].args[0] = va; +/** + * Per-instance configuration data. + */ +struct xnb_softc { + /** NewBus device corresponding to this instance. */ + device_t dev; + + /* Media related fields */ + + /** Generic network media state */ + struct ifmedia sc_media; + + /** Media carrier info */ + struct ifnet *xnb_ifp; + + /** Our own private carrier state */ + unsigned carrier; + + /** Device MAC Address */ + uint8_t mac[ETHER_ADDR_LEN]; + + /* Xen related fields */ + + /** + * \brief The netif protocol abi in effect. + * + * There are situations where the back and front ends can + * have a different, native abi (e.g. intel x86_64 and + * 32bit x86 domains on the same machine). The back-end + * always accomodates the front-end's native abi. That + * value is pulled from the XenStore and recorded here. + */ + int abi; - mfn_list[j++] = vtomach(va) >> PAGE_SHIFT; + /** + * Name of the bridge to which this VIF is connected, if any + * This field is dynamically allocated by xenbus and must be free()ed + * when no longer needed + */ + char *bridge; - xen_phys_machine[(vtophys(va) >> PAGE_SHIFT)] = INVALID_P2M_ENTRY; + /** The interrupt driven even channel used to signal ring events. */ + evtchn_port_t evtchn; - if (j == 16 || i == nr_pages) { - mcl[j-1].args[MULTI_UVMFLAGS_INDEX] = UVMF_TLB_FLUSH|UVMF_LOCAL; + /** Xen device handle.*/ + long handle; - reservation.nr_extents = j; + /** IRQ mapping for the communication ring event channel. */ + int irq; + + /** + * \brief Cached value of the front-end's domain id. + * + * This value is used at once for each mapped page in + * a transaction. We cache it to avoid incuring the + * cost of an ivar access every time this is needed. + */ + domid_t otherend_id; - mcl[j].op = __HYPERVISOR_memory_op; - mcl[j].args[0] = XENMEM_decrease_reservation; - mcl[j].args[1] = (unsigned long)&reservation; - - (void)HYPERVISOR_multicall(mcl, j+1); + /** + * Undocumented frontend feature. Has something to do with + * scatter/gather IO + */ + uint8_t can_sg; + /** Undocumented frontend feature */ + uint8_t gso; + /** Undocumented frontend feature */ + uint8_t gso_prefix; + /** Can checksum TCP/UDP over IPv4 */ + uint8_t ip_csum; + + /* Implementation related fields */ + /** + * Preallocated grant table copy descriptor for RX operations. + * Access must be protected by rx_lock + */ + gnttab_copy_table rx_gnttab; - mcl[j-1].args[MULTI_UVMFLAGS_INDEX] = 0; - j = 0; - } - } + /** + * Preallocated grant table copy descriptor for TX operations. + * Access must be protected by tx_lock + */ + gnttab_copy_table tx_gnttab; - return (unsigned long)pages; -} +#ifdef XENHVM + /** + * Resource representing allocated physical address space + * associated with our per-instance kva region. + */ + struct resource *pseudo_phys_res; -#ifdef XEN_NETBACK_FIXUP_CSUM -static void -fixup_checksum(struct mbuf *m) -{ - struct ether_header *eh = mtod(m, struct ether_header *); - struct ip *ip = (struct ip *)(eh + 1); - int iphlen = ip->ip_hl << 2; - int iplen = ntohs(ip->ip_len); - - if ((m->m_pkthdr.csum_flags & CSUM_TCP)) { - struct tcphdr *th = (struct tcphdr *)((caddr_t)ip + iphlen); - th->th_sum = in_pseudo(ip->ip_src.s_addr, ip->ip_dst.s_addr, - htons(IPPROTO_TCP + (iplen - iphlen))); - th->th_sum = in_cksum_skip(m, iplen + sizeof(*eh), sizeof(*eh) + iphlen); - m->m_pkthdr.csum_flags &= ~CSUM_TCP; -#ifdef SCTP - } else if (sw_csum & CSUM_SCTP) { - sctp_delayed_cksum(m, iphlen); - sw_csum &= ~CSUM_SCTP; -#endif - } else { - u_short csum; - struct udphdr *uh = (struct udphdr *)((caddr_t)ip + iphlen); - uh->uh_sum = in_pseudo(ip->ip_src.s_addr, ip->ip_dst.s_addr, - htons(IPPROTO_UDP + (iplen - iphlen))); - if ((csum = in_cksum_skip(m, iplen + sizeof(*eh), sizeof(*eh) + iphlen)) == 0) - csum = 0xffff; - uh->uh_sum = csum; - m->m_pkthdr.csum_flags &= ~CSUM_UDP; - } -} + /** Resource id for allocated physical address space. */ + int pseudo_phys_res_id; #endif -/* Add the interface to the specified bridge */ -static int -add_to_bridge(struct ifnet *ifp, char *bridge) -{ - struct ifdrv ifd; - struct ifbreq ifb; - struct ifnet *ifp_bridge = ifunit(bridge); + /** Ring mapping and interrupt configuration data. */ + struct xnb_ring_config ring_configs[XNB_NUM_RING_TYPES]; - if (!ifp_bridge) - return ENOENT; + /** + * Global pool of kva used for mapping remote domain ring + * and I/O transaction data. + */ + vm_offset_t kva; - bzero(&ifd, sizeof(ifd)); - bzero(&ifb, sizeof(ifb)); + /** Psuedo-physical address corresponding to kva. */ + uint64_t gnt_base_addr; - strcpy(ifb.ifbr_ifsname, ifp->if_xname); - strcpy(ifd.ifd_name, ifp->if_xname); - ifd.ifd_cmd = BRDGADD; - ifd.ifd_len = sizeof(ifb); - ifd.ifd_data = &ifb; + /** Various configuration and state bit flags. */ + xnb_flag_t flags; - return bridge_ioctl_kern(ifp_bridge, SIOCSDRVSPEC, &ifd); - -} + /** Mutex protecting per-instance data in the receive path. */ + struct mtx rx_lock; -static int -netif_create(int handle, struct xenbus_device *xdev, char *bridge) -{ - netif_t *netif; - struct ifnet *ifp; + /** Mutex protecting per-instance data in the softc structure. */ + struct mtx sc_lock; - netif = (netif_t *)malloc(sizeof(*netif), M_DEVBUF, M_NOWAIT | M_ZERO); - if (!netif) - return ENOMEM; + /** Mutex protecting per-instance data in the transmit path. */ + struct mtx tx_lock; - netif->ref_cnt = 1; - netif->handle = handle; - netif->domid = xdev->otherend_id; - netif->xdev = xdev; - netif->bridge = bridge; - xdev->data = netif; - - /* Set up ifnet structure */ - ifp = netif->ifp = if_alloc(IFT_ETHER); - if (!ifp) { - if (bridge) - free(bridge, M_DEVBUF); - free(netif, M_DEVBUF); - return ENOMEM; - } + /** The size of the global kva pool. */ + int kva_size; +}; - ifp->if_softc = netif; - if_initname(ifp, "vif", - atomic_fetchadd_int(&vif_unit_maker, 1) /* ifno */ ); - ifp->if_flags = IFF_BROADCAST | IFF_SIMPLEX; - ifp->if_output = ether_output; - ifp->if_start = netback_start; - ifp->if_ioctl = netback_ioctl; - ifp->if_mtu = ETHERMTU; - ifp->if_snd.ifq_maxlen = NET_TX_RING_SIZE - 1; - - DPRINTF("Created %s for domid=%d handle=%d\n", IFNAME(netif), netif->domid, netif->handle); +/*---------------------------- Debugging functions ---------------------------*/ +#ifdef XNB_DEBUG +static void __unused *** DIFF OUTPUT TRUNCATED AT 1000 LINES ***
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201202141800.q1EI0bHb011087>