diff options
Diffstat (limited to 'net')
56 files changed, 1071 insertions, 444 deletions
diff --git a/net/8021q/vlan.c b/net/8021q/vlan.c index 2a739adaa92..51961300b58 100644 --- a/net/8021q/vlan.c +++ b/net/8021q/vlan.c @@ -382,6 +382,24 @@ static void vlan_sync_address(struct net_device *dev, memcpy(vlan->real_dev_addr, dev->dev_addr, ETH_ALEN); } +static void vlan_transfer_features(struct net_device *dev, + struct net_device *vlandev) +{ + unsigned long old_features = vlandev->features; + + if (dev->features & NETIF_F_VLAN_TSO) { + vlandev->features &= ~VLAN_TSO_FEATURES; + vlandev->features |= dev->features & VLAN_TSO_FEATURES; + } + if (dev->features & NETIF_F_VLAN_CSUM) { + vlandev->features &= ~NETIF_F_ALL_CSUM; + vlandev->features |= dev->features & NETIF_F_ALL_CSUM; + } + + if (old_features != vlandev->features) + netdev_features_change(vlandev); +} + static void __vlan_device_event(struct net_device *dev, unsigned long event) { switch (event) { @@ -410,10 +428,8 @@ static int vlan_device_event(struct notifier_block *unused, unsigned long event, int i, flgs; struct net_device *vlandev; - if (is_vlan_dev(dev)) { + if (is_vlan_dev(dev)) __vlan_device_event(dev, event); - goto out; - } grp = __vlan_find_group(dev); if (!grp) @@ -450,6 +466,18 @@ static int vlan_device_event(struct notifier_block *unused, unsigned long event, } break; + case NETDEV_FEAT_CHANGE: + /* Propagate device features to underlying device */ + for (i = 0; i < VLAN_GROUP_ARRAY_LEN; i++) { + vlandev = vlan_group_get_device(grp, i); + if (!vlandev) + continue; + + vlan_transfer_features(dev, vlandev); + } + + break; + case NETDEV_DOWN: /* Put all VLANs for this dev in the down state too. */ for (i = 0; i < VLAN_GROUP_ARRAY_LEN; i++) { diff --git a/net/8021q/vlan.h b/net/8021q/vlan.h index 5229a72c7ea..79625696e86 100644 --- a/net/8021q/vlan.h +++ b/net/8021q/vlan.h @@ -7,6 +7,8 @@ #define VLAN_GRP_HASH_SIZE (1 << VLAN_GRP_HASH_SHIFT) #define VLAN_GRP_HASH_MASK (VLAN_GRP_HASH_SIZE - 1) +#define VLAN_TSO_FEATURES (NETIF_F_TSO | NETIF_F_TSO6 | NETIF_F_SG) + /* Find a VLAN device by the MAC address of its Ethernet device, and * it's VLAN ID. The default configuration is to have VLAN's scope * to be box-wide, so the MAC will be ignored. The mac will only be diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c index c961f082600..b1cfbaa88db 100644 --- a/net/8021q/vlan_dev.c +++ b/net/8021q/vlan_dev.c @@ -663,6 +663,11 @@ static int vlan_dev_init(struct net_device *dev) (1<<__LINK_STATE_DORMANT))) | (1<<__LINK_STATE_PRESENT); + if (real_dev->features & NETIF_F_VLAN_TSO) + dev->features |= real_dev->features & VLAN_TSO_FEATURES; + if (real_dev->features & NETIF_F_VLAN_CSUM) + dev->features |= real_dev->features & NETIF_F_ALL_CSUM; + /* ipv6 shared card related stuff */ dev->dev_id = real_dev->dev_id; diff --git a/net/9p/Kconfig b/net/9p/Kconfig index bafc50c9e6f..ff34c5acc13 100644 --- a/net/9p/Kconfig +++ b/net/9p/Kconfig @@ -13,16 +13,6 @@ menuconfig NET_9P If unsure, say N. -config NET_9P_FD - depends on NET_9P - default y if NET_9P - tristate "9P File Descriptor Transports (Experimental)" - help - This builds support for file descriptor transports for 9p - which includes support for TCP/IP, named pipes, or passed - file descriptors. TCP/IP is the default transport for 9p, - so if you are going to use 9p, you'll likely want this. - config NET_9P_VIRTIO depends on NET_9P && EXPERIMENTAL && VIRTIO tristate "9P Virtio Transport (Experimental)" diff --git a/net/9p/Makefile b/net/9p/Makefile index 8a105110189..519219480db 100644 --- a/net/9p/Makefile +++ b/net/9p/Makefile @@ -1,5 +1,4 @@ obj-$(CONFIG_NET_9P) := 9pnet.o -obj-$(CONFIG_NET_9P_FD) += 9pnet_fd.o obj-$(CONFIG_NET_9P_VIRTIO) += 9pnet_virtio.o 9pnet-objs := \ @@ -9,8 +8,6 @@ obj-$(CONFIG_NET_9P_VIRTIO) += 9pnet_virtio.o error.o \ fcprint.o \ util.o \ - -9pnet_fd-objs := \ trans_fd.o \ 9pnet_virtio-objs := \ diff --git a/net/9p/client.c b/net/9p/client.c index 84e087e2414..2ffe40cf2f0 100644 --- a/net/9p/client.c +++ b/net/9p/client.c @@ -64,21 +64,30 @@ static match_table_t tokens = { * @options: options string passed from mount * @v9ses: existing v9fs session information * + * Return 0 upon success, -ERRNO upon failure */ -static void parse_opts(char *options, struct p9_client *clnt) +static int parse_opts(char *opts, struct p9_client *clnt) { + char *options; char *p; substring_t args[MAX_OPT_ARGS]; int option; - int ret; + int ret = 0; clnt->trans_mod = v9fs_default_trans(); clnt->dotu = 1; clnt->msize = 8192; - if (!options) - return; + if (!opts) + return 0; + + options = kstrdup(opts, GFP_KERNEL); + if (!options) { + P9_DPRINTK(P9_DEBUG_ERROR, + "failed to allocate copy of option string\n"); + return -ENOMEM; + } while ((p = strsep(&options, ",")) != NULL) { int token; @@ -86,10 +95,11 @@ static void parse_opts(char *options, struct p9_client *clnt) continue; token = match_token(p, tokens, args); if (token < Opt_trans) { - ret = match_int(&args[0], &option); - if (ret < 0) { + int r = match_int(&args[0], &option); + if (r < 0) { P9_DPRINTK(P9_DEBUG_ERROR, "integer field, but no integer?\n"); + ret = r; continue; } } @@ -107,6 +117,8 @@ static void parse_opts(char *options, struct p9_client *clnt) continue; } } + kfree(options); + return ret; } @@ -138,16 +150,20 @@ struct p9_client *p9_client_create(const char *dev_name, char *options) if (!clnt) return ERR_PTR(-ENOMEM); + clnt->trans = NULL; spin_lock_init(&clnt->lock); INIT_LIST_HEAD(&clnt->fidlist); clnt->fidpool = p9_idpool_create(); - if (!clnt->fidpool) { + if (IS_ERR(clnt->fidpool)) { err = PTR_ERR(clnt->fidpool); clnt->fidpool = NULL; goto error; } - parse_opts(options, clnt); + err = parse_opts(options, clnt); + if (err < 0) + goto error; + if (clnt->trans_mod == NULL) { err = -EPROTONOSUPPORT; P9_DPRINTK(P9_DEBUG_ERROR, diff --git a/net/9p/conv.c b/net/9p/conv.c index 3fe35d532c8..44547201f5b 100644 --- a/net/9p/conv.c +++ b/net/9p/conv.c @@ -197,7 +197,7 @@ static void buf_get_qid(struct cbuf *bufp, struct p9_qid *qid) /** * p9_size_wstat - calculate the size of a variable length stat struct - * @stat: metadata (stat) structure + * @wstat: metadata (stat) structure * @dotu: non-zero if 9P2000.u * */ @@ -511,6 +511,12 @@ p9_create_common(struct cbuf *bufp, u32 size, u8 id) return fc; } +/** + * p9_set_tag - set the tag field of an &p9_fcall structure + * @fc: fcall structure to set tag within + * @tag: tag id to set + */ + void p9_set_tag(struct p9_fcall *fc, u16 tag) { fc->tag = tag; @@ -518,6 +524,12 @@ void p9_set_tag(struct p9_fcall *fc, u16 tag) } EXPORT_SYMBOL(p9_set_tag); +/** + * p9_create_tversion - allocates and creates a T_VERSION request + * @msize: requested maximum data size + * @version: version string to negotiate + * + */ struct p9_fcall *p9_create_tversion(u32 msize, char *version) { int size; @@ -542,6 +554,16 @@ error: } EXPORT_SYMBOL(p9_create_tversion); +/** + * p9_create_tauth - allocates and creates a T_AUTH request + * @afid: handle to use for authentication protocol + * @uname: user name attempting to authenticate + * @aname: mount specifier for remote server + * @n_uname: numeric id for user attempting to authneticate + * @dotu: 9P2000.u extension flag + * + */ + struct p9_fcall *p9_create_tauth(u32 afid, char *uname, char *aname, u32 n_uname, int dotu) { @@ -580,6 +602,18 @@ error: } EXPORT_SYMBOL(p9_create_tauth); +/** + * p9_create_tattach - allocates and creates a T_ATTACH request + * @fid: handle to use for the new mount point + * @afid: handle to use for authentication protocol + * @uname: user name attempting to attach + * @aname: mount specifier for remote server + * @n_uname: numeric id for user attempting to attach + * @n_uname: numeric id for user attempting to attach + * @dotu: 9P2000.u extension flag + * + */ + struct p9_fcall * p9_create_tattach(u32 fid, u32 afid, char *uname, char *aname, u32 n_uname, int dotu) @@ -616,6 +650,12 @@ error: } EXPORT_SYMBOL(p9_create_tattach); +/** + * p9_create_tflush - allocates and creates a T_FLUSH request + * @oldtag: tag id for the transaction we are attempting to cancel + * + */ + struct p9_fcall *p9_create_tflush(u16 oldtag) { int size; @@ -639,6 +679,15 @@ error: } EXPORT_SYMBOL(p9_create_tflush); +/** + * p9_create_twalk - allocates and creates a T_FLUSH request + * @fid: handle we are traversing from + * @newfid: a new handle for this transaction + * @nwname: number of path elements to traverse + * @wnames: array of path elements + * + */ + struct p9_fcall *p9_create_twalk(u32 fid, u32 newfid, u16 nwname, char **wnames) { @@ -677,6 +726,13 @@ error: } EXPORT_SYMBOL(p9_create_twalk); +/** + * p9_create_topen - allocates and creates a T_OPEN request + * @fid: handle we are trying to open + * @mode: what mode we are trying to open the file in + * + */ + struct p9_fcall *p9_create_topen(u32 fid, u8 mode) { int size; @@ -701,6 +757,19 @@ error: } EXPORT_SYMBOL(p9_create_topen); +/** + * p9_create_tcreate - allocates and creates a T_CREATE request + * @fid: handle of directory we are trying to create in + * @name: name of the file we are trying to create + * @perm: permissions for the file we are trying to create + * @mode: what mode we are trying to open the file in + * @extension: 9p2000.u extension string (for special files) + * @dotu: 9p2000.u enabled flag + * + * Note: Plan 9 create semantics include opening the resulting file + * which is why mode is included. + */ + struct p9_fcall *p9_create_tcreate(u32 fid, char *name, u32 perm, u8 mode, char *extension, int dotu) { @@ -736,6 +805,13 @@ error: } EXPORT_SYMBOL(p9_create_tcreate); +/** + * p9_create_tread - allocates and creates a T_READ request + * @fid: handle of the file we are trying to read + * @offset: offset to start reading from + * @count: how many bytes to read + */ + struct p9_fcall *p9_create_tread(u32 fid, u64 offset, u32 count) { int size; @@ -761,6 +837,17 @@ error: } EXPORT_SYMBOL(p9_create_tread); +/** + * p9_create_twrite - allocates and creates a T_WRITE request from the kernel + * @fid: handle of the file we are trying to write + * @offset: offset to start writing at + * @count: how many bytes to write + * @data: data to write + * + * This function will create a requst with data buffers from the kernel + * such as the page cache. + */ + struct p9_fcall *p9_create_twrite(u32 fid, u64 offset, u32 count, const char *data) { @@ -794,6 +881,16 @@ error: } EXPORT_SYMBOL(p9_create_twrite); +/** + * p9_create_twrite_u - allocates and creates a T_WRITE request from userspace + * @fid: handle of the file we are trying to write + * @offset: offset to start writing at + * @count: how many bytes to write + * @data: data to write + * + * This function will create a request with data buffers from userspace + */ + struct p9_fcall *p9_create_twrite_u(u32 fid, u64 offset, u32 count, const char __user *data) { @@ -827,6 +924,14 @@ error: } EXPORT_SYMBOL(p9_create_twrite_u); +/** + * p9_create_tclunk - allocate a request to forget about a file handle + * @fid: handle of the file we closing or forgetting about + * + * clunk is used both to close open files and to discard transient handles + * which may be created during meta-data operations and hierarchy traversal. + */ + struct p9_fcall *p9_create_tclunk(u32 fid) { int size; @@ -850,6 +955,12 @@ error: } EXPORT_SYMBOL(p9_create_tclunk); +/** + * p9_create_tremove - allocate and create a request to remove a file + * @fid: handle of the file or directory we are removing + * + */ + struct p9_fcall *p9_create_tremove(u32 fid) { int size; @@ -873,6 +984,12 @@ error: } EXPORT_SYMBOL(p9_create_tremove); +/** + * p9_create_tstat - allocate and populate a request for attributes + * @fid: handle of the file or directory we are trying to get the attributes of + * + */ + struct p9_fcall *p9_create_tstat(u32 fid) { int size; @@ -896,6 +1013,14 @@ error: } EXPORT_SYMBOL(p9_create_tstat); +/** + * p9_create_tstat - allocate and populate a request to change attributes + * @fid: handle of the file or directory we are trying to change + * @wstat: &p9_stat structure with attributes we wish to set + * @dotu: 9p2000.u enabled flag + * + */ + struct p9_fcall *p9_create_twstat(u32 fid, struct p9_wstat *wstat, int dotu) { @@ -922,3 +1047,4 @@ error: return fc; } EXPORT_SYMBOL(p9_create_twstat); + diff --git a/net/9p/error.c b/net/9p/error.c index 64104b9cb42..fdebe431406 100644 --- a/net/9p/error.c +++ b/net/9p/error.c @@ -33,6 +33,13 @@ #include <linux/errno.h> #include <net/9p/9p.h> +/** + * struct errormap - map string errors from Plan 9 to Linux numeric ids + * @name: string sent over 9P + * @val: numeric id most closely representing @name + * @namelen: length of string + * @list: hash-table list for string lookup + */ struct errormap { char *name; int val; @@ -177,8 +184,7 @@ static struct errormap errmap[] = { }; /** - * p9_error_init - preload - * @errstr: error string + * p9_error_init - preload mappings into hash list * */ @@ -206,6 +212,7 @@ EXPORT_SYMBOL(p9_error_init); /** * errstr2errno - convert error string to error number * @errstr: error string + * @len: length of error string * */ @@ -230,8 +237,8 @@ int p9_errstr2errno(char *errstr, int len) if (errno == 0) { /* TODO: if error isn't found, add it dynamically */ errstr[len] = 0; - printk(KERN_ERR "%s: errstr :%s: not found\n", __func__, - errstr); + printk(KERN_ERR "%s: server reported unknown error %s\n", + __func__, errstr); errno = 1; } diff --git a/net/9p/fcprint.c b/net/9p/fcprint.c index 40244fbd9b0..53dd8e28dd8 100644 --- a/net/9p/fcprint.c +++ b/net/9p/fcprint.c @@ -142,6 +142,14 @@ p9_printdata(char *buf, int buflen, u8 *data, int datalen) return p9_dumpdata(buf, buflen, data, datalen < 16?datalen:16); } +/** + * p9_printfcall - decode and print a protocol structure into a buffer + * @buf: buffer to deposit decoded structure into + * @buflen: available space in buffer + * @fc: protocol rpc structure of type &p9_fcall + * @extended: whether or not session is operating with extended protocol + */ + int p9_printfcall(char *buf, int buflen, struct p9_fcall *fc, int extended) { diff --git a/net/9p/mod.c b/net/9p/mod.c index c285aab2af0..bdee1fb7cc6 100644 --- a/net/9p/mod.c +++ b/net/9p/mod.c @@ -39,9 +39,6 @@ module_param_named(debug, p9_debug_level, uint, 0); MODULE_PARM_DESC(debug, "9P debugging level"); #endif -extern int p9_mux_global_init(void); -extern void p9_mux_global_exit(void); - /* * Dynamic Transport Registration Routines * @@ -52,7 +49,7 @@ static struct p9_trans_module *v9fs_default_transport; /** * v9fs_register_trans - register a new transport with 9p - * @m - structure describing the transport module and entry points + * @m: structure describing the transport module and entry points * */ void v9fs_register_trans(struct p9_trans_module *m) @@ -65,7 +62,7 @@ EXPORT_SYMBOL(v9fs_register_trans); /** * v9fs_match_trans - match transport versus registered transports - * @arg: string identifying transport + * @name: string identifying transport * */ struct p9_trans_module *v9fs_match_trans(const substring_t *name) @@ -110,6 +107,7 @@ static int __init init_p9(void) p9_error_init(); printk(KERN_INFO "Installing 9P2000 support\n"); + p9_trans_fd_init(); return ret; } diff --git a/net/9p/trans_fd.c b/net/9p/trans_fd.c index f624dff7685..4507f744f44 100644 --- a/net/9p/trans_fd.c +++ b/net/9p/trans_fd.c @@ -47,12 +47,29 @@ #define SCHED_TIMEOUT 10 #define MAXPOLLWADDR 2 +/** + * struct p9_fd_opts - per-transport options + * @rfd: file descriptor for reading (trans=fd) + * @wfd: file descriptor for writing (trans=fd) + * @port: port to connect to (trans=tcp) + * + */ + struct p9_fd_opts { int rfd; int wfd; u16 port; }; + +/** + * struct p9_trans_fd - transport state + * @rd: reference to file to read from + * @wr: reference of file to write to + * @conn: connection state reference + * + */ + struct p9_trans_fd { struct file *rd; struct file *wr; @@ -90,10 +107,24 @@ enum { }; struct p9_req; - typedef void (*p9_conn_req_callback)(struct p9_req *req, void *a); + +/** + * struct p9_req - fd mux encoding of an rpc transaction + * @lock: protects req_list + * @tag: numeric tag for rpc transaction + * @tcall: request &p9_fcall structure + * @rcall: response &p9_fcall structure + * @err: error state + * @cb: callback for when response is received + * @cba: argument to pass to callback + * @flush: flag to indicate RPC has been flushed + * @req_list: list link for higher level objects to chain requests + * + */ + struct p9_req { - spinlock_t lock; /* protect request structure */ + spinlock_t lock; int tag; struct p9_fcall *tcall; struct p9_fcall *rcall; @@ -104,7 +135,39 @@ struct p9_req { struct list_head req_list; }; -struct p9_mux_poll_task; +struct p9_mux_poll_task { + struct task_struct *task; + struct list_head mux_list; + int muxnum; +}; + +/** + * struct p9_conn - fd mux connection state information + * @lock: protects mux_list (?) + * @mux_list: list link for mux to manage multiple connections (?) + * @poll_task: task polling on this connection + * @msize: maximum size for connection (dup) + * @extended: 9p2000.u flag (dup) + * @trans: reference to transport instance for this connection + * @tagpool: id accounting for transactions + * @err: error state + * @equeue: event wait_q (?) + * @req_list: accounting for requests which have been sent + * @unsent_req_list: accounting for requests that haven't been sent + * @rcall: current response &p9_fcall structure + * @rpos: read position in current frame + * @rbuf: current read buffer + * @wpos: write position for current frame + * @wsize: amount of data to write for current frame + * @wbuf: current write buffer + * @poll_wait: array of wait_q's for various worker threads + * @poll_waddr: ???? + * @pt: poll state + * @rq: current read work + * @wq: current write work + * @wsched: ???? + * + */ struct p9_conn { spinlock_t lock; /* protect lock structure */ @@ -132,11 +195,16 @@ struct p9_conn { unsigned long wsched; }; -struct p9_mux_poll_task { - struct task_struct *task; - struct list_head mux_list; - int muxnum; -}; +/** + * struct p9_mux_rpc - fd mux rpc accounting structure + * @m: connection this request was issued on + * @err: error state + * @tcall: request &p9_fcall + * @rcall: response &p9_fcall + * @wqueue: wait queue that client is blocked on for this rpc + * + * Bug: isn't this information duplicated elsewhere like &p9_req + */ struct p9_mux_rpc { struct p9_conn *m; @@ -207,10 +275,12 @@ static void p9_mux_put_tag(struct p9_conn *m, u16 tag) /** * p9_mux_calc_poll_procs - calculates the number of polling procs - * based on the number of mounted v9fs filesystems. + * @muxnum: number of mounts * + * Calculation is based on the number of mounted v9fs filesystems. * The current implementation returns sqrt of the number of mounts. */ + static int p9_mux_calc_poll_procs(int muxnum) { int n; @@ -331,12 +401,11 @@ static void p9_mux_poll_stop(struct p9_conn *m) /** * p9_conn_create - allocate and initialize the per-session mux data - * Creates the polling task if this is the first session. + * @trans: transport structure * - * @trans - transport structure - * @msize - maximum message size - * @extended - extended flag + * Note: Creates the polling task if this is the first session. */ + static struct p9_conn *p9_conn_create(struct p9_trans *trans) { int i, n; @@ -406,7 +475,10 @@ static struct p9_conn *p9_conn_create(struct p9_trans *trans) /** * p9_mux_destroy - cancels all pending requests and frees mux resources + * @m: mux to destroy + * */ + static void p9_conn_destroy(struct p9_conn *m) { P9_DPRINTK(P9_DEBUG_MUX, "mux %p prev %p next %p\n", m, @@ -429,9 +501,14 @@ static void p9_conn_destroy(struct p9_conn *m) } /** - * p9_pollwait - called by files poll operation to add v9fs-poll task - * to files wait queue + * p9_pollwait - add poll task to the wait queue + * @filp: file pointer being polled + * @wait_address: wait_q to block on + * @p: poll state + * + * called by files poll operation to add v9fs-poll task to files wait queue */ + static void p9_pollwait(struct file *filp, wait_queue_head_t *wait_address, poll_table *p) { @@ -462,7 +539,10 @@ p9_pollwait(struct file *filp, wait_queue_head_t *wait_address, poll_table *p) /** * p9_poll_mux - polls a mux and schedules read or write works if necessary + * @m: connection to poll + * */ + static void p9_poll_mux(struct p9_conn *m) { int n; @@ -499,9 +579,14 @@ static void p9_poll_mux(struct p9_conn *m) } /** - * p9_poll_proc - polls all v9fs transports for new events and queues - * the appropriate work to the work queue + * p9_poll_proc - poll worker thread + * @a: thread state and arguments + * + * polls all v9fs transports for new events and queues the appropriate + * work to the work queue + * */ + static int p9_poll_proc(void *a) { struct p9_conn *m, *mtmp; @@ -527,7 +612,10 @@ static int p9_poll_proc(void *a) /** * p9_write_work - called when a transport can send some data + * @work: container for work to be done + * */ + static void p9_write_work(struct work_struct *work) { int n, err; @@ -638,7 +726,10 @@ static void process_request(struct p9_conn *m, struct p9_req *req) /** * p9_read_work - called when there is some data to be read from a transport + * @work: container of work to be done + * */ + static void p9_read_work(struct work_struct *work) { int n, err; @@ -793,7 +884,9 @@ error: * @tc: request to be sent * @cb: callback function to call when response is received * @cba: parameter to pass to the callback function + * */ + static struct p9_req *p9_send_request(struct p9_conn *m, struct p9_fcall *tc, p9_conn_req_callback cb, void *cba) @@ -961,10 +1054,12 @@ p9_conn_rpc_cb(struct p9_req *req, void *a) /** * p9_fd_rpc- sends 9P request and waits until a response is available. * The function can be interrupted. - * @m: mux data + * @t: transport data * @tc: request to be sent * @rc: pointer where a pointer to the response is stored + * */ + int p9_fd_rpc(struct p9_trans *t, struct p9_fcall *tc, struct p9_fcall **rc) { @@ -1041,8 +1136,10 @@ p9_fd_rpc(struct p9_trans *t, struct p9_fcall *tc, struct p9_fcall **rc) * @m: mux data * @tc: request to be sent * @cb: callback function to be called when response arrives - * @cba: value to pass to the callback function + * @a: value to pass to the callback function + * */ + int p9_conn_rpcnb(struct p9_conn *m, struct p9_fcall *tc, p9_conn_req_callback cb, void *a) { @@ -1065,7 +1162,9 @@ int p9_conn_rpcnb(struct p9_conn *m, struct p9_fcall *tc, * p9_conn_cancel - cancel all pending requests with error * @m: mux data * @err: error code + * */ + void p9_conn_cancel(struct p9_conn *m, int err) { struct p9_req *req, *rtmp; @@ -1097,35 +1196,46 @@ void p9_conn_cancel(struct p9_conn *m, int err) } /** - * v9fs_parse_options - parse mount options into session structure + * parse_options - parse mount options into session structure * @options: options string passed from mount - * @v9ses: existing v9fs session information + * @opts: transport-specific structure to parse options into * + * Returns 0 upon success, -ERRNO upon failure */ -static void parse_opts(char *options, struct p9_fd_opts *opts) +static int parse_opts(char *params, struct p9_fd_opts *opts) { char *p; substring_t args[MAX_OPT_ARGS]; int option; + char *options; int ret; opts->port = P9_PORT; opts->rfd = ~0; opts->wfd = ~0; - if (!options) - return; + if (!params) + return 0; + + options = kstrdup(params, GFP_KERNEL); + if (!options) { + P9_DPRINTK(P9_DEBUG_ERROR, + "failed to allocate copy of option string\n"); + return -ENOMEM; + } while ((p = strsep(&options, ",")) != NULL) { int token; + int r; if (!*p) continue; token = match_token(p, tokens, args); - ret = match_int(&args[0], &option); - if (ret < 0) { + r = match_int(&args[0], &option); + if (r < 0) { P9_DPRINTK(P9_DEBUG_ERROR, "integer field, but no integer?\n"); + ret = r; continue; } switch (token) { @@ -1142,6 +1252,8 @@ static void parse_opts(char *options, struct p9_fd_opts *opts) continue; } } + kfree(options); + return 0; } static int p9_fd_open(struct p9_trans *trans, int rfd, int wfd) @@ -1193,11 +1305,12 @@ static int p9_socket_open(struct p9_trans *trans, struct socket *csocket) /** * p9_fd_read- read from a fd - * @v9ses: session information + * @trans: transport instance state * @v: buffer to receive data into * @len: size of receive buffer * */ + static int p9_fd_read(struct p9_trans *trans, void *v, int len) { int ret; @@ -1220,11 +1333,12 @@ static int p9_fd_read(struct p9_trans *trans, void *v, int len) /** * p9_fd_write - write to a socket - * @v9ses: session information + * @trans: transport instance state * @v: buffer to send data from * @len: size of send buffer * */ + static int p9_fd_write(struct p9_trans *trans, void *v, int len) { int ret; @@ -1296,6 +1410,7 @@ end: * @trans: private socket structure * */ + static void p9_fd_close(struct p9_trans *trans) { struct p9_trans_fd *ts; @@ -1318,6 +1433,23 @@ static void p9_fd_close(struct p9_trans *trans) kfree(ts); } +/* + * stolen from NFS - maybe should be made a generic function? + */ +static inline int valid_ipaddr4(const char *buf) +{ + int rc, count, in[4]; + + rc = sscanf(buf, "%d.%d.%d.%d", &in[0], &in[1], &in[2], &in[3]); + if (rc != 4) + return -EINVAL; + for (count = 0; count < 4; count++) { + if (in[count] > 255) + return -EINVAL; + } + return 0; +} + static struct p9_trans * p9_trans_create_tcp(const char *addr, char *args, int msize, unsigned char dotu) { @@ -1328,7 +1460,12 @@ p9_trans_create_tcp(const char *addr, char *args, int msize, unsigned char dotu) struct p9_fd_opts opts; struct p9_trans_fd *p; - parse_opts(args, &opts); + err = parse_opts(args, &opts); + if (err < 0) + return ERR_PTR(err); + + if (valid_ipaddr4(addr) < 0) + return ERR_PTR(-EINVAL); csocket = NULL; trans = kmalloc(sizeof(struct p9_trans), GFP_KERNEL); @@ -1508,7 +1645,7 @@ static struct p9_trans_module p9_fd_trans = { .create = p9_trans_create_fd, }; -static int __init p9_trans_fd_init(void) +int p9_trans_fd_init(void) { int ret = p9_mux_global_init(); if (ret) { @@ -1522,9 +1659,4 @@ static int __init p9_trans_fd_init(void) return 0; } - -module_init(p9_trans_fd_init); - -MODULE_AUTHOR("Latchesar Ionkov <lucho@ionkov.net>"); -MODULE_AUTHOR("Eric Van Hensbergen <ericvh@gmail.com>"); -MODULE_LICENSE("GPL"); +EXPORT_SYMBOL(p9_trans_fd_init); diff --git a/net/9p/trans_virtio.c b/net/9p/trans_virtio.c index de7a9f532ed..42adc052b14 100644 --- a/net/9p/trans_virtio.c +++ b/net/9p/trans_virtio.c @@ -49,29 +49,75 @@ #define VIRTQUEUE_NUM 128 /* a single mutex to manage channel initialization and attachment */ -static DECLARE_MUTEX(virtio_9p_lock); +static DEFINE_MUTEX(virtio_9p_lock); /* global which tracks highest initialized channel */ static int chan_index; #define P9_INIT_MAXTAG 16 -#define REQ_STATUS_IDLE 0 -#define REQ_STATUS_SENT 1 -#define REQ_STATUS_RCVD 2 -#define REQ_STATUS_FLSH 3 + +/** + * enum p9_req_status_t - virtio request status + * @REQ_STATUS_IDLE: request slot unused + * @REQ_STATUS_SENT: request sent to server + * @REQ_STATUS_RCVD: response received from server + * @REQ_STATUS_FLSH: request has been flushed + * + * The @REQ_STATUS_IDLE state is used to mark a request slot as unused + * but use is actually tracked by the idpool structure which handles tag + * id allocation. + * + */ + +enum p9_req_status_t { + REQ_STATUS_IDLE, + REQ_STATUS_SENT, + REQ_STATUS_RCVD, + REQ_STATUS_FLSH, +}; + +/** + * struct p9_req_t - virtio request slots + * @status: status of this request slot + * @wq: wait_queue for the client to block on for this request + * + * The virtio transport uses an array to track outstanding requests + * instead of a list. While this may incurr overhead during initial + * allocation or expansion, it makes request lookup much easier as the + * tag id is a index into an array. (We use tag+1 so that we can accomodate + * the -1 tag for the T_VERSION request). + * This also has the nice effect of only having to allocate wait_queues + * once, instead of constantly allocating and freeing them. Its possible + * other resources could benefit from this scheme as well. + * + */ struct p9_req_t { int status; wait_queue_head_t *wq; }; -/* We keep all per-channel information in a structure. +/** + * struct virtio_chan - per-instance transport information + * @initialized: whether the channel is initialized + * @inuse: whether the channel is in use + * @lock: protects multiple elements within this structure + * @vdev: virtio dev associated with this channel + * @vq: virtio queue associated with this channel + * @tagpool: accounting for tag ids (and request slots) + * @reqs: array of request slots + * @max_tag: current number of request_slots allocated + * @sg: scatter gather list which is used to pack a request (protected?) + * + * We keep all per-channel information in a structure. * This structure is allocated within the devices dev->mem space. * A pointer to the structure will get put in the transport private. + * */ + static struct virtio_chan { - bool initialized; /* channel is initialized */ - bool inuse; /* channel is in use */ + bool initialized; + bool inuse; spinlock_t lock; @@ -86,7 +132,19 @@ static struct virtio_chan { struct scatterlist sg[VIRTQUEUE_NUM]; } channels[MAX_9P_CHAN]; -/* Lookup requests by tag */ +/** + * p9_lookup_tag - Lookup requests by tag + * @c: virtio channel to lookup tag within + * @tag: numeric id for transaction + * + * this is a simple array lookup, but will grow the + * request_slots as necessary to accomodate transaction + * ids which did not previously have a slot. + * + * Bugs: there is currently no upper limit on request slots set + * here, but that should be constrained by the id accounting. + */ + static struct p9_req_t *p9_lookup_tag(struct virtio_chan *c, u16 tag) { /* This looks up the original request by tag so we know which @@ -130,11 +188,20 @@ static unsigned int rest_of_page(void *data) return PAGE_SIZE - ((unsigned long)data % PAGE_SIZE); } +/** + * p9_virtio_close - reclaim resources of a channel + * @trans: transport state + * + * This reclaims a channel by freeing its resources and + * reseting its inuse flag. + * + */ + static void p9_virtio_close(struct p9_trans *trans) { struct virtio_chan *chan = trans->priv; int count; - unsigned int flags; + unsigned long flags; spin_lock_irqsave(&chan->lock, flags); p9_idpool_destroy(chan->tagpool); @@ -144,13 +211,26 @@ static void p9_virtio_close(struct p9_trans *trans) chan->max_tag = 0; spin_unlock_irqrestore(&chan->lock, flags); - down(&virtio_9p_lock); + mutex_lock(&virtio_9p_lock); chan->inuse = false; - up(&virtio_9p_lock); + mutex_unlock(&virtio_9p_lock); kfree(trans); } +/** + * req_done - callback which signals activity from the server + * @vq: virtio queue activity was received on + * + * This notifies us that the server has triggered some activity + * on the virtio channel - most likely a response to request we + * sent. Figure out which requests now have responses and wake up + * those threads. + * + * Bugs: could do with some additional sanity checking, but appears to work. + * + */ + static void req_done(struct virtqueue *vq) { struct virtio_chan *chan = vq->vdev->priv; @@ -169,6 +249,20 @@ static void req_done(struct virtqueue *vq) spin_unlock_irqrestore(&chan->lock, flags); } +/** + * pack_sg_list - pack a scatter gather list from a linear buffer + * @sg: scatter/gather list to pack into + * @start: which segment of the sg_list to start at + * @limit: maximum segment to pack data to + * @data: data to pack into scatter/gather list + * @count: amount of data to pack into the scatter/gather list + * + * sg_lists have multiple segments of various sizes. This will pack + * arbitrary data into an existing scatter gather list, segmenting the + * data as necessary within constraints. + * + */ + static int pack_sg_list(struct scatterlist *sg, int start, int limit, char *data, int count) @@ -189,6 +283,14 @@ pack_sg_list(struct scatterlist *sg, int start, int limit, char *data, return index-start; } +/** + * p9_virtio_rpc - issue a request and wait for a response + * @t: transport state + * @tc: &p9_fcall request to transmit + * @rc: &p9_fcall to put reponse into + * + */ + static int p9_virtio_rpc(struct p9_trans *t, struct p9_fcall *tc, struct p9_fcall **rc) { @@ -263,16 +365,26 @@ p9_virtio_rpc(struct p9_trans *t, struct p9_fcall *tc, struct p9_fcall **rc) return 0; } +/** + * p9_virtio_probe - probe for existence of 9P virtio channels + * @vdev: virtio device to probe + * + * This probes for existing virtio channels. At present only + * a single channel is in use, so in the future more work may need + * to be done here. + * + */ + static int p9_virtio_probe(struct virtio_device *vdev) { int err; struct virtio_chan *chan; int index; - down(&virtio_9p_lock); + mutex_lock(&virtio_9p_lock); index = chan_index++; chan = &channels[index]; - up(&virtio_9p_lock); + mutex_unlock(&virtio_9p_lock); if (chan_index > MAX_9P_CHAN) { printk(KERN_ERR "9p: virtio: Maximum channels exceeded\n"); @@ -301,17 +413,34 @@ static int p9_virtio_probe(struct virtio_device *vdev) out_free_vq: vdev->config->del_vq(chan->vq); fail: - down(&virtio_9p_lock); + mutex_lock(&virtio_9p_lock); chan_index--; - up(&virtio_9p_lock); + mutex_unlock(&virtio_9p_lock); return err; } -/* This sets up a transport channel for 9p communication. Right now + +/** + * p9_virtio_create - allocate a new virtio channel + * @devname: string identifying the channel to connect to (unused) + * @args: args passed from sys_mount() for per-transport options (unused) + * @msize: requested maximum packet size + * @extended: 9p2000.u enabled flag + * + * This sets up a transport channel for 9p communication. Right now * we only match the first available channel, but eventually we couldlook up * alternate channels by matching devname versus a virtio_config entry. * We use a simple reference count mechanism to ensure that only a single - * mount has a channel open at a time. */ + * mount has a channel open at a time. + * + * Bugs: doesn't allow identification of a specific channel + * to allocate, channels are allocated sequentially. This was + * a pragmatic decision to get things rolling, but ideally some + * way of identifying the channel to attach to would be nice + * if we are going to support multiple channels. + * + */ + static struct p9_trans * p9_virtio_create(const char *devname, char *args, int msize, unsigned char extended) @@ -320,7 +449,7 @@ p9_virtio_create(const char *devname, char *args, int msize, struct virtio_chan *chan = channels; int index = 0; - down(&virtio_9p_lock); + mutex_lock(&virtio_9p_lock); while (index < MAX_9P_CHAN) { if (chan->initialized && !chan->inuse) { chan->inuse = true; @@ -330,7 +459,7 @@ p9_virtio_create(const char *devname, char *args, int msize, chan = &channels[index]; } } - up(&virtio_9p_lock); + mutex_unlock(&virtio_9p_lock); if (index >= MAX_9P_CHAN) { printk(KERN_ERR "9p: no channels available\n"); @@ -360,6 +489,12 @@ p9_virtio_create(const char *devname, char *args, int msize, return trans; } +/** + * p9_virtio_remove - clean up resources associated with a virtio device + * @vdev: virtio device to remove + * + */ + static void p9_virtio_remove(struct virtio_device *vdev) { struct virtio_chan *chan = vdev->priv; diff --git a/net/9p/util.c b/net/9p/util.c index ef7215565d8..958fc58cd1f 100644 --- a/net/9p/util.c +++ b/net/9p/util.c @@ -32,11 +32,23 @@ #include <linux/idr.h> #include <net/9p/9p.h> +/** + * struct p9_idpool - per-connection accounting for tag idpool + * @lock: protects the pool + * @pool: idr to allocate tag id from + * + */ + struct p9_idpool { spinlock_t lock; struct idr pool; }; +/** + * p9_idpool_create - create a new per-connection id pool + * + */ + struct p9_idpool *p9_idpool_create(void) { struct p9_idpool *p; @@ -52,6 +64,11 @@ struct p9_idpool *p9_idpool_create(void) } EXPORT_SYMBOL(p9_idpool_create); +/** + * p9_idpool_destroy - create a new per-connection id pool + * @p: idpool to destory + */ + void p9_idpool_destroy(struct p9_idpool *p) { idr_destroy(&p->pool); @@ -61,9 +78,9 @@ EXPORT_SYMBOL(p9_idpool_destroy); /** * p9_idpool_get - allocate numeric id from pool - * @p - pool to allocate from + * @p: pool to allocate from * - * XXX - This seems to be an awful generic function, should it be in idr.c with + * Bugs: This seems to be an awful generic function, should it be in idr.c with * the lock included in struct idr? */ @@ -71,7 +88,7 @@ int p9_idpool_get(struct p9_idpool *p) { int i = 0; int error; - unsigned int flags; + unsigned long flags; retry: if (idr_pre_get(&p->pool, GFP_KERNEL) == 0) @@ -94,15 +111,16 @@ EXPORT_SYMBOL(p9_idpool_get); /** * p9_idpool_put - release numeric id from pool - * @p - pool to allocate from + * @id: numeric id which is being released + * @p: pool to release id into * - * XXX - This seems to be an awful generic function, should it be in idr.c with + * Bugs: This seems to be an awful generic function, should it be in idr.c with * the lock included in struct idr? */ void p9_idpool_put(int id, struct p9_idpool *p) { - unsigned int flags; + unsigned long flags; spin_lock_irqsave(&p->lock, flags); idr_remove(&p->pool, id); spin_unlock_irqrestore(&p->lock, flags); @@ -111,11 +129,13 @@ EXPORT_SYMBOL(p9_idpool_put); /** * p9_idpool_check - check if the specified id is available - * @id - id to check - * @p - pool + * @id: id to check + * @p: pool to check */ + int p9_idpool_check(int id, struct p9_idpool *p) { return idr_find(&p->pool, id) != NULL; } EXPORT_SYMBOL(p9_idpool_check); + diff --git a/net/core/dev.c b/net/core/dev.c index a1607bc0cd4..58296307787 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -903,7 +903,11 @@ int dev_change_name(struct net_device *dev, char *newname) strlcpy(dev->name, newname, IFNAMSIZ); rollback: - device_rename(&dev->dev, dev->name); + err = device_rename(&dev->dev, dev->name); + if (err) { + memcpy(dev->name, oldname, IFNAMSIZ); + return err; + } write_lock_bh(&dev_base_lock); hlist_del(&dev->name_hlist); @@ -3137,7 +3141,7 @@ int dev_change_flags(struct net_device *dev, unsigned flags) * Load in the correct multicast list now the flags have changed. */ - if (dev->change_rx_flags && (dev->flags ^ flags) & IFF_MULTICAST) + if (dev->change_rx_flags && (old_flags ^ flags) & IFF_MULTICAST) dev->change_rx_flags(dev, IFF_MULTICAST); dev_set_rx_mode(dev); diff --git a/net/core/netpoll.c b/net/core/netpoll.c index b04d643fc3c..8fb134da034 100644 --- a/net/core/netpoll.c +++ b/net/core/netpoll.c @@ -419,7 +419,7 @@ static void arp_reply(struct sk_buff *skb) return; size = arp_hdr_len(skb->dev); - send_skb = find_skb(np, size + LL_RESERVED_SPACE(np->dev), + send_skb = find_skb(np, size + LL_ALLOCATED_SPACE(np->dev), LL_RESERVED_SPACE(np->dev)); if (!send_skb) diff --git a/net/core/pktgen.c b/net/core/pktgen.c index 8dca2111049..fdf537707e5 100644 --- a/net/core/pktgen.c +++ b/net/core/pktgen.c @@ -390,6 +390,7 @@ struct pktgen_thread { int cpu; wait_queue_head_t queue; + struct completion start_done; }; #define REMOVE 1 @@ -3414,6 +3415,7 @@ static int pktgen_thread_worker(void *arg) BUG_ON(smp_processor_id() != cpu); init_waitqueue_head(&t->queue); + complete(&t->start_done); pr_debug("pktgen: starting pktgen/%d: pid=%d\n", cpu, task_pid_nr(current)); @@ -3615,6 +3617,7 @@ static int __init pktgen_create_thread(int cpu) INIT_LIST_HEAD(&t->if_list); list_add_tail(&t->th_list, &pktgen_threads); + init_completion(&t->start_done); p = kthread_create(pktgen_thread_worker, t, "kpktgend_%d", cpu); if (IS_ERR(p)) { @@ -3639,6 +3642,7 @@ static int __init pktgen_create_thread(int cpu) } wake_up_process(p); + wait_for_completion(&t->start_done); return 0; } diff --git a/net/core/sock.c b/net/core/sock.c index fa76f04fa9c..88094cb09c0 100644 --- a/net/core/sock.c +++ b/net/core/sock.c @@ -270,7 +270,7 @@ int sock_queue_rcv_skb(struct sock *sk, struct sk_buff *skb) int err = 0; int skb_len; - /* Cast skb->rcvbuf to unsigned... It's pointless, but reduces + /* Cast sk->rcvbuf to unsigned... It's pointless, but reduces number of warnings when compiling with -W --ANK */ if (atomic_read(&sk->sk_rmem_alloc) + skb->truesize >= diff --git a/net/econet/af_econet.c b/net/econet/af_econet.c index 68d15448004..7c9bb13b153 100644 --- a/net/econet/af_econet.c +++ b/net/econet/af_econet.c @@ -340,7 +340,7 @@ static int econet_sendmsg(struct kiocb *iocb, struct socket *sock, dev_hold(dev); - skb = sock_alloc_send_skb(sk, len+LL_RESERVED_SPACE(dev), + skb = sock_alloc_send_skb(sk, len+LL_ALLOCATED_SPACE(dev), msg->msg_flags & MSG_DONTWAIT, &err); if (skb==NULL) goto out_unlock; diff --git a/net/ipv4/arp.c b/net/ipv4/arp.c index 68b72a7a180..418862f1bf2 100644 --- a/net/ipv4/arp.c +++ b/net/ipv4/arp.c @@ -570,7 +570,7 @@ struct sk_buff *arp_create(int type, int ptype, __be32 dest_ip, * Allocate a buffer */ - skb = alloc_skb(arp_hdr_len(dev) + LL_RESERVED_SPACE(dev), GFP_ATOMIC); + skb = alloc_skb(arp_hdr_len(dev) + LL_ALLOCATED_SPACE(dev), GFP_ATOMIC); if (skb == NULL) return NULL; diff --git a/net/ipv4/cipso_ipv4.c b/net/ipv4/cipso_ipv4.c index 05afb576d93..2c0e4572cc9 100644 --- a/net/ipv4/cipso_ipv4.c +++ b/net/ipv4/cipso_ipv4.c @@ -338,7 +338,7 @@ static int cipso_v4_cache_check(const unsigned char *key, return -ENOENT; hash = cipso_v4_map_cache_hash(key, key_len); - bkt = hash & (CIPSO_V4_CACHE_BUCKETBITS - 1); + bkt = hash & (CIPSO_V4_CACHE_BUCKETS - 1); spin_lock_bh(&cipso_v4_cache[bkt].lock); list_for_each_entry(entry, &cipso_v4_cache[bkt].list, list) { if (entry->hash == hash && @@ -417,7 +417,7 @@ int cipso_v4_cache_add(const struct sk_buff *skb, atomic_inc(&secattr->cache->refcount); entry->lsm_data = secattr->cache; - bkt = entry->hash & (CIPSO_V4_CACHE_BUCKETBITS - 1); + bkt = entry->hash & (CIPSO_V4_CACHE_BUCKETS - 1); spin_lock_bh(&cipso_v4_cache[bkt].lock); if (cipso_v4_cache[bkt].size < cipso_v4_cache_bucketsize) { list_add(&entry->list, &cipso_v4_cache[bkt].list); diff --git a/net/ipv4/igmp.c b/net/ipv4/igmp.c index 6250f4239b6..2769dc4a4c8 100644 --- a/net/ipv4/igmp.c +++ b/net/ipv4/igmp.c @@ -292,7 +292,7 @@ static struct sk_buff *igmpv3_newpack(struct net_device *dev, int size) struct iphdr *pip; struct igmpv3_report *pig; - skb = alloc_skb(size + LL_RESERVED_SPACE(dev), GFP_ATOMIC); + skb = alloc_skb(size + LL_ALLOCATED_SPACE(dev), GFP_ATOMIC); if (skb == NULL) return NULL; @@ -653,7 +653,7 @@ static int igmp_send_report(struct in_device *in_dev, struct ip_mc_list *pmc, return -1; } - skb=alloc_skb(IGMP_SIZE+LL_RESERVED_SPACE(dev), GFP_ATOMIC); + skb=alloc_skb(IGMP_SIZE+LL_ALLOCATED_SPACE(dev), GFP_ATOMIC); if (skb == NULL) { ip_rt_put(rt); return -1; diff --git a/net/ipv4/ipconfig.c b/net/ipv4/ipconfig.c index 89dee4346f6..ed45037ce9b 100644 --- a/net/ipv4/ipconfig.c +++ b/net/ipv4/ipconfig.c @@ -710,14 +710,14 @@ static void __init ic_bootp_send_if(struct ic_device *d, unsigned long jiffies_d struct net_device *dev = d->dev; struct sk_buff *skb; struct bootp_pkt *b; - int hh_len = LL_RESERVED_SPACE(dev); struct iphdr *h; /* Allocate packet */ - skb = alloc_skb(sizeof(struct bootp_pkt) + hh_len + 15, GFP_KERNEL); + skb = alloc_skb(sizeof(struct bootp_pkt) + LL_ALLOCATED_SPACE(dev) + 15, + GFP_KERNEL); if (!skb) return; - skb_reserve(skb, hh_len); + skb_reserve(skb, LL_RESERVED_SPACE(dev)); b = (struct bootp_pkt *) skb_put(skb, sizeof(struct bootp_pkt)); memset(b, 0, sizeof(struct bootp_pkt)); diff --git a/net/ipv4/raw.c b/net/ipv4/raw.c index 11d7f753a82..fead049daf4 100644 --- a/net/ipv4/raw.c +++ b/net/ipv4/raw.c @@ -322,7 +322,6 @@ static int raw_send_hdrinc(struct sock *sk, void *from, size_t length, unsigned int flags) { struct inet_sock *inet = inet_sk(sk); - int hh_len; struct iphdr *iph; struct sk_buff *skb; unsigned int iphlen; @@ -336,13 +335,12 @@ static int raw_send_hdrinc(struct sock *sk, void *from, size_t length, if (flags&MSG_PROBE) goto out; - hh_len = LL_RESERVED_SPACE(rt->u.dst.dev); - - skb = sock_alloc_send_skb(sk, length+hh_len+15, - flags&MSG_DONTWAIT, &err); + skb = sock_alloc_send_skb(sk, + length + LL_ALLOCATED_SPACE(rt->u.dst.dev) + 15, + flags & MSG_DONTWAIT, &err); if (skb == NULL) goto error; - skb_reserve(skb, hh_len); + skb_reserve(skb, LL_RESERVED_SPACE(rt->u.dst.dev)); skb->priority = sk->sk_priority; skb->mark = sk->sk_mark; diff --git a/net/ipv4/route.c b/net/ipv4/route.c index 92f90ae46f4..df41026b60d 100644 --- a/net/ipv4/route.c +++ b/net/ipv4/route.c @@ -160,7 +160,7 @@ static struct dst_ops ipv4_dst_ops = { .negative_advice = ipv4_negative_advice, .link_failure = ipv4_link_failure, .update_pmtu = ip_rt_update_pmtu, - .local_out = ip_local_out, + .local_out = __ip_local_out, .entry_size = sizeof(struct rtable), .entries = ATOMIC_INIT(0), }; diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index 26c936930e9..b54d9d37b63 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -1842,9 +1842,16 @@ static void tcp_enter_frto_loss(struct sock *sk, int allowed_segments, int flag) TCP_SKB_CB(skb)->sacked &= ~TCPCB_SACKED_RETRANS; } - /* Don't lost mark skbs that were fwd transmitted after RTO */ - if (!(TCP_SKB_CB(skb)->sacked & TCPCB_SACKED_ACKED) && - !after(TCP_SKB_CB(skb)->end_seq, tp->frto_highmark)) { + /* Marking forward transmissions that were made after RTO lost + * can cause unnecessary retransmissions in some scenarios, + * SACK blocks will mitigate that in some but not in all cases. + * We used to not mark them but it was causing break-ups with + * receivers that do only in-order receival. + * + * TODO: we could detect presence of such receiver and select + * different behavior per flow. + */ + if (!(TCP_SKB_CB(skb)->sacked & TCPCB_SACKED_ACKED)) { TCP_SKB_CB(skb)->sacked |= TCPCB_LOST; tp->lost_out += tcp_skb_pcount(skb); } @@ -1860,7 +1867,7 @@ static void tcp_enter_frto_loss(struct sock *sk, int allowed_segments, int flag) tp->reordering = min_t(unsigned int, tp->reordering, sysctl_tcp_reordering); tcp_set_ca_state(sk, TCP_CA_Loss); - tp->high_seq = tp->frto_highmark; + tp->high_seq = tp->snd_nxt; TCP_ECN_queue_cwr(tp); tcp_clear_retrans_hints_partial(tp); @@ -2482,7 +2489,7 @@ static void tcp_try_to_open(struct sock *sk, int flag) tcp_verify_left_out(tp); - if (tp->retrans_out == 0) + if (!tp->frto_counter && tp->retrans_out == 0) tp->retrans_stamp = 0; if (flag & FLAG_ECE) diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c index e591e09e5e4..3a835578fd1 100644 --- a/net/ipv6/addrconf.c +++ b/net/ipv6/addrconf.c @@ -1764,14 +1764,16 @@ void addrconf_prefix_rcv(struct net_device *dev, u8 *opt, int len) * 2) Configure prefixes with the auto flag set */ - /* Avoid arithmetic overflow. Really, we could - save rt_expires in seconds, likely valid_lft, - but it would require division in fib gc, that it - not good. - */ - if (valid_lft >= 0x7FFFFFFF/HZ) + if (valid_lft == INFINITY_LIFE_TIME) + rt_expires = ~0UL; + else if (valid_lft >= 0x7FFFFFFF/HZ) { + /* Avoid arithmetic overflow. Really, we could + * save rt_expires in seconds, likely valid_lft, + * but it would require division in fib gc, that it + * not good. + */ rt_expires = 0x7FFFFFFF - (0x7FFFFFFF % HZ); - else + } else rt_expires = valid_lft * HZ; /* @@ -1779,7 +1781,7 @@ void addrconf_prefix_rcv(struct net_device *dev, u8 *opt, int len) * Avoid arithmetic overflow there as well. * Overflow can happen only if HZ < USER_HZ. */ - if (HZ < USER_HZ && rt_expires > 0x7FFFFFFF / USER_HZ) + if (HZ < USER_HZ && ~rt_expires && rt_expires > 0x7FFFFFFF / USER_HZ) rt_expires = 0x7FFFFFFF / USER_HZ; if (pinfo->onlink) { @@ -1788,17 +1790,28 @@ void addrconf_prefix_rcv(struct net_device *dev, u8 *opt, int len) dev->ifindex, 1); if (rt && ((rt->rt6i_flags & (RTF_GATEWAY | RTF_DEFAULT)) == 0)) { - if (rt->rt6i_flags&RTF_EXPIRES) { - if (valid_lft == 0) { - ip6_del_rt(rt); - rt = NULL; - } else { - rt->rt6i_expires = jiffies + rt_expires; - } + /* Autoconf prefix route */ + if (valid_lft == 0) { + ip6_del_rt(rt); + rt = NULL; + } else if (~rt_expires) { + /* not infinity */ + rt->rt6i_expires = jiffies + rt_expires; + rt->rt6i_flags |= RTF_EXPIRES; + } else { + rt->rt6i_flags &= ~RTF_EXPIRES; + rt->rt6i_expires = 0; } } else if (valid_lft) { + int flags = RTF_ADDRCONF | RTF_PREFIX_RT; + clock_t expires = 0; + if (~rt_expires) { + /* not infinity */ + flags |= RTF_EXPIRES; + expires = jiffies_to_clock_t(rt_expires); + } addrconf_prefix_route(&pinfo->prefix, pinfo->prefix_len, - dev, jiffies_to_clock_t(rt_expires), RTF_ADDRCONF|RTF_EXPIRES|RTF_PREFIX_RT); + dev, expires, flags); } if (rt) dst_release(&rt->u.dst); @@ -2021,7 +2034,8 @@ static int inet6_addr_add(struct net *net, int ifindex, struct in6_addr *pfx, struct inet6_dev *idev; struct net_device *dev; int scope; - u32 flags = RTF_EXPIRES; + u32 flags; + clock_t expires; ASSERT_RTNL(); @@ -2041,8 +2055,13 @@ static int inet6_addr_add(struct net *net, int ifindex, struct in6_addr *pfx, if (valid_lft == INFINITY_LIFE_TIME) { ifa_flags |= IFA_F_PERMANENT; flags = 0; - } else if (valid_lft >= 0x7FFFFFFF/HZ) - valid_lft = 0x7FFFFFFF/HZ; + expires = 0; + } else { + if (valid_lft >= 0x7FFFFFFF/HZ) + valid_lft = 0x7FFFFFFF/HZ; + flags = RTF_EXPIRES; + expires = jiffies_to_clock_t(valid_lft * HZ); + } if (prefered_lft == 0) ifa_flags |= IFA_F_DEPRECATED; @@ -2060,7 +2079,7 @@ static int inet6_addr_add(struct net *net, int ifindex, struct in6_addr *pfx, spin_unlock_bh(&ifp->lock); addrconf_prefix_route(&ifp->addr, ifp->prefix_len, dev, - jiffies_to_clock_t(valid_lft * HZ), flags); + expires, flags); /* * Note that section 3.1 of RFC 4429 indicates * that the Optimistic flag should not be set for @@ -3148,7 +3167,8 @@ inet6_rtm_deladdr(struct sk_buff *skb, struct nlmsghdr *nlh, void *arg) static int inet6_addr_modify(struct inet6_ifaddr *ifp, u8 ifa_flags, u32 prefered_lft, u32 valid_lft) { - u32 flags = RTF_EXPIRES; + u32 flags; + clock_t expires; if (!valid_lft || (prefered_lft > valid_lft)) return -EINVAL; @@ -3156,8 +3176,13 @@ static int inet6_addr_modify(struct inet6_ifaddr *ifp, u8 ifa_flags, if (valid_lft == INFINITY_LIFE_TIME) { ifa_flags |= IFA_F_PERMANENT; flags = 0; - } else if (valid_lft >= 0x7FFFFFFF/HZ) - valid_lft = 0x7FFFFFFF/HZ; + expires = 0; + } else { + if (valid_lft >= 0x7FFFFFFF/HZ) + valid_lft = 0x7FFFFFFF/HZ; + flags = RTF_EXPIRES; + expires = jiffies_to_clock_t(valid_lft * HZ); + } if (prefered_lft == 0) ifa_flags |= IFA_F_DEPRECATED; @@ -3176,7 +3201,7 @@ static int inet6_addr_modify(struct inet6_ifaddr *ifp, u8 ifa_flags, ipv6_ifa_notify(0, ifp); addrconf_prefix_route(&ifp->addr, ifp->prefix_len, ifp->idev->dev, - jiffies_to_clock_t(valid_lft * HZ), flags); + expires, flags); addrconf_verify(0); return 0; @@ -4242,7 +4267,7 @@ static void addrconf_sysctl_register(struct inet6_dev *idev) neigh_sysctl_register(idev->dev, idev->nd_parms, NET_IPV6, NET_IPV6_NEIGH, "ipv6", &ndisc_ifinfo_sysctl_change, - NULL); + ndisc_ifinfo_sysctl_strategy); __addrconf_sysctl_register(dev_net(idev->dev), idev->dev->name, idev->dev->ifindex, idev, &idev->cnf); } diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c index 0af2e055f88..48cdce9c696 100644 --- a/net/ipv6/ip6_output.c +++ b/net/ipv6/ip6_output.c @@ -780,7 +780,7 @@ slow_path: * Allocate buffer. */ - if ((frag = alloc_skb(len+hlen+sizeof(struct frag_hdr)+LL_RESERVED_SPACE(rt->u.dst.dev), GFP_ATOMIC)) == NULL) { + if ((frag = alloc_skb(len+hlen+sizeof(struct frag_hdr)+LL_ALLOCATED_SPACE(rt->u.dst.dev), GFP_ATOMIC)) == NULL) { NETDEBUG(KERN_INFO "IPv6: frag: no memory for new fragment!\n"); IP6_INC_STATS(ip6_dst_idev(skb->dst), IPSTATS_MIB_FRAGFAILS); diff --git a/net/ipv6/mcast.c b/net/ipv6/mcast.c index 54f91efdae5..fd632dd7f98 100644 --- a/net/ipv6/mcast.c +++ b/net/ipv6/mcast.c @@ -1411,7 +1411,7 @@ static struct sk_buff *mld_newpack(struct net_device *dev, int size) IPV6_TLV_PADN, 0 }; /* we assume size > sizeof(ra) here */ - skb = sock_alloc_send_skb(sk, size + LL_RESERVED_SPACE(dev), 1, &err); + skb = sock_alloc_send_skb(sk, size + LL_ALLOCATED_SPACE(dev), 1, &err); if (!skb) return NULL; @@ -1790,7 +1790,7 @@ static void igmp6_send(struct in6_addr *addr, struct net_device *dev, int type) payload_len = len + sizeof(ra); full_len = sizeof(struct ipv6hdr) + payload_len; - skb = sock_alloc_send_skb(sk, LL_RESERVED_SPACE(dev) + full_len, 1, &err); + skb = sock_alloc_send_skb(sk, LL_ALLOCATED_SPACE(dev) + full_len, 1, &err); if (skb == NULL) { rcu_read_lock(); diff --git a/net/ipv6/ndisc.c b/net/ipv6/ndisc.c index 2c74885f835..282fdb31f8e 100644 --- a/net/ipv6/ndisc.c +++ b/net/ipv6/ndisc.c @@ -479,7 +479,7 @@ static void __ndisc_send(struct net_device *dev, skb = sock_alloc_send_skb(sk, (MAX_HEADER + sizeof(struct ipv6hdr) + - len + LL_RESERVED_SPACE(dev)), + len + LL_ALLOCATED_SPACE(dev)), 1, &err); if (!skb) { ND_PRINTK0(KERN_ERR @@ -1521,7 +1521,7 @@ void ndisc_send_redirect(struct sk_buff *skb, struct neighbour *neigh, buff = sock_alloc_send_skb(sk, (MAX_HEADER + sizeof(struct ipv6hdr) + - len + LL_RESERVED_SPACE(dev)), + len + LL_ALLOCATED_SPACE(dev)), 1, &err); if (buff == NULL) { ND_PRINTK0(KERN_ERR @@ -1727,10 +1727,10 @@ int ndisc_ifinfo_sysctl_change(struct ctl_table *ctl, int write, struct file * f return ret; } -static int ndisc_ifinfo_sysctl_strategy(ctl_table *ctl, int __user *name, - int nlen, void __user *oldval, - size_t __user *oldlenp, - void __user *newval, size_t newlen) +int ndisc_ifinfo_sysctl_strategy(ctl_table *ctl, int __user *name, + int nlen, void __user *oldval, + size_t __user *oldlenp, + void __user *newval, size_t newlen) { struct net_device *dev = ctl->extra1; struct inet6_dev *idev; diff --git a/net/ipv6/raw.c b/net/ipv6/raw.c index 396f0ea1109..232e0dc45bf 100644 --- a/net/ipv6/raw.c +++ b/net/ipv6/raw.c @@ -609,7 +609,6 @@ static int rawv6_send_hdrinc(struct sock *sk, void *from, int length, struct ipv6_pinfo *np = inet6_sk(sk); struct ipv6hdr *iph; struct sk_buff *skb; - unsigned int hh_len; int err; if (length > rt->u.dst.dev->mtu) { @@ -619,13 +618,12 @@ static int rawv6_send_hdrinc(struct sock *sk, void *from, int length, if (flags&MSG_PROBE) goto out; - hh_len = LL_RESERVED_SPACE(rt->u.dst.dev); - - skb = sock_alloc_send_skb(sk, length+hh_len+15, - flags&MSG_DONTWAIT, &err); + skb = sock_alloc_send_skb(sk, + length + LL_ALLOCATED_SPACE(rt->u.dst.dev) + 15, + flags & MSG_DONTWAIT, &err); if (skb == NULL) goto error; - skb_reserve(skb, hh_len); + skb_reserve(skb, LL_RESERVED_SPACE(rt->u.dst.dev)); skb->priority = sk->sk_priority; skb->mark = sk->sk_mark; diff --git a/net/ipv6/route.c b/net/ipv6/route.c index 12bba088034..48534c6c073 100644 --- a/net/ipv6/route.c +++ b/net/ipv6/route.c @@ -109,7 +109,7 @@ static struct dst_ops ip6_dst_ops_template = { .negative_advice = ip6_negative_advice, .link_failure = ip6_link_failure, .update_pmtu = ip6_rt_update_pmtu, - .local_out = ip6_local_out, + .local_out = __ip6_local_out, .entry_size = sizeof(struct rt6_info), .entries = ATOMIC_INIT(0), }; @@ -475,7 +475,7 @@ int rt6_route_rcv(struct net_device *dev, u8 *opt, int len, lifetime = ntohl(rinfo->lifetime); if (lifetime == 0xffffffff) { /* infinity */ - } else if (lifetime > 0x7fffffff/HZ) { + } else if (lifetime > 0x7fffffff/HZ - 1) { /* Avoid arithmetic overflow */ lifetime = 0x7fffffff/HZ - 1; } @@ -1106,7 +1106,9 @@ int ip6_route_add(struct fib6_config *cfg) } rt->u.dst.obsolete = -1; - rt->rt6i_expires = jiffies + clock_t_to_jiffies(cfg->fc_expires); + rt->rt6i_expires = (cfg->fc_flags & RTF_EXPIRES) ? + jiffies + clock_t_to_jiffies(cfg->fc_expires) : + 0; if (cfg->fc_protocol == RTPROT_UNSPEC) cfg->fc_protocol = RTPROT_BOOT; @@ -2200,7 +2202,9 @@ static int rt6_fill_node(struct sk_buff *skb, struct rt6_info *rt, NLA_PUT_U32(skb, RTA_PRIORITY, rt->rt6i_metric); - expires = rt->rt6i_expires ? rt->rt6i_expires - jiffies : 0; + expires = (rt->rt6i_flags & RTF_EXPIRES) ? + rt->rt6i_expires - jiffies : 0; + if (rtnl_put_cacheinfo(skb, &rt->u.dst, 0, 0, 0, expires, rt->u.dst.error) < 0) goto nla_put_failure; diff --git a/net/irda/discovery.c b/net/irda/discovery.c index bfacef8b76f..a6f99b5a149 100644 --- a/net/irda/discovery.c +++ b/net/irda/discovery.c @@ -40,6 +40,8 @@ #include <net/irda/discovery.h> +#include <asm/unaligned.h> + /* * Function irlmp_add_discovery (cachelog, discovery) * @@ -87,7 +89,7 @@ void irlmp_add_discovery(hashbin_t *cachelog, discovery_t *new) */ hashbin_remove_this(cachelog, (irda_queue_t *) node); /* Check if hints bits are unchanged */ - if(u16ho(node->data.hints) == u16ho(new->data.hints)) + if (get_unaligned((__u16 *)node->data.hints) == get_unaligned((__u16 *)new->data.hints)) /* Set time of first discovery for this node */ new->firststamp = node->firststamp; kfree(node); @@ -281,9 +283,9 @@ struct irda_device_info *irlmp_copy_discoveries(hashbin_t *log, int *pn, /* Mask out the ones we don't want : * We want to match the discovery mask, and to get only * the most recent one (unless we want old ones) */ - if ((u16ho(discovery->data.hints) & mask) && + if ((get_unaligned((__u16 *)discovery->data.hints) & mask) && ((old_entries) || - ((jiffies - discovery->firststamp) < j_timeout)) ) { + ((jiffies - discovery->firststamp) < j_timeout))) { /* Create buffer as needed. * As this function get called a lot and most time * we don't have anything to put in the log (we are diff --git a/net/irda/irlmp.c b/net/irda/irlmp.c index 1f81f8e7c61..7bf5b913828 100644 --- a/net/irda/irlmp.c +++ b/net/irda/irlmp.c @@ -1062,7 +1062,8 @@ void irlmp_discovery_expiry(discinfo_t *expiries, int number) for(i = 0; i < number; i++) { /* Check if we should notify client */ if ((client->expir_callback) && - (client->hint_mask.word & u16ho(expiries[i].hints) + (client->hint_mask.word & + get_unaligned((__u16 *)expiries[i].hints) & 0x7f7f) ) client->expir_callback(&(expiries[i]), EXPIRY_TIMEOUT, @@ -1086,7 +1087,7 @@ discovery_t *irlmp_get_discovery_response(void) IRDA_ASSERT(irlmp != NULL, return NULL;); - u16ho(irlmp->discovery_rsp.data.hints) = irlmp->hints.word; + put_unaligned(irlmp->hints.word, (__u16 *)irlmp->discovery_rsp.data.hints); /* * Set character set for device name (we use ASCII), and diff --git a/net/irda/irnet/irnet_irda.c b/net/irda/irnet/irnet_irda.c index 75497e55927..cf9a4b531a9 100644 --- a/net/irda/irnet/irnet_irda.c +++ b/net/irda/irnet/irnet_irda.c @@ -10,6 +10,7 @@ #include "irnet_irda.h" /* Private header */ #include <linux/seq_file.h> +#include <asm/unaligned.h> /* * PPP disconnect work: we need to make sure we're in @@ -1673,7 +1674,7 @@ irnet_discovery_indication(discinfo_t * discovery, /* Notify the control channel */ irnet_post_event(NULL, IRNET_DISCOVER, discovery->saddr, discovery->daddr, discovery->info, - u16ho(discovery->hints)); + get_unaligned((__u16 *)discovery->hints)); DEXIT(IRDA_OCB_TRACE, "\n"); } @@ -1704,7 +1705,7 @@ irnet_expiry_indication(discinfo_t * expiry, /* Notify the control channel */ irnet_post_event(NULL, IRNET_EXPIRE, expiry->saddr, expiry->daddr, expiry->info, - u16ho(expiry->hints)); + get_unaligned((__u16 *)expiry->hints)); DEXIT(IRDA_OCB_TRACE, "\n"); } diff --git a/net/mac80211/debugfs_key.c b/net/mac80211/debugfs_key.c index 879e7210458..19efc3a6a93 100644 --- a/net/mac80211/debugfs_key.c +++ b/net/mac80211/debugfs_key.c @@ -255,14 +255,23 @@ void ieee80211_debugfs_key_remove(struct ieee80211_key *key) void ieee80211_debugfs_key_add_default(struct ieee80211_sub_if_data *sdata) { char buf[50]; + struct ieee80211_key *key; if (!sdata->debugfsdir) return; - sprintf(buf, "../keys/%d", sdata->default_key->debugfs.cnt); - sdata->debugfs.default_key = - debugfs_create_symlink("default_key", sdata->debugfsdir, buf); + /* this is running under the key lock */ + + key = sdata->default_key; + if (key) { + sprintf(buf, "../keys/%d", key->debugfs.cnt); + sdata->debugfs.default_key = + debugfs_create_symlink("default_key", + sdata->debugfsdir, buf); + } else + ieee80211_debugfs_key_remove_default(sdata); } + void ieee80211_debugfs_key_remove_default(struct ieee80211_sub_if_data *sdata) { if (!sdata) diff --git a/net/mac80211/iface.c b/net/mac80211/iface.c index 80954a51218..06e88a5a036 100644 --- a/net/mac80211/iface.c +++ b/net/mac80211/iface.c @@ -54,6 +54,15 @@ int ieee80211_if_add(struct net_device *dev, const char *name, if (!ndev) return -ENOMEM; + ndev->needed_headroom = local->tx_headroom + + 4*6 /* four MAC addresses */ + + 2 + 2 + 2 + 2 /* ctl, dur, seq, qos */ + + 6 /* mesh */ + + 8 /* rfc1042/bridge tunnel */ + - ETH_HLEN /* ethernet hard_header_len */ + + IEEE80211_ENCRYPT_HEADROOM; + ndev->needed_tailroom = IEEE80211_ENCRYPT_TAILROOM; + ret = dev_alloc_name(ndev, ndev->name); if (ret < 0) goto fail; diff --git a/net/mac80211/mesh.c b/net/mac80211/mesh.c index f76bc26ae4d..697ef67f96b 100644 --- a/net/mac80211/mesh.c +++ b/net/mac80211/mesh.c @@ -397,7 +397,7 @@ int ieee80211_new_mesh_header(struct ieee80211s_hdr *meshhdr, put_unaligned(cpu_to_le32(sdata->u.sta.mesh_seqnum), &meshhdr->seqnum); sdata->u.sta.mesh_seqnum++; - return 5; + return 6; } void ieee80211_mesh_init_sdata(struct ieee80211_sub_if_data *sdata) diff --git a/net/mac80211/mesh_hwmp.c b/net/mac80211/mesh_hwmp.c index 3df809222d1..af0cd1e3e21 100644 --- a/net/mac80211/mesh_hwmp.c +++ b/net/mac80211/mesh_hwmp.c @@ -120,7 +120,7 @@ static int mesh_path_sel_frame_tx(enum mpath_frame_type action, u8 flags, *pos++ = WLAN_EID_PREP; break; default: - kfree(skb); + kfree_skb(skb); return -ENOTSUPP; break; } diff --git a/net/mac80211/mesh_pathtbl.c b/net/mac80211/mesh_pathtbl.c index 5845dc21ce8..99c2d360888 100644 --- a/net/mac80211/mesh_pathtbl.c +++ b/net/mac80211/mesh_pathtbl.c @@ -158,19 +158,25 @@ int mesh_path_add(u8 *dst, struct net_device *dev) if (atomic_add_unless(&sdata->u.sta.mpaths, 1, MESH_MAX_MPATHS) == 0) return -ENOSPC; - read_lock(&pathtbl_resize_lock); - new_mpath = kzalloc(sizeof(struct mesh_path), GFP_KERNEL); if (!new_mpath) { atomic_dec(&sdata->u.sta.mpaths); err = -ENOMEM; goto endadd2; } + new_node = kmalloc(sizeof(struct mpath_node), GFP_KERNEL); + if (!new_node) { + kfree(new_mpath); + atomic_dec(&sdata->u.sta.mpaths); + err = -ENOMEM; + goto endadd2; + } + + read_lock(&pathtbl_resize_lock); memcpy(new_mpath->dst, dst, ETH_ALEN); new_mpath->dev = dev; new_mpath->flags = 0; skb_queue_head_init(&new_mpath->frame_queue); - new_node = kmalloc(sizeof(struct mpath_node), GFP_KERNEL); new_node->mpath = new_mpath; new_mpath->timer.data = (unsigned long) new_mpath; new_mpath->timer.function = mesh_path_timer; @@ -202,7 +208,6 @@ int mesh_path_add(u8 *dst, struct net_device *dev) endadd: spin_unlock(&mesh_paths->hashwlock[hash_idx]); -endadd2: read_unlock(&pathtbl_resize_lock); if (!err && grow) { struct mesh_table *oldtbl, *newtbl; @@ -215,10 +220,12 @@ endadd2: return -ENOMEM; } rcu_assign_pointer(mesh_paths, newtbl); + write_unlock(&pathtbl_resize_lock); + synchronize_rcu(); mesh_table_free(oldtbl, false); - write_unlock(&pathtbl_resize_lock); } +endadd2: return err; } diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index a5e5c31c23a..e470bf12b76 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -665,6 +665,26 @@ static void ieee80211_authenticate(struct net_device *dev, mod_timer(&ifsta->timer, jiffies + IEEE80211_AUTH_TIMEOUT); } +static int ieee80211_compatible_rates(struct ieee80211_sta_bss *bss, + struct ieee80211_supported_band *sband, + u64 *rates) +{ + int i, j, count; + *rates = 0; + count = 0; + for (i = 0; i < bss->supp_rates_len; i++) { + int rate = (bss->supp_rates[i] & 0x7F) * 5; + + for (j = 0; j < sband->n_bitrates; j++) + if (sband->bitrates[j].bitrate == rate) { + *rates |= BIT(j); + count++; + break; + } + } + + return count; +} static void ieee80211_send_assoc(struct net_device *dev, struct ieee80211_if_sta *ifsta) @@ -673,11 +693,12 @@ static void ieee80211_send_assoc(struct net_device *dev, struct sk_buff *skb; struct ieee80211_mgmt *mgmt; u8 *pos, *ies; - int i, len; + int i, len, count, rates_len, supp_rates_len; u16 capab; struct ieee80211_sta_bss *bss; int wmm = 0; struct ieee80211_supported_band *sband; + u64 rates = 0; skb = dev_alloc_skb(local->hw.extra_tx_headroom + sizeof(*mgmt) + 200 + ifsta->extra_ie_len + @@ -740,24 +761,39 @@ static void ieee80211_send_assoc(struct net_device *dev, *pos++ = ifsta->ssid_len; memcpy(pos, ifsta->ssid, ifsta->ssid_len); + /* all supported rates should be added here but some APs + * (e.g. D-Link DAP 1353 in b-only mode) don't like that + * Therefore only add rates the AP supports */ + rates_len = ieee80211_compatible_rates(bss, sband, &rates); + supp_rates_len = rates_len; + if (supp_rates_len > 8) + supp_rates_len = 8; + len = sband->n_bitrates; - if (len > 8) - len = 8; - pos = skb_put(skb, len + 2); + pos = skb_put(skb, supp_rates_len + 2); *pos++ = WLAN_EID_SUPP_RATES; - *pos++ = len; - for (i = 0; i < len; i++) { - int rate = sband->bitrates[i].bitrate; - *pos++ = (u8) (rate / 5); - } + *pos++ = supp_rates_len; - if (sband->n_bitrates > len) { - pos = skb_put(skb, sband->n_bitrates - len + 2); - *pos++ = WLAN_EID_EXT_SUPP_RATES; - *pos++ = sband->n_bitrates - len; - for (i = len; i < sband->n_bitrates; i++) { + count = 0; + for (i = 0; i < sband->n_bitrates; i++) { + if (BIT(i) & rates) { int rate = sband->bitrates[i].bitrate; *pos++ = (u8) (rate / 5); + if (++count == 8) + break; + } + } + + if (count == 8) { + pos = skb_put(skb, rates_len - count + 2); + *pos++ = WLAN_EID_EXT_SUPP_RATES; + *pos++ = rates_len - count; + + for (i++; i < sband->n_bitrates; i++) { + if (BIT(i) & rates) { + int rate = sband->bitrates[i].bitrate; + *pos++ = (u8) (rate / 5); + } } } @@ -3410,21 +3446,17 @@ static int ieee80211_sta_config_auth(struct net_device *dev, struct ieee80211_sta_bss *bss, *selected = NULL; int top_rssi = 0, freq; - if (!(ifsta->flags & (IEEE80211_STA_AUTO_SSID_SEL | - IEEE80211_STA_AUTO_BSSID_SEL | IEEE80211_STA_AUTO_CHANNEL_SEL))) { - ifsta->state = IEEE80211_AUTHENTICATE; - ieee80211_sta_reset_auth(dev, ifsta); - return 0; - } - spin_lock_bh(&local->sta_bss_lock); freq = local->oper_channel->center_freq; list_for_each_entry(bss, &local->sta_bss_list, list) { if (!(bss->capability & WLAN_CAPABILITY_ESS)) continue; - if (!!(bss->capability & WLAN_CAPABILITY_PRIVACY) ^ - !!sdata->default_key) + if ((ifsta->flags & (IEEE80211_STA_AUTO_SSID_SEL | + IEEE80211_STA_AUTO_BSSID_SEL | + IEEE80211_STA_AUTO_CHANNEL_SEL)) && + (!!(bss->capability & WLAN_CAPABILITY_PRIVACY) ^ + !!sdata->default_key)) continue; if (!(ifsta->flags & IEEE80211_STA_AUTO_CHANNEL_SEL) && diff --git a/net/mac80211/rx.c b/net/mac80211/rx.c index 02f436a8606..1958bfb361c 100644 --- a/net/mac80211/rx.c +++ b/net/mac80211/rx.c @@ -1305,11 +1305,11 @@ ieee80211_deliver_skb(struct ieee80211_rx_data *rx) if (is_multicast_ether_addr(skb->data)) { if (*mesh_ttl > 0) { xmit_skb = skb_copy(skb, GFP_ATOMIC); - if (!xmit_skb && net_ratelimit()) + if (xmit_skb) + xmit_skb->pkt_type = PACKET_OTHERHOST; + else if (net_ratelimit()) printk(KERN_DEBUG "%s: failed to clone " "multicast frame\n", dev->name); - else - xmit_skb->pkt_type = PACKET_OTHERHOST; } else IEEE80211_IFSTA_MESH_CTR_INC(&sdata->u.sta, dropped_frames_ttl); @@ -1395,7 +1395,7 @@ ieee80211_rx_h_amsdu(struct ieee80211_rx_data *rx) padding = ((4 - subframe_len) & 0x3); /* the last MSDU has no padding */ if (subframe_len > remaining) { - printk(KERN_DEBUG "%s: wrong buffer size", dev->name); + printk(KERN_DEBUG "%s: wrong buffer size\n", dev->name); return RX_DROP_UNUSABLE; } @@ -1418,7 +1418,7 @@ ieee80211_rx_h_amsdu(struct ieee80211_rx_data *rx) eth = (struct ethhdr *) skb_pull(skb, ntohs(len) + padding); if (!eth) { - printk(KERN_DEBUG "%s: wrong buffer size ", + printk(KERN_DEBUG "%s: wrong buffer size\n", dev->name); dev_kfree_skb(frame); return RX_DROP_UNUSABLE; @@ -1952,7 +1952,7 @@ static void __ieee80211_rx_handle_packet(struct ieee80211_hw *hw, if (!skb_new) { if (net_ratelimit()) printk(KERN_DEBUG "%s: failed to copy " - "multicast frame for %s", + "multicast frame for %s\n", wiphy_name(local->hw.wiphy), prev->dev->name); continue; diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c index f35eaea98e7..1d7dd54aace 100644 --- a/net/mac80211/tx.c +++ b/net/mac80211/tx.c @@ -1562,13 +1562,13 @@ int ieee80211_subif_start_xmit(struct sk_buff *skb, * be cloned. This could happen, e.g., with Linux bridge code passing * us broadcast frames. */ - if (head_need > 0 || skb_cloned(skb)) { + if (head_need > 0 || skb_header_cloned(skb)) { #if 0 printk(KERN_DEBUG "%s: need to reallocate buffer for %d bytes " "of headroom\n", dev->name, head_need); #endif - if (skb_cloned(skb)) + if (skb_header_cloned(skb)) I802_DEBUG_INC(local->tx_expand_skb_head_cloned); else I802_DEBUG_INC(local->tx_expand_skb_head); @@ -1898,6 +1898,7 @@ struct sk_buff *ieee80211_beacon_get(struct ieee80211_hw *hw, control->flags |= IEEE80211_TXCTL_SHORT_PREAMBLE; control->antenna_sel_tx = local->hw.conf.antenna_sel_tx; control->flags |= IEEE80211_TXCTL_NO_ACK; + control->flags |= IEEE80211_TXCTL_DO_NOT_ENCRYPT; control->retry_limit = 1; control->flags |= IEEE80211_TXCTL_CLEAR_PS_FILT; } diff --git a/net/mac80211/util.c b/net/mac80211/util.c index cc9f715c7bf..131e9e6c8a3 100644 --- a/net/mac80211/util.c +++ b/net/mac80211/util.c @@ -153,15 +153,15 @@ int ieee80211_get_mesh_hdrlen(struct ieee80211s_hdr *meshhdr) /* 7.1.3.5a.2 */ switch (ae) { case 0: - return 5; + return 6; case 1: - return 11; + return 12; case 2: - return 17; + return 18; case 3: - return 23; + return 24; default: - return 5; + return 6; } } @@ -389,6 +389,41 @@ void ieee80211_iterate_active_interfaces( struct ieee80211_local *local = hw_to_local(hw); struct ieee80211_sub_if_data *sdata; + rtnl_lock(); + + list_for_each_entry(sdata, &local->interfaces, list) { + switch (sdata->vif.type) { + case IEEE80211_IF_TYPE_INVALID: + case IEEE80211_IF_TYPE_MNTR: + case IEEE80211_IF_TYPE_VLAN: + continue; + case IEEE80211_IF_TYPE_AP: + case IEEE80211_IF_TYPE_STA: + case IEEE80211_IF_TYPE_IBSS: + case IEEE80211_IF_TYPE_WDS: + case IEEE80211_IF_TYPE_MESH_POINT: + break; + } + if (sdata->dev == local->mdev) + continue; + if (netif_running(sdata->dev)) + iterator(data, sdata->dev->dev_addr, + &sdata->vif); + } + + rtnl_unlock(); +} +EXPORT_SYMBOL_GPL(ieee80211_iterate_active_interfaces); + +void ieee80211_iterate_active_interfaces_atomic( + struct ieee80211_hw *hw, + void (*iterator)(void *data, u8 *mac, + struct ieee80211_vif *vif), + void *data) +{ + struct ieee80211_local *local = hw_to_local(hw); + struct ieee80211_sub_if_data *sdata; + rcu_read_lock(); list_for_each_entry_rcu(sdata, &local->interfaces, list) { @@ -413,4 +448,4 @@ void ieee80211_iterate_active_interfaces( rcu_read_unlock(); } -EXPORT_SYMBOL_GPL(ieee80211_iterate_active_interfaces); +EXPORT_SYMBOL_GPL(ieee80211_iterate_active_interfaces_atomic); diff --git a/net/mac80211/wme.c b/net/mac80211/wme.c index 64faa3dc488..dc1598b8600 100644 --- a/net/mac80211/wme.c +++ b/net/mac80211/wme.c @@ -394,7 +394,8 @@ static int wme_qdiscop_init(struct Qdisc *qd, struct nlattr *opt) qd->handle); if (!q->queues[i]) { q->queues[i] = &noop_qdisc; - printk(KERN_ERR "%s child qdisc %i creation failed", dev->name, i); + printk(KERN_ERR "%s child qdisc %i creation failed\n", + dev->name, i); } } diff --git a/net/netfilter/nf_conntrack_netlink.c b/net/netfilter/nf_conntrack_netlink.c index 16774ecd1c4..0edefcfc594 100644 --- a/net/netfilter/nf_conntrack_netlink.c +++ b/net/netfilter/nf_conntrack_netlink.c @@ -472,6 +472,9 @@ static int ctnetlink_conntrack_event(struct notifier_block *this, goto nla_put_failure; nla_nest_end(skb, nest_parms); + if (ctnetlink_dump_id(skb, ct) < 0) + goto nla_put_failure; + if (events & IPCT_DESTROY) { if (ctnetlink_dump_counters(skb, ct, IP_CT_DIR_ORIGINAL) < 0 || ctnetlink_dump_counters(skb, ct, IP_CT_DIR_REPLY) < 0) diff --git a/net/netfilter/xt_iprange.c b/net/netfilter/xt_iprange.c index 500528d60cd..c63e9333c75 100644 --- a/net/netfilter/xt_iprange.c +++ b/net/netfilter/xt_iprange.c @@ -179,3 +179,5 @@ module_exit(iprange_mt_exit); MODULE_LICENSE("GPL"); MODULE_AUTHOR("Jozsef Kadlecsik <kadlec@blackhole.kfki.hu>, Jan Engelhardt <jengelh@computergmbh.de>"); MODULE_DESCRIPTION("Xtables: arbitrary IPv4 range matching"); +MODULE_ALIAS("ipt_iprange"); +MODULE_ALIAS("ip6t_iprange"); diff --git a/net/packet/af_packet.c b/net/packet/af_packet.c index 25070240d4a..2cee87da444 100644 --- a/net/packet/af_packet.c +++ b/net/packet/af_packet.c @@ -743,7 +743,7 @@ static int packet_sendmsg(struct kiocb *iocb, struct socket *sock, if (len > dev->mtu+reserve) goto out_unlock; - skb = sock_alloc_send_skb(sk, len + LL_RESERVED_SPACE(dev), + skb = sock_alloc_send_skb(sk, len + LL_ALLOCATED_SPACE(dev), msg->msg_flags & MSG_DONTWAIT, &err); if (skb==NULL) goto out_unlock; diff --git a/net/sched/cls_api.c b/net/sched/cls_api.c index 1086df7478b..9360fc81e8c 100644 --- a/net/sched/cls_api.c +++ b/net/sched/cls_api.c @@ -220,7 +220,7 @@ replay: tp = kzalloc(sizeof(*tp), GFP_KERNEL); if (tp == NULL) goto errout; - err = -EINVAL; + err = -ENOENT; tp_ops = tcf_proto_lookup_ops(tca[TCA_KIND]); if (tp_ops == NULL) { #ifdef CONFIG_KMOD diff --git a/net/sctp/sm_make_chunk.c b/net/sctp/sm_make_chunk.c index 81b606424e1..bbc7107c86c 100644 --- a/net/sctp/sm_make_chunk.c +++ b/net/sctp/sm_make_chunk.c @@ -2418,7 +2418,8 @@ static int sctp_process_param(struct sctp_association *asoc, break; case SCTP_PARAM_IPV6_ADDRESS: - asoc->peer.ipv6_address = 1; + if (PF_INET6 == asoc->base.sk->sk_family) + asoc->peer.ipv6_address = 1; break; case SCTP_PARAM_HOST_NAME_ADDRESS: @@ -2829,6 +2830,19 @@ static __be16 sctp_process_asconf_param(struct sctp_association *asoc, addr_param = (union sctp_addr_param *) ((void *)asconf_param + sizeof(sctp_addip_param_t)); + switch (addr_param->v4.param_hdr.type) { + case SCTP_PARAM_IPV6_ADDRESS: + if (!asoc->peer.ipv6_address) + return SCTP_ERROR_INV_PARAM; + break; + case SCTP_PARAM_IPV4_ADDRESS: + if (!asoc->peer.ipv4_address) + return SCTP_ERROR_INV_PARAM; + break; + default: + return SCTP_ERROR_INV_PARAM; + } + af = sctp_get_af_specific(param_type2af(addr_param->v4.param_hdr.type)); if (unlikely(!af)) return SCTP_ERROR_INV_PARAM; diff --git a/net/sunrpc/auth_generic.c b/net/sunrpc/auth_generic.c index d927d9f5741..744b79fdcb1 100644 --- a/net/sunrpc/auth_generic.c +++ b/net/sunrpc/auth_generic.c @@ -17,8 +17,8 @@ # define RPCDBG_FACILITY RPCDBG_AUTH #endif -#define RPC_ANONYMOUS_USERID ((uid_t)-2) -#define RPC_ANONYMOUS_GROUPID ((gid_t)-2) +#define RPC_MACHINE_CRED_USERID ((uid_t)0) +#define RPC_MACHINE_CRED_GROUPID ((gid_t)0) struct generic_cred { struct rpc_cred gc_base; @@ -44,8 +44,8 @@ EXPORT_SYMBOL_GPL(rpc_lookup_cred); struct rpc_cred *rpc_lookup_machine_cred(void) { struct auth_cred acred = { - .uid = RPC_ANONYMOUS_USERID, - .gid = RPC_ANONYMOUS_GROUPID, + .uid = RPC_MACHINE_CRED_USERID, + .gid = RPC_MACHINE_CRED_GROUPID, .machine_cred = 1, }; diff --git a/net/sunrpc/svc_xprt.c b/net/sunrpc/svc_xprt.c index d8e8d79a845..e46c825f495 100644 --- a/net/sunrpc/svc_xprt.c +++ b/net/sunrpc/svc_xprt.c @@ -6,30 +6,9 @@ #include <linux/sched.h> #include <linux/errno.h> -#include <linux/fcntl.h> -#include <linux/net.h> -#include <linux/in.h> -#include <linux/inet.h> -#include <linux/udp.h> -#include <linux/tcp.h> -#include <linux/unistd.h> -#include <linux/slab.h> -#include <linux/netdevice.h> -#include <linux/skbuff.h> -#include <linux/file.h> #include <linux/freezer.h> #include <linux/kthread.h> #include <net/sock.h> -#include <net/checksum.h> -#include <net/ip.h> -#include <net/ipv6.h> -#include <net/tcp_states.h> -#include <linux/uaccess.h> -#include <asm/ioctls.h> - -#include <linux/sunrpc/types.h> -#include <linux/sunrpc/clnt.h> -#include <linux/sunrpc/xdr.h> #include <linux/sunrpc/stats.h> #include <linux/sunrpc/svc_xprt.h> @@ -296,8 +275,6 @@ void svc_xprt_enqueue(struct svc_xprt *xprt) if (!(xprt->xpt_flags & ((1<<XPT_CONN)|(1<<XPT_DATA)|(1<<XPT_CLOSE)|(1<<XPT_DEFERRED)))) return; - if (test_bit(XPT_DEAD, &xprt->xpt_flags)) - return; cpu = get_cpu(); pool = svc_pool_for_cpu(xprt->xpt_server, cpu); diff --git a/net/sunrpc/svcauth_unix.c b/net/sunrpc/svcauth_unix.c index 3f30ee6006a..f24800f2c09 100644 --- a/net/sunrpc/svcauth_unix.c +++ b/net/sunrpc/svcauth_unix.c @@ -278,7 +278,7 @@ static int ip_map_show(struct seq_file *m, dom = im->m_client->h.name; if (ipv6_addr_v4mapped(&addr)) { - seq_printf(m, "%s" NIPQUAD_FMT "%s\n", + seq_printf(m, "%s " NIPQUAD_FMT " %s\n", im->m_class, ntohl(addr.s6_addr32[3]) >> 24 & 0xff, ntohl(addr.s6_addr32[3]) >> 16 & 0xff, @@ -286,7 +286,7 @@ static int ip_map_show(struct seq_file *m, ntohl(addr.s6_addr32[3]) >> 0 & 0xff, dom); } else { - seq_printf(m, "%s" NIP6_FMT "%s\n", + seq_printf(m, "%s " NIP6_FMT " %s\n", im->m_class, NIP6(addr), dom); } return 0; diff --git a/net/sunrpc/xprtrdma/svc_rdma_recvfrom.c b/net/sunrpc/xprtrdma/svc_rdma_recvfrom.c index c22d6b6f2db..06ab4841537 100644 --- a/net/sunrpc/xprtrdma/svc_rdma_recvfrom.c +++ b/net/sunrpc/xprtrdma/svc_rdma_recvfrom.c @@ -260,11 +260,16 @@ static int rdma_read_max_sge(struct svcxprt_rdma *xprt, int sge_count) * On our side, we need to read into a pagelist. The first page immediately * follows the RPC header. * - * This function returns 1 to indicate success. The data is not yet in + * This function returns: + * 0 - No error and no read-list found. + * + * 1 - Successful read-list processing. The data is not yet in * the pagelist and therefore the RPC request must be deferred. The * I/O completion will enqueue the transport again and * svc_rdma_recvfrom will complete the request. * + * <0 - Error processing/posting read-list. + * * NOTE: The ctxt must not be touched after the last WR has been posted * because the I/O completion processing may occur on another * processor and free / modify the context. Ne touche pas! @@ -284,7 +289,6 @@ static int rdma_read_xdr(struct svcxprt_rdma *xprt, u64 sgl_offset; struct rpcrdma_read_chunk *ch; struct svc_rdma_op_ctxt *ctxt = NULL; - struct svc_rdma_op_ctxt *head; struct svc_rdma_op_ctxt *tmp_sge_ctxt; struct svc_rdma_op_ctxt *tmp_ch_ctxt; struct chunk_sge *ch_sge_ary; @@ -302,25 +306,19 @@ static int rdma_read_xdr(struct svcxprt_rdma *xprt, ch_sge_ary = (struct chunk_sge *)tmp_ch_ctxt->sge; svc_rdma_rcl_chunk_counts(ch, &ch_count, &byte_count); + if (ch_count > RPCSVC_MAXPAGES) + return -EINVAL; sge_count = rdma_rcl_to_sge(xprt, rqstp, hdr_ctxt, rmsgp, sge, ch_sge_ary, ch_count, byte_count); - head = svc_rdma_get_context(xprt); sgl_offset = 0; ch_no = 0; for (ch = (struct rpcrdma_read_chunk *)&rmsgp->rm_body.rm_chunks[0]; ch->rc_discrim != 0; ch++, ch_no++) { next_sge: - if (!ctxt) - ctxt = head; - else { - ctxt->next = svc_rdma_get_context(xprt); - ctxt = ctxt->next; - } - ctxt->next = NULL; + ctxt = svc_rdma_get_context(xprt); ctxt->direction = DMA_FROM_DEVICE; - clear_bit(RDMACTXT_F_READ_DONE, &ctxt->flags); clear_bit(RDMACTXT_F_LAST_CTXT, &ctxt->flags); /* Prepare READ WR */ @@ -347,20 +345,15 @@ next_sge: * the client and the RPC needs to be enqueued. */ set_bit(RDMACTXT_F_LAST_CTXT, &ctxt->flags); - ctxt->next = hdr_ctxt; - hdr_ctxt->next = head; + ctxt->read_hdr = hdr_ctxt; } /* Post the read */ err = svc_rdma_send(xprt, &read_wr); if (err) { - printk(KERN_ERR "svcrdma: Error posting send = %d\n", + printk(KERN_ERR "svcrdma: Error %d posting RDMA_READ\n", err); - /* - * Break the circular list so free knows when - * to stop if the error happened to occur on - * the last read - */ - ctxt->next = NULL; + set_bit(XPT_CLOSE, &xprt->sc_xprt.xpt_flags); + svc_rdma_put_context(ctxt, 0); goto out; } atomic_inc(&rdma_stat_read); @@ -371,7 +364,7 @@ next_sge: goto next_sge; } sgl_offset = 0; - err = 0; + err = 1; } out: @@ -389,25 +382,12 @@ next_sge: while (rqstp->rq_resused) rqstp->rq_respages[--rqstp->rq_resused] = NULL; - if (err) { - printk(KERN_ERR "svcrdma : RDMA_READ error = %d\n", err); - set_bit(XPT_CLOSE, &xprt->sc_xprt.xpt_flags); - /* Free the linked list of read contexts */ - while (head != NULL) { - ctxt = head->next; - svc_rdma_put_context(head, 1); - head = ctxt; - } - return 0; - } - - return 1; + return err; } static int rdma_read_complete(struct svc_rqst *rqstp, - struct svc_rdma_op_ctxt *data) + struct svc_rdma_op_ctxt *head) { - struct svc_rdma_op_ctxt *head = data->next; int page_no; int ret; @@ -433,21 +413,12 @@ static int rdma_read_complete(struct svc_rqst *rqstp, rqstp->rq_arg.len = head->arg.len; rqstp->rq_arg.buflen = head->arg.buflen; + /* Free the context */ + svc_rdma_put_context(head, 0); + /* XXX: What should this be? */ rqstp->rq_prot = IPPROTO_MAX; - - /* - * Free the contexts we used to build the RDMA_READ. We have - * to be careful here because the context list uses the same - * next pointer used to chain the contexts associated with the - * RDMA_READ - */ - data->next = NULL; /* terminate circular list */ - do { - data = head->next; - svc_rdma_put_context(head, 0); - head = data; - } while (head != NULL); + svc_xprt_copy_addrs(rqstp, rqstp->rq_xprt); ret = rqstp->rq_arg.head[0].iov_len + rqstp->rq_arg.page_len @@ -457,8 +428,6 @@ static int rdma_read_complete(struct svc_rqst *rqstp, ret, rqstp->rq_arg.len, rqstp->rq_arg.head[0].iov_base, rqstp->rq_arg.head[0].iov_len); - /* Indicate that we've consumed an RQ credit */ - rqstp->rq_xprt_ctxt = rqstp->rq_xprt; svc_xprt_received(rqstp->rq_xprt); return ret; } @@ -480,13 +449,6 @@ int svc_rdma_recvfrom(struct svc_rqst *rqstp) dprintk("svcrdma: rqstp=%p\n", rqstp); - /* - * The rq_xprt_ctxt indicates if we've consumed an RQ credit - * or not. It is used in the rdma xpo_release_rqst function to - * determine whether or not to return an RQ WQE to the RQ. - */ - rqstp->rq_xprt_ctxt = NULL; - spin_lock_bh(&rdma_xprt->sc_read_complete_lock); if (!list_empty(&rdma_xprt->sc_read_complete_q)) { ctxt = list_entry(rdma_xprt->sc_read_complete_q.next, @@ -537,21 +499,22 @@ int svc_rdma_recvfrom(struct svc_rqst *rqstp) /* If the request is invalid, reply with an error */ if (len < 0) { if (len == -ENOSYS) - (void)svc_rdma_send_error(rdma_xprt, rmsgp, ERR_VERS); + svc_rdma_send_error(rdma_xprt, rmsgp, ERR_VERS); goto close_out; } - /* Read read-list data. If we would need to wait, defer - * it. Not that in this case, we don't return the RQ credit - * until after the read completes. - */ - if (rdma_read_xdr(rdma_xprt, rmsgp, rqstp, ctxt)) { + /* Read read-list data. */ + ret = rdma_read_xdr(rdma_xprt, rmsgp, rqstp, ctxt); + if (ret > 0) { + /* read-list posted, defer until data received from client. */ svc_xprt_received(xprt); return 0; } - - /* Indicate we've consumed an RQ credit */ - rqstp->rq_xprt_ctxt = rqstp->rq_xprt; + if (ret < 0) { + /* Post of read-list failed, free context. */ + svc_rdma_put_context(ctxt, 1); + return 0; + } ret = rqstp->rq_arg.head[0].iov_len + rqstp->rq_arg.page_len @@ -569,11 +532,8 @@ int svc_rdma_recvfrom(struct svc_rqst *rqstp) return ret; close_out: - if (ctxt) { + if (ctxt) svc_rdma_put_context(ctxt, 1); - /* Indicate we've consumed an RQ credit */ - rqstp->rq_xprt_ctxt = rqstp->rq_xprt; - } dprintk("svcrdma: transport %p is closing\n", xprt); /* * Set the close bit and enqueue it. svc_recv will see the diff --git a/net/sunrpc/xprtrdma/svc_rdma_sendto.c b/net/sunrpc/xprtrdma/svc_rdma_sendto.c index 981f190c1b3..fb82b1b683f 100644 --- a/net/sunrpc/xprtrdma/svc_rdma_sendto.c +++ b/net/sunrpc/xprtrdma/svc_rdma_sendto.c @@ -389,6 +389,17 @@ static int send_reply(struct svcxprt_rdma *rdma, int page_no; int ret; + /* Post a recv buffer to handle another request. */ + ret = svc_rdma_post_recv(rdma); + if (ret) { + printk(KERN_INFO + "svcrdma: could not post a receive buffer, err=%d." + "Closing transport %p.\n", ret, rdma); + set_bit(XPT_CLOSE, &rdma->sc_xprt.xpt_flags); + svc_rdma_put_context(ctxt, 0); + return -ENOTCONN; + } + /* Prepare the context */ ctxt->pages[0] = page; ctxt->count = 1; diff --git a/net/sunrpc/xprtrdma/svc_rdma_transport.c b/net/sunrpc/xprtrdma/svc_rdma_transport.c index af408fc1263..e132509d1db 100644 --- a/net/sunrpc/xprtrdma/svc_rdma_transport.c +++ b/net/sunrpc/xprtrdma/svc_rdma_transport.c @@ -103,8 +103,8 @@ static int rdma_bump_context_cache(struct svcxprt_rdma *xprt) spin_lock_bh(&xprt->sc_ctxt_lock); if (ctxt) { at_least_one = 1; - ctxt->next = xprt->sc_ctxt_head; - xprt->sc_ctxt_head = ctxt; + INIT_LIST_HEAD(&ctxt->free_list); + list_add(&ctxt->free_list, &xprt->sc_ctxt_free); } else { /* kmalloc failed...give up for now */ xprt->sc_ctxt_cnt--; @@ -123,7 +123,7 @@ struct svc_rdma_op_ctxt *svc_rdma_get_context(struct svcxprt_rdma *xprt) while (1) { spin_lock_bh(&xprt->sc_ctxt_lock); - if (unlikely(xprt->sc_ctxt_head == NULL)) { + if (unlikely(list_empty(&xprt->sc_ctxt_free))) { /* Try to bump my cache. */ spin_unlock_bh(&xprt->sc_ctxt_lock); @@ -136,12 +136,15 @@ struct svc_rdma_op_ctxt *svc_rdma_get_context(struct svcxprt_rdma *xprt) schedule_timeout_uninterruptible(msecs_to_jiffies(500)); continue; } - ctxt = xprt->sc_ctxt_head; - xprt->sc_ctxt_head = ctxt->next; + ctxt = list_entry(xprt->sc_ctxt_free.next, + struct svc_rdma_op_ctxt, + free_list); + list_del_init(&ctxt->free_list); spin_unlock_bh(&xprt->sc_ctxt_lock); ctxt->xprt = xprt; INIT_LIST_HEAD(&ctxt->dto_q); ctxt->count = 0; + atomic_inc(&xprt->sc_ctxt_used); break; } return ctxt; @@ -159,14 +162,15 @@ void svc_rdma_put_context(struct svc_rdma_op_ctxt *ctxt, int free_pages) put_page(ctxt->pages[i]); for (i = 0; i < ctxt->count; i++) - dma_unmap_single(xprt->sc_cm_id->device->dma_device, - ctxt->sge[i].addr, - ctxt->sge[i].length, - ctxt->direction); + ib_dma_unmap_single(xprt->sc_cm_id->device, + ctxt->sge[i].addr, + ctxt->sge[i].length, + ctxt->direction); + spin_lock_bh(&xprt->sc_ctxt_lock); - ctxt->next = xprt->sc_ctxt_head; - xprt->sc_ctxt_head = ctxt; + list_add(&ctxt->free_list, &xprt->sc_ctxt_free); spin_unlock_bh(&xprt->sc_ctxt_lock); + atomic_dec(&xprt->sc_ctxt_used); } /* ib_cq event handler */ @@ -228,23 +232,8 @@ static void dto_tasklet_func(unsigned long data) list_del_init(&xprt->sc_dto_q); spin_unlock_irqrestore(&dto_lock, flags); - if (test_and_clear_bit(RDMAXPRT_RQ_PENDING, &xprt->sc_flags)) { - ib_req_notify_cq(xprt->sc_rq_cq, IB_CQ_NEXT_COMP); - rq_cq_reap(xprt); - set_bit(XPT_DATA, &xprt->sc_xprt.xpt_flags); - /* - * If data arrived before established event, - * don't enqueue. This defers RPC I/O until the - * RDMA connection is complete. - */ - if (!test_bit(RDMAXPRT_CONN_PENDING, &xprt->sc_flags)) - svc_xprt_enqueue(&xprt->sc_xprt); - } - - if (test_and_clear_bit(RDMAXPRT_SQ_PENDING, &xprt->sc_flags)) { - ib_req_notify_cq(xprt->sc_sq_cq, IB_CQ_NEXT_COMP); - sq_cq_reap(xprt); - } + rq_cq_reap(xprt); + sq_cq_reap(xprt); svc_xprt_put(&xprt->sc_xprt); spin_lock_irqsave(&dto_lock, flags); @@ -263,11 +252,15 @@ static void rq_comp_handler(struct ib_cq *cq, void *cq_context) struct svcxprt_rdma *xprt = cq_context; unsigned long flags; + /* Guard against unconditional flush call for destroyed QP */ + if (atomic_read(&xprt->sc_xprt.xpt_ref.refcount)==0) + return; + /* * Set the bit regardless of whether or not it's on the list * because it may be on the list already due to an SQ * completion. - */ + */ set_bit(RDMAXPRT_RQ_PENDING, &xprt->sc_flags); /* @@ -290,6 +283,8 @@ static void rq_comp_handler(struct ib_cq *cq, void *cq_context) * * Take all completing WC off the CQE and enqueue the associated DTO * context on the dto_q for the transport. + * + * Note that caller must hold a transport reference. */ static void rq_cq_reap(struct svcxprt_rdma *xprt) { @@ -297,29 +292,47 @@ static void rq_cq_reap(struct svcxprt_rdma *xprt) struct ib_wc wc; struct svc_rdma_op_ctxt *ctxt = NULL; + if (!test_and_clear_bit(RDMAXPRT_RQ_PENDING, &xprt->sc_flags)) + return; + + ib_req_notify_cq(xprt->sc_rq_cq, IB_CQ_NEXT_COMP); atomic_inc(&rdma_stat_rq_poll); - spin_lock_bh(&xprt->sc_rq_dto_lock); while ((ret = ib_poll_cq(xprt->sc_rq_cq, 1, &wc)) > 0) { ctxt = (struct svc_rdma_op_ctxt *)(unsigned long)wc.wr_id; ctxt->wc_status = wc.status; ctxt->byte_len = wc.byte_len; if (wc.status != IB_WC_SUCCESS) { /* Close the transport */ + dprintk("svcrdma: transport closing putting ctxt %p\n", ctxt); set_bit(XPT_CLOSE, &xprt->sc_xprt.xpt_flags); svc_rdma_put_context(ctxt, 1); + svc_xprt_put(&xprt->sc_xprt); continue; } + spin_lock_bh(&xprt->sc_rq_dto_lock); list_add_tail(&ctxt->dto_q, &xprt->sc_rq_dto_q); + spin_unlock_bh(&xprt->sc_rq_dto_lock); + svc_xprt_put(&xprt->sc_xprt); } - spin_unlock_bh(&xprt->sc_rq_dto_lock); if (ctxt) atomic_inc(&rdma_stat_rq_prod); + + set_bit(XPT_DATA, &xprt->sc_xprt.xpt_flags); + /* + * If data arrived before established event, + * don't enqueue. This defers RPC I/O until the + * RDMA connection is complete. + */ + if (!test_bit(RDMAXPRT_CONN_PENDING, &xprt->sc_flags)) + svc_xprt_enqueue(&xprt->sc_xprt); } /* * Send Queue Completion Handler - potentially called on interrupt context. + * + * Note that caller must hold a transport reference. */ static void sq_cq_reap(struct svcxprt_rdma *xprt) { @@ -328,6 +341,11 @@ static void sq_cq_reap(struct svcxprt_rdma *xprt) struct ib_cq *cq = xprt->sc_sq_cq; int ret; + + if (!test_and_clear_bit(RDMAXPRT_SQ_PENDING, &xprt->sc_flags)) + return; + + ib_req_notify_cq(xprt->sc_sq_cq, IB_CQ_NEXT_COMP); atomic_inc(&rdma_stat_sq_poll); while ((ret = ib_poll_cq(cq, 1, &wc)) > 0) { ctxt = (struct svc_rdma_op_ctxt *)(unsigned long)wc.wr_id; @@ -349,14 +367,16 @@ static void sq_cq_reap(struct svcxprt_rdma *xprt) case IB_WR_RDMA_READ: if (test_bit(RDMACTXT_F_LAST_CTXT, &ctxt->flags)) { + struct svc_rdma_op_ctxt *read_hdr = ctxt->read_hdr; + BUG_ON(!read_hdr); set_bit(XPT_DATA, &xprt->sc_xprt.xpt_flags); - set_bit(RDMACTXT_F_READ_DONE, &ctxt->flags); spin_lock_bh(&xprt->sc_read_complete_lock); - list_add_tail(&ctxt->dto_q, + list_add_tail(&read_hdr->dto_q, &xprt->sc_read_complete_q); spin_unlock_bh(&xprt->sc_read_complete_lock); svc_xprt_enqueue(&xprt->sc_xprt); } + svc_rdma_put_context(ctxt, 0); break; default: @@ -365,6 +385,7 @@ static void sq_cq_reap(struct svcxprt_rdma *xprt) wc.opcode, wc.status); break; } + svc_xprt_put(&xprt->sc_xprt); } if (ctxt) @@ -376,11 +397,15 @@ static void sq_comp_handler(struct ib_cq *cq, void *cq_context) struct svcxprt_rdma *xprt = cq_context; unsigned long flags; + /* Guard against unconditional flush call for destroyed QP */ + if (atomic_read(&xprt->sc_xprt.xpt_ref.refcount)==0) + return; + /* * Set the bit regardless of whether or not it's on the list * because it may be on the list already due to an RQ * completion. - */ + */ set_bit(RDMAXPRT_SQ_PENDING, &xprt->sc_flags); /* @@ -407,28 +432,29 @@ static void create_context_cache(struct svcxprt_rdma *xprt, xprt->sc_ctxt_max = ctxt_max; xprt->sc_ctxt_bump = ctxt_bump; xprt->sc_ctxt_cnt = 0; - xprt->sc_ctxt_head = NULL; + atomic_set(&xprt->sc_ctxt_used, 0); + + INIT_LIST_HEAD(&xprt->sc_ctxt_free); for (i = 0; i < ctxt_count; i++) { ctxt = kmalloc(sizeof(*ctxt), GFP_KERNEL); if (ctxt) { - ctxt->next = xprt->sc_ctxt_head; - xprt->sc_ctxt_head = ctxt; + INIT_LIST_HEAD(&ctxt->free_list); + list_add(&ctxt->free_list, &xprt->sc_ctxt_free); xprt->sc_ctxt_cnt++; } } } -static void destroy_context_cache(struct svc_rdma_op_ctxt *ctxt) +static void destroy_context_cache(struct svcxprt_rdma *xprt) { - struct svc_rdma_op_ctxt *next; - if (!ctxt) - return; - - do { - next = ctxt->next; + while (!list_empty(&xprt->sc_ctxt_free)) { + struct svc_rdma_op_ctxt *ctxt; + ctxt = list_entry(xprt->sc_ctxt_free.next, + struct svc_rdma_op_ctxt, + free_list); + list_del_init(&ctxt->free_list); kfree(ctxt); - ctxt = next; - } while (next); + } } static struct svcxprt_rdma *rdma_create_xprt(struct svc_serv *serv, @@ -465,7 +491,7 @@ static struct svcxprt_rdma *rdma_create_xprt(struct svc_serv *serv, reqs + cma_xprt->sc_sq_depth + RPCRDMA_MAX_THREADS + 1); /* max */ - if (!cma_xprt->sc_ctxt_head) { + if (list_empty(&cma_xprt->sc_ctxt_free)) { kfree(cma_xprt); return NULL; } @@ -520,7 +546,12 @@ int svc_rdma_post_recv(struct svcxprt_rdma *xprt) recv_wr.num_sge = ctxt->count; recv_wr.wr_id = (u64)(unsigned long)ctxt; + svc_xprt_get(&xprt->sc_xprt); ret = ib_post_recv(xprt->sc_qp, &recv_wr, &bad_recv_wr); + if (ret) { + svc_xprt_put(&xprt->sc_xprt); + svc_rdma_put_context(ctxt, 1); + } return ret; } @@ -539,6 +570,7 @@ static void handle_connect_req(struct rdma_cm_id *new_cma_id) { struct svcxprt_rdma *listen_xprt = new_cma_id->context; struct svcxprt_rdma *newxprt; + struct sockaddr *sa; /* Create a new transport */ newxprt = rdma_create_xprt(listen_xprt->sc_xprt.xpt_server, 0); @@ -551,6 +583,12 @@ static void handle_connect_req(struct rdma_cm_id *new_cma_id) dprintk("svcrdma: Creating newxprt=%p, cm_id=%p, listenxprt=%p\n", newxprt, newxprt->sc_cm_id, listen_xprt); + /* Set the local and remote addresses in the transport */ + sa = (struct sockaddr *)&newxprt->sc_cm_id->route.addr.dst_addr; + svc_xprt_set_remote(&newxprt->sc_xprt, sa, svc_addr_len(sa)); + sa = (struct sockaddr *)&newxprt->sc_cm_id->route.addr.src_addr; + svc_xprt_set_local(&newxprt->sc_xprt, sa, svc_addr_len(sa)); + /* * Enqueue the new transport on the accept queue of the listening * transport @@ -627,6 +665,7 @@ static int rdma_cma_handler(struct rdma_cm_id *cma_id, if (xprt) { set_bit(XPT_CLOSE, &xprt->xpt_flags); svc_xprt_enqueue(xprt); + svc_xprt_put(xprt); } break; case RDMA_CM_EVENT_DEVICE_REMOVAL: @@ -661,31 +700,27 @@ static struct svc_xprt *svc_rdma_create(struct svc_serv *serv, cma_xprt = rdma_create_xprt(serv, 1); if (!cma_xprt) - return ERR_PTR(ENOMEM); + return ERR_PTR(-ENOMEM); xprt = &cma_xprt->sc_xprt; listen_id = rdma_create_id(rdma_listen_handler, cma_xprt, RDMA_PS_TCP); if (IS_ERR(listen_id)) { - svc_xprt_put(&cma_xprt->sc_xprt); - dprintk("svcrdma: rdma_create_id failed = %ld\n", - PTR_ERR(listen_id)); - return (void *)listen_id; + ret = PTR_ERR(listen_id); + dprintk("svcrdma: rdma_create_id failed = %d\n", ret); + goto err0; } + ret = rdma_bind_addr(listen_id, sa); if (ret) { - rdma_destroy_id(listen_id); - svc_xprt_put(&cma_xprt->sc_xprt); dprintk("svcrdma: rdma_bind_addr failed = %d\n", ret); - return ERR_PTR(ret); + goto err1; } cma_xprt->sc_cm_id = listen_id; ret = rdma_listen(listen_id, RPCRDMA_LISTEN_BACKLOG); if (ret) { - rdma_destroy_id(listen_id); - svc_xprt_put(&cma_xprt->sc_xprt); dprintk("svcrdma: rdma_listen failed = %d\n", ret); - return ERR_PTR(ret); + goto err1; } /* @@ -696,6 +731,12 @@ static struct svc_xprt *svc_rdma_create(struct svc_serv *serv, svc_xprt_set_local(&cma_xprt->sc_xprt, sa, salen); return &cma_xprt->sc_xprt; + + err1: + rdma_destroy_id(listen_id); + err0: + kfree(cma_xprt); + return ERR_PTR(ret); } /* @@ -716,7 +757,6 @@ static struct svc_xprt *svc_rdma_accept(struct svc_xprt *xprt) struct rdma_conn_param conn_param; struct ib_qp_init_attr qp_attr; struct ib_device_attr devattr; - struct sockaddr *sa; int ret; int i; @@ -826,7 +866,6 @@ static struct svc_xprt *svc_rdma_accept(struct svc_xprt *xprt) newxprt->sc_sq_depth = qp_attr.cap.max_send_wr; newxprt->sc_max_requests = qp_attr.cap.max_recv_wr; } - svc_xprt_get(&newxprt->sc_xprt); newxprt->sc_qp = newxprt->sc_cm_id->qp; /* Register all of physical memory */ @@ -850,6 +889,13 @@ static struct svc_xprt *svc_rdma_accept(struct svc_xprt *xprt) /* Swap out the handler */ newxprt->sc_cm_id->event_handler = rdma_cma_handler; + /* + * Arm the CQs for the SQ and RQ before accepting so we can't + * miss the first message + */ + ib_req_notify_cq(newxprt->sc_sq_cq, IB_CQ_NEXT_COMP); + ib_req_notify_cq(newxprt->sc_rq_cq, IB_CQ_NEXT_COMP); + /* Accept Connection */ set_bit(RDMAXPRT_CONN_PENDING, &newxprt->sc_flags); memset(&conn_param, 0, sizeof conn_param); @@ -886,58 +932,26 @@ static struct svc_xprt *svc_rdma_accept(struct svc_xprt *xprt) newxprt->sc_max_requests, newxprt->sc_ord); - /* Set the local and remote addresses in the transport */ - sa = (struct sockaddr *)&newxprt->sc_cm_id->route.addr.dst_addr; - svc_xprt_set_remote(&newxprt->sc_xprt, sa, svc_addr_len(sa)); - sa = (struct sockaddr *)&newxprt->sc_cm_id->route.addr.src_addr; - svc_xprt_set_local(&newxprt->sc_xprt, sa, svc_addr_len(sa)); - - ib_req_notify_cq(newxprt->sc_sq_cq, IB_CQ_NEXT_COMP); - ib_req_notify_cq(newxprt->sc_rq_cq, IB_CQ_NEXT_COMP); return &newxprt->sc_xprt; errout: dprintk("svcrdma: failure accepting new connection rc=%d.\n", ret); /* Take a reference in case the DTO handler runs */ svc_xprt_get(&newxprt->sc_xprt); - if (newxprt->sc_qp && !IS_ERR(newxprt->sc_qp)) { + if (newxprt->sc_qp && !IS_ERR(newxprt->sc_qp)) ib_destroy_qp(newxprt->sc_qp); - svc_xprt_put(&newxprt->sc_xprt); - } rdma_destroy_id(newxprt->sc_cm_id); /* This call to put will destroy the transport */ svc_xprt_put(&newxprt->sc_xprt); return NULL; } -/* - * Post an RQ WQE to the RQ when the rqst is being released. This - * effectively returns an RQ credit to the client. The rq_xprt_ctxt - * will be null if the request is deferred due to an RDMA_READ or the - * transport had no data ready (EAGAIN). Note that an RPC deferred in - * svc_process will still return the credit, this is because the data - * is copied and no longer consume a WQE/WC. - */ static void svc_rdma_release_rqst(struct svc_rqst *rqstp) { - int err; - struct svcxprt_rdma *rdma = - container_of(rqstp->rq_xprt, struct svcxprt_rdma, sc_xprt); - if (rqstp->rq_xprt_ctxt) { - BUG_ON(rqstp->rq_xprt_ctxt != rdma); - err = svc_rdma_post_recv(rdma); - if (err) - dprintk("svcrdma: failed to post an RQ WQE error=%d\n", - err); - } - rqstp->rq_xprt_ctxt = NULL; } /* - * When connected, an svc_xprt has at least three references: - * - * - A reference held by the QP. We still hold that here because this - * code deletes the QP and puts the reference. + * When connected, an svc_xprt has at least two references: * * - A reference held by the cm_id between the ESTABLISHED and * DISCONNECTED events. If the remote peer disconnected first, this @@ -946,7 +960,7 @@ static void svc_rdma_release_rqst(struct svc_rqst *rqstp) * - A reference held by the svc_recv code that called this function * as part of close processing. * - * At a minimum two references should still be held. + * At a minimum one references should still be held. */ static void svc_rdma_detach(struct svc_xprt *xprt) { @@ -956,23 +970,53 @@ static void svc_rdma_detach(struct svc_xprt *xprt) /* Disconnect and flush posted WQE */ rdma_disconnect(rdma->sc_cm_id); - - /* Destroy the QP if present (not a listener) */ - if (rdma->sc_qp && !IS_ERR(rdma->sc_qp)) { - ib_destroy_qp(rdma->sc_qp); - svc_xprt_put(xprt); - } - - /* Destroy the CM ID */ - rdma_destroy_id(rdma->sc_cm_id); } -static void svc_rdma_free(struct svc_xprt *xprt) +static void __svc_rdma_free(struct work_struct *work) { - struct svcxprt_rdma *rdma = (struct svcxprt_rdma *)xprt; + struct svcxprt_rdma *rdma = + container_of(work, struct svcxprt_rdma, sc_work); dprintk("svcrdma: svc_rdma_free(%p)\n", rdma); + /* We should only be called from kref_put */ - BUG_ON(atomic_read(&xprt->xpt_ref.refcount) != 0); + BUG_ON(atomic_read(&rdma->sc_xprt.xpt_ref.refcount) != 0); + + /* + * Destroy queued, but not processed read completions. Note + * that this cleanup has to be done before destroying the + * cm_id because the device ptr is needed to unmap the dma in + * svc_rdma_put_context. + */ + spin_lock_bh(&rdma->sc_read_complete_lock); + while (!list_empty(&rdma->sc_read_complete_q)) { + struct svc_rdma_op_ctxt *ctxt; + ctxt = list_entry(rdma->sc_read_complete_q.next, + struct svc_rdma_op_ctxt, + dto_q); + list_del_init(&ctxt->dto_q); + svc_rdma_put_context(ctxt, 1); + } + spin_unlock_bh(&rdma->sc_read_complete_lock); + + /* Destroy queued, but not processed recv completions */ + spin_lock_bh(&rdma->sc_rq_dto_lock); + while (!list_empty(&rdma->sc_rq_dto_q)) { + struct svc_rdma_op_ctxt *ctxt; + ctxt = list_entry(rdma->sc_rq_dto_q.next, + struct svc_rdma_op_ctxt, + dto_q); + list_del_init(&ctxt->dto_q); + svc_rdma_put_context(ctxt, 1); + } + spin_unlock_bh(&rdma->sc_rq_dto_lock); + + /* Warn if we leaked a resource or under-referenced */ + WARN_ON(atomic_read(&rdma->sc_ctxt_used) != 0); + + /* Destroy the QP if present (not a listener) */ + if (rdma->sc_qp && !IS_ERR(rdma->sc_qp)) + ib_destroy_qp(rdma->sc_qp); + if (rdma->sc_sq_cq && !IS_ERR(rdma->sc_sq_cq)) ib_destroy_cq(rdma->sc_sq_cq); @@ -985,10 +1029,21 @@ static void svc_rdma_free(struct svc_xprt *xprt) if (rdma->sc_pd && !IS_ERR(rdma->sc_pd)) ib_dealloc_pd(rdma->sc_pd); - destroy_context_cache(rdma->sc_ctxt_head); + /* Destroy the CM ID */ + rdma_destroy_id(rdma->sc_cm_id); + + destroy_context_cache(rdma); kfree(rdma); } +static void svc_rdma_free(struct svc_xprt *xprt) +{ + struct svcxprt_rdma *rdma = + container_of(xprt, struct svcxprt_rdma, sc_xprt); + INIT_WORK(&rdma->sc_work, __svc_rdma_free); + schedule_work(&rdma->sc_work); +} + static int svc_rdma_has_wspace(struct svc_xprt *xprt) { struct svcxprt_rdma *rdma = @@ -1018,7 +1073,7 @@ int svc_rdma_send(struct svcxprt_rdma *xprt, struct ib_send_wr *wr) int ret; if (test_bit(XPT_CLOSE, &xprt->sc_xprt.xpt_flags)) - return 0; + return -ENOTCONN; BUG_ON(wr->send_flags != IB_SEND_SIGNALED); BUG_ON(((struct svc_rdma_op_ctxt *)(unsigned long)wr->wr_id)->wr_op != @@ -1029,7 +1084,8 @@ int svc_rdma_send(struct svcxprt_rdma *xprt, struct ib_send_wr *wr) if (xprt->sc_sq_depth == atomic_read(&xprt->sc_sq_count)) { spin_unlock_bh(&xprt->sc_lock); atomic_inc(&rdma_stat_sq_starve); - /* See if we can reap some SQ WR */ + + /* See if we can opportunistically reap SQ WR to make room */ sq_cq_reap(xprt); /* Wait until SQ WR available if SQ still full */ @@ -1041,22 +1097,25 @@ int svc_rdma_send(struct svcxprt_rdma *xprt, struct ib_send_wr *wr) continue; } /* Bumped used SQ WR count and post */ + svc_xprt_get(&xprt->sc_xprt); ret = ib_post_send(xprt->sc_qp, wr, &bad_wr); if (!ret) atomic_inc(&xprt->sc_sq_count); - else + else { + svc_xprt_put(&xprt->sc_xprt); dprintk("svcrdma: failed to post SQ WR rc=%d, " "sc_sq_count=%d, sc_sq_depth=%d\n", ret, atomic_read(&xprt->sc_sq_count), xprt->sc_sq_depth); + } spin_unlock_bh(&xprt->sc_lock); break; } return ret; } -int svc_rdma_send_error(struct svcxprt_rdma *xprt, struct rpcrdma_msg *rmsgp, - enum rpcrdma_errcode err) +void svc_rdma_send_error(struct svcxprt_rdma *xprt, struct rpcrdma_msg *rmsgp, + enum rpcrdma_errcode err) { struct ib_send_wr err_wr; struct ib_sge sge; @@ -1094,9 +1153,8 @@ int svc_rdma_send_error(struct svcxprt_rdma *xprt, struct rpcrdma_msg *rmsgp, /* Post It */ ret = svc_rdma_send(xprt, &err_wr); if (ret) { - dprintk("svcrdma: Error posting send = %d\n", ret); + dprintk("svcrdma: Error %d posting send for protocol error\n", + ret); svc_rdma_put_context(ctxt, 1); } - - return ret; } diff --git a/net/xfrm/xfrm_output.c b/net/xfrm/xfrm_output.c index 09cd9c0c2d8..3f964db908a 100644 --- a/net/xfrm/xfrm_output.c +++ b/net/xfrm/xfrm_output.c @@ -25,11 +25,11 @@ static int xfrm_state_check_space(struct xfrm_state *x, struct sk_buff *skb) struct dst_entry *dst = skb->dst; int nhead = dst->header_len + LL_RESERVED_SPACE(dst->dev) - skb_headroom(skb); + int ntail = dst->dev->needed_tailroom - skb_tailroom(skb); - if (nhead > 0) - return pskb_expand_head(skb, nhead, 0, GFP_ATOMIC); + if (nhead > 0 || ntail > 0) + return pskb_expand_head(skb, nhead, ntail, GFP_ATOMIC); - /* Check tail too... */ return 0; } |