/*
|
/*
|
* NET4: Implementation of BSD Unix domain sockets.
|
* NET4: Implementation of BSD Unix domain sockets.
|
*
|
*
|
* Authors: Alan Cox, <alan.cox@linux.org>
|
* Authors: Alan Cox, <alan.cox@linux.org>
|
*
|
*
|
* This program is free software; you can redistribute it and/or
|
* This program is free software; you can redistribute it and/or
|
* modify it under the terms of the GNU General Public License
|
* modify it under the terms of the GNU General Public License
|
* as published by the Free Software Foundation; either version
|
* as published by the Free Software Foundation; either version
|
* 2 of the License, or (at your option) any later version.
|
* 2 of the License, or (at your option) any later version.
|
*
|
*
|
* Version: $Id: af_unix.c,v 1.1.1.1 2004-04-15 01:16:03 phoenix Exp $
|
* Version: $Id: af_unix.c,v 1.1.1.1 2004-04-15 01:16:03 phoenix Exp $
|
*
|
*
|
* Fixes:
|
* Fixes:
|
* Linus Torvalds : Assorted bug cures.
|
* Linus Torvalds : Assorted bug cures.
|
* Niibe Yutaka : async I/O support.
|
* Niibe Yutaka : async I/O support.
|
* Carsten Paeth : PF_UNIX check, address fixes.
|
* Carsten Paeth : PF_UNIX check, address fixes.
|
* Alan Cox : Limit size of allocated blocks.
|
* Alan Cox : Limit size of allocated blocks.
|
* Alan Cox : Fixed the stupid socketpair bug.
|
* Alan Cox : Fixed the stupid socketpair bug.
|
* Alan Cox : BSD compatibility fine tuning.
|
* Alan Cox : BSD compatibility fine tuning.
|
* Alan Cox : Fixed a bug in connect when interrupted.
|
* Alan Cox : Fixed a bug in connect when interrupted.
|
* Alan Cox : Sorted out a proper draft version of
|
* Alan Cox : Sorted out a proper draft version of
|
* file descriptor passing hacked up from
|
* file descriptor passing hacked up from
|
* Mike Shaver's work.
|
* Mike Shaver's work.
|
* Marty Leisner : Fixes to fd passing
|
* Marty Leisner : Fixes to fd passing
|
* Nick Nevin : recvmsg bugfix.
|
* Nick Nevin : recvmsg bugfix.
|
* Alan Cox : Started proper garbage collector
|
* Alan Cox : Started proper garbage collector
|
* Heiko EiBfeldt : Missing verify_area check
|
* Heiko EiBfeldt : Missing verify_area check
|
* Alan Cox : Started POSIXisms
|
* Alan Cox : Started POSIXisms
|
* Andreas Schwab : Replace inode by dentry for proper
|
* Andreas Schwab : Replace inode by dentry for proper
|
* reference counting
|
* reference counting
|
* Kirk Petersen : Made this a module
|
* Kirk Petersen : Made this a module
|
* Christoph Rohland : Elegant non-blocking accept/connect algorithm.
|
* Christoph Rohland : Elegant non-blocking accept/connect algorithm.
|
* Lots of bug fixes.
|
* Lots of bug fixes.
|
* Alexey Kuznetosv : Repaired (I hope) bugs introduces
|
* Alexey Kuznetosv : Repaired (I hope) bugs introduces
|
* by above two patches.
|
* by above two patches.
|
* Andrea Arcangeli : If possible we block in connect(2)
|
* Andrea Arcangeli : If possible we block in connect(2)
|
* if the max backlog of the listen socket
|
* if the max backlog of the listen socket
|
* is been reached. This won't break
|
* is been reached. This won't break
|
* old apps and it will avoid huge amount
|
* old apps and it will avoid huge amount
|
* of socks hashed (this for unix_gc()
|
* of socks hashed (this for unix_gc()
|
* performances reasons).
|
* performances reasons).
|
* Security fix that limits the max
|
* Security fix that limits the max
|
* number of socks to 2*max_files and
|
* number of socks to 2*max_files and
|
* the number of skb queueable in the
|
* the number of skb queueable in the
|
* dgram receiver.
|
* dgram receiver.
|
* Artur Skawina : Hash function optimizations
|
* Artur Skawina : Hash function optimizations
|
* Alexey Kuznetsov : Full scale SMP. Lot of bugs are introduced 8)
|
* Alexey Kuznetsov : Full scale SMP. Lot of bugs are introduced 8)
|
* Malcolm Beattie : Set peercred for socketpair
|
* Malcolm Beattie : Set peercred for socketpair
|
* Michal Ostrowski : Module initialization cleanup.
|
* Michal Ostrowski : Module initialization cleanup.
|
*
|
*
|
*
|
*
|
* Known differences from reference BSD that was tested:
|
* Known differences from reference BSD that was tested:
|
*
|
*
|
* [TO FIX]
|
* [TO FIX]
|
* ECONNREFUSED is not returned from one end of a connected() socket to the
|
* ECONNREFUSED is not returned from one end of a connected() socket to the
|
* other the moment one end closes.
|
* other the moment one end closes.
|
* fstat() doesn't return st_dev=NODEV, and give the blksize as high water mark
|
* fstat() doesn't return st_dev=NODEV, and give the blksize as high water mark
|
* and a fake inode identifier (nor the BSD first socket fstat twice bug).
|
* and a fake inode identifier (nor the BSD first socket fstat twice bug).
|
* [NOT TO FIX]
|
* [NOT TO FIX]
|
* accept() returns a path name even if the connecting socket has closed
|
* accept() returns a path name even if the connecting socket has closed
|
* in the meantime (BSD loses the path and gives up).
|
* in the meantime (BSD loses the path and gives up).
|
* accept() returns 0 length path for an unbound connector. BSD returns 16
|
* accept() returns 0 length path for an unbound connector. BSD returns 16
|
* and a null first byte in the path (but not for gethost/peername - BSD bug ??)
|
* and a null first byte in the path (but not for gethost/peername - BSD bug ??)
|
* socketpair(...SOCK_RAW..) doesn't panic the kernel.
|
* socketpair(...SOCK_RAW..) doesn't panic the kernel.
|
* BSD af_unix apparently has connect forgetting to block properly.
|
* BSD af_unix apparently has connect forgetting to block properly.
|
* (need to check this with the POSIX spec in detail)
|
* (need to check this with the POSIX spec in detail)
|
*
|
*
|
* Differences from 2.0.0-11-... (ANK)
|
* Differences from 2.0.0-11-... (ANK)
|
* Bug fixes and improvements.
|
* Bug fixes and improvements.
|
* - client shutdown killed server socket.
|
* - client shutdown killed server socket.
|
* - removed all useless cli/sti pairs.
|
* - removed all useless cli/sti pairs.
|
*
|
*
|
* Semantic changes/extensions.
|
* Semantic changes/extensions.
|
* - generic control message passing.
|
* - generic control message passing.
|
* - SCM_CREDENTIALS control message.
|
* - SCM_CREDENTIALS control message.
|
* - "Abstract" (not FS based) socket bindings.
|
* - "Abstract" (not FS based) socket bindings.
|
* Abstract names are sequences of bytes (not zero terminated)
|
* Abstract names are sequences of bytes (not zero terminated)
|
* started by 0, so that this name space does not intersect
|
* started by 0, so that this name space does not intersect
|
* with BSD names.
|
* with BSD names.
|
*/
|
*/
|
|
|
#include <linux/module.h>
|
#include <linux/module.h>
|
#include <linux/config.h>
|
#include <linux/config.h>
|
#include <linux/kernel.h>
|
#include <linux/kernel.h>
|
#include <linux/major.h>
|
#include <linux/major.h>
|
#include <linux/signal.h>
|
#include <linux/signal.h>
|
#include <linux/sched.h>
|
#include <linux/sched.h>
|
#include <linux/errno.h>
|
#include <linux/errno.h>
|
#include <linux/string.h>
|
#include <linux/string.h>
|
#include <linux/stat.h>
|
#include <linux/stat.h>
|
#include <linux/socket.h>
|
#include <linux/socket.h>
|
#include <linux/un.h>
|
#include <linux/un.h>
|
#include <linux/fcntl.h>
|
#include <linux/fcntl.h>
|
#include <linux/termios.h>
|
#include <linux/termios.h>
|
#include <linux/sockios.h>
|
#include <linux/sockios.h>
|
#include <linux/net.h>
|
#include <linux/net.h>
|
#include <linux/in.h>
|
#include <linux/in.h>
|
#include <linux/fs.h>
|
#include <linux/fs.h>
|
#include <linux/slab.h>
|
#include <linux/slab.h>
|
#include <asm/uaccess.h>
|
#include <asm/uaccess.h>
|
#include <linux/skbuff.h>
|
#include <linux/skbuff.h>
|
#include <linux/netdevice.h>
|
#include <linux/netdevice.h>
|
#include <net/sock.h>
|
#include <net/sock.h>
|
#include <linux/tcp.h>
|
#include <linux/tcp.h>
|
#include <net/af_unix.h>
|
#include <net/af_unix.h>
|
#include <linux/proc_fs.h>
|
#include <linux/proc_fs.h>
|
#include <net/scm.h>
|
#include <net/scm.h>
|
#include <linux/init.h>
|
#include <linux/init.h>
|
#include <linux/poll.h>
|
#include <linux/poll.h>
|
#include <linux/smp_lock.h>
|
#include <linux/smp_lock.h>
|
#include <linux/rtnetlink.h>
|
#include <linux/rtnetlink.h>
|
|
|
#include <asm/checksum.h>
|
#include <asm/checksum.h>
|
|
|
int sysctl_unix_max_dgram_qlen = 10;
|
int sysctl_unix_max_dgram_qlen = 10;
|
|
|
unix_socket *unix_socket_table[UNIX_HASH_SIZE+1];
|
unix_socket *unix_socket_table[UNIX_HASH_SIZE+1];
|
rwlock_t unix_table_lock = RW_LOCK_UNLOCKED;
|
rwlock_t unix_table_lock = RW_LOCK_UNLOCKED;
|
static atomic_t unix_nr_socks = ATOMIC_INIT(0);
|
static atomic_t unix_nr_socks = ATOMIC_INIT(0);
|
|
|
#define unix_sockets_unbound (unix_socket_table[UNIX_HASH_SIZE])
|
#define unix_sockets_unbound (unix_socket_table[UNIX_HASH_SIZE])
|
|
|
#define UNIX_ABSTRACT(sk) ((sk)->protinfo.af_unix.addr->hash!=UNIX_HASH_SIZE)
|
#define UNIX_ABSTRACT(sk) ((sk)->protinfo.af_unix.addr->hash!=UNIX_HASH_SIZE)
|
|
|
/*
|
/*
|
* SMP locking strategy:
|
* SMP locking strategy:
|
* hash table is protected with rwlock unix_table_lock
|
* hash table is protected with rwlock unix_table_lock
|
* each socket state is protected by separate rwlock.
|
* each socket state is protected by separate rwlock.
|
*/
|
*/
|
|
|
static inline unsigned unix_hash_fold(unsigned hash)
|
static inline unsigned unix_hash_fold(unsigned hash)
|
{
|
{
|
hash ^= hash>>16;
|
hash ^= hash>>16;
|
hash ^= hash>>8;
|
hash ^= hash>>8;
|
return hash&(UNIX_HASH_SIZE-1);
|
return hash&(UNIX_HASH_SIZE-1);
|
}
|
}
|
|
|
#define unix_peer(sk) ((sk)->pair)
|
#define unix_peer(sk) ((sk)->pair)
|
|
|
static inline int unix_our_peer(unix_socket *sk, unix_socket *osk)
|
static inline int unix_our_peer(unix_socket *sk, unix_socket *osk)
|
{
|
{
|
return unix_peer(osk) == sk;
|
return unix_peer(osk) == sk;
|
}
|
}
|
|
|
static inline int unix_may_send(unix_socket *sk, unix_socket *osk)
|
static inline int unix_may_send(unix_socket *sk, unix_socket *osk)
|
{
|
{
|
return (unix_peer(osk) == NULL || unix_our_peer(sk, osk));
|
return (unix_peer(osk) == NULL || unix_our_peer(sk, osk));
|
}
|
}
|
|
|
static inline unix_socket * unix_peer_get(unix_socket *s)
|
static inline unix_socket * unix_peer_get(unix_socket *s)
|
{
|
{
|
unix_socket *peer;
|
unix_socket *peer;
|
|
|
unix_state_rlock(s);
|
unix_state_rlock(s);
|
peer = unix_peer(s);
|
peer = unix_peer(s);
|
if (peer)
|
if (peer)
|
sock_hold(peer);
|
sock_hold(peer);
|
unix_state_runlock(s);
|
unix_state_runlock(s);
|
return peer;
|
return peer;
|
}
|
}
|
|
|
extern inline void unix_release_addr(struct unix_address *addr)
|
extern inline void unix_release_addr(struct unix_address *addr)
|
{
|
{
|
if (atomic_dec_and_test(&addr->refcnt))
|
if (atomic_dec_and_test(&addr->refcnt))
|
kfree(addr);
|
kfree(addr);
|
}
|
}
|
|
|
/*
|
/*
|
* Check unix socket name:
|
* Check unix socket name:
|
* - should be not zero length.
|
* - should be not zero length.
|
* - if started by not zero, should be NULL terminated (FS object)
|
* - if started by not zero, should be NULL terminated (FS object)
|
* - if started by zero, it is abstract name.
|
* - if started by zero, it is abstract name.
|
*/
|
*/
|
|
|
static int unix_mkname(struct sockaddr_un * sunaddr, int len, unsigned *hashp)
|
static int unix_mkname(struct sockaddr_un * sunaddr, int len, unsigned *hashp)
|
{
|
{
|
if (len <= sizeof(short) || len > sizeof(*sunaddr))
|
if (len <= sizeof(short) || len > sizeof(*sunaddr))
|
return -EINVAL;
|
return -EINVAL;
|
if (!sunaddr || sunaddr->sun_family != AF_UNIX)
|
if (!sunaddr || sunaddr->sun_family != AF_UNIX)
|
return -EINVAL;
|
return -EINVAL;
|
if (sunaddr->sun_path[0])
|
if (sunaddr->sun_path[0])
|
{
|
{
|
/*
|
/*
|
* This may look like an off by one error but it is
|
* This may look like an off by one error but it is
|
* a bit more subtle. 108 is the longest valid AF_UNIX
|
* a bit more subtle. 108 is the longest valid AF_UNIX
|
* path for a binding. sun_path[108] doesn't as such
|
* path for a binding. sun_path[108] doesn't as such
|
* exist. However in kernel space we are guaranteed that
|
* exist. However in kernel space we are guaranteed that
|
* it is a valid memory location in our kernel
|
* it is a valid memory location in our kernel
|
* address buffer.
|
* address buffer.
|
*/
|
*/
|
if (len > sizeof(*sunaddr))
|
if (len > sizeof(*sunaddr))
|
len = sizeof(*sunaddr);
|
len = sizeof(*sunaddr);
|
((char *)sunaddr)[len]=0;
|
((char *)sunaddr)[len]=0;
|
len = strlen(sunaddr->sun_path)+1+sizeof(short);
|
len = strlen(sunaddr->sun_path)+1+sizeof(short);
|
return len;
|
return len;
|
}
|
}
|
|
|
*hashp = unix_hash_fold(csum_partial((char*)sunaddr, len, 0));
|
*hashp = unix_hash_fold(csum_partial((char*)sunaddr, len, 0));
|
return len;
|
return len;
|
}
|
}
|
|
|
static void __unix_remove_socket(unix_socket *sk)
|
static void __unix_remove_socket(unix_socket *sk)
|
{
|
{
|
unix_socket **list = sk->protinfo.af_unix.list;
|
unix_socket **list = sk->protinfo.af_unix.list;
|
if (list) {
|
if (list) {
|
if (sk->next)
|
if (sk->next)
|
sk->next->prev = sk->prev;
|
sk->next->prev = sk->prev;
|
if (sk->prev)
|
if (sk->prev)
|
sk->prev->next = sk->next;
|
sk->prev->next = sk->next;
|
if (*list == sk)
|
if (*list == sk)
|
*list = sk->next;
|
*list = sk->next;
|
sk->protinfo.af_unix.list = NULL;
|
sk->protinfo.af_unix.list = NULL;
|
sk->prev = NULL;
|
sk->prev = NULL;
|
sk->next = NULL;
|
sk->next = NULL;
|
__sock_put(sk);
|
__sock_put(sk);
|
}
|
}
|
}
|
}
|
|
|
static void __unix_insert_socket(unix_socket **list, unix_socket *sk)
|
static void __unix_insert_socket(unix_socket **list, unix_socket *sk)
|
{
|
{
|
BUG_TRAP(sk->protinfo.af_unix.list==NULL);
|
BUG_TRAP(sk->protinfo.af_unix.list==NULL);
|
|
|
sk->protinfo.af_unix.list = list;
|
sk->protinfo.af_unix.list = list;
|
sk->prev = NULL;
|
sk->prev = NULL;
|
sk->next = *list;
|
sk->next = *list;
|
if (*list)
|
if (*list)
|
(*list)->prev = sk;
|
(*list)->prev = sk;
|
*list=sk;
|
*list=sk;
|
sock_hold(sk);
|
sock_hold(sk);
|
}
|
}
|
|
|
static inline void unix_remove_socket(unix_socket *sk)
|
static inline void unix_remove_socket(unix_socket *sk)
|
{
|
{
|
write_lock(&unix_table_lock);
|
write_lock(&unix_table_lock);
|
__unix_remove_socket(sk);
|
__unix_remove_socket(sk);
|
write_unlock(&unix_table_lock);
|
write_unlock(&unix_table_lock);
|
}
|
}
|
|
|
static inline void unix_insert_socket(unix_socket **list, unix_socket *sk)
|
static inline void unix_insert_socket(unix_socket **list, unix_socket *sk)
|
{
|
{
|
write_lock(&unix_table_lock);
|
write_lock(&unix_table_lock);
|
__unix_insert_socket(list, sk);
|
__unix_insert_socket(list, sk);
|
write_unlock(&unix_table_lock);
|
write_unlock(&unix_table_lock);
|
}
|
}
|
|
|
static unix_socket *__unix_find_socket_byname(struct sockaddr_un *sunname,
|
static unix_socket *__unix_find_socket_byname(struct sockaddr_un *sunname,
|
int len, int type, unsigned hash)
|
int len, int type, unsigned hash)
|
{
|
{
|
unix_socket *s;
|
unix_socket *s;
|
|
|
for (s=unix_socket_table[hash^type]; s; s=s->next) {
|
for (s=unix_socket_table[hash^type]; s; s=s->next) {
|
if(s->protinfo.af_unix.addr->len==len &&
|
if(s->protinfo.af_unix.addr->len==len &&
|
memcmp(s->protinfo.af_unix.addr->name, sunname, len) == 0)
|
memcmp(s->protinfo.af_unix.addr->name, sunname, len) == 0)
|
return s;
|
return s;
|
}
|
}
|
return NULL;
|
return NULL;
|
}
|
}
|
|
|
static inline unix_socket *
|
static inline unix_socket *
|
unix_find_socket_byname(struct sockaddr_un *sunname,
|
unix_find_socket_byname(struct sockaddr_un *sunname,
|
int len, int type, unsigned hash)
|
int len, int type, unsigned hash)
|
{
|
{
|
unix_socket *s;
|
unix_socket *s;
|
|
|
read_lock(&unix_table_lock);
|
read_lock(&unix_table_lock);
|
s = __unix_find_socket_byname(sunname, len, type, hash);
|
s = __unix_find_socket_byname(sunname, len, type, hash);
|
if (s)
|
if (s)
|
sock_hold(s);
|
sock_hold(s);
|
read_unlock(&unix_table_lock);
|
read_unlock(&unix_table_lock);
|
return s;
|
return s;
|
}
|
}
|
|
|
static unix_socket *unix_find_socket_byinode(struct inode *i)
|
static unix_socket *unix_find_socket_byinode(struct inode *i)
|
{
|
{
|
unix_socket *s;
|
unix_socket *s;
|
|
|
read_lock(&unix_table_lock);
|
read_lock(&unix_table_lock);
|
for (s=unix_socket_table[i->i_ino & (UNIX_HASH_SIZE-1)]; s; s=s->next)
|
for (s=unix_socket_table[i->i_ino & (UNIX_HASH_SIZE-1)]; s; s=s->next)
|
{
|
{
|
struct dentry *dentry = s->protinfo.af_unix.dentry;
|
struct dentry *dentry = s->protinfo.af_unix.dentry;
|
|
|
if(dentry && dentry->d_inode == i)
|
if(dentry && dentry->d_inode == i)
|
{
|
{
|
sock_hold(s);
|
sock_hold(s);
|
break;
|
break;
|
}
|
}
|
}
|
}
|
read_unlock(&unix_table_lock);
|
read_unlock(&unix_table_lock);
|
return s;
|
return s;
|
}
|
}
|
|
|
static inline int unix_writable(struct sock *sk)
|
static inline int unix_writable(struct sock *sk)
|
{
|
{
|
return ((atomic_read(&sk->wmem_alloc)<<2) <= sk->sndbuf);
|
return ((atomic_read(&sk->wmem_alloc)<<2) <= sk->sndbuf);
|
}
|
}
|
|
|
static void unix_write_space(struct sock *sk)
|
static void unix_write_space(struct sock *sk)
|
{
|
{
|
read_lock(&sk->callback_lock);
|
read_lock(&sk->callback_lock);
|
if (unix_writable(sk)) {
|
if (unix_writable(sk)) {
|
if (sk->sleep && waitqueue_active(sk->sleep))
|
if (sk->sleep && waitqueue_active(sk->sleep))
|
wake_up_interruptible(sk->sleep);
|
wake_up_interruptible(sk->sleep);
|
sk_wake_async(sk, 2, POLL_OUT);
|
sk_wake_async(sk, 2, POLL_OUT);
|
}
|
}
|
read_unlock(&sk->callback_lock);
|
read_unlock(&sk->callback_lock);
|
}
|
}
|
|
|
/* When dgram socket disconnects (or changes its peer), we clear its receive
|
/* When dgram socket disconnects (or changes its peer), we clear its receive
|
* queue of packets arrived from previous peer. First, it allows to do
|
* queue of packets arrived from previous peer. First, it allows to do
|
* flow control based only on wmem_alloc; second, sk connected to peer
|
* flow control based only on wmem_alloc; second, sk connected to peer
|
* may receive messages only from that peer. */
|
* may receive messages only from that peer. */
|
static void unix_dgram_disconnected(struct sock *sk, struct sock *other)
|
static void unix_dgram_disconnected(struct sock *sk, struct sock *other)
|
{
|
{
|
if (skb_queue_len(&sk->receive_queue)) {
|
if (skb_queue_len(&sk->receive_queue)) {
|
skb_queue_purge(&sk->receive_queue);
|
skb_queue_purge(&sk->receive_queue);
|
wake_up_interruptible_all(&sk->protinfo.af_unix.peer_wait);
|
wake_up_interruptible_all(&sk->protinfo.af_unix.peer_wait);
|
|
|
/* If one link of bidirectional dgram pipe is disconnected,
|
/* If one link of bidirectional dgram pipe is disconnected,
|
* we signal error. Messages are lost. Do not make this,
|
* we signal error. Messages are lost. Do not make this,
|
* when peer was not connected to us.
|
* when peer was not connected to us.
|
*/
|
*/
|
if (!other->dead && unix_peer(other) == sk) {
|
if (!other->dead && unix_peer(other) == sk) {
|
other->err = ECONNRESET;
|
other->err = ECONNRESET;
|
other->error_report(other);
|
other->error_report(other);
|
}
|
}
|
}
|
}
|
}
|
}
|
|
|
static void unix_sock_destructor(struct sock *sk)
|
static void unix_sock_destructor(struct sock *sk)
|
{
|
{
|
skb_queue_purge(&sk->receive_queue);
|
skb_queue_purge(&sk->receive_queue);
|
|
|
BUG_TRAP(atomic_read(&sk->wmem_alloc) == 0);
|
BUG_TRAP(atomic_read(&sk->wmem_alloc) == 0);
|
BUG_TRAP(sk->protinfo.af_unix.list==NULL);
|
BUG_TRAP(sk->protinfo.af_unix.list==NULL);
|
BUG_TRAP(sk->socket==NULL);
|
BUG_TRAP(sk->socket==NULL);
|
if (sk->dead==0) {
|
if (sk->dead==0) {
|
printk("Attempt to release alive unix socket: %p\n", sk);
|
printk("Attempt to release alive unix socket: %p\n", sk);
|
return;
|
return;
|
}
|
}
|
|
|
if (sk->protinfo.af_unix.addr)
|
if (sk->protinfo.af_unix.addr)
|
unix_release_addr(sk->protinfo.af_unix.addr);
|
unix_release_addr(sk->protinfo.af_unix.addr);
|
|
|
atomic_dec(&unix_nr_socks);
|
atomic_dec(&unix_nr_socks);
|
#ifdef UNIX_REFCNT_DEBUG
|
#ifdef UNIX_REFCNT_DEBUG
|
printk(KERN_DEBUG "UNIX %p is destroyed, %d are still alive.\n", sk, atomic_read(&unix_nr_socks));
|
printk(KERN_DEBUG "UNIX %p is destroyed, %d are still alive.\n", sk, atomic_read(&unix_nr_socks));
|
#endif
|
#endif
|
MOD_DEC_USE_COUNT;
|
MOD_DEC_USE_COUNT;
|
}
|
}
|
|
|
static int unix_release_sock (unix_socket *sk, int embrion)
|
static int unix_release_sock (unix_socket *sk, int embrion)
|
{
|
{
|
struct dentry *dentry;
|
struct dentry *dentry;
|
struct vfsmount *mnt;
|
struct vfsmount *mnt;
|
unix_socket *skpair;
|
unix_socket *skpair;
|
struct sk_buff *skb;
|
struct sk_buff *skb;
|
int state;
|
int state;
|
|
|
unix_remove_socket(sk);
|
unix_remove_socket(sk);
|
|
|
/* Clear state */
|
/* Clear state */
|
unix_state_wlock(sk);
|
unix_state_wlock(sk);
|
sock_orphan(sk);
|
sock_orphan(sk);
|
sk->shutdown = SHUTDOWN_MASK;
|
sk->shutdown = SHUTDOWN_MASK;
|
dentry = sk->protinfo.af_unix.dentry;
|
dentry = sk->protinfo.af_unix.dentry;
|
sk->protinfo.af_unix.dentry=NULL;
|
sk->protinfo.af_unix.dentry=NULL;
|
mnt = sk->protinfo.af_unix.mnt;
|
mnt = sk->protinfo.af_unix.mnt;
|
sk->protinfo.af_unix.mnt=NULL;
|
sk->protinfo.af_unix.mnt=NULL;
|
state = sk->state;
|
state = sk->state;
|
sk->state = TCP_CLOSE;
|
sk->state = TCP_CLOSE;
|
unix_state_wunlock(sk);
|
unix_state_wunlock(sk);
|
|
|
wake_up_interruptible_all(&sk->protinfo.af_unix.peer_wait);
|
wake_up_interruptible_all(&sk->protinfo.af_unix.peer_wait);
|
|
|
skpair=unix_peer(sk);
|
skpair=unix_peer(sk);
|
|
|
if (skpair!=NULL) {
|
if (skpair!=NULL) {
|
if (sk->type==SOCK_STREAM) {
|
if (sk->type==SOCK_STREAM) {
|
unix_state_wlock(skpair);
|
unix_state_wlock(skpair);
|
skpair->shutdown=SHUTDOWN_MASK; /* No more writes*/
|
skpair->shutdown=SHUTDOWN_MASK; /* No more writes*/
|
if (!skb_queue_empty(&sk->receive_queue) || embrion)
|
if (!skb_queue_empty(&sk->receive_queue) || embrion)
|
skpair->err = ECONNRESET;
|
skpair->err = ECONNRESET;
|
unix_state_wunlock(skpair);
|
unix_state_wunlock(skpair);
|
skpair->state_change(skpair);
|
skpair->state_change(skpair);
|
read_lock(&skpair->callback_lock);
|
read_lock(&skpair->callback_lock);
|
sk_wake_async(skpair,1,POLL_HUP);
|
sk_wake_async(skpair,1,POLL_HUP);
|
read_unlock(&skpair->callback_lock);
|
read_unlock(&skpair->callback_lock);
|
}
|
}
|
sock_put(skpair); /* It may now die */
|
sock_put(skpair); /* It may now die */
|
unix_peer(sk) = NULL;
|
unix_peer(sk) = NULL;
|
}
|
}
|
|
|
/* Try to flush out this socket. Throw out buffers at least */
|
/* Try to flush out this socket. Throw out buffers at least */
|
|
|
while((skb=skb_dequeue(&sk->receive_queue))!=NULL)
|
while((skb=skb_dequeue(&sk->receive_queue))!=NULL)
|
{
|
{
|
if (state==TCP_LISTEN)
|
if (state==TCP_LISTEN)
|
unix_release_sock(skb->sk, 1);
|
unix_release_sock(skb->sk, 1);
|
/* passed fds are erased in the kfree_skb hook */
|
/* passed fds are erased in the kfree_skb hook */
|
kfree_skb(skb);
|
kfree_skb(skb);
|
}
|
}
|
|
|
if (dentry) {
|
if (dentry) {
|
dput(dentry);
|
dput(dentry);
|
mntput(mnt);
|
mntput(mnt);
|
}
|
}
|
|
|
sock_put(sk);
|
sock_put(sk);
|
|
|
/* ---- Socket is dead now and most probably destroyed ---- */
|
/* ---- Socket is dead now and most probably destroyed ---- */
|
|
|
/*
|
/*
|
* Fixme: BSD difference: In BSD all sockets connected to use get
|
* Fixme: BSD difference: In BSD all sockets connected to use get
|
* ECONNRESET and we die on the spot. In Linux we behave
|
* ECONNRESET and we die on the spot. In Linux we behave
|
* like files and pipes do and wait for the last
|
* like files and pipes do and wait for the last
|
* dereference.
|
* dereference.
|
*
|
*
|
* Can't we simply set sock->err?
|
* Can't we simply set sock->err?
|
*
|
*
|
* What the above comment does talk about? --ANK(980817)
|
* What the above comment does talk about? --ANK(980817)
|
*/
|
*/
|
|
|
if (atomic_read(&unix_tot_inflight))
|
if (atomic_read(&unix_tot_inflight))
|
unix_gc(); /* Garbage collect fds */
|
unix_gc(); /* Garbage collect fds */
|
|
|
return 0;
|
return 0;
|
}
|
}
|
|
|
static int unix_listen(struct socket *sock, int backlog)
|
static int unix_listen(struct socket *sock, int backlog)
|
{
|
{
|
int err;
|
int err;
|
struct sock *sk = sock->sk;
|
struct sock *sk = sock->sk;
|
|
|
err = -EOPNOTSUPP;
|
err = -EOPNOTSUPP;
|
if (sock->type!=SOCK_STREAM)
|
if (sock->type!=SOCK_STREAM)
|
goto out; /* Only stream sockets accept */
|
goto out; /* Only stream sockets accept */
|
err = -EINVAL;
|
err = -EINVAL;
|
if (!sk->protinfo.af_unix.addr)
|
if (!sk->protinfo.af_unix.addr)
|
goto out; /* No listens on an unbound socket */
|
goto out; /* No listens on an unbound socket */
|
unix_state_wlock(sk);
|
unix_state_wlock(sk);
|
if (sk->state != TCP_CLOSE && sk->state != TCP_LISTEN)
|
if (sk->state != TCP_CLOSE && sk->state != TCP_LISTEN)
|
goto out_unlock;
|
goto out_unlock;
|
if (backlog > sk->max_ack_backlog)
|
if (backlog > sk->max_ack_backlog)
|
wake_up_interruptible_all(&sk->protinfo.af_unix.peer_wait);
|
wake_up_interruptible_all(&sk->protinfo.af_unix.peer_wait);
|
sk->max_ack_backlog=backlog;
|
sk->max_ack_backlog=backlog;
|
sk->state=TCP_LISTEN;
|
sk->state=TCP_LISTEN;
|
/* set credentials so connect can copy them */
|
/* set credentials so connect can copy them */
|
sk->peercred.pid = current->pid;
|
sk->peercred.pid = current->pid;
|
sk->peercred.uid = current->euid;
|
sk->peercred.uid = current->euid;
|
sk->peercred.gid = current->egid;
|
sk->peercred.gid = current->egid;
|
err = 0;
|
err = 0;
|
|
|
out_unlock:
|
out_unlock:
|
unix_state_wunlock(sk);
|
unix_state_wunlock(sk);
|
out:
|
out:
|
return err;
|
return err;
|
}
|
}
|
|
|
extern struct proto_ops unix_stream_ops;
|
extern struct proto_ops unix_stream_ops;
|
extern struct proto_ops unix_dgram_ops;
|
extern struct proto_ops unix_dgram_ops;
|
|
|
static struct sock * unix_create1(struct socket *sock)
|
static struct sock * unix_create1(struct socket *sock)
|
{
|
{
|
struct sock *sk;
|
struct sock *sk;
|
|
|
if (atomic_read(&unix_nr_socks) >= 2*files_stat.max_files)
|
if (atomic_read(&unix_nr_socks) >= 2*files_stat.max_files)
|
return NULL;
|
return NULL;
|
|
|
MOD_INC_USE_COUNT;
|
MOD_INC_USE_COUNT;
|
sk = sk_alloc(PF_UNIX, GFP_KERNEL, 1);
|
sk = sk_alloc(PF_UNIX, GFP_KERNEL, 1);
|
if (!sk) {
|
if (!sk) {
|
MOD_DEC_USE_COUNT;
|
MOD_DEC_USE_COUNT;
|
return NULL;
|
return NULL;
|
}
|
}
|
|
|
atomic_inc(&unix_nr_socks);
|
atomic_inc(&unix_nr_socks);
|
|
|
sock_init_data(sock,sk);
|
sock_init_data(sock,sk);
|
|
|
sk->write_space = unix_write_space;
|
sk->write_space = unix_write_space;
|
|
|
sk->max_ack_backlog = sysctl_unix_max_dgram_qlen;
|
sk->max_ack_backlog = sysctl_unix_max_dgram_qlen;
|
sk->destruct = unix_sock_destructor;
|
sk->destruct = unix_sock_destructor;
|
sk->protinfo.af_unix.dentry=NULL;
|
sk->protinfo.af_unix.dentry=NULL;
|
sk->protinfo.af_unix.mnt=NULL;
|
sk->protinfo.af_unix.mnt=NULL;
|
sk->protinfo.af_unix.lock = RW_LOCK_UNLOCKED;
|
sk->protinfo.af_unix.lock = RW_LOCK_UNLOCKED;
|
atomic_set(&sk->protinfo.af_unix.inflight, sock ? 0 : -1);
|
atomic_set(&sk->protinfo.af_unix.inflight, sock ? 0 : -1);
|
init_MUTEX(&sk->protinfo.af_unix.readsem);/* single task reading lock */
|
init_MUTEX(&sk->protinfo.af_unix.readsem);/* single task reading lock */
|
init_waitqueue_head(&sk->protinfo.af_unix.peer_wait);
|
init_waitqueue_head(&sk->protinfo.af_unix.peer_wait);
|
sk->protinfo.af_unix.list=NULL;
|
sk->protinfo.af_unix.list=NULL;
|
unix_insert_socket(&unix_sockets_unbound, sk);
|
unix_insert_socket(&unix_sockets_unbound, sk);
|
|
|
return sk;
|
return sk;
|
}
|
}
|
|
|
static int unix_create(struct socket *sock, int protocol)
|
static int unix_create(struct socket *sock, int protocol)
|
{
|
{
|
if (protocol && protocol != PF_UNIX)
|
if (protocol && protocol != PF_UNIX)
|
return -EPROTONOSUPPORT;
|
return -EPROTONOSUPPORT;
|
|
|
sock->state = SS_UNCONNECTED;
|
sock->state = SS_UNCONNECTED;
|
|
|
switch (sock->type) {
|
switch (sock->type) {
|
case SOCK_STREAM:
|
case SOCK_STREAM:
|
sock->ops = &unix_stream_ops;
|
sock->ops = &unix_stream_ops;
|
break;
|
break;
|
/*
|
/*
|
* Believe it or not BSD has AF_UNIX, SOCK_RAW though
|
* Believe it or not BSD has AF_UNIX, SOCK_RAW though
|
* nothing uses it.
|
* nothing uses it.
|
*/
|
*/
|
case SOCK_RAW:
|
case SOCK_RAW:
|
sock->type=SOCK_DGRAM;
|
sock->type=SOCK_DGRAM;
|
case SOCK_DGRAM:
|
case SOCK_DGRAM:
|
sock->ops = &unix_dgram_ops;
|
sock->ops = &unix_dgram_ops;
|
break;
|
break;
|
default:
|
default:
|
return -ESOCKTNOSUPPORT;
|
return -ESOCKTNOSUPPORT;
|
}
|
}
|
|
|
return unix_create1(sock) ? 0 : -ENOMEM;
|
return unix_create1(sock) ? 0 : -ENOMEM;
|
}
|
}
|
|
|
static int unix_release(struct socket *sock)
|
static int unix_release(struct socket *sock)
|
{
|
{
|
unix_socket *sk = sock->sk;
|
unix_socket *sk = sock->sk;
|
|
|
if (!sk)
|
if (!sk)
|
return 0;
|
return 0;
|
|
|
sock->sk = NULL;
|
sock->sk = NULL;
|
|
|
return unix_release_sock (sk, 0);
|
return unix_release_sock (sk, 0);
|
}
|
}
|
|
|
static int unix_autobind(struct socket *sock)
|
static int unix_autobind(struct socket *sock)
|
{
|
{
|
struct sock *sk = sock->sk;
|
struct sock *sk = sock->sk;
|
static u32 ordernum = 1;
|
static u32 ordernum = 1;
|
struct unix_address * addr;
|
struct unix_address * addr;
|
int err;
|
int err;
|
|
|
down(&sk->protinfo.af_unix.readsem);
|
down(&sk->protinfo.af_unix.readsem);
|
|
|
err = 0;
|
err = 0;
|
if (sk->protinfo.af_unix.addr)
|
if (sk->protinfo.af_unix.addr)
|
goto out;
|
goto out;
|
|
|
err = -ENOMEM;
|
err = -ENOMEM;
|
addr = kmalloc(sizeof(*addr) + sizeof(short) + 16, GFP_KERNEL);
|
addr = kmalloc(sizeof(*addr) + sizeof(short) + 16, GFP_KERNEL);
|
if (!addr)
|
if (!addr)
|
goto out;
|
goto out;
|
|
|
memset(addr, 0, sizeof(*addr) + sizeof(short) + 16);
|
memset(addr, 0, sizeof(*addr) + sizeof(short) + 16);
|
addr->name->sun_family = AF_UNIX;
|
addr->name->sun_family = AF_UNIX;
|
atomic_set(&addr->refcnt, 1);
|
atomic_set(&addr->refcnt, 1);
|
|
|
retry:
|
retry:
|
addr->len = sprintf(addr->name->sun_path+1, "%05x", ordernum) + 1 + sizeof(short);
|
addr->len = sprintf(addr->name->sun_path+1, "%05x", ordernum) + 1 + sizeof(short);
|
addr->hash = unix_hash_fold(csum_partial((void*)addr->name, addr->len, 0));
|
addr->hash = unix_hash_fold(csum_partial((void*)addr->name, addr->len, 0));
|
|
|
write_lock(&unix_table_lock);
|
write_lock(&unix_table_lock);
|
ordernum = (ordernum+1)&0xFFFFF;
|
ordernum = (ordernum+1)&0xFFFFF;
|
|
|
if (__unix_find_socket_byname(addr->name, addr->len, sock->type,
|
if (__unix_find_socket_byname(addr->name, addr->len, sock->type,
|
addr->hash)) {
|
addr->hash)) {
|
write_unlock(&unix_table_lock);
|
write_unlock(&unix_table_lock);
|
/* Sanity yield. It is unusual case, but yet... */
|
/* Sanity yield. It is unusual case, but yet... */
|
if (!(ordernum&0xFF))
|
if (!(ordernum&0xFF))
|
yield();
|
yield();
|
goto retry;
|
goto retry;
|
}
|
}
|
addr->hash ^= sk->type;
|
addr->hash ^= sk->type;
|
|
|
__unix_remove_socket(sk);
|
__unix_remove_socket(sk);
|
sk->protinfo.af_unix.addr = addr;
|
sk->protinfo.af_unix.addr = addr;
|
__unix_insert_socket(&unix_socket_table[addr->hash], sk);
|
__unix_insert_socket(&unix_socket_table[addr->hash], sk);
|
write_unlock(&unix_table_lock);
|
write_unlock(&unix_table_lock);
|
err = 0;
|
err = 0;
|
|
|
out:
|
out:
|
up(&sk->protinfo.af_unix.readsem);
|
up(&sk->protinfo.af_unix.readsem);
|
return err;
|
return err;
|
}
|
}
|
|
|
static unix_socket *unix_find_other(struct sockaddr_un *sunname, int len,
|
static unix_socket *unix_find_other(struct sockaddr_un *sunname, int len,
|
int type, unsigned hash, int *error)
|
int type, unsigned hash, int *error)
|
{
|
{
|
unix_socket *u;
|
unix_socket *u;
|
struct nameidata nd;
|
struct nameidata nd;
|
int err = 0;
|
int err = 0;
|
|
|
if (sunname->sun_path[0]) {
|
if (sunname->sun_path[0]) {
|
if (path_init(sunname->sun_path,
|
if (path_init(sunname->sun_path,
|
LOOKUP_POSITIVE|LOOKUP_FOLLOW, &nd))
|
LOOKUP_POSITIVE|LOOKUP_FOLLOW, &nd))
|
err = path_walk(sunname->sun_path, &nd);
|
err = path_walk(sunname->sun_path, &nd);
|
if (err)
|
if (err)
|
goto fail;
|
goto fail;
|
err = permission(nd.dentry->d_inode,MAY_WRITE);
|
err = permission(nd.dentry->d_inode,MAY_WRITE);
|
if (err)
|
if (err)
|
goto put_fail;
|
goto put_fail;
|
|
|
err = -ECONNREFUSED;
|
err = -ECONNREFUSED;
|
if (!S_ISSOCK(nd.dentry->d_inode->i_mode))
|
if (!S_ISSOCK(nd.dentry->d_inode->i_mode))
|
goto put_fail;
|
goto put_fail;
|
u=unix_find_socket_byinode(nd.dentry->d_inode);
|
u=unix_find_socket_byinode(nd.dentry->d_inode);
|
if (!u)
|
if (!u)
|
goto put_fail;
|
goto put_fail;
|
|
|
if (u->type == type)
|
if (u->type == type)
|
UPDATE_ATIME(nd.dentry->d_inode);
|
UPDATE_ATIME(nd.dentry->d_inode);
|
|
|
path_release(&nd);
|
path_release(&nd);
|
|
|
err=-EPROTOTYPE;
|
err=-EPROTOTYPE;
|
if (u->type != type) {
|
if (u->type != type) {
|
sock_put(u);
|
sock_put(u);
|
goto fail;
|
goto fail;
|
}
|
}
|
} else {
|
} else {
|
err = -ECONNREFUSED;
|
err = -ECONNREFUSED;
|
u=unix_find_socket_byname(sunname, len, type, hash);
|
u=unix_find_socket_byname(sunname, len, type, hash);
|
if (u) {
|
if (u) {
|
struct dentry *dentry;
|
struct dentry *dentry;
|
dentry = u->protinfo.af_unix.dentry;
|
dentry = u->protinfo.af_unix.dentry;
|
if (dentry)
|
if (dentry)
|
UPDATE_ATIME(dentry->d_inode);
|
UPDATE_ATIME(dentry->d_inode);
|
} else
|
} else
|
goto fail;
|
goto fail;
|
}
|
}
|
return u;
|
return u;
|
|
|
put_fail:
|
put_fail:
|
path_release(&nd);
|
path_release(&nd);
|
fail:
|
fail:
|
*error=err;
|
*error=err;
|
return NULL;
|
return NULL;
|
}
|
}
|
|
|
|
|
static int unix_bind(struct socket *sock, struct sockaddr *uaddr, int addr_len)
|
static int unix_bind(struct socket *sock, struct sockaddr *uaddr, int addr_len)
|
{
|
{
|
struct sock *sk = sock->sk;
|
struct sock *sk = sock->sk;
|
struct sockaddr_un *sunaddr=(struct sockaddr_un *)uaddr;
|
struct sockaddr_un *sunaddr=(struct sockaddr_un *)uaddr;
|
struct dentry * dentry = NULL;
|
struct dentry * dentry = NULL;
|
struct nameidata nd;
|
struct nameidata nd;
|
int err;
|
int err;
|
unsigned hash;
|
unsigned hash;
|
struct unix_address *addr;
|
struct unix_address *addr;
|
unix_socket **list;
|
unix_socket **list;
|
|
|
err = -EINVAL;
|
err = -EINVAL;
|
if (sunaddr->sun_family != AF_UNIX)
|
if (sunaddr->sun_family != AF_UNIX)
|
goto out;
|
goto out;
|
|
|
if (addr_len==sizeof(short)) {
|
if (addr_len==sizeof(short)) {
|
err = unix_autobind(sock);
|
err = unix_autobind(sock);
|
goto out;
|
goto out;
|
}
|
}
|
|
|
err = unix_mkname(sunaddr, addr_len, &hash);
|
err = unix_mkname(sunaddr, addr_len, &hash);
|
if (err < 0)
|
if (err < 0)
|
goto out;
|
goto out;
|
addr_len = err;
|
addr_len = err;
|
|
|
down(&sk->protinfo.af_unix.readsem);
|
down(&sk->protinfo.af_unix.readsem);
|
|
|
err = -EINVAL;
|
err = -EINVAL;
|
if (sk->protinfo.af_unix.addr)
|
if (sk->protinfo.af_unix.addr)
|
goto out_up;
|
goto out_up;
|
|
|
err = -ENOMEM;
|
err = -ENOMEM;
|
addr = kmalloc(sizeof(*addr)+addr_len, GFP_KERNEL);
|
addr = kmalloc(sizeof(*addr)+addr_len, GFP_KERNEL);
|
if (!addr)
|
if (!addr)
|
goto out_up;
|
goto out_up;
|
|
|
memcpy(addr->name, sunaddr, addr_len);
|
memcpy(addr->name, sunaddr, addr_len);
|
addr->len = addr_len;
|
addr->len = addr_len;
|
addr->hash = hash^sk->type;
|
addr->hash = hash^sk->type;
|
atomic_set(&addr->refcnt, 1);
|
atomic_set(&addr->refcnt, 1);
|
|
|
if (sunaddr->sun_path[0]) {
|
if (sunaddr->sun_path[0]) {
|
unsigned int mode;
|
unsigned int mode;
|
err = 0;
|
err = 0;
|
/*
|
/*
|
* Get the parent directory, calculate the hash for last
|
* Get the parent directory, calculate the hash for last
|
* component.
|
* component.
|
*/
|
*/
|
if (path_init(sunaddr->sun_path, LOOKUP_PARENT, &nd))
|
if (path_init(sunaddr->sun_path, LOOKUP_PARENT, &nd))
|
err = path_walk(sunaddr->sun_path, &nd);
|
err = path_walk(sunaddr->sun_path, &nd);
|
if (err)
|
if (err)
|
goto out_mknod_parent;
|
goto out_mknod_parent;
|
/*
|
/*
|
* Yucky last component or no last component at all?
|
* Yucky last component or no last component at all?
|
* (foo/., foo/.., /////)
|
* (foo/., foo/.., /////)
|
*/
|
*/
|
err = -EEXIST;
|
err = -EEXIST;
|
if (nd.last_type != LAST_NORM)
|
if (nd.last_type != LAST_NORM)
|
goto out_mknod;
|
goto out_mknod;
|
/*
|
/*
|
* Lock the directory.
|
* Lock the directory.
|
*/
|
*/
|
down(&nd.dentry->d_inode->i_sem);
|
down(&nd.dentry->d_inode->i_sem);
|
/*
|
/*
|
* Do the final lookup.
|
* Do the final lookup.
|
*/
|
*/
|
dentry = lookup_hash(&nd.last, nd.dentry);
|
dentry = lookup_hash(&nd.last, nd.dentry);
|
err = PTR_ERR(dentry);
|
err = PTR_ERR(dentry);
|
if (IS_ERR(dentry))
|
if (IS_ERR(dentry))
|
goto out_mknod_unlock;
|
goto out_mknod_unlock;
|
err = -ENOENT;
|
err = -ENOENT;
|
/*
|
/*
|
* Special case - lookup gave negative, but... we had foo/bar/
|
* Special case - lookup gave negative, but... we had foo/bar/
|
* From the vfs_mknod() POV we just have a negative dentry -
|
* From the vfs_mknod() POV we just have a negative dentry -
|
* all is fine. Let's be bastards - you had / on the end, you've
|
* all is fine. Let's be bastards - you had / on the end, you've
|
* been asking for (non-existent) directory. -ENOENT for you.
|
* been asking for (non-existent) directory. -ENOENT for you.
|
*/
|
*/
|
if (nd.last.name[nd.last.len] && !dentry->d_inode)
|
if (nd.last.name[nd.last.len] && !dentry->d_inode)
|
goto out_mknod_dput;
|
goto out_mknod_dput;
|
/*
|
/*
|
* All right, let's create it.
|
* All right, let's create it.
|
*/
|
*/
|
mode = S_IFSOCK | (sock->inode->i_mode & ~current->fs->umask);
|
mode = S_IFSOCK | (sock->inode->i_mode & ~current->fs->umask);
|
err = vfs_mknod(nd.dentry->d_inode, dentry, mode, 0);
|
err = vfs_mknod(nd.dentry->d_inode, dentry, mode, 0);
|
if (err)
|
if (err)
|
goto out_mknod_dput;
|
goto out_mknod_dput;
|
up(&nd.dentry->d_inode->i_sem);
|
up(&nd.dentry->d_inode->i_sem);
|
dput(nd.dentry);
|
dput(nd.dentry);
|
nd.dentry = dentry;
|
nd.dentry = dentry;
|
|
|
addr->hash = UNIX_HASH_SIZE;
|
addr->hash = UNIX_HASH_SIZE;
|
}
|
}
|
|
|
write_lock(&unix_table_lock);
|
write_lock(&unix_table_lock);
|
|
|
if (!sunaddr->sun_path[0]) {
|
if (!sunaddr->sun_path[0]) {
|
err = -EADDRINUSE;
|
err = -EADDRINUSE;
|
if (__unix_find_socket_byname(sunaddr, addr_len,
|
if (__unix_find_socket_byname(sunaddr, addr_len,
|
sk->type, hash)) {
|
sk->type, hash)) {
|
unix_release_addr(addr);
|
unix_release_addr(addr);
|
goto out_unlock;
|
goto out_unlock;
|
}
|
}
|
|
|
list = &unix_socket_table[addr->hash];
|
list = &unix_socket_table[addr->hash];
|
} else {
|
} else {
|
list = &unix_socket_table[dentry->d_inode->i_ino & (UNIX_HASH_SIZE-1)];
|
list = &unix_socket_table[dentry->d_inode->i_ino & (UNIX_HASH_SIZE-1)];
|
sk->protinfo.af_unix.dentry = nd.dentry;
|
sk->protinfo.af_unix.dentry = nd.dentry;
|
sk->protinfo.af_unix.mnt = nd.mnt;
|
sk->protinfo.af_unix.mnt = nd.mnt;
|
}
|
}
|
|
|
err = 0;
|
err = 0;
|
__unix_remove_socket(sk);
|
__unix_remove_socket(sk);
|
sk->protinfo.af_unix.addr = addr;
|
sk->protinfo.af_unix.addr = addr;
|
__unix_insert_socket(list, sk);
|
__unix_insert_socket(list, sk);
|
|
|
out_unlock:
|
out_unlock:
|
write_unlock(&unix_table_lock);
|
write_unlock(&unix_table_lock);
|
out_up:
|
out_up:
|
up(&sk->protinfo.af_unix.readsem);
|
up(&sk->protinfo.af_unix.readsem);
|
out:
|
out:
|
return err;
|
return err;
|
|
|
out_mknod_dput:
|
out_mknod_dput:
|
dput(dentry);
|
dput(dentry);
|
out_mknod_unlock:
|
out_mknod_unlock:
|
up(&nd.dentry->d_inode->i_sem);
|
up(&nd.dentry->d_inode->i_sem);
|
out_mknod:
|
out_mknod:
|
path_release(&nd);
|
path_release(&nd);
|
out_mknod_parent:
|
out_mknod_parent:
|
if (err==-EEXIST)
|
if (err==-EEXIST)
|
err=-EADDRINUSE;
|
err=-EADDRINUSE;
|
unix_release_addr(addr);
|
unix_release_addr(addr);
|
goto out_up;
|
goto out_up;
|
}
|
}
|
|
|
static int unix_dgram_connect(struct socket *sock, struct sockaddr *addr,
|
static int unix_dgram_connect(struct socket *sock, struct sockaddr *addr,
|
int alen, int flags)
|
int alen, int flags)
|
{
|
{
|
struct sock *sk = sock->sk;
|
struct sock *sk = sock->sk;
|
struct sockaddr_un *sunaddr=(struct sockaddr_un*)addr;
|
struct sockaddr_un *sunaddr=(struct sockaddr_un*)addr;
|
struct sock *other;
|
struct sock *other;
|
unsigned hash;
|
unsigned hash;
|
int err;
|
int err;
|
|
|
if (addr->sa_family != AF_UNSPEC) {
|
if (addr->sa_family != AF_UNSPEC) {
|
err = unix_mkname(sunaddr, alen, &hash);
|
err = unix_mkname(sunaddr, alen, &hash);
|
if (err < 0)
|
if (err < 0)
|
goto out;
|
goto out;
|
alen = err;
|
alen = err;
|
|
|
if (sock->passcred && !sk->protinfo.af_unix.addr &&
|
if (sock->passcred && !sk->protinfo.af_unix.addr &&
|
(err = unix_autobind(sock)) != 0)
|
(err = unix_autobind(sock)) != 0)
|
goto out;
|
goto out;
|
|
|
other=unix_find_other(sunaddr, alen, sock->type, hash, &err);
|
other=unix_find_other(sunaddr, alen, sock->type, hash, &err);
|
if (!other)
|
if (!other)
|
goto out;
|
goto out;
|
|
|
unix_state_wlock(sk);
|
unix_state_wlock(sk);
|
|
|
err = -EPERM;
|
err = -EPERM;
|
if (!unix_may_send(sk, other))
|
if (!unix_may_send(sk, other))
|
goto out_unlock;
|
goto out_unlock;
|
} else {
|
} else {
|
/*
|
/*
|
* 1003.1g breaking connected state with AF_UNSPEC
|
* 1003.1g breaking connected state with AF_UNSPEC
|
*/
|
*/
|
other = NULL;
|
other = NULL;
|
unix_state_wlock(sk);
|
unix_state_wlock(sk);
|
}
|
}
|
|
|
/*
|
/*
|
* If it was connected, reconnect.
|
* If it was connected, reconnect.
|
*/
|
*/
|
if (unix_peer(sk)) {
|
if (unix_peer(sk)) {
|
struct sock *old_peer = unix_peer(sk);
|
struct sock *old_peer = unix_peer(sk);
|
unix_peer(sk)=other;
|
unix_peer(sk)=other;
|
unix_state_wunlock(sk);
|
unix_state_wunlock(sk);
|
|
|
if (other != old_peer)
|
if (other != old_peer)
|
unix_dgram_disconnected(sk, old_peer);
|
unix_dgram_disconnected(sk, old_peer);
|
sock_put(old_peer);
|
sock_put(old_peer);
|
} else {
|
} else {
|
unix_peer(sk)=other;
|
unix_peer(sk)=other;
|
unix_state_wunlock(sk);
|
unix_state_wunlock(sk);
|
}
|
}
|
return 0;
|
return 0;
|
|
|
out_unlock:
|
out_unlock:
|
unix_state_wunlock(sk);
|
unix_state_wunlock(sk);
|
sock_put(other);
|
sock_put(other);
|
out:
|
out:
|
return err;
|
return err;
|
}
|
}
|
|
|
static long unix_wait_for_peer(unix_socket *other, long timeo)
|
static long unix_wait_for_peer(unix_socket *other, long timeo)
|
{
|
{
|
int sched;
|
int sched;
|
DECLARE_WAITQUEUE(wait, current);
|
DECLARE_WAITQUEUE(wait, current);
|
|
|
__set_current_state(TASK_INTERRUPTIBLE);
|
__set_current_state(TASK_INTERRUPTIBLE);
|
add_wait_queue_exclusive(&other->protinfo.af_unix.peer_wait, &wait);
|
add_wait_queue_exclusive(&other->protinfo.af_unix.peer_wait, &wait);
|
|
|
sched = (!other->dead &&
|
sched = (!other->dead &&
|
!(other->shutdown&RCV_SHUTDOWN) &&
|
!(other->shutdown&RCV_SHUTDOWN) &&
|
skb_queue_len(&other->receive_queue) > other->max_ack_backlog);
|
skb_queue_len(&other->receive_queue) > other->max_ack_backlog);
|
|
|
unix_state_runlock(other);
|
unix_state_runlock(other);
|
|
|
if (sched)
|
if (sched)
|
timeo = schedule_timeout(timeo);
|
timeo = schedule_timeout(timeo);
|
|
|
__set_current_state(TASK_RUNNING);
|
__set_current_state(TASK_RUNNING);
|
remove_wait_queue(&other->protinfo.af_unix.peer_wait, &wait);
|
remove_wait_queue(&other->protinfo.af_unix.peer_wait, &wait);
|
return timeo;
|
return timeo;
|
}
|
}
|
|
|
static int unix_stream_connect(struct socket *sock, struct sockaddr *uaddr,
|
static int unix_stream_connect(struct socket *sock, struct sockaddr *uaddr,
|
int addr_len, int flags)
|
int addr_len, int flags)
|
{
|
{
|
struct sockaddr_un *sunaddr=(struct sockaddr_un *)uaddr;
|
struct sockaddr_un *sunaddr=(struct sockaddr_un *)uaddr;
|
struct sock *sk = sock->sk;
|
struct sock *sk = sock->sk;
|
struct sock *newsk = NULL;
|
struct sock *newsk = NULL;
|
unix_socket *other = NULL;
|
unix_socket *other = NULL;
|
struct sk_buff *skb = NULL;
|
struct sk_buff *skb = NULL;
|
unsigned hash;
|
unsigned hash;
|
int st;
|
int st;
|
int err;
|
int err;
|
long timeo;
|
long timeo;
|
|
|
err = unix_mkname(sunaddr, addr_len, &hash);
|
err = unix_mkname(sunaddr, addr_len, &hash);
|
if (err < 0)
|
if (err < 0)
|
goto out;
|
goto out;
|
addr_len = err;
|
addr_len = err;
|
|
|
if (sock->passcred && !sk->protinfo.af_unix.addr &&
|
if (sock->passcred && !sk->protinfo.af_unix.addr &&
|
(err = unix_autobind(sock)) != 0)
|
(err = unix_autobind(sock)) != 0)
|
goto out;
|
goto out;
|
|
|
timeo = sock_sndtimeo(sk, flags & O_NONBLOCK);
|
timeo = sock_sndtimeo(sk, flags & O_NONBLOCK);
|
|
|
/* First of all allocate resources.
|
/* First of all allocate resources.
|
If we will make it after state is locked,
|
If we will make it after state is locked,
|
we will have to recheck all again in any case.
|
we will have to recheck all again in any case.
|
*/
|
*/
|
|
|
err = -ENOMEM;
|
err = -ENOMEM;
|
|
|
/* create new sock for complete connection */
|
/* create new sock for complete connection */
|
newsk = unix_create1(NULL);
|
newsk = unix_create1(NULL);
|
if (newsk == NULL)
|
if (newsk == NULL)
|
goto out;
|
goto out;
|
|
|
/* Allocate skb for sending to listening sock */
|
/* Allocate skb for sending to listening sock */
|
skb = sock_wmalloc(newsk, 1, 0, GFP_KERNEL);
|
skb = sock_wmalloc(newsk, 1, 0, GFP_KERNEL);
|
if (skb == NULL)
|
if (skb == NULL)
|
goto out;
|
goto out;
|
|
|
restart:
|
restart:
|
/* Find listening sock. */
|
/* Find listening sock. */
|
other=unix_find_other(sunaddr, addr_len, sk->type, hash, &err);
|
other=unix_find_other(sunaddr, addr_len, sk->type, hash, &err);
|
if (!other)
|
if (!other)
|
goto out;
|
goto out;
|
|
|
/* Latch state of peer */
|
/* Latch state of peer */
|
unix_state_rlock(other);
|
unix_state_rlock(other);
|
|
|
/* Apparently VFS overslept socket death. Retry. */
|
/* Apparently VFS overslept socket death. Retry. */
|
if (other->dead) {
|
if (other->dead) {
|
unix_state_runlock(other);
|
unix_state_runlock(other);
|
sock_put(other);
|
sock_put(other);
|
goto restart;
|
goto restart;
|
}
|
}
|
|
|
err = -ECONNREFUSED;
|
err = -ECONNREFUSED;
|
if (other->state != TCP_LISTEN)
|
if (other->state != TCP_LISTEN)
|
goto out_unlock;
|
goto out_unlock;
|
|
|
if (skb_queue_len(&other->receive_queue) > other->max_ack_backlog) {
|
if (skb_queue_len(&other->receive_queue) > other->max_ack_backlog) {
|
err = -EAGAIN;
|
err = -EAGAIN;
|
if (!timeo)
|
if (!timeo)
|
goto out_unlock;
|
goto out_unlock;
|
|
|
timeo = unix_wait_for_peer(other, timeo);
|
timeo = unix_wait_for_peer(other, timeo);
|
|
|
err = sock_intr_errno(timeo);
|
err = sock_intr_errno(timeo);
|
if (signal_pending(current))
|
if (signal_pending(current))
|
goto out;
|
goto out;
|
sock_put(other);
|
sock_put(other);
|
goto restart;
|
goto restart;
|
}
|
}
|
|
|
/* Latch our state.
|
/* Latch our state.
|
|
|
It is tricky place. We need to grab write lock and cannot
|
It is tricky place. We need to grab write lock and cannot
|
drop lock on peer. It is dangerous because deadlock is
|
drop lock on peer. It is dangerous because deadlock is
|
possible. Connect to self case and simultaneous
|
possible. Connect to self case and simultaneous
|
attempt to connect are eliminated by checking socket
|
attempt to connect are eliminated by checking socket
|
state. other is TCP_LISTEN, if sk is TCP_LISTEN we
|
state. other is TCP_LISTEN, if sk is TCP_LISTEN we
|
check this before attempt to grab lock.
|
check this before attempt to grab lock.
|
|
|
Well, and we have to recheck the state after socket locked.
|
Well, and we have to recheck the state after socket locked.
|
*/
|
*/
|
st = sk->state;
|
st = sk->state;
|
|
|
switch (st) {
|
switch (st) {
|
case TCP_CLOSE:
|
case TCP_CLOSE:
|
/* This is ok... continue with connect */
|
/* This is ok... continue with connect */
|
break;
|
break;
|
case TCP_ESTABLISHED:
|
case TCP_ESTABLISHED:
|
/* Socket is already connected */
|
/* Socket is already connected */
|
err = -EISCONN;
|
err = -EISCONN;
|
goto out_unlock;
|
goto out_unlock;
|
default:
|
default:
|
err = -EINVAL;
|
err = -EINVAL;
|
goto out_unlock;
|
goto out_unlock;
|
}
|
}
|
|
|
unix_state_wlock(sk);
|
unix_state_wlock(sk);
|
|
|
if (sk->state != st) {
|
if (sk->state != st) {
|
unix_state_wunlock(sk);
|
unix_state_wunlock(sk);
|
unix_state_runlock(other);
|
unix_state_runlock(other);
|
sock_put(other);
|
sock_put(other);
|
goto restart;
|
goto restart;
|
}
|
}
|
|
|
/* The way is open! Fastly set all the necessary fields... */
|
/* The way is open! Fastly set all the necessary fields... */
|
|
|
sock_hold(sk);
|
sock_hold(sk);
|
unix_peer(newsk)=sk;
|
unix_peer(newsk)=sk;
|
newsk->state=TCP_ESTABLISHED;
|
newsk->state=TCP_ESTABLISHED;
|
newsk->type=SOCK_STREAM;
|
newsk->type=SOCK_STREAM;
|
newsk->peercred.pid = current->pid;
|
newsk->peercred.pid = current->pid;
|
newsk->peercred.uid = current->euid;
|
newsk->peercred.uid = current->euid;
|
newsk->peercred.gid = current->egid;
|
newsk->peercred.gid = current->egid;
|
newsk->sleep = &newsk->protinfo.af_unix.peer_wait;
|
newsk->sleep = &newsk->protinfo.af_unix.peer_wait;
|
|
|
/* copy address information from listening to new sock*/
|
/* copy address information from listening to new sock*/
|
if (other->protinfo.af_unix.addr)
|
if (other->protinfo.af_unix.addr)
|
{
|
{
|
atomic_inc(&other->protinfo.af_unix.addr->refcnt);
|
atomic_inc(&other->protinfo.af_unix.addr->refcnt);
|
newsk->protinfo.af_unix.addr=other->protinfo.af_unix.addr;
|
newsk->protinfo.af_unix.addr=other->protinfo.af_unix.addr;
|
}
|
}
|
if (other->protinfo.af_unix.dentry) {
|
if (other->protinfo.af_unix.dentry) {
|
newsk->protinfo.af_unix.dentry=dget(other->protinfo.af_unix.dentry);
|
newsk->protinfo.af_unix.dentry=dget(other->protinfo.af_unix.dentry);
|
newsk->protinfo.af_unix.mnt=mntget(other->protinfo.af_unix.mnt);
|
newsk->protinfo.af_unix.mnt=mntget(other->protinfo.af_unix.mnt);
|
}
|
}
|
|
|
/* Set credentials */
|
/* Set credentials */
|
sk->peercred = other->peercred;
|
sk->peercred = other->peercred;
|
|
|
sock_hold(newsk);
|
sock_hold(newsk);
|
unix_peer(sk)=newsk;
|
unix_peer(sk)=newsk;
|
sock->state=SS_CONNECTED;
|
sock->state=SS_CONNECTED;
|
sk->state=TCP_ESTABLISHED;
|
sk->state=TCP_ESTABLISHED;
|
|
|
unix_state_wunlock(sk);
|
unix_state_wunlock(sk);
|
|
|
/* take ten and and send info to listening sock */
|
/* take ten and and send info to listening sock */
|
spin_lock(&other->receive_queue.lock);
|
spin_lock(&other->receive_queue.lock);
|
__skb_queue_tail(&other->receive_queue,skb);
|
__skb_queue_tail(&other->receive_queue,skb);
|
/* Undo artificially decreased inflight after embrion
|
/* Undo artificially decreased inflight after embrion
|
* is installed to listening socket. */
|
* is installed to listening socket. */
|
atomic_inc(&newsk->protinfo.af_unix.inflight);
|
atomic_inc(&newsk->protinfo.af_unix.inflight);
|
spin_unlock(&other->receive_queue.lock);
|
spin_unlock(&other->receive_queue.lock);
|
unix_state_runlock(other);
|
unix_state_runlock(other);
|
other->data_ready(other, 0);
|
other->data_ready(other, 0);
|
sock_put(other);
|
sock_put(other);
|
return 0;
|
return 0;
|
|
|
out_unlock:
|
out_unlock:
|
if (other)
|
if (other)
|
unix_state_runlock(other);
|
unix_state_runlock(other);
|
|
|
out:
|
out:
|
if (skb)
|
if (skb)
|
kfree_skb(skb);
|
kfree_skb(skb);
|
if (newsk)
|
if (newsk)
|
unix_release_sock(newsk, 0);
|
unix_release_sock(newsk, 0);
|
if (other)
|
if (other)
|
sock_put(other);
|
sock_put(other);
|
return err;
|
return err;
|
}
|
}
|
|
|
static int unix_socketpair(struct socket *socka, struct socket *sockb)
|
static int unix_socketpair(struct socket *socka, struct socket *sockb)
|
{
|
{
|
struct sock *ska=socka->sk, *skb = sockb->sk;
|
struct sock *ska=socka->sk, *skb = sockb->sk;
|
|
|
/* Join our sockets back to back */
|
/* Join our sockets back to back */
|
sock_hold(ska);
|
sock_hold(ska);
|
sock_hold(skb);
|
sock_hold(skb);
|
unix_peer(ska)=skb;
|
unix_peer(ska)=skb;
|
unix_peer(skb)=ska;
|
unix_peer(skb)=ska;
|
ska->peercred.pid = skb->peercred.pid = current->pid;
|
ska->peercred.pid = skb->peercred.pid = current->pid;
|
ska->peercred.uid = skb->peercred.uid = current->euid;
|
ska->peercred.uid = skb->peercred.uid = current->euid;
|
ska->peercred.gid = skb->peercred.gid = current->egid;
|
ska->peercred.gid = skb->peercred.gid = current->egid;
|
|
|
if (ska->type != SOCK_DGRAM)
|
if (ska->type != SOCK_DGRAM)
|
{
|
{
|
ska->state=TCP_ESTABLISHED;
|
ska->state=TCP_ESTABLISHED;
|
skb->state=TCP_ESTABLISHED;
|
skb->state=TCP_ESTABLISHED;
|
socka->state=SS_CONNECTED;
|
socka->state=SS_CONNECTED;
|
sockb->state=SS_CONNECTED;
|
sockb->state=SS_CONNECTED;
|
}
|
}
|
return 0;
|
return 0;
|
}
|
}
|
|
|
static int unix_accept(struct socket *sock, struct socket *newsock, int flags)
|
static int unix_accept(struct socket *sock, struct socket *newsock, int flags)
|
{
|
{
|
unix_socket *sk = sock->sk;
|
unix_socket *sk = sock->sk;
|
unix_socket *tsk;
|
unix_socket *tsk;
|
struct sk_buff *skb;
|
struct sk_buff *skb;
|
int err;
|
int err;
|
|
|
err = -EOPNOTSUPP;
|
err = -EOPNOTSUPP;
|
if (sock->type!=SOCK_STREAM)
|
if (sock->type!=SOCK_STREAM)
|
goto out;
|
goto out;
|
|
|
err = -EINVAL;
|
err = -EINVAL;
|
if (sk->state!=TCP_LISTEN)
|
if (sk->state!=TCP_LISTEN)
|
goto out;
|
goto out;
|
|
|
/* If socket state is TCP_LISTEN it cannot change (for now...),
|
/* If socket state is TCP_LISTEN it cannot change (for now...),
|
* so that no locks are necessary.
|
* so that no locks are necessary.
|
*/
|
*/
|
|
|
skb = skb_recv_datagram(sk, 0, flags&O_NONBLOCK, &err);
|
skb = skb_recv_datagram(sk, 0, flags&O_NONBLOCK, &err);
|
if (!skb) {
|
if (!skb) {
|
/* This means receive shutdown. */
|
/* This means receive shutdown. */
|
if (err == 0)
|
if (err == 0)
|
err = -EINVAL;
|
err = -EINVAL;
|
goto out;
|
goto out;
|
}
|
}
|
|
|
tsk = skb->sk;
|
tsk = skb->sk;
|
skb_free_datagram(sk, skb);
|
skb_free_datagram(sk, skb);
|
wake_up_interruptible(&sk->protinfo.af_unix.peer_wait);
|
wake_up_interruptible(&sk->protinfo.af_unix.peer_wait);
|
|
|
/* attach accepted sock to socket */
|
/* attach accepted sock to socket */
|
unix_state_wlock(tsk);
|
unix_state_wlock(tsk);
|
newsock->state = SS_CONNECTED;
|
newsock->state = SS_CONNECTED;
|
sock_graft(tsk, newsock);
|
sock_graft(tsk, newsock);
|
unix_state_wunlock(tsk);
|
unix_state_wunlock(tsk);
|
return 0;
|
return 0;
|
|
|
out:
|
out:
|
return err;
|
return err;
|
}
|
}
|
|
|
|
|
static int unix_getname(struct socket *sock, struct sockaddr *uaddr, int *uaddr_len, int peer)
|
static int unix_getname(struct socket *sock, struct sockaddr *uaddr, int *uaddr_len, int peer)
|
{
|
{
|
struct sock *sk = sock->sk;
|
struct sock *sk = sock->sk;
|
struct sockaddr_un *sunaddr=(struct sockaddr_un *)uaddr;
|
struct sockaddr_un *sunaddr=(struct sockaddr_un *)uaddr;
|
int err = 0;
|
int err = 0;
|
|
|
if (peer) {
|
if (peer) {
|
sk = unix_peer_get(sk);
|
sk = unix_peer_get(sk);
|
|
|
err = -ENOTCONN;
|
err = -ENOTCONN;
|
if (!sk)
|
if (!sk)
|
goto out;
|
goto out;
|
err = 0;
|
err = 0;
|
} else {
|
} else {
|
sock_hold(sk);
|
sock_hold(sk);
|
}
|
}
|
|
|
unix_state_rlock(sk);
|
unix_state_rlock(sk);
|
if (!sk->protinfo.af_unix.addr) {
|
if (!sk->protinfo.af_unix.addr) {
|
sunaddr->sun_family = AF_UNIX;
|
sunaddr->sun_family = AF_UNIX;
|
sunaddr->sun_path[0] = 0;
|
sunaddr->sun_path[0] = 0;
|
*uaddr_len = sizeof(short);
|
*uaddr_len = sizeof(short);
|
} else {
|
} else {
|
struct unix_address *addr = sk->protinfo.af_unix.addr;
|
struct unix_address *addr = sk->protinfo.af_unix.addr;
|
|
|
*uaddr_len = addr->len;
|
*uaddr_len = addr->len;
|
memcpy(sunaddr, addr->name, *uaddr_len);
|
memcpy(sunaddr, addr->name, *uaddr_len);
|
}
|
}
|
unix_state_runlock(sk);
|
unix_state_runlock(sk);
|
sock_put(sk);
|
sock_put(sk);
|
out:
|
out:
|
return err;
|
return err;
|
}
|
}
|
|
|
static void unix_detach_fds(struct scm_cookie *scm, struct sk_buff *skb)
|
static void unix_detach_fds(struct scm_cookie *scm, struct sk_buff *skb)
|
{
|
{
|
int i;
|
int i;
|
|
|
scm->fp = UNIXCB(skb).fp;
|
scm->fp = UNIXCB(skb).fp;
|
skb->destructor = sock_wfree;
|
skb->destructor = sock_wfree;
|
UNIXCB(skb).fp = NULL;
|
UNIXCB(skb).fp = NULL;
|
|
|
for (i=scm->fp->count-1; i>=0; i--)
|
for (i=scm->fp->count-1; i>=0; i--)
|
unix_notinflight(scm->fp->fp[i]);
|
unix_notinflight(scm->fp->fp[i]);
|
}
|
}
|
|
|
static void unix_destruct_fds(struct sk_buff *skb)
|
static void unix_destruct_fds(struct sk_buff *skb)
|
{
|
{
|
struct scm_cookie scm;
|
struct scm_cookie scm;
|
memset(&scm, 0, sizeof(scm));
|
memset(&scm, 0, sizeof(scm));
|
unix_detach_fds(&scm, skb);
|
unix_detach_fds(&scm, skb);
|
|
|
/* Alas, it calls VFS */
|
/* Alas, it calls VFS */
|
/* So fscking what? fput() had been SMP-safe since the last Summer */
|
/* So fscking what? fput() had been SMP-safe since the last Summer */
|
scm_destroy(&scm);
|
scm_destroy(&scm);
|
sock_wfree(skb);
|
sock_wfree(skb);
|
}
|
}
|
|
|
static void unix_attach_fds(struct scm_cookie *scm, struct sk_buff *skb)
|
static void unix_attach_fds(struct scm_cookie *scm, struct sk_buff *skb)
|
{
|
{
|
int i;
|
int i;
|
for (i=scm->fp->count-1; i>=0; i--)
|
for (i=scm->fp->count-1; i>=0; i--)
|
unix_inflight(scm->fp->fp[i]);
|
unix_inflight(scm->fp->fp[i]);
|
UNIXCB(skb).fp = scm->fp;
|
UNIXCB(skb).fp = scm->fp;
|
skb->destructor = unix_destruct_fds;
|
skb->destructor = unix_destruct_fds;
|
scm->fp = NULL;
|
scm->fp = NULL;
|
}
|
}
|
|
|
/*
|
/*
|
* Send AF_UNIX data.
|
* Send AF_UNIX data.
|
*/
|
*/
|
|
|
static int unix_dgram_sendmsg(struct socket *sock, struct msghdr *msg, int len,
|
static int unix_dgram_sendmsg(struct socket *sock, struct msghdr *msg, int len,
|
struct scm_cookie *scm)
|
struct scm_cookie *scm)
|
{
|
{
|
struct sock *sk = sock->sk;
|
struct sock *sk = sock->sk;
|
struct sockaddr_un *sunaddr=msg->msg_name;
|
struct sockaddr_un *sunaddr=msg->msg_name;
|
unix_socket *other = NULL;
|
unix_socket *other = NULL;
|
int namelen = 0; /* fake GCC */
|
int namelen = 0; /* fake GCC */
|
int err;
|
int err;
|
unsigned hash;
|
unsigned hash;
|
struct sk_buff *skb;
|
struct sk_buff *skb;
|
long timeo;
|
long timeo;
|
|
|
err = -EOPNOTSUPP;
|
err = -EOPNOTSUPP;
|
if (msg->msg_flags&MSG_OOB)
|
if (msg->msg_flags&MSG_OOB)
|
goto out;
|
goto out;
|
|
|
if (msg->msg_namelen) {
|
if (msg->msg_namelen) {
|
err = unix_mkname(sunaddr, msg->msg_namelen, &hash);
|
err = unix_mkname(sunaddr, msg->msg_namelen, &hash);
|
if (err < 0)
|
if (err < 0)
|
goto out;
|
goto out;
|
namelen = err;
|
namelen = err;
|
} else {
|
} else {
|
sunaddr = NULL;
|
sunaddr = NULL;
|
err = -ENOTCONN;
|
err = -ENOTCONN;
|
other = unix_peer_get(sk);
|
other = unix_peer_get(sk);
|
if (!other)
|
if (!other)
|
goto out;
|
goto out;
|
}
|
}
|
|
|
if (sock->passcred && !sk->protinfo.af_unix.addr &&
|
if (sock->passcred && !sk->protinfo.af_unix.addr &&
|
(err = unix_autobind(sock)) != 0)
|
(err = unix_autobind(sock)) != 0)
|
goto out;
|
goto out;
|
|
|
err = -EMSGSIZE;
|
err = -EMSGSIZE;
|
if ((unsigned)len > sk->sndbuf - 32)
|
if ((unsigned)len > sk->sndbuf - 32)
|
goto out;
|
goto out;
|
|
|
skb = sock_alloc_send_skb(sk, len, msg->msg_flags&MSG_DONTWAIT, &err);
|
skb = sock_alloc_send_skb(sk, len, msg->msg_flags&MSG_DONTWAIT, &err);
|
if (skb==NULL)
|
if (skb==NULL)
|
goto out;
|
goto out;
|
|
|
memcpy(UNIXCREDS(skb), &scm->creds, sizeof(struct ucred));
|
memcpy(UNIXCREDS(skb), &scm->creds, sizeof(struct ucred));
|
if (scm->fp)
|
if (scm->fp)
|
unix_attach_fds(scm, skb);
|
unix_attach_fds(scm, skb);
|
|
|
skb->h.raw = skb->data;
|
skb->h.raw = skb->data;
|
err = memcpy_fromiovec(skb_put(skb,len), msg->msg_iov, len);
|
err = memcpy_fromiovec(skb_put(skb,len), msg->msg_iov, len);
|
if (err)
|
if (err)
|
goto out_free;
|
goto out_free;
|
|
|
timeo = sock_sndtimeo(sk, msg->msg_flags & MSG_DONTWAIT);
|
timeo = sock_sndtimeo(sk, msg->msg_flags & MSG_DONTWAIT);
|
|
|
restart:
|
restart:
|
if (!other) {
|
if (!other) {
|
err = -ECONNRESET;
|
err = -ECONNRESET;
|
if (sunaddr == NULL)
|
if (sunaddr == NULL)
|
goto out_free;
|
goto out_free;
|
|
|
other = unix_find_other(sunaddr, namelen, sk->type, hash, &err);
|
other = unix_find_other(sunaddr, namelen, sk->type, hash, &err);
|
if (other==NULL)
|
if (other==NULL)
|
goto out_free;
|
goto out_free;
|
}
|
}
|
|
|
unix_state_rlock(other);
|
unix_state_rlock(other);
|
err = -EPERM;
|
err = -EPERM;
|
if (!unix_may_send(sk, other))
|
if (!unix_may_send(sk, other))
|
goto out_unlock;
|
goto out_unlock;
|
|
|
if (other->dead) {
|
if (other->dead) {
|
/*
|
/*
|
* Check with 1003.1g - what should
|
* Check with 1003.1g - what should
|
* datagram error
|
* datagram error
|
*/
|
*/
|
unix_state_runlock(other);
|
unix_state_runlock(other);
|
sock_put(other);
|
sock_put(other);
|
|
|
err = 0;
|
err = 0;
|
unix_state_wlock(sk);
|
unix_state_wlock(sk);
|
if (unix_peer(sk) == other) {
|
if (unix_peer(sk) == other) {
|
unix_peer(sk)=NULL;
|
unix_peer(sk)=NULL;
|
unix_state_wunlock(sk);
|
unix_state_wunlock(sk);
|
|
|
unix_dgram_disconnected(sk, other);
|
unix_dgram_disconnected(sk, other);
|
sock_put(other);
|
sock_put(other);
|
err = -ECONNREFUSED;
|
err = -ECONNREFUSED;
|
} else {
|
} else {
|
unix_state_wunlock(sk);
|
unix_state_wunlock(sk);
|
}
|
}
|
|
|
other = NULL;
|
other = NULL;
|
if (err)
|
if (err)
|
goto out_free;
|
goto out_free;
|
goto restart;
|
goto restart;
|
}
|
}
|
|
|
err = -EPIPE;
|
err = -EPIPE;
|
if (other->shutdown&RCV_SHUTDOWN)
|
if (other->shutdown&RCV_SHUTDOWN)
|
goto out_unlock;
|
goto out_unlock;
|
|
|
if (unix_peer(other) != sk &&
|
if (unix_peer(other) != sk &&
|
skb_queue_len(&other->receive_queue) > other->max_ack_backlog) {
|
skb_queue_len(&other->receive_queue) > other->max_ack_backlog) {
|
if (!timeo) {
|
if (!timeo) {
|
err = -EAGAIN;
|
err = -EAGAIN;
|
goto out_unlock;
|
goto out_unlock;
|
}
|
}
|
|
|
timeo = unix_wait_for_peer(other, timeo);
|
timeo = unix_wait_for_peer(other, timeo);
|
|
|
err = sock_intr_errno(timeo);
|
err = sock_intr_errno(timeo);
|
if (signal_pending(current))
|
if (signal_pending(current))
|
goto out_free;
|
goto out_free;
|
|
|
goto restart;
|
goto restart;
|
}
|
}
|
|
|
skb_queue_tail(&other->receive_queue, skb);
|
skb_queue_tail(&other->receive_queue, skb);
|
unix_state_runlock(other);
|
unix_state_runlock(other);
|
other->data_ready(other, len);
|
other->data_ready(other, len);
|
sock_put(other);
|
sock_put(other);
|
return len;
|
return len;
|
|
|
out_unlock:
|
out_unlock:
|
unix_state_runlock(other);
|
unix_state_runlock(other);
|
out_free:
|
out_free:
|
kfree_skb(skb);
|
kfree_skb(skb);
|
out:
|
out:
|
if (other)
|
if (other)
|
sock_put(other);
|
sock_put(other);
|
return err;
|
return err;
|
}
|
}
|
|
|
|
|
static int unix_stream_sendmsg(struct socket *sock, struct msghdr *msg, int len,
|
static int unix_stream_sendmsg(struct socket *sock, struct msghdr *msg, int len,
|
struct scm_cookie *scm)
|
struct scm_cookie *scm)
|
{
|
{
|
struct sock *sk = sock->sk;
|
struct sock *sk = sock->sk;
|
unix_socket *other = NULL;
|
unix_socket *other = NULL;
|
struct sockaddr_un *sunaddr=msg->msg_name;
|
struct sockaddr_un *sunaddr=msg->msg_name;
|
int err,size;
|
int err,size;
|
struct sk_buff *skb;
|
struct sk_buff *skb;
|
int sent=0;
|
int sent=0;
|
|
|
err = -EOPNOTSUPP;
|
err = -EOPNOTSUPP;
|
if (msg->msg_flags&MSG_OOB)
|
if (msg->msg_flags&MSG_OOB)
|
goto out_err;
|
goto out_err;
|
|
|
if (msg->msg_namelen) {
|
if (msg->msg_namelen) {
|
err = (sk->state==TCP_ESTABLISHED ? -EISCONN : -EOPNOTSUPP);
|
err = (sk->state==TCP_ESTABLISHED ? -EISCONN : -EOPNOTSUPP);
|
goto out_err;
|
goto out_err;
|
} else {
|
} else {
|
sunaddr = NULL;
|
sunaddr = NULL;
|
err = -ENOTCONN;
|
err = -ENOTCONN;
|
other = unix_peer_get(sk);
|
other = unix_peer_get(sk);
|
if (!other)
|
if (!other)
|
goto out_err;
|
goto out_err;
|
}
|
}
|
|
|
if (sk->shutdown&SEND_SHUTDOWN)
|
if (sk->shutdown&SEND_SHUTDOWN)
|
goto pipe_err;
|
goto pipe_err;
|
|
|
while(sent < len)
|
while(sent < len)
|
{
|
{
|
/*
|
/*
|
* Optimisation for the fact that under 0.01% of X messages typically
|
* Optimisation for the fact that under 0.01% of X messages typically
|
* need breaking up.
|
* need breaking up.
|
*/
|
*/
|
|
|
size=len-sent;
|
size=len-sent;
|
|
|
/* Keep two messages in the pipe so it schedules better */
|
/* Keep two messages in the pipe so it schedules better */
|
if (size > sk->sndbuf/2 - 64)
|
if (size > sk->sndbuf/2 - 64)
|
size = sk->sndbuf/2 - 64;
|
size = sk->sndbuf/2 - 64;
|
|
|
if (size > SKB_MAX_ALLOC)
|
if (size > SKB_MAX_ALLOC)
|
size = SKB_MAX_ALLOC;
|
size = SKB_MAX_ALLOC;
|
|
|
/*
|
/*
|
* Grab a buffer
|
* Grab a buffer
|
*/
|
*/
|
|
|
skb=sock_alloc_send_skb(sk,size,msg->msg_flags&MSG_DONTWAIT, &err);
|
skb=sock_alloc_send_skb(sk,size,msg->msg_flags&MSG_DONTWAIT, &err);
|
|
|
if (skb==NULL)
|
if (skb==NULL)
|
goto out_err;
|
goto out_err;
|
|
|
/*
|
/*
|
* If you pass two values to the sock_alloc_send_skb
|
* If you pass two values to the sock_alloc_send_skb
|
* it tries to grab the large buffer with GFP_NOFS
|
* it tries to grab the large buffer with GFP_NOFS
|
* (which can fail easily), and if it fails grab the
|
* (which can fail easily), and if it fails grab the
|
* fallback size buffer which is under a page and will
|
* fallback size buffer which is under a page and will
|
* succeed. [Alan]
|
* succeed. [Alan]
|
*/
|
*/
|
size = min_t(int, size, skb_tailroom(skb));
|
size = min_t(int, size, skb_tailroom(skb));
|
|
|
memcpy(UNIXCREDS(skb), &scm->creds, sizeof(struct ucred));
|
memcpy(UNIXCREDS(skb), &scm->creds, sizeof(struct ucred));
|
if (scm->fp)
|
if (scm->fp)
|
unix_attach_fds(scm, skb);
|
unix_attach_fds(scm, skb);
|
|
|
if ((err = memcpy_fromiovec(skb_put(skb,size), msg->msg_iov, size)) != 0) {
|
if ((err = memcpy_fromiovec(skb_put(skb,size), msg->msg_iov, size)) != 0) {
|
kfree_skb(skb);
|
kfree_skb(skb);
|
goto out_err;
|
goto out_err;
|
}
|
}
|
|
|
unix_state_rlock(other);
|
unix_state_rlock(other);
|
|
|
if (other->dead || (other->shutdown & RCV_SHUTDOWN))
|
if (other->dead || (other->shutdown & RCV_SHUTDOWN))
|
goto pipe_err_free;
|
goto pipe_err_free;
|
|
|
skb_queue_tail(&other->receive_queue, skb);
|
skb_queue_tail(&other->receive_queue, skb);
|
unix_state_runlock(other);
|
unix_state_runlock(other);
|
other->data_ready(other, size);
|
other->data_ready(other, size);
|
sent+=size;
|
sent+=size;
|
}
|
}
|
sock_put(other);
|
sock_put(other);
|
return sent;
|
return sent;
|
|
|
pipe_err_free:
|
pipe_err_free:
|
unix_state_runlock(other);
|
unix_state_runlock(other);
|
kfree_skb(skb);
|
kfree_skb(skb);
|
pipe_err:
|
pipe_err:
|
if (sent==0 && !(msg->msg_flags&MSG_NOSIGNAL))
|
if (sent==0 && !(msg->msg_flags&MSG_NOSIGNAL))
|
send_sig(SIGPIPE,current,0);
|
send_sig(SIGPIPE,current,0);
|
err = -EPIPE;
|
err = -EPIPE;
|
out_err:
|
out_err:
|
if (other)
|
if (other)
|
sock_put(other);
|
sock_put(other);
|
return sent ? : err;
|
return sent ? : err;
|
}
|
}
|
|
|
static void unix_copy_addr(struct msghdr *msg, struct sock *sk)
|
static void unix_copy_addr(struct msghdr *msg, struct sock *sk)
|
{
|
{
|
msg->msg_namelen = 0;
|
msg->msg_namelen = 0;
|
if (sk->protinfo.af_unix.addr) {
|
if (sk->protinfo.af_unix.addr) {
|
msg->msg_namelen=sk->protinfo.af_unix.addr->len;
|
msg->msg_namelen=sk->protinfo.af_unix.addr->len;
|
memcpy(msg->msg_name,
|
memcpy(msg->msg_name,
|
sk->protinfo.af_unix.addr->name,
|
sk->protinfo.af_unix.addr->name,
|
sk->protinfo.af_unix.addr->len);
|
sk->protinfo.af_unix.addr->len);
|
}
|
}
|
}
|
}
|
|
|
static int unix_dgram_recvmsg(struct socket *sock, struct msghdr *msg, int size,
|
static int unix_dgram_recvmsg(struct socket *sock, struct msghdr *msg, int size,
|
int flags, struct scm_cookie *scm)
|
int flags, struct scm_cookie *scm)
|
{
|
{
|
struct sock *sk = sock->sk;
|
struct sock *sk = sock->sk;
|
int noblock = flags & MSG_DONTWAIT;
|
int noblock = flags & MSG_DONTWAIT;
|
struct sk_buff *skb;
|
struct sk_buff *skb;
|
int err;
|
int err;
|
|
|
err = -EOPNOTSUPP;
|
err = -EOPNOTSUPP;
|
if (flags&MSG_OOB)
|
if (flags&MSG_OOB)
|
goto out;
|
goto out;
|
|
|
msg->msg_namelen = 0;
|
msg->msg_namelen = 0;
|
|
|
skb = skb_recv_datagram(sk, flags, noblock, &err);
|
skb = skb_recv_datagram(sk, flags, noblock, &err);
|
if (!skb)
|
if (!skb)
|
goto out;
|
goto out;
|
|
|
wake_up_interruptible(&sk->protinfo.af_unix.peer_wait);
|
wake_up_interruptible(&sk->protinfo.af_unix.peer_wait);
|
|
|
if (msg->msg_name)
|
if (msg->msg_name)
|
unix_copy_addr(msg, skb->sk);
|
unix_copy_addr(msg, skb->sk);
|
|
|
if (size > skb->len)
|
if (size > skb->len)
|
size = skb->len;
|
size = skb->len;
|
else if (size < skb->len)
|
else if (size < skb->len)
|
msg->msg_flags |= MSG_TRUNC;
|
msg->msg_flags |= MSG_TRUNC;
|
|
|
err = skb_copy_datagram_iovec(skb, 0, msg->msg_iov, size);
|
err = skb_copy_datagram_iovec(skb, 0, msg->msg_iov, size);
|
if (err)
|
if (err)
|
goto out_free;
|
goto out_free;
|
|
|
scm->creds = *UNIXCREDS(skb);
|
scm->creds = *UNIXCREDS(skb);
|
|
|
if (!(flags & MSG_PEEK))
|
if (!(flags & MSG_PEEK))
|
{
|
{
|
if (UNIXCB(skb).fp)
|
if (UNIXCB(skb).fp)
|
unix_detach_fds(scm, skb);
|
unix_detach_fds(scm, skb);
|
}
|
}
|
else
|
else
|
{
|
{
|
/* It is questionable: on PEEK we could:
|
/* It is questionable: on PEEK we could:
|
- do not return fds - good, but too simple 8)
|
- do not return fds - good, but too simple 8)
|
- return fds, and do not return them on read (old strategy,
|
- return fds, and do not return them on read (old strategy,
|
apparently wrong)
|
apparently wrong)
|
- clone fds (I choosed it for now, it is the most universal
|
- clone fds (I choosed it for now, it is the most universal
|
solution)
|
solution)
|
|
|
POSIX 1003.1g does not actually define this clearly
|
POSIX 1003.1g does not actually define this clearly
|
at all. POSIX 1003.1g doesn't define a lot of things
|
at all. POSIX 1003.1g doesn't define a lot of things
|
clearly however!
|
clearly however!
|
|
|
*/
|
*/
|
if (UNIXCB(skb).fp)
|
if (UNIXCB(skb).fp)
|
scm->fp = scm_fp_dup(UNIXCB(skb).fp);
|
scm->fp = scm_fp_dup(UNIXCB(skb).fp);
|
}
|
}
|
err = size;
|
err = size;
|
|
|
out_free:
|
out_free:
|
skb_free_datagram(sk,skb);
|
skb_free_datagram(sk,skb);
|
out:
|
out:
|
return err;
|
return err;
|
}
|
}
|
|
|
/*
|
/*
|
* Sleep until data has arrive. But check for races..
|
* Sleep until data has arrive. But check for races..
|
*/
|
*/
|
|
|
static long unix_stream_data_wait(unix_socket * sk, long timeo)
|
static long unix_stream_data_wait(unix_socket * sk, long timeo)
|
{
|
{
|
DECLARE_WAITQUEUE(wait, current);
|
DECLARE_WAITQUEUE(wait, current);
|
|
|
unix_state_rlock(sk);
|
unix_state_rlock(sk);
|
|
|
add_wait_queue(sk->sleep, &wait);
|
add_wait_queue(sk->sleep, &wait);
|
|
|
for (;;) {
|
for (;;) {
|
set_current_state(TASK_INTERRUPTIBLE);
|
set_current_state(TASK_INTERRUPTIBLE);
|
|
|
if (skb_queue_len(&sk->receive_queue) ||
|
if (skb_queue_len(&sk->receive_queue) ||
|
sk->err ||
|
sk->err ||
|
(sk->shutdown & RCV_SHUTDOWN) ||
|
(sk->shutdown & RCV_SHUTDOWN) ||
|
signal_pending(current) ||
|
signal_pending(current) ||
|
!timeo)
|
!timeo)
|
break;
|
break;
|
|
|
set_bit(SOCK_ASYNC_WAITDATA, &sk->socket->flags);
|
set_bit(SOCK_ASYNC_WAITDATA, &sk->socket->flags);
|
unix_state_runlock(sk);
|
unix_state_runlock(sk);
|
timeo = schedule_timeout(timeo);
|
timeo = schedule_timeout(timeo);
|
unix_state_rlock(sk);
|
unix_state_rlock(sk);
|
clear_bit(SOCK_ASYNC_WAITDATA, &sk->socket->flags);
|
clear_bit(SOCK_ASYNC_WAITDATA, &sk->socket->flags);
|
}
|
}
|
|
|
__set_current_state(TASK_RUNNING);
|
__set_current_state(TASK_RUNNING);
|
remove_wait_queue(sk->sleep, &wait);
|
remove_wait_queue(sk->sleep, &wait);
|
unix_state_runlock(sk);
|
unix_state_runlock(sk);
|
return timeo;
|
return timeo;
|
}
|
}
|
|
|
|
|
|
|
static int unix_stream_recvmsg(struct socket *sock, struct msghdr *msg, int size,
|
static int unix_stream_recvmsg(struct socket *sock, struct msghdr *msg, int size,
|
int flags, struct scm_cookie *scm)
|
int flags, struct scm_cookie *scm)
|
{
|
{
|
struct sock *sk = sock->sk;
|
struct sock *sk = sock->sk;
|
struct sockaddr_un *sunaddr=msg->msg_name;
|
struct sockaddr_un *sunaddr=msg->msg_name;
|
int copied = 0;
|
int copied = 0;
|
int check_creds = 0;
|
int check_creds = 0;
|
int target;
|
int target;
|
int err = 0;
|
int err = 0;
|
long timeo;
|
long timeo;
|
|
|
err = -EINVAL;
|
err = -EINVAL;
|
if (sk->state != TCP_ESTABLISHED)
|
if (sk->state != TCP_ESTABLISHED)
|
goto out;
|
goto out;
|
|
|
err = -EOPNOTSUPP;
|
err = -EOPNOTSUPP;
|
if (flags&MSG_OOB)
|
if (flags&MSG_OOB)
|
goto out;
|
goto out;
|
|
|
target = sock_rcvlowat(sk, flags&MSG_WAITALL, size);
|
target = sock_rcvlowat(sk, flags&MSG_WAITALL, size);
|
timeo = sock_rcvtimeo(sk, flags&MSG_DONTWAIT);
|
timeo = sock_rcvtimeo(sk, flags&MSG_DONTWAIT);
|
|
|
msg->msg_namelen = 0;
|
msg->msg_namelen = 0;
|
|
|
/* Lock the socket to prevent queue disordering
|
/* Lock the socket to prevent queue disordering
|
* while sleeps in memcpy_tomsg
|
* while sleeps in memcpy_tomsg
|
*/
|
*/
|
|
|
down(&sk->protinfo.af_unix.readsem);
|
down(&sk->protinfo.af_unix.readsem);
|
|
|
do
|
do
|
{
|
{
|
int chunk;
|
int chunk;
|
struct sk_buff *skb;
|
struct sk_buff *skb;
|
|
|
skb=skb_dequeue(&sk->receive_queue);
|
skb=skb_dequeue(&sk->receive_queue);
|
if (skb==NULL)
|
if (skb==NULL)
|
{
|
{
|
if (copied >= target)
|
if (copied >= target)
|
break;
|
break;
|
|
|
/*
|
/*
|
* POSIX 1003.1g mandates this order.
|
* POSIX 1003.1g mandates this order.
|
*/
|
*/
|
|
|
if ((err = sock_error(sk)) != 0)
|
if ((err = sock_error(sk)) != 0)
|
break;
|
break;
|
if (sk->shutdown & RCV_SHUTDOWN)
|
if (sk->shutdown & RCV_SHUTDOWN)
|
break;
|
break;
|
err = -EAGAIN;
|
err = -EAGAIN;
|
if (!timeo)
|
if (!timeo)
|
break;
|
break;
|
up(&sk->protinfo.af_unix.readsem);
|
up(&sk->protinfo.af_unix.readsem);
|
|
|
timeo = unix_stream_data_wait(sk, timeo);
|
timeo = unix_stream_data_wait(sk, timeo);
|
|
|
if (signal_pending(current)) {
|
if (signal_pending(current)) {
|
err = sock_intr_errno(timeo);
|
err = sock_intr_errno(timeo);
|
goto out;
|
goto out;
|
}
|
}
|
down(&sk->protinfo.af_unix.readsem);
|
down(&sk->protinfo.af_unix.readsem);
|
continue;
|
continue;
|
}
|
}
|
|
|
if (check_creds) {
|
if (check_creds) {
|
/* Never glue messages from different writers */
|
/* Never glue messages from different writers */
|
if (memcmp(UNIXCREDS(skb), &scm->creds, sizeof(scm->creds)) != 0) {
|
if (memcmp(UNIXCREDS(skb), &scm->creds, sizeof(scm->creds)) != 0) {
|
skb_queue_head(&sk->receive_queue, skb);
|
skb_queue_head(&sk->receive_queue, skb);
|
break;
|
break;
|
}
|
}
|
} else {
|
} else {
|
/* Copy credentials */
|
/* Copy credentials */
|
scm->creds = *UNIXCREDS(skb);
|
scm->creds = *UNIXCREDS(skb);
|
check_creds = 1;
|
check_creds = 1;
|
}
|
}
|
|
|
/* Copy address just once */
|
/* Copy address just once */
|
if (sunaddr)
|
if (sunaddr)
|
{
|
{
|
unix_copy_addr(msg, skb->sk);
|
unix_copy_addr(msg, skb->sk);
|
sunaddr = NULL;
|
sunaddr = NULL;
|
}
|
}
|
|
|
chunk = min_t(unsigned int, skb->len, size);
|
chunk = min_t(unsigned int, skb->len, size);
|
if (memcpy_toiovec(msg->msg_iov, skb->data, chunk)) {
|
if (memcpy_toiovec(msg->msg_iov, skb->data, chunk)) {
|
skb_queue_head(&sk->receive_queue, skb);
|
skb_queue_head(&sk->receive_queue, skb);
|
if (copied == 0)
|
if (copied == 0)
|
copied = -EFAULT;
|
copied = -EFAULT;
|
break;
|
break;
|
}
|
}
|
copied += chunk;
|
copied += chunk;
|
size -= chunk;
|
size -= chunk;
|
|
|
/* Mark read part of skb as used */
|
/* Mark read part of skb as used */
|
if (!(flags & MSG_PEEK))
|
if (!(flags & MSG_PEEK))
|
{
|
{
|
skb_pull(skb, chunk);
|
skb_pull(skb, chunk);
|
|
|
if (UNIXCB(skb).fp)
|
if (UNIXCB(skb).fp)
|
unix_detach_fds(scm, skb);
|
unix_detach_fds(scm, skb);
|
|
|
/* put the skb back if we didn't use it up.. */
|
/* put the skb back if we didn't use it up.. */
|
if (skb->len)
|
if (skb->len)
|
{
|
{
|
skb_queue_head(&sk->receive_queue, skb);
|
skb_queue_head(&sk->receive_queue, skb);
|
break;
|
break;
|
}
|
}
|
|
|
kfree_skb(skb);
|
kfree_skb(skb);
|
|
|
if (scm->fp)
|
if (scm->fp)
|
break;
|
break;
|
}
|
}
|
else
|
else
|
{
|
{
|
/* It is questionable, see note in unix_dgram_recvmsg.
|
/* It is questionable, see note in unix_dgram_recvmsg.
|
*/
|
*/
|
if (UNIXCB(skb).fp)
|
if (UNIXCB(skb).fp)
|
scm->fp = scm_fp_dup(UNIXCB(skb).fp);
|
scm->fp = scm_fp_dup(UNIXCB(skb).fp);
|
|
|
/* put message back and return */
|
/* put message back and return */
|
skb_queue_head(&sk->receive_queue, skb);
|
skb_queue_head(&sk->receive_queue, skb);
|
break;
|
break;
|
}
|
}
|
} while (size);
|
} while (size);
|
|
|
up(&sk->protinfo.af_unix.readsem);
|
up(&sk->protinfo.af_unix.readsem);
|
out:
|
out:
|
return copied ? : err;
|
return copied ? : err;
|
}
|
}
|
|
|
static int unix_shutdown(struct socket *sock, int mode)
|
static int unix_shutdown(struct socket *sock, int mode)
|
{
|
{
|
struct sock *sk = sock->sk;
|
struct sock *sk = sock->sk;
|
unix_socket *other;
|
unix_socket *other;
|
|
|
mode = (mode+1)&(RCV_SHUTDOWN|SEND_SHUTDOWN);
|
mode = (mode+1)&(RCV_SHUTDOWN|SEND_SHUTDOWN);
|
|
|
if (mode) {
|
if (mode) {
|
unix_state_wlock(sk);
|
unix_state_wlock(sk);
|
sk->shutdown |= mode;
|
sk->shutdown |= mode;
|
other=unix_peer(sk);
|
other=unix_peer(sk);
|
if (other)
|
if (other)
|
sock_hold(other);
|
sock_hold(other);
|
unix_state_wunlock(sk);
|
unix_state_wunlock(sk);
|
sk->state_change(sk);
|
sk->state_change(sk);
|
|
|
if (other && sk->type == SOCK_STREAM) {
|
if (other && sk->type == SOCK_STREAM) {
|
int peer_mode = 0;
|
int peer_mode = 0;
|
|
|
if (mode&RCV_SHUTDOWN)
|
if (mode&RCV_SHUTDOWN)
|
peer_mode |= SEND_SHUTDOWN;
|
peer_mode |= SEND_SHUTDOWN;
|
if (mode&SEND_SHUTDOWN)
|
if (mode&SEND_SHUTDOWN)
|
peer_mode |= RCV_SHUTDOWN;
|
peer_mode |= RCV_SHUTDOWN;
|
unix_state_wlock(other);
|
unix_state_wlock(other);
|
other->shutdown |= peer_mode;
|
other->shutdown |= peer_mode;
|
unix_state_wunlock(other);
|
unix_state_wunlock(other);
|
other->state_change(other);
|
other->state_change(other);
|
read_lock(&other->callback_lock);
|
read_lock(&other->callback_lock);
|
if (peer_mode == SHUTDOWN_MASK)
|
if (peer_mode == SHUTDOWN_MASK)
|
sk_wake_async(other,1,POLL_HUP);
|
sk_wake_async(other,1,POLL_HUP);
|
else if (peer_mode & RCV_SHUTDOWN)
|
else if (peer_mode & RCV_SHUTDOWN)
|
sk_wake_async(other,1,POLL_IN);
|
sk_wake_async(other,1,POLL_IN);
|
read_unlock(&other->callback_lock);
|
read_unlock(&other->callback_lock);
|
}
|
}
|
if (other)
|
if (other)
|
sock_put(other);
|
sock_put(other);
|
}
|
}
|
return 0;
|
return 0;
|
}
|
}
|
|
|
static int unix_ioctl(struct socket *sock, unsigned int cmd, unsigned long arg)
|
static int unix_ioctl(struct socket *sock, unsigned int cmd, unsigned long arg)
|
{
|
{
|
struct sock *sk = sock->sk;
|
struct sock *sk = sock->sk;
|
long amount=0;
|
long amount=0;
|
int err;
|
int err;
|
|
|
switch(cmd)
|
switch(cmd)
|
{
|
{
|
case SIOCOUTQ:
|
case SIOCOUTQ:
|
amount = atomic_read(&sk->wmem_alloc);
|
amount = atomic_read(&sk->wmem_alloc);
|
err = put_user(amount, (int *)arg);
|
err = put_user(amount, (int *)arg);
|
break;
|
break;
|
case SIOCINQ:
|
case SIOCINQ:
|
{
|
{
|
struct sk_buff *skb;
|
struct sk_buff *skb;
|
if (sk->state==TCP_LISTEN) {
|
if (sk->state==TCP_LISTEN) {
|
err = -EINVAL;
|
err = -EINVAL;
|
break;
|
break;
|
}
|
}
|
|
|
spin_lock(&sk->receive_queue.lock);
|
spin_lock(&sk->receive_queue.lock);
|
if((skb=skb_peek(&sk->receive_queue))!=NULL)
|
if((skb=skb_peek(&sk->receive_queue))!=NULL)
|
amount=skb->len;
|
amount=skb->len;
|
spin_unlock(&sk->receive_queue.lock);
|
spin_unlock(&sk->receive_queue.lock);
|
err = put_user(amount, (int *)arg);
|
err = put_user(amount, (int *)arg);
|
break;
|
break;
|
}
|
}
|
|
|
default:
|
default:
|
err = dev_ioctl(cmd, (void *)arg);
|
err = dev_ioctl(cmd, (void *)arg);
|
break;
|
break;
|
}
|
}
|
return err;
|
return err;
|
}
|
}
|
|
|
static unsigned int unix_poll(struct file * file, struct socket *sock, poll_table *wait)
|
static unsigned int unix_poll(struct file * file, struct socket *sock, poll_table *wait)
|
{
|
{
|
struct sock *sk = sock->sk;
|
struct sock *sk = sock->sk;
|
unsigned int mask;
|
unsigned int mask;
|
|
|
poll_wait(file, sk->sleep, wait);
|
poll_wait(file, sk->sleep, wait);
|
mask = 0;
|
mask = 0;
|
|
|
/* exceptional events? */
|
/* exceptional events? */
|
if (sk->err)
|
if (sk->err)
|
mask |= POLLERR;
|
mask |= POLLERR;
|
if (sk->shutdown == SHUTDOWN_MASK)
|
if (sk->shutdown == SHUTDOWN_MASK)
|
mask |= POLLHUP;
|
mask |= POLLHUP;
|
|
|
/* readable? */
|
/* readable? */
|
if (!skb_queue_empty(&sk->receive_queue) || (sk->shutdown&RCV_SHUTDOWN))
|
if (!skb_queue_empty(&sk->receive_queue) || (sk->shutdown&RCV_SHUTDOWN))
|
mask |= POLLIN | POLLRDNORM;
|
mask |= POLLIN | POLLRDNORM;
|
|
|
/* Connection-based need to check for termination and startup */
|
/* Connection-based need to check for termination and startup */
|
if (sk->type == SOCK_STREAM && sk->state==TCP_CLOSE)
|
if (sk->type == SOCK_STREAM && sk->state==TCP_CLOSE)
|
mask |= POLLHUP;
|
mask |= POLLHUP;
|
|
|
/*
|
/*
|
* we set writable also when the other side has shut down the
|
* we set writable also when the other side has shut down the
|
* connection. This prevents stuck sockets.
|
* connection. This prevents stuck sockets.
|
*/
|
*/
|
if (unix_writable(sk))
|
if (unix_writable(sk))
|
mask |= POLLOUT | POLLWRNORM | POLLWRBAND;
|
mask |= POLLOUT | POLLWRNORM | POLLWRBAND;
|
|
|
return mask;
|
return mask;
|
}
|
}
|
|
|
|
|
#ifdef CONFIG_PROC_FS
|
#ifdef CONFIG_PROC_FS
|
static int unix_read_proc(char *buffer, char **start, off_t offset,
|
static int unix_read_proc(char *buffer, char **start, off_t offset,
|
int length, int *eof, void *data)
|
int length, int *eof, void *data)
|
{
|
{
|
off_t pos=0;
|
off_t pos=0;
|
off_t begin=0;
|
off_t begin=0;
|
int len=0;
|
int len=0;
|
int i;
|
int i;
|
unix_socket *s;
|
unix_socket *s;
|
|
|
len+= sprintf(buffer,"Num RefCount Protocol Flags Type St "
|
len+= sprintf(buffer,"Num RefCount Protocol Flags Type St "
|
"Inode Path\n");
|
"Inode Path\n");
|
|
|
read_lock(&unix_table_lock);
|
read_lock(&unix_table_lock);
|
forall_unix_sockets (i,s)
|
forall_unix_sockets (i,s)
|
{
|
{
|
unix_state_rlock(s);
|
unix_state_rlock(s);
|
|
|
len+=sprintf(buffer+len,"%p: %08X %08X %08X %04X %02X %5lu",
|
len+=sprintf(buffer+len,"%p: %08X %08X %08X %04X %02X %5lu",
|
s,
|
s,
|
atomic_read(&s->refcnt),
|
atomic_read(&s->refcnt),
|
0,
|
0,
|
s->state == TCP_LISTEN ? __SO_ACCEPTCON : 0,
|
s->state == TCP_LISTEN ? __SO_ACCEPTCON : 0,
|
s->type,
|
s->type,
|
s->socket ?
|
s->socket ?
|
(s->state == TCP_ESTABLISHED ? SS_CONNECTED : SS_UNCONNECTED) :
|
(s->state == TCP_ESTABLISHED ? SS_CONNECTED : SS_UNCONNECTED) :
|
(s->state == TCP_ESTABLISHED ? SS_CONNECTING : SS_DISCONNECTING),
|
(s->state == TCP_ESTABLISHED ? SS_CONNECTING : SS_DISCONNECTING),
|
sock_i_ino(s));
|
sock_i_ino(s));
|
|
|
if (s->protinfo.af_unix.addr)
|
if (s->protinfo.af_unix.addr)
|
{
|
{
|
buffer[len++] = ' ';
|
buffer[len++] = ' ';
|
memcpy(buffer+len, s->protinfo.af_unix.addr->name->sun_path,
|
memcpy(buffer+len, s->protinfo.af_unix.addr->name->sun_path,
|
s->protinfo.af_unix.addr->len-sizeof(short));
|
s->protinfo.af_unix.addr->len-sizeof(short));
|
if (!UNIX_ABSTRACT(s))
|
if (!UNIX_ABSTRACT(s))
|
len--;
|
len--;
|
else
|
else
|
buffer[len] = '@';
|
buffer[len] = '@';
|
len += s->protinfo.af_unix.addr->len - sizeof(short);
|
len += s->protinfo.af_unix.addr->len - sizeof(short);
|
}
|
}
|
unix_state_runlock(s);
|
unix_state_runlock(s);
|
|
|
buffer[len++]='\n';
|
buffer[len++]='\n';
|
|
|
pos = begin + len;
|
pos = begin + len;
|
if(pos<offset)
|
if(pos<offset)
|
{
|
{
|
len=0;
|
len=0;
|
begin=pos;
|
begin=pos;
|
}
|
}
|
if(pos>offset+length)
|
if(pos>offset+length)
|
goto done;
|
goto done;
|
}
|
}
|
*eof = 1;
|
*eof = 1;
|
done:
|
done:
|
read_unlock(&unix_table_lock);
|
read_unlock(&unix_table_lock);
|
*start=buffer+(offset-begin);
|
*start=buffer+(offset-begin);
|
len-=(offset-begin);
|
len-=(offset-begin);
|
if(len>length)
|
if(len>length)
|
len=length;
|
len=length;
|
if (len < 0)
|
if (len < 0)
|
len = 0;
|
len = 0;
|
return len;
|
return len;
|
}
|
}
|
#endif
|
#endif
|
|
|
struct proto_ops unix_stream_ops = {
|
struct proto_ops unix_stream_ops = {
|
family: PF_UNIX,
|
family: PF_UNIX,
|
|
|
release: unix_release,
|
release: unix_release,
|
bind: unix_bind,
|
bind: unix_bind,
|
connect: unix_stream_connect,
|
connect: unix_stream_connect,
|
socketpair: unix_socketpair,
|
socketpair: unix_socketpair,
|
accept: unix_accept,
|
accept: unix_accept,
|
getname: unix_getname,
|
getname: unix_getname,
|
poll: unix_poll,
|
poll: unix_poll,
|
ioctl: unix_ioctl,
|
ioctl: unix_ioctl,
|
listen: unix_listen,
|
listen: unix_listen,
|
shutdown: unix_shutdown,
|
shutdown: unix_shutdown,
|
setsockopt: sock_no_setsockopt,
|
setsockopt: sock_no_setsockopt,
|
getsockopt: sock_no_getsockopt,
|
getsockopt: sock_no_getsockopt,
|
sendmsg: unix_stream_sendmsg,
|
sendmsg: unix_stream_sendmsg,
|
recvmsg: unix_stream_recvmsg,
|
recvmsg: unix_stream_recvmsg,
|
mmap: sock_no_mmap,
|
mmap: sock_no_mmap,
|
sendpage: sock_no_sendpage,
|
sendpage: sock_no_sendpage,
|
};
|
};
|
|
|
struct proto_ops unix_dgram_ops = {
|
struct proto_ops unix_dgram_ops = {
|
family: PF_UNIX,
|
family: PF_UNIX,
|
|
|
release: unix_release,
|
release: unix_release,
|
bind: unix_bind,
|
bind: unix_bind,
|
connect: unix_dgram_connect,
|
connect: unix_dgram_connect,
|
socketpair: unix_socketpair,
|
socketpair: unix_socketpair,
|
accept: sock_no_accept,
|
accept: sock_no_accept,
|
getname: unix_getname,
|
getname: unix_getname,
|
poll: datagram_poll,
|
poll: datagram_poll,
|
ioctl: unix_ioctl,
|
ioctl: unix_ioctl,
|
listen: sock_no_listen,
|
listen: sock_no_listen,
|
shutdown: unix_shutdown,
|
shutdown: unix_shutdown,
|
setsockopt: sock_no_setsockopt,
|
setsockopt: sock_no_setsockopt,
|
getsockopt: sock_no_getsockopt,
|
getsockopt: sock_no_getsockopt,
|
sendmsg: unix_dgram_sendmsg,
|
sendmsg: unix_dgram_sendmsg,
|
recvmsg: unix_dgram_recvmsg,
|
recvmsg: unix_dgram_recvmsg,
|
mmap: sock_no_mmap,
|
mmap: sock_no_mmap,
|
sendpage: sock_no_sendpage,
|
sendpage: sock_no_sendpage,
|
};
|
};
|
|
|
struct net_proto_family unix_family_ops = {
|
struct net_proto_family unix_family_ops = {
|
family: PF_UNIX,
|
family: PF_UNIX,
|
create: unix_create
|
create: unix_create
|
};
|
};
|
|
|
#ifdef CONFIG_SYSCTL
|
#ifdef CONFIG_SYSCTL
|
extern void unix_sysctl_register(void);
|
extern void unix_sysctl_register(void);
|
extern void unix_sysctl_unregister(void);
|
extern void unix_sysctl_unregister(void);
|
#else
|
#else
|
static inline void unix_sysctl_register(void) {}
|
static inline void unix_sysctl_register(void) {}
|
static inline void unix_sysctl_unregister(void) {}
|
static inline void unix_sysctl_unregister(void) {}
|
#endif
|
#endif
|
|
|
static char banner[] __initdata = KERN_INFO "NET4: Unix domain sockets 1.0/SMP for Linux NET4.0.\n";
|
static char banner[] __initdata = KERN_INFO "NET4: Unix domain sockets 1.0/SMP for Linux NET4.0.\n";
|
|
|
static int __init af_unix_init(void)
|
static int __init af_unix_init(void)
|
{
|
{
|
struct sk_buff *dummy_skb;
|
struct sk_buff *dummy_skb;
|
|
|
printk(banner);
|
printk(banner);
|
if (sizeof(struct unix_skb_parms) > sizeof(dummy_skb->cb))
|
if (sizeof(struct unix_skb_parms) > sizeof(dummy_skb->cb))
|
{
|
{
|
printk(KERN_CRIT "unix_proto_init: panic\n");
|
printk(KERN_CRIT "unix_proto_init: panic\n");
|
return -1;
|
return -1;
|
}
|
}
|
sock_register(&unix_family_ops);
|
sock_register(&unix_family_ops);
|
#ifdef CONFIG_PROC_FS
|
#ifdef CONFIG_PROC_FS
|
create_proc_read_entry("net/unix", 0, 0, unix_read_proc, NULL);
|
create_proc_read_entry("net/unix", 0, 0, unix_read_proc, NULL);
|
#endif
|
#endif
|
unix_sysctl_register();
|
unix_sysctl_register();
|
return 0;
|
return 0;
|
}
|
}
|
|
|
static void __exit af_unix_exit(void)
|
static void __exit af_unix_exit(void)
|
{
|
{
|
sock_unregister(PF_UNIX);
|
sock_unregister(PF_UNIX);
|
unix_sysctl_unregister();
|
unix_sysctl_unregister();
|
remove_proc_entry("net/unix", 0);
|
remove_proc_entry("net/unix", 0);
|
}
|
}
|
|
|
module_init(af_unix_init);
|
module_init(af_unix_init);
|
module_exit(af_unix_exit);
|
module_exit(af_unix_exit);
|
|
|
MODULE_LICENSE("GPL");
|
MODULE_LICENSE("GPL");
|
|
|