linux/net/rds
David Howells cdfbabfb2f net: Work around lockdep limitation in sockets that use sockets
Lockdep issues a circular dependency warning when AFS issues an operation
through AF_RXRPC from a context in which the VFS/VM holds the mmap_sem.

The theory lockdep comes up with is as follows:

 (1) If the pagefault handler decides it needs to read pages from AFS, it
     calls AFS with mmap_sem held and AFS begins an AF_RXRPC call, but
     creating a call requires the socket lock:

	mmap_sem must be taken before sk_lock-AF_RXRPC

 (2) afs_open_socket() opens an AF_RXRPC socket and binds it.  rxrpc_bind()
     binds the underlying UDP socket whilst holding its socket lock.
     inet_bind() takes its own socket lock:

	sk_lock-AF_RXRPC must be taken before sk_lock-AF_INET

 (3) Reading from a TCP socket into a userspace buffer might cause a fault
     and thus cause the kernel to take the mmap_sem, but the TCP socket is
     locked whilst doing this:

	sk_lock-AF_INET must be taken before mmap_sem

However, lockdep's theory is wrong in this instance because it deals only
with lock classes and not individual locks.  The AF_INET lock in (2) isn't
really equivalent to the AF_INET lock in (3) as the former deals with a
socket entirely internal to the kernel that never sees userspace.  This is
a limitation in the design of lockdep.

Fix the general case by:

 (1) Double up all the locking keys used in sockets so that one set are
     used if the socket is created by userspace and the other set is used
     if the socket is created by the kernel.

 (2) Store the kern parameter passed to sk_alloc() in a variable in the
     sock struct (sk_kern_sock).  This informs sock_lock_init(),
     sock_init_data() and sk_clone_lock() as to the lock keys to be used.

     Note that the child created by sk_clone_lock() inherits the parent's
     kern setting.

 (3) Add a 'kern' parameter to ->accept() that is analogous to the one
     passed in to ->create() that distinguishes whether kernel_accept() or
     sys_accept4() was the caller and can be passed to sk_alloc().

     Note that a lot of accept functions merely dequeue an already
     allocated socket.  I haven't touched these as the new socket already
     exists before we get the parameter.

     Note also that there are a couple of places where I've made the accepted
     socket unconditionally kernel-based:

	irda_accept()
	rds_rcp_accept_one()
	tcp_accept_from_sock()

     because they follow a sock_create_kern() and accept off of that.

Whilst creating this, I noticed that lustre and ocfs don't create sockets
through sock_create_kern() and thus they aren't marked as for-kernel,
though they appear to be internal.  I wonder if these should do that so
that they use the new set of lock keys.

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2017-03-09 18:23:27 -08:00
..
af_rds.c RDS: validate the requested traces user input against max supported 2017-01-06 22:14:26 -05:00
bind.c RDS: log the address on bind failure 2017-01-02 14:02:39 -08:00
cong.c RDS: split out connection specific state from rds_connection to rds_conn_path 2016-06-14 23:50:41 -07:00
connection.c rds: tcp: Take explicit refcounts on struct net 2017-03-07 14:09:59 -08:00
ib_cm.c rds: ib: add error handle 2017-03-09 13:09:18 -08:00
ib_fmr.c RDS: IB: move FMR code to its own file 2016-03-02 14:13:18 -05:00
ib_frmr.c RDS: IB: split the mr registration and invalidation path 2017-01-02 14:02:47 -08:00
ib_mr.h rds: ib: add the static type to the variables 2017-03-01 09:50:58 -08:00
ib_rdma.c RDS: split out connection specific state from rds_connection to rds_conn_path 2016-06-14 23:50:41 -07:00
ib_recv.c RDS: add receive message trace used by application 2017-01-02 14:02:59 -08:00
ib_ring.c
ib_send.c rds:Remove unnecessary ib_ring unalloc 2017-02-17 15:19:51 -05:00
ib_stats.c RDS: IB: add few useful cache stasts 2017-01-02 14:02:51 -08:00
ib_sysctl.c net: Convert uses of typedef ctl_table to struct ctl_table 2013-06-13 02:36:09 -07:00
ib.c rds: remove unnecessary returned value check 2017-03-03 09:51:22 -08:00
ib.h This is a tree wide change and has been kept separate for that reason. 2017-02-25 13:45:43 -08:00
info.c rds: fix an integer overflow test in rds_info_getsockopt() 2015-08-03 15:20:16 -07:00
info.h
Kconfig RDS: Drop stale iWARP RDMA transport 2016-03-02 14:13:17 -05:00
loop.c RDS: TCP: Hooks to set up a single connection path 2016-07-01 16:45:17 -04:00
loop.h
Makefile rds: debug messages are enabled by default 2016-10-29 15:55:57 -04:00
message.c RDS: TCP: Track peer's connection generation number 2016-11-17 13:35:18 -05:00
page.c rds: remove dead code 2016-12-26 21:35:39 -05:00
rdma_transport.c RDS: RDMA: start rdma listening after init 2017-01-02 14:02:45 -08:00
rdma_transport.h RDS: Drop stale iWARP RDMA transport 2016-03-02 14:13:17 -05:00
rdma.c RDS: make message size limit compliant with spec 2017-01-02 14:02:57 -08:00
rds_single_path.h RDS: split out connection specific state from rds_connection to rds_conn_path 2016-06-14 23:50:41 -07:00
rds.h rds: tcp: Take explicit refcounts on struct net 2017-03-07 14:09:59 -08:00
recv.c RDS: add receive message trace used by application 2017-01-02 14:02:59 -08:00
send.c RDS: make message size limit compliant with spec 2017-01-02 14:02:57 -08:00
stats.c net/rds: zero last byte for strncpy 2013-03-08 00:35:44 -05:00
sysctl.c net: rds: fix coding style issues 2016-06-18 21:34:09 -07:00
tcp_connect.c RDS: TCP: Force every connection to be initiated by numerically smaller IP address 2016-11-17 13:35:18 -05:00
tcp_listen.c net: Work around lockdep limitation in sockets that use sockets 2017-03-09 18:23:27 -08:00
tcp_recv.c RDS: add receive message trace used by application 2017-01-02 14:02:59 -08:00
tcp_send.c RDS: TCP: set RDS_FLAG_RETRANSMITTED in cp_retrans list 2016-11-17 13:35:18 -05:00
tcp_stats.c net: rds: fix const array syntax 2011-07-01 16:16:19 -07:00
tcp.c rds: tcp: Sequence teardown of listen and acceptor sockets to avoid races 2017-03-07 14:09:59 -08:00
tcp.h rds: tcp: Sequence teardown of listen and acceptor sockets to avoid races 2017-03-07 14:09:59 -08:00
threads.c rds: Remove duplicate prefix from rds_conn_path_error use 2016-10-17 11:07:22 -04:00
transport.c rds: remove unnecessary returned value check 2017-03-03 09:51:22 -08:00