Return to sys.uipc.t CVS log

Up to [CSRG BSD Unix] / 43BSDReno / share / doc / smm / 13.kchanges

Annotation of 43BSDReno/share/doc/smm/13.kchanges/sys.uipc.t, revision 1.1

1.1     ! root        1: .\" Copyright (c) 1986 Regents of the University of California.
        !             2: .\" All rights reserved.  The Berkeley software License Agreement
        !             3: .\" specifies the terms and conditions for redistribution.
        !             4: .\"
        !             5: .\"    @(#)sys.uipc.t  1.7 (Berkeley) 4/11/86
        !             6: .\"
        !             7: .NH 2
        !             8: Changes in Interprocess Communication support
        !             9: .XP uipc_domain.c
        !            10: The skeletal support for the PUP-1 protocol has been removed.
        !            11: A domain for Xerox NS is now in use.
        !            12: The per-domain data structure allows a per-domain initialization routine
        !            13: to be called at boot time.
        !            14: .XP
        !            15: The \fIpffindproto\fP routine, used in creating a socket to support
        !            16: a specified protocol,
        !            17: takes an additional argument, the type of the socket.
        !            18: It checks both the protocol and type, useful when the same protocol
        !            19: implements multiple socket types.
        !            20: If the type is SOCK_RAW and no exact match is found,
        !            21: a \fIprotosw\fP entry for raw support and a wildcard protocol (number zero)
        !            22: will be used.
        !            23: This allows for a generic raw socket that passes
        !            24: through packets for any given protocol.
        !            25: .XP
        !            26: The second argument to \fIpfctlinput\fP, the generic error-reporting
        !            27: routine, is now declared as a \fIsockaddr\fP pointer.
        !            28: .XP  uipc_mbuf.c
        !            29: The mbuf support routines now use the \fIwait\fP flag passed to \fIm_get\fP
        !            30: or MGET.
        !            31: If M_WAIT is specified, the allocator may wait for free memory,
        !            32: and the allocation is guaranteed to return an mbuf if it returns.
        !            33: In order to prevent the system from slowly going to sleep after
        !            34: exhausting the mbuf pool by losing the mbufs to a leak,
        !            35: the allocator will panic after creating the maximum allocation of mbufs
        !            36: (by default, 256K).
        !            37: Redundant \fIspl\fP's have been removed; most internal routines must
        !            38: be called at \fIsplimp\fP, the highest priority at which mbuf and memory
        !            39: allocation occur.
        !            40: .XP
        !            41: When copying mbuf chains \fIm_copy\fP now preserves the type of each mbuf.
        !            42: There were problems in \fIm_adj\fP, in particular assumptions
        !            43: that there would be no zero-length mbufs within the chain;
        !            44: this was corrected by changing its \fIn\fP-pass algorithm for trimming
        !            45: from the tail of the chain
        !            46: to either one- or two-pass, depending on whether the correction was entirely
        !            47: within the last mbuf.
        !            48: In order to avoid return business, \fIm_pullup\fP was changed
        !            49: to pull additional data (MPULL_EXTRA, defined in \fImbuf.h\fP)
        !            50: into the contiguous area in the first mbuf, if convenient.
        !            51: \fIm_pullup\fP will use the first mbuf of the chain rather then a new one
        !            52: if it can avoid copying.
        !            53: .XP uipc_pipe.c
        !            54: This ``temporary'' file has been removed;
        !            55: pipe now uses \fIsocketpair\fP.
        !            56: .XP uipc_proto.c
        !            57: New entries in the protocol switch for externalization and disposal
        !            58: of access rights are initialized for the Unix domain protocols.
        !            59: .XP uipc_socket.c
        !            60: The \fIsocreate\fP function uses the new interface to \fIpffindproto\fP
        !            61: described above if the protocol is specified by the caller.
        !            62: The \fIsoconnect\fP routine will now try to disconnect a connected socket
        !            63: before reconnecting.
        !            64: This is only allowed if the protocol itself is not connection oriented.
        !            65: Datagram sockets may connect to specify
        !            66: a default destination, then later connect to another destination
        !            67: or to a null destination to disconnect.
        !            68: The \fIsodisconnect\fP routine never used its second argument, and it has
        !            69: been removed.
        !            70: .XP
        !            71: The \fIsosend\fP routine, which implements write and send on sockets,
        !            72: has been restructured for clarity.
        !            73: The old routine had the main loop upside down, first emptying and then filling
        !            74: the buffers.
        !            75: The new implementation also makes it possible to send zero-length datagrams.
        !            76: The maximum length calculation was simplified to avoid problems
        !            77: trying to account for both mbufs and characters of buffer space used.
        !            78: Because of the large improvement in speed of data handling when large
        !            79: buffers are used, \fIsosend\fP will use page clusters if it can use
        !            80: at least half of the cluster.
        !            81: Also, if not using nonblocking I/O,
        !            82: it will wait for output to drain if it has enough data
        !            83: to fill an mbuf cluster but not enough space in the output queue for one,
        !            84: instead of fragmenting the write into small mbufs.
        !            85: A bug allowing access rights to be sent more than once when using scatter-gather
        !            86: I/O (\fIsendmsg\fP) was fixed.
        !            87: A race that occurred when \fIuiomove\fP blocked during a page fault
        !            88: was corrected by allowing the protocol send routines to report disconnection
        !            89: errors; as with disconnection detected earlier, \fIsosend\fP returns
        !            90: EPIPE and sends a SIGPIPE signal to the process.
        !            91: .XP
        !            92: The receive side of socket operations, \fIsoreceive\fP, has also been reworked.
        !            93: The major changes are a reflection of the way that datagrams are now queued;
        !            94: see uipc_socket2.c for further information.
        !            95: The MSG_PEEK flag is passed to the protocol's \fIusrreq\fP routine
        !            96: when requesting out-of-band data so that the protocol may know
        !            97: when the out-of-band data has been consumed.
        !            98: Another bug in access-rights passing was corrected here; the protocol
        !            99: is not called to externalize the data when PEEKing.
        !           100: .XP
        !           101: The \fIsosetopt\fP and \fIsogetopt\fP functions have been expanded
        !           102: considerably.
        !           103: The options that existed in 4.2BSD all set some flag at the socket level.
        !           104: The corresponding options in 4.3BSD use the value argument as a boolean,
        !           105: turning the flag off or on as appropriate.
        !           106: There are a number of additional options at the socket level.
        !           107: Most importantly, it is possible to adjust the send or receive buffer
        !           108: allocation so that higher throughput may be achieved, or that temporary
        !           109: peaks in datagram arrival are less likely to result in datagram loss.
        !           110: The linger option is now set with a structure including a boolean
        !           111: (whether or not to linger) and a time to linger if the boolean is true.
        !           112: Other options have been added to determine the type of a socket
        !           113: (eg, SOCK_STREAM, SOCK_DGRAM), and to collect any outstanding error status.
        !           114: If an option is not destined for the socket level itself,
        !           115: the option is passed to the protocol using the \fIctloutput\fP entry.
        !           116: \fIGetopt\fP's last argument was changed from \fImbuf *\fP to \fImbuf **\fP
        !           117: for consistency with \fIsetopt\fP and the 
        !           118: new \fIctloutput\fP calling convention.
        !           119: .XP
        !           120: \fISelect\fP for exceptional conditions on sockets is now possible,
        !           121: and this returns true when out-of-band data is pending.
        !           122: This is true from the time that the socket layer is notified
        !           123: that the OOB data is on its way until the OOB data has been consumed.
        !           124: The interpretation of socket process groups in 4.2BSD was inconsistent
        !           125: with that of ttys and with the \fIfcntl\fP documentation.
        !           126: This was corrected; positive numbers refer to processes, negative numbers
        !           127: to process groups.
        !           128: The socket process group is used when posting a SIGURG to notify
        !           129: processes of pending out-of-band data.
        !           130: .XP uipc_socket2.c
        !           131: Signal-driven I/O now works with sockets as well as with ttys;
        !           132: \fIsorwakeup\fP and \fIsowwakeup\fP call the new routine \fIsowakeup\fP
        !           133: which calls \fIsbwakeup\fP as before and also sends SIGIO as appropriate.
        !           134: Process groups are interpreted in the same manner as for SIGURG.
        !           135: .XP
        !           136: Larger socket buffers may be used with 4.3BSD than with 4.2BSD;
        !           137: socket buffers (\fIsockbuf\fPs) have been modified to use unsigned short
        !           138: rather than short integers for character counts and mbuf counts.
        !           139: This increases the maximum buffer size to 64K\-1.  These fields
        !           140: should really be unsigned longs, but a socket would no longer fit
        !           141: in an mbuf.
        !           142: So that as much as possible of the allotment may be used,
        !           143: \fIsbreserve\fP allows the high-water mark for data to be set as high as 80%
        !           144: of the maximum value (64K), and sets the high-water mark on mbuf allocation
        !           145: to the smaller of twice the character limit and 64K.
        !           146: .XP
        !           147: In 4.2BSD, datagrams queued in sockbufs were linked through the mbuf
        !           148: \fIm_next\fP field, with \fIm_act\fP set to 1 in the last mbuf
        !           149: of each datagram.
        !           150: Also, each datagram was required to have one mbuf to contain an address,
        !           151: another to contain access rights, and at least one additional mbuf of data.
        !           152: In 4.3BSD, the mbufs comprising a datagram are linked through \fIm_next\fP,
        !           153: and different datagrams are linked through the \fIm_act\fP field of the first
        !           154: mbuf in each.
        !           155: No mbuf is used to represent missing components of a datagram,
        !           156: but the ordering of the mbufs remains important.
        !           157: The components are distinguished by the mbuf type.
        !           158: Any address must be in the first mbuf.
        !           159: Access rights follow the address if present, otherwise they may be first.
        !           160: Data mbufs follow; at least one data buffer will be present
        !           161: if there is no address or access rights.
        !           162: The routines \fIsbappend\fP, \fIsbappendaddr\fP, \fIsbappendrights\fP
        !           163: and \fIsbappendrecord\fP are used to add new data to a sockbuf.
        !           164: The first of these appends to an existing record, and is commonly
        !           165: used for stream sockets.
        !           166: The other three begin new records with address, optional rights, and data
        !           167: (\fIsbappendaddr\fP), with rights and data (\fIsbappendrights\fP),
        !           168: or data only (\fIsbappendrecord\fP).
        !           169: A new internal routine, \fIsbcompress\fP, is used by these functions
        !           170: to compress and append data mbufs to a record.
        !           171: These changes improve the functionality of this layer
        !           172: and in addition make it faster to find the end of a queue.
        !           173: .XP
        !           174: An occasional ``panic: sbdrop'' was due to zero-length mbufs at the end
        !           175: of a chain.
        !           176: Although these should no longer be found in a sockbuf queue,
        !           177: \fIsbdrop\fP was fixed to free empty buffers at the end of the last
        !           178: record.
        !           179: Similarly, \fIsbfree\fP continues to empty a sockbuf as long as mbufs
        !           180: remain, as zero-length packets might be present.
        !           181: \fISbdroprecord\fP was added to free exactly one record from the front
        !           182: of a sockbuf queue.
        !           183: .XP uipc_syscalls.c
        !           184: Errors reported during an \fIaccept\fP call are cleared so that
        !           185: subsequent \fIaccept\fP calls may succeed.
        !           186: A failed attempt to \fIconnect\fP returns the error once only,
        !           187: and SOISCONNECTING is cleared,
        !           188: so that additional connect calls may be attempted.
        !           189: (Lower level protocols may or may not allow this, depending
        !           190: on the nature of the failure.)
        !           191: The \fIsocketpair\fP system call has been fixed to work
        !           192: with datagram sockets as well as with streams,
        !           193: and to clean up properly upon failure.
        !           194: Pipes are now created using \fIconnect2\fP.
        !           195: An additional argument, the type of the data to be fetched,
        !           196: is passed to \fIsockargs\fP. 
        !           197: .XP uipc_usrreq.c
        !           198: The binding and connection of Unix domain sockets has
        !           199: been cleaned up so that \fIrecvfrom\fP and \fIaccept\fP get the address 
        !           200: of the peer (if bound) rather than their own.
        !           201: The Unix-domain connection block records the bound address of a socket,
        !           202: not the address of the socket to which it is connected.
        !           203: For stream sockets, back pressure to implement flow control
        !           204: is now handled by adjusting the limits in the send buffer
        !           205: without overloading the normal count fields; the flow control
        !           206: information was moved to the connection block.
        !           207: Access rights are checked now when connecting; the connected-to socket
        !           208: must be writable by the caller, or the connection request is denied.
        !           209: In order to test one previously unused
        !           210: routine, the Unix domain stream support was modified
        !           211: to support the passage of access rights.
        !           212: Problems with access-rights garbage collection were also noted and fixed,
        !           213: and a count is kept of rights outstanding so that garbage collection
        !           214: is done only when needed.
        !           215: Garbage collection is triggered by socket shutdown now
        !           216: rather than file close; in 4.2BSD, it happened prematurely.
        !           217: The PRU_SENSE \fIusrreq\fP entry, used by \fIstat\fP, has been added.
        !           218: It returns the write buffer size as the ``blocksize,'' and generates
        !           219: a fake inode number and device for the benefit of those programs
        !           220: that use \fIfstat\fP information to determine whether file descriptors refer
        !           221: to the same file.
        !           222: Unimplemented requests have been carefully checked to see that they properly
        !           223: free mbufs when required and never otherwise.
        !           224: Larger buffers are allocated for both stream and datagram sockets.
        !           225: A number of minor bugs have been corrected: the back pointer from an inode
        !           226: to a socket needed to be cleared before release of the inode when detaching;
        !           227: sockets can only be bound once, rather than losing inodes; datagram
        !           228: sockets are correctly marked as connected and disconnected; several mbuf
        !           229: leaks were plugged.
        !           230: A serious problem was corrected in \fIunp_drop\fP: it did not properly
        !           231: abort pending connections, with the result that closing a socket with
        !           232: unaccepted connections would cause an infinite loop trying to drop them.

This archive runs on limited infrastructure. Preserving old code on modern bandwidth. Automated agents are requested to crawl responsibly.