- Devel - libvirt List Archives

[Libvir] Release of libvirt-0.2.0
by Daniel Veillard 14 Feb '07

14 Feb '07

That's it, it's out ! This include the KVM and QEmu support, and the networking that was discussed for some time now. Since that's a serious set of changes I bumped the intermediate version number, beware ! Available as usual from ftp://libvirt.org/libvirt/ * Various internal cleanups (Mark McLoughlin, Richard Jones, Daniel Berrange, Karel Zak) * Bug fixes: - avoid a crash in connect (Daniel Berrange) - virsh args parsing (Richard Jones) * Add support for QEmu and KVM virtualization (Daniel Berrange) * Add support for network configuration (Mark McLoughlin) * Minor improvements: - regression testing (Daniel Berrange) - localization string updates Thanks to all who contributed, enjoy ! Daniel -- Red Hat Virtualization group http://redhat.com/virtualization/ Daniel Veillard | virtualization library http://libvirt.org/ veillard(a)redhat.com | libxml GNOME XML XSLT toolkit http://xmlsoft.org/ http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/

1 0

[Libvir] libmm
by Richard W.M. Jones 14 Feb '07

14 Feb '07

Does anyone object to adding a dep to libmm (http://www.ossp.org/man/man.cgi/pkg/lib/mm/mm.pod) to libvirt? It simplifies sharing of data between preforked libvirtd instances, and allows us to persist that data (while at the same time not requiring it). Rich. -- Emerging Technologies, Red Hat http://et.redhat.com/~rjones/ 64 Baker Street, London, W1U 7DF Mobile: +44 7866 314 421 "[Negative numbers] darken the very whole doctrines of the equations and make dark of the things which are in their nature excessively obvious and simple" (Francis Maseres FRS, mathematician, 1759)

2 1

[Libvir] PATCH 0/2 QEMU driver
by Daniel P. Berrange 14 Feb '07

14 Feb '07

What follows is yet another iteration of the QEMU patches. Based on recent discussions about a generic remote management daemon, this patch set has been stripped to remove all TCP functionality, and TLS encryption. It is now local only unix domain sockets, with no encyption. THe idea being that the generic libvirtd will be able to expose this over the network securely. The patchset is split into two halves - the driver and the daemon. There will be further explanation alongside each patch. Regards, Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=|

3 7

[Libvir] Preliminary patch to support remote driver / libvirtd
by Richard W.M. Jones 13 Feb '07

13 Feb '07

This patch is just for discussion. It's not in a state to be applied, even if it were accepted (which is a long-shot at present anyway). When looking at the patch, a good starting point is to search for "Architecture and notes" and read from there. Supports: * remote driver (just does the "open", "close", "type" and "version" calls at present as a proof of concept) * TLS transport (built using GnuTLS) * SSH transport (forks external ssh process) * TCP transport (unencrypted - just for testing) * Unix domain socket transport * arbitrary external program / shell script transport * IPv6-ready on client & server I've tested all the transports, and in limited tests they all seem to work. ie. You really can do: virsh -c remote:tls:server version Shortcomings in this version: * in "open" call, name must be non-NULL (this is just a bug) * doesn't actually invoke libvirt on the server side; just prints out messages and returns dummy values * "ssh" not recognised as a service name by getaddrinfo, so you must always give a port number, ie. remote:ssh:server:22 * /tmp/socket should be cleaned up when the server exits Potential problems: * SunRPC is stateless so we need to hand out a cookie to represent the virConnectPtr handle on the server side. However if the client dies without explicitly calling close, we have no way to know, and so the cookie/handle on the server side lives forever. * There's some confusion about the level of abstraction. At the moment I'm abstracting at the driver level, but that may be wrong and possibly I should be abstracting at the level of vir* calls. On the other hand, there's not a huge amount of difference. * Security: Is it safe for libvirt to be connecting to arbitrary TCP sockets? Is it safe for libvirt to be able to run arbitrary programs? Rich. -- Emerging Technologies, Red Hat http://et.redhat.com/~rjones/ 64 Baker Street, London, W1U 7DF Mobile: +44 7866 314 421 "[Negative numbers] darken the very whole doctrines of the equations and make dark of the things which are in their nature excessively obvious and simple" (Francis Maseres FRS, mathematician, 1759) ? src/libvirtd.c ? src/remote_internal.c ? src/remote_internal.h ? src/remote_rpc.x ? src/sunrpc/README ? src/sunrpc/clnt_ext.c ? src/sunrpc/clnt_ext.h ? src/sunrpc/clnt_gnutls.c ? src/sunrpc/clnt_gnutls.h ? src/sunrpc/clnt_tcp2.c ? src/sunrpc/clnt_tcp2.h ? src/sunrpc/create_xid.c ? src/sunrpc/svc_gnutls.c ? src/sunrpc/svc_gnutls.h ? src/sunrpc/svc_tcp2.c ? src/sunrpc/svc_tcp2.h Index: config.h.in =================================================================== RCS file: /data/cvs/libvirt/config.h.in,v retrieving revision 1.7 diff -u -r1.7 config.h.in --- config.h.in 31 Oct 2006 10:25:13 -0000 1.7 +++ config.h.in 30 Jan 2007 17:22:07 -0000 @@ -28,6 +28,9 @@ /* Define to 1 if you have the <inttypes.h> header file. */ #undef HAVE_INTTYPES_H +/* Define to 1 if you have the `gnutls' library (-lgnutls). */ +#undef HAVE_LIBGNUTLS + /* Define to 1 if you have the <memory.h> header file. */ #undef HAVE_MEMORY_H Index: configure.in =================================================================== RCS file: /data/cvs/libvirt/configure.in,v retrieving revision 1.52 diff -u -r1.52 configure.in --- configure.in 22 Jan 2007 15:31:00 -0000 1.52 +++ configure.in 30 Jan 2007 17:22:07 -0000 @@ -146,6 +146,11 @@ AC_SUBST(LIBXML_CONFIG) AC_SUBST(LIBXML_MIN_VERSION) +dnl GnuTLS library +AC_CHECK_LIB(gnutls, gnutls_handshake, + [], + [AC_MSG_ERROR([gnutls library not found])]) + dnl virsh libraries AC_CHECK_LIB(curses, initscr, [VIRSH_LIBS="$VIRSH_LIBS -lcurses"], Index: include/libvirt/virterror.h =================================================================== RCS file: /data/cvs/libvirt/include/libvirt/virterror.h,v retrieving revision 1.17 diff -u -r1.17 virterror.h --- include/libvirt/virterror.h 8 Nov 2006 16:55:20 -0000 1.17 +++ include/libvirt/virterror.h 30 Jan 2007 17:22:07 -0000 @@ -46,7 +46,8 @@ VIR_FROM_DOM, /* Error when operating on a domain */ VIR_FROM_RPC, /* Error in the XML-RPC code */ VIR_FROM_PROXY, /* Error in the proxy code */ - VIR_FROM_CONF /* Error in the configuration file handling */ + VIR_FROM_CONF, /* Error in the configuration file handling */ + VIR_FROM_REMOTE /* Error from remote driver */ } virErrorDomain; @@ -113,7 +114,8 @@ VIR_ERR_PARSE_FAILED, /* failed to parse a conf file */ VIR_ERR_CONF_SYNTAX, /* failed to parse the syntax of a conf file */ VIR_ERR_WRITE_FAILED, /* failed to write a conf file */ - VIR_ERR_XML_DETAIL /* detail of an XML error */ + VIR_ERR_XML_DETAIL, /* detail of an XML error */ + VIR_ERR_RPC /* some sort of RPC error */ } virErrorNumber; /** Index: src/.cvsignore =================================================================== RCS file: /data/cvs/libvirt/src/.cvsignore,v retrieving revision 1.2 diff -u -r1.2 .cvsignore --- src/.cvsignore 5 Jul 2006 21:52:52 -0000 1.2 +++ src/.cvsignore 30 Jan 2007 17:22:07 -0000 @@ -5,3 +5,8 @@ *.lo *.la virsh +libvirtd +remote_rpc_clnt.c +remote_rpc_svc.c +remote_rpc_xdr.c +remote_rpc.h \ No newline at end of file Index: src/Makefile.am =================================================================== RCS file: /data/cvs/libvirt/src/Makefile.am,v retrieving revision 1.31 diff -u -r1.31 Makefile.am --- src/Makefile.am 26 Jan 2007 11:54:29 -0000 1.31 +++ src/Makefile.am 30 Jan 2007 17:22:07 -0000 @@ -28,15 +28,41 @@ driver.h \ proxy_internal.c proxy_internal.h \ conf.c conf.h \ - xm_internal.c xm_internal.h + xm_internal.c xm_internal.h \ + sunrpc/create_xid.c remote_rpc_xdr.c \ + remote_rpc_clnt.c remote_rpc.h \ + sunrpc/clnt_ext.h sunrpc/clnt_gnutls.h \ + sunrpc/clnt_tcp2.h sunrpc/clnt_ext.c \ + sunrpc/clnt_gnutls.c sunrpc/clnt_tcp2.c \ + remote_internal.c remote_internal.h bin_PROGRAMS = virsh +sbin_PROGRAMS = libvirtd virsh_SOURCES = virsh.c console.c console.h virsh_LDFLAGS = virsh_DEPENDENCIES = $(DEPS) virsh_LDADD = $(LDADDS) $(VIRSH_LIBS) +libvirtd_SOURCES = \ + remote_rpc_svc.c \ + sunrpc/svc_gnutls.c sunrpc/svc_gnutls.h \ + sunrpc/svc_tcp2.c sunrpc/svc_tcp2.h \ + libvirtd.c +libvirtd_LDFLAGS = +libvirtd_DEPENDENCIES = $(DEPS) +libvirtd_LDADD = $(LDADDS) remote_rpc_xdr.o + +# Build client and server stubs. +# 'rpcgen' program comes with glibc. +# This is convoluted because we need to build the server stubs +# without main() (-m option), but you can't just do that in a +# simple way. +remote_rpc_clnt.c remote_rpc.h remote_rpc_svc.c remote_rpc_xdr.c: remote_rpc.x + rpcgen remote_rpc.x + rm -f remote_rpc_svc.c + rpcgen -m remote_rpc.x > remote_rpc_svc.c + # # target to ease building test programs # Index: src/driver.h =================================================================== RCS file: /data/cvs/libvirt/src/driver.h,v retrieving revision 1.16 diff -u -r1.16 driver.h --- src/driver.h 22 Jan 2007 16:21:27 -0000 1.16 +++ src/driver.h 30 Jan 2007 17:22:07 -0000 @@ -22,7 +22,8 @@ VIR_DRV_XEN_DAEMON = 3, VIR_DRV_TEST = 4, VIR_DRV_XEN_PROXY = 5, - VIR_DRV_XEN_XM = 6 + VIR_DRV_XEN_XM = 6, + VIR_DRV_REMOTE = 7 } virDrvNo; Index: src/internal.h =================================================================== RCS file: /data/cvs/libvirt/src/internal.h,v retrieving revision 1.28 diff -u -r1.28 internal.h --- src/internal.h 23 Jan 2007 14:39:45 -0000 1.28 +++ src/internal.h 30 Jan 2007 17:22:07 -0000 @@ -117,6 +117,13 @@ struct sockaddr_un addr_un; /* the unix address */ struct sockaddr_in addr_in; /* the inet address */ + /* driver private data + * (Ought to replace the above ad-hoc Xen data, IMHO anyway. + * Currently only the 'remote' driver uses this. + * - RWMJ). + */ + void *private; + /* error stuff */ virError err; /* the last error */ virErrorFunc handler; /* associated handlet */ Index: src/libvirt.c =================================================================== RCS file: /data/cvs/libvirt/src/libvirt.c,v retrieving revision 1.53 diff -u -r1.53 libvirt.c --- src/libvirt.c 23 Jan 2007 14:39:45 -0000 1.53 +++ src/libvirt.c 30 Jan 2007 17:22:09 -0000 @@ -30,6 +30,7 @@ #include "xs_internal.h" #include "xm_internal.h" #include "proxy_internal.h" +#include "remote_internal.h" #include "xml.h" #include "test.h" @@ -79,6 +80,7 @@ xenStoreRegister(); xenXMRegister(); testRegister(); + remoteRegister (); return(0); } Index: src/virsh.c =================================================================== RCS file: /data/cvs/libvirt/src/virsh.c,v retrieving revision 1.47 diff -u -r1.47 virsh.c --- src/virsh.c 28 Jan 2007 19:47:36 -0000 1.47 +++ src/virsh.c 30 Jan 2007 17:22:10 -0000 @@ -2760,7 +2760,7 @@ end = end ? : argc; /* standard (non-command) options */ - while ((arg = getopt_long(end, argv, "d:hqtcv", opt, &idx)) != -1) { + while ((arg = getopt_long(end, argv, "d:hqtc:v", opt, &idx)) != -1) { switch (arg) { case 'd': ctl->debug = atoi(optarg); Index: src/virterror.c =================================================================== RCS file: /data/cvs/libvirt/src/virterror.c,v retrieving revision 1.19 diff -u -r1.19 virterror.c --- src/virterror.c 8 Nov 2006 16:55:20 -0000 1.19 +++ src/virterror.c 30 Jan 2007 17:22:10 -0000 @@ -268,6 +268,9 @@ case VIR_FROM_RPC: dom = "XML-RPC "; break; + case VIR_FROM_REMOTE: + dom = "Remote "; + break; } if ((err->dom != NULL) && (err->code != VIR_ERR_INVALID_DOMAIN)) { domain = err->dom->name; @@ -582,6 +585,12 @@ else errmsg = "%s"; break; + case VIR_ERR_RPC: + if (info == NULL) + errmsg = _("RPC error"); + else + errmsg = "%s"; + break; } return (errmsg); } --- /tmp/newfile 2007-01-30 17:23:46.000000000 +0000 +++ src/libvirtd.c 2007-01-30 16:58:42.000000000 +0000 @@ -0,0 +1,257 @@ +/* + * libvirtd: This is a small server to be used in conjunction with + * the "remote" driver (remote_internal.c). + * + * Copyright (C) 2007 Red Hat, Inc. + * + * See COPYING.LIB for the license of this software. + * + * Richard Jones <rjones(a)redhat.com> + */ + +#include <stdio.h> +#include <stdlib.h> +#include <unistd.h> +#include <malloc.h> +#include <time.h> +#include <rpc/rpc.h> +#include <sys/types.h> +#include <sys/socket.h> +#include <netdb.h> +#include <errno.h> +#include <gnutls/gnutls.h> + +#include "sunrpc/svc_gnutls.h" +#include "sunrpc/svc_tcp2.h" +#include "remote_internal.h" +#include "remote_rpc.h" + +// XXX Replace with variables from a configuration file +// XXX TCP services should be disabled by default +#define listen_tls 1 +#define listen_tcp 1 +#define listen_unix 1 + +// XXX Still to decide where these certificates should be located. +#define KEYFILE "newkey.pem" +#define CERTFILE "newcert.pem" +#define CAFILE "demoCA/cacert.pem" +//#define CRLFILE "crl.pem" + +// This is autogenerated, in remote_rpc_svc.c +extern void libvirtremote_1 (struct svc_req *rqstp, register SVCXPRT *transp); + +/*----------------------------------------------------------------------*/ +/* Server side of the remote procedure calls. */ + +int * +remote_rpc_open_1_svc (char **name, + struct svc_req *req __attribute__((unused))) +{ + static int retcode = 0; + + // Before this is going to work, we will have to assign a cookie + // to each caller to represent their server-side virConnectPtr. + // XXX XXX XXX + //virConnectOpen (name); + + printf ("libvirtd: open: name = %s\n", *name); + return &retcode; +} + +int * +remote_rpc_close_1_svc (void *p __attribute__((unused)), + struct svc_req *req __attribute__((unused))) +{ + static int retcode = 0; + + printf ("libvirtd: close\n"); + return &retcode; +} + +char ** +remote_rpc_type_1_svc (void *p __attribute__((unused)), + struct svc_req *req __attribute__((unused))) +{ + static char *dummy = "remote"; + + printf ("libvirtd: type\n"); + return &dummy; +} + +struct version_ret * +remote_rpc_version_1_svc (void *p __attribute__((unused)), + struct svc_req *req __attribute__((unused))) +{ + static struct version_ret ret = { .retcode = 0, .hvVer = 1000000 }; + + printf ("libvirtd: version\n"); + return &ret; +} + +/*----------------------------------------------------------------------*/ +/* Main function. */ + +static gnutls_certificate_credentials_t x509_cred; +static gnutls_dh_params_t dh_params; + +static void generate_dh_params (void); +static int make_sockets (int *fds, int max_fds, int *nfds_r, + const char *service); + +int +main (int argc, char *argv[]) +{ + if (!listen_tls || !listen_tcp || !listen_unix) { + fprintf (stderr, "libvirtd: you need to enable at least one service in the configuration file\n"); + exit (1); + } + + if (listen_tls) { + /* Initialise GnuTLS. */ + gnutls_global_init (); + + gnutls_certificate_allocate_credentials (&x509_cred); + gnutls_certificate_set_x509_trust_file (x509_cred, CAFILE, + GNUTLS_X509_FMT_PEM); + + // gnutls_certificate_set_x509_crl_file (x509_cred, CRLFILE, + // GNUTLS_X509_FMT_PEM); + + gnutls_certificate_set_x509_key_file (x509_cred, CERTFILE, KEYFILE, + GNUTLS_X509_FMT_PEM); + + generate_dh_params (); + gnutls_certificate_set_dh_params (x509_cred, dh_params); + + int fds[2]; + int nfds = 0; + if (make_sockets (fds, 2, &nfds, LIBVIRTD_GNUTLS_PORT) == -1) + exit (1); + + int i; + for (i = 0; i < nfds; ++i) { + SVCXPRT *transp = svcgnutls_create (fds[i], 0, 0, x509_cred); + if (!transp) { + fprintf (stderr, "libvirtd: cannot create TLS service\n"); + exit (1); + } + + /* Because final arg is 0, this will not register with portmap. */ + if (!svc_register (transp, LIBVIRTREMOTE, LIBVIRTREMOTE_VERS1, + libvirtremote_1, 0)) { + fprintf (stderr, "libvirtd: unable to register (LIBVIRTREMOTE, LIBVIRTREMOTE_VERS1, 0)\n"); + exit (1); + } + } + } + + if (listen_tcp) { + int fds[2]; + int nfds = 0; + if (make_sockets (fds, 2, &nfds, LIBVIRTD_TCP_PORT) == -1) + exit (1); + + int i; + for (i = 0; i < nfds; ++i) { + SVCXPRT *transp = svctcp2_create (fds[i], 0, 0); + if (!transp) { + fprintf (stderr, "libvirtd: cannot create TCP service\n"); + exit (1); + } + + /* Because final arg is 0, this will not register with portmap. */ + if (!svc_register (transp, LIBVIRTREMOTE, LIBVIRTREMOTE_VERS1, + libvirtremote_1, 0)) { + fprintf (stderr, "libvirtd: unable to register (LIBVIRTREMOTE, LIBVIRTREMOTE_VERS1, 0)\n"); + exit (1); + } + } + } + + if (listen_unix) { + char *sockname = LIBVIRTD_UNIX_SOCKET; + int sock = RPC_ANYSOCK; + SVCXPRT *transp = svcunix_create (sock, 0, 0, sockname); + if (!transp) { + fprintf (stderr, "libvirtd: cannot create Unix domain socket service\n"); + exit (1); + } + + /* Because final arg is 0, this will not register with portmap. */ + if (!svc_register (transp, LIBVIRTREMOTE, LIBVIRTREMOTE_VERS1, + libvirtremote_1, 0)) { + fprintf (stderr, "libvirtd: unable to register (LIBVIRTREMOTE, LIBVIRTREMOTE_VERS1, 0)\n"); + exit (1); + } + } + + svc_run (); + fprintf (stderr, "libvirtd: svc_run should not return\n"); + exit (1); +} + +// XXX DH_BITS has to match the value define in svc_gnutls.c +#define DH_BITS 1024 + +static void +generate_dh_params (void) +{ + /* Generate Diffie Hellman parameters - for use with DHE + * kx algorithms. These should be discarded and regenerated + * once a day, once a week or once a month. Depending on the + * security requirements. + */ + gnutls_dh_params_init (&dh_params); + gnutls_dh_params_generate2 (dh_params, DH_BITS); +} + +// See: http://people.redhat.com/drepper/userapi-ipv6.html +static int +make_sockets (int *fds, int max_fds, int *nfds_r, const char *service) +{ + struct addrinfo *ai; + struct addrinfo hints; + memset (&hints, 0, sizeof hints); + hints.ai_flags = AI_PASSIVE | AI_ADDRCONFIG; + hints.ai_socktype = SOCK_STREAM; + + int e = getaddrinfo (NULL, service, &hints, &ai); + if (e != 0) { + fprintf (stderr, "getaddrinfo: %s\n", gai_strerror (e)); + return -1; + } + + struct addrinfo *runp = ai; + while (runp && *nfds_r < max_fds) { + fds[*nfds_r] = socket (runp->ai_family, runp->ai_socktype, + runp->ai_protocol); + if (fds[*nfds_r] == -1) { + perror ("socket"); + return -1; + } + + int opt = 1; + setsockopt (fds[*nfds_r], SOL_SOCKET, SO_REUSEADDR, &opt, sizeof opt); + + if (bind (fds[*nfds_r], runp->ai_addr, runp->ai_addrlen) == -1) { + if (errno != EADDRINUSE) { + perror ("bind"); + return -1; + } + close (fds[*nfds_r]); + } + else { + if (listen (fds[*nfds_r], SOMAXCONN) == -1) { + perror ("listen"); + return -1; + } + ++*nfds_r; + } + runp = runp->ai_next; + } + + freeaddrinfo (ai); + + return 0; +} --- /tmp/newfile 2007-01-30 17:23:46.000000000 +0000 +++ src/remote_internal.c 2007-01-30 16:31:59.000000000 +0000 @@ -0,0 +1,663 @@ +/* + * remote_internal.c: driver to provide access to libvirtd running + * on a remote machine. + * + * Copyright (C) 2007 Red Hat, Inc. + * + * See COPYING.LIB for the license of this software. + * + * Richard Jones <rjones(a)redhat.com> + */ + +/* Architecture and notes: + * + * virConnectOpen ("remote:....") invokes this driver. Depending + * on the exact contents of the ellipsis "...." in the name string + * we will try some method to connect to a libvirtd daemon, running + * on a remote machine (or sometimes running on the local machine). + * + * All other vir* calls made on this connection are forwarded + * to the libvirtd daemon which carries out the requested action. + * So for example if you call virDomainCreateLinux, then the + * domain gets created on the remote machine, and virConnectListDomains + * lists domains running on the remote machine. + * + * Connections can be authenticated and encrypted -- it depends + * on the transport selected by the name string. + * + * The current implementation uses SunRPC layered over one of: + * - GnuTLS (an SSL/TLS library providing enterprise-level + * authentication and encryption) + * - a local Unix domain socket + * - ssh or another external program such as rsh + * - a plain TCP socket (unencrypted, not recommended for production) + * + * See http://et.redhat.com/~rjones/secure_rpc for an insight into + * the thinking that went into the selection of SunRPC. In + * the future we may use a different RPC system - for example + * XML-RPC would be a logical choice - so for now you should regard + * the protocol used as private and subject to change in future + * versions of libvirt without notice. + * + * The name string selects the transport to use and the type of + * virtualisation at the remote end. The general format is: + * + * "remote:<protocol>:<path> var=value var=value ..." + * + * Some examples: + * + * "remote:unix:/var/run/libvirtd/socket" + * "remote:tls:myxenserver" + * "remote:ssh:myserver name=qemud" + * "remote:ssh:myserver command=/opt/openssh/bin/ssh" + * + * The <protocol> is one of: tls, unix, ssh, ext or tcp. + * The <path> is protocol specific: + * + * Protocol Path-format + * ----------------------------------------- + * tls hostname[:port] + * unix Path to local socket + * ssh hostname[:port] + * ext Name or path of external program + * tcp hostname[:port] + * + * For tls, the default port is 16514. For tcp, the default port is + * 16509 (but note that tcp is almost never enabled because it is + * insecure - it's only there for testing). + * + * For ssh: The default port for ssh is 22. You should configure ssh + * so that it doesn't ask for a password (eg. using ssh-agent). The + * remote server should have a recent version of the the netcat program + * installed as 'nc', and the remote libvirtd must be configured to + * listen on a Unix domain socket. The following full command is run: + * ssh -p $port $hostname nc -U /var/run/libvirtd/socket + * + * For ext: Only the command you specify is run. It is up to you to + * write this command so that it somehow makes a connection to a + * remote libvirtd, and passes input and output over its stdin/stdout. + * + * The var=value pairs provide optional extra information: + * + * Variable Protocols Meaning + * ----------------------------------------- + * name (all) Name used in remote virConnectOpen + * (default is NULL). + * command ssh Name or path of external program (instead + * of "ssh"). + * + * The value is %-escaped (just like URL encoding), so if you want it + * to contain a literal space use "%20" or "+", if you want it to have + * a literal + character use "%2b", and for a literal % character use "%25". + * + * To provide some forwards compatibility, variables which are not + * understood are ignored (but a warning is printed on stderr). + * + * Several shorthand syntaxes are available: + * + * "remote:/var/run/libvirtd/socket" connect to Unix domain socket + * "remote://server" connect to TLS socket on server + * "remote://server:9000" connect to TLS server port 9000 + * + * For the details of the implementation of SunRPC over GnuTLS, etc. + * please see http://et.redhat.com/~rjones/secure_rpc which contains + * simple code samples which will allow you to understand what's + * going on here. + * + * - Richard Jones <rjones(a)redhat.com> + */ + +#define _GNU_SOURCE + +#include <stdio.h> +#include <stdlib.h> +#include <string.h> +#include <rpc/rpc.h> +#include <sys/socket.h> +#include <sys/un.h> +#include <netdb.h> +#include <assert.h> +#include <gnutls/gnutls.h> + +#include "sunrpc/clnt_ext.h" +#include "sunrpc/clnt_gnutls.h" +#include "sunrpc/clnt_tcp2.h" + +#include "internal.h" +#include "driver.h" +#include "remote_internal.h" +#include "remote_rpc.h" + +#define VERSION 1 /* Doesn't really mean anything. */ +#define DEBUG 1 /* Enable verbose messages on stderr. */ + +#define MAGIC 999 /* private_data->magic if OK */ +#define DEAD 998 /* private_data->magic if dead/closed */ + +struct private_data { + int magic; /* Should be MAGIC or DEAD. */ + int sock; /* Socket. */ + CLIENT *cl; /* SunRPC client. */ +}; + +#define GET_PRIVATE(conn,retcode) \ + struct private_data *private = (struct private_data *) (conn)->private; \ + assert (private); \ + if (private->magic == DEAD) { \ + error (conn, VIR_ERR_INVALID_ARG, \ + "tried to use a closed or uninitialised handle"); \ + return (retcode); \ + } \ + assert (private->magic == MAGIC) + +#define CAFILE "demoCA/cacert.pem" /* XXX */ + +static gnutls_certificate_credentials_t xcred; + +static void +initialise_gnutls (void) +{ + static int initialised = 0; + + if (initialised) return; + + gnutls_global_init (); + + /* X509 stuff */ + gnutls_certificate_allocate_credentials (&xcred); + + /* sets the trusted cas file + */ + gnutls_certificate_set_x509_trust_file (xcred, CAFILE, GNUTLS_X509_FMT_PEM); + + initialised = 1; +} + +enum protocol { + proto_unknown, + proto_tls, + proto_unix, + proto_ssh, + proto_ext, + proto_tcp +}; + +#define whitespace " \t\n" + +static char *unescape (const char *); +static int parse_hostname_port (const char *, char **hostname_r, + char **port_r, const char *default_port); +static void error (virConnectPtr conn, virErrorNumber code, const char *info); + +static int +remote_open (virConnectPtr conn, const char *name, + int flags __attribute__((unused))) +{ + int retcode = -1; /* Return code from this function. */ + enum protocol proto = proto_unknown; + char *path = 0; /* or hostname */ + char *port = 0; + char *remote_name = 0; /* Name to use at remote end of connection. */ + char *command = 0; /* External command. */ + char **cmd_argv = 0; /* External command argv[] array. */ + struct private_data private = { .magic = DEAD }; + /* Private data - copied to conn->private at the + * end of this function. + */ + + initialise_gnutls (); + + if (strncasecmp (name, "remote:", 7) != 0) + return -1; /* Not for me. */ + + /* Split the name at whitespace and parse it. */ + const char *p; + for (p = name; *p;) { + size_t len = strcspn (p, whitespace); + char *token = strndup (p, len); +#if DEBUG + fprintf (stderr, "token = %s\n", token); +#endif + if (p == name) { + /* First parameter is remote:protocol:path or a shortcut. */ + if (strncasecmp (token, "remote://", 9) == 0) { + proto = proto_tls; + if (parse_hostname_port (token+9, &path, &port, LIBVIRTD_GNUTLS_PORT) + == -1) { + error (conn, VIR_ERR_INVALID_ARG, + "remote_open: cannot parse port number"); + goto failed; + } + } else if (strncasecmp (token, "remote:/", 8) == 0) { + proto = proto_unix; + path = strdup (token+7); // include the initial slash + } else if (strncasecmp (token, "remote:tls:", 11) == 0) { + proto = proto_tls; + if (parse_hostname_port (token+11, &path, &port, LIBVIRTD_GNUTLS_PORT) + == -1) { + error (conn, VIR_ERR_INVALID_ARG, + "remote_open: cannot parse port number"); + goto failed; + } + } else if (strncasecmp (token, "remote:unix:", 12) == 0) { + proto = proto_unix; + path = strdup (token+12); + } else if (strncasecmp (token, "remote:ssh:", 11) == 0) { + proto = proto_ssh; + if (parse_hostname_port (token+11, &path, &port, "ssh") == -1) { + error (conn, VIR_ERR_INVALID_ARG, + "remote_open: cannot parse port number"); + goto failed; + } + command = strdup ("ssh"); + } else if (strncasecmp (token, "remote:ext:", 11) == 0) { + proto = proto_ext; + command = strdup (token+11); + } else if (strncasecmp (token, "remote:tcp:", 11) == 0) { + proto = proto_tcp; + if (parse_hostname_port (token+11, &path, &port, LIBVIRTD_TCP_PORT) + == -1) { + error (conn, VIR_ERR_INVALID_ARG, + "remote_open: cannot parse port number"); + goto failed; + } + } else { + error (conn, VIR_ERR_INVALID_ARG, + "remote_open: expecting 'remote:protocol:path' - please read the manual for the remote driver / libvirtd"); + goto failed; + } + } else { + /* Variable=value. Value is URL-escaped. */ + char *value = strchr (token, '='); + if (!value) { + error (conn, VIR_ERR_INVALID_ARG, + "remote_open: expecting 'variable=value'"); + goto failed; + } + *value++ = '\0'; /* So now token = variable, value = value. */ + /* Unescape value - this always makes a copy. */ + value = unescape (value); + + if (strcasecmp (token, "name") == 0) { + if (remote_name) { + free (remote_name); + fprintf (stderr, "libvir: warning: multiple name parameters. All but final will be ignored.\n"); + } + remote_name = value; + } else if (strcasecmp (token, "command") == 0) { + if (command) free (command); + command = value; + } else + /* For forwards compatibility, just warn about variables we + * don't understand. + */ + fprintf (stderr, "libvir: warning: variable '%s' ignored\n", token); + } + + free (token); + + /* Skip to next token. */ + p += len; + p += strspn (p, whitespace); + } + + /* Connect to the remote service. */ + switch (proto) { + case proto_unknown: + abort (); /* Internal error in this function. */ + + case proto_tls: + case proto_tcp: { + // http://people.redhat.com/drepper/userapi-ipv6.html + struct addrinfo *res, *r; + struct addrinfo hints; + memset (&hints, 0, sizeof hints); + hints.ai_socktype = SOCK_STREAM; + hints.ai_flags = AI_ADDRCONFIG; + int e = getaddrinfo (path, port, &hints, &res); + if (e != 0) { + error (conn, VIR_ERR_INVALID_ARG, gai_strerror (e)); + goto failed; + } + + /* Try to connect to each returned address in turn. */ + for (r = res; r; r = r->ai_next) { + private.sock = RPC_ANYSOCK; + private.cl = + proto == proto_tls + ? clntgnutls_create (r->ai_family, r->ai_addr, r->ai_addrlen, xcred, + LIBVIRTREMOTE, LIBVIRTREMOTE_VERS1, + &private.sock, 0, 0) + : clnttcp2_create (r->ai_family, r->ai_addr, r->ai_addrlen, + LIBVIRTREMOTE, LIBVIRTREMOTE_VERS1, + &private.sock, 0, 0); + if (private.cl) + goto tcp_connected; + } + + freeaddrinfo (res); + error (conn, VIR_ERR_RPC, + clnt_spcreateerror ("could not create SunRPC client")); + goto failed; + + tcp_connected: + freeaddrinfo (res); + + // NB. All versioning is done by SunRPC so we don't need to worry + // that we are connected to an incompatible daemon. + break; + } + + case proto_unix: { + // 108 is hard-coded into the header files as well. +#define UNIX_PATH_MAX 108 + struct sockaddr_un addr; + memset (&addr, 0, sizeof addr); + addr.sun_family = AF_UNIX; + strncpy (addr.sun_path, path, UNIX_PATH_MAX); + + private.sock = RPC_ANYSOCK; + private.cl = + clntunix_create (&addr, + LIBVIRTREMOTE, LIBVIRTREMOTE_VERS1, + &private.sock, 0, 0); + if (!private.cl) { + error (conn, VIR_ERR_RPC, + clnt_spcreateerror ("could not create SunRPC client")); + goto failed; + } + + break; + } + + case proto_ssh: + // Generate the final command argv[] array. + // ssh -p $port $hostname nc -U $socket [NULL] + cmd_argv = malloc (8 * sizeof (char *)); + cmd_argv[0] = strdup (command); + cmd_argv[1] = strdup ("-p"); + cmd_argv[2] = strdup (port); + cmd_argv[3] = strdup (path); + cmd_argv[4] = strdup ("nc"); + cmd_argv[5] = strdup ("-U"); + cmd_argv[6] = strdup (LIBVIRTD_UNIX_SOCKET); + cmd_argv[7] = 0; + /*FALLTHROUGH*/ + case proto_ext: + private.sock = RPC_ANYSOCK; + private.cl = + clntext_create (command, cmd_argv, + LIBVIRTREMOTE, LIBVIRTREMOTE_VERS1, + &private.sock, 0, 0); + if (!private.cl) { + error (conn, VIR_ERR_RPC, + clnt_spcreateerror ("could not create SunRPC client")); + goto failed; + } + } + + // Send name (make the actual driver open RPC). + int *retcode_ptr = remote_rpc_open_1 (&remote_name, private.cl); + if (!retcode_ptr || *retcode_ptr == -1) { + error (conn, VIR_ERR_RPC, + clnt_sperror (private.cl, "remote_rpc_open")); + clnt_destroy (private.cl); + goto failed; + } + + conn->private = malloc (sizeof private); + if (!conn->private) { + error (conn, VIR_ERR_NO_MEMORY, "malloc"); + clnt_destroy (private.cl); + goto failed; + } + private.magic = MAGIC; + memcpy (conn->private, &private, sizeof private); + + retcode = 0; /* Success. */ + /*FALLTHROUGH*/ + + failed: + if (path) free (path); + if (port) free (port); + if (remote_name) free (remote_name); + if (command) free (command); + if (cmd_argv) { + char **a = cmd_argv; + while (*a) { free (*a); a++; } + free (cmd_argv); + } + + return retcode; +} + +/* Un-%-escape the argument string. Note that this always makes + * a copy, and that is intentional. + */ +static inline int +xdigit (char c) +{ + switch (c) { + case '0'...'9': return c - '0'; + case 'a'...'f': return c - 'a' + 10; + case 'A'...'F': return c - 'A' + 10; + default: return -1; + } +} + +static char * +unescape (const char *str) +{ + // Returned string will always be same length or shorter than input. + int n = strlen (str); + + char *ret = malloc (n+1); + if (ret == 0) return 0; // although _I_ think we should abort(). + + int i; + char *p = ret; + for (i = 0; i < n; ++i) { + if (str[i] == '+') { + *p++ = ' '; + } else if (str[i] == '%') { + if (i+2 < n) { + int c1 = xdigit (str[i+1]), c2 = xdigit (str[i+2]); + if (c1 == -1 || c2 == -1) { + fprintf (stderr, "remote_open: incorrect %%-hex sequence in name\n"); + return 0; + } + *p++ = xdigit (c1) << 4 | xdigit (c2); + } else { + fprintf (stderr, "remote_open: short %%-hex sequence in name\n"); + return 0; + } + } else + *p++ = str[i]; + } + *p = '\0'; + return ret; +} + +/* Parse a string which may be either "hostname" or "hostname:port". + * The hostname may contain colons (eg. if it's an IPv6 name). + * Note that this always makes a copy of the hostname and port number, + * and that is intentional. + */ +static int +parse_hostname_port (const char *str, char **hostname_r, + char **port_r, const char *default_port) +{ + char *p = strrchr (str, ':'); + if (p) { + *hostname_r = strndup (str, p-str); + *port_r = strdup (p+1); + } else { + // No :port, so just copy the hostname. + *hostname_r = strdup (str); + *port_r = strdup (default_port); + } + return 0; +} + +static int +remote_close (virConnectPtr conn) +{ + GET_PRIVATE (conn, -1); + + int *retcode = remote_rpc_close_1 (NULL, private->cl); + if (retcode == 0) { + error (conn, VIR_ERR_RPC, + clnt_sperror (private->cl, "remote_rpc_open")); + return -1; + } + if (*retcode == -1) return -1; + + // XXX freeres + + // NB. clnt_destroy should close the socket (private->sock) too. + clnt_destroy (private->cl); + // Force errors if anyone tries to reuse the closed connection. + private->magic = DEAD; + + return *retcode; +} + +/* Remote_open and remote_close functions above are the complex ones. The + * rest just shuffle arguments and pass them along to the remote libvirtd. + */ + +// Should we return our local type (ie. "remote"), or the type +// of the remote HV, or the type of the remote HV + some flag? +// I took the view that we should just transparently shuffle. +static const char * +remote_type (virConnectPtr conn) +{ + GET_PRIVATE (conn, NULL); + + char **type = remote_rpc_type_1 (NULL, private->cl); + if (type == 0) { + error (conn, VIR_ERR_RPC, + clnt_sperror (private->cl, "remote_rpc_open")); + return NULL; + } + + // XXX freeres + + return *type; +} + +static int +remote_version (virConnectPtr conn, unsigned long *hvVer) +{ + GET_PRIVATE (conn, -1); + + struct version_ret *ret = remote_rpc_version_1 (NULL, private->cl); + if (ret == 0) { + error (conn, VIR_ERR_RPC, + clnt_sperror (private->cl, "remote_rpc_open")); + return -1; + } + + // XXX freeres + + if (ret->retcode == -1) return -1; + *hvVer = ret->hvVer; + return ret->retcode; +} + +/* + .nodeGetInfo = remote_nodeGetInfo, + .listDomains = remote_listDomains, + .numOfDomains = remote_numOfDomains, + .domainCreateLinux = remote_domainCreateLinux, + .domainLookupByID = remote_domainLookupByID, + .domainLookupByUUID = remote_domainLookupByUUID, + .domainLookupByName = remote_domainLookupByName, + .domainSuspend = remote_domainSuspend, + .domainResume = remote_domainResume, + .domainShutdown = remote_domainShutdown, + .domainReboot = remote_domainReboot, + .domainDestroy = remote_domainDestroy, + .domainGetOSType = remote_domainGetOSType, + .domainGetMaxMemory = remote_domainGetMaxMemory, + .domainSetMaxMemory = remote_domainSetMaxMemory, + .domainSetMemory = remote_domainSetMemory, + .domainGetInfo = remote_domainGetInfo, + .domainSave = remote_domainSave, + .domainRestore = remote_domainRestore, + .domainCoreDump = remote_domainCoreDump, + .domainSetVcpus = remote_domainSetVcpus, + .domainPinVcpu = remote_domainPinVcpu, + .domainGetVcpus = remote_domainGetVcpus, + .domainDumpXML = remote_domainDumpXML, + .listDefinedDomains = remote_listDefinedDomains, + .numOfDefinedDomains = remote_numOfDefinedDomains, + .domainCreate = remote_domainCreate, + .domainDefineXML = remote_domainDefineXML, + .domainUndefine = remote_domainUndefine, + .domainAttachDevice = remote_domainAttachDevice, + .domainDetachDevice = remote_domainDetachDevice, +*/ + +/* Error handling. This error handling is on crack. */ +static void +error (virConnectPtr conn, virErrorNumber code, const char *info) +{ + const char *errmsg; + + errmsg = __virErrorMsg (code, info); + __virRaiseError (conn, NULL, VIR_FROM_REMOTE, + code, VIR_ERR_ERROR, errmsg, info, NULL, 0, 0, + errmsg, info); +} + +static virDriver driver = { + .no = VIR_DRV_REMOTE, + .name = "remote", + .ver = VERSION, + //.init = remote_init, + .open = remote_open, + .close = remote_close, + .type = remote_type, + .version = remote_version, +#if 0 + .nodeGetInfo = remote_nodeGetInfo, + .listDomains = remote_listDomains, + .numOfDomains = remote_numOfDomains, + .domainCreateLinux = remote_domainCreateLinux, + .domainLookupByID = remote_domainLookupByID, + .domainLookupByUUID = remote_domainLookupByUUID, + .domainLookupByName = remote_domainLookupByName, + .domainSuspend = remote_domainSuspend, + .domainResume = remote_domainResume, + .domainShutdown = remote_domainShutdown, + .domainReboot = remote_domainReboot, + .domainDestroy = remote_domainDestroy, + .domainGetOSType = remote_domainGetOSType, + .domainGetMaxMemory = remote_domainGetMaxMemory, + .domainSetMaxMemory = remote_domainSetMaxMemory, + .domainSetMemory = remote_domainSetMemory, + .domainGetInfo = remote_domainGetInfo, + .domainSave = remote_domainSave, + .domainRestore = remote_domainRestore, + .domainCoreDump = remote_domainCoreDump, + .domainSetVcpus = remote_domainSetVcpus, + .domainPinVcpu = remote_domainPinVcpu, + .domainGetVcpus = remote_domainGetVcpus, + .domainDumpXML = remote_domainDumpXML, + .listDefinedDomains = remote_listDefinedDomains, + .numOfDefinedDomains = remote_numOfDefinedDomains, + .domainCreate = remote_domainCreate, + .domainDefineXML = remote_domainDefineXML, + .domainUndefine = remote_domainUndefine, + .domainAttachDevice = remote_domainAttachDevice, + .domainDetachDevice = remote_domainDetachDevice, +#endif +}; + +/* Register driver. */ +void +remoteRegister (void) +{ + virRegisterDriver (&driver); +} --- /tmp/newfile 2007-01-30 17:23:46.000000000 +0000 +++ src/remote_internal.h 2007-01-30 16:21:19.000000000 +0000 @@ -0,0 +1,34 @@ +/* + * remote_internal.h: driver to provide access to libvirtd running + * on a remote machine. + * + * Copyright (C) 2007 Red Hat, Inc. + * + * See COPYING.LIB for the license of this software. + * + * Richard Jones <rjones(a)redhat.com> + */ + +#ifndef __VIR_REMOTE_INTERNAL_H__ +#define __VIR_REMOTE_INTERNAL_H__ + +#ifdef __cplusplus +extern "C" { +#endif + +void remoteRegister (void); + +/* The port numbers are strings because you can also use + * service names here. + */ +#define LIBVIRTD_GNUTLS_PORT "16514" +#define LIBVIRTD_TCP_PORT "16509" + //#define LIBVIRTD_UNIX_SOCKET "/var/run/libvirtd/socket" +#define LIBVIRTD_UNIX_SOCKET "/tmp/socket" // Just for testing + +#ifdef __cplusplus +} +#endif + + +#endif /* __VIR_REMOTE_INTERNAL_H__ */ --- /tmp/newfile 2007-01-30 17:23:46.000000000 +0000 +++ src/remote_rpc.x 2007-01-30 13:32:56.000000000 +0000 @@ -0,0 +1,38 @@ +/* -*- C -*- + * remote_rpc.x: Remote procedure call interface for the remote driver. + * Process this file with rpcgen to generate client and server stubs. + * + * Copyright (C) 2007 Red Hat, Inc. + * + * See COPYING.LIB for the license of this software. + * + * Richard Jones <rjones(a)redhat.com> + */ + +/* Structure returned from remote_rpc_version. */ +struct version_ret { + int retcode; + long hvVer; +}; + +program LIBVIRTREMOTE { + version LIBVIRTREMOTE_VERS1 { + /* XXX The open interface should return a cookie to represent + * the virConnectPtr on the server side. Note also that SunRPC + * is stateless so it's unclear when cookies can be garbage + * collected. + */ + int remote_rpc_open (string) = 1; + int remote_rpc_close (void) = 2; + string remote_rpc_type (void) = 3; + version_ret remote_rpc_version (void) = 4; + + /* etc */ + + } = 1; + /* It doesn't really matter what program number we choose here because + * there will only ever be one "program" listening on the assigned + * TCP port number. Nevertheless, choose one from the Sun private + * space. + */ +} = 0x20008080; --- /tmp/newfile 2007-01-30 17:23:46.000000000 +0000 +++ src/sunrpc/README 2007-01-30 16:15:43.000000000 +0000 @@ -0,0 +1,25 @@ +This directory contains modified SunRPC transports. They are based on +the standard transports from glibc 2.5 (in the sunrpc/ directory +there). + +clnt_ext.c + - Fork an external program, eg. ssh. + (Original: clnt_unix.c) + +clnt_gnutls.c + - Modified for IPv6 and GnuTLS support. + (Original: clnt_tcp.c) + +clnt_tcp2.c + - Modified for IPv6 support. + (Original: clnt_tcp.c) + +svc_gnutls.c + - Modified for IPv6 and GnuTLS support. + (Original: svc_tcp.c) + +svc_tcp2.c + - Modified for IPv6 support. + (Original: svc_tcp.c) + +$Id$ \ No newline at end of file --- /tmp/newfile 2007-01-30 17:23:46.000000000 +0000 +++ src/sunrpc/clnt_ext.c 2007-01-30 11:48:09.000000000 +0000 @@ -0,0 +1,645 @@ +/* + * clnt_ext.c: SunRPC over an external program. This is a modified version + * of clnt_unix.c from glibc which is written to run over a forked + * external program. + * + * Note that there is no corresponding svc_ext.c. It is expected that + * this client will talk to a remote Unix domain socket (ie. svc_unix.c). + * + * Modifications from glibc-2.5 base by Richard Jones <rjones(a)redhat.com>. + */ + +/* + * Sun RPC is a product of Sun Microsystems, Inc. and is provided for + * unrestricted use provided that this legend is included on all tape + * media and as a part of the software program in whole or part. Users + * may copy or modify Sun RPC without charge, but are not authorized + * to license or distribute it to anyone else except as part of a product or + * program developed by the user. + * + * SUN RPC IS PROVIDED AS IS WITH NO WARRANTIES OF ANY KIND INCLUDING THE + * WARRANTIES OF DESIGN, MERCHANTIBILITY AND FITNESS FOR A PARTICULAR + * PURPOSE, OR ARISING FROM A COURSE OF DEALING, USAGE OR TRADE PRACTICE. + * + * Sun RPC is provided with no support and without any obligation on the + * part of Sun Microsystems, Inc. to assist in its use, correction, + * modification or enhancement. + * + * SUN MICROSYSTEMS, INC. SHALL HAVE NO LIABILITY WITH RESPECT TO THE + * INFRINGEMENT OF COPYRIGHTS, TRADE SECRETS OR ANY PATENTS BY SUN RPC + * OR ANY PART THEREOF. + * + * In no event will Sun Microsystems, Inc. be liable for any lost revenue + * or profits or other special, indirect and consequential damages, even if + * Sun has been advised of the possibility of such damages. + * + * Sun Microsystems, Inc. + * 2550 Garcia Avenue + * Mountain View, California 94043 + */ + +/* + * clnt_unix.c, Implements a TCP/IP based, client side RPC. + * + * Copyright (C) 1984, Sun Microsystems, Inc. + * + * TCP based RPC supports 'batched calls'. + * A sequence of calls may be batched-up in a send buffer. The rpc call + * return immediately to the client even though the call was not necessarily + * sent. The batching occurs if the results' xdr routine is NULL (0) AND + * the rpc timeout value is zero (see clnt.h, rpc). + * + * Clients should NOT casually batch calls that in fact return results; that is, + * the server side should be aware that a call is batched and not produce any + * return message. Batched calls that produce many result messages can + * deadlock (netlock) the client and the server.... + * + * Now go hang yourself. + */ + +#include <netdb.h> +#include <errno.h> +#include <stdio.h> +#include <unistd.h> +#include <libintl.h> +#include <rpc/rpc.h> +#include <sys/uio.h> +#include <sys/poll.h> +#include <sys/socket.h> +#include <sys/types.h> +#include <sys/wait.h> +#include <rpc/pmap_clnt.h> +#ifdef USE_IN_LIBIO +# include <wchar.h> +#endif + +#include "clnt_ext.h" + +extern u_long _create_xid (void); + +#define MCALL_MSG_SIZE 24 + +struct ct_data + { + int ct_sock; + bool_t ct_closeit; + struct timeval ct_wait; + bool_t ct_waitset; /* wait set by clnt_control? */ + struct rpc_err ct_error; + char ct_mcall[MCALL_MSG_SIZE]; /* marshalled callmsg */ + u_int ct_mpos; /* pos after marshal */ + XDR ct_xdrs; + int ct_pid; /* Child PID. */ + }; + +static int readunix (char *, char *, int); +static int writeunix (char *, char *, int); + +static enum clnt_stat clntunix_call (CLIENT *, u_long, xdrproc_t, caddr_t, + xdrproc_t, caddr_t, struct timeval); +static void clntunix_abort (void); +static void clntunix_geterr (CLIENT *, struct rpc_err *); +static bool_t clntunix_freeres (CLIENT *, xdrproc_t, caddr_t); +static bool_t clntunix_control (CLIENT *, int, char *); +static void clntunix_destroy (CLIENT *); + +static const struct clnt_ops unix_ops = +{ + clntunix_call, + clntunix_abort, + clntunix_geterr, + clntunix_freeres, + clntunix_destroy, + clntunix_control +}; + +CLIENT * +clntext_create (char *filename, char *argv[], + u_long prog, u_long vers, + int *sockp, u_int sendsz, u_int recvsz) +{ + CLIENT *h; + struct ct_data *ct = (struct ct_data *) mem_alloc (sizeof (*ct)); + struct rpc_msg call_msg; + int len, sv[2]; + + h = (CLIENT *) mem_alloc (sizeof (*h)); + if (h == NULL || ct == NULL) + { + struct rpc_createerr *ce = &get_rpc_createerr (); + (void) fprintf (stderr, "clntext_create: out of memory\n"); + ce->cf_stat = RPC_SYSTEMERROR; + ce->cf_error.re_errno = ENOMEM; + goto fooy; + } + + /* Fork off the external process. Use socketpair to create a private + * (unnamed) Unix domain socket to the child process so we don't have + * to faff around with two file descriptors (a la 'pipe(2)'). + */ + if (*sockp < 0) + { + if (socketpair (PF_UNIX, SOCK_STREAM, 0, sv) == -1) + { + struct rpc_createerr *ce = &get_rpc_createerr (); + (void) fprintf (stderr, "clntext_create: socketpair\n"); + ce->cf_stat = RPC_SYSTEMERROR; + ce->cf_error.re_errno = errno; + goto fooy; + } + + ct->ct_pid = fork (); + if (ct->ct_pid == -1) { + struct rpc_createerr *ce = &get_rpc_createerr (); + (void) fprintf (stderr, "clntext_create: fork\n"); + ce->cf_stat = RPC_SYSTEMERROR; + ce->cf_error.re_errno = errno; + goto fooy; + } else if (ct->ct_pid == 0) { /* Child. */ + close (sv[0]); + // Connect socket (sv[1]) to stdin/stdout. + close (0); + dup (sv[1]); + close (1); + dup (sv[1]); + close (sv[1]); + + // Run the external process. + + if (!argv) { + argv = malloc (2 * sizeof (char *)); + argv[0] = filename; + argv[1] = 0; + } + execvp (filename, argv); + perror (filename); + _exit (1); + } + + /* Parent continues here. */ + close (sv[1]); + *sockp = sv[0]; + ct->ct_closeit = TRUE; + } + else + { + ct->ct_closeit = FALSE; + } + + /* + * Set up private data struct + */ + ct->ct_sock = *sockp; + ct->ct_wait.tv_usec = 0; + ct->ct_waitset = FALSE; + + /* + * Initialize call message + */ + call_msg.rm_xid = _create_xid (); + call_msg.rm_direction = CALL; + call_msg.rm_call.cb_rpcvers = RPC_MSG_VERSION; + call_msg.rm_call.cb_prog = prog; + call_msg.rm_call.cb_vers = vers; + + /* + * pre-serialize the static part of the call msg and stash it away + */ + xdrmem_create (&(ct->ct_xdrs), ct->ct_mcall, MCALL_MSG_SIZE, + XDR_ENCODE); + if (!xdr_callhdr (&(ct->ct_xdrs), &call_msg)) + { + if (ct->ct_closeit) + close (*sockp); + goto fooy; + } + ct->ct_mpos = XDR_GETPOS (&(ct->ct_xdrs)); + XDR_DESTROY (&(ct->ct_xdrs)); + + /* + * Create a client handle which uses xdrrec for serialization + * and authnone for authentication. + */ + xdrrec_create (&(ct->ct_xdrs), sendsz, recvsz, + (caddr_t) ct, readunix, writeunix); + h->cl_ops = (struct clnt_ops *) &unix_ops; + h->cl_private = (caddr_t) ct; + h->cl_auth = authnone_create (); + return h; + +fooy: + /* + * Something goofed, free stuff and barf + */ + mem_free ((caddr_t) ct, sizeof (struct ct_data)); + mem_free ((caddr_t) h, sizeof (CLIENT)); + return (CLIENT *) NULL; +} + +static enum clnt_stat +clntunix_call (h, proc, xdr_args, args_ptr, xdr_results, results_ptr, timeout) + CLIENT *h; + u_long proc; + xdrproc_t xdr_args; + caddr_t args_ptr; + xdrproc_t xdr_results; + caddr_t results_ptr; + struct timeval timeout; +{ + struct ct_data *ct = (struct ct_data *) h->cl_private; + XDR *xdrs = &(ct->ct_xdrs); + struct rpc_msg reply_msg; + u_long x_id; + u_int32_t *msg_x_id = (u_int32_t *) (ct->ct_mcall); /* yuk */ + bool_t shipnow; + int refreshes = 2; + + if (!ct->ct_waitset) + { + ct->ct_wait = timeout; + } + + shipnow = + (xdr_results == (xdrproc_t) 0 && ct->ct_wait.tv_sec == 0 + && ct->ct_wait.tv_usec == 0) ? FALSE : TRUE; + +call_again: + xdrs->x_op = XDR_ENCODE; + ct->ct_error.re_status = RPC_SUCCESS; + x_id = ntohl (--(*msg_x_id)); + if ((!XDR_PUTBYTES (xdrs, ct->ct_mcall, ct->ct_mpos)) || + (!XDR_PUTLONG (xdrs, (long *) &proc)) || + (!AUTH_MARSHALL (h->cl_auth, xdrs)) || + (!(*xdr_args) (xdrs, args_ptr))) + { + if (ct->ct_error.re_status == RPC_SUCCESS) + ct->ct_error.re_status = RPC_CANTENCODEARGS; + (void) xdrrec_endofrecord (xdrs, TRUE); + return ct->ct_error.re_status; + } + if (!xdrrec_endofrecord (xdrs, shipnow)) + return ct->ct_error.re_status = RPC_CANTSEND; + if (!shipnow) + return RPC_SUCCESS; + /* + * Hack to provide rpc-based message passing + */ + if (ct->ct_wait.tv_sec == 0 && ct->ct_wait.tv_usec == 0) + return ct->ct_error.re_status = RPC_TIMEDOUT; + + + /* + * Keep receiving until we get a valid transaction id + */ + xdrs->x_op = XDR_DECODE; + while (TRUE) + { + reply_msg.acpted_rply.ar_verf = _null_auth; + reply_msg.acpted_rply.ar_results.where = NULL; + reply_msg.acpted_rply.ar_results.proc = (xdrproc_t)xdr_void; + if (!xdrrec_skiprecord (xdrs)) + return ct->ct_error.re_status; + /* now decode and validate the response header */ + if (!xdr_replymsg (xdrs, &reply_msg)) + { + if (ct->ct_error.re_status == RPC_SUCCESS) + continue; + return ct->ct_error.re_status; + } + if (reply_msg.rm_xid == x_id) + break; + } + + /* + * process header + */ + _seterr_reply (&reply_msg, &(ct->ct_error)); + if (ct->ct_error.re_status == RPC_SUCCESS) + { + if (!AUTH_VALIDATE (h->cl_auth, &reply_msg.acpted_rply.ar_verf)) + { + ct->ct_error.re_status = RPC_AUTHERROR; + ct->ct_error.re_why = AUTH_INVALIDRESP; + } + else if (!(*xdr_results) (xdrs, results_ptr)) + { + if (ct->ct_error.re_status == RPC_SUCCESS) + ct->ct_error.re_status = RPC_CANTDECODERES; + } + /* free verifier ... */ + if (reply_msg.acpted_rply.ar_verf.oa_base != NULL) + { + xdrs->x_op = XDR_FREE; + (void) xdr_opaque_auth (xdrs, + &(reply_msg.acpted_rply.ar_verf)); + } + } /* end successful completion */ + else + { + /* maybe our credentials need to be refreshed ... */ + if (refreshes-- && AUTH_REFRESH (h->cl_auth)) + goto call_again; + } /* end of unsuccessful completion */ + return ct->ct_error.re_status; +} + +static void +clntunix_geterr (CLIENT *h, struct rpc_err *errp) +{ + struct ct_data *ct = (struct ct_data *) h->cl_private; + + *errp = ct->ct_error; +} + +static bool_t +clntunix_freeres (cl, xdr_res, res_ptr) + CLIENT *cl; + xdrproc_t xdr_res; + caddr_t res_ptr; +{ + struct ct_data *ct = (struct ct_data *) cl->cl_private; + XDR *xdrs = &(ct->ct_xdrs); + + xdrs->x_op = XDR_FREE; + return (*xdr_res) (xdrs, res_ptr); +} + +static void +clntunix_abort () +{ +} + +static bool_t +clntunix_control (CLIENT *cl, int request, char *info) +{ + struct ct_data *ct = (struct ct_data *) cl->cl_private; + + + switch (request) + { + case CLSET_FD_CLOSE: + ct->ct_closeit = TRUE; + break; + case CLSET_FD_NCLOSE: + ct->ct_closeit = FALSE; + break; + case CLSET_TIMEOUT: + ct->ct_wait = *(struct timeval *) info; + break; + case CLGET_TIMEOUT: + *(struct timeval *) info = ct->ct_wait; + break; + case CLGET_FD: + *(int *)info = ct->ct_sock; + break; + case CLGET_XID: + /* + * use the knowledge that xid is the + * first element in the call structure *. + * This will get the xid of the PREVIOUS call + */ + *(u_long *) info = ntohl (*(u_long *)ct->ct_mcall); + break; + case CLSET_XID: + /* This will set the xid of the NEXT call */ + *(u_long *) ct->ct_mcall = htonl (*(u_long *)info - 1); + /* decrement by 1 as clntunix_call() increments once */ + case CLGET_VERS: + /* + * This RELIES on the information that, in the call body, + * the version number field is the fifth field from the + * begining of the RPC header. MUST be changed if the + * call_struct is changed + */ + *(u_long *) info = ntohl (*(u_long *) (ct->ct_mcall + + 4 * BYTES_PER_XDR_UNIT)); + break; + case CLSET_VERS: + *(u_long *) (ct->ct_mcall + 4 * BYTES_PER_XDR_UNIT) + = htonl (*(u_long *) info); + break; + case CLGET_PROG: + /* + * This RELIES on the information that, in the call body, + * the program number field is the field from the + * begining of the RPC header. MUST be changed if the + * call_struct is changed + */ + *(u_long *) info = ntohl (*(u_long *) (ct->ct_mcall + + 3 * BYTES_PER_XDR_UNIT)); + break; + case CLSET_PROG: + *(u_long *) (ct->ct_mcall + 3 * BYTES_PER_XDR_UNIT) + = htonl(*(u_long *) info); + break; + /* The following are only possible with TI-RPC */ + case CLGET_RETRY_TIMEOUT: + case CLSET_RETRY_TIMEOUT: + case CLGET_SVC_ADDR: + case CLSET_SVC_ADDR: + case CLSET_PUSH_TIMOD: + case CLSET_POP_TIMOD: + default: + return FALSE; + } + return TRUE; +} + + +static void +clntunix_destroy (CLIENT *h) +{ + struct ct_data *ct = + (struct ct_data *) h->cl_private; + + if (ct->ct_closeit) + { + (void) close (ct->ct_sock); + } + XDR_DESTROY (&(ct->ct_xdrs)); + mem_free ((caddr_t) ct, sizeof (struct ct_data)); + mem_free ((caddr_t) h, sizeof (CLIENT)); + + // Wait for child to finish. Print any errors but don't stop. + int status; + if (waitpid (ct->ct_pid, &status, 0) == -1) { + perror ("waitpid"); + return; + } + if (WIFEXITED (status)) { + int s = WEXITSTATUS (status); + if (s != 0) + fprintf (stderr, "clntext_destroy: warning: external command exited with non-zero exit status %d\n", s); + } else if (WIFSIGNALED (status)) { + int s = WTERMSIG (status); + fprintf (stderr, "clntext_destroy: warning: external command died on signal %d\n", s); + } else if (WIFSTOPPED (status)) { + int s = WSTOPSIG (status); + fprintf (stderr, "clntext_destroy: warning: external command stopped on signal %d\n", s); + } +} + +static int +__msgread (int sock, void *data, size_t cnt) +{ + struct iovec iov; + struct msghdr msg; +#ifdef SCM_CREDENTIALS + static char cm[CMSG_SPACE(sizeof (struct ucred))]; +#endif + int len; + + iov.iov_base = data; + iov.iov_len = cnt; + + msg.msg_iov = &iov; + msg.msg_iovlen = 1; + msg.msg_name = NULL; + msg.msg_namelen = 0; +#ifdef SCM_CREDENTIALS + msg.msg_control = (caddr_t) &cm; + msg.msg_controllen = CMSG_SPACE(sizeof (struct ucred)); +#endif + msg.msg_flags = 0; + +#ifdef SO_PASSCRED + { + int on = 1; + if (setsockopt (sock, SOL_SOCKET, SO_PASSCRED, &on, sizeof (on))) + return -1; + } +#endif + + restart: + len = recvmsg (sock, &msg, 0); + if (len >= 0) + { + if (msg.msg_flags & MSG_CTRUNC || len == 0) + return 0; + else + return len; + } + if (errno == EINTR) + goto restart; + return -1; +} + +static int +__msgwrite (int sock, void *data, size_t cnt) +{ +#ifndef SCM_CREDENTIALS + /* We cannot implement this reliably. */ + __set_errno (ENOSYS); + return -1; +#else + struct iovec iov; + struct msghdr msg; + struct cmsghdr *cmsg = alloca (CMSG_SPACE(sizeof (struct ucred))); + struct ucred cred; + int len; + + /* XXX I'm not sure, if gete?id() is always correct, or if we should use + get?id(). But since keyserv needs geteuid(), we have no other chance. + It would be much better, if the kernel could pass both to the server. */ + cred.pid = getpid (); + cred.uid = geteuid (); + cred.gid = getegid (); + + memcpy (CMSG_DATA(cmsg), &cred, sizeof (struct ucred)); + cmsg->cmsg_level = SOL_SOCKET; + cmsg->cmsg_type = SCM_CREDENTIALS; + cmsg->cmsg_len = sizeof(*cmsg) + sizeof(struct ucred); + + iov.iov_base = data; + iov.iov_len = cnt; + + msg.msg_iov = &iov; + msg.msg_iovlen = 1; + msg.msg_name = NULL; + msg.msg_namelen = 0; + msg.msg_control = cmsg; + msg.msg_controllen = CMSG_ALIGN(cmsg->cmsg_len); + msg.msg_flags = 0; + + restart: + len = sendmsg (sock, &msg, 0); + if (len >= 0) + return len; + if (errno == EINTR) + goto restart; + return -1; + +#endif +} + + +/* + * Interface between xdr serializer and unix connection. + * Behaves like the system calls, read & write, but keeps some error state + * around for the rpc level. + */ +static int +readunix (char *ctptr, char *buf, int len) +{ + struct ct_data *ct = (struct ct_data *) ctptr; + struct pollfd fd; + int milliseconds = ((ct->ct_wait.tv_sec * 1000) + + (ct->ct_wait.tv_usec / 1000)); + + if (len == 0) + return 0; + + fd.fd = ct->ct_sock; + fd.events = POLLIN; + while (TRUE) + { + switch (poll (&fd, 1, milliseconds)) + { + case 0: + ct->ct_error.re_status = RPC_TIMEDOUT; + return -1; + + case -1: + if (errno == EINTR) + continue; + ct->ct_error.re_status = RPC_CANTRECV; + ct->ct_error.re_errno = errno; + return -1; + } + break; + } + switch (len = __msgread (ct->ct_sock, buf, len)) + { + + case 0: + /* premature eof */ + ct->ct_error.re_errno = ECONNRESET; + ct->ct_error.re_status = RPC_CANTRECV; + len = -1; /* it's really an error */ + break; + + case -1: + ct->ct_error.re_errno = errno; + ct->ct_error.re_status = RPC_CANTRECV; + break; + } + return len; +} + +static int +writeunix (char *ctptr, char *buf, int len) +{ + int i, cnt; + struct ct_data *ct = (struct ct_data *) ctptr; + + for (cnt = len; cnt > 0; cnt -= i, buf += i) + { + if ((i = __msgwrite (ct->ct_sock, buf, cnt)) == -1) + { + ct->ct_error.re_errno = errno; + ct->ct_error.re_status = RPC_CANTSEND; + return -1; + } + } + return len; +} --- /tmp/newfile 2007-01-30 17:23:46.000000000 +0000 +++ src/sunrpc/clnt_ext.h 2007-01-30 11:47:24.000000000 +0000 @@ -0,0 +1,18 @@ +/* + * clnt_ext.h: Interface to the SunRPC over external program. + * + * Copyright (C) 2007 Red Hat, Inc. + * + * See COPYING.LIB for the license of this software. + * + * Richard Jones <rjones(a)redhat.com> + */ + +#ifndef __CLNT_EXT_H__ +#define __CLNT_EXT_H__ + +extern CLIENT *clntext_create (char *filename, char *argv[], + u_long prog, u_long vers, + int *sockp, u_int sendsz, u_int recvsz); + +#endif /* __CLNT_EXT_H__ */ --- /tmp/newfile 2007-01-30 17:23:46.000000000 +0000 +++ src/sunrpc/clnt_gnutls.c 2007-01-30 17:03:31.000000000 +0000 @@ -0,0 +1,603 @@ +/* + * clnt_gnutls.c: SunRPC over GnuTLS client. This is a modified version + * of clnt_tcp.c from glibc which is written to run over GnuTLS. + * + * Modifications from glibc-2.5 base by Richard Jones <rjones(a)redhat.com>. + */ + +/* @(#)clnt_tcp.c 2.2 88/08/01 4.0 RPCSRC */ +/* + * Sun RPC is a product of Sun Microsystems, Inc. and is provided for + * unrestricted use provided that this legend is included on all tape + * media and as a part of the software program in whole or part. Users + * may copy or modify Sun RPC without charge, but are not authorized + * to license or distribute it to anyone else except as part of a product or + * program developed by the user. + * + * SUN RPC IS PROVIDED AS IS WITH NO WARRANTIES OF ANY KIND INCLUDING THE + * WARRANTIES OF DESIGN, MERCHANTIBILITY AND FITNESS FOR A PARTICULAR + * PURPOSE, OR ARISING FROM A COURSE OF DEALING, USAGE OR TRADE PRACTICE. + * + * Sun RPC is provided with no support and without any obligation on the + * part of Sun Microsystems, Inc. to assist in its use, correction, + * modification or enhancement. + * + * SUN MICROSYSTEMS, INC. SHALL HAVE NO LIABILITY WITH RESPECT TO THE + * INFRINGEMENT OF COPYRIGHTS, TRADE SECRETS OR ANY PATENTS BY SUN RPC + * OR ANY PART THEREOF. + * + * In no event will Sun Microsystems, Inc. be liable for any lost revenue + * or profits or other special, indirect and consequential damages, even if + * Sun has been advised of the possibility of such damages. + * + * Sun Microsystems, Inc. + * 2550 Garcia Avenue + * Mountain View, California 94043 + */ +#if !defined(lint) && defined(SCCSIDS) +static char sccsid[] = "@(#)clnt_tcp.c 1.37 87/10/05 Copyr 1984 Sun Micro"; +#endif + +/* + * clnt_tcp.c, Implements a TCP/IP based, client side RPC. + * + * Copyright (C) 1984, Sun Microsystems, Inc. + * + * TCP based RPC supports 'batched calls'. + * A sequence of calls may be batched-up in a send buffer. The rpc call + * return immediately to the client even though the call was not necessarily + * sent. The batching occurs if the results' xdr routine is NULL (0) AND + * the rpc timeout value is zero (see clnt.h, rpc). + * + * Clients should NOT casually batch calls that in fact return results; that is, + * the server side should be aware that a call is batched and not produce any + * return message. Batched calls that produce many result messages can + * deadlock (netlock) the client and the server.... + * + * Now go hang yourself. + */ + +#include <netdb.h> +#include <errno.h> +#include <stdio.h> +#include <unistd.h> +#include <libintl.h> +#include <rpc/rpc.h> +#include <sys/poll.h> +#include <sys/socket.h> +#include <rpc/pmap_clnt.h> +#ifdef USE_IN_LIBIO +# include <wchar.h> +#endif +#include <gnutls/gnutls.h> + +#include "clnt_gnutls.h" + +extern u_long _create_xid (void); + +#define MCALL_MSG_SIZE 24 + +struct ct_data + { + int ct_sock; + bool_t ct_closeit; + struct timeval ct_wait; + bool_t ct_waitset; /* wait set by clnt_control? */ + struct rpc_err ct_error; + char ct_mcall[MCALL_MSG_SIZE]; /* marshalled callmsg */ + u_int ct_mpos; /* pos after marshal */ + XDR ct_xdrs; + gnutls_session_t ct_session; /* GnuTLS session. */ + }; + +static int readtcp (char *, char *, int); +static int writetcp (char *, char *, int); + +static enum clnt_stat clnttcp_call (CLIENT *, u_long, xdrproc_t, caddr_t, + xdrproc_t, caddr_t, struct timeval); +static void clnttcp_abort (void); +static void clnttcp_geterr (CLIENT *, struct rpc_err *); +static bool_t clnttcp_freeres (CLIENT *, xdrproc_t, caddr_t); +static bool_t clnttcp_control (CLIENT *, int, char *); +static void clnttcp_destroy (CLIENT *); + +static const struct clnt_ops tcp_ops = +{ + clnttcp_call, + clnttcp_abort, + clnttcp_geterr, + clnttcp_freeres, + clnttcp_destroy, + clnttcp_control +}; + +static int start_gnutls_session (gnutls_certificate_credentials_t xcred, + struct ct_data *ct, int sock); +/* + * Create a client handle for a tcp/ip connection. + * If *sockp<0, *sockp is set to a newly created TCP socket and it is + * connected to raddr. If *sockp non-negative then + * raddr is ignored. The rpc/tcp package does buffering + * similar to stdio, so the client must pick send and receive buffer sizes,]; + * 0 => use the default. + * If raddr->sin_port is 0, then a binder on the remote machine is + * consulted for the right port number. + * NB: *sockp is copied into a private area. + * NB: It is the clients responsibility to close *sockp. + * NB: The rpch->cl_auth is set null authentication. Caller may wish to set this + * something more useful. + */ +CLIENT * +clntgnutls_create (int pf, struct sockaddr *raddr, int raddr_size, + gnutls_certificate_credentials_t xcred, + u_long prog, u_long vers, + int *sockp, u_int sendsz, u_int recvsz) +{ + CLIENT *h; + struct ct_data *ct; + struct rpc_msg call_msg; + + h = (CLIENT *) mem_alloc (sizeof (*h)); + ct = (struct ct_data *) mem_alloc (sizeof (*ct)); + if (h == NULL || ct == NULL) + { + struct rpc_createerr *ce = &get_rpc_createerr (); + (void) fprintf (stderr, "clnttcp_create: out of memory\n"); + ce->cf_stat = RPC_SYSTEMERROR; + ce->cf_error.re_errno = ENOMEM; + goto fooy; + } + +#if 0 + // RWMJ: Note this will never be supported for portmap because the + // portmap protocol depends pretty fundamentally on IPv4. Don't + /// use portmapper anyway -- it's silly. + /* + * If no port number given ask the pmap for one + */ + if (port == 0) + { + u_short port; + if ((port = pmap_getport (raddr, prog, vers, IPPROTO_TCP)) == 0) + { + mem_free ((caddr_t) ct, sizeof (struct ct_data)); + mem_free ((caddr_t) h, sizeof (CLIENT)); + return ((CLIENT *) NULL); + } + raddr->sin_port = htons (port); + } +#endif + + /* + * If no socket given, open one + */ + if (*sockp < 0) + { + *sockp = socket (pf, SOCK_STREAM, 0); + // RWMJ: Why? + //(void) bindresvport (*sockp, (struct sockaddr_in *) 0); + if ((*sockp < 0) + || (connect (*sockp, raddr, raddr_size) < 0)) + { + struct rpc_createerr *ce = &get_rpc_createerr (); + ce->cf_stat = RPC_SYSTEMERROR; + ce->cf_error.re_errno = errno; + if (*sockp >= 0) + (void) close (*sockp); + goto fooy; + } + ct->ct_closeit = TRUE; + } + else + { + ct->ct_closeit = FALSE; + } + + /* + * Set up private data struct + */ + ct->ct_sock = *sockp; + ct->ct_wait.tv_usec = 0; + ct->ct_waitset = FALSE; + + /* + * Initialize call message + */ + call_msg.rm_xid = _create_xid (); + call_msg.rm_direction = CALL; + call_msg.rm_call.cb_rpcvers = RPC_MSG_VERSION; + call_msg.rm_call.cb_prog = prog; + call_msg.rm_call.cb_vers = vers; + + /* + * pre-serialize the static part of the call msg and stash it away + */ + xdrmem_create (&(ct->ct_xdrs), ct->ct_mcall, MCALL_MSG_SIZE, + XDR_ENCODE); + if (!xdr_callhdr (&(ct->ct_xdrs), &call_msg)) + { + if (ct->ct_closeit) + { + (void) close (*sockp); + } + goto fooy; + } + ct->ct_mpos = XDR_GETPOS (&(ct->ct_xdrs)); + XDR_DESTROY (&(ct->ct_xdrs)); + + /* + * Create a client handle which uses xdrrec for serialization + * and authnone for authentication. + */ + xdrrec_create (&(ct->ct_xdrs), sendsz, recvsz, + (caddr_t) ct, readtcp, writetcp); + h->cl_ops = (struct clnt_ops *) &tcp_ops; + h->cl_private = (caddr_t) ct; + h->cl_auth = authnone_create (); + + /* Start GnuTLS on this socket. */ + if (start_gnutls_session (xcred, ct, *sockp) == -1) { + if (ct->ct_closeit) + close (*sockp); + goto fooy; + } + + return h; + +fooy: + /* + * Something goofed, free stuff and barf + */ + mem_free ((caddr_t) ct, sizeof (struct ct_data)); + mem_free ((caddr_t) h, sizeof (CLIENT)); + return ((CLIENT *) NULL); +} + +static int +start_gnutls_session (gnutls_certificate_credentials_t xcred, + struct ct_data *ct, int sock) +{ + const int cert_type_priority[3] = { + GNUTLS_CRT_X509, + GNUTLS_CRT_OPENPGP, + 0 + }; + + /* Initialize TLS session + */ + gnutls_init (&ct->ct_session, GNUTLS_CLIENT); + + /* Use default priorities */ + gnutls_set_default_priority (ct->ct_session); + gnutls_certificate_type_set_priority (ct->ct_session, cert_type_priority); + + /* put the x509 credentials to the current session + */ + gnutls_credentials_set (ct->ct_session, GNUTLS_CRD_CERTIFICATE, xcred); + + gnutls_transport_set_ptr (ct->ct_session, + (gnutls_transport_ptr_t) (long) sock); + + /* Perform the TLS handshake + */ + int ret = gnutls_handshake (ct->ct_session); + + if (ret < 0) + { + fprintf (stderr, "clnt_gnutls: TLS handshake failed\n"); + gnutls_perror (ret); + exit (1); + } + + /* XXX You need to verify the peer's certificate matches its name. */ + printf ("XXX need to verify peer's certificate matches its name.\n"); + +#if 0 + /* Print session info. */ + print_info (session); +#endif + + return 0; +} + +static enum clnt_stat +clnttcp_call (h, proc, xdr_args, args_ptr, xdr_results, results_ptr, timeout) + CLIENT *h; + u_long proc; + xdrproc_t xdr_args; + caddr_t args_ptr; + xdrproc_t xdr_results; + caddr_t results_ptr; + struct timeval timeout; +{ + struct ct_data *ct = (struct ct_data *) h->cl_private; + XDR *xdrs = &(ct->ct_xdrs); + struct rpc_msg reply_msg; + u_long x_id; + u_int32_t *msg_x_id = (u_int32_t *) (ct->ct_mcall); /* yuk */ + bool_t shipnow; + int refreshes = 2; + + if (!ct->ct_waitset) + { + ct->ct_wait = timeout; + } + + shipnow = + (xdr_results == (xdrproc_t) 0 && ct->ct_wait.tv_sec == 0 + && ct->ct_wait.tv_usec == 0) ? FALSE : TRUE; + +call_again: + xdrs->x_op = XDR_ENCODE; + ct->ct_error.re_status = RPC_SUCCESS; + x_id = ntohl (--(*msg_x_id)); + if ((!XDR_PUTBYTES (xdrs, ct->ct_mcall, ct->ct_mpos)) || + (!XDR_PUTLONG (xdrs, (long *) &proc)) || + (!AUTH_MARSHALL (h->cl_auth, xdrs)) || + (!(*xdr_args) (xdrs, args_ptr))) + { + if (ct->ct_error.re_status == RPC_SUCCESS) + ct->ct_error.re_status = RPC_CANTENCODEARGS; + (void) xdrrec_endofrecord (xdrs, TRUE); + return (ct->ct_error.re_status); + } + if (!xdrrec_endofrecord (xdrs, shipnow)) + return ct->ct_error.re_status = RPC_CANTSEND; + if (!shipnow) + return RPC_SUCCESS; + /* + * Hack to provide rpc-based message passing + */ + if (ct->ct_wait.tv_sec == 0 && ct->ct_wait.tv_usec == 0) + { + return ct->ct_error.re_status = RPC_TIMEDOUT; + } + + + /* + * Keep receiving until we get a valid transaction id + */ + xdrs->x_op = XDR_DECODE; + while (TRUE) + { + reply_msg.acpted_rply.ar_verf = _null_auth; + reply_msg.acpted_rply.ar_results.where = NULL; + reply_msg.acpted_rply.ar_results.proc = (xdrproc_t)xdr_void; + if (!xdrrec_skiprecord (xdrs)) + return (ct->ct_error.re_status); + /* now decode and validate the response header */ + if (!xdr_replymsg (xdrs, &reply_msg)) + { + if (ct->ct_error.re_status == RPC_SUCCESS) + continue; + return ct->ct_error.re_status; + } + if ((u_int32_t) reply_msg.rm_xid == (u_int32_t) x_id) + break; + } + + /* + * process header + */ + _seterr_reply (&reply_msg, &(ct->ct_error)); + if (ct->ct_error.re_status == RPC_SUCCESS) + { + if (!AUTH_VALIDATE (h->cl_auth, &reply_msg.acpted_rply.ar_verf)) + { + ct->ct_error.re_status = RPC_AUTHERROR; + ct->ct_error.re_why = AUTH_INVALIDRESP; + } + else if (!(*xdr_results) (xdrs, results_ptr)) + { + if (ct->ct_error.re_status == RPC_SUCCESS) + ct->ct_error.re_status = RPC_CANTDECODERES; + } + /* free verifier ... */ + if (reply_msg.acpted_rply.ar_verf.oa_base != NULL) + { + xdrs->x_op = XDR_FREE; + (void) xdr_opaque_auth (xdrs, + &(reply_msg.acpted_rply.ar_verf)); + } + } /* end successful completion */ + else + { + /* maybe our credentials need to be refreshed ... */ + if (refreshes-- && AUTH_REFRESH (h->cl_auth)) + goto call_again; + } /* end of unsuccessful completion */ + return ct->ct_error.re_status; +} + +static void +clnttcp_geterr (h, errp) + CLIENT *h; + struct rpc_err *errp; +{ + struct ct_data *ct = + (struct ct_data *) h->cl_private; + + *errp = ct->ct_error; +} + +static bool_t +clnttcp_freeres (cl, xdr_res, res_ptr) + CLIENT *cl; + xdrproc_t xdr_res; + caddr_t res_ptr; +{ + struct ct_data *ct = (struct ct_data *) cl->cl_private; + XDR *xdrs = &(ct->ct_xdrs); + + xdrs->x_op = XDR_FREE; + return (*xdr_res) (xdrs, res_ptr); +} + +static void +clnttcp_abort () +{ +} + +static bool_t +clnttcp_control (CLIENT *cl, int request, char *info) +{ + struct ct_data *ct = (struct ct_data *) cl->cl_private; + + + switch (request) + { + case CLSET_FD_CLOSE: + ct->ct_closeit = TRUE; + break; + case CLSET_FD_NCLOSE: + ct->ct_closeit = FALSE; + break; + case CLSET_TIMEOUT: + ct->ct_wait = *(struct timeval *) info; + ct->ct_waitset = TRUE; + break; + case CLGET_TIMEOUT: + *(struct timeval *) info = ct->ct_wait; + break; + case CLGET_FD: + *(int *)info = ct->ct_sock; + break; + case CLGET_XID: + /* + * use the knowledge that xid is the + * first element in the call structure *. + * This will get the xid of the PREVIOUS call + */ + *(u_long *)info = ntohl (*(u_long *)ct->ct_mcall); + break; + case CLSET_XID: + /* This will set the xid of the NEXT call */ + *(u_long *)ct->ct_mcall = htonl (*(u_long *)info - 1); + /* decrement by 1 as clnttcp_call() increments once */ + case CLGET_VERS: + /* + * This RELIES on the information that, in the call body, + * the version number field is the fifth field from the + * begining of the RPC header. MUST be changed if the + * call_struct is changed + */ + *(u_long *)info = ntohl (*(u_long *)(ct->ct_mcall + + 4 * BYTES_PER_XDR_UNIT)); + break; + case CLSET_VERS: + *(u_long *)(ct->ct_mcall + 4 * BYTES_PER_XDR_UNIT) + = htonl (*(u_long *)info); + break; + case CLGET_PROG: + /* + * This RELIES on the information that, in the call body, + * the program number field is the field from the + * begining of the RPC header. MUST be changed if the + * call_struct is changed + */ + *(u_long *)info = ntohl(*(u_long *)(ct->ct_mcall + + 3 * BYTES_PER_XDR_UNIT)); + break; + case CLSET_PROG: + *(u_long *)(ct->ct_mcall + 3 * BYTES_PER_XDR_UNIT) + = htonl(*(u_long *)info); + break; + /* The following are only possible with TI-RPC */ + case CLGET_RETRY_TIMEOUT: + case CLSET_RETRY_TIMEOUT: + case CLGET_SVC_ADDR: + case CLSET_SVC_ADDR: + case CLSET_PUSH_TIMOD: + case CLSET_POP_TIMOD: + default: + return FALSE; + } + return TRUE; +} + + +static void +clnttcp_destroy (CLIENT *h) +{ + struct ct_data *ct = + (struct ct_data *) h->cl_private; + + gnutls_bye (ct->ct_session, GNUTLS_SHUT_RDWR); + gnutls_deinit (ct->ct_session); + + if (ct->ct_closeit) + { + (void) close (ct->ct_sock); + } + XDR_DESTROY (&(ct->ct_xdrs)); + mem_free ((caddr_t) ct, sizeof (struct ct_data)); + mem_free ((caddr_t) h, sizeof (CLIENT)); +} + +/* + * Interface between xdr serializer and tcp connection. + * Behaves like the system calls, read & write, but keeps some error state + * around for the rpc level. + */ +static int +readtcp (char *ctptr, char *buf, int len) +{ + struct ct_data *ct = (struct ct_data *)ctptr; + struct pollfd fd; + int milliseconds = (ct->ct_wait.tv_sec * 1000) + + (ct->ct_wait.tv_usec / 1000); + + if (len == 0) + return 0; + + fd.fd = ct->ct_sock; + fd.events = POLLIN; + while (TRUE) + { + switch (poll(&fd, 1, milliseconds)) + { + case 0: + ct->ct_error.re_status = RPC_TIMEDOUT; + return -1; + + case -1: + if (errno == EINTR) + continue; + ct->ct_error.re_status = RPC_CANTRECV; + ct->ct_error.re_errno = errno; + return -1; + } + break; + } + + switch (len = gnutls_record_recv (ct->ct_session, buf, len)) + { + case 0: + /* premature eof */ + ct->ct_error.re_errno = ECONNRESET; + ct->ct_error.re_status = RPC_CANTRECV; + len = -1; /* it's really an error */ + break; + + case -1: + ct->ct_error.re_errno = errno; + ct->ct_error.re_status = RPC_CANTRECV; + break; + } + return len; +} + +static int +writetcp (char *ctptr, char *buf, int len) +{ + struct ct_data *ct = (struct ct_data*)ctptr; + + if (gnutls_record_send (ct->ct_session, buf, len) < 0) + { + ct->ct_error.re_errno = errno; + ct->ct_error.re_status = RPC_CANTSEND; + return -1; + } + + return len; +} --- /tmp/newfile 2007-01-30 17:23:46.000000000 +0000 +++ src/sunrpc/clnt_gnutls.h 2007-01-30 11:59:17.000000000 +0000 @@ -0,0 +1,20 @@ +/* + * clnt_gnutls.h: Interface to the SunRPC over GnuTLS. + * + * Copyright (C) 2007 Red Hat, Inc. + * + * See COPYING.LIB for the license of this software. + * + * Richard Jones <rjones(a)redhat.com> + */ + +#ifndef __CLNT_GNUTLS_H__ +#define __CLNT_GNUTLS_H__ + +extern CLIENT *clntgnutls_create (int pf, + struct sockaddr *raddr, int raddr_size, + gnutls_certificate_credentials_t xcred, + u_long prog, u_long vers, + int *sockp, u_int sendsz, u_int recvsz); + +#endif /* __CLNT_GNUTLS_H__ */ --- /tmp/newfile 2007-01-30 17:23:46.000000000 +0000 +++ src/sunrpc/clnt_tcp2.c 2007-01-30 16:08:20.000000000 +0000 @@ -0,0 +1,544 @@ +/* + * clnt_tcp2.c: SunRPC TCP client. This is a very slightly modified + * version of clnt_tcp.c from glibc which removes some of the IPv4 + * assumptions, so this version can be used over IPv4 or IPv6 sockets. + * + * Modifications from glibc-2.5 base by Richard Jones <rjones(a)redhat.com>. + */ + +/* @(#)clnt_tcp.c 2.2 88/08/01 4.0 RPCSRC */ +/* + * Sun RPC is a product of Sun Microsystems, Inc. and is provided for + * unrestricted use provided that this legend is included on all tape + * media and as a part of the software program in whole or part. Users + * may copy or modify Sun RPC without charge, but are not authorized + * to license or distribute it to anyone else except as part of a product or + * program developed by the user. + * + * SUN RPC IS PROVIDED AS IS WITH NO WARRANTIES OF ANY KIND INCLUDING THE + * WARRANTIES OF DESIGN, MERCHANTIBILITY AND FITNESS FOR A PARTICULAR + * PURPOSE, OR ARISING FROM A COURSE OF DEALING, USAGE OR TRADE PRACTICE. + * + * Sun RPC is provided with no support and without any obligation on the + * part of Sun Microsystems, Inc. to assist in its use, correction, + * modification or enhancement. + * + * SUN MICROSYSTEMS, INC. SHALL HAVE NO LIABILITY WITH RESPECT TO THE + * INFRINGEMENT OF COPYRIGHTS, TRADE SECRETS OR ANY PATENTS BY SUN RPC + * OR ANY PART THEREOF. + * + * In no event will Sun Microsystems, Inc. be liable for any lost revenue + * or profits or other special, indirect and consequential damages, even if + * Sun has been advised of the possibility of such damages. + * + * Sun Microsystems, Inc. + * 2550 Garcia Avenue + * Mountain View, California 94043 + */ +#if !defined(lint) && defined(SCCSIDS) +static char sccsid[] = "@(#)clnt_tcp.c 1.37 87/10/05 Copyr 1984 Sun Micro"; +#endif + +/* + * clnt_tcp.c, Implements a TCP/IP based, client side RPC. + * + * Copyright (C) 1984, Sun Microsystems, Inc. + * + * TCP based RPC supports 'batched calls'. + * A sequence of calls may be batched-up in a send buffer. The rpc call + * return immediately to the client even though the call was not necessarily + * sent. The batching occurs if the results' xdr routine is NULL (0) AND + * the rpc timeout value is zero (see clnt.h, rpc). + * + * Clients should NOT casually batch calls that in fact return results; that is, + * the server side should be aware that a call is batched and not produce any + * return message. Batched calls that produce many result messages can + * deadlock (netlock) the client and the server.... + * + * Now go hang yourself. + */ + +#include <netdb.h> +#include <errno.h> +#include <stdio.h> +#include <unistd.h> +#include <libintl.h> +#include <rpc/rpc.h> +#include <sys/poll.h> +#include <sys/socket.h> +#include <rpc/pmap_clnt.h> +#ifdef USE_IN_LIBIO +# include <wchar.h> +#endif + +#include "clnt_tcp2.h" + +extern u_long _create_xid (void); + +#define MCALL_MSG_SIZE 24 + +struct ct_data + { + int ct_sock; + bool_t ct_closeit; + struct timeval ct_wait; + bool_t ct_waitset; /* wait set by clnt_control? */ + struct rpc_err ct_error; + char ct_mcall[MCALL_MSG_SIZE]; /* marshalled callmsg */ + u_int ct_mpos; /* pos after marshal */ + XDR ct_xdrs; + }; + +static int readtcp (char *, char *, int); +static int writetcp (char *, char *, int); + +static enum clnt_stat clnttcp_call (CLIENT *, u_long, xdrproc_t, caddr_t, + xdrproc_t, caddr_t, struct timeval); +static void clnttcp_abort (void); +static void clnttcp_geterr (CLIENT *, struct rpc_err *); +static bool_t clnttcp_freeres (CLIENT *, xdrproc_t, caddr_t); +static bool_t clnttcp_control (CLIENT *, int, char *); +static void clnttcp_destroy (CLIENT *); + +static const struct clnt_ops tcp_ops = +{ + clnttcp_call, + clnttcp_abort, + clnttcp_geterr, + clnttcp_freeres, + clnttcp_destroy, + clnttcp_control +}; + +/* + * Create a client handle for a tcp/ip connection. + * If *sockp<0, *sockp is set to a newly created TCP socket and it is + * connected to raddr. If *sockp non-negative then + * raddr is ignored. The rpc/tcp package does buffering + * similar to stdio, so the client must pick send and receive buffer sizes,]; + * 0 => use the default. + * If raddr->sin_port is 0, then a binder on the remote machine is + * consulted for the right port number. + * NB: *sockp is copied into a private area. + * NB: It is the clients responsibility to close *sockp. + * NB: The rpch->cl_auth is set null authentication. Caller may wish to set this + * something more useful. + */ +CLIENT * +clnttcp2_create (int af, struct sockaddr *raddr, int raddr_size, + u_long prog, u_long vers, + int *sockp, u_int sendsz, u_int recvsz) +{ + CLIENT *h; + struct ct_data *ct; + struct rpc_msg call_msg; + + h = (CLIENT *) mem_alloc (sizeof (*h)); + ct = (struct ct_data *) mem_alloc (sizeof (*ct)); + if (h == NULL || ct == NULL) + { + struct rpc_createerr *ce = &get_rpc_createerr (); + (void) fprintf (stderr, "clnttcp_create: out of memory\n"); + ce->cf_stat = RPC_SYSTEMERROR; + ce->cf_error.re_errno = ENOMEM; + goto fooy; + } + +#if 0 + // RWMJ: Note this will never be supported for portmap because the + // portmap protocol depends pretty fundamentally on IPv4. Don't + /// use portmapper anyway -- it's silly. + /* + * If no port number given ask the pmap for one + */ + if (port == 0) + { + u_short port; + if ((port = pmap_getport (raddr, prog, vers, IPPROTO_TCP)) == 0) + { + mem_free ((caddr_t) ct, sizeof (struct ct_data)); + mem_free ((caddr_t) h, sizeof (CLIENT)); + return ((CLIENT *) NULL); + } + raddr->sin_port = htons (port); + } +#endif + + /* + * If no socket given, open one + */ + if (*sockp < 0) + { + *sockp = socket (af, SOCK_STREAM, 0); + // RWMJ: Why? + //(void) bindresvport (*sockp, (struct sockaddr_in *) 0); + if ((*sockp < 0) + || (connect (*sockp, raddr, raddr_size) < 0)) + { + struct rpc_createerr *ce = &get_rpc_createerr (); + ce->cf_stat = RPC_SYSTEMERROR; + ce->cf_error.re_errno = errno; + if (*sockp >= 0) + (void) close (*sockp); + goto fooy; + } + ct->ct_closeit = TRUE; + } + else + { + ct->ct_closeit = FALSE; + } + + /* + * Set up private data struct + */ + ct->ct_sock = *sockp; + ct->ct_wait.tv_usec = 0; + ct->ct_waitset = FALSE; + + /* + * Initialize call message + */ + call_msg.rm_xid = _create_xid (); + call_msg.rm_direction = CALL; + call_msg.rm_call.cb_rpcvers = RPC_MSG_VERSION; + call_msg.rm_call.cb_prog = prog; + call_msg.rm_call.cb_vers = vers; + + /* + * pre-serialize the static part of the call msg and stash it away + */ + xdrmem_create (&(ct->ct_xdrs), ct->ct_mcall, MCALL_MSG_SIZE, + XDR_ENCODE); + if (!xdr_callhdr (&(ct->ct_xdrs), &call_msg)) + { + if (ct->ct_closeit) + { + (void) close (*sockp); + } + goto fooy; + } + ct->ct_mpos = XDR_GETPOS (&(ct->ct_xdrs)); + XDR_DESTROY (&(ct->ct_xdrs)); + + /* + * Create a client handle which uses xdrrec for serialization + * and authnone for authentication. + */ + xdrrec_create (&(ct->ct_xdrs), sendsz, recvsz, + (caddr_t) ct, readtcp, writetcp); + h->cl_ops = (struct clnt_ops *) &tcp_ops; + h->cl_private = (caddr_t) ct; + h->cl_auth = authnone_create (); + return h; + +fooy: + /* + * Something goofed, free stuff and barf + */ + mem_free ((caddr_t) ct, sizeof (struct ct_data)); + mem_free ((caddr_t) h, sizeof (CLIENT)); + return ((CLIENT *) NULL); +} + +static enum clnt_stat +clnttcp_call (h, proc, xdr_args, args_ptr, xdr_results, results_ptr, timeout) + CLIENT *h; + u_long proc; + xdrproc_t xdr_args; + caddr_t args_ptr; + xdrproc_t xdr_results; + caddr_t results_ptr; + struct timeval timeout; +{ + struct ct_data *ct = (struct ct_data *) h->cl_private; + XDR *xdrs = &(ct->ct_xdrs); + struct rpc_msg reply_msg; + u_long x_id; + u_int32_t *msg_x_id = (u_int32_t *) (ct->ct_mcall); /* yuk */ + bool_t shipnow; + int refreshes = 2; + + if (!ct->ct_waitset) + { + ct->ct_wait = timeout; + } + + shipnow = + (xdr_results == (xdrproc_t) 0 && ct->ct_wait.tv_sec == 0 + && ct->ct_wait.tv_usec == 0) ? FALSE : TRUE; + +call_again: + xdrs->x_op = XDR_ENCODE; + ct->ct_error.re_status = RPC_SUCCESS; + x_id = ntohl (--(*msg_x_id)); + if ((!XDR_PUTBYTES (xdrs, ct->ct_mcall, ct->ct_mpos)) || + (!XDR_PUTLONG (xdrs, (long *) &proc)) || + (!AUTH_MARSHALL (h->cl_auth, xdrs)) || + (!(*xdr_args) (xdrs, args_ptr))) + { + if (ct->ct_error.re_status == RPC_SUCCESS) + ct->ct_error.re_status = RPC_CANTENCODEARGS; + (void) xdrrec_endofrecord (xdrs, TRUE); + return (ct->ct_error.re_status); + } + if (!xdrrec_endofrecord (xdrs, shipnow)) + return ct->ct_error.re_status = RPC_CANTSEND; + if (!shipnow) + return RPC_SUCCESS; + /* + * Hack to provide rpc-based message passing + */ + if (ct->ct_wait.tv_sec == 0 && ct->ct_wait.tv_usec == 0) + { + return ct->ct_error.re_status = RPC_TIMEDOUT; + } + + + /* + * Keep receiving until we get a valid transaction id + */ + xdrs->x_op = XDR_DECODE; + while (TRUE) + { + reply_msg.acpted_rply.ar_verf = _null_auth; + reply_msg.acpted_rply.ar_results.where = NULL; + reply_msg.acpted_rply.ar_results.proc = (xdrproc_t)xdr_void; + if (!xdrrec_skiprecord (xdrs)) + return (ct->ct_error.re_status); + /* now decode and validate the response header */ + if (!xdr_replymsg (xdrs, &reply_msg)) + { + if (ct->ct_error.re_status == RPC_SUCCESS) + continue; + return ct->ct_error.re_status; + } + if ((u_int32_t) reply_msg.rm_xid == (u_int32_t) x_id) + break; + } + + /* + * process header + */ + _seterr_reply (&reply_msg, &(ct->ct_error)); + if (ct->ct_error.re_status == RPC_SUCCESS) + { + if (!AUTH_VALIDATE (h->cl_auth, &reply_msg.acpted_rply.ar_verf)) + { + ct->ct_error.re_status = RPC_AUTHERROR; + ct->ct_error.re_why = AUTH_INVALIDRESP; + } + else if (!(*xdr_results) (xdrs, results_ptr)) + { + if (ct->ct_error.re_status == RPC_SUCCESS) + ct->ct_error.re_status = RPC_CANTDECODERES; + } + /* free verifier ... */ + if (reply_msg.acpted_rply.ar_verf.oa_base != NULL) + { + xdrs->x_op = XDR_FREE; + (void) xdr_opaque_auth (xdrs, + &(reply_msg.acpted_rply.ar_verf)); + } + } /* end successful completion */ + else + { + /* maybe our credentials need to be refreshed ... */ + if (refreshes-- && AUTH_REFRESH (h->cl_auth)) + goto call_again; + } /* end of unsuccessful completion */ + return ct->ct_error.re_status; +} + +static void +clnttcp_geterr (h, errp) + CLIENT *h; + struct rpc_err *errp; +{ + struct ct_data *ct = + (struct ct_data *) h->cl_private; + + *errp = ct->ct_error; +} + +static bool_t +clnttcp_freeres (cl, xdr_res, res_ptr) + CLIENT *cl; + xdrproc_t xdr_res; + caddr_t res_ptr; +{ + struct ct_data *ct = (struct ct_data *) cl->cl_private; + XDR *xdrs = &(ct->ct_xdrs); + + xdrs->x_op = XDR_FREE; + return (*xdr_res) (xdrs, res_ptr); +} + +static void +clnttcp_abort () +{ +} + +static bool_t +clnttcp_control (CLIENT *cl, int request, char *info) +{ + struct ct_data *ct = (struct ct_data *) cl->cl_private; + + + switch (request) + { + case CLSET_FD_CLOSE: + ct->ct_closeit = TRUE; + break; + case CLSET_FD_NCLOSE: + ct->ct_closeit = FALSE; + break; + case CLSET_TIMEOUT: + ct->ct_wait = *(struct timeval *) info; + ct->ct_waitset = TRUE; + break; + case CLGET_TIMEOUT: + *(struct timeval *) info = ct->ct_wait; + break; + case CLGET_FD: + *(int *)info = ct->ct_sock; + break; + case CLGET_XID: + /* + * use the knowledge that xid is the + * first element in the call structure *. + * This will get the xid of the PREVIOUS call + */ + *(u_long *)info = ntohl (*(u_long *)ct->ct_mcall); + break; + case CLSET_XID: + /* This will set the xid of the NEXT call */ + *(u_long *)ct->ct_mcall = htonl (*(u_long *)info - 1); + /* decrement by 1 as clnttcp_call() increments once */ + case CLGET_VERS: + /* + * This RELIES on the information that, in the call body, + * the version number field is the fifth field from the + * begining of the RPC header. MUST be changed if the + * call_struct is changed + */ + *(u_long *)info = ntohl (*(u_long *)(ct->ct_mcall + + 4 * BYTES_PER_XDR_UNIT)); + break; + case CLSET_VERS: + *(u_long *)(ct->ct_mcall + 4 * BYTES_PER_XDR_UNIT) + = htonl (*(u_long *)info); + break; + case CLGET_PROG: + /* + * This RELIES on the information that, in the call body, + * the program number field is the field from the + * begining of the RPC header. MUST be changed if the + * call_struct is changed + */ + *(u_long *)info = ntohl(*(u_long *)(ct->ct_mcall + + 3 * BYTES_PER_XDR_UNIT)); + break; + case CLSET_PROG: + *(u_long *)(ct->ct_mcall + 3 * BYTES_PER_XDR_UNIT) + = htonl(*(u_long *)info); + break; + /* The following are only possible with TI-RPC */ + case CLGET_RETRY_TIMEOUT: + case CLSET_RETRY_TIMEOUT: + case CLGET_SVC_ADDR: + case CLSET_SVC_ADDR: + case CLSET_PUSH_TIMOD: + case CLSET_POP_TIMOD: + default: + return FALSE; + } + return TRUE; +} + + +static void +clnttcp_destroy (CLIENT *h) +{ + struct ct_data *ct = + (struct ct_data *) h->cl_private; + + if (ct->ct_closeit) + { + (void) close (ct->ct_sock); + } + XDR_DESTROY (&(ct->ct_xdrs)); + mem_free ((caddr_t) ct, sizeof (struct ct_data)); + mem_free ((caddr_t) h, sizeof (CLIENT)); +} + +/* + * Interface between xdr serializer and tcp connection. + * Behaves like the system calls, read & write, but keeps some error state + * around for the rpc level. + */ +static int +readtcp (char *ctptr, char *buf, int len) +{ + struct ct_data *ct = (struct ct_data *)ctptr; + struct pollfd fd; + int milliseconds = (ct->ct_wait.tv_sec * 1000) + + (ct->ct_wait.tv_usec / 1000); + + if (len == 0) + return 0; + + fd.fd = ct->ct_sock; + fd.events = POLLIN; + while (TRUE) + { + switch (poll(&fd, 1, milliseconds)) + { + case 0: + ct->ct_error.re_status = RPC_TIMEDOUT; + return -1; + + case -1: + if (errno == EINTR) + continue; + ct->ct_error.re_status = RPC_CANTRECV; + ct->ct_error.re_errno = errno; + return -1; + } + break; + } + switch (len = read (ct->ct_sock, buf, len)) + { + + case 0: + /* premature eof */ + ct->ct_error.re_errno = ECONNRESET; + ct->ct_error.re_status = RPC_CANTRECV; + len = -1; /* it's really an error */ + break; + + case -1: + ct->ct_error.re_errno = errno; + ct->ct_error.re_status = RPC_CANTRECV; + break; + } + return len; +} + +static int +writetcp (char *ctptr, char *buf, int len) +{ + int i, cnt; + struct ct_data *ct = (struct ct_data*)ctptr; + + for (cnt = len; cnt > 0; cnt -= i, buf += i) + { + if ((i = write (ct->ct_sock, buf, cnt)) == -1) + { + ct->ct_error.re_errno = errno; + ct->ct_error.re_status = RPC_CANTSEND; + return -1; + } + } + return len; +} --- /tmp/newfile 2007-01-30 17:23:46.000000000 +0000 +++ src/sunrpc/clnt_tcp2.h 2007-01-30 15:26:09.000000000 +0000 @@ -0,0 +1,18 @@ +/* + * clnt_tcp2.h: Interface to the SunRPC over TCP. + * + * Copyright (C) 2007 Red Hat, Inc. + * + * See COPYING.LIB for the license of this software. + * + * Richard Jones <rjones(a)redhat.com> + */ + +#ifndef __CLNT_TCP2_H__ +#define __CLNT_TCP2_H__ + +extern CLIENT *clnttcp2_create (int af, struct sockaddr *raddr, int raddr_size, + u_long prog, u_long vers, + int *sockp, u_int sendsz, u_int recvsz); + +#endif /* __CLNT_TCP2_H__ */ --- /tmp/newfile 2007-01-30 17:23:46.000000000 +0000 +++ src/sunrpc/create_xid.c 2007-01-30 13:21:07.000000000 +0000 @@ -0,0 +1,55 @@ +/* + * create_xid.c: Makes unique, unguessable identifiers. This is modified + * from the original source found in glibc-2.5. The original version + * is thread safe (but the rest of SunRPC isn't). This version is + * not thread safe. + * + * Modifications from glibc-2.5 base by Richard Jones <rjones(a)redhat.com>. + */ + +/* Copyright (c) 1998, 2000 Free Software Foundation, Inc. + This file is part of the GNU C Library. + Contributed by Thorsten Kukuk <kukuk(a)vt.uni-paderborn.de>, 1998. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, write to the Free + Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA + 02111-1307 USA. */ + +#include <unistd.h> +#include <stdlib.h> +#include <sys/time.h> +#include <bits/libc-lock.h> +#include <rpc/rpc.h> + +static int is_initialized; +static struct drand48_data __rpc_lrand48_data; + +unsigned long +_create_xid (void) +{ + long int res; + + if (!is_initialized) + { + struct timeval now; + + gettimeofday (&now, (struct timezone *) 0); + srand48_r (now.tv_sec ^ now.tv_usec, &__rpc_lrand48_data); + is_initialized = 1; + } + + lrand48_r (&__rpc_lrand48_data, &res); + + return res; +} --- /tmp/newfile 2007-01-30 17:23:46.000000000 +0000 +++ src/sunrpc/svc_gnutls.c 2007-01-30 17:03:35.000000000 +0000 @@ -0,0 +1,477 @@ +/* + * svc_gnutls.c: SunRPC over GnuTLS server. This is a modified version + * of svc_tcp.c from glibc which is written to run over GnuTLS. + * + * Modifications from glibc-2.5 base by Richard Jones <rjones(a)redhat.com>. + */ + +/* @(#)svc_tcp.c 2.2 88/08/01 4.0 RPCSRC */ +/* + * Sun RPC is a product of Sun Microsystems, Inc. and is provided for + * unrestricted use provided that this legend is included on all tape + * media and as a part of the software program in whole or part. Users + * may copy or modify Sun RPC without charge, but are not authorized + * to license or distribute it to anyone else except as part of a product or + * program developed by the user. + * + * SUN RPC IS PROVIDED AS IS WITH NO WARRANTIES OF ANY KIND INCLUDING THE + * WARRANTIES OF DESIGN, MERCHANTIBILITY AND FITNESS FOR A PARTICULAR + * PURPOSE, OR ARISING FROM A COURSE OF DEALING, USAGE OR TRADE PRACTICE. + * + * Sun RPC is provided with no support and without any obligation on the + * part of Sun Microsystems, Inc. to assist in its use, correction, + * modification or enhancement. + * + * SUN MICROSYSTEMS, INC. SHALL HAVE NO LIABILITY WITH RESPECT TO THE + * INFRINGEMENT OF COPYRIGHTS, TRADE SECRETS OR ANY PATENTS BY SUN RPC + * OR ANY PART THEREOF. + * + * In no event will Sun Microsystems, Inc. be liable for any lost revenue + * or profits or other special, indirect and consequential damages, even if + * Sun has been advised of the possibility of such damages. + * + * Sun Microsystems, Inc. + * 2550 Garcia Avenue + * Mountain View, California 94043 + */ +#if !defined(lint) && defined(SCCSIDS) +static char sccsid[] = "@(#)svc_tcp.c 1.21 87/08/11 Copyr 1984 Sun Micro"; +#endif + +/* + * svc_tcp.c, Server side for TCP/IP based RPC. + * + * Copyright (C) 1984, Sun Microsystems, Inc. + * + * Actually implements two flavors of transporter - + * a tcp rendezvouser (a listener and connection establisher) + * and a record/tcp stream. + */ + +#include <stdio.h> +#include <unistd.h> +#include <string.h> +#include <libintl.h> +#include <rpc/rpc.h> +#include <sys/socket.h> +#include <sys/poll.h> +#include <errno.h> +#include <stdlib.h> + +#ifdef USE_IN_LIBIO +# include <wchar.h> +# include <libio/iolibio.h> +#endif + +#include <gnutls/gnutls.h> +#include "svc_gnutls.h" + +/* + * Ops vector for TCP/IP based rpc service handle + */ +static bool_t svctcp_recv (SVCXPRT *, struct rpc_msg *); +static enum xprt_stat svctcp_stat (SVCXPRT *); +static bool_t svctcp_getargs (SVCXPRT *, xdrproc_t, caddr_t); +static bool_t svctcp_reply (SVCXPRT *, struct rpc_msg *); +static bool_t svctcp_freeargs (SVCXPRT *, xdrproc_t, caddr_t); +static void svctcp_destroy (SVCXPRT *); + +static const struct xp_ops svctcp_op = +{ + svctcp_recv, + svctcp_stat, + svctcp_getargs, + svctcp_reply, + svctcp_freeargs, + svctcp_destroy +}; + +/* + * Ops vector for TCP/IP rendezvous handler + */ +static bool_t rendezvous_request (SVCXPRT *, struct rpc_msg *); +static enum xprt_stat rendezvous_stat (SVCXPRT *); +static void svctcp_rendezvous_abort (void) __attribute__ ((__noreturn__)); + +/* This function makes sure abort() relocation goes through PLT + and thus can be lazy bound. */ +static void +svctcp_rendezvous_abort (void) +{ + abort (); +}; + +static const struct xp_ops svctcp_rendezvous_op = +{ + rendezvous_request, + rendezvous_stat, + (bool_t (*) (SVCXPRT *, xdrproc_t, caddr_t)) svctcp_rendezvous_abort, + (bool_t (*) (SVCXPRT *, struct rpc_msg *)) svctcp_rendezvous_abort, + (bool_t (*) (SVCXPRT *, xdrproc_t, caddr_t)) svctcp_rendezvous_abort, + svctcp_destroy +}; + +static int readtcp (char*, char *, int); +static int writetcp (char *, char *, int); +static SVCXPRT *makefd_xprt (int, u_int, u_int, gnutls_session_t); + +struct tcp_rendezvous + { /* kept in xprt->xp_p1 */ + u_int sendsize; + u_int recvsize; + gnutls_certificate_credentials_t x509_cred; + }; + +struct tcp_conn + { /* kept in xprt->xp_p1 */ + enum xprt_stat strm_stat; + u_long x_id; + XDR xdrs; + char verf_body[MAX_AUTH_BYTES]; + }; + +/* + * Usage: + * xprt = svcgnutls_create(sock, send_buf_size, recv_buf_size, + * gnutls_certificate_credentials_t x509_cred); + * + * Creates, registers, and returns a (rpc) tcp based transporter. + * Once *xprt is initialized, it is registered as a transporter + * see (svc.h, xprt_register). This routine returns + * a NULL if a problem occurred. + * + * If sock<0 then a socket is created, else sock is used. + * If the socket, sock is not bound to a port then svctcp_create + * binds it to an arbitrary port. The routine then starts a tcp + * listener on the socket's associated port. In any (successful) case, + * xprt->xp_sock is the registered socket number and xprt->xp_port is the + * associated port number. + * + * Since tcp streams do buffered io similar to stdio, the caller can specify + * how big the send and receive buffers are via the second and third parms; + * 0 => use the system default. + */ +SVCXPRT * +svcgnutls_create (int sock, u_int sendsize, u_int recvsize, + gnutls_certificate_credentials_t x509_cred) +{ + bool_t madesock = FALSE; + SVCXPRT *xprt; + struct tcp_rendezvous *r; + struct sockaddr_in addr; + socklen_t len = sizeof (struct sockaddr_in); + + if (sock == RPC_ANYSOCK) + { + if ((sock = socket (AF_INET, SOCK_STREAM, IPPROTO_TCP)) < 0) + { + perror ("svc_gnutls.c - tcp socket creation problem"); + return (SVCXPRT *) NULL; + } + madesock = TRUE; + } + bzero ((char *) &addr, sizeof (addr)); + addr.sin_family = AF_INET; + if (bindresvport (sock, &addr)) + { + addr.sin_port = 0; + (void) bind (sock, (struct sockaddr *) &addr, len); + } + if ((getsockname (sock, (struct sockaddr *) &addr, &len) != 0) || + (listen (sock, SOMAXCONN) != 0)) + { + perror ("svc_gnutls.c - cannot getsockname or listen"); + if (madesock) + (void) close (sock); + return (SVCXPRT *) NULL; + } + r = (struct tcp_rendezvous *) mem_alloc (sizeof (*r)); + xprt = (SVCXPRT *) mem_alloc (sizeof (SVCXPRT)); + if (r == NULL || xprt == NULL) + { + (void) fprintf (stderr, "svcgnutls_create: out of memory\n"); + mem_free (r, sizeof (*r)); + mem_free (xprt, sizeof (SVCXPRT)); + return NULL; + } + r->sendsize = sendsize; + r->recvsize = recvsize; + r->x509_cred = x509_cred; + xprt->xp_p2 = NULL; + xprt->xp_p1 = (caddr_t) r; + xprt->xp_verf = _null_auth; + xprt->xp_ops = &svctcp_rendezvous_op; + xprt->xp_port = ntohs (addr.sin_port); + xprt->xp_sock = sock; + xprt_register (xprt); + return xprt; +} + +#if 0 +/* + * Like svtcp_create(), except the routine takes any *open* UNIX file + * descriptor as its first input. + */ +SVCXPRT * +svcfd_create (int fd, u_int sendsize, u_int recvsize) +{ + return makefd_xprt (fd, sendsize, recvsize); +} +#endif + +static SVCXPRT * +makefd_xprt (int fd, u_int sendsize, u_int recvsize, + gnutls_session_t session) +{ + SVCXPRT *xprt; + struct tcp_conn *cd; + + xprt = (SVCXPRT *) mem_alloc (sizeof (SVCXPRT)); + cd = (struct tcp_conn *) mem_alloc (sizeof (struct tcp_conn)); + if (xprt == (SVCXPRT *) NULL || cd == NULL) + { + (void) fprintf (stderr, "svc_gnutls: makefd_xprt: out of memory\n"); + mem_free (xprt, sizeof (SVCXPRT)); + mem_free (cd, sizeof (struct tcp_conn)); + return NULL; + } + cd->strm_stat = XPRT_IDLE; + xdrrec_create (&(cd->xdrs), sendsize, recvsize, + (caddr_t) xprt, readtcp, writetcp); + xprt->xp_p2 = (caddr_t) session; + xprt->xp_p1 = (caddr_t) cd; + xprt->xp_verf.oa_base = cd->verf_body; + xprt->xp_addrlen = 0; + xprt->xp_ops = &svctcp_op; /* truly deals with calls */ + xprt->xp_port = 0; /* this is a connection, not a rendezvouser */ + xprt->xp_sock = fd; + xprt_register (xprt); + return xprt; +} + +// XXX DH_BITS must be defined the same in the server main loop too. +#define DH_BITS 1024 + +static gnutls_session_t +initialize_tls_session (gnutls_certificate_credentials_t x509_cred) +{ + gnutls_session_t session; + + gnutls_init (&session, GNUTLS_SERVER); + + /* avoid calling all the priority functions, since the defaults + * are adequate. + */ + gnutls_set_default_priority (session); + + gnutls_credentials_set (session, GNUTLS_CRD_CERTIFICATE, x509_cred); + + /* request client certificate if any. + */ + gnutls_certificate_server_set_request (session, GNUTLS_CERT_REQUEST); + + gnutls_dh_set_prime_bits (session, DH_BITS); + + return session; +} + +static bool_t +rendezvous_request (SVCXPRT *xprt, + struct rpc_msg *errmsg __attribute__((unused))) +{ + int sock; + struct tcp_rendezvous *r; + struct sockaddr_in addr; + socklen_t len; + + r = (struct tcp_rendezvous *) xprt->xp_p1; + gnutls_session_t session = initialize_tls_session (r->x509_cred); + +again: + len = sizeof (struct sockaddr_in); + if ((sock = accept (xprt->xp_sock, (struct sockaddr *) &addr, &len)) < 0) + { + if (errno == EINTR) + goto again; + return FALSE; + } + + gnutls_transport_set_ptr (session, (gnutls_transport_ptr_t) (long) sock); + int ret = gnutls_handshake (session); + if (ret < 0) + { + close (sock); + gnutls_deinit (session); + fprintf (stderr, "svc_gnutls: TLS handshake failed (%s)\n\n", + gnutls_strerror (ret)); + return FALSE; + } + + // XXX verify peer + fprintf (stderr, "XXX you must verify the peer\n"); + + /* + * make a new transporter (re-uses xprt) + */ + xprt = makefd_xprt (sock, r->sendsize, r->recvsize, session); +#if 0 + memcpy (&xprt->xp_raddr, &addr, sizeof (addr)); + xprt->xp_addrlen = len; +#endif + return FALSE; /* there is never an rpc msg to be processed */ +} + +static enum xprt_stat +rendezvous_stat (SVCXPRT *xprt __attribute__((unused))) +{ + return XPRT_IDLE; +} + +static void +svctcp_destroy (SVCXPRT *xprt) +{ + struct tcp_conn *cd = (struct tcp_conn *) xprt->xp_p1; + gnutls_session_t session = (gnutls_session_t) xprt->xp_p2; + + gnutls_bye (session, GNUTLS_SHUT_WR); + + xprt_unregister (xprt); + (void) close (xprt->xp_sock); + if (xprt->xp_port != 0) + { + /* a rendezvouser socket */ + xprt->xp_port = 0; + } + else + { + /* an actual connection socket */ + XDR_DESTROY (&(cd->xdrs)); + } + + gnutls_deinit (session); + + mem_free ((caddr_t) cd, sizeof (struct tcp_conn)); + mem_free ((caddr_t) xprt, sizeof (SVCXPRT)); +} + + +/* + * reads data from the tcp connection. + * any error is fatal and the connection is closed. + * (And a read of zero bytes is a half closed stream => error.) + */ +static int +readtcp (char *xprtptr, char *buf, int len) +{ + SVCXPRT *xprt = (SVCXPRT *)xprtptr; + gnutls_session_t session = (gnutls_session_t) xprt->xp_p2; + int sock = xprt->xp_sock; + int milliseconds = 35 * 1000; + struct pollfd pollfd; + + do + { + pollfd.fd = sock; + pollfd.events = POLLIN; + switch (poll (&pollfd, 1, milliseconds)) + { + case -1: + if (errno == EINTR) + continue; + /*FALLTHROUGH*/ + case 0: + goto fatal_err; + default: + if ((pollfd.revents & POLLERR) || (pollfd.revents & POLLHUP) + || (pollfd.revents & POLLNVAL)) + goto fatal_err; + break; + } + } + while ((pollfd.revents & POLLIN) == 0); + + if ((len = gnutls_record_recv (session, buf, len)) > 0) + return len; + + fatal_err: + ((struct tcp_conn *) (xprt->xp_p1))->strm_stat = XPRT_DIED; + return -1; +} + +/* + * writes data to the tcp connection. + * Any error is fatal and the connection is closed. + */ +static int +writetcp (char *xprtptr, char * buf, int len) +{ + SVCXPRT *xprt = (SVCXPRT *)xprtptr; + gnutls_session_t session = (gnutls_session_t) xprt->xp_p2; + + if (gnutls_record_send (session, buf, len) < 0) { + ((struct tcp_conn *) (xprt->xp_p1))->strm_stat = XPRT_DIED; + return -1; + } + + return len; +} + +static enum xprt_stat +svctcp_stat (SVCXPRT *xprt) +{ + struct tcp_conn *cd = + (struct tcp_conn *) (xprt->xp_p1); + + if (cd->strm_stat == XPRT_DIED) + return XPRT_DIED; + if (!xdrrec_eof (&(cd->xdrs))) + return XPRT_MOREREQS; + return XPRT_IDLE; +} + +static bool_t +svctcp_recv (SVCXPRT *xprt, struct rpc_msg *msg) +{ + struct tcp_conn *cd = (struct tcp_conn *) (xprt->xp_p1); + XDR *xdrs = &(cd->xdrs); + + xdrs->x_op = XDR_DECODE; + (void) xdrrec_skiprecord (xdrs); + if (xdr_callmsg (xdrs, msg)) + { + cd->x_id = msg->rm_xid; + return TRUE; + } + cd->strm_stat = XPRT_DIED; /* XXXX */ + return FALSE; +} + +static bool_t +svctcp_getargs (SVCXPRT *xprt, xdrproc_t xdr_args, caddr_t args_ptr) +{ + return ((*xdr_args) (&(((struct tcp_conn *) + (xprt->xp_p1))->xdrs), args_ptr)); +} + +static bool_t +svctcp_freeargs (SVCXPRT *xprt, xdrproc_t xdr_args, caddr_t args_ptr) +{ + XDR *xdrs = &(((struct tcp_conn *) (xprt->xp_p1))->xdrs); + + xdrs->x_op = XDR_FREE; + return ((*xdr_args) (xdrs, args_ptr)); +} + +static bool_t +svctcp_reply (SVCXPRT *xprt, struct rpc_msg *msg) +{ + struct tcp_conn *cd = (struct tcp_conn *) (xprt->xp_p1); + XDR *xdrs = &(cd->xdrs); + bool_t stat; + + xdrs->x_op = XDR_ENCODE; + msg->rm_xid = cd->x_id; + stat = xdr_replymsg (xdrs, msg); + (void) xdrrec_endofrecord (xdrs, TRUE); + return stat; +} --- /tmp/newfile 2007-01-30 17:23:46.000000000 +0000 +++ src/sunrpc/svc_gnutls.h 2007-01-29 17:03:24.000000000 +0000 @@ -0,0 +1,17 @@ +/* + * svc_gnutls.h: Interface to the SunRPC over GnuTLS server. + * + * Copyright (C) 2007 Red Hat, Inc. + * + * See COPYING.LIB for the license of this software. + * + * Richard Jones <rjones(a)redhat.com> + */ + +#ifndef __SVC_GNUTLS_H__ +#define __SVC_GNUTLS_H__ + +extern SVCXPRT *svcgnutls_create (int sock, u_int sendsize, u_int recvsize, + gnutls_certificate_credentials_t x509_cred); + +#endif /* __SVC_GNUTLS_H__ */ --- /tmp/newfile 2007-01-30 17:23:46.000000000 +0000 +++ src/sunrpc/svc_tcp2.c 2007-01-30 16:16:32.000000000 +0000 @@ -0,0 +1,426 @@ +/* + * svc_tcp2.c: SunRPC TCP server. This is a very slightly modified + * version of svc_tcp.c from glibc which removes some of the IPv4 + * assumptions, so this version can be used over IPv4 or IPv6 sockets. + * + * Modifications from glibc-2.5 base by Richard Jones <rjones(a)redhat.com>. + */ + +/* @(#)svc_tcp.c 2.2 88/08/01 4.0 RPCSRC */ +/* + * Sun RPC is a product of Sun Microsystems, Inc. and is provided for + * unrestricted use provided that this legend is included on all tape + * media and as a part of the software program in whole or part. Users + * may copy or modify Sun RPC without charge, but are not authorized + * to license or distribute it to anyone else except as part of a product or + * program developed by the user. + * + * SUN RPC IS PROVIDED AS IS WITH NO WARRANTIES OF ANY KIND INCLUDING THE + * WARRANTIES OF DESIGN, MERCHANTIBILITY AND FITNESS FOR A PARTICULAR + * PURPOSE, OR ARISING FROM A COURSE OF DEALING, USAGE OR TRADE PRACTICE. + * + * Sun RPC is provided with no support and without any obligation on the + * part of Sun Microsystems, Inc. to assist in its use, correction, + * modification or enhancement. + * + * SUN MICROSYSTEMS, INC. SHALL HAVE NO LIABILITY WITH RESPECT TO THE + * INFRINGEMENT OF COPYRIGHTS, TRADE SECRETS OR ANY PATENTS BY SUN RPC + * OR ANY PART THEREOF. + * + * In no event will Sun Microsystems, Inc. be liable for any lost revenue + * or profits or other special, indirect and consequential damages, even if + * Sun has been advised of the possibility of such damages. + * + * Sun Microsystems, Inc. + * 2550 Garcia Avenue + * Mountain View, California 94043 + */ +#if !defined(lint) && defined(SCCSIDS) +static char sccsid[] = "@(#)svc_tcp.c 1.21 87/08/11 Copyr 1984 Sun Micro"; +#endif + +/* + * svc_tcp.c, Server side for TCP/IP based RPC. + * + * Copyright (C) 1984, Sun Microsystems, Inc. + * + * Actually implements two flavors of transporter - + * a tcp rendezvouser (a listener and connection establisher) + * and a record/tcp stream. + */ + +#include <stdio.h> +#include <unistd.h> +#include <string.h> +#include <libintl.h> +#include <rpc/rpc.h> +#include <sys/socket.h> +#include <sys/poll.h> +#include <errno.h> +#include <stdlib.h> + +#ifdef USE_IN_LIBIO +# include <wchar.h> +# include <libio/iolibio.h> +#endif + +#include "svc_tcp2.h" + +/* + * Ops vector for TCP/IP based rpc service handle + */ +static bool_t svctcp_recv (SVCXPRT *, struct rpc_msg *); +static enum xprt_stat svctcp_stat (SVCXPRT *); +static bool_t svctcp_getargs (SVCXPRT *, xdrproc_t, caddr_t); +static bool_t svctcp_reply (SVCXPRT *, struct rpc_msg *); +static bool_t svctcp_freeargs (SVCXPRT *, xdrproc_t, caddr_t); +static void svctcp_destroy (SVCXPRT *); + +static const struct xp_ops svctcp_op = +{ + svctcp_recv, + svctcp_stat, + svctcp_getargs, + svctcp_reply, + svctcp_freeargs, + svctcp_destroy +}; + +/* + * Ops vector for TCP/IP rendezvous handler + */ +static bool_t rendezvous_request (SVCXPRT *, struct rpc_msg *); +static enum xprt_stat rendezvous_stat (SVCXPRT *); +static void svctcp_rendezvous_abort (void) __attribute__ ((__noreturn__)); + +/* This function makes sure abort() relocation goes through PLT + and thus can be lazy bound. */ +static void +svctcp_rendezvous_abort (void) +{ + abort (); +}; + +static const struct xp_ops svctcp_rendezvous_op = +{ + rendezvous_request, + rendezvous_stat, + (bool_t (*) (SVCXPRT *, xdrproc_t, caddr_t)) svctcp_rendezvous_abort, + (bool_t (*) (SVCXPRT *, struct rpc_msg *)) svctcp_rendezvous_abort, + (bool_t (*) (SVCXPRT *, xdrproc_t, caddr_t)) svctcp_rendezvous_abort, + svctcp_destroy +}; + +static int readtcp (char*, char *, int); +static int writetcp (char *, char *, int); +static SVCXPRT *makefd_xprt (int, u_int, u_int); + +struct tcp_rendezvous + { /* kept in xprt->xp_p1 */ + u_int sendsize; + u_int recvsize; + }; + +struct tcp_conn + { /* kept in xprt->xp_p1 */ + enum xprt_stat strm_stat; + u_long x_id; + XDR xdrs; + char verf_body[MAX_AUTH_BYTES]; + }; + +/* + * Usage: + * xprt = svctcp_create(sock, send_buf_size, recv_buf_size); + * + * Creates, registers, and returns a (rpc) tcp based transporter. + * Once *xprt is initialized, it is registered as a transporter + * see (svc.h, xprt_register). This routine returns + * a NULL if a problem occurred. + * + * If sock<0 then a socket is created, else sock is used. + * If the socket, sock is not bound to a port then svctcp_create + * binds it to an arbitrary port. The routine then starts a tcp + * listener on the socket's associated port. In any (successful) case, + * xprt->xp_sock is the registered socket number and xprt->xp_port is the + * associated port number. + * + * Since tcp streams do buffered io similar to stdio, the caller can specify + * how big the send and receive buffers are via the second and third parms; + * 0 => use the system default. + */ +SVCXPRT * +svctcp2_create (int sock, u_int sendsize, u_int recvsize) +{ + bool_t madesock = FALSE; + SVCXPRT *xprt; + struct tcp_rendezvous *r; + struct sockaddr_in addr; + socklen_t len = sizeof (struct sockaddr_in); + + if (sock == RPC_ANYSOCK) + { + if ((sock = socket (AF_INET, SOCK_STREAM, IPPROTO_TCP)) < 0) + { + perror ("svc_tcp.c - tcp socket creation problem"); + return (SVCXPRT *) NULL; + } + madesock = TRUE; + + bzero ((char *) &addr, sizeof (addr)); + addr.sin_family = AF_INET; + if (bindresvport (sock, &addr)) + { + addr.sin_port = 0; + (void) bind (sock, (struct sockaddr *) &addr, len); + } + if ((getsockname (sock, (struct sockaddr *) &addr, &len) != 0) || + (listen (sock, SOMAXCONN) != 0)) + { + perror ("svc_tcp.c - cannot getsockname or listen"); + if (madesock) + (void) close (sock); + return (SVCXPRT *) NULL; + } + } + r = (struct tcp_rendezvous *) mem_alloc (sizeof (*r)); + xprt = (SVCXPRT *) mem_alloc (sizeof (SVCXPRT)); + if (r == NULL || xprt == NULL) + { + (void) fprintf (stderr, "svctcp_create: out of memory\n"); + mem_free (r, sizeof (*r)); + mem_free (xprt, sizeof (SVCXPRT)); + return NULL; + } + r->sendsize = sendsize; + r->recvsize = recvsize; + xprt->xp_p2 = NULL; + xprt->xp_p1 = (caddr_t) r; + xprt->xp_verf = _null_auth; + xprt->xp_ops = &svctcp_rendezvous_op; + xprt->xp_port = ntohs (addr.sin_port); + xprt->xp_sock = sock; + xprt_register (xprt); + return xprt; +} + +#if 0 +/* + * Like svtcp_create(), except the routine takes any *open* UNIX file + * descriptor as its first input. + */ +SVCXPRT * +svcfd_create (int fd, u_int sendsize, u_int recvsize) +{ + return makefd_xprt (fd, sendsize, recvsize); +} +#endif + +static SVCXPRT * +makefd_xprt (int fd, u_int sendsize, u_int recvsize) +{ + SVCXPRT *xprt; + struct tcp_conn *cd; + + xprt = (SVCXPRT *) mem_alloc (sizeof (SVCXPRT)); + cd = (struct tcp_conn *) mem_alloc (sizeof (struct tcp_conn)); + if (xprt == (SVCXPRT *) NULL || cd == NULL) + { + (void) fprintf (stderr, "svc_tcp: makefd_xprt: out of memory\n"); + mem_free (xprt, sizeof (SVCXPRT)); + mem_free (cd, sizeof (struct tcp_conn)); + return NULL; + } + cd->strm_stat = XPRT_IDLE; + xdrrec_create (&(cd->xdrs), sendsize, recvsize, + (caddr_t) xprt, readtcp, writetcp); + xprt->xp_p2 = NULL; + xprt->xp_p1 = (caddr_t) cd; + xprt->xp_verf.oa_base = cd->verf_body; + xprt->xp_addrlen = 0; + xprt->xp_ops = &svctcp_op; /* truly deals with calls */ + xprt->xp_port = 0; /* this is a connection, not a rendezvouser */ + xprt->xp_sock = fd; + xprt_register (xprt); + return xprt; +} + +static bool_t +rendezvous_request (SVCXPRT *xprt, + struct rpc_msg *errmsg __attribute__((unused))) +{ + int sock; + struct tcp_rendezvous *r; + struct sockaddr_in addr; + socklen_t len; + + r = (struct tcp_rendezvous *) xprt->xp_p1; +again: + len = sizeof (struct sockaddr_in); + if ((sock = accept (xprt->xp_sock, (struct sockaddr *) &addr, &len)) < 0) + { + if (errno == EINTR) + goto again; + return FALSE; + } + /* + * make a new transporter (re-uses xprt) + */ + xprt = makefd_xprt (sock, r->sendsize, r->recvsize); +#if 0 + memcpy (&xprt->xp_raddr, &addr, sizeof (addr)); + xprt->xp_addrlen = len; +#endif + return FALSE; /* there is never an rpc msg to be processed */ +} + +static enum xprt_stat +rendezvous_stat (SVCXPRT *xprt __attribute__((unused))) +{ + return XPRT_IDLE; +} + +static void +svctcp_destroy (SVCXPRT *xprt) +{ + struct tcp_conn *cd = (struct tcp_conn *) xprt->xp_p1; + + xprt_unregister (xprt); + (void) close (xprt->xp_sock); + if (xprt->xp_port != 0) + { + /* a rendezvouser socket */ + xprt->xp_port = 0; + } + else + { + /* an actual connection socket */ + XDR_DESTROY (&(cd->xdrs)); + } + mem_free ((caddr_t) cd, sizeof (struct tcp_conn)); + mem_free ((caddr_t) xprt, sizeof (SVCXPRT)); +} + + +/* + * reads data from the tcp connection. + * any error is fatal and the connection is closed. + * (And a read of zero bytes is a half closed stream => error.) + */ +static int +readtcp (char *xprtptr, char *buf, int len) +{ + SVCXPRT *xprt = (SVCXPRT *)xprtptr; + int sock = xprt->xp_sock; + int milliseconds = 35 * 1000; + struct pollfd pollfd; + + do + { + pollfd.fd = sock; + pollfd.events = POLLIN; + switch (poll (&pollfd, 1, milliseconds)) + { + case -1: + if (errno == EINTR) + continue; + /*FALLTHROUGH*/ + case 0: + goto fatal_err; + default: + if ((pollfd.revents & POLLERR) || (pollfd.revents & POLLHUP) + || (pollfd.revents & POLLNVAL)) + goto fatal_err; + break; + } + } + while ((pollfd.revents & POLLIN) == 0); + + if ((len = read (sock, buf, len)) > 0) + return len; + + fatal_err: + ((struct tcp_conn *) (xprt->xp_p1))->strm_stat = XPRT_DIED; + return -1; +} + +/* + * writes data to the tcp connection. + * Any error is fatal and the connection is closed. + */ +static int +writetcp (char *xprtptr, char * buf, int len) +{ + SVCXPRT *xprt = (SVCXPRT *)xprtptr; + int i, cnt; + + for (cnt = len; cnt > 0; cnt -= i, buf += i) + { + if ((i = write (xprt->xp_sock, buf, cnt)) < 0) + { + ((struct tcp_conn *) (xprt->xp_p1))->strm_stat = XPRT_DIED; + return -1; + } + } + return len; +} + +static enum xprt_stat +svctcp_stat (SVCXPRT *xprt) +{ + struct tcp_conn *cd = + (struct tcp_conn *) (xprt->xp_p1); + + if (cd->strm_stat == XPRT_DIED) + return XPRT_DIED; + if (!xdrrec_eof (&(cd->xdrs))) + return XPRT_MOREREQS; + return XPRT_IDLE; +} + +static bool_t +svctcp_recv (SVCXPRT *xprt, struct rpc_msg *msg) +{ + struct tcp_conn *cd = (struct tcp_conn *) (xprt->xp_p1); + XDR *xdrs = &(cd->xdrs); + + xdrs->x_op = XDR_DECODE; + (void) xdrrec_skiprecord (xdrs); + if (xdr_callmsg (xdrs, msg)) + { + cd->x_id = msg->rm_xid; + return TRUE; + } + cd->strm_stat = XPRT_DIED; /* XXXX */ + return FALSE; +} + +static bool_t +svctcp_getargs (SVCXPRT *xprt, xdrproc_t xdr_args, caddr_t args_ptr) +{ + return ((*xdr_args) (&(((struct tcp_conn *) + (xprt->xp_p1))->xdrs), args_ptr)); +} + +static bool_t +svctcp_freeargs (SVCXPRT *xprt, xdrproc_t xdr_args, caddr_t args_ptr) +{ + XDR *xdrs = &(((struct tcp_conn *) (xprt->xp_p1))->xdrs); + + xdrs->x_op = XDR_FREE; + return ((*xdr_args) (xdrs, args_ptr)); +} + +static bool_t +svctcp_reply (SVCXPRT *xprt, struct rpc_msg *msg) +{ + struct tcp_conn *cd = (struct tcp_conn *) (xprt->xp_p1); + XDR *xdrs = &(cd->xdrs); + bool_t stat; + + xdrs->x_op = XDR_ENCODE; + msg->rm_xid = cd->x_id; + stat = xdr_replymsg (xdrs, msg); + (void) xdrrec_endofrecord (xdrs, TRUE); + return stat; +} --- /tmp/newfile 2007-01-30 17:23:46.000000000 +0000 +++ src/sunrpc/svc_tcp2.h 2007-01-30 16:14:22.000000000 +0000 @@ -0,0 +1,16 @@ +/* + * svc_tcp2.h: Interface to the SunRPC over TCP server. + * + * Copyright (C) 2007 Red Hat, Inc. + * + * See COPYING.LIB for the license of this software. + * + * Richard Jones <rjones(a)redhat.com> + */ + +#ifndef __SVC_TCP2_H__ +#define __SVC_TCP2_H__ + +extern SVCXPRT *svctcp2_create (int sock, u_int sendsize, u_int recvsize); + +#endif /* __SVC_TCP2_H__ */

5 14

[Libvir] Features and roadmap tentative draft
by Daniel Veillard 09 Feb '07

09 Feb '07

We had a meeting this week in the Red Hat team and looked at the work ahead for libvirt. We identified a set of feature that we would like to get in, the time frame (mostly based on Fedora releases schedules) and tried to identify remaining issues: Features when Issues --------------------------------------------------------------------------- KEmu & KVM FC7 Wire API , the way the process is controlled mid feb will probably change, but requires QEmu upstream extensions of the monitor remote access FC8 Naming (URIs), Authentication and wire protocol inetd hook, XML format slight change for dhcp Virtual Network FC7 Sharing QEmu, root priviledge for the daemon isolation Migration FC8 API design Access Control ??? depends on Xen Security Module, KVM equivalent Storage ??? Early concept, will be discussed separately on the mailing-list Sync virsh shutdown FC7 Easy to add in virsh I expect to make a new libvirt release with the KEmu and KVM support as well as a first version of the virtual network support next week. Then the focus will be on remote support and adding migration to the API. Other various issues raised: - URL/drivers, making sure that the last proposal for URI used in remote connection are okay - python bindings: disable automatic generation, handle manually, extend - one Xen driver: currently implemented on top of 3 drivers, clean it up - go though errors, cleaning up duplicates - compile without Xen, once KEmu and KVM support gets added - documentation cleanup: generation and CVS, checkout on libvirt.org - export the config parser API at the shared library with __ prefix since that's used by the daemons but not - semantic of virConnect(NULL) is a bit fuzzy, Xen or default engine ? - Need to plug libxml2 memory routines for memory debug Of course the details will be discussed on the list, but at least it gets a bit clearer where we expect to spend time, but as always other features and feedback is more than welcome, it's needed, so if there is any issue or comments on our viewpoint of the roadmap and feature, please share them :-) Daniel -- Red Hat Virtualization group http://redhat.com/virtualization/ Daniel Veillard | virtualization library http://libvirt.org/ veillard(a)redhat.com | libxml GNOME XML XSLT toolkit http://xmlsoft.org/ http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/

2 1

[Libvir] [PATCH] Remove dead code
by Daniel P. Berrange 07 Feb '07

07 Feb '07

The attached patch removes 1000 lines of dead code from xml.c and xend_internal.c. It is not clear that this code was ever used at all in the past, and we certainly don't use or test it today. If we truely need it in the future, we can trivially get it back from the CVS repo. KIlling it now makes its easier to maintain the code base because we're not obscuring active code with historical cruft. Finally it will also make the code coverage reporting more useful since it won't be complaining about untested code we never use. Any objections to me applying this ? Regards, Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=|

2 1

Re: [Libvir] what is virt-install doing differently?
by Aron Griffis 31 Jan '07

31 Jan '07

Hi Daniel, Daniel P. Berrange wrote: [Tue Jan 23 2007, 10:49:07AM EST] > For HVM installs, the different should only be the addition of the > CDROM device. If you take a look in /var/log/xen/xend.log it should > show you the final SEXPR that XenD is getting when creating the > guest. Comparing the SEXPR seen when booting the config file, vs > that generated by virt-intsall may shed some more light on the > problem. > > Also, virt-install has a --debug flag which will make it show the libvirt > XML description it used to boot. Thanks for these suggestions. Yongkang looked at xend.log and found this: You, Yongkang wrote: [Thu Jan 25 2007, 10:26:33AM EST] > Thanks for the information. I have check /var/log/xen/xend.log. > Found something interesting. When using virt-install to install VTI > domain, it seemed didn't parse the right vcpu number to qemu-dm. > Please look at the following log: > > 1. The spawning qemu log, when installation by virt-intall with vcpu=4: > --------------------- > [2007-01-25 23:14:13 xend 2817] INFO (image:418) spawning device models: /usr/lib/xen/bin/qemu-dm ['/usr/lib/xen/bin/qemu-dm', '-d', '9', '-m', '1024', '-boot', 'c', '-acpi', '-domain-name', 'test', '-net', 'nic,vlan=1,macaddr=00:16:3e:0a:d2:86,model=rtl8139', '-net', 'tap,vlan=1,bridge=xenbr0', '-vncunused', '-vnclisten', '127.0.0.1'] > ~~~~~~~~~~~~ > > 2. The spawning qemu log, when use command "xm create": > ------------------------ > [2007-01-25 17:13:13 xend 2817] INFO (image:418) spawning device models: /usr/lib/xen/bin/qemu-dm ['/usr/lib/xen/bin/qemu-dm', '-d', '7', '-m', '1024', '-boot', 'd', '-vcpus', '4', '-acpi', '-domain-name', 'ExampleVTIDomain', '-net', 'nic,vlan=1,macaddr=00:16:3e:31:42:c4,model=rtl8139', '-net', 'tap,vlan=1,bridge=xenbr0', '-vncviewer'] > ~~~~~~~~~~~ > > There is "-vcpus 4" in the 2nd log. It means qemu would create up > a 4 vpucs VTI domain. > > For the situation 1, although qemu doesn't get any information about > "vcpu=4". But xen has already prepared 2 vcpus to qemu. Will it be > the problem? I'm looking through the code now to see if I can find where this is going awry, but while I'm looking, does this ring a bell for you? Thanks, Aron

2 2

[Libvir] libvirtd - configuration file
by Richard W.M. Jones 31 Jan '07

31 Jan '07

Libvirtd needs a configuration file. It's very simple - just to store the services on which it's listening, and optionally to configure port numbers and paths. Libvirt so far doesn't have a concept of a configuration file, except that obviously domain descriptions come out of XML files. Do people on list have a preference for a style of configuration file? An alternative is to configure entirely via the command line. Rich. PS: Personally when administering a machine there is nothing worse than finding an XML configuration file ... -- Emerging Technologies, Red Hat http://et.redhat.com/~rjones/ 64 Baker Street, London, W1U 7DF Mobile: +44 7866 314 421 "[Negative numbers] darken the very whole doctrines of the equations and make dark of the things which are in their nature excessively obvious and simple" (Francis Maseres FRS, mathematician, 1759)

4 3

[Libvir] Problem in configuring libvirt (Xen store library not found)
by aditya aekkaldevi 31 Jan '07

31 Jan '07

Hello, First of all I am a newbie to linux. I installed Red hat enterprise 5 beta 2 which has xen. So in order to start xen I had to install libvirt as it is needed by virt-install command. But I am getting an error while configuring libvirt software. The error is "Xen store library not found". I am getting this error after a check for library containing xs_read failed. So how can I configure and install libvirt now. Please let me know the solution for this ASAP. Also, if I am doing anything wrong, please let me know. Thanks in advance Aditya

3 2

[Libvir] Automated build script
by Daniel P. Berrange 31 Jan '07

31 Jan '07

I'm going to setup libvirt to be part of an automated build system[1] which will checkout the code, perform a build, run tests, etc, etc on a 24x7 basis. This merely requires adding a single file 'autobuild.sh' to the top level directory to perform the neccessary build/test tasks (see attached file). As part of this automation I'll have it setup to run all the test suites under valgrind which should let us identify memory leaks sooner. Finally I also plan to add in the Makefile rules[2] neccessary to run gcov so as to enable reports to be generated on test suite coverage. Regards, Dan. [1] http://autobuild.org/ [2] Hmm, i'm sure i sent such a patch before but can't find any trace of it in the mail archives. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=|

1 0

[Libvir] Nothing ever calls the driver ->init function, right?
by Richard W.M. Jones 31 Jan '07

31 Jan '07

... or at least I can't see it being called anywhere. Rich. -- Emerging Technologies, Red Hat http://et.redhat.com/~rjones/ 64 Baker Street, London, W1U 7DF Mobile: +44 7866 314 421 "[Negative numbers] darken the very whole doctrines of the equations and make dark of the things which are in their nature excessively obvious and simple" (Francis Maseres FRS, mathematician, 1759)

3 5

[Libvir] Issue with creating domains that use pygrub
by Chris Halstead 30 Jan '07

30 Jan '07

I'm stumped. I'm trying to automate creation of domains through libvirt and the Python bindings, but I'm having trouble getting things to work when using pygrub. If I define an XML-described domain and use <bootloader>/usr/bin/pygrub</bootloader> the domain tries to start and then immediately crashes. Using explicit local-pathed vmlinuz/initrd without the bootloader element works just fine. Am I missing something? The version of libvirt on my test system in 0.1.9. Thanks! C. Halstead <chris(a)sourcelabs.com> SourceLabs - http://www.sourcelabs.com Dependable Open Source Systems

2 2

[Libvir] Option string in getopt_long wrong?
by Richard W.M. Jones 30 Jan '07

30 Jan '07

while ((arg = getopt_long(end, argv, "d:hqtv", opt, &idx)) != -1) { ------ switch (arg) { case 'd': ctl->debug = atoi(optarg); break; case 'h': help = 1; break; case 'q': ctl->quiet = TRUE; break; case 't': ctl->timing = TRUE; break; case 'c': ctl->name = vshStrdup(ctl, optarg); break; // etc Shurley shome mishtake? Rich. -- Emerging Technologies, Red Hat http://et.redhat.com/~rjones/ 64 Baker Street, London, W1U 7DF Mobile: +44 7866 314 421 "[Negative numbers] darken the very whole doctrines of the equations and make dark of the things which are in their nature excessively obvious and simple" (Francis Maseres FRS, mathematician, 1759)

3 3

[Libvir] Virtual networking
by Mark McLoughlin 29 Jan '07

29 Jan '07

Hi, Dan and I have been discussing how to "fix networking", not just Xen's networking but also getting something sane wrt. QEMU/KVM etc. Comments very welcome on the writeup below. The libvirt stuff is towards the end, but I think all of it is probably useful to this list. Cheers, Mark. Virtual Networking The ability to manage virtual machines is something which is receiving a lot of focus right now. Xen, KVM, QEMU and others provide the infrastructure required to run a virtual machine, and each can provide guests with a virtual network interface. This proposal addresses the problem of how guests are networked together. We aim: * To make virtual networking "just work". Guests should be able to communicate with each other, their host and the Internet without any fuss or configuration. This should be the case even with laptops and offline machines. * To allow a greater flexibily with how guests are networked. It should be possible to isolate groups of guests in different networks, allow guests on different physical machines to communicate, firewall guests' networks from physical networks or allow guests to appear just like physical machines on physical networks. * To make networking virtual machines analogous with networking physical machines. * To support inter-networking between virtualisation technologies. User Visible Concepts ===================== It's important to consider the manner in which we expose the functionality of virtual networking. What concepts will be exposing through the UI? Are those concepts well defined and consistent? Are those concepts more complex than neccessary? Or are the too simple to be able to support the functionality we want? Real world, or "physical", concepts[1]: * Network - a number of interconnected machines. * Network Interface - hardware which enables a machine to connect to a network. * Bridge - hardware which allows enables the interconnection of machines to form a network. Bridges can also be connected to other bridges to form a larger network. * Router - hardware which connects two or more distinct networks, allowing machines on different networks to communicate with one another. Sometimes a router and a bridge are available as a combined piece of hardware - the bridge forms a network and the router connects that network to another distinct network. * Firewall - software on a router which can be used to control how machines on an "external" network (e.g. the Internet) can communicate with machines on an "internal" network. For a given type of connection, you can choose to disallow connections of a that type or forward them to a specific internal machine. Can also be used to control how internal machines can communicate with external machines. With virtual networking, we will be exposing the following "virtual" concepts: * Virtual Network - a number of interconnected virtual machines. * Virtual Network Interface - a network interface in a virtual machine. * Virtual Bridge - allows the interconnection of virtual machines to form a virtual network. A virtual bridge may be configured to also act as a virtual router and firewall. A virtual bridge may also be connected to another virtual bridge (perhaps on another physical machine) to create a larger virtual network. (Note, unprivileged users may create any of the above) Finally, where the physical world meets the virtual world: * Shared Physical Interface - if a physical interface is configured to be "shared", then any number of virtual interfaces may be connected to it allowing virtual machines to be connected to the same physical network which the physical interface is connected to. Only privileged users may configure a physical interface to be shared and/or connect guests to it. There are a few problems with all of the above: 1. The distinction between a bridge and a router requires a lot of technical knowledge to fully understand. However, the model of e.g. a LinkSys router is familiar to a lot of people - a box which allows you to network your machines together and connect that network to (and firewall off) the Internet. 2. This "shared physical interface" notion is very "makey upey". We could perhaps talk about the idea in terms of connecting a physical interface to a virtual bridge, but it exposes the bridge vs. router distinction more than we'd like. 3. Guests are connected to a specific physical interface, whereas perhaps users wish guests to be connected to "the network" - i.e. if NetworkManager switched from wireless to wired while remaining on the same subnet, perhaps we'd like to automatically switch the bridge to the new network. In reality, though, bridged networking is only really sane for machines on a fairly static network connection. [1] - Yes, these definitions aren't entirely accurate, but they describe the kind of understanding a moderately technical user might have of the concepts. Example Networks ================ Below are some example networks users may configure and an explanation of how that network would be implemented in practice. 1. A privileged user creates two (Xen) guests, each with a Virtual Network Interface. Without any special networking configuration, these two guests are connected to a default Virtual Network which contains a combined Virtual Bridge/Router/Firewall. +-----------+ D +-----------+ | Guest | N D H | Guest | | A | A N C | B | | +---+ | T S P | +---+ | | |NIC| | ^ ^ ^ | |NIC| | +---+-+-+---+ +---+---+ +---+-+-+---+ ^ | ^ | +--------+ +---+---+ +--------+ | +-->+ vif1.0 +----+ vnbr0 +----+ vif2.0 +<--+ +--------+ +-------+ +--------+ Notes: * "vnbr0" is a bridge device with it's own IP address on the same subnet as the guests. * IP forwarding is enabled in Dom0. Masquerading and DNAT is implemented using iptables. * We run a DHCP server and a DNS proxy in Dom0 (e.g. dnsmasq) 2. A privileged user does exactly the same thing as (1), but with QEMU guests. D N D H A N C T S P ^ ^ ^ +---+---+ | +---+---+ +-----------+ | vnbr0 | +-----------+ | Guest | +---+---+ | Guest | | A | | | B | | +---+ | +---+---+ | +---+ | | |NIC| | | vtap0 | | |NIC| | +---+-+-+---+ +---+---+ +---+-+-+---+ ^ +-------+ | +-------+ ^ | | | +---+---+ | | | +------>+ VLAN0 +-+ VDE +-+ VLAN0 +<------+ | | +-------+ | | +-------+ +-------+ Notes: * VDE is a userspace ethernet bridge implemented using vde_switch * "vtap0" is a TAP device created by vde_switch * Everything else is the same as (1) * This could be done without vde_switch by having Guest A create vtap0 and have Guest B connect directly to Guest A's VLAN. However, if Guest A is shut down, Guest B's network would go down. 3. An unprivileged user does exactly the same thing as (2). +-----------+ +-----------+ | Guest | +----+----+ | Guest | | A | |userspace| | B | | +---+ | | network | | +---+ | | |NIC| | | stack | | |NIC| | +---+-+-+---+ +----+----+ +---+-+-+---+ ^ +-------+ | +-------+ ^ | | | +---+---+ | | | +------>+ VLAN0 +-+ VDE +-+ VLAN0 +<------+ | | +-------+ | | +-------+ +-------+ Notes: * Similar to (2) except there is can be no TAP device or bridge * The userspace network stack is implemented using slirpvde to provide a DHCP server and DNS proxy to the network, but also effectively a SNAT and DNAT router. * slirpvde implements ethernet, ip, tcp, udp, icmp, dhcp, tftp (etc.) in userspace. Completely crazy, but since the kernel apparently has no secure way to allow unprivileged users to leverage the kernel's network stack for this, then it must be done in userspace. 4. Same as (2), except the user also creates two Xen guests. +-----------+ D +-----------+ | Guest | N D H | Guest | | A | A N C | B | | +---+ | T S P | +---+ | | |NIC| | ^ ^ ^ | |NIC| | +---+-+-+---+ +---+---+ +---+-+-+---+ ^ | ^ | +--------+ +---+---+ +--------+ | +-->+ vif1.0 +----+ vnbr0 +----+ vif2.0 +<--+ +--------+ +---+---+ +--------+ | +---+---+ | vtap0 | +---+---+ | +-------+ +--+--+ +-------+ +---->+ VLAN0 +----+ VDE +---+ VLAN0 +<-----+ | +-------+ +-----+ +-------+ | V V +---+-+-+---+ +---+-+-+---+ | |NIC| | | |NIC| | | +---+ | | +---+ | | Guest | | Guest | | C | | D | +-----------+ +-----------+ Notes: * In this case we could do away with VDE and have each QEMU guest use its own TAP device. 5. Same as (3) except Guests A and C are connected to a Shared Physical Interface. +-----------+ | D +-----------+ | Guest | ^ | N D H | Guest | | A | | | A N C | B | | +---+ | +---+---+ | T S P | +---+ | | |NIC| | | eth0 | | ^ ^ ^ | |NIC| | +---+-+-+---+ +---+---+ | +---+---+ +---+-+-+---+ ^ | | | ^ | +--------+ +---+---+ | +---+---+ +--------+ | +>+ vif1.0 +-+ ebr0 + | + vnbr0 +-+ vif2.0 +<-+ +--------+ +---+---+ | +---+---+ +--------+ | | | +---+---+ | +---+---+ | vtap1 | | | vtap0 | +---+---+ | +---+---+ | | | +-------+ +--+--+ | +--+--+ +-------+ +->+ VLAN0 +--+ VDE + | + VDE +--+ VLAN0 +<-+ | +-------+ +-----+ | +-----+ +-------+ | V | V +---+-+-+---+ | +---+-+-+---+ | |NIC| | | | |NIC| | | +---+ | | | +---+ | | Guest | | | Guest | | C | | | D | +-----------+ | +-----------+ Notes: * The idea here is that when the admin configures eth0 to be shareable, eth0 is configured as an addressless NIC enslaved to a bridge which has the MAC address and IP address that eth0 should have * Again, VDE is redundant here. 6. Same as 2) except the QEMU guests are on a Virtual Network on another physical machine which is, in turn, connected to the Virtual Network on the first physical machine +-----------+ D +-----------+ | Guest | N D H | Guest | | A | A N C | B | | +---+ | T S P | +---+ | | |NIC| | ^ ^ ^ | |NIC| | +---+-+-+---+ +---+---+ +---+-+-+---+ ^ | ^ | +--------+ +---+---+ +--------+ | +-->+ vif1.0 +----+ vnbr0 +----+ vif2.0 +<--+ +--------+ +---+---+ +--------+ | +---+---+ | vtap0 | +---+---+ | +--+--+ | VDE | +--+--+ | First Physical Machine V ------------------------------------------------------------- Second Physical Machine ^ | +-------+ +--+--+ +-------+ +---->+ VLAN0 +----+ VDE +---+ VLAN0 +<-----+ | +-------+ +-----+ +-------+ | V V +---+-+-+---+ +---+-+-+---+ | |NIC| | | |NIC| | | +---+ | | +---+ | | Guest | | Guest | | C | | D | +-----------+ +-----------+ Notes: * What's going on here is that the two VDEs are connected over the network, either via a plan socket or perhaps encapsulated in another protocol like SSH or TLS One interesting thing to note from all of those examples is that although QEMU's networking options are very interesting, it doesn't actually make sense for a network to be implemented inside a guest. The network needs to be external to any guests, and so we use VDE to offer similar networking options to the ones QEMU provides. All QEMU needs to be able to do is to connect to VDE. User Interface ============== This isn't meant a UI specification, but just some notes on how this stuff might be exposed in virt-manager. * Networks List: * Name * Virtual/Physical * Status * Activity/traffic * Virtual Network Configuration: * Name * List of connected guests * Allow other Virtual Networks to connect to this (defaults to no) * Connect to other Virtual Network (defaults to none) * DHCP enabled - DHCP configuration: * IP range (optional) * Router IP address (optional) * Guest IP address/hostname assignment (optional) * Forwarding enabled - firewall configuration: * Incoming ports list and destination guest+port for each (defaults to empty) * Blocked outgoing ports lists (defaults to empty) * Virtual NICs list: * Guest interface name * Virtual Network/Shared Physical Interface * Hostname (defaults to guest name) * IP address (if assigned) * MAC address (if assigned) * Virtual NIC Configuration: * Random MAC address, or user-supplied MAC address. * Virtual Network or Shared Physical Interface to connect to. Implementation ============== Parity with the current state of networking with Xen will be achieved by: * Implementing "shared physical interface" support in Fedora's initscripts and network configuration tool. It boils down to configuring the interface (e.g. eth0) something like: ifcfg-peth0: DEVICE=peth0 ONBOOT=yes Bridge=eth0 HWADDR=00:30:48:30:73:19 ifcfg-eth0 DEVICE=eth0 Type=Bridge ONBOOT=yes BOOTPROTO=dhcp * Fixing Xen so that netloop is no longer required. Upstream have ideas about how to make Xen automatically copy any frames that are destined for Dom0 so that the netback driver doesn't run out of shared pages if Dom0 doesn't process the frames quickly enough. * Create new network/vif scripts for Xen which will connect guests to a shared physical interface's bridge. Virtual Networks will be implemented in libvirt. First, there will be an XML description of Virtual Networks e.g.: <network id="0"> <name>Foo</name> <uuid>596a5d2171f48fb2e068e2386a5c413e</uuid> <listen address="172.31.0.5" port="1234" /> <connections> <connection address="172.31.0.6" port="4321" /> </conections> <dhcp enabled="true"> <ip address="10.0.0.1" netmask="255.255.255.0" start="10.0.0.128" end="10.0.0.254" /> </dhcp> <forwarding enabled="true"> <incoming default="deny"> <allow port="123" domain="foobar" destport="321" /> </incoming> <outgoing default="allow"> <deny port="25" /> </outgoing> </forwarding> <network> In a manner similar to libvirt's QEMU support, there will be a daemon to manage Virtual Networks. The daemon will have access to a store of network definitions. The daemon will be responsible for managing the bridge devices, vde_switch/dhcp/dnses processes and the iptables rules needed for SNAT/DNAT etc. virsh command line interface would look like: $> virsh network-create foo.xml $> virsh network-dumpxml > foo.xml $> virsh network-define foo.xml $> virsh network-list $> virsh network-start Foo $> virsh network-stop Foo $> virsh network-restart Foo The libvirt API for virtual networks would be modelled on the API for virtual machines: /* * Virtual Networks API */ /** * virNetwork: * * a virNetwork is a private structure representing a virtual network. */ typedef struct _virNetwork virNetwork; /** * virNetworkPtr: * * a virNetworkPtr is pointer to a virNetwork private structure, this is the * type used to reference a virtual network in the API. */ typedef virNetwork *virNetworkPtr; /** * virNetworkCreateFlags: * * Flags OR'ed together to provide specific behaviour when creating a * Network. */ typedef enum { VIR_NETWORK_NONE = 0 } virNetworkCreateFlags; /* * List active networks */ int virConnectNumOfNetworks (virConnectPtr conn); int virConnectListNetworks (virConnectPtr conn, int *ids, int maxids); /* * List inactive networks */ int virConnectNumOfDefinedNetworks (virConnectPtr conn); int virConnectListDefinedNetworks (virConnectPtr conn, const char **names, int maxnames); /* * Lookup network by name, id or uuid */ virNetworkPtr virNetworkLookupByName (virConnectPtr conn, const char *name); virNetworkPtr virNetworkLookupByID (virConnectPtr conn, int id); virNetworkPtr virNetworkLookupByUUID (virConnectPtr conn, const unsigned char *uuid); virNetworkPtr virNetworkLookupByUUIDString (virConnectPtr conn, const char *uuid); /* * Create active transient network */ virNetworkPtr virNetworkCreateXML (virConnectPtr conn, const char *xmlDesc, unsigned int flags); /* * Define inactive persistent network */ virNetworkPtr virNetworkDefineXML (virConnectPtr conn, const char *xmlDesc); /* * Delete persistent network */ int virNetworkUndefine (virNetworkPtr network); /* * Activate persistent network */ int virNetworkCreate (virNetworkPtr network); /* * Network destroy/free */ int virNetworkDestroy (virNetworkPtr network); int virNetworkFree (virNetworkPtr network); /* * Network informations */ const char* virNetworkGetName (virNetworkPtr network); unsigned int virNetworkGetID (virNetworkPtr network); int virNetworkGetUUID (virNetworkPtr network, unsigned char *uuid); int virNetworkGetUUIDString (virNetworkPtr network, char *buf); char * virNetworkGetXMLDesc (virNetworkPtr network, int flags); Discussion points on the XML format and API: * The XML format isn't thought out at all, but briefly: * The <listen> and <connections> elements describe networks connected across physical machine boundaries. * The <dhcp> element describes the configuration of the DHCP server on the network. * The <forwarding> element describes how incoming and outgoing connections are forwarded. * Since virConnect is supposed to be a connection to a specific hypervisor, does it make sense to create networks (which should be hypervisor agnostic) through virConnect? * Are we needlessly replicating any mistakes from the domains API here? e.g. is the transient vs. persistent distinction useful for networks? * Is a UUID useful for networks? Yes, because it distinguishes between networks of the same name on different hosts? * Where is the connection between domains and networks in either the API or the XML format? How is a domain associated with a network? You put a bridge name in the <network&gtl definition and use that in the domains <interface> definition? Or you put the network name in the interface definition and have libvirt look up the bridge name when creating the guest? * Should it be possible to stop/start/restart a network? What for? If something breaks the user restarts it to see if that will fix it?

7 59

[Libvir] Add 'console' and 'vncdisplay' commands to virsh
by Daniel P. Berrange 29 Jan '07

29 Jan '07

The attached patch adds two new commands to virsh: - console - connects to the guest's serial console - vncdisplay - outputs the ip address & port number of a guest vnc display The former is another stage in eliminating the need to run 'xm' - replacing the Xen specific 'xm console' code, with the added advantage of working with any libvirt backend driver. The vncdisplay is intended to make it easier for people to launch a VNC viewer process. It prints out a IP address & port number in a format suitable for passing to vncviewer on the command line, eg vncviewer `virsh vncdisplay myguest` Regards, Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=|

3 3

[Libvir] Dom0 hypercall ABI changes in xen-unstable ?
by Daniel P. Berrange 28 Jan '07

28 Jan '07

This changeset concerns me in today's xen-unstable http://xenbits2.xensource.com/xen-unstable.hg?rev/30af6cfdb05c "Make domctl/sysctl interfaces 32-/64-bit invariant. This kills off a fair amount of unpleasant CONFIG_COMPAT shimming and avoids needing to keep the compat paths in sync as these interfaces continue to develop." For example @@ -81,9 +81,9 @@ struct xen_sysctl_physinfo { uint32_t sockets_per_node; uint32_t nr_nodes; uint32_t cpu_khz; - uint64_t total_pages; - uint64_t free_pages; - uint64_t scrub_pages; + uint64_aligned_t total_pages; + uint64_aligned_t free_pages; + uint64_aligned_t scrub_pages; uint32_t hw_cap[8]; Suggests to me that alignment of thse fields may now have changed on 32-bit hosts. Well, at least it would if it were not for public/xen.h doing: #define uint64_aligned_t uint64_t So has it changed or not ? The interface version was incremented at least... -#define XEN_DOMCTL_INTERFACE_VERSION 0x00000004 +#define XEN_DOMCTL_INTERFACE_VERSION 0x00000005 -#define XEN_SYSCTL_INTERFACE_VERSION 0x00000002 +#define XEN_SYSCTL_INTERFACE_VERSION 0x00000003 I'm expecting the worst in terms of ABI compat, but I'm really not at all certain yet. Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=|

1 1

[Libvir] Patch: In autogen.sh, if configure fails, don't continue with instructions
by Richard W.M. Jones 26 Jan '07

26 Jan '07

http://et.redhat.com/~rjones/patches/libvirt-autogen-configure-failure.patch (Very trivial ...) Rich. -- Emerging Technologies, Red Hat http://et.redhat.com/~rjones/ 64 Baker Street, London, W1U 7DF Mobile: +44 7866 314 421 "[Negative numbers] darken the very whole doctrines of the equations and make dark of the things which are in their nature excessively obvious and simple" (Francis Maseres FRS, mathematician, 1759)

2 1

[Libvir] libvirt and accessing remote systems
by Richard W.M. Jones 25 Jan '07

25 Jan '07

This is a follow on to this thread: https://www.redhat.com/archives/libvir-list/2007-January/thread.html#00064 but I think it deserves a thread of its own for discussion. Background: Dan drew this diagram proposing a way to include remote access to systems from within libvirt: http://people.redhat.com/berrange/libvirt/libvirt-arch-remote-2.png libvirt would continue as now to provide direct hypervisor calls, direct access to xend and so on. But in addition, a new backend would be written ("remote") which could talk to a remote daemon ("libvirtd") using some sort of RPC mechanism. Position: I gave this architecture some thought over the weekend, and I like it for the following reasons (some not very technical): * Authentication and encryption is handled entirely within the libvirt / libvirtd library, allowing us to use whatever RPC mechanism we like on top of a selection of transports of our choosing (eg. GnuTLS, ssh, unencrypted TCP sockets, ...) * We don't need to modify xend at all, and additionally we won't need to modify future flavour-of-the-month virtual machine monitors. I have a particular issue with xend (written in Python) because in my own tests I've seen my Python XMLRPC/SSL server actually segfault. It doesn't inspire me that this Python solution is adding anything more than apparent security. * The architecture is very flexible: It allows virt-manager to run as root or as non-root, according to customer wishes. virt-manager can make direct HV calls, or everything can be remoted, and it's easy to explain to the user about the performance vs management trade-offs. * It's relatively easy to implement. Note that libvirtd is just a thin server layer linked to its own copy of libvirt. * Another proposal was to make all libvirt calls remote (http://people.redhat.com/berrange/libvirt/libvirt-arch-remote-3.png) but I don't think this is a going concern because (1) it requires a daemon always be run, which is another installation problem and another chance for sysadmins to give up, and (2) the perception will be that this is slow, whether or not that is actually true. Now some concerns: * libvirtd will likely need to be run as root, so another root daemon written in C listening on a public port. (On the other hand, xend listening on a public port also isn't too desirable, even with authentication). * If Xen upstream in the meantime come up with a secure remote access method then potentially this means clients could have to choose between the two, or run two services (libvirtd + Xen/remote). * There are issues with versioning the remote API. Do we allow different versions of libvirt/libvirtd to talk to each other? Do we provide backwards compatibility when we move to a new API? * Do we allow more than one client to talk to a libvirtd daemon (no | multiple readers one writer | multiple readers & writers). * What's the right level to make a remote API? Should we batch calls up together? RPC mechanism: I've been investigating RPC mechanisms and there seem to be two reasonable possibilities, SunRPC and XMLRPC. (Both would need to run over some sort of secure connection, so there is a layer below both). My analysis of those is here: http://et.redhat.com/~rjones/secure_rpc/ Rich. -- Red Hat UK Ltd. 64 Baker Street, London, W1U 7DF Mobile: +44 7866 314 421 (will change soon)

4 10

[Libvir] what is virt-install doing differently?
by Aron Griffis 23 Jan '07

23 Jan '07

I'm looking at a problem where HVM xen/ia64 domains hang on boot using virt-install, specifically at this point: ACPI: Core revision 20060707 Boot processor id 0x0/0x0 Given VT-I capable hardware, this can be demonstrated with: virt-install -n rhel5hvm1 -v -r 1024 --vcpus=2 -c \ /root/RHEL5-Server-20070112.3-ia64-DVD.iso -f \ /var/lib/xen/images/rhel5hvm1 -s 20 --vnc The really strange thing is that the boot works fine, using the same configuration, if virt-install is not involved. Actually, I modify the configuration slightly to include the CD-ROM, but otherwise it is the same: # cat /etc/xen/rhel5rc1s7hvm2 # Automatically generated xen config file name = "rhel5rc1s7hvm2" builder = "hvm" memory = "1024" disk = [ 'phy:/dev/cciss/c0d1p1,hda,w', 'file:/root/RHEL5-Server-20070112.3-ia64-DVD.iso,hdc:cdrom,r' ] vif = [ 'type=ioemu, mac=00:16:3e:50:4c:95, bridge=xenbr1', ] uuid = "c228739b-9e6c-bb11-47d8-281ca2edf750" device_model = "/usr/lib/xen/bin/qemu-dm" kernel = "/usr/lib/xen/boot/hvmloader" vnc=1 vncunused=1 apic=1 acpi=1 vcpus=2 serial = "pty" # enable serial console on_reboot = 'restart' on_crash = 'restart' I don't understand what would be different about the virt-install boot vs. booting straight from the generated configuration. Can somebody shed some light on this? Thanks, Aron

2 1

[Libvir] virtinst 0.100.0 and virt-manager 0.3.0 releases
by Daniel P. Berrange 23 Jan '07

23 Jan '07

Following on from the recent libvirt 0.1.11 release, I'm pleased to announce new releases of the virtinst & virt-manager applications. The virtinst 0.100.0 release is a major re-factoring to make use of the libvirt inactive domain management APIs when provisioning guests. It also tries to clean up many of the Xen specific bits to facilitate future use with the QEMU backend of libvirt. From a usability point of view it also now displays progress information when downloading kernel & initrd and creating the filesystem images. The virt-manager 0.3.0 release brings a major functionality update, enabling of inactive domain management. This requires libvirt at least version 0.1.11 to provide implementations of inactive domain management for Xen 3.0.3 and Xen 3.0.4. With this new functionality the display will list all guests which are in the 'shutoff' state. The guest can be started with the 'Run' button in the virtual console window. The vistinst package must also be updated to at least version 0.100.0 to ensure that during provisioning of guests it uses the new inactive domain management APIs. Finally there have been a variety of minor UI fixes & enhancements such as progress bars during guest creation, reliability fixes to the virtual console and even greater coverage for translations. As of today, virtinst is now formally a part of the virt-manager project, so downloads for both virtinst & virt-manager are available from the same download page: http://virt-manager.et.redhat.com/downloads.html All historical release from virtinst 0.95.0 onwards are available there. In addition, to encourage / facilitate broader community development & feedback on the virtinst/virt-manager applications we have decided to designate et-mgmt-tools(a)redhat.com as the project primary mailing list: http://www.redhat.com/mailman/listinfo/et-mgmt-tools We welcome feedback on what additional capabilities / features people would like to see in future virtinst / virt-manager releases. As a taste of things to come, we have active developments plans for: - Secure authenticated remote management - Support for QEMU & KVM virtualization Regards, Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=|

1 1

[Libvir] Release of libvirt-0.1.11
by Daniel Veillard 22 Jan '07

22 Jan '07

I just made a release, last one was one month ago, there is a leak fixed and this make it simpler to commit new code for QEmu and remote support: * Finish XML XM config files support (Daniel Berrange) * Remove memory leak when freeing virConf objects (Daniel Berrange) * Finishing inactive domain support (Daniel Berrange) * Added a Relax-NG schemas to check XML instances Available as usual at ftp://libvirt.org/libvirt/ thanks everybody ! Daniel -- Red Hat Virtualization group http://redhat.com/virtualization/ Daniel Veillard | virtualization library http://libvirt.org/ veillard(a)redhat.com | libxml GNOME XML XSLT toolkit http://xmlsoft.org/ http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/

1 0

[Libvir] Tricky provisioning problem with inactive domains
by Daniel P. Berrange 17 Jan '07

17 Jan '07

Adding support for inactive domains was supposed to make everyone's life easier, but as luck would have it, its actually made one thing very much harder. In the virt-inst/virt-manager tools provisioning works something like this: In paravirt case: - Create a guest using an explicit kernel/initrd from the images/xen directory on the install CD - Write a config file to /etc/xen setup to boot using pygrub In fullvirt case: - Create a guest booting directly off a CDROM - Write a config file to /etc/xen setup to boot off the harddisk So in both these cases, the libvirt XML config for the very first boot of the guest is different, from the XML config for subsequent boots. With the new inactive domain support in libvirt & xend, we can't write out config files directly, instead there is the virDomainDefine() API, which calls to appropriate APIs in XenD. And this is where the problem arises: 1. If we call virDomainDefine() to write the long term config, then virDomainStart() will not be using the correct boot method for the initial install. 2. If we call virDomainDefine() to write the initial install config, then virDomainStart() will kick off the install correctly, but on subsequent boots we'll end up booting the installer again instead of the just installed OS. 3. We could just use virDomainCreate() to start installer, and try to use virDomainDefine() to write the long term config - the latter call will fail though because there will already be a running guest with that name. 4. Conversely if we use virDomainDefine() to write the config, and then tried to create a one-off guest with virDomainCreate() the latter will fail due to duplicate names. So, thus far the only way out of the trap I can think of is: 1. Use virDomainCreate() to kick off the initial install 2. Poll virDomainLookupByXXX() to watch for this initial guest shutting down 3. Write out the persistent config using virDomainDefine() The big problem with this, is that if the user were to exit virt-manager sometime after the guest install starts, but before step 3, the config for the guest will never be written, even though it has successfully installed. There are two further ideas I've had - both requiring additional APIs in libvirt & probably XenD too. - Make it possible change the boot configuration of an existing guest. This would let us do: 1. Use virDomainDefine() to define a config file suitable for installing the guest, ie using explicit kernel/initrd 2. Use virDomainStart() to kick off the installer 3. Uew new API to change the guest config to remove explicit kernel & initrd config, and add a bootloader for pygrub. Or in HVM case, switch boot order to use harddisk instead of CDROM & detach the CDROM device. - Make it possible to start an existing inactive guest using an alternative one-off configuration. This would let us do: 1. Use virDomainDefine() to define a config file suitable for running the guest during normal operation. 2. Use virDomainStartConfig(xml) to start the guest with a special config suitable for installing the guest. Ultimately I think we do need the means to change arbitrary parts of a guest configuration, so I think the first option would be the preferred approach. The trouble is that I think implementing this would require using the new XenAPI, or adding a number of new methods to the legacy SXPR API which is not really very desirable. Regards, Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=|

5 7

Re: [Libvir] Virtual networking
by Richard W.M. Jones 16 Jan '07

16 Jan '07

Hugh Brock wrote: > Daniel P. Berrange wrote: > >> 3. The way I think you re suggesting - a libvirt server on every remote >> host which calls into the regular libvirt internal driver model to >> proxy remote calls. So even if the hypervisor in question provides a >> remote network management API, we will always use the local API and >> do *all* remote networking via the libvirt server >> >> http://people.redhat.com/berrange/libvirt/libvirt-arch-remote-2.png >> > This strikes me as *much* easier to manage, and the most consistent > thus far with the idea that libvirt should remain as > hypervisor-neutral as possible. I guess the management issue is going to be versioning the protocol. If the protocol is just a direct mapping of vir* calls and structures then you'll quickly end up in a situation where even the smallest change requires you to upgrade the world or old versions have to be maintained indefinitely. That's not saying I don't like the idea. Rich. -- Red Hat UK Ltd. 64 Baker Street, London, W1U 7DF Mobile: +44 7866 314 421 (will change soon)

2 1

Re: [Libvir] Virtual networking
by Hugh Brock 16 Jan '07

16 Jan '07

Richard W.M. Jones wrote: > Hugh Brock wrote: >> Daniel P. Berrange wrote: >>> 3. The way I think you re suggesting - a libvirt server on every remote >>> host which calls into the regular libvirt internal driver model to >>> proxy remote calls. So even if the hypervisor in question provides a >>> remote network management API, we will always use the local API and >>> do *all* remote networking via the libvirt server >>> >>> http://people.redhat.com/berrange/libvirt/libvirt-arch-remote-2.png >>> >> This strikes me as *much* easier to manage, and the most consistent >> thus far with the idea that libvirt should remain as >> hypervisor-neutral as possible. > > I guess the management issue is going to be versioning the protocol. If > the protocol is just a direct mapping of vir* calls and structures then > you'll quickly end up in a situation where even the smallest change > requires you to upgrade the world or old versions have to be maintained > indefinitely. > > That's not saying I don't like the idea. > True enough... we're guaranteeing we're going to have backwards compatibility problems. On the other hand the libvirt API is supposed to be held pretty stable. DV, any thoughts? --H

1 0

Re: [Libvir] Certificate management APIs ?
by Richard W.M. Jones 16 Jan '07

16 Jan '07

[Apologies also that this is not threaded with the original post] > $HOME/.libvirt/tls/ > | > +- ca > | | > | +- cert.pem > | +- ca-crl.pem Note that there are standard locations for CA certs. On my Debian box the standard locations appear to be /etc/ca-certificates.conf and /usr/share/ca-certificates. Not sure yet about Fedora/RHEL. I suppose you hope that people will be using formal CA's rather than their own, or at least have a CA certificate issued by a formal CA from which they can issue their own client & server certs. Rich. -- Red Hat UK Ltd. 64 Baker Street, London, W1U 7DF Mobile: +44 7866 314 421 (will change soon)

3 4