X-Git-Url: http://git.onelab.eu/?a=blobdiff_plain;f=python%2Fvserverimpl.c;h=17d1482a36672451d93471de0c7059dbc48e3acc;hb=df6666ed5ab57b500cdd0a0f3ef610025fa10b4f;hp=b5b9c5bce844d3fe4103367ab8b94d987acd8df1;hpb=9f4de1c49ce5a14e8ac580c5b93c88cb5d964e23;p=util-vserver.git diff --git a/python/vserverimpl.c b/python/vserverimpl.c index b5b9c5b..17d1482 100644 --- a/python/vserverimpl.c +++ b/python/vserverimpl.c @@ -4,17 +4,17 @@ Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met: - * Redistributions of source code must retain the above copyright - notice, this list of conditions and the following disclaimer. +* Redistributions of source code must retain the above copyright +notice, this list of conditions and the following disclaimer. - * Redistributions in binary form must reproduce the above - copyright notice, this list of conditions and the following - disclaimer in the documentation and/or other materials provided - with the distribution. +* Redistributions in binary form must reproduce the above +copyright notice, this list of conditions and the following +disclaimer in the documentation and/or other materials provided +with the distribution. - * Neither the name of the copyright holder nor the names of its - contributors may be used to endorse or promote products derived - from this software without specific prior written permission. +* Neither the name of the copyright holder nor the names of its +contributors may be used to endorse or promote products derived +from this software without specific prior written permission. THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT @@ -35,210 +35,514 @@ POSSIBILITY OF SUCH DAMAGE. #include #include +#include +#include +#include #include +#include +#include +#include +#include #include "config.h" #include "pathconfig.h" -#include "planetlab.h" #include "virtual.h" #include "vserver.h" +#include "planetlab.h" #include "vserver-internal.h" +#define NONE ({ Py_INCREF(Py_None); Py_None; }) + /* * context create */ static PyObject * vserver_chcontext(PyObject *self, PyObject *args) { + int ctx_is_new; xid_t ctx; - uint32_t flags = 0; - uint32_t bcaps = ~vc_get_insecurebcaps(); - /* XXX - use appropriate resource values here */ - rspec_t rspec = { 32, VC_VXF_SCHED_FLAGS, -1, -1 }; + uint_least64_t bcaps = 0; - if (!PyArg_ParseTuple(args, "I|K", &ctx, &flags)) + if (!PyArg_ParseTuple(args, "I|K", &ctx, &bcaps)) return NULL; + bcaps |= ~vc_get_insecurebcaps(); - if (pl_chcontext(ctx, flags, bcaps, &rspec)) - PyErr_SetFromErrno(PyExc_OSError); + if ((ctx_is_new = pl_chcontext(ctx, bcaps, 0)) < 0) + return PyErr_SetFromErrno(PyExc_OSError); - return Py_None; + return PyBool_FromLong(ctx_is_new); } static PyObject * -vserver_set_rlimit(PyObject *self, PyObject *args) { - struct vc_rlimit limits; - xid_t xid; - int resource; - PyObject *ret; - - limits.min = VC_LIM_KEEP; - limits.soft = VC_LIM_KEEP; - limits.hard = VC_LIM_KEEP; - - if (!PyArg_ParseTuple(args, "IiL", &xid, &resource, &limits.hard)) - return NULL; - - ret = Py_None; - if (vc_set_rlimit(xid, resource, &limits)) - ret = PyErr_SetFromErrno(PyExc_OSError); - else if (vc_get_rlimit(xid, resource, &limits)==-1) - ret = PyErr_SetFromErrno(PyExc_OSError); - else - ret = Py_BuildValue("L",limits.hard); - - return ret; +vserver_setup_done(PyObject *self, PyObject *args) +{ + xid_t ctx; + + if (!PyArg_ParseTuple(args, "I", &ctx)) + return NULL; + + if (pl_setup_done(ctx) < 0) + return PyErr_SetFromErrno(PyExc_OSError); + + return NONE; +} + +static PyObject * +vserver_isrunning(PyObject *self, PyObject *args) +{ + xid_t ctx; + PyObject *ret; + struct stat statbuf; + char fname[64]; + + if (!PyArg_ParseTuple(args, "I", &ctx)) + return NULL; + + sprintf(fname,"/proc/virtual/%d", ctx); + + if(stat(&fname[0],&statbuf)==0) + ret = PyBool_FromLong(1); + else + ret = PyBool_FromLong(0); + + return ret; +} + +static PyObject * +__vserver_get_rlimit(xid_t xid, int resource) { + struct vc_rlimit limits; + PyObject *ret; + + errno = 0; + if (vc_get_rlimit(xid, resource, &limits)==-1) + ret = PyErr_SetFromErrno(PyExc_OSError); + else + ret = Py_BuildValue("LLL",limits.hard, limits.soft, limits.min); + + return ret; } static PyObject * vserver_get_rlimit(PyObject *self, PyObject *args) { - struct vc_rlimit limits; - xid_t xid; - int resource; - PyObject *ret; + xid_t xid; + int resource; + PyObject *ret; - limits.min = VC_LIM_KEEP; - limits.soft = VC_LIM_KEEP; - limits.hard = VC_LIM_KEEP; + if (!PyArg_ParseTuple(args, "Ii", &xid, &resource)) + ret = NULL; + else + ret = __vserver_get_rlimit(xid, resource); - if (!PyArg_ParseTuple(args, "Ii", &xid, &resource)) - return NULL; + return ret; +} - ret = Py_None; - if (vc_get_rlimit(xid, resource, &limits)==-1) - ret = PyErr_SetFromErrno(PyExc_OSError); - else - ret = Py_BuildValue("L",limits.hard); +static PyObject * +vserver_set_rlimit(PyObject *self, PyObject *args) { + struct vc_rlimit limits; + struct rlimit lim; + xid_t xid; + int resource, lresource; + PyObject *ret; + + limits.min = VC_LIM_KEEP; + limits.soft = VC_LIM_KEEP; + limits.hard = VC_LIM_KEEP; + + if (!PyArg_ParseTuple(args, "IiLLL", &xid, &resource, &limits.hard, &limits.soft, &limits.min)) + return NULL; + + lresource = resource; + switch (resource) { + case VC_VLIMIT_NSOCK: + case VC_VLIMIT_ANON: + case VC_VLIMIT_SHMEM: + goto do_vc_set_rlimit; + case VC_VLIMIT_OPENFD: + lresource = RLIMIT_NOFILE; + break; + default: + break; + } + + getrlimit(lresource,&lim); + if (adjust_lim(&limits,&lim)) { + setrlimit(lresource, &lim); + } - return ret; + do_vc_set_rlimit: + errno = 0; + if (vc_set_rlimit(xid, resource, &limits)==-1) + ret = PyErr_SetFromErrno(PyExc_OSError); + else + ret = __vserver_get_rlimit(xid, resource); + + return ret; } -#if 0 /* * setsched */ static PyObject * vserver_setsched(PyObject *self, PyObject *args) { - xid_t xid; - struct vc_set_sched sched; - struct vc_ctx_flags flags; - unsigned cpuguaranteed = 0; - - sched.set_mask = (VC_VXSM_FILL_RATE | - VC_VXSM_INTERVAL | - VC_VXSM_TOKENS_MIN | - VC_VXSM_TOKENS_MAX); - - if (!PyArg_ParseTuple(args, "I|I|I|I|I|I|I", &xid, - &sched.fill_rate, - &sched.interval, - &sched.tokens, - &sched.tokens_min, - &sched.tokens_max, - &cpuguaranteed)) + xid_t ctx; + uint32_t cpu_share; + uint32_t cpu_sched_flags = VC_VXF_SCHED_FLAGS; + + if (!PyArg_ParseTuple(args, "II|I", &ctx, &cpu_share, &cpu_sched_flags)) return NULL; - flags.flagword = VC_VXF_SCHED_HARD; - flags.mask |= VC_VXF_SCHED_HARD; -#define VC_VXF_SCHED_SHARE 0x00000800ull - if (cpuguaranteed==0) { - flags.flagword |= VC_VXF_SCHED_SHARE; - flags.mask |= VC_VXF_SCHED_SHARE; + /* ESRCH indicates that there are no processes in the context */ + if (pl_setsched(ctx, cpu_share, cpu_sched_flags) && + errno != ESRCH) + return PyErr_SetFromErrno(PyExc_OSError); + + return NONE; +} + +static PyObject * +vserver_get_dlimit(PyObject *self, PyObject *args) +{ + PyObject *res; + char* path; + unsigned xid; + struct vc_ctx_dlimit data; + int r; + + if (!PyArg_ParseTuple(args, "si", &path,&xid)) + return NULL; + + memset(&data, 0, sizeof(data)); + r = vc_get_dlimit(path, xid, 0, &data); + if (r>=0) { + res = Py_BuildValue("(i,i,i,i,i)", + data.space_used, + data.space_total, + data.inodes_used, + data.inodes_total, + data.reserved); + } else { + res = PyErr_SetFromErrno(PyExc_OSError); } - if (vc_set_cflags(xid, &flags) == -1) - return PyErr_SetFromErrno(PyExc_OSError); + return res; +} + + +static PyObject * +vserver_set_dlimit(PyObject *self, PyObject *args) +{ + char* path; + unsigned xid; + struct vc_ctx_dlimit data; + + memset(&data,0,sizeof(data)); + if (!PyArg_ParseTuple(args, "siiiiii", &path, + &xid, + &data.space_used, + &data.space_total, + &data.inodes_used, + &data.inodes_total, + &data.reserved)) + return NULL; - if (vc_set_sched(xid, &sched) == -1) - return PyErr_SetFromErrno(PyExc_OSError); + if ((vc_add_dlimit(path, xid, 0) && errno != EEXIST) || + vc_set_dlimit(path, xid, 0, &data)) + return PyErr_SetFromErrno(PyExc_OSError); - return Py_None; + return NONE; } -/* - * setsched - */ -#endif +static PyObject * +vserver_unset_dlimit(PyObject *self, PyObject *args) +{ + char *path; + unsigned xid; + + if (!PyArg_ParseTuple(args, "si", &path, &xid)) + return NULL; + + if (vc_rem_dlimit(path, xid, 0) && errno != ESRCH) + return PyErr_SetFromErrno(PyExc_OSError); + + return NONE; +} static PyObject * -vserver_get_dlimit(PyObject *self, PyObject *args) +vserver_killall(PyObject *self, PyObject *args) { - PyObject *res; - char* path; - unsigned xid; - struct vcmd_ctx_dlimit_v0 data; - int r; - - if (!PyArg_ParseTuple(args, "si", &path,&xid)) - return NULL; - - memset(&data, 0, sizeof(data)); - data.name = path; - data.flags = 0; - r = vserver(VCMD_get_dlimit, xid, &data); - if (r>=0) { - res = Py_BuildValue("(i,i,i,i,i)", - data.space_used, - data.space_total, - data.inodes_used, - data.inodes_total, - data.reserved); - } else { - res = PyErr_SetFromErrno(PyExc_OSError); - } + xid_t ctx; + int sig; + struct vc_ctx_flags cflags = { + .flagword = 0, + .mask = VC_VXF_PERSISTENT + }; + struct vc_net_flags nflags = { + .flagword = 0, + .mask = VC_NXF_PERSISTENT + }; + + if (!PyArg_ParseTuple(args, "Ii", &ctx, &sig)) + return NULL; + + if (vc_ctx_kill(ctx, 0, sig) && errno != ESRCH) + return PyErr_SetFromErrno(PyExc_OSError); + + if (vc_set_cflags(ctx, &cflags) && errno != ESRCH) + return PyErr_SetFromErrno(PyExc_OSError); - return res; + if (vc_set_nflags(ctx, &nflags) && errno != ESRCH) + return PyErr_SetFromErrno(PyExc_OSError); + + return NONE; } +static PyObject * +vserver_set_bcaps(PyObject *self, PyObject *args) +{ + xid_t ctx; + struct vc_ctx_caps caps; + + if (!PyArg_ParseTuple(args, "IK", &ctx, &caps.bcaps)) + return NULL; + + caps.bmask = vc_get_insecurebcaps(); + caps.cmask = caps.ccaps = 0; + if (vc_set_ccaps(ctx, &caps) == -1 && errno != ESRCH) + return PyErr_SetFromErrno(PyExc_OSError); + + return NONE; +} static PyObject * -vserver_set_dlimit(PyObject *self, PyObject *args) +vserver_text2bcaps(PyObject *self, PyObject *args) { - char* path; - unsigned xid; - struct vcmd_ctx_dlimit_base_v0 init; - struct vcmd_ctx_dlimit_v0 data; - - memset(&data,0,sizeof(data)); - if (!PyArg_ParseTuple(args, "siiiiii", &path, - &xid, - &data.space_used, - &data.space_total, - &data.inodes_used, - &data.inodes_total, - &data.reserved)) - return NULL; - - data.name = path; - data.flags = 0; - - memset(&init, 0, sizeof(init)); - init.name = path; - init.flags = 0; - - if ((vserver(VCMD_add_dlimit, xid, &init) && errno != EEXIST) || - vserver(VCMD_set_dlimit, xid, &data)) - return PyErr_SetFromErrno(PyExc_OSError); - - return Py_None; + struct vc_ctx_caps caps = { .bcaps = 0 }; + const char *list; + int len; + struct vc_err_listparser err; + + if (!PyArg_ParseTuple(args, "s#", &list, &len)) + return NULL; + + vc_list2bcap(list, len, &err, &caps); + + return Py_BuildValue("K", caps.bcaps); +} + +static PyObject * +vserver_get_bcaps(PyObject *self, PyObject *args) +{ + xid_t ctx; + struct vc_ctx_caps caps; + + if (!PyArg_ParseTuple(args, "I", &ctx)) + return NULL; + + if (vc_get_ccaps(ctx, &caps) == -1) { + if (errno != -ESRCH) + return PyErr_SetFromErrno(PyExc_OSError); + else + caps.bcaps = 0; + } + + return Py_BuildValue("K", caps.bcaps & vc_get_insecurebcaps()); +} + +static PyObject * +vserver_bcaps2text(PyObject *self, PyObject *args) +{ + struct vc_ctx_caps caps = { .bcaps = 0 }; + PyObject *list; + const char *cap; + + if (!PyArg_ParseTuple(args, "K", &caps.bcaps)) + return NULL; + + list = PyString_FromString(""); + + while ((cap = vc_lobcap2text(&caps.bcaps)) != NULL) { + if (list == NULL) + break; + PyString_ConcatAndDel(&list, PyString_FromFormat( + (PyString_Size(list) > 0 ? ",CAP_%s" : "CAP_%s" ), + cap)); + } + + return list; +} + +static const struct AF_to_vcNET { + int af; + vc_net_nx_type vc_net; + size_t len; + size_t offset; +} converter[] = { + { AF_INET, vcNET_IPV4, sizeof(struct in_addr), offsetof(struct sockaddr_in, sin_addr.s_addr) }, + { AF_INET6, vcNET_IPV6, sizeof(struct in6_addr), offsetof(struct sockaddr_in6, sin6_addr.s6_addr) }, + { 0, 0 } +}; + +static inline int +convert_address(const char *str, vc_net_nx_type *type, void *dst) +{ + const struct AF_to_vcNET *i; + for (i = converter; i->af; i++) { + if (inet_pton(i->af, str, dst)) { + *type = i->vc_net; + return 0; + } + } + return -1; +} + +static int +get_mask(struct vc_net_nx *addr) +{ + const struct AF_to_vcNET *i; + struct ifaddrs *head, *ifa; + int ret = 0; + + for (i = converter; i->af; i++) { + if (i->vc_net == addr->type) + break; + } + if (!i) { + errno = EINVAL; + return -1; + } + + if (getifaddrs(&head) == -1) + return -1; + for (ifa = head; ifa; ifa = ifa->ifa_next) { + if (ifa->ifa_addr->sa_family == i->af && + memcmp((char *) ifa->ifa_addr + i->offset, addr->ip, i->len) == 0) { + switch (addr->type) { + case vcNET_IPV4: + memcpy(&addr->mask[0], ifa->ifa_netmask + i->offset, i->len); + break; + case vcNET_IPV6: { + uint32_t *m = ((struct sockaddr_in6 *) ifa->ifa_netmask)->sin6_addr.s6_addr32; + /* optimization for the common case */ + if ((m[1] & 1) == 1 && (m[2] & 0x80000000) == 0) + addr->mask[0] = 64; + else { + addr->mask[0] = 0; + while (m[addr->mask[0] / 32] & (addr->mask[0] % 32)) + addr->mask[0]++; + } + break; + } + ret = 1; + break; + } + } + /* no match, use a default */ + if (!ret) { + switch (addr->type) { + case vcNET_IPV4: addr->mask[0] = htonl(0xffffff00); break; + case vcNET_IPV6: addr->mask[0] = 64; break; + default: addr->mask[0] = 0; break; + } + } + freeifaddrs(head); + return ret; +} + +/* XXX These two functions are really similar */ +static PyObject * +vserver_net_add(PyObject *self, PyObject *args) +{ + struct vc_net_nx addr; + nid_t nid; + const char *ip; + + if (!PyArg_ParseTuple(args, "Is", &nid, &ip)) + return NULL; + + if (convert_address(ip, &addr.type, &addr.ip) == -1) + return PyErr_Format(PyExc_ValueError, "%s is not a valid IP address", ip); + + switch (get_mask(&addr)) { + case -1: + return PyErr_SetFromErrno(PyExc_OSError); + case 0: + /* XXX error here? */ + break; + } + addr.count = 1; + + if (vc_net_add(nid, &addr) == -1 && errno != ESRCH) + return PyErr_SetFromErrno(PyExc_OSError); + + return NONE; +} + +static PyObject * +vserver_net_remove(PyObject *self, PyObject *args) +{ + struct vc_net_nx addr; + nid_t nid; + const char *ip; + + if (!PyArg_ParseTuple(args, "Is", &nid, &ip)) + return NULL; + + if (strcmp(ip, "all") == 0) + addr.type = vcNET_ANY; + else if (strcmp(ip, "all4") == 0) + addr.type = vcNET_IPV4A; + else if (strcmp(ip, "all6") == 0) + addr.type = vcNET_IPV6A; + else + if (convert_address(ip, &addr.type, &addr.ip) == -1) + return PyErr_Format(PyExc_ValueError, "%s is not a valid IP address", ip); + + switch (get_mask(&addr)) { + case -1: + return PyErr_SetFromErrno(PyExc_OSError); + } + addr.count = 1; + + if (vc_net_remove(nid, &addr) == -1 && errno != ESRCH) + return PyErr_SetFromErrno(PyExc_OSError); + + return NONE; } static PyMethodDef methods[] = { { "chcontext", vserver_chcontext, METH_VARARGS, "chcontext to vserver with provided flags" }, -#if 0 + { "setup_done", vserver_setup_done, METH_VARARGS, + "Release vserver setup lock" }, { "setsched", vserver_setsched, METH_VARARGS, "Change vserver scheduling attributes for given vserver context" }, -#endif { "setdlimit", vserver_set_dlimit, METH_VARARGS, "Set disk limits for given vserver context" }, + { "unsetdlimit", vserver_unset_dlimit, METH_VARARGS, + "Remove disk limits for given vserver context" }, { "getdlimit", vserver_get_dlimit, METH_VARARGS, "Get disk limits for given vserver context" }, { "setrlimit", vserver_set_rlimit, METH_VARARGS, "Set resource limits for given resource of a vserver context" }, { "getrlimit", vserver_get_rlimit, METH_VARARGS, "Get resource limits for given resource of a vserver context" }, + { "killall", vserver_killall, METH_VARARGS, + "Send signal to all processes in vserver context" }, + { "isrunning", vserver_isrunning, METH_VARARGS, + "Check if vserver is running"}, + { "setbcaps", vserver_set_bcaps, METH_VARARGS, + "Set POSIX capabilities of a vserver context" }, + { "getbcaps", vserver_get_bcaps, METH_VARARGS, + "Get POSIX capabilities of a vserver context" }, + { "text2bcaps", vserver_text2bcaps, METH_VARARGS, + "Translate a string of capabilities to a bitmap" }, + { "bcaps2text", vserver_bcaps2text, METH_VARARGS, + "Translate a capability-bitmap into a string" }, + { "netadd", vserver_net_add, METH_VARARGS, + "Assign an IP address to a context" }, + { "netremove", vserver_net_remove, METH_VARARGS, + "Remove IP address(es) from a context" }, { NULL, NULL, 0, NULL } }; @@ -256,6 +560,28 @@ initvserverimpl(void) PyModule_AddStringConstant(mod, "VSERVER_BASEDIR", DEFAULT_VSERVERDIR); /* export limit-related constants */ - PyModule_AddIntConstant(mod, "DLIMIT_KEEP", (int)CDLIM_KEEP); - PyModule_AddIntConstant(mod, "DLIMIT_INF", (int)CDLIM_INFINITY); + PyModule_AddIntConstant(mod, "DLIMIT_KEEP", (int)VC_CDLIM_KEEP); + PyModule_AddIntConstant(mod, "DLIMIT_INF", (int)VC_CDLIM_INFINITY); + PyModule_AddIntConstant(mod, "VC_LIM_KEEP", (int)VC_LIM_KEEP); + + PyModule_AddIntConstant(mod, "RLIMIT_CPU", (int)RLIMIT_CPU); + PyModule_AddIntConstant(mod, "RLIMIT_RSS", (int)RLIMIT_RSS); + PyModule_AddIntConstant(mod, "RLIMIT_NPROC", (int)RLIMIT_NPROC); + PyModule_AddIntConstant(mod, "RLIMIT_NOFILE", (int)RLIMIT_NOFILE); + PyModule_AddIntConstant(mod, "RLIMIT_MEMLOCK", (int)RLIMIT_MEMLOCK); + PyModule_AddIntConstant(mod, "RLIMIT_AS", (int)RLIMIT_AS); + PyModule_AddIntConstant(mod, "RLIMIT_LOCKS", (int)RLIMIT_LOCKS); + + PyModule_AddIntConstant(mod, "RLIMIT_SIGPENDING", (int)RLIMIT_SIGPENDING); + PyModule_AddIntConstant(mod, "RLIMIT_MSGQUEUE", (int)RLIMIT_MSGQUEUE); + + PyModule_AddIntConstant(mod, "VLIMIT_NSOCK", (int)VC_VLIMIT_NSOCK); + PyModule_AddIntConstant(mod, "VLIMIT_OPENFD", (int)VC_VLIMIT_OPENFD); + PyModule_AddIntConstant(mod, "VLIMIT_ANON", (int)VC_VLIMIT_ANON); + PyModule_AddIntConstant(mod, "VLIMIT_SHMEM", (int)VC_VLIMIT_SHMEM); + + /* scheduler flags */ + PyModule_AddIntConstant(mod, + "VS_SCHED_CPU_GUARANTEED", + VS_SCHED_CPU_GUARANTEED); }