path: root/xlators/mount/fuse/src/fuse-helpers.c

/*
   Copyright (c) 2010-2012 Red Hat, Inc. <http://www.redhat.com>
   This file is part of GlusterFS.

   This file is licensed to you under your choice of the GNU Lesser
   General Public License, version 3 or any later version (LGPLv3 or
   later), or the GNU General Public License, version 2 (GPLv2), in all
   cases as published by the Free Software Foundation.
*/
#ifdef __NetBSD__
#define _KMEMUSER
#endif

#include "fuse-bridge.h"
#if defined(GF_SOLARIS_HOST_OS)
#include <sys/procfs.h>
#else
#include <sys/sysctl.h>
#endif


static void
fuse_resolve_wipe (fuse_resolve_t *resolve)
{
        GF_FREE ((void *)resolve->path);

        GF_FREE ((void *)resolve->bname);

        if (resolve->fd)
                fd_unref (resolve->fd);

        loc_wipe (&resolve->resolve_loc);

	if (resolve->hint) {
		inode_unref (resolve->hint);
		resolve->hint = 0;
	}

	if (resolve->parhint) {
		inode_unref (resolve->parhint);
		resolve->parhint = 0;
	}
}


void
free_fuse_state (fuse_state_t *state)
{
        xlator_t       *this     = NULL;
        fuse_private_t *priv     = NULL;
        uint64_t        winds    = 0;
        char            switched = 0;

        this = state->this;

        priv = this->private;

        loc_wipe (&state->loc);

        loc_wipe (&state->loc2);

        if (state->xdata) {
                dict_unref (state->xdata);
                state->xdata = (void *)0xaaaaeeee;
        }
        if (state->xattr)
                dict_unref (state->xattr);

        if (state->name) {
                GF_FREE (state->name);
                state->name = NULL;
        }
        if (state->fd) {
                fd_unref (state->fd);
                state->fd = (void *)0xfdfdfdfd;
        }
        if (state->finh) {
                GF_FREE (state->finh);
                state->finh = NULL;
        }

        fuse_resolve_wipe (&state->resolve);
        fuse_resolve_wipe (&state->resolve2);

        pthread_mutex_lock (&priv->sync_mutex);
        {
                winds = --state->active_subvol->winds;
                switched = state->active_subvol->switched;
        }
        pthread_mutex_unlock (&priv->sync_mutex);

        if ((winds == 0) && (switched)) {
                xlator_notify (state->active_subvol, GF_EVENT_PARENT_DOWN,
                               state->active_subvol, NULL);
        }

#ifdef DEBUG
        memset (state, 0x90, sizeof (*state));
#endif
        GF_FREE (state);
        state = NULL;
}


fuse_state_t *
get_fuse_state (xlator_t *this, fuse_in_header_t *finh)
{
        fuse_state_t   *state         = NULL;
	xlator_t       *active_subvol = NULL;
        fuse_private_t *priv          = NULL;

        state = (void *)GF_CALLOC (1, sizeof (*state),
                                   gf_fuse_mt_fuse_state_t);
        if (!state)
                return NULL;

	state->this = THIS;
        priv = this->private;

        pthread_mutex_lock (&priv->sync_mutex);
        {
                active_subvol = fuse_active_subvol (state->this);
                active_subvol->winds++;
        }
        pthread_mutex_unlock (&priv->sync_mutex);

	state->active_subvol = active_subvol;
	state->itable = active_subvol->itable;

        state->pool = this->ctx->pool;
        state->finh = finh;
        state->this = this;

        LOCK_INIT (&state->lock);

        return state;
}


#define FUSE_MAX_AUX_GROUPS 32 /* We can get only up to 32 aux groups from /proc */
void
frame_fill_groups (call_frame_t *frame)
{
#if defined(GF_LINUX_HOST_OS)
        char         filename[32];
        char         line[4096];
        char        *ptr = NULL;
        FILE        *fp = NULL;
        int          idx = 0;
        long int     id = 0;
        char        *saveptr = NULL;
        char        *endptr = NULL;
        int          ret = 0;

        ret = snprintf (filename, sizeof filename, "/proc/%d/status", frame->root->pid);
        if (ret >= sizeof filename)
                goto out;

        fp = fopen (filename, "r");
        if (!fp)
                goto out;

	if (call_stack_alloc_groups (frame->root, FUSE_MAX_AUX_GROUPS) != 0)
		goto out;

        while ((ptr = fgets (line, sizeof line, fp))) {
                if (strncmp (ptr, "Groups:", 7) != 0)
                        continue;

                ptr = line + 8;

                for (ptr = strtok_r (ptr, " \t\r\n", &saveptr);
                     ptr;
                     ptr = strtok_r (NULL, " \t\r\n", &saveptr)) {
                        errno = 0;
                        id = strtol (ptr, &endptr, 0);
                        if (errno == ERANGE)
                                break;
                        if (!endptr || *endptr)
                                break;
                        frame->root->groups[idx++] = id;
                        if (idx == FUSE_MAX_AUX_GROUPS)
                                break;
                }

                frame->root->ngrps = idx;
                break;
        }
out:
        if (fp)
                fclose (fp);
#elif defined(GF_SOLARIS_HOST_OS)
        char         filename[32];
        char         scratch[128];
        prcred_t    *prcred = (prcred_t *) scratch;
        FILE        *fp = NULL;
        int          ret = 0;
	int          ngrps;

        ret = snprintf (filename, sizeof filename,
                        "/proc/%d/cred", frame->root->pid);

        if (ret < sizeof filename) {
                fp = fopen (filename, "r");
                if (fp != NULL) {
                        if (fgets (scratch, sizeof scratch, fp) != NULL) {
                                ngrps = MIN(prcred->pr_ngroups,
					    GF_MAX_AUX_GROUPS);
				if (call_stack_alloc_groups (frame->root,
							     ngrps) != 0)
					return;
                        }
                        fclose (fp);
                 }
         }
#elif defined(CTL_KERN) /* DARWIN and *BSD */
        /* 
           N.B. CTL_KERN is an enum on Linux. (Meaning, if it's not
           obvious, that it's not subject to preprocessor directives 
           like '#if defined'.)
           Unlike Linux, on Mac OS and the BSDs it is a #define. We
           could test to see that KERN_PROC is defined, but, barring any 
           evidence to the contrary, I think that's overkill.
           We might also test that GF_DARWIN_HOST_OS is defined, why
           limit this to just Mac OS. It's equally valid for the BSDs
           and we do have people building on NetBSD and FreeBSD.
        */
        int name[] = { CTL_KERN, KERN_PROC, KERN_PROC_PID, frame->root->pid };
        size_t namelen = sizeof name / sizeof name[0];
        struct kinfo_proc kp;
        size_t kplen = sizeof(kp);
        int i, ngroups;

        if (sysctl(name, namelen, &kp, &kplen, NULL, 0) != 0)
                return;
        ngroups = MIN(kp.kp_eproc.e_ucred.cr_ngroups, GF_MAX_AUX_GROUPS);
	if (call_stack_alloc_groups (frame->root, ngroups) != 0)
		return;
        for (i = 0; i < ngroups; i++)
                frame->root->groups[i] = kp.kp_eproc.e_ucred.cr_groups[i];
        frame->root->ngrps = ngroups;
#else
        frame->root->ngrps = 0;
#endif /* GF_LINUX_HOST_OS */
}

/*
 * Get the groups for the PID associated with this frame. If enabled,
 * use the gid cache to reduce group list collection.
 */
static void get_groups(fuse_private_t *priv, call_frame_t *frame)
{
	int i;
	const gid_list_t *gl;
	gid_list_t agl;

        if (-1 == priv->gid_cache_timeout) {
                frame->root->ngrps = 0;
                return;
        }

	if (!priv->gid_cache_timeout) {
		frame_fill_groups(frame);
		return;
	}

	gl = gid_cache_lookup(&priv->gid_cache, frame->root->pid,
			      frame->root->uid, frame->root->gid);
	if (gl) {
		if (call_stack_alloc_groups (frame->root, gl->gl_count) != 0)
			return;
		frame->root->ngrps = gl->gl_count;
		for (i = 0; i < gl->gl_count; i++)
			frame->root->groups[i] = gl->gl_list[i];
		gid_cache_release(&priv->gid_cache, gl);
		return;
	}

	frame_fill_groups (frame);

	agl.gl_id = frame->root->pid;
	agl.gl_uid = frame->root->uid;
	agl.gl_gid = frame->root->gid;
	agl.gl_count = frame->root->ngrps;
	agl.gl_list = GF_CALLOC(frame->root->ngrps, sizeof(gid_t),
			gf_fuse_mt_gids_t);
	if (!agl.gl_list)
		return;

	for (i = 0; i < frame->root->ngrps; i++)
		agl.gl_list[i] = frame->root->groups[i];

	if (gid_cache_add(&priv->gid_cache, &agl) != 1)
		GF_FREE(agl.gl_list);
}

call_frame_t *
get_call_frame_for_req (fuse_state_t *state)
{
        call_pool_t           *pool = NULL;
        fuse_in_header_t      *finh = NULL;
        call_frame_t          *frame = NULL;
        xlator_t              *this = NULL;
        fuse_private_t        *priv = NULL;

        pool = state->pool;
        finh = state->finh;
        this = state->this;
        priv = this->private;

        frame = create_frame (this, pool);
        if (!frame)
                return NULL;

        if (finh) {
                frame->root->uid      = finh->uid;
                frame->root->gid      = finh->gid;
                frame->root->pid      = finh->pid;
                frame->root->unique   = finh->unique;
                set_lk_owner_from_uint64 (&frame->root->lk_owner,
                                          state->lk_owner);
        }

	get_groups(priv, frame);

        if (priv && priv->client_pid_set)
                frame->root->pid = priv->client_pid;

        frame->root->type = GF_OP_TYPE_FOP;

        return frame;
}


inode_t *
fuse_ino_to_inode (uint64_t ino, xlator_t *fuse)
{
        inode_t  *inode = NULL;
        xlator_t *active_subvol = NULL;

        if (ino == 1) {
                active_subvol = fuse_active_subvol (fuse);
                if (active_subvol)
                        inode = active_subvol->itable->root;
        } else {
                inode = (inode_t *) (unsigned long) ino;
                inode_ref (inode);
        }

        return inode;
}

uint64_t
inode_to_fuse_nodeid (inode_t *inode)
{
        if (!inode)
		return 0;
	if (__is_root_gfid (inode->gfid))
                return 1;

        return (unsigned long) inode;
}


GF_MUST_CHECK int32_t
fuse_loc_fill (loc_t *loc, fuse_state_t *state, ino_t ino,
               ino_t par, const char *name)
{
        inode_t  *inode = NULL;
        inode_t  *parent = NULL;
        int32_t   ret = -1;
        char     *path = NULL;
        uuid_t    null_gfid = {0,};

        /* resistance against multiple invocation of loc_fill not to get
           reference leaks via inode_search() */

        if (name) {
                parent = loc->parent;
                if (!parent) {
                        parent = fuse_ino_to_inode (par, state->this);
                        loc->parent = parent;
                        if (parent)
                                uuid_copy (loc->pargfid, parent->gfid);
                }

                inode = loc->inode;
                if (!inode) {
                        inode = inode_grep (parent->table, parent, name);
                        loc->inode = inode;
                }

                ret = inode_path (parent, name, &path);
                if (ret <= 0) {
                        gf_log ("glusterfs-fuse", GF_LOG_DEBUG,
                                "inode_path failed for %s/%s",
                                (parent)?uuid_utoa (parent->gfid):"0", name);
                        goto fail;
                }
                loc->path = path;
        } else {
                inode = loc->inode;
                if (!inode) {
                        inode = fuse_ino_to_inode (ino, state->this);
                        loc->inode = inode;
                        if (inode)
                                uuid_copy (loc->gfid, inode->gfid);
                }

                parent = loc->parent;
                if (!parent) {
                        parent = inode_parent (inode, null_gfid, NULL);
                        loc->parent = parent;
                        if (parent)
                                uuid_copy (loc->pargfid, parent->gfid);

                }

                ret = inode_path (inode, NULL, &path);
                if (ret <= 0) {
                        gf_log ("glusterfs-fuse", GF_LOG_DEBUG,
                                "inode_path failed for %s",
                                (inode) ? uuid_utoa (inode->gfid) : "0");
                        goto fail;
                }
                loc->path = path;
        }

        if (loc->path) {
                loc->name = strrchr (loc->path, '/');
                if (loc->name)
                        loc->name++;
                else
                        loc->name = "";
        }

        if ((ino != 1) && (parent == NULL)) {
                gf_log ("fuse-bridge", GF_LOG_DEBUG,
                        "failed to search parent for %"PRId64"/%s (%"PRId64")",
                        (ino_t)par, name, (ino_t)ino);
                ret = -1;
                goto fail;
        }
        ret = 0;
fail:
        /* this should not happen as inode_path returns -1 when buf is NULL
           for sure */
        if (path && !loc->path)
                GF_FREE (path);
        return ret;
}

/* Use the same logic as the Linux NFS-client */
#define GF_FUSE_SQUASH_INO(ino) ((uint32_t) ino) ^ (ino >> 32)

/* courtesy of folly */
void
gf_fuse_stat2attr (struct iatt *st, struct fuse_attr *fa, gf_boolean_t enable_ino32)
{
        if (enable_ino32)
                fa->ino = GF_FUSE_SQUASH_INO(st->ia_ino);
        else
                fa->ino = st->ia_ino;

        fa->size        = st->ia_size;
        fa->blocks      = st->ia_blocks;
        fa->atime       = st->ia_atime;
        fa->mtime       = st->ia_mtime;
        fa->ctime       = st->ia_ctime;
        fa->atimensec   = st->ia_atime_nsec;
        fa->mtimensec   = st->ia_mtime_nsec;
        fa->ctimensec   = st->ia_ctime_nsec;
        fa->mode        = st_mode_from_ia (st->ia_prot, st->ia_type);
        fa->nlink       = st->ia_nlink;
        fa->uid         = st->ia_uid;
        fa->gid         = st->ia_gid;
        fa->rdev        = makedev (ia_major (st->ia_rdev),
                                   ia_minor (st->ia_rdev));
#if FUSE_KERNEL_MINOR_VERSION >= 9
        fa->blksize     = st->ia_blksize;
#endif
#ifdef GF_DARWIN_HOST_OS
        fa->crtime      = (uint64_t)-1;
        fa->crtimensec  = (uint32_t)-1;
        fa->flags       = 0;
#endif
}

void
gf_fuse_fill_dirent (gf_dirent_t *entry, struct fuse_dirent *fde, gf_boolean_t enable_ino32)
{
        if (enable_ino32)
                fde->ino = GF_FUSE_SQUASH_INO(entry->d_ino);
        else
                fde->ino = entry->d_ino;

        fde->off         = entry->d_off;
        fde->type        = entry->d_type;
        fde->namelen     = strlen (entry->d_name);
        strncpy (fde->name, entry->d_name, fde->namelen);
}

static int
fuse_do_flip_xattr_ns (char *okey, const char *nns, char **nkey)
{
        int   ret = 0;
        char *key = NULL;

        okey = strchr (okey, '.');
        GF_ASSERT (okey);

        key = GF_CALLOC (1, strlen (nns) + strlen(okey) + 1,
                         gf_common_mt_char);
        if (!key) {
                ret = -1;
                goto out;
        }

        strcpy (key, nns);
        strcat (key, okey);

        *nkey = key;

 out:
        return ret;
}

static int
fuse_xattr_alloc_default (char *okey, char **nkey)
{
        int ret = 0;

        *nkey = gf_strdup (okey);
        if (!*nkey)
                ret = -1;
        return ret;
}

#define PRIV_XA_NS   "trusted"
#define UNPRIV_XA_NS "system"

int
fuse_flip_xattr_ns (fuse_private_t *priv, char *okey, char **nkey)
{
        int             ret       = 0;
        gf_boolean_t    need_flip = _gf_false;

        switch (priv->client_pid) {
        case GF_CLIENT_PID_GSYNCD:
                /* valid xattr(s): *xtime, volume-mark* */
                gf_log("glusterfs-fuse", GF_LOG_DEBUG, "PID: %d, checking xattr(s): "
                       "volume-mark*, *xtime", priv->client_pid);
                if ( (strcmp (okey, UNPRIV_XA_NS".glusterfs.volume-mark") == 0)
                     || (fnmatch (UNPRIV_XA_NS".glusterfs.volume-mark.*", okey, FNM_PERIOD) == 0)
                     || (fnmatch (UNPRIV_XA_NS".glusterfs.*.xtime", okey, FNM_PERIOD) == 0) )
                        need_flip = _gf_true;
                break;

        case GF_CLIENT_PID_HADOOP:
                /* valid xattr(s): pathinfo */
                gf_log("glusterfs-fuse", GF_LOG_DEBUG, "PID: %d, checking xattr(s): "
                       "pathinfo", priv->client_pid);
                if (strcmp (okey, UNPRIV_XA_NS".glusterfs.pathinfo") == 0)
                        need_flip = _gf_true;
                break;
        }

        if (need_flip) {
                gf_log ("glusterfs-fuse", GF_LOG_DEBUG, "flipping %s to "PRIV_XA_NS" equivalent",
                        okey);
                ret = fuse_do_flip_xattr_ns (okey, PRIV_XA_NS, nkey);
        } else {
                /* if we cannot match, continue with what we got */
                ret = fuse_xattr_alloc_default (okey, nkey);
        }

        return ret;
}

int
fuse_ignore_xattr_set (fuse_private_t *priv, char *key)
{
        int ret = 0;

        /* don't mess with user namespace */
        if (fnmatch ("user.*", key, FNM_PERIOD) == 0)
                goto out;

        if (priv->client_pid != GF_CLIENT_PID_GSYNCD)
                goto out;

        /* trusted NS check */
        if (!((fnmatch ("*.glusterfs.*.xtime", key, FNM_PERIOD) == 0)
              || (fnmatch ("*.glusterfs.volume-mark",
                           key, FNM_PERIOD) == 0)
              || (fnmatch ("*.glusterfs.volume-mark.*",
                           key, FNM_PERIOD) == 0)
              || (fnmatch ("glusterfs.gfid.newfile",
                           key, FNM_PERIOD) == 0)))
                ret = -1;

 out:
        gf_log ("glusterfs-fuse", GF_LOG_DEBUG, "%s setxattr: key [%s], "
                " client pid [%d]", (ret ? "disallowing" : "allowing"), key,
                priv->client_pid);

        return ret;
}