From 18ef831cac81a6bd2336c73dda357d9d69f8fd25 Mon Sep 17 00:00:00 2001

From: "Dr. David Alan Gilbert" <dgilbert@redhat.com>

Date: Mon, 27 Jan 2020 19:00:43 +0100

Subject: [PATCH 012/116] virtiofsd: Add passthrough_ll

MIME-Version: 1.0

Content-Type: text/plain; charset=UTF-8

Content-Transfer-Encoding: 8bit

RH-Author: Dr. David Alan Gilbert <dgilbert@redhat.com>

Message-id: <20200127190227.40942-9-dgilbert@redhat.com>

Patchwork-id: 93462

O-Subject: [RHEL-AV-8.2 qemu-kvm PATCH 008/112] virtiofsd: Add passthrough_ll

Bugzilla: 1694164

RH-Acked-by: Philippe Mathieu-Daudé <philmd@redhat.com>

RH-Acked-by: Stefan Hajnoczi <stefanha@redhat.com>

RH-Acked-by: Sergio Lopez Pascual <slp@redhat.com>

From: "Dr. David Alan Gilbert" <dgilbert@redhat.com>

passthrough_ll is one of the examples in the upstream fuse project

and is the main part of our daemon here.  It passes through requests

from fuse to the underlying filesystem, using syscalls as directly

as possible.

>From libfuse fuse-3.8.0

Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>

  Fixed up 'GPL' to 'GPLv2' as per Dan's comments and consistent

  with the 'LICENSE' file in libfuse;  patch sent to libfuse to fix

  it upstream.

Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>

Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>

(cherry picked from commit 7c6b66027241f41720240fc6ee1021cdbd975b2e)

Signed-off-by: Miroslav Rezanina <mrezanin@redhat.com>

---

 tools/virtiofsd/passthrough_ll.c | 1338 ++++++++++++++++++++++++++++++++++++++

 1 file changed, 1338 insertions(+)

 create mode 100644 tools/virtiofsd/passthrough_ll.c

diff --git a/tools/virtiofsd/passthrough_ll.c b/tools/virtiofsd/passthrough_ll.c

new file mode 100644

index 0000000..e1a6056

--- /dev/null

+++ b/tools/virtiofsd/passthrough_ll.c

@@ -0,0 +1,1338 @@

+/*

+  FUSE: Filesystem in Userspace

+  Copyright (C) 2001-2007  Miklos Szeredi <miklos@szeredi.hu>

+

+  This program can be distributed under the terms of the GNU GPLv2.

+  See the file COPYING.

+*/

+

+/** @file

+ *

+ * This file system mirrors the existing file system hierarchy of the

+ * system, starting at the root file system. This is implemented by

+ * just "passing through" all requests to the corresponding user-space

+ * libc functions. In contrast to passthrough.c and passthrough_fh.c,

+ * this implementation uses the low-level API. Its performance should

+ * be the least bad among the three, but many operations are not

+ * implemented. In particular, it is not possible to remove files (or

+ * directories) because the code necessary to defer actual removal

+ * until the file is not opened anymore would make the example much

+ * more complicated.

+ *

+ * When writeback caching is enabled (-o writeback mount option), it

+ * is only possible to write to files for which the mounting user has

+ * read permissions. This is because the writeback cache requires the

+ * kernel to be able to issue read requests for all files (which the

+ * passthrough filesystem cannot satisfy if it can't read the file in

+ * the underlying filesystem).

+ *

+ * Compile with:

+ *

+ *     gcc -Wall passthrough_ll.c `pkg-config fuse3 --cflags --libs` -o passthrough_ll

+ *

+ * ## Source code ##

+ * \include passthrough_ll.c

+ */

+

+#define _GNU_SOURCE

+#define FUSE_USE_VERSION 31

+

+#include "config.h"

+

+#include <fuse_lowlevel.h>

+#include <unistd.h>

+#include <stdlib.h>

+#include <stdio.h>

+#include <stddef.h>

+#include <stdbool.h>

+#include <string.h>

+#include <limits.h>

+#include <dirent.h>

+#include <assert.h>

+#include <errno.h>

+#include <inttypes.h>

+#include <pthread.h>

+#include <sys/file.h>

+#include <sys/xattr.h>

+

+#include "passthrough_helpers.h"

+

+/* We are re-using pointers to our `struct lo_inode` and `struct

+   lo_dirp` elements as inodes. This means that we must be able to

+   store uintptr_t values in a fuse_ino_t variable. The following

+   incantation checks this condition at compile time. */

+#if defined(__GNUC__) && (__GNUC__ > 4 || __GNUC__ == 4 && __GNUC_MINOR__ >= 6) && !defined __cplusplus

+_Static_assert(sizeof(fuse_ino_t) >= sizeof(uintptr_t),

+	       "fuse_ino_t too small to hold uintptr_t values!");

+#else

+struct _uintptr_to_must_hold_fuse_ino_t_dummy_struct \

+	{ unsigned _uintptr_to_must_hold_fuse_ino_t:

+			((sizeof(fuse_ino_t) >= sizeof(uintptr_t)) ? 1 : -1); };

+#endif

+

+struct lo_inode {

+	struct lo_inode *next; /* protected by lo->mutex */

+	struct lo_inode *prev; /* protected by lo->mutex */

+	int fd;

+	bool is_symlink;

+	ino_t ino;

+	dev_t dev;

+	uint64_t refcount; /* protected by lo->mutex */

+};

+

+enum {

+	CACHE_NEVER,

+	CACHE_NORMAL,

+	CACHE_ALWAYS,

+};

+

+struct lo_data {

+	pthread_mutex_t mutex;

+	int debug;

+	int writeback;

+	int flock;

+	int xattr;

+	const char *source;

+	double timeout;

+	int cache;

+	int timeout_set;

+	struct lo_inode root; /* protected by lo->mutex */

+};

+

+static const struct fuse_opt lo_opts[] = {

+	{ "writeback",

+	  offsetof(struct lo_data, writeback), 1 },

+	{ "no_writeback",

+	  offsetof(struct lo_data, writeback), 0 },

+	{ "source=%s",

+	  offsetof(struct lo_data, source), 0 },

+	{ "flock",

+	  offsetof(struct lo_data, flock), 1 },

+	{ "no_flock",

+	  offsetof(struct lo_data, flock), 0 },

+	{ "xattr",

+	  offsetof(struct lo_data, xattr), 1 },

+	{ "no_xattr",

+	  offsetof(struct lo_data, xattr), 0 },

+	{ "timeout=%lf",

+	  offsetof(struct lo_data, timeout), 0 },

+	{ "timeout=",

+	  offsetof(struct lo_data, timeout_set), 1 },

+	{ "cache=never",

+	  offsetof(struct lo_data, cache), CACHE_NEVER },

+	{ "cache=auto",

+	  offsetof(struct lo_data, cache), CACHE_NORMAL },

+	{ "cache=always",

+	  offsetof(struct lo_data, cache), CACHE_ALWAYS },

+

+	FUSE_OPT_END

+};

+

+static struct lo_data *lo_data(fuse_req_t req)

+{

+	return (struct lo_data *) fuse_req_userdata(req);

+}

+

+static struct lo_inode *lo_inode(fuse_req_t req, fuse_ino_t ino)

+{

+	if (ino == FUSE_ROOT_ID)

+		return &lo_data(req)->root;

+	else

+		return (struct lo_inode *) (uintptr_t) ino;

+}

+

+static int lo_fd(fuse_req_t req, fuse_ino_t ino)

+{

+	return lo_inode(req, ino)->fd;

+}

+

+static bool lo_debug(fuse_req_t req)

+{

+	return lo_data(req)->debug != 0;

+}

+

+static void lo_init(void *userdata,

+		    struct fuse_conn_info *conn)

+{

+	struct lo_data *lo = (struct lo_data*) userdata;

+

+	if(conn->capable & FUSE_CAP_EXPORT_SUPPORT)

+		conn->want |= FUSE_CAP_EXPORT_SUPPORT;

+

+	if (lo->writeback &&

+	    conn->capable & FUSE_CAP_WRITEBACK_CACHE) {

+		if (lo->debug)

+			fuse_log(FUSE_LOG_DEBUG, "lo_init: activating writeback\n");

+		conn->want |= FUSE_CAP_WRITEBACK_CACHE;

+	}

+	if (lo->flock && conn->capable & FUSE_CAP_FLOCK_LOCKS) {

+		if (lo->debug)

+			fuse_log(FUSE_LOG_DEBUG, "lo_init: activating flock locks\n");

+		conn->want |= FUSE_CAP_FLOCK_LOCKS;

+	}

+}

+

+static void lo_getattr(fuse_req_t req, fuse_ino_t ino,

+			     struct fuse_file_info *fi)

+{

+	int res;

+	struct stat buf;

+	struct lo_data *lo = lo_data(req);

+

+	(void) fi;

+

+	res = fstatat(lo_fd(req, ino), "", &buf, AT_EMPTY_PATH | AT_SYMLINK_NOFOLLOW);

+	if (res == -1)

+		return (void) fuse_reply_err(req, errno);

+

+	fuse_reply_attr(req, &buf, lo->timeout);

+}

+

+static int utimensat_empty_nofollow(struct lo_inode *inode,

+				    const struct timespec *tv)

+{

+	int res;

+	char procname[64];

+

+	if (inode->is_symlink) {

+		res = utimensat(inode->fd, "", tv,

+				AT_EMPTY_PATH | AT_SYMLINK_NOFOLLOW);

+		if (res == -1 && errno == EINVAL) {

+			/* Sorry, no race free way to set times on symlink. */

+			errno = EPERM;

+		}

+		return res;

+	}

+	sprintf(procname, "/proc/self/fd/%i", inode->fd);

+

+	return utimensat(AT_FDCWD, procname, tv, 0);

+}

+

+static void lo_setattr(fuse_req_t req, fuse_ino_t ino, struct stat *attr,

+		       int valid, struct fuse_file_info *fi)

+{

+	int saverr;

+	char procname[64];

+	struct lo_inode *inode = lo_inode(req, ino);

+	int ifd = inode->fd;

+	int res;

+

+	if (valid & FUSE_SET_ATTR_MODE) {

+		if (fi) {

+			res = fchmod(fi->fh, attr->st_mode);

+		} else {

+			sprintf(procname, "/proc/self/fd/%i", ifd);

+			res = chmod(procname, attr->st_mode);

+		}

+		if (res == -1)

+			goto out_err;

+	}

+	if (valid & (FUSE_SET_ATTR_UID | FUSE_SET_ATTR_GID)) {

+		uid_t uid = (valid & FUSE_SET_ATTR_UID) ?

+			attr->st_uid : (uid_t) -1;

+		gid_t gid = (valid & FUSE_SET_ATTR_GID) ?

+			attr->st_gid : (gid_t) -1;

+

+		res = fchownat(ifd, "", uid, gid,

+			       AT_EMPTY_PATH | AT_SYMLINK_NOFOLLOW);

+		if (res == -1)

+			goto out_err;

+	}

+	if (valid & FUSE_SET_ATTR_SIZE) {

+		if (fi) {

+			res = ftruncate(fi->fh, attr->st_size);

+		} else {

+			sprintf(procname, "/proc/self/fd/%i", ifd);

+			res = truncate(procname, attr->st_size);

+		}

+		if (res == -1)

+			goto out_err;

+	}

+	if (valid & (FUSE_SET_ATTR_ATIME | FUSE_SET_ATTR_MTIME)) {

+		struct timespec tv[2];

+

+		tv[0].tv_sec = 0;

+		tv[1].tv_sec = 0;

+		tv[0].tv_nsec = UTIME_OMIT;

+		tv[1].tv_nsec = UTIME_OMIT;

+

+		if (valid & FUSE_SET_ATTR_ATIME_NOW)

+			tv[0].tv_nsec = UTIME_NOW;

+		else if (valid & FUSE_SET_ATTR_ATIME)

+			tv[0] = attr->st_atim;

+

+		if (valid & FUSE_SET_ATTR_MTIME_NOW)

+			tv[1].tv_nsec = UTIME_NOW;

+		else if (valid & FUSE_SET_ATTR_MTIME)

+			tv[1] = attr->st_mtim;

+

+		if (fi)

+			res = futimens(fi->fh, tv);

+		else

+			res = utimensat_empty_nofollow(inode, tv);

+		if (res == -1)

+			goto out_err;

+	}

+

+	return lo_getattr(req, ino, fi);

+

+out_err:

+	saverr = errno;

+	fuse_reply_err(req, saverr);

+}

+

+static struct lo_inode *lo_find(struct lo_data *lo, struct stat *st)

+{

+	struct lo_inode *p;

+	struct lo_inode *ret = NULL;

+

+	pthread_mutex_lock(&lo->mutex);

+	for (p = lo->root.next; p != &lo->root; p = p->next) {

+		if (p->ino == st->st_ino && p->dev == st->st_dev) {

+			assert(p->refcount > 0);

+			ret = p;

+			ret->refcount++;

+			break;

+		}

+	}

+	pthread_mutex_unlock(&lo->mutex);

+	return ret;

+}

+

+static int lo_do_lookup(fuse_req_t req, fuse_ino_t parent, const char *name,

+			 struct fuse_entry_param *e)

+{

+	int newfd;

+	int res;

+	int saverr;

+	struct lo_data *lo = lo_data(req);

+	struct lo_inode *inode;

+

+	memset(e, 0, sizeof(*e));

+	e->attr_timeout = lo->timeout;

+	e->entry_timeout = lo->timeout;

+

+	newfd = openat(lo_fd(req, parent), name, O_PATH | O_NOFOLLOW);

+	if (newfd == -1)

+		goto out_err;

+

+	res = fstatat(newfd, "", &e->attr, AT_EMPTY_PATH | AT_SYMLINK_NOFOLLOW);

+	if (res == -1)

+		goto out_err;

+

+	inode = lo_find(lo_data(req), &e->attr);

+	if (inode) {

+		close(newfd);

+		newfd = -1;

+	} else {

+		struct lo_inode *prev, *next;

+

+		saverr = ENOMEM;

+		inode = calloc(1, sizeof(struct lo_inode));

+		if (!inode)

+			goto out_err;

+

+		inode->is_symlink = S_ISLNK(e->attr.st_mode);

+		inode->refcount = 1;

+		inode->fd = newfd;

+		inode->ino = e->attr.st_ino;

+		inode->dev = e->attr.st_dev;

+

+		pthread_mutex_lock(&lo->mutex);

+		prev = &lo->root;

+		next = prev->next;

+		next->prev = inode;

+		inode->next = next;

+		inode->prev = prev;

+		prev->next = inode;

+		pthread_mutex_unlock(&lo->mutex);

+	}

+	e->ino = (uintptr_t) inode;

+

+	if (lo_debug(req))

+		fuse_log(FUSE_LOG_DEBUG, "  %lli/%s -> %lli\n",

+			(unsigned long long) parent, name, (unsigned long long) e->ino);

+

+	return 0;

+

+out_err:

+	saverr = errno;

+	if (newfd != -1)

+		close(newfd);

+	return saverr;

+}

+

+static void lo_lookup(fuse_req_t req, fuse_ino_t parent, const char *name)

+{

+	struct fuse_entry_param e;

+	int err;

+

+	if (lo_debug(req))

+		fuse_log(FUSE_LOG_DEBUG, "lo_lookup(parent=%" PRIu64 ", name=%s)\n",

+			parent, name);

+

+	err = lo_do_lookup(req, parent, name, &e);

+	if (err)

+		fuse_reply_err(req, err);

+	else

+		fuse_reply_entry(req, &e);

+}

+

+static void lo_mknod_symlink(fuse_req_t req, fuse_ino_t parent,

+			     const char *name, mode_t mode, dev_t rdev,

+			     const char *link)

+{

+	int res;

+	int saverr;

+	struct lo_inode *dir = lo_inode(req, parent);

+	struct fuse_entry_param e;

+

+	saverr = ENOMEM;

+

+	res = mknod_wrapper(dir->fd, name, link, mode, rdev);

+

+	saverr = errno;

+	if (res == -1)

+		goto out;

+

+	saverr = lo_do_lookup(req, parent, name, &e);

+	if (saverr)

+		goto out;

+

+	if (lo_debug(req))

+		fuse_log(FUSE_LOG_DEBUG, "  %lli/%s -> %lli\n",

+			(unsigned long long) parent, name, (unsigned long long) e.ino);

+

+	fuse_reply_entry(req, &e);

+	return;

+

+out:

+	fuse_reply_err(req, saverr);

+}

+

+static void lo_mknod(fuse_req_t req, fuse_ino_t parent,

+		     const char *name, mode_t mode, dev_t rdev)

+{

+	lo_mknod_symlink(req, parent, name, mode, rdev, NULL);

+}

+

+static void lo_mkdir(fuse_req_t req, fuse_ino_t parent, const char *name,

+		     mode_t mode)

+{

+	lo_mknod_symlink(req, parent, name, S_IFDIR | mode, 0, NULL);

+}

+

+static void lo_symlink(fuse_req_t req, const char *link,

+		       fuse_ino_t parent, const char *name)

+{

+	lo_mknod_symlink(req, parent, name, S_IFLNK, 0, link);

+}

+

+static int linkat_empty_nofollow(struct lo_inode *inode, int dfd,

+				 const char *name)

+{

+	int res;

+	char procname[64];

+

+	if (inode->is_symlink) {

+		res = linkat(inode->fd, "", dfd, name, AT_EMPTY_PATH);

+		if (res == -1 && (errno == ENOENT || errno == EINVAL)) {

+			/* Sorry, no race free way to hard-link a symlink. */

+			errno = EPERM;

+		}

+		return res;

+	}

+

+	sprintf(procname, "/proc/self/fd/%i", inode->fd);

+

+	return linkat(AT_FDCWD, procname, dfd, name, AT_SYMLINK_FOLLOW);

+}

+

+static void lo_link(fuse_req_t req, fuse_ino_t ino, fuse_ino_t parent,

+		    const char *name)

+{

+	int res;

+	struct lo_data *lo = lo_data(req);

+	struct lo_inode *inode = lo_inode(req, ino);

+	struct fuse_entry_param e;

+	int saverr;

+

+	memset(&e, 0, sizeof(struct fuse_entry_param));

+	e.attr_timeout = lo->timeout;

+	e.entry_timeout = lo->timeout;

+

+	res = linkat_empty_nofollow(inode, lo_fd(req, parent), name);

+	if (res == -1)

+		goto out_err;

+

+	res = fstatat(inode->fd, "", &e.attr, AT_EMPTY_PATH | AT_SYMLINK_NOFOLLOW);

+	if (res == -1)

+		goto out_err;

+

+	pthread_mutex_lock(&lo->mutex);

+	inode->refcount++;

+	pthread_mutex_unlock(&lo->mutex);

+	e.ino = (uintptr_t) inode;

+

+	if (lo_debug(req))

+		fuse_log(FUSE_LOG_DEBUG, "  %lli/%s -> %lli\n",

+			(unsigned long long) parent, name,

+			(unsigned long long) e.ino);

+

+	fuse_reply_entry(req, &e);

+	return;

+

+out_err:

+	saverr = errno;

+	fuse_reply_err(req, saverr);

+}

+

+static void lo_rmdir(fuse_req_t req, fuse_ino_t parent, const char *name)

+{

+	int res;

+

+	res = unlinkat(lo_fd(req, parent), name, AT_REMOVEDIR);

+

+	fuse_reply_err(req, res == -1 ? errno : 0);

+}

+

+static void lo_rename(fuse_req_t req, fuse_ino_t parent, const char *name,

+		      fuse_ino_t newparent, const char *newname,

+		      unsigned int flags)

+{

+	int res;

+

+	if (flags) {

+		fuse_reply_err(req, EINVAL);

+		return;

+	}

+

+	res = renameat(lo_fd(req, parent), name,

+			lo_fd(req, newparent), newname);

+

+	fuse_reply_err(req, res == -1 ? errno : 0);

+}

+

+static void lo_unlink(fuse_req_t req, fuse_ino_t parent, const char *name)

+{

+	int res;

+

+	res = unlinkat(lo_fd(req, parent), name, 0);

+

+	fuse_reply_err(req, res == -1 ? errno : 0);

+}

+

+static void unref_inode(struct lo_data *lo, struct lo_inode *inode, uint64_t n)

+{

+	if (!inode)

+		return;

+

+	pthread_mutex_lock(&lo->mutex);

+	assert(inode->refcount >= n);

+	inode->refcount -= n;

+	if (!inode->refcount) {

+		struct lo_inode *prev, *next;

+

+		prev = inode->prev;

+		next = inode->next;

+		next->prev = prev;

+		prev->next = next;

+

+		pthread_mutex_unlock(&lo->mutex);

+		close(inode->fd);

+		free(inode);

+

+	} else {

+		pthread_mutex_unlock(&lo->mutex);

+	}

+}

+

+static void lo_forget_one(fuse_req_t req, fuse_ino_t ino, uint64_t nlookup)

+{

+	struct lo_data *lo = lo_data(req);

+	struct lo_inode *inode = lo_inode(req, ino);

+

+	if (lo_debug(req)) {

+		fuse_log(FUSE_LOG_DEBUG, "  forget %lli %lli -%lli\n",

+			(unsigned long long) ino,

+			(unsigned long long) inode->refcount,

+			(unsigned long long) nlookup);

+	}

+

+	unref_inode(lo, inode, nlookup);

+}

+

+static void lo_forget(fuse_req_t req, fuse_ino_t ino, uint64_t nlookup)

+{

+	lo_forget_one(req, ino, nlookup);

+	fuse_reply_none(req);

+}

+

+static void lo_forget_multi(fuse_req_t req, size_t count,

+				struct fuse_forget_data *forgets)

+{

+	int i;

+

+	for (i = 0; i < count; i++)

+		lo_forget_one(req, forgets[i].ino, forgets[i].nlookup);

+	fuse_reply_none(req);

+}

+

+static void lo_readlink(fuse_req_t req, fuse_ino_t ino)

+{

+	char buf[PATH_MAX + 1];

+	int res;

+

+	res = readlinkat(lo_fd(req, ino), "", buf, sizeof(buf));

+	if (res == -1)

+		return (void) fuse_reply_err(req, errno);

+

+	if (res == sizeof(buf))

+		return (void) fuse_reply_err(req, ENAMETOOLONG);

+

+	buf[res] = '\0';

+

+	fuse_reply_readlink(req, buf);

+}

+

+struct lo_dirp {

+	DIR *dp;

+	struct dirent *entry;

+	off_t offset;

+};

+

+static struct lo_dirp *lo_dirp(struct fuse_file_info *fi)

+{

+	return (struct lo_dirp *) (uintptr_t) fi->fh;

+}

+

+static void lo_opendir(fuse_req_t req, fuse_ino_t ino, struct fuse_file_info *fi)

+{

+	int error = ENOMEM;

+	struct lo_data *lo = lo_data(req);

+	struct lo_dirp *d;

+	int fd;

+

+	d = calloc(1, sizeof(struct lo_dirp));

+	if (d == NULL)

+		goto out_err;

+

+	fd = openat(lo_fd(req, ino), ".", O_RDONLY);

+	if (fd == -1)

+		goto out_errno;

+

+	d->dp = fdopendir(fd);

+	if (d->dp == NULL)

+		goto out_errno;

+

+	d->offset = 0;

+	d->entry = NULL;

+

+	fi->fh = (uintptr_t) d;

+	if (lo->cache == CACHE_ALWAYS)

+		fi->keep_cache = 1;

+	fuse_reply_open(req, fi);

+	return;

+

+out_errno:

+	error = errno;

+out_err:

+	if (d) {

+		if (fd != -1)

+			close(fd);

+		free(d);

+	}

+	fuse_reply_err(req, error);

+}

+

+static int is_dot_or_dotdot(const char *name)

+{

+	return name[0] == '.' && (name[1] == '\0' ||

+				  (name[1] == '.' && name[2] == '\0'));

+}

+

+static void lo_do_readdir(fuse_req_t req, fuse_ino_t ino, size_t size,

+			  off_t offset, struct fuse_file_info *fi, int plus)

+{

+	struct lo_dirp *d = lo_dirp(fi);

+	char *buf;

+	char *p;

+	size_t rem = size;

+	int err;

+

+	(void) ino;

+

+	buf = calloc(1, size);

+	if (!buf) {

+		err = ENOMEM;

+		goto error;

+	}

+	p = buf;

+

+	if (offset != d->offset) {

+		seekdir(d->dp, offset);