samba-mirror/source3/modules/vfs_aio_pthread.c

/*
 * Simulate Posix AIO using pthreads.
 *
 * Based on the aio_fork work from Volker and Volker's pthreadpool library.
 *
 * Copyright (C) Volker Lendecke 2008
 * Copyright (C) Jeremy Allison 2012
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 3 of the License, or
 * (at your option) any later version.
 *
 * This program 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 General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
 */

#include "includes.h"
#include "system/filesys.h"
#include "system/shmem.h"
#include "smbd/smbd.h"
#include "smbd/globals.h"
#include "../lib/pthreadpool/pthreadpool_tevent.h"
#ifdef HAVE_LINUX_FALLOC_H
#include <linux/falloc.h>
#endif

#if defined(HAVE_OPENAT) && defined(HAVE_LINUX_THREAD_CREDENTIALS)

/*
 * We must have openat() to do any thread-based
 * asynchronous opens. We also must be using
 * thread-specific credentials (Linux-only
 * for now).
 */

struct aio_open_private_data {
	struct aio_open_private_data *prev, *next;
	/* Inputs. */
	int dir_fd;
	bool opened_dir_fd;
	int flags;
	mode_t mode;
	uint64_t mid;
	bool in_progress;
	struct smb_filename *fsp_name;
	struct smb_filename *smb_fname;
	connection_struct *conn;
	struct smbXsrv_connection *xconn;
	const struct security_unix_token *ux_tok;
	uint64_t initial_allocation_size;
	/* Returns. */
	int ret_fd;
	int ret_errno;
};

/* List of outstanding requests we have. */
static struct aio_open_private_data *open_pd_list;

static void aio_open_do(struct aio_open_private_data *opd);
static void opd_free(struct aio_open_private_data *opd);

/************************************************************************
 Find the open private data by mid.
***********************************************************************/

static struct aio_open_private_data *find_open_private_data_by_mid(uint64_t mid)
{
	struct aio_open_private_data *opd;

	for (opd = open_pd_list; opd != NULL; opd = opd->next) {
		if (opd->mid == mid) {
			return opd;
		}
	}

	return NULL;
}

/************************************************************************
 Callback when an open completes.
***********************************************************************/

static void aio_open_handle_completion(struct tevent_req *subreq)
{
	struct aio_open_private_data *opd =
		tevent_req_callback_data(subreq,
		struct aio_open_private_data);
	int ret;

	ret = pthreadpool_tevent_job_recv(subreq);
	TALLOC_FREE(subreq);

	/*
	 * We're no longer in flight. Remove the
	 * destructor used to preserve opd so
	 * a talloc_free actually removes it.
	 */
	talloc_set_destructor(opd, NULL);

	if (opd->conn == NULL) {
		/*
		 * We were shutdown closed in flight. No one
		 * wants the result, and state has been reparented
		 * to the NULL context, so just free it so we
		 * don't leak memory.
		 */
		DBG_NOTICE("aio open request for %s abandoned in flight\n",
			opd->fsp_name->base_name);
		if (opd->ret_fd != -1) {
			close(opd->ret_fd);
			opd->ret_fd = -1;
		}
		/*
		 * Find outstanding event and reschedule so the client
		 * gets an error message return from the open.
		 */
		schedule_deferred_open_message_smb(opd->xconn, opd->mid);
		opd_free(opd);
		return;
	}

	if (ret != 0) {
		bool ok;

		if (ret != EAGAIN) {
			smb_panic("aio_open_handle_completion");
			/* notreached. */
			return;
		}
		/*
		 * Make sure we run as the user again
		 */
		ok = change_to_user_and_service(opd->conn, opd->conn->vuid);
		if (!ok) {
			smb_panic("Can't change to user");
			return;
		}
		/*
		 * If we get EAGAIN from pthreadpool_tevent_job_recv() this
		 * means the lower level pthreadpool failed to create a new
		 * thread. Fallback to sync processing in that case to allow
		 * some progress for the client.
		 */
		aio_open_do(opd);
	}

	DEBUG(10,("aio_open_handle_completion: mid %llu "
		"for file %s completed\n",
		(unsigned long long)opd->mid,
		opd->fsp_name->base_name));

	opd->in_progress = false;

	/* Find outstanding event and reschedule. */
	if (!schedule_deferred_open_message_smb(opd->xconn, opd->mid)) {
		/*
		 * Outstanding event didn't exist or was
		 * cancelled. Free up the fd and throw
		 * away the result.
		 */
		if (opd->ret_fd != -1) {
			close(opd->ret_fd);
			opd->ret_fd = -1;
		}
		opd_free(opd);
	}
}

/*****************************************************************
 The core of the async open code - the worker function. Note we
 use the new openat() system call to avoid any problems with
 current working directory changes plus we change credentials
 on the thread to prevent any security race conditions.
*****************************************************************/

static void aio_open_worker(void *private_data)
{
	struct aio_open_private_data *opd =
		(struct aio_open_private_data *)private_data;

	/* Become the correct credential on this thread. */
	if (set_thread_credentials(opd->ux_tok->uid,
				opd->ux_tok->gid,
				(size_t)opd->ux_tok->ngroups,
				opd->ux_tok->groups) != 0) {
		opd->ret_fd = -1;
		opd->ret_errno = errno;
		return;
	}

	aio_open_do(opd);
}

static void aio_open_do(struct aio_open_private_data *opd)
{
	opd->ret_fd = openat(opd->dir_fd,
			opd->smb_fname->base_name,
			opd->flags,
			opd->mode);

	if (opd->ret_fd == -1) {
		opd->ret_errno = errno;
	} else {
		/* Create was successful. */
		opd->ret_errno = 0;

#if defined(HAVE_LINUX_FALLOCATE)
		/*
		 * See if we can set the initial
		 * allocation size. We don't record
		 * the return for this as it's an
		 * optimization - the upper layer
		 * will also do this for us once
		 * the open returns.
		 */
		if (opd->initial_allocation_size) {
			(void)fallocate(opd->ret_fd,
					FALLOC_FL_KEEP_SIZE,
					0,
					(off_t)opd->initial_allocation_size);
		}
#endif
	}
}

/************************************************************************
 Open private data teardown.
***********************************************************************/

static void opd_free(struct aio_open_private_data *opd)
{
	if (opd->opened_dir_fd && opd->dir_fd != -1) {
		close(opd->dir_fd);
	}
	DLIST_REMOVE(open_pd_list, opd);
	TALLOC_FREE(opd);
}

/************************************************************************
 Create and initialize a private data struct for async open.
***********************************************************************/

static struct aio_open_private_data *create_private_open_data(
	TALLOC_CTX *ctx,
	const struct files_struct *dirfsp,
	const struct smb_filename *smb_fname,
	const files_struct *fsp,
	int flags,
	mode_t mode)
{
	struct aio_open_private_data *opd = talloc_zero(ctx,
					struct aio_open_private_data);

	if (!opd) {
		return NULL;
	}

	*opd = (struct aio_open_private_data) {
		.dir_fd = -1,
		.ret_fd = -1,
		.ret_errno = EINPROGRESS,
		.flags = flags,
		.mode = mode,
		.mid = fsp->mid,
		.in_progress = true,
		.conn = fsp->conn,
		/*
		 * TODO: In future we need a proper algorithm
		 * to find the correct connection for a fsp.
		 * For now we only have one connection, so this is correct...
		 */
		.xconn = fsp->conn->sconn->client->connections,
		.initial_allocation_size = fsp->initial_allocation_size,
	};

	/* Copy our current credentials. */
	opd->ux_tok = copy_unix_token(opd, get_current_utok(fsp->conn));
	if (opd->ux_tok == NULL) {
		opd_free(opd);
		return NULL;
	}

	/*
	 * Copy the full fsp_name and smb_fname which is the basename.
	 */
	opd->smb_fname = cp_smb_filename(opd, smb_fname);
	if (opd->smb_fname == NULL) {
		opd_free(opd);
		return NULL;
	}

	opd->fsp_name = cp_smb_filename(opd, fsp->fsp_name);
	if (opd->fsp_name == NULL) {
		opd_free(opd);
		return NULL;
	}

	if (fsp_get_pathref_fd(dirfsp) != AT_FDCWD) {
		opd->dir_fd = fsp_get_pathref_fd(dirfsp);
	} else {
#if defined(O_DIRECTORY)
		opd->dir_fd = open(".", O_RDONLY|O_DIRECTORY);
#else
		opd->dir_fd = open(".", O_RDONLY);
#endif
		opd->opened_dir_fd = true;
	}
	if (opd->dir_fd == -1) {
		opd_free(opd);
		return NULL;
	}

	DLIST_ADD_END(open_pd_list, opd);
	return opd;
}

static int opd_inflight_destructor(struct aio_open_private_data *opd)
{
	/*
	 * Setting conn to NULL allows us to
	 * discover the connection was torn
	 * down which kills the fsp that owns
	 * opd.
	 */
	DBG_NOTICE("aio open request for %s cancelled\n",
		opd->fsp_name->base_name);
	opd->conn = NULL;
	/* Don't let opd go away. */
	return -1;
}

/*****************************************************************
 Setup an async open.
*****************************************************************/

static int open_async(const struct files_struct *dirfsp,
		      const struct smb_filename *smb_fname,
		      const files_struct *fsp,
		      int flags,
		      mode_t mode)
{
	struct aio_open_private_data *opd = NULL;
	struct tevent_req *subreq = NULL;

	/*
	 * Allocate off fsp->conn, not NULL or fsp. As we're going
	 * async fsp will get talloc_free'd when we return
	 * EINPROGRESS/NT_STATUS_MORE_PROCESSING_REQUIRED. A new fsp
	 * pointer gets allocated on every re-run of the
	 * open code path. Allocating on fsp->conn instead
	 * of NULL allows use to get notified via destructor
	 * if the conn is force-closed or we shutdown.
	 * opd is always safely freed in all codepath so no
	 * memory leaks.
	 */
	opd = create_private_open_data(fsp->conn,
				       dirfsp,
				       smb_fname,
				       fsp,
				       flags,
				       mode);
	if (opd == NULL) {
		DEBUG(10, ("open_async: Could not create private data.\n"));
		return -1;
	}

	subreq = pthreadpool_tevent_job_send(opd,
					     fsp->conn->sconn->ev_ctx,
					     fsp->conn->sconn->pool,
					     aio_open_worker, opd);
	if (subreq == NULL) {
		opd_free(opd);
		return -1;
	}
	tevent_req_set_callback(subreq, aio_open_handle_completion, opd);

	DEBUG(5,("open_async: mid %llu created for file %s\n",
		(unsigned long long)opd->mid,
		opd->fsp_name->base_name));

	/*
	 * Add a destructor to protect us from connection
	 * teardown whilst the open thread is in flight.
	 */
	talloc_set_destructor(opd, opd_inflight_destructor);

	/* Cause the calling code to reschedule us. */
	errno = EINPROGRESS; /* Maps to NT_STATUS_MORE_PROCESSING_REQUIRED. */
	return -1;
}

/*****************************************************************
 Look for a matching SMB2 mid. If we find it we're rescheduled,
 just return the completed open.
*****************************************************************/

static bool find_completed_open(files_struct *fsp,
				int *p_fd,
				int *p_errno)
{
	struct aio_open_private_data *opd;

	opd = find_open_private_data_by_mid(fsp->mid);
	if (!opd) {
		return false;
	}

	if (opd->in_progress) {
		DEBUG(0,("find_completed_open: mid %llu "
			"still in progress for "
			"file %s. PANIC !\n",
			(unsigned long long)opd->mid,
			opd->fsp_name->base_name));
		/* Disaster ! This is an open timeout. Just panic. */
		smb_panic("find_completed_open - in_progress\n");
		/* notreached. */
		return false;
	}

	*p_fd = opd->ret_fd;
	*p_errno = opd->ret_errno;

	DEBUG(5,("find_completed_open: mid %llu returning "
		"fd = %d, errno = %d (%s) "
		"for file %s\n",
		(unsigned long long)opd->mid,
		opd->ret_fd,
		opd->ret_errno,
		strerror(opd->ret_errno),
		smb_fname_str_dbg(fsp->fsp_name)));

	/* Now we can free the opd. */
	opd_free(opd);
	return true;
}

/*****************************************************************
 The core open function. Only go async on O_CREAT|O_EXCL
 opens to prevent any race conditions.
*****************************************************************/

static int aio_pthread_openat_fn(vfs_handle_struct *handle,
				 const struct files_struct *dirfsp,
				 const struct smb_filename *smb_fname,
				 struct files_struct *fsp,
				 const struct vfs_open_how *how)
{
	int my_errno = 0;
	int fd = -1;
	bool aio_allow_open = lp_parm_bool(
		SNUM(handle->conn), "aio_pthread", "aio open", false);

	if (how->resolve != 0) {
		errno = ENOSYS;
		return -1;
	}

	if (is_named_stream(smb_fname)) {
		/* Don't handle stream opens. */
		errno = ENOENT;
		return -1;
	}

	if (fsp->conn->sconn->pool == NULL) {
		/*
		 * a threadpool is required for async support
		 */
		aio_allow_open = false;
	}

	if (fsp->conn->sconn->client != NULL &&
	    fsp->conn->sconn->client->server_multi_channel_enabled) {
		/*
		 * This module is not compatible with multi channel yet.
		 */
		aio_allow_open = false;
	}

	if (fsp->fsp_flags.is_pathref) {
		/* Use SMB_VFS_NEXT_OPENAT() to call openat() with O_PATH. */
		aio_allow_open = false;
	}

	if (!(how->flags & O_CREAT)) {
		/* Only creates matter. */
		aio_allow_open = false;
	}

	if (!(how->flags & O_EXCL)) {
		/* Only creates with O_EXCL matter. */
		aio_allow_open = false;
	}

	if (!aio_allow_open) {
		/* aio opens turned off. */
		return SMB_VFS_NEXT_OPENAT(handle,
			      dirfsp,
			      smb_fname,
			      fsp,
			      how);
	}

	/*
	 * See if this is a reentrant call - i.e. is this a
	 * restart of an existing open that just completed.
	 */

	if (find_completed_open(fsp,
				&fd,
				&my_errno)) {
		errno = my_errno;
		return fd;
	}

	/* Ok, it's a create exclusive call - pass it to a thread helper. */
	return open_async(dirfsp, smb_fname, fsp, how->flags, how->mode);
}
#endif

static struct vfs_fn_pointers vfs_aio_pthread_fns = {
#if defined(HAVE_OPENAT) && defined(HAVE_LINUX_THREAD_CREDENTIALS)
	.openat_fn = aio_pthread_openat_fn,
#endif
};

static_decl_vfs;
NTSTATUS vfs_aio_pthread_init(TALLOC_CTX *ctx)
{
	return smb_register_vfs(SMB_VFS_INTERFACE_VERSION,
				"aio_pthread", &vfs_aio_pthread_fns);
}