Add initial AIO/LIO kernel thread support files. This is preliminary, and

further features will be added.

2 files changed, 432 insertions, 0 deletions

diff --git a/sys/kern/kern_threads.c b/sys/kern/kern_threads.c
new file mode 100644
index 0000000..30e15c9
--- /dev/null
+++ b/sys/kern/kern_threads.c
@@ -0,0 +1,168 @@
+/*
+ *
+ * Portions of this code was derived from the file kern_fork.c and as such
+ * is subject to the copyrights below.
+ *
+ * Copyright (c) 1982, 1986, 1989, 1991, 1993
+ *	The Regents of the University of California.  All rights reserved.
+ * (c) UNIX System Laboratories, Inc.
+ * All or some portions of this file are derived from material licensed
+ * to the University of California by American Telephone and Telegraph
+ * Co. or Unix System Laboratories, Inc. and are reproduced herein with
+ * the permission of UNIX System Laboratories, Inc.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. 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.
+ * 3. All advertising materials mentioning features or use of this software
+ *    must display the following acknowledgement:
+ *	This product includes software developed by the University of
+ *	California, Berkeley and its contributors.
+ * 4. Neither the name of the University 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 REGENTS AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ *
+ * Copyright (c) 1996 Douglas Santry
+ *
+ * This code is subject to the beer copyright.  If I chance to meet you in a
+ * bar and this code helped you in some way, you owe me a beer.  Only
+ * in Germany will I accept domestic beer.  This code may or may not work
+ * and I certainly make no claims as to its fitness for *any* purpose.
+ * 
+ * $Id$
+ */
+
+#include <sys/param.h>
+#include <sys/systm.h>
+#include <sys/filedesc.h>
+#include <sys/kernel.h>
+#include <sys/malloc.h>
+#include <sys/proc.h>
+#include <sys/resourcevar.h>
+#include <sys/vnode.h>
+#include <sys/file.h>
+#include <sys/acct.h>
+#include <sys/ktrace.h>
+#include <sys/sysproto.h>
+#include <sys/unistd.h>
+
+#include <vm/vm.h>
+
+#include <machine/cpu.h>
+
+/*
+ * Low level support for sleep/wakeup paradigm
+ * If a timeout is specified:
+ *	returns 0 if wakeup
+ *	returns EAGAIN if timed out
+ *	returns EINVAL if error
+ *
+ * If a timeout is not specified:
+ *
+ *	returns time waiting in ticks.
+ */
+int
+thr_sleep(struct proc *p, struct thr_sleep_args *uap, int *retval) {
+	int sleepStart;
+	long long sleeptime;
+	int sleepclocks;
+	struct timespec ts;
+	int error;
+
+	sleepclocks = 0;
+	if (uap->timeout != 0) {
+		/*
+		 * Get timespec struct
+		 */
+		if (error = copyin((caddr_t) uap->timeout, (caddr_t) &ts, sizeof ts)) {
+			p->p_wakeup = 0;
+			return error;
+		}
+		sleeptime = (long long) (hz * ts.tv_nsec);
+		sleeptime /= 1000000000LL;
+		sleeptime += ts.tv_sec * hz;
+		sleepclocks = sleeptime;
+		if (sleepclocks != sleeptime) {
+			p->p_wakeup = 0;
+			retval[0] = EINVAL;
+			return 0;
+		}
+		if (sleepclocks == 0)
+			sleepclocks = 1;
+	}
+
+	retval[0] = 0;
+	if (p->p_wakeup == 0) {
+		sleepStart = ticks;
+		p->p_flag |= P_SINTR;
+		error = tsleep(p, PUSER, "thrslp", sleepclocks);
+		p->p_flag &= ~P_SINTR;
+		if (error == EWOULDBLOCK) {
+			p->p_wakeup = 0;
+			retval[0] = EAGAIN;
+			return 0;
+		}
+		if (uap->timeout == 0)
+			retval[0] = ticks - sleepStart;
+	}
+	p->p_wakeup = 0;
+	return (0);
+}
+
+int
+thr_wakeup(struct proc *p, struct thr_wakeup_args *uap, int *retval) {
+	struct proc *pSlave = p->p_leader;
+
+	while(pSlave && (pSlave->p_pid != uap->pid))
+		pSlave = pSlave->p_peers;
+
+	if(pSlave == 0) {
+		retval[0] = ESRCH;
+		return(0);
+	}
+
+	pSlave->p_wakeup++;
+	if((pSlave->p_stat == SSLEEP) && (pSlave->p_wchan == pSlave)) {
+		wakeup(pSlave);
+		return(0);
+	}
+
+	retval[0] = EAGAIN;
+	return 0;
+}
+
+/*
+ * General purpose yield system call
+ */
+int
+yield(struct proc *p, struct yield_args *uap, int *retval) {
+	int s;
+
+	retval[0] = 0;
+
+	s = splhigh();
+	p->p_priority = MAXPRI;
+	setrunqueue(p);
+	mi_switch();
+	splx(s);
+
+	return(0);
+}
+
diff --git a/sys/kern/vfs_aio.c b/sys/kern/vfs_aio.c
new file mode 100644
index 0000000..282eee4
--- /dev/null
+++ b/sys/kern/vfs_aio.c
@@ -0,0 +1,264 @@
+
+/*
+ * Copyright (c) 1997 John S. Dyson.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. John S. Dyson's name may not be used to endorse or promote products
+ *    derived from this software without specific prior written permission.
+ *
+ * DISCLAIMER:  This code isn't warranted to do anything useful.  Anything
+ * bad that happens because of using this software isn't the responsibility
+ * of the author.  This software is distributed AS-IS.
+ *
+ * $Id$
+ */
+
+/*
+ * This file contains support for the POSIX.4 AIO facility.
+ *
+ * The initial version provides only the (bogus) synchronous semantics
+ * but will support async in the future.  Note that a bit
+ * in a private field allows the user mode subroutine to adapt
+ * the kernel operations to true POSIX.4 for future compatibility.
+ *
+ * This code is used to support true POSIX.4 AIO/LIO with the help
+ * of a user mode subroutine package.  Note that eventually more support
+ * will be pushed into the kernel.
+ */
+
+#include <sys/param.h>
+#include <sys/systm.h>
+#include <sys/sysent.h>
+#include <sys/sysproto.h>
+#include <sys/namei.h>
+#include <sys/filedesc.h>
+#include <sys/kernel.h>
+#include <sys/fcntl.h>
+#include <sys/file.h>
+#include <sys/stat.h>
+#include <sys/unistd.h>
+#include <sys/vnode.h>
+#include <sys/mount.h>
+#include <sys/proc.h>
+#include <sys/uio.h>
+#include <sys/malloc.h>
+#include <sys/dirent.h>
+#include <sys/signalvar.h>
+
+#include <vm/vm.h>
+#include <vm/vm_param.h>
+#include <vm/vm_object.h>
+#include <vm/vm_extern.h>
+#include <sys/sysctl.h>
+#include <sys/aio.h>
+
+
+/*
+ * aio_cancel at the kernel level is a NOOP right now.  It
+ * might be possible to support it partially in user mode, or
+ * in kernel mode later on.
+ */
+int
+aio_cancel(struct proc *p, struct aio_cancel_args *uap, int *retval) {
+	return AIO_NOTCANCELLED;
+}
+
+
+/*
+ * aio_error is implemented in the kernel level for compatibility
+ * purposes only.  For a user mode async implementation, it would be
+ * best to do it in a userland subroutine.
+ */
+int
+aio_error(struct proc *p, struct aio_error_args *uap, int *retval) {
+	int activeflag, errorcode;
+	struct aiocb iocb;
+	int error;
+
+	/*
+	 * Get control block
+	 */
+	if (error = copyin((caddr_t) uap->aiocbp, (caddr_t) &iocb, sizeof iocb))
+		return error;
+	if (iocb._aiocb_private.active == -1)
+		return EFAULT;
+
+	if (iocb._aiocb_private.active != AIO_PMODE_ACTIVE) {
+		retval[0] = EINVAL;
+		return(0);
+	}
+
+	retval[0] = iocb._aiocb_private.error;
+	return(0);
+}
+
+int
+aio_read(struct proc *p, struct aio_read_args *uap, int *retval) {
+	struct filedesc *fdp;
+	struct file *fp;
+	struct uio auio;
+	struct iovec aiov;
+	unsigned int fd;
+	int cnt;
+	struct aiocb iocb;
+	int error;
+
+
+	/*
+	 * Get control block
+	 */
+	if (error = copyin((caddr_t) uap->aiocbp, (caddr_t) &iocb, sizeof iocb))
+		return error;
+
+	/*
+	 * We support sync only for now.
+	 */
+	if ((iocb._aiocb_private.privatemodes & AIO_PMODE_SYNC) == 0)
+		return ENOSYS;
+
+	/*
+	 * Get the fd info for process
+	 */
+	fdp = p->p_fd;
+
+	/*
+	 * Range check file descriptor
+	 */
+	fd = iocb.aio_fildes;
+	if (fd >= fdp->fd_nfiles)
+		return EBADF;
+	fp = fdp->fd_ofiles[fd];
+	if ((fp == NULL) || ((fp->f_flag & FREAD) == 0))
+		return EBADF;
+	if (((int) iocb.aio_offset) == -1)
+		return EINVAL;
+
+	aiov.iov_base = iocb.aio_buf;
+	aiov.iov_len = iocb.aio_nbytes;
+	auio.uio_iov = &aiov;
+	auio.uio_iovcnt = 1;
+	auio.uio_offset = iocb.aio_offset;
+
+	auio.uio_resid = iocb.aio_nbytes;
+	if (auio.uio_resid < 0)
+		return (EINVAL);
+
+	auio.uio_rw = UIO_READ;
+	auio.uio_segflg = UIO_USERSPACE;
+	auio.uio_procp = p;
+
+	cnt = iocb.aio_nbytes;
+	error = (*fp->f_ops->fo_read)(fp, &auio, fp->f_cred);
+	if (error &&
+		(auio.uio_resid != cnt) &&
+		(error == ERESTART || error == EINTR || error == EWOULDBLOCK))
+			error = 0;
+	cnt -= auio.uio_resid;
+	*retval = cnt;
+	return error;
+}
+
+
+/*
+ * Return and suspend aren't supported (yet).
+ */
+int
+aio_return(struct proc *p, struct aio_return_args *uap, int *retval) {
+	return (0);
+}
+
+int
+aio_suspend(struct proc *p, struct aio_suspend_args *uap, int *retval) {
+	return (0);
+}
+
+int
+aio_write(struct proc *p, struct aio_write_args *uap, int *retval) {
+	struct filedesc *fdp;
+	struct file *fp;
+	struct uio auio;
+	struct iovec aiov;
+	unsigned int fd;
+	int cnt;
+	struct aiocb iocb;
+	int error;
+
+	if (error = copyin((caddr_t) uap->aiocbp, (caddr_t) &iocb, sizeof iocb))
+		return error;
+
+	/*
+	 * We support sync only for now.
+	 */
+	if ((iocb._aiocb_private.privatemodes & AIO_PMODE_SYNC) == 0)
+		return ENOSYS;
+
+	/*
+	 * Get the fd info for process
+	 */
+	fdp = p->p_fd;
+
+	/*
+	 * Range check file descriptor
+	 */
+	fd = iocb.aio_fildes;
+	if (fd >= fdp->fd_nfiles)
+		return EBADF;
+	fp = fdp->fd_ofiles[fd];
+	if ((fp == NULL) || ((fp->f_flag & FWRITE) == 0))
+		return EBADF;
+	if (((int) iocb.aio_offset) == -1)
+		return EINVAL;
+
+	aiov.iov_base = iocb.aio_buf;
+	aiov.iov_len = iocb.aio_nbytes;
+	auio.uio_iov = &aiov;
+	auio.uio_iovcnt = 1;
+	auio.uio_offset = iocb.aio_offset;
+
+	auio.uio_resid = iocb.aio_nbytes;
+	if (auio.uio_resid < 0)
+		return (EINVAL);
+
+	auio.uio_rw = UIO_WRITE;
+	auio.uio_segflg = UIO_USERSPACE;
+	auio.uio_procp = p;
+
+	cnt = iocb.aio_nbytes;
+	error = (*fp->f_ops->fo_write)(fp, &auio, fp->f_cred);
+	if (error) {
+		if (auio.uio_resid != cnt) {
+			if (error == ERESTART || error == EINTR || error == EWOULDBLOCK)
+				error = 0;
+			if (error == EPIPE)
+				psignal(p, SIGPIPE);
+		}
+	}
+	cnt -= auio.uio_resid;
+	*retval = cnt;
+	return error;
+}
+
+int
+lio_listio(struct proc *p, struct lio_listio_args *uap, int *retval) {
+	struct filedesc *fdp;
+	struct file *fp;
+	struct uio auio;
+	struct iovec aiov;
+	unsigned int fd;
+	int cnt;
+	unsigned int iocblen, iocbcnt;
+	struct aiocb *iocb;
+	int error;
+	int i;
+
+	if (uap->mode == LIO_NOWAIT)
+		return ENOSYS;
+	iocbcnt = uap->nent;
+	if (iocbcnt > AIO_LISTIO_MAX)
+		return EINVAL;
+	return ENOSYS;
+}