int mmap_exec(u_int k, int envid, int execonly) {
/* munmap all non-MAP_INHERIT regions */
struct Mmap *m2, *m = mmap_list.lh_first;
if (!execonly) return 0;
while (m)
if (m->mmap_flags & MAP_INHERIT) {
/* not implemented - need to set up a region of vm to mmap data */
assert(0);
m = m->mmap_link.le_next;
}
else {
m2 = m->mmap_link.le_next;
assert(msync(m->mmap_addr, m->mmap_len, 0) == 0);
if (m->mmap_filp) {
lock_filp(m->mmap_filp);
filp_refcount_dec(m->mmap_filp);
if (filp_refcount_get(m->mmap_filp) == 0) {
unlock_filp(m->mmap_filp);
close_filp(m->mmap_filp);
} else
unlock_filp(m->mmap_filp);
}
m = m2;
}
return 0;
}
/*===========================================================================*
* do_lseek *
*===========================================================================*/
int do_lseek()
{
/* Perform the lseek(ls_fd, offset, whence) system call. */
register struct filp *rfilp;
int r = OK, seekfd, seekwhence;
off_t offset;
u64_t pos, newpos;
seekfd = job_m_in.ls_fd;
seekwhence = job_m_in.whence;
offset = (off_t) job_m_in.offset_lo;
/* Check to see if the file descriptor is valid. */
if ( (rfilp = get_filp(seekfd, VNODE_READ)) == NULL) return(err_code);
/* No lseek on pipes. */
if (S_ISFIFO(rfilp->filp_vno->v_mode)) {
unlock_filp(rfilp);
return(ESPIPE);
}
/* The value of 'whence' determines the start position to use. */
switch(seekwhence) {
case SEEK_SET: pos = cvu64(0); break;
case SEEK_CUR: pos = rfilp->filp_pos; break;
case SEEK_END: pos = cvul64(rfilp->filp_vno->v_size); break;
default: unlock_filp(rfilp); return(EINVAL);
}
if (offset >= 0)
newpos = add64ul(pos, offset);
else
newpos = sub64ul(pos, -offset);
/* Check for overflow. */
if (ex64hi(newpos) != 0) {
r = EOVERFLOW;
} else if ((off_t) ex64lo(newpos) < 0) { /* no negative file size */
r = EOVERFLOW;
} else {
/* insert the new position into the output message */
m_out.reply_l1 = ex64lo(newpos);
if (cmp64(newpos, rfilp->filp_pos) != 0) {
rfilp->filp_pos = newpos;
/* Inhibit read ahead request */
r = req_inhibread(rfilp->filp_vno->v_fs_e,
rfilp->filp_vno->v_inode_nr);
}
}
unlock_filp(rfilp);
return(r);
}
/*===========================================================================*
* actual_llseek *
*===========================================================================*/
int actual_llseek(struct fproc *rfp, message *m_out, int seekfd, int seekwhence,
u64_t offset)
{
/* Perform the llseek(ls_fd, offset, whence) system call. */
register struct filp *rfilp;
u64_t pos, newpos;
int r = OK;
long off_hi = ex64hi(offset);
/* Check to see if the file descriptor is valid. */
if ( (rfilp = get_filp2(rfp, seekfd, VNODE_READ)) == NULL) {
return(err_code);
}
/* No lseek on pipes. */
if (S_ISFIFO(rfilp->filp_vno->v_mode)) {
unlock_filp(rfilp);
return(ESPIPE);
}
/* The value of 'whence' determines the start position to use. */
switch(seekwhence) {
case SEEK_SET: pos = cvu64(0); break;
case SEEK_CUR: pos = rfilp->filp_pos; break;
case SEEK_END: pos = cvul64(rfilp->filp_vno->v_size); break;
default: unlock_filp(rfilp); return(EINVAL);
}
newpos = pos + offset;
/* Check for overflow. */
if ((off_hi > 0) && cmp64(newpos, pos) < 0)
r = EINVAL;
else if ((off_hi < 0) && cmp64(newpos, pos) > 0)
r = EINVAL;
else {
/* insert the new position into the output message */
m_out->reply_l1 = ex64lo(newpos);
m_out->reply_l2 = ex64hi(newpos);
if (cmp64(newpos, rfilp->filp_pos) != 0) {
rfilp->filp_pos = newpos;
/* Inhibit read ahead request */
r = req_inhibread(rfilp->filp_vno->v_fs_e,
rfilp->filp_vno->v_inode_nr);
}
}
unlock_filp(rfilp);
return(r);
}
int dupvm(struct fproc *rfp, int pfd, int *vmfd, struct filp **newfilp)
{
int result, procfd;
struct filp *f = NULL;
struct fproc *vmf = &fproc[VM_PROC_NR];
*newfilp = NULL;
if ((f = get_filp2(rfp, pfd, VNODE_READ)) == NULL) {
printf("VFS dupvm: get_filp2 failed\n");
return EBADF;
}
if(!f->filp_vno->v_vmnt->m_haspeek) {
unlock_filp(f);
printf("VFS dupvm: no peek available\n");
return EINVAL;
}
assert(f->filp_vno);
assert(f->filp_vno->v_vmnt);
if (!S_ISREG(f->filp_vno->v_mode) && !S_ISBLK(f->filp_vno->v_mode)) {
printf("VFS: mmap regular/blockdev only; dev 0x%x ino %d has mode 0%o\n",
(int) f->filp_vno->v_dev, (int) f->filp_vno->v_inode_nr, (int) f->filp_vno->v_mode);
unlock_filp(f);
return EINVAL;
}
/* get free FD in VM */
if((result=get_fd(vmf, 0, 0, &procfd, NULL)) != OK) {
unlock_filp(f);
printf("VFS dupvm: getfd failed\n");
return result;
}
*vmfd = procfd;
f->filp_count++;
assert(f->filp_count > 0);
vmf->fp_filp[procfd] = f;
/* mmap FD's are inuse */
FD_SET(procfd, &vmf->fp_filp_inuse);
*newfilp = f;
return OK;
}
/*===========================================================================*
* actual_lseek *
*===========================================================================*/
int actual_lseek(struct fproc *rfp, int seekfd, int seekwhence, off_t offset,
off_t *newposp)
{
register struct filp *rfilp;
int r = OK;
off_t pos, newpos;
/* Check to see if the file descriptor is valid. */
if ( (rfilp = get_filp2(rfp, seekfd, VNODE_READ)) == NULL) {
return(err_code);
}
/* No lseek on pipes. */
if (S_ISFIFO(rfilp->filp_vno->v_mode)) {
unlock_filp(rfilp);
return(ESPIPE);
}
/* The value of 'whence' determines the start position to use. */
switch(seekwhence) {
case SEEK_SET: pos = 0; break;
case SEEK_CUR: pos = rfilp->filp_pos; break;
case SEEK_END: pos = rfilp->filp_vno->v_size; break;
default: unlock_filp(rfilp); return(EINVAL);
}
newpos = pos + offset;
/* Check for overflow. */
if ((offset > 0) && (newpos <= pos)) {
r = EOVERFLOW;
} else if ((offset < 0) && (newpos >= pos)) {
r = EOVERFLOW;
} else {
if (newposp != NULL) *newposp = newpos;
if (newpos != rfilp->filp_pos) {
rfilp->filp_pos = newpos;
/* Inhibit read ahead request */
r = req_inhibread(rfilp->filp_vno->v_fs_e,
rfilp->filp_vno->v_inode_nr);
}
}
unlock_filp(rfilp);
return(r);
}
/*===========================================================================*
* do_dup *
*===========================================================================*/
int do_dup()
{
/* Perform the dup(fd) or dup2(fd,fd2) system call. These system calls are
* obsolete. In fact, it is not even possible to invoke them using the
* current library because the library routines call fcntl(). They are
* provided to permit old binary programs to continue to run.
*/
int rfd, rfd2;
struct filp *f;
int r = OK;
scratch(fp).file.fd_nr = job_m_in.fd;
rfd2 = job_m_in.fd2;
/* Is the file descriptor valid? */
rfd = scratch(fp).file.fd_nr & ~DUP_MASK; /* kill off dup2 bit, if on */
if ((f = get_filp(rfd, VNODE_READ)) == NULL) return(err_code);
/* Distinguish between dup and dup2. */
if (!(scratch(fp).file.fd_nr & DUP_MASK)) { /* bit not on */
/* dup(fd) */
r = get_fd(0, 0, &rfd2, NULL);
} else {
/* dup2(old_fd, new_fd) */
if (rfd2 < 0 || rfd2 >= OPEN_MAX) {
r = EBADF;
} else if (rfd == rfd2) { /* ignore the call: dup2(x, x) */
r = rfd2;
} else {
/* All is fine, close new_fd if necessary */
unlock_filp(f); /* or it might deadlock on do_close */
(void) close_fd(fp, rfd2); /* cannot fail */
f = get_filp(rfd, VNODE_READ); /* lock old_fd again */
}
}
if (r == OK) {
/* Success. Set up new file descriptors. */
f->filp_count++;
fp->fp_filp[rfd2] = f;
FD_SET(rfd2, &fp->fp_filp_inuse);
r = rfd2;
}
unlock_filp(f);
return(r);
}
int dupvm(struct fproc *rfp, int pfd, int *vmfd, struct filp **newfilp)
{
int result, procfd;
struct filp *f = NULL;
struct fproc *vmf = fproc_addr(VM_PROC_NR);
*newfilp = NULL;
if ((f = get_filp2(rfp, pfd, VNODE_READ)) == NULL) {
printf("VFS dupvm: get_filp2 failed\n");
return EBADF;
}
if(!(f->filp_vno->v_vmnt->m_fs_flags & RES_HASPEEK)) {
unlock_filp(f);
#if 0 /* Noisy diagnostic for mmap() by ld.so */
printf("VFS dupvm: no peek available\n");
#endif
return EINVAL;
}
assert(f->filp_vno);
assert(f->filp_vno->v_vmnt);
if (!S_ISREG(f->filp_vno->v_mode) && !S_ISBLK(f->filp_vno->v_mode)) {
printf("VFS: mmap regular/blockdev only; dev 0x%llx ino %llu has mode 0%o\n",
f->filp_vno->v_dev, f->filp_vno->v_inode_nr, f->filp_vno->v_mode);
unlock_filp(f);
return EINVAL;
}
/* get free FD in VM */
if((result=get_fd(vmf, 0, 0, &procfd, NULL)) != OK) {
unlock_filp(f);
printf("VFS dupvm: getfd failed\n");
return result;
}
*vmfd = procfd;
f->filp_count++;
assert(f->filp_count > 0);
vmf->fp_filp[procfd] = f;
*newfilp = f;
return OK;
}
/*===========================================================================*
* do_llseek *
*===========================================================================*/
PUBLIC int do_llseek()
{
/* Perform the llseek(ls_fd, offset, whence) system call. */
register struct filp *rfilp;
u64_t pos, newpos;
int r = OK;
/* Check to see if the file descriptor is valid. */
if ( (rfilp = get_filp(m_in.ls_fd, VNODE_READ)) == NULL) return(err_code);
/* No lseek on pipes. */
if (rfilp->filp_vno->v_pipe == I_PIPE) {
unlock_filp(rfilp);
return(ESPIPE);
}
/* The value of 'whence' determines the start position to use. */
switch(m_in.whence) {
case SEEK_SET: pos = cvu64(0); break;
case SEEK_CUR: pos = rfilp->filp_pos; break;
case SEEK_END: pos = cvul64(rfilp->filp_vno->v_size); break;
default: unlock_filp(rfilp); return(EINVAL);
}
newpos = add64(pos, make64(m_in.offset_lo, m_in.offset_high));
/* Check for overflow. */
if (( (long) m_in.offset_high > 0) && cmp64(newpos, pos) < 0)
r = EINVAL;
else if (( (long) m_in.offset_high < 0) && cmp64(newpos, pos) > 0)
r = EINVAL;
else {
rfilp->filp_pos = newpos;
/* insert the new position into the output message */
m_out.reply_l1 = ex64lo(newpos);
m_out.reply_l2 = ex64hi(newpos);
if (cmp64(newpos, rfilp->filp_pos) != 0) {
/* Inhibit read ahead request */
r = req_inhibread(rfilp->filp_vno->v_fs_e,
rfilp->filp_vno->v_inode_nr);
}
}
unlock_filp(rfilp);
return(r);
}
/*===========================================================================*
* do_pending_pipe *
*===========================================================================*/
static void do_pending_pipe(void)
{
int r, op;
struct filp *f;
tll_access_t locktype;
f = fp->fp_filp[fp->fp_fd];
assert(f != NULL);
locktype = (job_call_nr == VFS_READ) ? VNODE_READ : VNODE_WRITE;
op = (job_call_nr == VFS_READ) ? READING : WRITING;
lock_filp(f, locktype);
r = rw_pipe(op, who_e, f, fp->fp_io_buffer, fp->fp_io_nbytes);
if (r != SUSPEND) { /* Do we have results to report? */
/* Process is writing, but there is no reader. Send a SIGPIPE signal.
* This should match the corresponding code in read_write().
*/
if (r == EPIPE && op == WRITING) {
if (!(f->filp_flags & O_NOSIGPIPE))
sys_kill(fp->fp_endpoint, SIGPIPE);
}
replycode(fp->fp_endpoint, r);
}
unlock_filp(f);
}
/*===========================================================================*
* do_fsync *
*===========================================================================*/
PUBLIC int do_fsync()
{
/* Perform the fsync() system call. */
struct filp *rfilp;
struct vmnt *vmp;
dev_t dev;
int r = OK;
if ((rfilp = get_filp(m_in.m1_i1, VNODE_READ)) == NULL) return(err_code);
dev = rfilp->filp_vno->v_dev;
for (vmp = &vmnt[0]; vmp < &vmnt[NR_MNTS]; ++vmp) {
if (vmp->m_dev != NO_DEV && vmp->m_dev == dev &&
vmp->m_fs_e != NONE && vmp->m_root_node != NULL) {
if ((r = lock_vmnt(vmp, VMNT_EXCL)) != OK)
break;
req_sync(vmp->m_fs_e);
unlock_vmnt(vmp);
}
}
unlock_filp(rfilp);
return(r);
}
/*===========================================================================*
* do_getdents *
*===========================================================================*/
int do_getdents(void)
{
/* Perform the getdents(fd, buf, size) system call. */
int r = OK;
off_t new_pos;
register struct filp *rfilp;
scratch(fp).file.fd_nr = job_m_in.m_lc_vfs_readwrite.fd;
scratch(fp).io.io_buffer = job_m_in.m_lc_vfs_readwrite.buf;
scratch(fp).io.io_nbytes = job_m_in.m_lc_vfs_readwrite.len;
/* Is the file descriptor valid? */
if ( (rfilp = get_filp(scratch(fp).file.fd_nr, VNODE_READ)) == NULL)
return(err_code);
if (!(rfilp->filp_mode & R_BIT))
r = EBADF;
else if (!S_ISDIR(rfilp->filp_vno->v_mode))
r = EBADF;
if (r == OK) {
r = req_getdents(rfilp->filp_vno->v_fs_e, rfilp->filp_vno->v_inode_nr,
rfilp->filp_pos, scratch(fp).io.io_buffer,
scratch(fp).io.io_nbytes, &new_pos, 0);
if (r > 0) rfilp->filp_pos = new_pos;
}
unlock_filp(rfilp);
return(r);
}
/*===========================================================================*
* do_fsync *
*===========================================================================*/
int do_fsync(void)
{
/* Perform the fsync() system call. */
struct filp *rfilp;
struct vmnt *vmp;
dev_t dev;
int r = OK;
scratch(fp).file.fd_nr = job_m_in.m_lc_vfs_fsync.fd;
if ((rfilp = get_filp(scratch(fp).file.fd_nr, VNODE_READ)) == NULL)
return(err_code);
dev = rfilp->filp_vno->v_dev;
unlock_filp(rfilp);
for (vmp = &vmnt[0]; vmp < &vmnt[NR_MNTS]; ++vmp) {
if (vmp->m_dev != dev) continue;
if ((r = lock_vmnt(vmp, VMNT_READ)) != OK)
break;
if (vmp->m_dev != NO_DEV && vmp->m_dev == dev &&
vmp->m_fs_e != NONE && vmp->m_root_node != NULL) {
req_sync(vmp->m_fs_e);
}
unlock_vmnt(vmp);
}
return(r);
}
/*===========================================================================*
* do_pending_pipe *
*===========================================================================*/
static void *do_pending_pipe(void *arg)
{
int r, op;
struct job my_job;
struct filp *f;
tll_access_t locktype;
my_job = *((struct job *) arg);
fp = my_job.j_fp;
lock_proc(fp, 1 /* force lock */);
f = scratch(fp).file.filp;
assert(f != NULL);
scratch(fp).file.filp = NULL;
locktype = (job_call_nr == READ) ? VNODE_READ : VNODE_WRITE;
op = (job_call_nr == READ) ? READING : WRITING;
lock_filp(f, locktype);
r = rw_pipe(op, who_e, f, scratch(fp).io.io_buffer, scratch(fp).io.io_nbytes);
if (r != SUSPEND) /* Do we have results to report? */
reply(fp->fp_endpoint, r);
unlock_filp(f);
thread_cleanup(fp);
return(NULL);
}
/*===========================================================================*
* do_fstat *
*===========================================================================*/
int do_fstat()
{
/* Perform the fstat(fd, buf) system call. */
register struct filp *rfilp;
int r, pipe_pos = 0, old_stat = 0, rfd;
vir_bytes statbuf;
statbuf = (vir_bytes) job_m_in.buffer;
rfd = job_m_in.fd;
if (job_call_nr == PREV_FSTAT)
old_stat = 1;
/* Is the file descriptor valid? */
if ((rfilp = get_filp(rfd, VNODE_READ)) == NULL) return(err_code);
/* If we read from a pipe, send position too */
if (rfilp->filp_vno->v_pipe == I_PIPE) {
if (rfilp->filp_mode & R_BIT)
if (ex64hi(rfilp->filp_pos) != 0) {
panic("do_fstat: bad position in pipe");
}
pipe_pos = ex64lo(rfilp->filp_pos);
}
r = req_stat(rfilp->filp_vno->v_fs_e, rfilp->filp_vno->v_inode_nr,
who_e, statbuf, pipe_pos, old_stat);
unlock_filp(rfilp);
return(r);
}
static int
cdev_ioctl(struct file *filp, unsigned int request, char *argp) {
int status;
dev_t device = filp->f_dev;
unlock_filp(filp);
signals_off();
status = (*cdevsw[major(device)].d_ioctl)(device, request, argp, GENFLAG(filp), curproc);
if (status == 0) {
#if 1
if (request == TIOCSCTTY) {
{int ret = proc_controlt(-1,-1,(int)filp); assert(ret == 0);}
}
#endif
} else {
errno = status; status = -1;
}
signals_on();
/* HBXX - Race condition, can receive a signal during the return and next
unlock_filp() */
lock_filp(filp);
return status;
}
void mmap_exit(void *arg) {
/* munmap all regions */
struct Mmap *m2, *m = mmap_list.lh_first;
while (m) {
m2 = m->mmap_link.le_next;
assert(msync(m->mmap_addr, m->mmap_len, 0) == 0);
if (m->mmap_filp) {
lock_filp(m->mmap_filp);
filp_refcount_dec(m->mmap_filp);
if (filp_refcount_get(m->mmap_filp) == 0) {
unlock_filp(m->mmap_filp);
close_filp(m->mmap_filp);
} else
unlock_filp(m->mmap_filp);
}
m = m2;
}
}
/*===========================================================================*
* do_fchdir *
*===========================================================================*/
int do_fchdir()
{
/* Change directory on already-opened fd. */
struct filp *rfilp;
int r, rfd;
rfd = job_m_in.fd;
/* Is the file descriptor valid? */
if ((rfilp = get_filp(rfd, VNODE_READ)) == NULL) return(err_code);
r = change_into(&fp->fp_wd, rfilp->filp_vno);
unlock_filp(rfilp);
return(r);
}
/*===========================================================================*
* actual_read_write_peek *
*===========================================================================*/
int actual_read_write_peek(struct fproc *rfp, int rw_flag, int io_fd,
vir_bytes io_buf, size_t io_nbytes)
{
/* Perform read(fd, buffer, nbytes) or write(fd, buffer, nbytes) call. */
struct filp *f;
tll_access_t locktype;
int r;
int ro = 1;
if(rw_flag == WRITING) ro = 0;
scratch(rfp).file.fd_nr = io_fd;
scratch(rfp).io.io_buffer = io_buf;
scratch(rfp).io.io_nbytes = io_nbytes;
locktype = rw_flag == WRITING ? VNODE_WRITE : VNODE_READ;
if ((f = get_filp2(rfp, scratch(rfp).file.fd_nr, locktype)) == NULL)
return(err_code);
assert(f->filp_count > 0);
if (((f->filp_mode) & (ro ? R_BIT : W_BIT)) == 0) {
unlock_filp(f);
return(EBADF);
}
if (scratch(rfp).io.io_nbytes == 0) {
unlock_filp(f);
return(0); /* so char special files need not check for 0*/
}
r = read_write(rfp, rw_flag, f, scratch(rfp).io.io_buffer,
scratch(rfp).io.io_nbytes, who_e);
unlock_filp(f);
return(r);
}
static inline void close_filp(struct file *filp) {
int fd;
for (fd = NR_OPEN - 1; fd >= 0; fd--)
if (__current->fd[fd] == NULL) {
__current->fd[fd] = filp;
break;
}
assert(fd >= 0);
__current->cloexec_flag[fd] = 0;
lock_filp(filp);
filp_refcount_inc(filp);
unlock_filp(filp);
close(fd);
}
int mmap_fork(u_int k, int envid, int NewPid) {
/* increase the refcounts */
struct Mmap *m = mmap_list.lh_first;
while (m) {
if (m->mmap_filp) {
lock_filp(m->mmap_filp);
filp_refcount_inc(m->mmap_filp);
unlock_filp(m->mmap_filp);
}
m = m->mmap_link.le_next;
}
return 0;
}
static int
cdev_write(struct file *filp, char *buffer, int nbyte, int blocking) {
int status;
dev_t dev = filp->f_dev;
struct uio uio;
struct iovec iov[4];
char buf[CLALLOCSZ*8];
int i;
demand(filp, bogus filp);
unlock_filp(filp);
signals_off();
DPRINTF(CLU_LEVEL,
("cdev_write: filp: %08x offset: %qd nbyte: %d\n",
(int)filp, filp->f_pos, nbyte));
assert(nbyte <= CLALLOCSZ*8);
memcpy(buf,buffer,nbyte);
iov[0].iov_base = buf;
iov[0].iov_len = nbyte;
uio.uio_iov = iov;
uio.uio_iovcnt = 1;
uio.uio_offset = 0;
uio.uio_resid = nbyte;
uio.uio_rw = UIO_WRITE;
k0printf("Write: %d: ",uio.uio_resid);
for (i = 0; i < uio.uio_resid; i++)
k0printf(">%d (%c)",(unsigned int)buf[i],buf[i]);
EnterCritical();
status = (*cdevsw[major(dev)].d_write)(dev, &uio, GENFLAG(filp));
ExitCritical();
k0printf("Read: %d: ",nbyte - uio.uio_resid);
if (status == 0) {
status = nbyte - uio.uio_resid;
} else {
errno = status; status = -1;
}
signals_on();
/* HBXX - Race condition, can receive a signal during the return and next
unlock_filp() */
lock_filp(filp);
return status;
}
static int
cdev_read(struct file *filp, char *buffer, int nbyte, int blocking) {
int status;
dev_t dev = filp->f_dev;
struct uio uio;
struct iovec iov[4];
demand(filp, bogus filp);
DPRINTF(CLU_LEVEL,
("cdev_read: filp: %08x offset: %qd nbyte: %d\n",
(int)filp, filp->f_pos, nbyte));
/* if (nbyte > CLALLOCSZ) {fprintf(stderr,"ncdev_read, warn large nbyte\n");} */
iov[0].iov_base = buffer;
iov[0].iov_len = nbyte;
uio.uio_iov = iov;
uio.uio_iovcnt = 1;
uio.uio_offset = 0;
uio.uio_resid = nbyte;
uio.uio_rw = UIO_READ;
signals_off();
unlock_filp(filp);
EnterCritical();
status = (*cdevsw[major(dev)].d_read)(dev, &uio, GENFLAG(filp));
ExitCritical();
signals_on();
k0printf("Read: %d: ",nbyte - uio.uio_resid);
if (status == 0) {
status = nbyte - uio.uio_resid;
} else {
errno = status;
status = -1;
}
/* HBXX - Race condition, can receive a signal during the return and next
unlock_filp() */
lock_filp(filp);
#if 0
if (status >= 0) {
extern void pr_uio();
kprintf("read(%d,%d) ",major(dev),minor(dev));
pr_uio(&uio);
}
#endif
return status;
}
/*===========================================================================*
* do_fstatvfs *
*===========================================================================*/
int do_fstatvfs()
{
/* Perform the fstat(fd, buf) system call. */
register struct filp *rfilp;
int r, rfd;
vir_bytes statbuf;
rfd = job_m_in.fd;
statbuf = (vir_bytes) job_m_in.name2;
/* Is the file descriptor valid? */
if ((rfilp = get_filp(rfd, VNODE_READ)) == NULL) return(err_code);
r = req_statvfs(rfilp->filp_vno->v_fs_e, who_e, statbuf);
unlock_filp(rfilp);
return(r);
}
/*===========================================================================*
* do_pending_pipe *
*===========================================================================*/
static void do_pending_pipe(void)
{
vir_bytes buf;
size_t nbytes, cum_io;
int r, op, fd;
struct filp *f;
tll_access_t locktype;
assert(fp->fp_blocked_on == FP_BLOCKED_ON_NONE);
/*
* We take all our needed resumption state from the m_in message, which is
* filled by unblock(). Since this is an internal resumption, there is no
* need to perform extensive checks on the message fields.
*/
fd = job_m_in.m_lc_vfs_readwrite.fd;
buf = job_m_in.m_lc_vfs_readwrite.buf;
nbytes = job_m_in.m_lc_vfs_readwrite.len;
cum_io = job_m_in.m_lc_vfs_readwrite.cum_io;
f = fp->fp_filp[fd];
assert(f != NULL);
locktype = (job_call_nr == VFS_READ) ? VNODE_READ : VNODE_WRITE;
op = (job_call_nr == VFS_READ) ? READING : WRITING;
lock_filp(f, locktype);
r = rw_pipe(op, who_e, f, job_call_nr, fd, buf, nbytes, cum_io);
if (r != SUSPEND) { /* Do we have results to report? */
/* Process is writing, but there is no reader. Send a SIGPIPE signal.
* This should match the corresponding code in read_write().
*/
if (r == EPIPE && op == WRITING) {
if (!(f->filp_flags & O_NOSIGPIPE))
sys_kill(fp->fp_endpoint, SIGPIPE);
}
replycode(fp->fp_endpoint, r);
}
unlock_filp(f);
}
/*===========================================================================*
* do_fstatvfs *
*===========================================================================*/
int do_fstatvfs(void)
{
/* Perform the fstatvfs1(fd, buf, flags) system call. */
register struct filp *rfilp;
int r, rfd, flags;
vir_bytes statbuf;
rfd = job_m_in.m_lc_vfs_statvfs1.fd;
statbuf = job_m_in.m_lc_vfs_statvfs1.buf;
flags = job_m_in.m_lc_vfs_statvfs1.flags;
/* Is the file descriptor valid? */
if ((rfilp = get_filp(rfd, VNODE_READ)) == NULL) return(err_code);
r = fill_statvfs(rfilp->filp_vno->v_vmnt, who_e, statbuf, flags);
unlock_filp(rfilp);
return(r);
}
/*===========================================================================*
* select_cancel_filp *
*===========================================================================*/
static void select_cancel_filp(struct filp *f)
{
/* Reduce number of select users of this filp */
assert(f);
assert(f->filp_selectors >= 0);
if (f->filp_selectors == 0) return;
if (f->filp_count == 0) return;
select_lock_filp(f, f->filp_select_ops);
f->filp_selectors--;
if (f->filp_selectors == 0) {
/* No one selecting on this filp anymore, forget about select state */
f->filp_select_ops = 0;
f->filp_select_flags = 0;
f->filp_pipe_select_ops = 0;
}
unlock_filp(f);
}
/*===========================================================================*
* do_select *
*===========================================================================*/
int do_select(void)
{
/* Implement the select(nfds, readfds, writefds, errorfds, timeout) system
* call. First we copy the arguments and verify their sanity. Then we check
* whether there are file descriptors that satisfy the select call right of the
* bat. If so, or if there are no ready file descriptors but the process
* requested to return immediately, we return the result. Otherwise we set a
* timeout and wait for either the file descriptors to become ready or the
* timer to go off. If no timeout value was provided, we wait indefinitely. */
int r, nfds, do_timeout = 0, fd, s;
struct timeval timeout;
struct selectentry *se;
vir_bytes vtimeout;
nfds = job_m_in.SEL_NFDS;
vtimeout = (vir_bytes) job_m_in.SEL_TIMEOUT;
/* Sane amount of file descriptors? */
if (nfds < 0 || nfds > OPEN_MAX) return(EINVAL);
/* Find a slot to store this select request */
for (s = 0; s < MAXSELECTS; s++)
if (selecttab[s].requestor == NULL) /* Unused slot */
break;
if (s >= MAXSELECTS) return(ENOSPC);
se = &selecttab[s];
wipe_select(se); /* Clear results of previous usage */
se->requestor = fp;
se->req_endpt = who_e;
se->vir_readfds = (fd_set *) job_m_in.SEL_READFDS;
se->vir_writefds = (fd_set *) job_m_in.SEL_WRITEFDS;
se->vir_errorfds = (fd_set *) job_m_in.SEL_ERRORFDS;
/* Copy fdsets from the process */
if ((r = copy_fdsets(se, nfds, FROM_PROC)) != OK) {
se->requestor = NULL;
return(r);
}
/* Did the process set a timeout value? If so, retrieve it. */
if (vtimeout != 0) {
do_timeout = 1;
r = sys_vircopy(who_e, (vir_bytes) vtimeout, SELF,
(vir_bytes) &timeout, sizeof(timeout));
if (r != OK) {
se->requestor = NULL;
return(r);
}
}
/* No nonsense in the timeval */
if (do_timeout && (timeout.tv_sec < 0 || timeout.tv_usec < 0)) {
se->requestor = NULL;
return(EINVAL);
}
/* If there is no timeout, we block forever. Otherwise, we block up to the
* specified time interval.
*/
if (!do_timeout) /* No timeout value set */
se->block = 1;
else if (do_timeout && (timeout.tv_sec > 0 || timeout.tv_usec > 0))
se->block = 1;
else /* timeout set as (0,0) - this effects a poll */
se->block = 0;
se->expiry = 0; /* no timer set (yet) */
/* Verify that file descriptors are okay to select on */
for (fd = 0; fd < nfds; fd++) {
struct filp *f;
unsigned int type, ops;
/* Because the select() interface implicitly includes file descriptors
* you might not want to select on, we have to figure out whether we're
* interested in them. Typically, these file descriptors include fd's
* inherited from the parent proc and file descriptors that have been
* close()d, but had a lower fd than one in the current set.
*/
if (!(ops = tab2ops(fd, se)))
continue; /* No operations set; nothing to do for this fd */
/* Get filp belonging to this fd */
f = se->filps[fd] = get_filp(fd, VNODE_READ);
if (f == NULL) {
if (err_code == EBADF)
r = err_code;
else /* File descriptor is 'ready' to return EIO */
r = EINTR;
se->requestor = NULL;
return(r);
}
/* Check file types. According to POSIX 2008:
* "The pselect() and select() functions shall support regular files,
* terminal and pseudo-terminal devices, FIFOs, pipes, and sockets. The
* behavior of pselect() and select() on file descriptors that refer to
* other types of file is unspecified."
*
* In our case, terminal and pseudo-terminal devices are handled by the
* TTY major and sockets by either INET major (socket type AF_INET) or
* PFS major (socket type AF_UNIX). PFS acts as an FS when it handles
* pipes and as a driver when it handles sockets. Additionally, we
* support select on the LOG major to handle kernel logging, which is
* beyond the POSIX spec. */
se->type[fd] = -1;
for (type = 0; type < SEL_FDS; type++) {
if (fdtypes[type].type_match(f)) {
se->type[fd] = type;
se->nfds = fd+1;
se->filps[fd]->filp_selectors++;
break;
}
}
unlock_filp(f);
if (se->type[fd] == -1) { /* Type not found */
se->requestor = NULL;
return(EBADF);
}
}
/* Check all file descriptors in the set whether one is 'ready' now */
for (fd = 0; fd < nfds; fd++) {
int ops, r;
struct filp *f;
/* Again, check for involuntarily selected fd's */
if (!(ops = tab2ops(fd, se)))
continue; /* No operations set; nothing to do for this fd */
/* Test filp for select operations if not already done so. e.g.,
* processes sharing a filp and both doing a select on that filp. */
f = se->filps[fd];
if ((f->filp_select_ops & ops) != ops) {
int wantops;
wantops = (f->filp_select_ops |= ops);
r = do_select_request(se, fd, &wantops);
if (r != OK && r != SUSPEND)
break; /* Error or bogus return code; abort */
/* The select request above might have turned on/off some
* operations because they were 'ready' or not meaningful.
* Either way, we might have a result and we need to store them
* in the select table entry. */
if (wantops & ops) ops2tab(wantops, fd, se);
}
}
if ((se->nreadyfds > 0 || !se->block) && !is_deferred(se)) {
/* fd's were found that were ready to go right away, and/or
* we were instructed not to block at all. Must return
* immediately.
*/
r = copy_fdsets(se, se->nfds, TO_PROC);
select_cancel_all(se);
se->requestor = NULL;
if (r != OK)
return(r);
else if (se->error != OK)
return(se->error);
return(se->nreadyfds);
}
/* Convert timeval to ticks and set the timer. If it fails, undo
* all, return error.
*/
if (do_timeout) {
int ticks;
/* Open Group:
* "If the requested timeout interval requires a finer
* granularity than the implementation supports, the
* actual timeout interval shall be rounded up to the next
* supported value."
*/
#define USECPERSEC 1000000
while(timeout.tv_usec >= USECPERSEC) {
/* this is to avoid overflow with *system_hz below */
timeout.tv_usec -= USECPERSEC;
timeout.tv_sec++;
}
ticks = timeout.tv_sec * system_hz +
(timeout.tv_usec * system_hz + USECPERSEC-1) / USECPERSEC;
se->expiry = ticks;
set_timer(&se->timer, ticks, select_timeout_check, s);
}
/* process now blocked */
suspend(FP_BLOCKED_ON_SELECT);
return(SUSPEND);
}
/*===========================================================================*
* common_open *
*===========================================================================*/
int common_open(char path[PATH_MAX], int oflags, mode_t omode)
{
/* Common code from do_creat and do_open. */
int b, r, exist = TRUE, major_dev;
dev_t dev;
mode_t bits;
struct filp *filp, *filp2;
struct vnode *vp;
struct vmnt *vmp;
struct dmap *dp;
struct lookup resolve;
/* Remap the bottom two bits of oflags. */
bits = (mode_t) mode_map[oflags & O_ACCMODE];
if (!bits) return(EINVAL);
/* See if file descriptor and filp slots are available. */
if ((r = get_fd(0, bits, &(scratch(fp).file.fd_nr), &filp)) != OK) return(r);
lookup_init(&resolve, path, PATH_NOFLAGS, &vmp, &vp);
/* If O_CREATE is set, try to make the file. */
if (oflags & O_CREAT) {
omode = I_REGULAR | (omode & ALLPERMS & fp->fp_umask);
vp = new_node(&resolve, oflags, omode);
r = err_code;
if (r == OK) exist = FALSE; /* We just created the file */
else if (r != EEXIST) { /* other error */
if (vp) unlock_vnode(vp);
unlock_filp(filp);
return(r);
}
else exist = !(oflags & O_EXCL);/* file exists, if the O_EXCL
flag is set this is an error */
} else {
/* Scan path name */
resolve.l_vmnt_lock = VMNT_READ;
resolve.l_vnode_lock = VNODE_OPCL;
if ((vp = eat_path(&resolve, fp)) == NULL) {
unlock_filp(filp);
return(err_code);
}
if (vmp != NULL) unlock_vmnt(vmp);
}
/* Claim the file descriptor and filp slot and fill them in. */
fp->fp_filp[scratch(fp).file.fd_nr] = filp;
FD_SET(scratch(fp).file.fd_nr, &fp->fp_filp_inuse);
filp->filp_count = 1;
filp->filp_vno = vp;
filp->filp_flags = oflags;
/* Only do the normal open code if we didn't just create the file. */
if (exist) {
/* Check protections. */
if ((r = forbidden(fp, vp, bits)) == OK) {
/* Opening reg. files, directories, and special files differ */
switch (vp->v_mode & S_IFMT) {
case S_IFREG:
/* Truncate regular file if O_TRUNC. */
if (oflags & O_TRUNC) {
if ((r = forbidden(fp, vp, W_BIT)) != OK)
break;
truncate_vnode(vp, 0);
}
break;
case S_IFDIR:
/* Directories may be read but not written. */
r = (bits & W_BIT ? EISDIR : OK);
break;
case S_IFCHR:
/* Invoke the driver for special processing. */
dev = (dev_t) vp->v_sdev;
/* TTY needs to know about the O_NOCTTY flag. */
r = dev_open(dev, who_e, bits | (oflags & O_NOCTTY));
if (r == SUSPEND) suspend(FP_BLOCKED_ON_DOPEN);
else vp = filp->filp_vno; /* Might be updated by
* dev_open/clone_opcl */
break;
case S_IFBLK:
lock_bsf();
/* Invoke the driver for special processing. */
dev = (dev_t) vp->v_sdev;
r = bdev_open(dev, bits);
if (r != OK) {
unlock_bsf();
break;
}
major_dev = major(vp->v_sdev);
dp = &dmap[major_dev];
if (dp->dmap_driver == NONE) {
printf("VFS: block driver disappeared!\n");
unlock_bsf();
r = ENXIO;
break;
}
/* Check whether the device is mounted or not. If so,
* then that FS is responsible for this device.
* Otherwise we default to ROOT_FS.
*/
vp->v_bfs_e = ROOT_FS_E; /* By default */
for (vmp = &vmnt[0]; vmp < &vmnt[NR_MNTS]; ++vmp)
if (vmp->m_dev == vp->v_sdev &&
!(vmp->m_flags & VMNT_FORCEROOTBSF)) {
vp->v_bfs_e = vmp->m_fs_e;
}
/* Send the driver label to the file system that will
* handle the block I/O requests (even when its label
* and endpoint are known already), but only when it is
* the root file system. Other file systems will
* already have it anyway.
*/
if (vp->v_bfs_e != ROOT_FS_E) {
unlock_bsf();
break;
}
if (req_newdriver(vp->v_bfs_e, vp->v_sdev,
dp->dmap_label) != OK) {
printf("VFS: error sending driver label\n");
bdev_close(dev);
r = ENXIO;
}
unlock_bsf();
break;
case S_IFIFO:
/* Create a mapped inode on PFS which handles reads
and writes to this named pipe. */
tll_upgrade(&vp->v_lock);
r = map_vnode(vp, PFS_PROC_NR);
if (r == OK) {
if (vp->v_ref_count == 1) {
vp->v_pipe_rd_pos = 0;
vp->v_pipe_wr_pos = 0;
if (vp->v_size != 0)
r = truncate_vnode(vp, 0);
}
oflags |= O_APPEND; /* force append mode */
filp->filp_flags = oflags;
}
if (r == OK) {
r = pipe_open(vp, bits, oflags);
}
if (r != ENXIO) {
/* See if someone else is doing a rd or wt on
* the FIFO. If so, use its filp entry so the
* file position will be automatically shared.
*/
b = (bits & R_BIT ? R_BIT : W_BIT);
filp->filp_count = 0; /* don't find self */
if ((filp2 = find_filp(vp, b)) != NULL) {
/* Co-reader or writer found. Use it.*/
fp->fp_filp[scratch(fp).file.fd_nr] = filp2;
filp2->filp_count++;
filp2->filp_vno = vp;
filp2->filp_flags = oflags;
/* v_count was incremented after the vnode
* has been found. i_count was incremented
* incorrectly in FS, not knowing that we
* were going to use an existing filp
* entry. Correct this error.
*/
unlock_vnode(vp);
put_vnode(vp);
} else {
/* Nobody else found. Restore filp. */
filp->filp_count = 1;
}
}
break;
}
}
}
unlock_filp(filp);
/* If error, release inode. */
if (r != OK) {
if (r != SUSPEND) {
fp->fp_filp[scratch(fp).file.fd_nr] = NULL;
FD_CLR(scratch(fp).file.fd_nr, &fp->fp_filp_inuse);
filp->filp_count = 0;
filp->filp_vno = NULL;
put_vnode(vp);
}
} else {
r = scratch(fp).file.fd_nr;
}
return(r);
}
/*===========================================================================*
* free_proc *
*===========================================================================*/
static void free_proc(int flags)
{
int i;
register struct fproc *rfp;
register struct filp *rfilp;
register struct vnode *vp;
dev_t dev;
if (fp->fp_endpoint == NONE)
panic("free_proc: already free");
if (fp_is_blocked(fp))
unpause();
/* Loop on file descriptors, closing any that are open. */
for (i = 0; i < OPEN_MAX; i++) {
(void) close_fd(fp, i);
}
/* Release root and working directories. */
if (fp->fp_rd) { put_vnode(fp->fp_rd); fp->fp_rd = NULL; }
if (fp->fp_wd) { put_vnode(fp->fp_wd); fp->fp_wd = NULL; }
/* The rest of these actions is only done when processes actually exit. */
if (!(flags & FP_EXITING)) return;
fp->fp_flags |= FP_EXITING;
/* Check if any process is SUSPENDed on this driver.
* If a driver exits, unmap its entries in the dmap table.
* (unmapping has to be done after the first step, because the
* dmap table is used in the first step.)
*/
unsuspend_by_endpt(fp->fp_endpoint);
dmap_unmap_by_endpt(fp->fp_endpoint);
worker_stop_by_endpt(fp->fp_endpoint); /* Unblock waiting threads */
vmnt_unmap_by_endpt(fp->fp_endpoint); /* Invalidate open files if this
* was an active FS */
/* If a session leader exits and it has a controlling tty, then revoke
* access to its controlling tty from all other processes using it.
*/
if ((fp->fp_flags & FP_SESLDR) && fp->fp_tty != 0) {
dev = fp->fp_tty;
for (rfp = &fproc[0]; rfp < &fproc[NR_PROCS]; rfp++) {
if(rfp->fp_pid == PID_FREE) continue;
if (rfp->fp_tty == dev) rfp->fp_tty = 0;
for (i = 0; i < OPEN_MAX; i++) {
if ((rfilp = rfp->fp_filp[i]) == NULL) continue;
if (rfilp->filp_mode == FILP_CLOSED) continue;
vp = rfilp->filp_vno;
if (!S_ISCHR(vp->v_mode)) continue;
if (vp->v_sdev != dev) continue;
lock_filp(rfilp, VNODE_READ);
(void) cdev_close(dev); /* Ignore any errors. */
/* FIXME: missing select check */
rfilp->filp_mode = FILP_CLOSED;
unlock_filp(rfilp);
}
}
}
/* Exit done. Mark slot as free. */
fp->fp_endpoint = NONE;
fp->fp_pid = PID_FREE;
fp->fp_flags = FP_NOFLAGS;
}
/*===========================================================================*
* do_fcntl *
*===========================================================================*/
int do_fcntl(void)
{
/* Perform the fcntl(fd, cmd, ...) system call. */
register struct filp *f;
int new_fd, fl, r = OK, fcntl_req, fcntl_argx;
tll_access_t locktype;
scratch(fp).file.fd_nr = job_m_in.m_lc_vfs_fcntl.fd;
scratch(fp).io.io_buffer = job_m_in.m_lc_vfs_fcntl.arg_ptr;
scratch(fp).io.io_nbytes = job_m_in.m_lc_vfs_fcntl.cmd;
fcntl_req = job_m_in.m_lc_vfs_fcntl.cmd;
fcntl_argx = job_m_in.m_lc_vfs_fcntl.arg_int;
/* Is the file descriptor valid? */
locktype = (fcntl_req == F_FREESP) ? VNODE_WRITE : VNODE_READ;
if ((f = get_filp(scratch(fp).file.fd_nr, locktype)) == NULL)
return(err_code);
switch (fcntl_req) {
case F_DUPFD:
/* This replaces the old dup() system call. */
if (fcntl_argx < 0 || fcntl_argx >= OPEN_MAX) r = EINVAL;
else if ((r = get_fd(fp, fcntl_argx, 0, &new_fd, NULL)) == OK) {
f->filp_count++;
fp->fp_filp[new_fd] = f;
r = new_fd;
}
break;
case F_GETFD:
/* Get close-on-exec flag (FD_CLOEXEC in POSIX Table 6-2). */
r = 0;
if (FD_ISSET(scratch(fp).file.fd_nr, &fp->fp_cloexec_set))
r = FD_CLOEXEC;
break;
case F_SETFD:
/* Set close-on-exec flag (FD_CLOEXEC in POSIX Table 6-2). */
if (fcntl_argx & FD_CLOEXEC)
FD_SET(scratch(fp).file.fd_nr, &fp->fp_cloexec_set);
else
FD_CLR(scratch(fp).file.fd_nr, &fp->fp_cloexec_set);
break;
case F_GETFL:
/* Get file status flags (O_NONBLOCK and O_APPEND). */
fl = f->filp_flags & (O_NONBLOCK | O_APPEND | O_ACCMODE);
r = fl;
break;
case F_SETFL:
/* Set file status flags (O_NONBLOCK and O_APPEND). */
fl = O_NONBLOCK | O_APPEND;
f->filp_flags = (f->filp_flags & ~fl) | (fcntl_argx & fl);
break;
case F_GETLK:
case F_SETLK:
case F_SETLKW:
/* Set or clear a file lock. */
r = lock_op(f, fcntl_req);
break;
case F_FREESP:
{
/* Free a section of a file */
off_t start, end, offset;
struct flock flock_arg;
/* Check if it's a regular file. */
if (!S_ISREG(f->filp_vno->v_mode)) r = EINVAL;
else if (!(f->filp_mode & W_BIT)) r = EBADF;
else {
/* Copy flock data from userspace. */
r = sys_datacopy_wrapper(who_e, scratch(fp).io.io_buffer,
SELF, (vir_bytes) &flock_arg, sizeof(flock_arg));
}
if (r != OK) break;
/* Convert starting offset to signed. */
offset = (off_t) flock_arg.l_start;
/* Figure out starting position base. */
switch(flock_arg.l_whence) {
case SEEK_SET: start = 0; break;
case SEEK_CUR: start = f->filp_pos; break;
case SEEK_END: start = f->filp_vno->v_size; break;
default: r = EINVAL;
}
if (r != OK) break;
/* Check for overflow or underflow. */
if (offset > 0 && start + offset < start) r = EINVAL;
else if (offset < 0 && start + offset > start) r = EINVAL;
else {
start += offset;
if (start < 0) r = EINVAL;
}
if (r != OK) break;
if (flock_arg.l_len != 0) {
if (start >= f->filp_vno->v_size) r = EINVAL;
else if ((end = start + flock_arg.l_len) <= start) r = EINVAL;
else if (end > f->filp_vno->v_size) end = f->filp_vno->v_size;
} else {
end = 0;
}
if (r != OK) break;
r = req_ftrunc(f->filp_vno->v_fs_e, f->filp_vno->v_inode_nr,start,end);
if (r == OK && flock_arg.l_len == 0)
f->filp_vno->v_size = start;
break;
}
case F_GETNOSIGPIPE:
r = !!(f->filp_flags & O_NOSIGPIPE);
break;
case F_SETNOSIGPIPE:
if (fcntl_argx)
f->filp_flags |= O_NOSIGPIPE;
else
f->filp_flags &= ~O_NOSIGPIPE;
break;
case F_FLUSH_FS_CACHE:
{
struct vnode *vn = f->filp_vno;
mode_t mode = f->filp_vno->v_mode;
if (!super_user) {
r = EPERM;
} else if (S_ISBLK(mode)) {
/* Block device; flush corresponding device blocks. */
r = req_flush(vn->v_bfs_e, vn->v_sdev);
} else if (S_ISREG(mode) || S_ISDIR(mode)) {
/* Directory or regular file; flush hosting FS blocks. */
r = req_flush(vn->v_fs_e, vn->v_dev);
} else {
/* Remaining cases.. Meaning unclear. */
r = ENODEV;
}
break;
}
default:
r = EINVAL;
}
unlock_filp(f);
return(r);
}