root/sys/compat/linux32/common/linux32_dirent.c
/*      $NetBSD: linux32_dirent.c,v 1.20 2019/08/23 07:53:36 maxv Exp $ */

/*-
 * Copyright (c) 2006 Emmanuel Dreyfus, 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. 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 Emmanuel Dreyfus
 * 4. The name of the author may not be used to endorse or promote 
 *    products derived from this software without specific prior written 
 *    permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE THE AUTHOR 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 AUTHOR 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.
 */

#include <sys/cdefs.h>

__KERNEL_RCSID(0, "$NetBSD: linux32_dirent.c,v 1.20 2019/08/23 07:53:36 maxv Exp $");

#include <sys/types.h>
#include <sys/param.h>
#include <sys/fstypes.h>
#include <sys/signal.h>
#include <sys/dirent.h>
#include <sys/kernel.h>
#include <sys/fcntl.h>
#include <sys/file.h>
#include <sys/filedesc.h>
#include <sys/malloc.h>
#include <sys/select.h>
#include <sys/proc.h>
#include <sys/ucred.h>
#include <sys/vnode.h>
#include <sys/swap.h>

#include <machine/types.h>

#include <sys/syscallargs.h>

#include <compat/netbsd32/netbsd32.h>
#include <compat/netbsd32/netbsd32_conv.h>
#include <compat/netbsd32/netbsd32_syscallargs.h>

#include <compat/linux/common/linux_types.h>
#include <compat/linux/common/linux_signal.h>
#include <compat/linux/common/linux_machdep.h>
#include <compat/linux/common/linux_misc.h>
#include <compat/linux/common/linux_oldolduname.h>
#include <compat/linux/common/linux_dirent.h>
#include <compat/linux/common/linux_ipc.h>
#include <compat/linux/common/linux_sem.h>
#include <compat/linux/linux_syscallargs.h>

#include <compat/linux32/common/linux32_types.h>
#include <compat/linux32/common/linux32_signal.h>
#include <compat/linux32/common/linux32_machdep.h>
#include <compat/linux32/common/linux32_sysctl.h>
#include <compat/linux32/common/linux32_socketcall.h>
#include <compat/linux32/linux32_syscallargs.h>

/*
 * Linux 'readdir' call. This code is mostly taken from the
 * SunOS getdents call (see compat/sunos/sunos_misc.c), though
 * an attempt has been made to keep it a little cleaner (failing
 * miserably, because of the cruft needed if count 1 is passed).
 *
 * The d_off field should contain the offset of the next valid entry,
 * but in Linux it has the offset of the entry itself. We emulate
 * that bug here.
 *
 * Read in BSD-style entries, convert them, and copy them out.
 *
 * Note that this doesn't handle union-mounted filesystems.
 */

int
linux32_sys_getdents(struct lwp *l, const struct linux32_sys_getdents_args *uap, register_t *retval)
{
        /* {
                syscallarg(int) fd;
                syscallarg(linux32_direntp_t) dent;
                syscallarg(unsigned int) count;
        } */
        struct dirent *bdp;
        struct vnode *vp;
        char *inp, *tbuf;               /* BSD-format */
        int len, reclen;                /* BSD-format */
        char *outp;                     /* Linux-format */
        int resid, linux32_reclen = 0;  /* Linux-format */
        struct file *fp;
        struct uio auio;
        struct iovec aiov;
        struct linux32_dirent idb;
        off_t off;              /* true file offset */
        int buflen, error, eofflag, nbytes, oldcall;
        struct vattr va;
        off_t *cookiebuf = NULL, *cookie;
        int ncookies;

        /* fd_getvnode() will use the descriptor for us */
        if ((error = fd_getvnode(SCARG(uap, fd), &fp)) != 0)
                return (error);

        if ((fp->f_flag & FREAD) == 0) {
                error = EBADF;
                goto out1;
        }

        vp = (struct vnode *)fp->f_data;
        if (vp->v_type != VDIR) {
                error = ENOTDIR;
                goto out1;
        }

        vn_lock(vp, LK_SHARED | LK_RETRY);
        error = VOP_GETATTR(vp, &va, l->l_cred);
        VOP_UNLOCK(vp);
        if (error)
                goto out1;

        nbytes = SCARG(uap, count);
        if (nbytes == 1) {      /* emulating old, broken behaviour */
                nbytes = sizeof (idb);
                buflen = uimax(va.va_blocksize, nbytes);
                oldcall = 1;
        } else {
                buflen = uimin(MAXBSIZE, nbytes);
                if (buflen < va.va_blocksize)
                        buflen = va.va_blocksize;
                oldcall = 0;
        }
        tbuf = malloc(buflen, M_TEMP, M_WAITOK);

        vn_lock(vp, LK_EXCLUSIVE | LK_RETRY);
        off = fp->f_offset;
again:
        aiov.iov_base = tbuf;
        aiov.iov_len = buflen;
        auio.uio_iov = &aiov;
        auio.uio_iovcnt = 1;
        auio.uio_rw = UIO_READ;
        auio.uio_resid = buflen;
        auio.uio_offset = off;
        UIO_SETUP_SYSSPACE(&auio);
        /*
         * First we read into the malloc'ed buffer, then
         * we massage it into user space, one record at a time.
         */
        error = VOP_READDIR(vp, &auio, fp->f_cred, &eofflag, &cookiebuf,
            &ncookies);
        if (error)
                goto out;

        inp = tbuf;
        outp = (void *)SCARG_P32(uap, dent);
        resid = nbytes;
        if ((len = buflen - auio.uio_resid) == 0)
                goto eof;

        for (cookie = cookiebuf; len > 0; len -= reclen) {
                bdp = (struct dirent *)inp;
                reclen = bdp->d_reclen;
                if (reclen & 3) {
                        error = EIO;
                        goto out;
                }
                if (bdp->d_fileno == 0) {
                        inp += reclen;  /* it is a hole; squish it out */
                        if (cookie)
                                off = *cookie++;
                        else
                                off += reclen;
                        continue;
                }
                linux32_reclen = LINUX_RECLEN(&idb, bdp->d_namlen);
                if (reclen > len || resid < linux32_reclen) {
                        /* entry too big for buffer, so just stop */
                        outp++;
                        break;
                }
                /*
                 * Massage in place to make a Linux-shaped dirent (otherwise
                 * we have to worry about touching user memory outside of
                 * the copyout() call).
                 */
                memset(&idb, 0, sizeof(idb));
                idb.d_ino = bdp->d_fileno;
                /*
                 * The old readdir() call misuses the offset and reclen fields.
                 */
                if (oldcall) {
                        idb.d_off = (linux32_off_t)linux32_reclen;
                        idb.d_reclen = (u_short)bdp->d_namlen;
                } else {
                        if (sizeof (idb.d_off) <= 4 && (off >> 32) != 0) {
                                compat_offseterr(vp, "linux32_getdents");
                                error = EINVAL;
                                goto out;
                        }
                        idb.d_off = (linux32_off_t)off;
                        idb.d_reclen = (u_short)linux32_reclen;
                }
                size_t dirl = MIN(sizeof(idb.d_name) - 1, bdp->d_namlen + 1);
                memcpy(idb.d_name, bdp->d_name, dirl);
                idb.d_name[dirl + 1] = bdp->d_type;
                if ((error = copyout((void *)&idb, outp, linux32_reclen)))
                        goto out;
                /* advance past this real entry */
                inp += reclen;
                if (cookie)
                        off = *cookie++; /* each entry points to itself */
                else
                        off += reclen;
                /* advance output past Linux-shaped entry */
                outp += linux32_reclen;
                resid -= linux32_reclen;
                if (oldcall)
                        break;
        }

        /* if we squished out the whole block, try again */
        if (outp == (void *)SCARG_P32(uap, dent)) {
                if (cookiebuf)
                        free(cookiebuf, M_TEMP);
                cookiebuf = NULL;
                goto again;
        }
        fp->f_offset = off;     /* update the vnode offset */

        if (oldcall)
                nbytes = resid + linux32_reclen;

eof:
        *retval = nbytes - resid;
out:
        VOP_UNLOCK(vp);
        if (cookiebuf)
                free(cookiebuf, M_TEMP);
        free(tbuf, M_TEMP);
out1:
        fd_putfile(SCARG(uap, fd));
        return error;
}

int
linux32_sys_getdents64(struct lwp *l, const struct linux32_sys_getdents64_args *uap, register_t *retval)
{
        /* {
                syscallarg(int) fd;
                syscallarg(linux32_dirent64p_t) dent;
                syscallarg(unsigned int) count;
        } */
        struct linux_sys_getdents64_args ua;

        NETBSD32TO64_UAP(fd);
        NETBSD32TOP_UAP(dent, struct linux_dirent64);
        NETBSD32TO64_UAP(count);

        return linux_sys_getdents64(l, &ua, retval);
}

int
linux32_sys_readdir(struct lwp *l, const struct linux32_sys_readdir_args *uap, register_t *retval)
{
        /* {
                syscallarg(int) fd;
                syscallarg(struct linux32_direntp_t) dent;
                syscallarg(unsigned int) count;
        } */
        int error;
        struct linux32_sys_getdents_args da;

        SCARG(&da, fd) = SCARG(uap, fd);
        SCARG(&da, dent) = SCARG(uap, dent);
        SCARG(&da, count) = 1;

        error = linux32_sys_getdents(l, &da, retval);
        if (error == 0 && *retval > 1)
                *retval = 1;

        return error;
}