Attachment 244334 Details for Bug 245865 – Patch for ports tree

[patch] Patch for ports tree

fix-vboxvfs-bugs-for-ports-tree.patch (text/plain), 17.24 KB, created by takahiro.kurosawa on 2023-08-25 10:43:08 UTC

(hide)

Description:

Filename:

MIME Type:

Creator: takahiro.kurosawa

Created: 2023-08-25 10:43:08 UTC

Size: 17.24 KB

patch

obsolete

>From 534194bdd3fec73bf66c4a747c79571f1f38d1e5 Mon Sep 17 00:00:00 2001
>From: KUROSAWA Takahiro <takahiro.kurosawa@gmail.com>
>Date: Fri, 25 Aug 2023 10:22:21 +0900
>Subject: [PATCH] emulators/virtualbox-ose-additions: fix vboxvfs bugs
>
>This commit fixes vboxvfs (shared foler) kernel module bugs:
>* fix open file leaks on creating files that causes file descriptor
>  leaks on host
>* fix halfway lookups
>---
> ...c_VBox_Additions_freebsd_vboxvfs_vboxvfs.h |   9 +-
> ..._Additions_freebsd_vboxvfs_vboxvfs__prov.c |  17 +-
> ...Additions_freebsd_vboxvfs_vboxvfs__vnops.c | 228 +++++++++++-------
> 3 files changed, 149 insertions(+), 105 deletions(-)
>
>diff --git a/emulators/virtualbox-ose/files/patch-src_VBox_Additions_freebsd_vboxvfs_vboxvfs.h b/emulators/virtualbox-ose/files/patch-src_VBox_Additions_freebsd_vboxvfs_vboxvfs.h
>index a4f94f92e148..6fdab60541c3 100644
>--- a/emulators/virtualbox-ose/files/patch-src_VBox_Additions_freebsd_vboxvfs_vboxvfs.h
>+++ b/emulators/virtualbox-ose/files/patch-src_VBox_Additions_freebsd_vboxvfs_vboxvfs.h
>@@ -9,7 +9,7 @@
>  /*
>   * Copyright (C) 2010-2020 Oracle Corporation
>   *
>-@@ -24,72 +19,371 @@
>+@@ -24,72 +19,372 @@
>  #define VBOXVFS_VFSNAME "vboxvfs"
>  #define VBOXVFS_VERSION 1
>  
>@@ -234,6 +234,7 @@
> +	struct vnode		*sf_vnode;	/* vnode if active */
> +	sfp_file_t		*sf_file;	/* non NULL if open */
> +	struct vboxfs_node	*sf_parent;	/* parent sfnode of this one */
>++	uint32_t		sf_opencnt;	/* sf_file reference counter */
> +	uint16_t		sf_children;	/* number of children sfnodes */
> +	uint8_t			sf_type;	/* VDIR or VREG */
> +	uint8_t			sf_vpstate;	/* XXX: ADD COMMENT */
>@@ -242,7 +243,7 @@
> +	uint64_t		sf_stat_time;	/* last-modified time of sf_stat */
> +	sffs_dirents_t		*sf_dir_list;	/* list of entries for this directory */
> +
>-+	/* interlock to protect sf_vpstate */
>++	/* interlock to protect sf_vpstate, sf_file and sf_opencnt */
> +	struct mtx		sf_interlock;
> +};
> +
>@@ -365,7 +366,7 @@
> +extern int sfprov_get_fsinfo(sfp_mount_t *, sffs_fsinfo_t *);
> +
> +extern int sfprov_create(sfp_mount_t *, char *path, mode_t mode,
>-+    sfp_file_t **fp, sffs_stat_t *stat);
>++    sffs_stat_t *stat);
> +extern int sfprov_open(sfp_mount_t *, char *path, sfp_file_t **fp);
> +extern int sfprov_close(sfp_file_t *fp);
> +extern int sfprov_read(sfp_file_t *, char * buffer, uint64_t offset,
>@@ -395,7 +396,7 @@
> +extern int sfprov_trunc(sfp_mount_t *, char *);
> +extern int sfprov_remove(sfp_mount_t *, char *path, u_int is_link);
> +extern int sfprov_mkdir(sfp_mount_t *, char *path, mode_t mode,
>-+    sfp_file_t **fp, sffs_stat_t *stat);
>++    sffs_stat_t *stat);
> +extern int sfprov_rmdir(sfp_mount_t *, char *path);
> +extern int sfprov_rename(sfp_mount_t *, char *from, char *to, u_int is_dir);
> +
>diff --git a/emulators/virtualbox-ose/files/patch-src_VBox_Additions_freebsd_vboxvfs_vboxvfs__prov.c b/emulators/virtualbox-ose/files/patch-src_VBox_Additions_freebsd_vboxvfs_vboxvfs__prov.c
>index bf094e6a854a..af2521f61379 100644
>--- a/emulators/virtualbox-ose/files/patch-src_VBox_Additions_freebsd_vboxvfs_vboxvfs__prov.c
>+++ b/emulators/virtualbox-ose/files/patch-src_VBox_Additions_freebsd_vboxvfs_vboxvfs__prov.c
>@@ -1,6 +1,6 @@
>---- src/VBox/Additions/freebsd/vboxvfs/vboxvfs_prov.c.orig	2020-06-26 09:59:35 UTC
>+--- src/VBox/Additions/freebsd/vboxvfs/vboxvfs_prov.c.orig	2023-08-24 23:34:12 UTC
> +++ src/VBox/Additions/freebsd/vboxvfs/vboxvfs_prov.c
>-@@ -0,0 +1,1020 @@
>+@@ -0,0 +1,1011 @@
> +/*
> + * Copyright (C) 2008-2016 Oracle Corporation
> + *
>@@ -308,7 +308,6 @@
> +	sfp_mount_t *mnt,
> +	char *path,
> +	mode_t mode,
>-+	sfp_file_t **fp,
> +	sffs_stat_t *stat)
> +{
> +	int rc;
>@@ -333,10 +332,7 @@
> +			return (EEXIST);
> +		return (ENOENT);
> +	}
>-+	newfp = malloc(sizeof(sfp_file_t), M_VBOXVFS, M_WAITOK | M_ZERO);
>-+	newfp->handle = parms.Handle;
>-+	newfp->map = mnt->map;
>-+	*fp = newfp;
>++	(void)VbglR0SfClose(&vbox_client, &mnt->map, parms.Handle);
> +	sfprov_stat_from_info(stat, &parms.Info);
> +	return (0);
> +}
>@@ -713,14 +709,12 @@
> +	sfp_mount_t *mnt,
> +	char *path,
> +	mode_t mode,
>-+	sfp_file_t **fp,
> +	sffs_stat_t *stat)
> +{
> +	int rc;
> +	SHFLCREATEPARMS parms;
> +	SHFLSTRING *str;
> +	int size;
>-+	sfp_file_t *newfp;
> +
> +	str = sfprov_string(path, &size);
> +	parms.Handle = SHFL_HANDLE_NIL;
>@@ -738,10 +732,7 @@
> +			return (EEXIST);
> +		return (ENOENT);
> +	}
>-+	newfp = malloc(sizeof(sfp_file_t), M_VBOXVFS, M_WAITOK | M_ZERO);
>-+	newfp->handle = parms.Handle;
>-+	newfp->map = mnt->map;
>-+	*fp = newfp;
>++	(void)VbglR0SfClose(&vbox_client, &mnt->map, parms.Handle);
> +	sfprov_stat_from_info(stat, &parms.Info);
> +	return (0);
> +}
>diff --git a/emulators/virtualbox-ose/files/patch-src_VBox_Additions_freebsd_vboxvfs_vboxvfs__vnops.c b/emulators/virtualbox-ose/files/patch-src_VBox_Additions_freebsd_vboxvfs_vboxvfs__vnops.c
>index b2f56a2bbd28..a7dedb49f979 100644
>--- a/emulators/virtualbox-ose/files/patch-src_VBox_Additions_freebsd_vboxvfs_vboxvfs__vnops.c
>+++ b/emulators/virtualbox-ose/files/patch-src_VBox_Additions_freebsd_vboxvfs_vboxvfs__vnops.c
>@@ -1,6 +1,6 @@
>---- src/VBox/Additions/freebsd/vboxvfs/vboxvfs_vnops.c.orig	2022-07-19 20:51:58 UTC
>+--- src/VBox/Additions/freebsd/vboxvfs/vboxvfs_vnops.c.orig	2023-07-12 15:59:35 UTC
> +++ src/VBox/Additions/freebsd/vboxvfs/vboxvfs_vnops.c
>-@@ -14,228 +14,1364 @@
>+@@ -14,228 +14,1416 @@
>   * VirtualBox OSE distribution. VirtualBox OSE is distributed in the
>   * hope that it will be useful, but WITHOUT ANY WARRANTY of any kind.
>   */
>@@ -83,7 +83,8 @@
> +static vop_fsync_t	vboxfs_fsync;
> +static vop_remove_t	vboxfs_remove;
> +static vop_link_t	vboxfs_link;
>-+static vop_cachedlookup_t	vboxfs_lookup;
>++static vop_lookup_t	vboxfs_lookup;
>++static vop_cachedlookup_t	vboxfs_cachedlookup;
> +static vop_rename_t	vboxfs_rename;
> +static vop_mkdir_t	vboxfs_mkdir;
> +static vop_rmdir_t	vboxfs_rmdir;
>@@ -140,8 +141,8 @@
> +	.vop_inactive	= vboxfs_inactive,
> +	.vop_ioctl	= vboxfs_ioctl,
> +	.vop_link	= vboxfs_link,
>-+	.vop_lookup	= vfs_cache_lookup,
>-+	.vop_cachedlookup	= vboxfs_lookup,
>++	.vop_lookup	= vboxfs_lookup,
>++	.vop_cachedlookup	= vboxfs_cachedlookup,
> +	.vop_mkdir	= vboxfs_mkdir,
> +	.vop_mknod	= VOP_EOPNOTSUPP,
> +	.vop_open	= vboxfs_open,
>@@ -462,13 +463,15 @@
>  {
> -    return 0;
> +	char *p;
>++	size_t dstsz;
> +
> +	if (len <= 2 && tail[0] == '.' && (len == 1 || tail[1] == '.'))
> +		panic("construct path for %s", tail);
>-+	p = malloc(strlen(node->sf_path) + 1 + len + 1, M_VBOXVFS, M_WAITOK);
>++	dstsz = strlen(node->sf_path) + 1 + len + 1;
>++	p = malloc(dstsz, M_VBOXVFS, M_WAITOK);
> +	strcpy(p, node->sf_path);
> +	strcat(p, "/");
>-+	strcat(p, tail);
>++	strlcat(p, tail, dstsz);
> +	return (p);
>  }
>  
>@@ -531,38 +534,85 @@
>  
> -static int vboxvfs_symlink(struct vop_symlink_args *ap)
> +static int
>-+vboxfs_open(struct vop_open_args *ap)
>++vboxfs_get_sfp_file(struct vboxfs_node *np)
>  {
> -    return EOPNOTSUPP;
>-+	struct vboxfs_node *np;
> +	sfp_file_t *fp;
> +	int error;
> +
>++	fp = NULL;
>++	VBOXFS_NODE_LOCK(np);
>++	for (;;) {
>++		if (np->sf_file != NULL) {
>++			if (fp != NULL)
>++				(void) sfprov_close(fp);
>++			np->sf_opencnt++;
>++			fp = np->sf_file;
>++			break;
>++		} else if (fp != NULL) {
>++			np->sf_file = fp;
>++			KASSERT(np->sf_opencnt == 0,
>++				("np %p opencnt (%d) must be zero.",
>++				 np, np->sf_opencnt));
>++			np->sf_opencnt = 1;
>++			break;
>++		}
>++		VBOXFS_NODE_UNLOCK(np);
>++		error = sfprov_open(np->vboxfsmp->sf_handle, np->sf_path, &fp);
>++		if (error != 0)
>++			return (error);
>++		VBOXFS_NODE_LOCK(np);
>++	}
>++	VBOXFS_NODE_UNLOCK(np);
>++
>++	return (0);
>+ }
>+ 
>+-static int vboxvfs_mknod(struct vop_mknod_args *ap)
>++static void
>++vboxfs_put_sfp_file(struct vboxfs_node *np)
>+ {
>+-    return EOPNOTSUPP;
>++	VBOXFS_NODE_LOCK(np);
>++	np->sf_opencnt--;
>++	if (np->sf_opencnt == 0) {
>++		(void) sfprov_close(np->sf_file);
>++		np->sf_file = NULL;
>++	}
>++	VBOXFS_NODE_UNLOCK(np);
>+ }
>+ 
>+-static int vboxvfs_mkdir(struct vop_mkdir_args *ap)
>++static int
>++vboxfs_open(struct vop_open_args *ap)
>+ {
>+-    return 0;
>++	struct vboxfs_node *np;
>++	int error;
>++
> +	MPASS(VOP_ISLOCKED(vp));
> +
> +	np = VP_TO_VBOXFS_NODE(ap->a_vp);
>-+	error = sfprov_open(np->vboxfsmp->sf_handle, np->sf_path, &fp);
>++	error = vboxfs_get_sfp_file(np);
> +	if (error != 0)
> +		goto out;
> +
>-+	np->sf_file = fp;
> +	vnode_create_vobject(ap->a_vp, 0, ap->a_td);
>-+
> +out:
> +	MPASS(VOP_ISLOCKED(vp));
> +
> +	return (error);
>  }
>  
>--static int vboxvfs_mknod(struct vop_mknod_args *ap)
>+-static int vboxvfs_rmdir(struct vop_rmdir_args *ap)
> +static void
> +vfsnode_invalidate_stat_cache(struct vboxfs_node *np)
>  {
>--    return EOPNOTSUPP;
>+-    return 0;
> +	np->sf_stat_time = 0;
>  }
>  
>--static int vboxvfs_mkdir(struct vop_mkdir_args *ap)
>+-static int vboxvfs_readdir(struct vop_readdir_args *ap)
> +static int
> +vboxfs_close(struct vop_close_args *ap)
>  {
>@@ -585,15 +635,12 @@
> +
> +	vfsnode_invalidate_stat_cache(np);
> +
>-+	if (np->sf_file != NULL && vp->v_usecount <= 1) {
>-+		(void) sfprov_close(np->sf_file);
>-+		np->sf_file = NULL;
>-+	}
>++	vboxfs_put_sfp_file(np);
> +
> +	return (0);
>  }
>  
>--static int vboxvfs_rmdir(struct vop_rmdir_args *ap)
>+-static int vboxvfs_fsync(struct vop_fsync_args *ap)
> +static int
> +vboxfs_getattr(struct vop_getattr_args *ap)
>  {
>@@ -674,7 +721,7 @@
> +	return (error);
>  }
>  
>--static int vboxvfs_readdir(struct vop_readdir_args *ap)
>+-static int vboxvfs_print (struct vop_print_args *ap)
> +static int
> +vboxfs_setattr(struct vop_setattr_args *ap)
>  {
>@@ -738,7 +785,7 @@
> +	return (error);
>  }
>  
>--static int vboxvfs_fsync(struct vop_fsync_args *ap)
>+-static int vboxvfs_pathconf (struct vop_pathconf_args *ap)
> +#define blkoff(vboxfsmp, loc)	((loc) & (vboxfsmp)->bmask)
> +
> +static int
>@@ -776,6 +823,13 @@
> +	if (tmpbuf == 0)
> +		return (ENOMEM);
> +
>++	/*
>++	 * XXX VOP_READ() is called without VOP_OPEN() on exec case.
>++	 * We need to ensure the file is opened here.
>++	 */
>++	error = vboxfs_get_sfp_file(np);
>++	if (error != 0)	/* Maybe removed on the host. */
>++		return (EIO);
> +	do {
> +		offset = uio->uio_offset;
> +		done = bytes = min(PAGE_SIZE, uio->uio_resid);
>@@ -784,6 +838,7 @@
> +		if (error == 0 && done > 0)
> +			error = uiomove(tmpbuf, done, uio);
> +	} while (error == 0 && uio->uio_resid > 0 && done > 0);
>++	vboxfs_put_sfp_file(np);
> +
> +	contigfree(tmpbuf, PAGE_SIZE, M_DEVBUF);
> +
>@@ -794,7 +849,7 @@
> +	return (error);
>  }
>  
>--static int vboxvfs_print (struct vop_print_args *ap)
>+-static int vboxvfs_strategy (struct vop_strategy_args *ap)
> +static int
> +vboxfs_write(struct vop_write_args *ap)
>  {
>@@ -855,11 +910,11 @@
> +	return (error);
>  }
>  
>--static int vboxvfs_pathconf (struct vop_pathconf_args *ap)
>+-static int vboxvfs_ioctl(struct vop_ioctl_args *ap)
> +static int
> +vboxfs_create(struct vop_create_args *ap)
>  {
>--    return 0;
>+-    return ENOTTY;
> +	struct vnode *dvp = ap->a_dvp;
> +	struct vnode **vpp = ap->a_vpp;
> +	struct componentname *cnp = ap->a_cnp;
>@@ -867,7 +922,6 @@
> +	sffs_stat_t	stat;
> +	char	*fullpath = NULL;
> +	struct vboxfs_node *dir = VP_TO_VBOXFS_NODE(dvp);
>-+	sfp_file_t *fp;
> +	int error;
> +	struct 	vboxfs_mnt *vboxfsmp = dir->vboxfsmp;
> +
>@@ -875,7 +929,7 @@
> +
> +	fullpath = sfnode_construct_path(dir, cnp->cn_nameptr, cnp->cn_namelen);
> +	error = sfprov_create(dir->vboxfsmp->sf_handle, fullpath, vap->va_mode,
>-+	    &fp, &stat);
>++	    &stat);
> +
> +	if (error)
> +		goto out;
>@@ -895,7 +949,7 @@
> +	return (error);
>  }
>  
>--static int vboxvfs_strategy (struct vop_strategy_args *ap)
>+-static int vboxvfs_getextattr(struct vop_getextattr_args *ap)
> +static int
> +vboxfs_remove(struct vop_remove_args *ap)
>  {
>@@ -914,22 +968,6 @@
> +	np = VP_TO_VBOXFS_NODE(vp);
> +	dir = VP_TO_VBOXFS_NODE(vp);
> +
>-+	/*
>-+	 * If anything else is using this vnode, then fail the remove.
>-+	 * Why?  Windows hosts can't sfprov_remove() a file that is open,
>-+	 * so we have to sfprov_close() it first.
>-+	 * There is no errno for this - since it's not a problem on UNIX,
>-+	 * but ETXTBSY is the closest.
>-+	 */
>-+	if (np->sf_file != NULL) {
>-+		if (vp->v_usecount > 1) {
>-+			error = ETXTBSY;
>-+			goto out;
>-+		}
>-+		sfprov_close(np->sf_file);
>-+		np->sf_file = NULL;
>-+	}
>-+
> +	error = sfprov_remove(np->vboxfsmp->sf_handle, np->sf_path,
> +	    np->sf_type == VLNK);
> +
>@@ -945,11 +983,11 @@
> +	return (error);
>  }
>  
>--static int vboxvfs_ioctl(struct vop_ioctl_args *ap)
>+-static int vboxvfs_advlock(struct vop_advlock_args *ap)
> +static int
> +vboxfs_rename(struct vop_rename_args *ap)
>  {
>--    return ENOTTY;
>+-    return 0;
> +	struct vnode *fvp;
> +	struct vnode *fdvp;
> +	struct vnode *tvp;
>@@ -989,7 +1027,7 @@
> +	return (ret);
>  }
>  
>--static int vboxvfs_getextattr(struct vop_getextattr_args *ap)
>+-static int vboxvfs_lookup(struct vop_lookup_args *ap)
> +static int
> +vboxfs_link(struct vop_link_args *ap)
>  {
>@@ -997,7 +1035,7 @@
> +	return (EOPNOTSUPP);
>  }
>  
>--static int vboxvfs_advlock(struct vop_advlock_args *ap)
>+-static int vboxvfs_inactive(struct vop_inactive_args *ap)
> +static int
> +vboxfs_symlink(struct vop_symlink_args *ap)
>  {
>@@ -1032,7 +1070,7 @@
> +	return (error);
>  }
>  
>--static int vboxvfs_lookup(struct vop_lookup_args *ap)
>+-static int vboxvfs_reclaim(struct vop_reclaim_args *ap)
> +static int
> +vboxfs_mkdir(struct vop_mkdir_args *ap)
>  {
>@@ -1044,7 +1082,6 @@
> +	sffs_stat_t	stat;
> +	char	*fullpath = NULL;
> +	struct vboxfs_node *dir = VP_TO_VBOXFS_NODE(dvp);
>-+	sfp_file_t *fp;
> +	int error;
> +	struct 	vboxfs_mnt *vboxfsmp = dir->vboxfsmp;
> +
>@@ -1052,7 +1089,7 @@
> +
> +	fullpath = sfnode_construct_path(dir, cnp->cn_nameptr, cnp->cn_namelen);
> +	error = sfprov_mkdir(dir->vboxfsmp->sf_handle, fullpath, vap->va_mode,
>-+	    &fp, &stat);
>++	    &stat);
> +
> +	if (error)
> +		goto out;
>@@ -1069,7 +1106,7 @@
> +	return (error);
>  }
>  
>--static int vboxvfs_inactive(struct vop_inactive_args *ap)
>+-static int vboxvfs_getpages(struct vop_getpages_args *ap)
> +static int
> +vboxfs_rmdir(struct vop_rmdir_args *ap)
>  {
>@@ -1088,22 +1125,6 @@
> +	np = VP_TO_VBOXFS_NODE(vp);
> +	dir = VP_TO_VBOXFS_NODE(vp);
> +
>-+	/*
>-+	 * If anything else is using this vnode, then fail the remove.
>-+	 * Why?  Windows hosts can't sfprov_remove() a file that is open,
>-+	 * so we have to sfprov_close() it first.
>-+	 * There is no errno for this - since it's not a problem on UNIX,
>-+	 * but ETXTBSY is the closest.
>-+	 */
>-+	if (np->sf_file != NULL) {
>-+		if (vp->v_usecount > 1) {
>-+			error = ETXTBSY;
>-+			goto out;
>-+		}
>-+		sfprov_close(np->sf_file);
>-+		np->sf_file = NULL;
>-+	}
>-+
> +	error = sfprov_rmdir(np->vboxfsmp->sf_handle, np->sf_path);
> +
> +#if 0
>@@ -1118,7 +1139,7 @@
> +	return (error);
>  }
>  
>--static int vboxvfs_reclaim(struct vop_reclaim_args *ap)
>+-static int vboxvfs_putpages(struct vop_putpages_args *ap)
> +static int
> +vboxfs_readdir(struct vop_readdir_args *ap)
>  {
>@@ -1243,11 +1264,9 @@
> +	return (error);
>  }
>  
>--static int vboxvfs_getpages(struct vop_getpages_args *ap)
> +static int
> +vboxfs_readlink(struct vop_readlink_args *v)
>- {
>--    return 0;
>++{
> +	struct vnode *vp = v->a_vp;
> +	struct uio *uio = v->a_uio;
> +
>@@ -1275,13 +1294,11 @@
> +	if (tmpbuf)
> +		contigfree(tmpbuf, MAXPATHLEN, M_DEVBUF);
> +	return (error);
>- }
>- 
>--static int vboxvfs_putpages(struct vop_putpages_args *ap)
>++}
>++
> +static int
> +vboxfs_fsync(struct vop_fsync_args *ap)
>- {
>--    return 0;
>++{
> +	struct vnode *vp;
> +	struct vboxfs_node *np;
> +	int ret;
>@@ -1292,8 +1309,8 @@
> +		return (0);
> +	ret = sfprov_fsync(np->sf_file);
> +	return (ret);
>- }
>- 
>++}
>++
> +static int
> +vboxfs_print(struct vop_print_args *ap)
> +{
>@@ -1349,17 +1366,10 @@
> + * Lookup an entry in a directory and create a new vnode if found.
> + */
> +static int
>-+vboxfs_lookup(struct vop_cachedlookup_args /* {
>-+		struct vnodeop_desc *a_desc;
>-+		struct vnode *a_dvp;
>-+		struct vnode **a_vpp;
>-+		struct componentname *a_cnp;
>-+	} */ *ap)
>++vboxfs_lookup1(struct vnode *dvp, struct vnode **vpp,
>++    struct componentname *cnp)
> +{
>-+	struct 	componentname *cnp = ap->a_cnp;
>-+	struct 	vnode *dvp = ap->a_dvp;		/* the directory vnode */
> +	char	*nameptr = cnp->cn_nameptr;	/* the name of the file or directory */
>-+	struct	vnode **vpp = ap->a_vpp;	/* the vnode we found or NULL */
> +	struct  vnode *tdp = NULL;
> +	struct 	vboxfs_node *node = VP_TO_VBOXFS_NODE(dvp);
> +	struct 	vboxfs_mnt *vboxfsmp = node->vboxfsmp;
>@@ -1372,6 +1382,7 @@
> +	int 	lkflags = cnp->cn_lkflags;
> +	char	*fullpath = NULL;
> +
>++	*vpp = NULLVP;
> +	error = ENOENT;
> +	if (cnp->cn_flags & ISDOTDOT) {
> +		error = vn_vget_ino_gen(dvp, vboxfs_vn_get_ino_alloc,
>@@ -1444,6 +1455,47 @@
> +}
> +
> +static int
>++vboxfs_cachedlookup(struct vop_cachedlookup_args *ap)
>++{
>++	return (vboxfs_lookup1(ap->a_dvp, ap->a_vpp, ap->a_cnp));
>++}
>++
>++static int
>++vboxfs_lookup(struct vop_lookup_args *ap)
>++{
>++	struct vnode *dvp = ap->a_dvp;
>++	struct componentname *cnp = ap->a_cnp;
>++	struct vboxfs_node *np = VP_TO_VBOXFS_NODE(dvp);
>++	struct timespec mtime;
>++	int flags = cnp->cn_flags;
>++	int error;
>++
>++	if (dvp->v_type != VDIR)
>++		return (ENOTDIR);
>++
>++	if ((flags & ISLASTCN) && (dvp->v_mount->mnt_flag & MNT_RDONLY) &&
>++	    (cnp->cn_nameiop == DELETE || cnp->cn_nameiop == RENAME))
>++		return (EROFS);
>++
>++	error = vn_dir_check_exec(dvp, cnp);
>++	if (error != 0)
>++		return (error);
>++
>++	/* Check if the directory is unmodified on the host. */
>++	mtime = np->sf_stat.sf_mtime;
>++	error = vsfnode_update_stat_cache(np);
>++	if (error == 0) {
>++		if (mtime.tv_sec == np->sf_stat.sf_mtime.tv_sec &&
>++		    mtime.tv_nsec == np->sf_stat.sf_mtime.tv_nsec)
>++			return (vfs_cache_lookup(ap));
>++	}
>++
>++	cache_purge(dvp);
>++
>++	return (vboxfs_lookup1(ap->a_dvp, ap->a_vpp, ap->a_cnp));
>++}
>++
>++static int
> +vboxfs_inactive(struct vop_inactive_args *ap)
> +{
> +   	return (0);
>-- 
>2.41.0
>

Actions: View | Diff

Attachments on bug 245865: 244334 | 244335