* afs_getattr
* afs_VAttrToAS
* afs_setattr
+ * afs_CreateAttr
+ * afs_DestroyAttr
*
*/
#include <afsconfig.h>
#include "afs/param.h"
-RCSID
- ("$Header$");
#include "afs/sysincludes.h" /* Standard vendor system headers */
#include "afsincludes.h" /* Afs-based standard headers */
extern struct vfs *afs_globalVFS;
#endif
+
/* copy out attributes from cache entry */
int
-afs_CopyOutAttrs(register struct vcache *avc, register struct vattr *attrs)
+afs_CopyOutAttrs(struct vcache *avc, struct vattr *attrs)
{
- register struct volume *tvp;
- register struct cell *tcell;
+ struct volume *tvp;
+ struct cell *tcell;
+#if defined(AFS_FBSD_ENV) || defined(AFS_DFBSD_ENV)
+ struct vnode *vp = AFSTOV(avc);
+#endif
int fakedir = 0;
AFS_STATCNT(afs_CopyOutAttrs);
- if (afs_fakestat_enable && avc->mvstat == 1)
+ if (afs_fakestat_enable && avc->mvstat == AFS_MVSTAT_MTPT)
fakedir = 1;
-#if defined(AFS_MACH_ENV )
- attrs->va_mode =
- fakedir ? VDIR | 0755 : vType(avc) | (avc->m.Mode & ~VFMT);
-#else /* AFS_MACH_ENV */
attrs->va_type = fakedir ? VDIR : vType(avc);
-#if defined(AFS_SGI_ENV) || defined(AFS_AIX32_ENV) || defined(AFS_SUN5_ENV)
- attrs->va_mode = fakedir ? 0755 : (mode_t) (avc->m.Mode & 0xffff);
+#if defined(AFS_SGI_ENV) || defined(AFS_AIX32_ENV) || defined(AFS_SUN5_ENV) || defined(AFS_DARWIN_ENV)
+ attrs->va_mode = fakedir ? S_IFDIR | 0755 : (mode_t) (avc->f.m.Mode & 0xffff);
#else
- attrs->va_mode = fakedir ? VDIR | 0755 : avc->m.Mode;
+ attrs->va_mode = fakedir ? VDIR | 0755 : avc->f.m.Mode;
#endif
-#endif /* AFS_MACH_ENV */
- if (avc->m.Mode & (VSUID | VSGID)) {
+ if (avc->f.m.Mode & (VSUID | VSGID)) {
/* setuid or setgid, make sure we're allowed to run them from this cell */
- tcell = afs_GetCell(avc->fid.Cell, 0);
+ tcell = afs_GetCell(avc->f.fid.Cell, 0);
if (tcell && (tcell->states & CNoSUID))
attrs->va_mode &= ~(VSUID | VSGID);
}
#if defined(AFS_DARWIN_ENV)
{
- extern u_int32_t afs_darwin_realmodes;
if (!afs_darwin_realmodes) {
/* Mac OS X uses the mode bits to determine whether a file or
* directory is accessible, and believes them, even though under
}
}
#endif /* AFS_DARWIN_ENV */
- attrs->va_uid = fakedir ? 0 : avc->m.Owner;
- attrs->va_gid = fakedir ? 0 : avc->m.Group; /* yeah! */
-#if defined(AFS_SUN56_ENV)
+ attrs->va_uid = fakedir ? 0 : avc->f.m.Owner;
+ attrs->va_gid = fakedir ? 0 : avc->f.m.Group; /* yeah! */
+#if defined(AFS_SUN5_ENV)
attrs->va_fsid = avc->v.v_vfsp->vfs_fsid.val[0];
-#else
-#ifdef AFS_SUN5_ENV
- /* XXX We try making this match the vfs's dev field XXX */
- attrs->va_fsid = 1;
-#else
-#ifdef AFS_OSF_ENV
- attrs->va_fsid = avc->v.v_mount->m_stat.f_fsid.val[0];
-#else
-#ifdef AFS_DARWIN70_ENV
- attrs->va_fsid = avc->v.v_mount->mnt_stat.f_fsid.val[0];
-#else /* ! AFS_DARWIN70_ENV */
+#elif defined(AFS_DARWIN80_ENV)
+ VATTR_RETURN(attrs, va_fsid, vfs_statfs(vnode_mount(AFSTOV(avc)))->f_fsid.val[0]);
+#elif defined(AFS_DARWIN_ENV)
+ attrs->va_fsid = avc->v->v_mount->mnt_stat.f_fsid.val[0];
+#else /* ! AFS_DARWIN_ENV */
attrs->va_fsid = 1;
-#endif /* AFS_DARWIN70_ENV */
-#endif
-#endif
-#endif /* AFS_SUN56_ENV */
- if (avc->mvstat == 2) {
- tvp = afs_GetVolume(&avc->fid, 0, READ_LOCK);
+#endif
+ if (avc->mvstat == AFS_MVSTAT_ROOT) {
+ tvp = afs_GetVolume(&avc->f.fid, 0, READ_LOCK);
/* The mount point's vnode. */
if (tvp) {
attrs->va_nodeid =
- tvp->mtpoint.Fid.Vnode + (tvp->mtpoint.Fid.Volume << 16);
- if (FidCmp(&afs_rootFid, &avc->fid) && !attrs->va_nodeid)
+ afs_calc_inum(tvp->mtpoint.Cell,
+ tvp->mtpoint.Fid.Volume,
+ tvp->mtpoint.Fid.Vnode);
+ if (FidCmp(&afs_rootFid, &avc->f.fid) && !attrs->va_nodeid)
attrs->va_nodeid = 2;
afs_PutVolume(tvp, READ_LOCK);
} else
attrs->va_nodeid = 2;
} else
- attrs->va_nodeid = avc->fid.Fid.Vnode + (avc->fid.Fid.Volume << 16);
+ attrs->va_nodeid =
+ afs_calc_inum(avc->f.fid.Cell,
+ avc->f.fid.Fid.Volume,
+ avc->f.fid.Fid.Vnode);
attrs->va_nodeid &= 0x7fffffff; /* Saber C hates negative inode #s! */
- attrs->va_nlink = fakedir ? 100 : avc->m.LinkCount;
- attrs->va_size = fakedir ? 4096 : avc->m.Length;
+ attrs->va_nlink = fakedir ? 100 : avc->f.m.LinkCount;
+ attrs->va_size = fakedir ? 4096 : avc->f.m.Length;
+#if defined(AFS_FBSD_ENV) || defined(AFS_DFBSD_ENV)
+ vnode_pager_setsize(vp, (u_long) attrs->va_size);
+#endif
attrs->va_atime.tv_sec = attrs->va_mtime.tv_sec = attrs->va_ctime.tv_sec =
- fakedir ? 0 : (int)avc->m.Date;
+ fakedir ? 0 : (int)avc->f.m.Date;
/* set microseconds to be dataversion # so that we approximate NFS-style
* use of mtime as a dataversion #. We take it mod 512K because
* microseconds *must* be less than a million, and 512K is the biggest
attrs->va_atime.tv_nsec = attrs->va_mtime.tv_nsec =
attrs->va_ctime.tv_nsec = 0;
- attrs->va_gen = hgetlo(avc->m.DataVersion);
-#elif defined(AFS_SGI_ENV) || defined(AFS_SUN5_ENV) || defined(AFS_AIX41_ENV) || defined(AFS_OBSD_ENV)
+ attrs->va_gen = hgetlo(avc->f.m.DataVersion);
+#elif defined(AFS_SGI_ENV) || defined(AFS_SUN5_ENV) || defined(AFS_AIX41_ENV) || defined(AFS_OBSD_ENV) || defined(AFS_NBSD_ENV)
attrs->va_atime.tv_nsec = attrs->va_mtime.tv_nsec =
attrs->va_ctime.tv_nsec =
- (hgetlo(avc->m.DataVersion) & 0x7ffff) * 1000;
+ (hgetlo(avc->f.m.DataVersion) & 0x7ffff) * 1000;
#else
attrs->va_atime.tv_usec = attrs->va_mtime.tv_usec =
- attrs->va_ctime.tv_usec = (hgetlo(avc->m.DataVersion) & 0x7ffff);
+ attrs->va_ctime.tv_usec = (hgetlo(avc->f.m.DataVersion) & 0x7ffff);
#endif
-#if defined(AFS_DARWIN_ENV) || defined(AFS_XBSD_ENV) || defined(AFS_OSF_ENV)
+#if defined(AFS_DARWIN_ENV) || defined(AFS_XBSD_ENV)
attrs->va_flags = 0;
#endif
#if defined(AFS_SGI_ENV) || defined(AFS_SUN5_ENV)
- attrs->va_blksize = PAGESIZE; /* XXX Was 8192 XXX */
+ attrs->va_blksize = AFS_BLKSIZE; /* XXX Was 8192 XXX */
#else
- attrs->va_blocksize = PAGESIZE; /* XXX Was 8192 XXX */
+ attrs->va_blocksize = AFS_BLKSIZE; /* XXX Was 8192 XXX */
#endif
-#ifdef AFS_DEC_ENV
- /* Have to use real device #s in Ultrix, since that's how FS type is
- * encoded. If rdev doesn't match Ultrix equivalent of statfs's rdev, then
- * "df ." doesn't work.
- */
- if (afs_globalVFS && afs_globalVFS->vfs_data)
- attrs->va_rdev = ((struct mount *)(afs_globalVFS->vfs_data))->m_dev;
- else
- attrs->va_rdev = 1; /* better than nothing */
-#else
attrs->va_rdev = 1;
-#endif
#if defined(AFS_HPUX110_ENV)
if (afs_globalVFS)
attrs->va_fstype = afs_globalVFS->vfs_mtype;
* Below return 0 (and not 1) blocks if the file is zero length. This conforms
* better with the other filesystems that do return 0.
*/
-#if !defined(AFS_OSF_ENV) && !defined(AFS_DARWIN_ENV) && !defined(AFS_XBSD_ENV)
-#if !defined(AFS_HPUX_ENV)
-#ifdef AFS_SUN5_ENV
- attrs->va_nblocks =
- (attrs->va_size ? ((attrs->va_size + 1023) >> 10) << 1 : 0);
-#elif defined(AFS_SGI_ENV)
- attrs->va_blocks = BTOBB(attrs->va_size);
-#else
- attrs->va_blocks =
- (attrs->va_size ? ((attrs->va_size + 1023) >> 10) << 1 : 0);
-#endif
-#else /* !defined(AFS_HPUX_ENV) */
- attrs->va_blocks = (attrs->va_size ? ((attrs->va_size + 1023) >> 10) : 0);
-#endif /* !defined(AFS_HPUX_ENV) */
-#else /* ! AFS_OSF_ENV && !AFS_DARWIN_ENV && !AFS_XBSD_ENV */
+#if defined(AFS_DARWIN_ENV) || defined(AFS_XBSD_ENV)
attrs->va_bytes = (attrs->va_size ? (attrs->va_size + 1023) : 1024);
#ifdef va_bytes_rsv
attrs->va_bytes_rsv = -1;
#endif
-#endif /* ! AFS_OSF_ENV && !AFS_DARWIN_ENV && !AFS_XBSD_ENV */
-
-#ifdef AFS_LINUX22_ENV
- /* And linux has its own stash as well. */
- vattr2inode(AFSTOV(avc), attrs);
-#endif
-#ifdef notdef
-#ifdef AFS_AIX51_ENV
- afs_Trace2(afs_iclSetp, CM_TRACE_STATACLX, ICL_TYPE_POINTER,
- attrs->va_acl, ICL_TYPE_INT32, attrs->va_aclsiz);
- if (attrs->va_acl && attrs->va_aclsiz >= 12) {
- struct acl *ap;
-
- ap = (struct acl *)attrs->va_acl;
- ap->acl_len = 8;
- ap->acl_mode = ACL_MODE;
- ap->acl_rsvd = 0;
- ap->u_access = 7;
- }
- /* temporary fix ? */
- attrs->va_aclsiz = 1;
-#endif
+#elif defined(AFS_HPUX_ENV)
+ attrs->va_blocks = (attrs->va_size ? ((attrs->va_size + 1023)>>10) : 0);
+#elif defined(AFS_SGI_ENV)
+ attrs->va_blocks = BTOBB(attrs->va_size);
+#elif defined(AFS_SUN5_ENV)
+ attrs->va_nblocks = (attrs->va_size ? ((attrs->va_size + 1023)>>10)<<1:0);
+#else /* everything else */
+ attrs->va_blocks = (attrs->va_size ? ((attrs->va_size + 1023)>>10)<<1:0);
#endif
return 0;
}
#if defined(AFS_SUN5_ENV) || defined(AFS_SGI_ENV)
int
afs_getattr(OSI_VC_DECL(avc), struct vattr *attrs, int flags,
- struct AFS_UCRED *acred)
+ afs_ucred_t *acred)
#else
int
-afs_getattr(OSI_VC_DECL(avc), struct vattr *attrs, struct AFS_UCRED *acred)
+afs_getattr(OSI_VC_DECL(avc), struct vattr *attrs, afs_ucred_t *acred)
#endif
{
afs_int32 code;
- struct vrequest treq;
- extern struct unixuser *afs_FindUser();
+ struct vrequest *treq = NULL;
struct unixuser *au;
int inited = 0;
OSI_VC_CONVERT(avc);
AFS_STATCNT(afs_getattr);
afs_Trace2(afs_iclSetp, CM_TRACE_GETATTR, ICL_TYPE_POINTER, avc,
- ICL_TYPE_OFFSET, ICL_HANDLE_OFFSET(avc->m.Length));
+ ICL_TYPE_OFFSET, ICL_HANDLE_OFFSET(avc->f.m.Length));
- if (afs_fakestat_enable && avc->mvstat == 1) {
+ if (afs_fakestat_enable && avc->mvstat == AFS_MVSTAT_MTPT) {
struct afs_fakestat_state fakestat;
+ struct vrequest *ureq = NULL;
- code = afs_InitReq(&treq, acred);
- if (code)
+ code = afs_CreateReq(&ureq, acred);
+ if (code) {
return code;
+ }
afs_InitFakeStat(&fakestat);
- code = afs_TryEvalFakeStat(&avc, &fakestat, &treq);
+ code = afs_TryEvalFakeStat(&avc, &fakestat, ureq);
if (code) {
afs_PutFakeStat(&fakestat);
+ afs_DestroyReq(ureq);
return code;
}
code = afs_CopyOutAttrs(avc, attrs);
afs_PutFakeStat(&fakestat);
+ afs_DestroyReq(ureq);
return code;
}
#if defined(AFS_SUN5_ENV)
return code;
}
#endif
-#if defined(AFS_DARWIN_ENV)
- if (avc->states & CUBCinit) {
+#if defined(AFS_DARWIN_ENV) && !defined(AFS_DARWIN80_ENV)
+ if (avc->f.states & CUBCinit) {
code = afs_CopyOutAttrs(avc, attrs);
return code;
}
#endif
-#if defined(AFS_SUN_ENV) || defined(AFS_ALPHA_ENV) || defined(AFS_SUN5_ENV)
- afs_BozonLock(&avc->pvnLock, avc);
-#endif
+ AFS_DISCON_LOCK();
- if (afs_shuttingdown)
+ if (afs_shuttingdown != AFS_RUNNING) {
+ AFS_DISCON_UNLOCK();
return EIO;
+ }
- if (!(avc->states & CStatd)) {
- if (!(code = afs_InitReq(&treq, acred))) {
- code = afs_VerifyVCache2(avc, &treq);
+ if (!(avc->f.states & CStatd)) {
+ if (!(code = afs_CreateReq(&treq, acred))) {
+ code = afs_VerifyVCache2(avc, treq);
inited = 1;
}
} else
code = 0;
-#if defined(AFS_SUN_ENV) || defined(AFS_ALPHA_ENV) || defined(AFS_SUN5_ENV)
+#if defined(AFS_SUN5_ENV)
if (code == 0)
osi_FlushPages(avc, acred);
- afs_BozonUnlock(&avc->pvnLock, avc);
#endif
-
if (code == 0) {
osi_FlushText(avc); /* only needed to flush text if text locked last time */
code = afs_CopyOutAttrs(avc, attrs);
if (afs_nfsexporter) {
if (!inited) {
- if ((code = afs_InitReq(&treq, acred)))
+ if ((code = afs_CreateReq(&treq, acred))) {
return code;
+ }
inited = 1;
}
if (AFS_NFSXLATORREQ(acred)) {
if ((vType(avc) != VDIR)
- && !afs_AccessOK(avc, PRSFS_READ, &treq,
+ && !afs_AccessOK(avc, PRSFS_READ, treq,
CHECK_MODE_BITS |
CMB_ALLOW_EXEC_AS_READ)) {
+ afs_DestroyReq(treq);
return EACCES;
}
- if (avc->mvstat == 2) {
-#if defined(AFS_SGI_ENV) || defined(AFS_SUN5_ENV) || defined(AFS_AIX41_ENV) || defined(AFS_DARWIN_ENV) || defined(AFS_XBSD_ENV)
- attrs->va_mtime.tv_nsec += ((++avc->xlatordv) * 1000);
-#else
- attrs->va_mtime.tv_usec += ++avc->xlatordv;
-#endif
- }
}
- if ((au = afs_FindUser(treq.uid, -1, READ_LOCK))) {
- register struct afs_exporter *exporter = au->exporter;
+ if ((au = afs_FindUser(treq->uid, -1, READ_LOCK))) {
+ struct afs_exporter *exporter = au->exporter;
if (exporter && !(afs_nfsexporter->exp_states & EXP_UNIXMODE)) {
unsigned int ubits;
* confuses getwd()... */
#ifdef AFS_LINUX22_ENV
if (avc == afs_globalVp) {
- struct inode *ip = avc->v.i_sb->s_root->d_inode;
- attrs->va_nodeid = ip->i_ino;
+ struct inode *ip = AFSTOV(avc)->i_sb->s_root->d_inode;
+ attrs->va_nodeid = ip->i_ino; /* VTOI()? */
}
#else
- if (AFSTOV(avc)->v_flag & VROOT) {
+ if (
+#if defined(AFS_DARWIN_ENV)
+ vnode_isvroot(AFSTOV(avc))
+#elif defined(AFS_NBSD50_ENV)
+ AFSTOV(avc)->v_vflag & VV_ROOT
+#else
+ AFSTOV(avc)->v_flag & VROOT
+#endif
+ ) {
struct vnode *vp = AFSTOV(avc);
+#ifdef AFS_DARWIN80_ENV
+ /* XXX vp = vnode_mount(vp)->mnt_vnodecovered; */
+ vp = 0;
+#else
vp = vp->v_vfsp->vfs_vnodecovered;
if (vp) { /* Ignore weird failures */
#ifdef AFS_SGI62_ENV
attrs->va_nodeid = ip->i_number;
#endif
}
+#endif
}
#endif /* AFS_LINUX22_ENV */
}
}
}
}
- if (!code)
+
+ AFS_DISCON_UNLOCK();
+
+ if (!code) {
+ afs_DestroyReq(treq);
return 0;
- code = afs_CheckCode(code, &treq, 14);
+ }
+ code = afs_CheckCode(code, treq, 14);
+ afs_DestroyReq(treq);
return code;
}
/* convert a Unix request into a status store request */
int
-afs_VAttrToAS(register struct vcache *avc, register struct vattr *av,
- register struct AFSStoreStatus *as)
+afs_VAttrToAS(struct vcache *avc, struct vattr *av,
+ struct AFSStoreStatus *as)
{
- register int mask;
+ int mask;
mask = 0;
+
AFS_STATCNT(afs_VAttrToAS);
-#if defined(AFS_AIX_ENV)
+#if defined(AFS_DARWIN80_ENV)
+ if (VATTR_IS_ACTIVE(av, va_mode)) {
+#elif defined(AFS_AIX_ENV)
/* Boy, was this machine dependent bogosity hard to swallow????.... */
if (av->va_mode != -1) {
-#else
-#if defined(AFS_SUN5_ENV) || defined(AFS_SGI_ENV) || defined(AFS_LINUX22_ENV)
+#elif defined(AFS_LINUX22_ENV) || defined(UKERNEL)
+ if (av->va_mask & ATTR_MODE) {
+#elif defined(AFS_SUN5_ENV) || defined(AFS_SGI_ENV)
if (av->va_mask & AT_MODE) {
+#elif defined(AFS_XBSD_ENV)
+ if (av->va_mode != (mode_t)VNOVAL) {
#else
if (av->va_mode != ((unsigned short)-1)) {
#endif
-#endif
mask |= AFS_SETMODE;
as->UnixModeBits = av->va_mode & 0xffff;
- if (avc->states & CForeign) {
+ if (avc->f.states & CForeign) {
ObtainWriteLock(&avc->lock, 127);
afs_FreeAllAxs(&(avc->Access));
ReleaseWriteLock(&avc->lock);
}
}
-#if defined(AFS_SUN5_ENV) || defined(AFS_SGI_ENV) || defined(AFS_LINUX22_ENV)
+#if defined(AFS_DARWIN80_ENV)
+ if (VATTR_IS_ACTIVE(av, va_gid)) {
+#elif defined(AFS_LINUX22_ENV) || defined(UKERNEL)
+ if (av->va_mask & ATTR_GID) {
+#elif defined(AFS_SUN5_ENV) || defined(AFS_SGI_ENV)
if (av->va_mask & AT_GID) {
-#else
-#if (defined(AFS_HPUX_ENV) || defined(AFS_SUN_ENV))
+#elif defined(AFS_HPUX_ENV)
#if defined(AFS_HPUX102_ENV)
if (av->va_gid != GID_NO_CHANGE) {
#else
if (av->va_gid != ((unsigned short)-1)) {
#endif
+#elif defined(AFS_XBSD_ENV)
+ if (av->va_gid != (gid_t)VNOVAL) {
#else
if (av->va_gid != -1) {
-#endif
-#endif /* AFS_SUN5_ENV */
+#endif /* AFS_LINUX22_ENV */
mask |= AFS_SETGROUP;
as->Group = av->va_gid;
}
-#if defined(AFS_SUN5_ENV) || defined(AFS_SGI_ENV) || defined(AFS_LINUX22_ENV)
+#if defined(AFS_DARWIN80_ENV)
+ if (VATTR_IS_ACTIVE(av, va_uid)) {
+#elif defined(AFS_LINUX22_ENV) || defined(UKERNEL)
+ if (av->va_mask & ATTR_UID) {
+#elif defined(AFS_SUN5_ENV) || defined(AFS_SGI_ENV)
if (av->va_mask & AT_UID) {
-#else
-#if (defined(AFS_HPUX_ENV) || defined(AFS_SUN_ENV))
+#elif defined(AFS_HPUX_ENV)
#if defined(AFS_HPUX102_ENV)
if (av->va_uid != UID_NO_CHANGE) {
+#elif defined(AFS_XBSD_ENV)
+ if (av->va_uid != (uid_t)VNOVAL) {
#else
if (av->va_uid != ((unsigned short)-1)) {
#endif
#else
if (av->va_uid != -1) {
-#endif
-#endif /* AFS_SUN5_ENV */
+#endif /* AFS_LINUX22_ENV */
mask |= AFS_SETOWNER;
as->Owner = av->va_uid;
}
-#if defined(AFS_SUN5_ENV) || defined(AFS_SGI_ENV) || defined(AFS_LINUX22_ENV)
+#if defined(AFS_DARWIN80_ENV)
+ if (VATTR_IS_ACTIVE(av, va_modify_time)) {
+#elif defined(AFS_LINUX22_ENV) || defined(UKERNEL)
+ if (av->va_mask & ATTR_MTIME) {
+#elif defined(AFS_SUN5_ENV) || defined(AFS_SGI_ENV)
if (av->va_mask & AT_MTIME) {
#else
if (av->va_mtime.tv_sec != -1) {
return 0;
}
-/* We don't set CDirty bit in avc->states because setattr calls WriteVCache
+/* We don't set CDirty bit in avc->f.states because setattr calls WriteVCache
* synchronously, therefore, it's not needed.
*/
#if defined(AFS_SUN5_ENV) || defined(AFS_SGI_ENV)
int
-afs_setattr(OSI_VC_DECL(avc), register struct vattr *attrs, int flags,
- struct AFS_UCRED *acred)
+afs_setattr(OSI_VC_DECL(avc), struct vattr *attrs, int flags,
+ afs_ucred_t *acred)
#else
int
-afs_setattr(OSI_VC_DECL(avc), register struct vattr *attrs,
- struct AFS_UCRED *acred)
+afs_setattr(OSI_VC_DECL(avc), struct vattr *attrs,
+ afs_ucred_t *acred)
#endif
{
- struct vrequest treq;
+ struct vrequest *treq = NULL;
struct AFSStoreStatus astat;
- register afs_int32 code;
+ afs_int32 code;
+#if defined(AFS_FBSD_ENV) || defined(AFS_DFBSD_ENV)
+ struct vnode *vp = AFSTOV(avc);
+#endif
struct afs_fakestat_state fakestate;
OSI_VC_CONVERT(avc);
afs_Trace4(afs_iclSetp, CM_TRACE_SETATTR, ICL_TYPE_POINTER, avc,
ICL_TYPE_INT32, attrs->va_mask, ICL_TYPE_OFFSET,
ICL_HANDLE_OFFSET(attrs->va_size), ICL_TYPE_OFFSET,
- ICL_HANDLE_OFFSET(avc->m.Length));
+ ICL_HANDLE_OFFSET(avc->f.m.Length));
#else
afs_Trace4(afs_iclSetp, CM_TRACE_SETATTR, ICL_TYPE_POINTER, avc,
ICL_TYPE_INT32, attrs->va_mode, ICL_TYPE_OFFSET,
ICL_HANDLE_OFFSET(attrs->va_size), ICL_TYPE_OFFSET,
- ICL_HANDLE_OFFSET(avc->m.Length));
+ ICL_HANDLE_OFFSET(avc->f.m.Length));
#endif
- if ((code = afs_InitReq(&treq, acred)))
+ if ((code = afs_CreateReq(&treq, acred)))
return code;
+ memset(&astat, 0, sizeof(astat));
+
+ AFS_DISCON_LOCK();
+
afs_InitFakeStat(&fakestate);
- code = afs_EvalFakeStat(&avc, &fakestate, &treq);
+ code = afs_EvalFakeStat(&avc, &fakestate, treq);
if (code)
goto done;
- if (avc->states & CRO) {
+ if (avc->f.states & CRO) {
code = EROFS;
goto done;
}
if (flags & ATTR_LAZY)
goto done;
#endif
-#ifndef AFS_DEC_ENV
/* if file size has changed, we need write access, otherwise (e.g.
* chmod) give it a shot; if it fails, we'll discard the status
* info.
- *
- * Note that Ultrix actually defines ftruncate of a file you have open to
- * be O.K., and does the proper access checks itself in the truncate
- * path (unlike BSD or SUNOS), so we skip this check for Ultrix.
- *
*/
-#if defined(AFS_SUN5_ENV) || defined(AFS_SGI_ENV) || defined(AFS_LINUX22_ENV)
+#if defined(AFS_DARWIN80_ENV)
+ if (VATTR_IS_ACTIVE(attrs, va_data_size)) {
+#elif defined(AFS_LINUX22_ENV) || defined(UKERNEL)
+ if (attrs->va_mask & ATTR_SIZE) {
+#elif defined(AFS_SUN5_ENV) || defined(AFS_SGI_ENV)
if (attrs->va_mask & AT_SIZE) {
-#else
-#ifdef AFS_OSF_ENV
+#elif defined(AFS_XBSD_ENV)
if (attrs->va_size != VNOVAL) {
-#else
-#ifdef AFS_AIX41_ENV
+#elif defined(AFS_AIX41_ENV)
if (attrs->va_size != -1) {
#else
if (attrs->va_size != ~0) {
#endif
-#endif
-#endif
- if (!afs_AccessOK(avc, PRSFS_WRITE, &treq, DONT_CHECK_MODE_BITS)) {
+ if (!afs_AccessOK(avc, PRSFS_WRITE, treq, DONT_CHECK_MODE_BITS)) {
code = EACCES;
goto done;
}
}
-#endif
+
+ if (AFS_IS_DISCONNECTED && !AFS_IS_DISCON_RW) {
+ code = ENETDOWN;
+ goto done;
+ }
afs_VAttrToAS(avc, attrs, &astat); /* interpret request */
code = 0;
-#if defined(AFS_SUN_ENV) || defined(AFS_ALPHA_ENV) || defined(AFS_SUN5_ENV)
- afs_BozonLock(&avc->pvnLock, avc);
-#endif
#if defined(AFS_SUN5_ENV) || defined(AFS_SGI_ENV)
if (AFS_NFSXLATORREQ(acred)) {
avc->execsOrWriters++;
#if defined(AFS_SGI_ENV)
AFS_RWLOCK((vnode_t *) avc, VRWLOCK_WRITE);
#endif
-#if defined(AFS_SUN5_ENV) || defined(AFS_SGI_ENV) || defined(AFS_LINUX22_ENV)
+#if defined(AFS_DARWIN80_ENV)
+ if (VATTR_IS_ACTIVE(attrs, va_data_size)) {
+#elif defined(AFS_LINUX22_ENV) || defined(UKERNEL)
+ if (attrs->va_mask & ATTR_SIZE) {
+#elif defined(AFS_SUN5_ENV) || defined(AFS_SGI_ENV)
if (attrs->va_mask & AT_SIZE) {
-#else
-#if defined(AFS_OSF_ENV)
+#elif defined(AFS_XBSD_ENV)
if (attrs->va_size != VNOVAL) {
-#else /* AFS_OSF_ENV */
+#elif defined(AFS_AIX41_ENV)
if (attrs->va_size != -1) {
-#endif
+#else
+ if (attrs->va_size != ~0) {
#endif
afs_size_t tsize = attrs->va_size;
ObtainWriteLock(&avc->lock, 128);
- avc->states |= CDirty;
- code = afs_TruncateAllSegments(avc, tsize, &treq, acred);
+ avc->f.states |= CDirty;
+
+ if (AFS_IS_DISCONNECTED && tsize >=avc->f.m.Length) {
+ /* If we're growing the file, and we're disconnected, we need
+ * to make the relevant dcache chunks appear ourselves. */
+ code = afs_ExtendSegments(avc, tsize, treq);
+ } else {
+ code = afs_TruncateAllSegments(avc, tsize, treq, acred);
+ }
+#ifdef AFS_LINUX26_ENV
+ /* We must update the Linux kernel's idea of file size as soon as
+ * possible, to avoid racing with delayed writepages delivered by
+ * pdflush */
+ if (code == 0)
+ i_size_write(AFSTOV(avc), tsize);
+#endif
+#if defined(AFS_FBSD_ENV) || defined(AFS_DFBSD_ENV)
+ vnode_pager_setsize(vp, (u_long) tsize);
+#endif
/* if date not explicitly set by this call, set it ourselves, since we
* changed the data */
if (!(astat.Mask & AFS_SETMODTIME)) {
astat.Mask |= AFS_SETMODTIME;
astat.ClientModTime = osi_Time();
}
+
if (code == 0) {
- if (((avc->execsOrWriters <= 0) && (avc->states & CCreating) == 0)
+ if (((avc->execsOrWriters <= 0) && (avc->f.states & CCreating) == 0)
|| (avc->execsOrWriters == 1 && AFS_NFSXLATORREQ(acred))) {
- code = afs_StoreAllSegments(avc, &treq, AFS_ASYNC);
- if (!code)
- avc->states &= ~CDirty;
+
+ /* Store files now if not disconnected. */
+ /* XXX: AFS_IS_DISCON_RW handled. */
+ if (!AFS_IS_DISCONNECTED) {
+ code = afs_StoreAllSegments(avc, treq, AFS_ASYNC);
+ if (!code)
+ avc->f.states &= ~CDirty;
+ }
}
} else
- avc->states &= ~CDirty;
+ avc->f.states &= ~CDirty;
ReleaseWriteLock(&avc->lock);
hzero(avc->flushDV);
osi_FlushText(avc); /* do this after releasing all locks */
-#ifdef AFS_DEC_ENV
- /* in case we changed the size here, propagate it to gp->g_size */
- afs_gfshack((struct gnode *)avc);
-#endif
- }
- if (code == 0) {
- ObtainSharedLock(&avc->lock, 16); /* lock entry */
- code = afs_WriteVCache(avc, &astat, &treq); /* send request */
- ReleaseSharedLock(&avc->lock); /* release lock */
- }
- if (code) {
- ObtainWriteLock(&afs_xcbhash, 487);
- afs_DequeueCallback(avc);
- avc->states &= ~CStatd;
- ReleaseWriteLock(&afs_xcbhash);
- if (avc->fid.Fid.Vnode & 1 || (vType(avc) == VDIR))
- osi_dnlc_purgedp(avc);
- /* error? erase any changes we made to vcache entry */
}
+
+ if (!AFS_IS_DISCONNECTED) {
+ if (code == 0) {
+ ObtainSharedLock(&avc->lock, 16); /* lock entry */
+ code = afs_WriteVCache(avc, &astat, treq); /* send request */
+ ReleaseSharedLock(&avc->lock); /* release lock */
+ }
+ if (code) {
+ ObtainWriteLock(&afs_xcbhash, 487);
+ afs_DequeueCallback(avc);
+ avc->f.states &= ~CStatd;
+ ReleaseWriteLock(&afs_xcbhash);
+ if (avc->f.fid.Fid.Vnode & 1 || (vType(avc) == VDIR))
+ osi_dnlc_purgedp(avc);
+ /* error? erase any changes we made to vcache entry */
+ }
+ } else {
+ ObtainSharedLock(&avc->lock, 712);
+ /* Write changes locally. */
+ code = afs_WriteVCacheDiscon(avc, &astat, attrs);
+ ReleaseSharedLock(&avc->lock);
+ } /* if (!AFS_IS_DISCONNECTED) */
+
#if defined(AFS_SUN5_ENV) || defined(AFS_SGI_ENV)
if (AFS_NFSXLATORREQ(acred)) {
avc->execsOrWriters--;
}
#endif
-#if defined(AFS_SUN_ENV) || defined(AFS_OSF_ENV) || defined(AFS_SUN5_ENV)
- afs_BozonUnlock(&avc->pvnLock, avc);
-#endif
#if defined(AFS_SGI_ENV)
AFS_RWUNLOCK((vnode_t *) avc, VRWLOCK_WRITE);
#endif
done:
afs_PutFakeStat(&fakestate);
- code = afs_CheckCode(code, &treq, 15);
+
+ AFS_DISCON_UNLOCK();
+ code = afs_CheckCode(code, treq, 15);
+ afs_DestroyReq(treq);
return code;
}
+
+/*!
+ * Allocate a vattr.
+ *
+ * \note The caller must free the allocated vattr with
+ * afs_DestroyAttr() if this function returns successfully (zero).
+ *
+ * \note The GLOCK must be held on platforms which require the GLOCK
+ * for osi_AllocSmallSpace() and osi_FreeSmallSpace().
+ *
+ * \param[out] out address of the vattr pointer
+ * \return 0 on success
+ */
+int
+afs_CreateAttr(struct vattr **out)
+{
+ struct vattr *vattr = NULL;
+
+ if (!out) {
+ return EINVAL;
+ }
+ vattr = osi_AllocSmallSpace(sizeof(struct vattr));
+ if (!vattr) {
+ return ENOMEM;
+ }
+ memset(vattr, 0, sizeof(struct vattr));
+ *out = vattr;
+ return 0;
+}
+
+/*!
+ * Deallocate a vattr.
+ *
+ * \note The GLOCK must be held on platforms which require the GLOCK
+ * for osi_FreeSmallSpace().
+ *
+ * \param[in] vattr pointer to the vattr to free; may be NULL
+ */
+void
+afs_DestroyAttr(struct vattr *vattr)
+{
+ if (vattr) {
+ osi_FreeSmallSpace(vattr);
+ }
+}
+