Whamcloud - gitweb
b=22615 fixes for regressions caused by 11063
[fs/lustre-release.git] / lustre / liblustre / super.c
index 9ecabb6..7be70b9 100644 (file)
@@ -65,8 +65,6 @@
 #include <file.h>
 #endif
 
-#undef LIST_HEAD
-
 #include "llite_lib.h"
 
 #ifndef MAY_EXEC
@@ -84,7 +82,7 @@ static int ll_permission(struct inode *inode, int mask)
 
         if (current->fsuid == st->st_uid)
                 mode >>= 6;
-        else if (in_group_p(st->st_gid))
+        else if (cfs_curproc_is_in_groups(st->st_gid))
                 mode >>= 3;
 
         if ((mode & mask & (MAY_READ|MAY_WRITE|MAY_EXEC)) == mask)
@@ -111,7 +109,7 @@ static void llu_fsop_gone(struct filesys *fs)
         int next = 0;
         ENTRY;
 
-        list_del(&sbi->ll_conn_chain);
+        cfs_list_del(&sbi->ll_conn_chain);
         cl_sb_fini(sbi);
         obd_disconnect(sbi->ll_dt_exp);
         obd_disconnect(sbi->ll_md_exp);
@@ -175,24 +173,22 @@ void llu_update_inode(struct inode *inode, struct lustre_md *md)
                 }
         }
 
-        if (body->valid & OBD_MD_FLMTIME &&
-            body->mtime > LTIME_S(st->st_mtime))
-                LTIME_S(st->st_mtime) = body->mtime;
-        if (body->valid & OBD_MD_FLATIME &&
-            body->atime > LTIME_S(st->st_atime))
-                LTIME_S(st->st_atime) = body->atime;
-
-        /* mtime is always updated with ctime, but can be set in past.
-           As write and utime(2) may happen within 1 second, and utime's
-           mtime has a priority over write's one, so take mtime from mds
-           for the same ctimes. */
-        if (body->valid & OBD_MD_FLCTIME &&
-            body->ctime >= LTIME_S(st->st_ctime)) {
-                LTIME_S(st->st_ctime) = body->ctime;
-                if (body->valid & OBD_MD_FLMTIME)
+        if (body->valid & OBD_MD_FLATIME) {
+                if (body->atime > LTIME_S(st->st_atime))
+                        LTIME_S(st->st_atime) = body->atime;
+                lli->lli_lvb.lvb_atime = body->atime;
+        }
+        if (body->valid & OBD_MD_FLMTIME) {
+                if (body->mtime > LTIME_S(st->st_mtime))
                         LTIME_S(st->st_mtime) = body->mtime;
+                lli->lli_lvb.lvb_mtime = body->mtime;
+        }
+        if (body->valid & OBD_MD_FLCTIME) {
+                if (body->ctime > LTIME_S(st->st_ctime))
+                        LTIME_S(st->st_ctime) = body->ctime;
+                lli->lli_lvb.lvb_ctime = body->ctime;
         }
-       if (S_ISREG(st->st_mode))
+        if (S_ISREG(st->st_mode))
                 st->st_blksize = min(2UL * PTLRPC_MAX_BRW_SIZE, LL_MAX_BLKSIZE);
         else
                 st->st_blksize = 4096;
@@ -468,7 +464,8 @@ static int llu_have_md_lock(struct inode *inode, __u64 lockpart)
 
 static int llu_inode_revalidate(struct inode *inode)
 {
-        struct lov_stripe_md *lsm = NULL;
+        struct llu_inode_info *lli = llu_i2info(inode);
+        struct intnl_stat *st = llu_i2stat(inode);
         ENTRY;
 
         if (!inode) {
@@ -480,20 +477,25 @@ static int llu_inode_revalidate(struct inode *inode)
                 struct lustre_md md;
                 struct ptlrpc_request *req = NULL;
                 struct llu_sb_info *sbi = llu_i2sbi(inode);
+                struct md_op_data op_data = { { 0 } };
                 unsigned long valid = OBD_MD_FLGETATTR;
                 int rc, ealen = 0;
 
                 /* Why don't we update all valid MDS fields here, if we're
                  * doing an RPC anyways?  -phil */
-                if (S_ISREG(llu_i2stat(inode)->st_mode)) {
+                if (S_ISREG(st->st_mode)) {
                         ealen = obd_size_diskmd(sbi->ll_dt_exp, NULL);
                         valid |= OBD_MD_FLEASIZE;
                 }
-                rc = md_getattr(sbi->ll_md_exp, ll_inode2fid(inode),
-                                NULL, valid, ealen, &req);
+
+                llu_prep_md_op_data(&op_data, inode, NULL, NULL, 0, ealen,
+                                    LUSTRE_OPC_ANY);
+                op_data.op_valid = valid;
+
+                rc = md_getattr(sbi->ll_md_exp, &op_data, &req);
                 if (rc) {
                         CERROR("failure %d inode %llu\n", rc,
-                               (long long)llu_i2stat(inode)->st_ino);
+                               (long long)st->st_ino);
                         RETURN(-abs(rc));
                 }
                 rc = md_get_lustre_md(sbi->ll_md_exp, req,
@@ -515,14 +517,18 @@ static int llu_inode_revalidate(struct inode *inode)
 
 
                 llu_update_inode(inode, &md);
-                if (md.lsm != NULL && llu_i2info(inode)->lli_smd != md.lsm)
+                if (md.lsm != NULL && lli->lli_smd != md.lsm)
                         obd_free_memmd(sbi->ll_dt_exp, &md.lsm);
                 ptlrpc_req_finished(req);
         }
 
-        lsm = llu_i2info(inode)->lli_smd;
-        if (!lsm)       /* object not yet allocated, don't validate size */
+        if (!lli->lli_smd) {
+                /* object not yet allocated, don't validate size */
+                st->st_atime = lli->lli_lvb.lvb_atime;
+                st->st_mtime = lli->lli_lvb.lvb_mtime;
+                st->st_ctime = lli->lli_lvb.lvb_ctime;
                 RETURN(0);
+        }
 
         /* ll_glimpse_size will prefer locally cached writes if they extend
          * the file */
@@ -648,7 +654,8 @@ static int inode_setattr(struct inode * inode, struct iattr * attr)
                 st->st_ctime = attr->ia_ctime;
         if (ia_valid & ATTR_MODE) {
                 st->st_mode = attr->ia_mode;
-                if (!in_group_p(st->st_gid) && !cfs_capable(CFS_CAP_FSETID))
+                if (!cfs_curproc_is_in_groups(st->st_gid) &&
+                    !cfs_capable(CFS_CAP_FSETID))
                         st->st_mode &= ~S_ISGID;
         }
         /* mark_inode_dirty(inode); */
@@ -766,33 +773,23 @@ int llu_setattr_raw(struct inode *inode, struct iattr *attr)
 
         /* We mark all of the fields "set" so MDS/OST does not re-set them */
         if (attr->ia_valid & ATTR_CTIME) {
-                attr->ia_ctime = CURRENT_TIME;
+                attr->ia_ctime = CFS_CURRENT_TIME;
                 attr->ia_valid |= ATTR_CTIME_SET;
         }
         if (!(ia_valid & ATTR_ATIME_SET) && (attr->ia_valid & ATTR_ATIME)) {
-                attr->ia_atime = CURRENT_TIME;
+                attr->ia_atime = CFS_CURRENT_TIME;
                 attr->ia_valid |= ATTR_ATIME_SET;
         }
         if (!(ia_valid & ATTR_MTIME_SET) && (attr->ia_valid & ATTR_MTIME)) {
-                attr->ia_mtime = CURRENT_TIME;
+                attr->ia_mtime = CFS_CURRENT_TIME;
                 attr->ia_valid |= ATTR_MTIME_SET;
         }
-        if ((attr->ia_valid & ATTR_CTIME) && !(attr->ia_valid & ATTR_MTIME)) {
-                /* To avoid stale mtime on mds, obtain it from ost and send
-                   to mds. */
-                rc = cl_glimpse_size(inode);
-                if (rc)
-                        RETURN(rc);
-
-                attr->ia_valid |= ATTR_MTIME_SET | ATTR_MTIME;
-                attr->ia_mtime = inode->i_stbuf.st_mtime;
-        }
 
         if (attr->ia_valid & (ATTR_MTIME | ATTR_CTIME))
                 CDEBUG(D_INODE, "setting mtime "CFS_TIME_T", ctime "CFS_TIME_T
                       ", now = "CFS_TIME_T"\n",
                        LTIME_S(attr->ia_mtime), LTIME_S(attr->ia_ctime),
-                       LTIME_S(CURRENT_TIME));
+                       LTIME_S(CFS_CURRENT_TIME));
 
         /* NB: ATTR_SIZE will only be set after this point if the size
          * resides on the MDS, ie, this file has no objects. */
@@ -843,13 +840,16 @@ int llu_setattr_raw(struct inode *inode, struct iattr *attr)
                 inode_setattr(inode, attr);
         }
 
-        if (ia_valid & ATTR_SIZE) {
-                rc = cl_setattr_do_truncate(inode, attr->ia_size, NULL);
-        } else if (ia_valid & (ATTR_MTIME | ATTR_MTIME_SET)) {
-                CDEBUG(D_INODE, "set mtime on OST inode %llu to %lu\n",
-                       (long long unsigned)st->st_ino, LTIME_S(attr->ia_mtime));
-                rc = cl_setattr_ost(inode, NULL);
-        }
+        if (ia_valid & ATTR_SIZE)
+                attr->ia_valid |= ATTR_SIZE;
+        if ((ia_valid & ATTR_SIZE) | 
+            ((ia_valid | ATTR_ATIME | ATTR_ATIME_SET) &&
+             LTIME_S(attr->ia_atime) < LTIME_S(attr->ia_ctime)) ||
+            ((ia_valid | ATTR_MTIME | ATTR_MTIME_SET) &&
+             LTIME_S(attr->ia_mtime) < LTIME_S(attr->ia_ctime)))
+                /* perform truncate and setting mtime/atime to past under PW
+                 * 0:EOF extent lock (new_size:EOF for truncate) */
+                rc = cl_setattr_ost(inode, attr, NULL);
         EXIT;
 out:
         if (op_data.op_ioepoch)
@@ -902,7 +902,7 @@ static int llu_iop_setattr(struct pnode *pno,
         }
 
         iattr.ia_valid |= ATTR_RAW | ATTR_CTIME;
-        iattr.ia_ctime = CURRENT_TIME;
+        iattr.ia_ctime = CFS_CURRENT_TIME;
 
         rc = llu_setattr_raw(ino, &iattr);
         liblustre_wait_idle();
@@ -946,6 +946,7 @@ static int llu_readlink_internal(struct inode *inode,
         struct llu_sb_info *sbi = llu_i2sbi(inode);
         struct mdt_body *body;
         struct intnl_stat *st = llu_i2stat(inode);
+        struct md_op_data op_data = {{ 0 }};
         int rc, symlen = st->st_size + 1;
         ENTRY;
 
@@ -958,8 +959,11 @@ static int llu_readlink_internal(struct inode *inode,
                 RETURN(0);
         }
 
-        rc = md_getattr(sbi->ll_md_exp, ll_inode2fid(inode), NULL,
-                        OBD_MD_LINKNAME, symlen, request);
+        llu_prep_md_op_data(&op_data, inode, NULL, NULL, 0, symlen,
+                            LUSTRE_OPC_ANY);
+        op_data.op_valid = OBD_MD_LINKNAME;
+
+        rc = md_getattr(sbi->ll_md_exp, &op_data, request);
         if (rc) {
                 CERROR("inode %llu: rc = %d\n", (long long)st->st_ino, rc);
                 RETURN(rc);
@@ -1211,7 +1215,7 @@ static int llu_statfs(struct llu_sb_info *sbi, struct statfs *sfs)
         /* For now we will always get up-to-date statfs values, but in the
          * future we may allow some amount of caching on the client (e.g.
          * from QOS or lprocfs updates). */
-        rc = llu_statfs_internal(sbi, &osfs, cfs_time_current_64() - HZ);
+        rc = llu_statfs_internal(sbi, &osfs, cfs_time_current_64() - CFS_HZ);
         if (rc)
                 return rc;
 
@@ -1672,7 +1676,7 @@ static int llu_lov_dir_setstripe(struct inode *ino, unsigned long arg)
         LASSERT(sizeof(lum) == sizeof(*lump));
         LASSERT(sizeof(lum.lmm_objects[0]) ==
                 sizeof(lump->lmm_objects[0]));
-        if (copy_from_user(&lum, lump, sizeof(lum)))
+        if (cfs_copy_from_user(&lum, lump, sizeof(lum)))
                 return(-EFAULT);
 
         switch (lum.lmm_magic) {
@@ -1780,7 +1784,7 @@ static int llu_lov_file_setstripe(struct inode *ino, unsigned long arg)
 
         LASSERT(sizeof(lum) == sizeof(*lump));
         LASSERT(sizeof(lum.lmm_objects[0]) == sizeof(lump->lmm_objects[0]));
-        if (copy_from_user(&lum, lump, sizeof(lum)))
+        if (cfs_copy_from_user(&lum, lump, sizeof(lum)))
                 RETURN(-EFAULT);
 
         rc = llu_lov_setstripe_ea_info(ino, flags, &lum, sizeof(lum));
@@ -1923,6 +1927,7 @@ llu_fsswop_mount(const char *source,
         char *osc = NULL, *mdc = NULL;
         int async = 1, err = -EINVAL;
         struct obd_connect_data ocd = {0,};
+        struct md_op_data op_data = {{0}};
 
         ENTRY;
 
@@ -2055,9 +2060,10 @@ llu_fsswop_mount(const char *source,
         }
         CDEBUG(D_SUPER, "rootfid "DFID"\n", PFID(&sbi->ll_root_fid));
 
+        op_data.op_fid1 = sbi->ll_root_fid;
+        op_data.op_valid = OBD_MD_FLGETATTR | OBD_MD_FLBLOCKS;
         /* fetch attr of root inode */
-        err = md_getattr(sbi->ll_md_exp, &sbi->ll_root_fid, NULL,
-                         OBD_MD_FLGETATTR | OBD_MD_FLBLOCKS, 0, &request);
+        err = md_getattr(sbi->ll_md_exp, &op_data, &request);
         if (err) {
                 CERROR("md_getattr failed for root: rc = %d\n", err);
                 GOTO(out_lock_cn_cb, err);