Add the rt linux 4.1.3-rt3 as base
[kvmfornfv.git] / kernel / fs / xfs / xfs_symlink.c
diff --git a/kernel/fs/xfs/xfs_symlink.c b/kernel/fs/xfs/xfs_symlink.c
new file mode 100644 (file)
index 0000000..3df411e
--- /dev/null
@@ -0,0 +1,608 @@
+/*
+ * Copyright (c) 2000-2006 Silicon Graphics, Inc.
+ * Copyright (c) 2012-2013 Red Hat, Inc.
+ * All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#include "xfs.h"
+#include "xfs_shared.h"
+#include "xfs_fs.h"
+#include "xfs_format.h"
+#include "xfs_log_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_bit.h"
+#include "xfs_mount.h"
+#include "xfs_da_format.h"
+#include "xfs_da_btree.h"
+#include "xfs_dir2.h"
+#include "xfs_inode.h"
+#include "xfs_ialloc.h"
+#include "xfs_alloc.h"
+#include "xfs_bmap.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_bmap_util.h"
+#include "xfs_error.h"
+#include "xfs_quota.h"
+#include "xfs_trans_space.h"
+#include "xfs_trace.h"
+#include "xfs_symlink.h"
+#include "xfs_trans.h"
+#include "xfs_log.h"
+
+/* ----- Kernel only functions below ----- */
+STATIC int
+xfs_readlink_bmap(
+       struct xfs_inode        *ip,
+       char                    *link)
+{
+       struct xfs_mount        *mp = ip->i_mount;
+       struct xfs_bmbt_irec    mval[XFS_SYMLINK_MAPS];
+       struct xfs_buf          *bp;
+       xfs_daddr_t             d;
+       char                    *cur_chunk;
+       int                     pathlen = ip->i_d.di_size;
+       int                     nmaps = XFS_SYMLINK_MAPS;
+       int                     byte_cnt;
+       int                     n;
+       int                     error = 0;
+       int                     fsblocks = 0;
+       int                     offset;
+
+       fsblocks = xfs_symlink_blocks(mp, pathlen);
+       error = xfs_bmapi_read(ip, 0, fsblocks, mval, &nmaps, 0);
+       if (error)
+               goto out;
+
+       offset = 0;
+       for (n = 0; n < nmaps; n++) {
+               d = XFS_FSB_TO_DADDR(mp, mval[n].br_startblock);
+               byte_cnt = XFS_FSB_TO_B(mp, mval[n].br_blockcount);
+
+               bp = xfs_buf_read(mp->m_ddev_targp, d, BTOBB(byte_cnt), 0,
+                                 &xfs_symlink_buf_ops);
+               if (!bp)
+                       return -ENOMEM;
+               error = bp->b_error;
+               if (error) {
+                       xfs_buf_ioerror_alert(bp, __func__);
+                       xfs_buf_relse(bp);
+
+                       /* bad CRC means corrupted metadata */
+                       if (error == -EFSBADCRC)
+                               error = -EFSCORRUPTED;
+                       goto out;
+               }
+               byte_cnt = XFS_SYMLINK_BUF_SPACE(mp, byte_cnt);
+               if (pathlen < byte_cnt)
+                       byte_cnt = pathlen;
+
+               cur_chunk = bp->b_addr;
+               if (xfs_sb_version_hascrc(&mp->m_sb)) {
+                       if (!xfs_symlink_hdr_ok(ip->i_ino, offset,
+                                                       byte_cnt, bp)) {
+                               error = -EFSCORRUPTED;
+                               xfs_alert(mp,
+"symlink header does not match required off/len/owner (0x%x/Ox%x,0x%llx)",
+                                       offset, byte_cnt, ip->i_ino);
+                               xfs_buf_relse(bp);
+                               goto out;
+
+                       }
+
+                       cur_chunk += sizeof(struct xfs_dsymlink_hdr);
+               }
+
+               memcpy(link + offset, bp->b_addr, byte_cnt);
+
+               pathlen -= byte_cnt;
+               offset += byte_cnt;
+
+               xfs_buf_relse(bp);
+       }
+       ASSERT(pathlen == 0);
+
+       link[ip->i_d.di_size] = '\0';
+       error = 0;
+
+ out:
+       return error;
+}
+
+int
+xfs_readlink(
+       struct xfs_inode *ip,
+       char            *link)
+{
+       struct xfs_mount *mp = ip->i_mount;
+       xfs_fsize_t     pathlen;
+       int             error = 0;
+
+       trace_xfs_readlink(ip);
+
+       if (XFS_FORCED_SHUTDOWN(mp))
+               return -EIO;
+
+       xfs_ilock(ip, XFS_ILOCK_SHARED);
+
+       pathlen = ip->i_d.di_size;
+       if (!pathlen)
+               goto out;
+
+       if (pathlen < 0 || pathlen > MAXPATHLEN) {
+               xfs_alert(mp, "%s: inode (%llu) bad symlink length (%lld)",
+                        __func__, (unsigned long long) ip->i_ino,
+                        (long long) pathlen);
+               ASSERT(0);
+               error = -EFSCORRUPTED;
+               goto out;
+       }
+
+
+       if (ip->i_df.if_flags & XFS_IFINLINE) {
+               memcpy(link, ip->i_df.if_u1.if_data, pathlen);
+               link[pathlen] = '\0';
+       } else {
+               error = xfs_readlink_bmap(ip, link);
+       }
+
+ out:
+       xfs_iunlock(ip, XFS_ILOCK_SHARED);
+       return error;
+}
+
+int
+xfs_symlink(
+       struct xfs_inode        *dp,
+       struct xfs_name         *link_name,
+       const char              *target_path,
+       umode_t                 mode,
+       struct xfs_inode        **ipp)
+{
+       struct xfs_mount        *mp = dp->i_mount;
+       struct xfs_trans        *tp = NULL;
+       struct xfs_inode        *ip = NULL;
+       int                     error = 0;
+       int                     pathlen;
+       struct xfs_bmap_free    free_list;
+       xfs_fsblock_t           first_block;
+       bool                    unlock_dp_on_error = false;
+       uint                    cancel_flags;
+       int                     committed;
+       xfs_fileoff_t           first_fsb;
+       xfs_filblks_t           fs_blocks;
+       int                     nmaps;
+       struct xfs_bmbt_irec    mval[XFS_SYMLINK_MAPS];
+       xfs_daddr_t             d;
+       const char              *cur_chunk;
+       int                     byte_cnt;
+       int                     n;
+       xfs_buf_t               *bp;
+       prid_t                  prid;
+       struct xfs_dquot        *udqp = NULL;
+       struct xfs_dquot        *gdqp = NULL;
+       struct xfs_dquot        *pdqp = NULL;
+       uint                    resblks;
+
+       *ipp = NULL;
+
+       trace_xfs_symlink(dp, link_name);
+
+       if (XFS_FORCED_SHUTDOWN(mp))
+               return -EIO;
+
+       /*
+        * Check component lengths of the target path name.
+        */
+       pathlen = strlen(target_path);
+       if (pathlen >= MAXPATHLEN)      /* total string too long */
+               return -ENAMETOOLONG;
+
+       udqp = gdqp = NULL;
+       prid = xfs_get_initial_prid(dp);
+
+       /*
+        * Make sure that we have allocated dquot(s) on disk.
+        */
+       error = xfs_qm_vop_dqalloc(dp,
+                       xfs_kuid_to_uid(current_fsuid()),
+                       xfs_kgid_to_gid(current_fsgid()), prid,
+                       XFS_QMOPT_QUOTALL | XFS_QMOPT_INHERIT,
+                       &udqp, &gdqp, &pdqp);
+       if (error)
+               return error;
+
+       tp = xfs_trans_alloc(mp, XFS_TRANS_SYMLINK);
+       cancel_flags = XFS_TRANS_RELEASE_LOG_RES;
+       /*
+        * The symlink will fit into the inode data fork?
+        * There can't be any attributes so we get the whole variable part.
+        */
+       if (pathlen <= XFS_LITINO(mp, dp->i_d.di_version))
+               fs_blocks = 0;
+       else
+               fs_blocks = xfs_symlink_blocks(mp, pathlen);
+       resblks = XFS_SYMLINK_SPACE_RES(mp, link_name->len, fs_blocks);
+       error = xfs_trans_reserve(tp, &M_RES(mp)->tr_symlink, resblks, 0);
+       if (error == -ENOSPC && fs_blocks == 0) {
+               resblks = 0;
+               error = xfs_trans_reserve(tp, &M_RES(mp)->tr_symlink, 0, 0);
+       }
+       if (error) {
+               cancel_flags = 0;
+               goto out_trans_cancel;
+       }
+
+       xfs_ilock(dp, XFS_ILOCK_EXCL | XFS_ILOCK_PARENT);
+       unlock_dp_on_error = true;
+
+       /*
+        * Check whether the directory allows new symlinks or not.
+        */
+       if (dp->i_d.di_flags & XFS_DIFLAG_NOSYMLINKS) {
+               error = -EPERM;
+               goto out_trans_cancel;
+       }
+
+       /*
+        * Reserve disk quota : blocks and inode.
+        */
+       error = xfs_trans_reserve_quota(tp, mp, udqp, gdqp,
+                                               pdqp, resblks, 1, 0);
+       if (error)
+               goto out_trans_cancel;
+
+       /*
+        * Check for ability to enter directory entry, if no space reserved.
+        */
+       if (!resblks) {
+               error = xfs_dir_canenter(tp, dp, link_name);
+               if (error)
+                       goto out_trans_cancel;
+       }
+       /*
+        * Initialize the bmap freelist prior to calling either
+        * bmapi or the directory create code.
+        */
+       xfs_bmap_init(&free_list, &first_block);
+
+       /*
+        * Allocate an inode for the symlink.
+        */
+       error = xfs_dir_ialloc(&tp, dp, S_IFLNK | (mode & ~S_IFMT), 1, 0,
+                              prid, resblks > 0, &ip, NULL);
+       if (error)
+               goto out_trans_cancel;
+
+       /*
+        * Now we join the directory inode to the transaction.  We do not do it
+        * earlier because xfs_dir_ialloc might commit the previous transaction
+        * (and release all the locks).  An error from here on will result in
+        * the transaction cancel unlocking dp so don't do it explicitly in the
+        * error path.
+        */
+       xfs_trans_ijoin(tp, dp, XFS_ILOCK_EXCL);
+       unlock_dp_on_error = false;
+
+       /*
+        * Also attach the dquot(s) to it, if applicable.
+        */
+       xfs_qm_vop_create_dqattach(tp, ip, udqp, gdqp, pdqp);
+
+       if (resblks)
+               resblks -= XFS_IALLOC_SPACE_RES(mp);
+       /*
+        * If the symlink will fit into the inode, write it inline.
+        */
+       if (pathlen <= XFS_IFORK_DSIZE(ip)) {
+               xfs_idata_realloc(ip, pathlen, XFS_DATA_FORK);
+               memcpy(ip->i_df.if_u1.if_data, target_path, pathlen);
+               ip->i_d.di_size = pathlen;
+
+               /*
+                * The inode was initially created in extent format.
+                */
+               ip->i_df.if_flags &= ~(XFS_IFEXTENTS | XFS_IFBROOT);
+               ip->i_df.if_flags |= XFS_IFINLINE;
+
+               ip->i_d.di_format = XFS_DINODE_FMT_LOCAL;
+               xfs_trans_log_inode(tp, ip, XFS_ILOG_DDATA | XFS_ILOG_CORE);
+
+       } else {
+               int     offset;
+
+               first_fsb = 0;
+               nmaps = XFS_SYMLINK_MAPS;
+
+               error = xfs_bmapi_write(tp, ip, first_fsb, fs_blocks,
+                                 XFS_BMAPI_METADATA, &first_block, resblks,
+                                 mval, &nmaps, &free_list);
+               if (error)
+                       goto out_bmap_cancel;
+
+               if (resblks)
+                       resblks -= fs_blocks;
+               ip->i_d.di_size = pathlen;
+               xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+
+               cur_chunk = target_path;
+               offset = 0;
+               for (n = 0; n < nmaps; n++) {
+                       char    *buf;
+
+                       d = XFS_FSB_TO_DADDR(mp, mval[n].br_startblock);
+                       byte_cnt = XFS_FSB_TO_B(mp, mval[n].br_blockcount);
+                       bp = xfs_trans_get_buf(tp, mp->m_ddev_targp, d,
+                                              BTOBB(byte_cnt), 0);
+                       if (!bp) {
+                               error = -ENOMEM;
+                               goto out_bmap_cancel;
+                       }
+                       bp->b_ops = &xfs_symlink_buf_ops;
+
+                       byte_cnt = XFS_SYMLINK_BUF_SPACE(mp, byte_cnt);
+                       byte_cnt = min(byte_cnt, pathlen);
+
+                       buf = bp->b_addr;
+                       buf += xfs_symlink_hdr_set(mp, ip->i_ino, offset,
+                                                  byte_cnt, bp);
+
+                       memcpy(buf, cur_chunk, byte_cnt);
+
+                       cur_chunk += byte_cnt;
+                       pathlen -= byte_cnt;
+                       offset += byte_cnt;
+
+                       xfs_trans_buf_set_type(tp, bp, XFS_BLFT_SYMLINK_BUF);
+                       xfs_trans_log_buf(tp, bp, 0, (buf + byte_cnt - 1) -
+                                                       (char *)bp->b_addr);
+               }
+               ASSERT(pathlen == 0);
+       }
+
+       /*
+        * Create the directory entry for the symlink.
+        */
+       error = xfs_dir_createname(tp, dp, link_name, ip->i_ino,
+                                       &first_block, &free_list, resblks);
+       if (error)
+               goto out_bmap_cancel;
+       xfs_trans_ichgtime(tp, dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
+       xfs_trans_log_inode(tp, dp, XFS_ILOG_CORE);
+
+       /*
+        * If this is a synchronous mount, make sure that the
+        * symlink transaction goes to disk before returning to
+        * the user.
+        */
+       if (mp->m_flags & (XFS_MOUNT_WSYNC|XFS_MOUNT_DIRSYNC)) {
+               xfs_trans_set_sync(tp);
+       }
+
+       error = xfs_bmap_finish(&tp, &free_list, &committed);
+       if (error)
+               goto out_bmap_cancel;
+
+       error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
+       if (error)
+               goto out_release_inode;
+
+       xfs_qm_dqrele(udqp);
+       xfs_qm_dqrele(gdqp);
+       xfs_qm_dqrele(pdqp);
+
+       *ipp = ip;
+       return 0;
+
+out_bmap_cancel:
+       xfs_bmap_cancel(&free_list);
+       cancel_flags |= XFS_TRANS_ABORT;
+out_trans_cancel:
+       xfs_trans_cancel(tp, cancel_flags);
+out_release_inode:
+       /*
+        * Wait until after the current transaction is aborted to finish the
+        * setup of the inode and release the inode.  This prevents recursive
+        * transactions and deadlocks from xfs_inactive.
+        */
+       if (ip) {
+               xfs_finish_inode_setup(ip);
+               IRELE(ip);
+       }
+
+       xfs_qm_dqrele(udqp);
+       xfs_qm_dqrele(gdqp);
+       xfs_qm_dqrele(pdqp);
+
+       if (unlock_dp_on_error)
+               xfs_iunlock(dp, XFS_ILOCK_EXCL);
+       return error;
+}
+
+/*
+ * Free a symlink that has blocks associated with it.
+ */
+STATIC int
+xfs_inactive_symlink_rmt(
+       struct xfs_inode *ip)
+{
+       xfs_buf_t       *bp;
+       int             committed;
+       int             done;
+       int             error;
+       xfs_fsblock_t   first_block;
+       xfs_bmap_free_t free_list;
+       int             i;
+       xfs_mount_t     *mp;
+       xfs_bmbt_irec_t mval[XFS_SYMLINK_MAPS];
+       int             nmaps;
+       int             size;
+       xfs_trans_t     *tp;
+
+       mp = ip->i_mount;
+       ASSERT(ip->i_df.if_flags & XFS_IFEXTENTS);
+       /*
+        * We're freeing a symlink that has some
+        * blocks allocated to it.  Free the
+        * blocks here.  We know that we've got
+        * either 1 or 2 extents and that we can
+        * free them all in one bunmapi call.
+        */
+       ASSERT(ip->i_d.di_nextents > 0 && ip->i_d.di_nextents <= 2);
+
+       tp = xfs_trans_alloc(mp, XFS_TRANS_INACTIVE);
+       error = xfs_trans_reserve(tp, &M_RES(mp)->tr_itruncate, 0, 0);
+       if (error) {
+               xfs_trans_cancel(tp, 0);
+               return error;
+       }
+
+       xfs_ilock(ip, XFS_ILOCK_EXCL);
+       xfs_trans_ijoin(tp, ip, 0);
+
+       /*
+        * Lock the inode, fix the size, and join it to the transaction.
+        * Hold it so in the normal path, we still have it locked for
+        * the second transaction.  In the error paths we need it
+        * held so the cancel won't rele it, see below.
+        */
+       size = (int)ip->i_d.di_size;
+       ip->i_d.di_size = 0;
+       xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+       /*
+        * Find the block(s) so we can inval and unmap them.
+        */
+       done = 0;
+       xfs_bmap_init(&free_list, &first_block);
+       nmaps = ARRAY_SIZE(mval);
+       error = xfs_bmapi_read(ip, 0, xfs_symlink_blocks(mp, size),
+                               mval, &nmaps, 0);
+       if (error)
+               goto error_trans_cancel;
+       /*
+        * Invalidate the block(s). No validation is done.
+        */
+       for (i = 0; i < nmaps; i++) {
+               bp = xfs_trans_get_buf(tp, mp->m_ddev_targp,
+                       XFS_FSB_TO_DADDR(mp, mval[i].br_startblock),
+                       XFS_FSB_TO_BB(mp, mval[i].br_blockcount), 0);
+               if (!bp) {
+                       error = -ENOMEM;
+                       goto error_bmap_cancel;
+               }
+               xfs_trans_binval(tp, bp);
+       }
+       /*
+        * Unmap the dead block(s) to the free_list.
+        */
+       error = xfs_bunmapi(tp, ip, 0, size, XFS_BMAPI_METADATA, nmaps,
+                           &first_block, &free_list, &done);
+       if (error)
+               goto error_bmap_cancel;
+       ASSERT(done);
+       /*
+        * Commit the first transaction.  This logs the EFI and the inode.
+        */
+       error = xfs_bmap_finish(&tp, &free_list, &committed);
+       if (error)
+               goto error_bmap_cancel;
+       /*
+        * The transaction must have been committed, since there were
+        * actually extents freed by xfs_bunmapi.  See xfs_bmap_finish.
+        * The new tp has the extent freeing and EFDs.
+        */
+       ASSERT(committed);
+       /*
+        * The first xact was committed, so add the inode to the new one.
+        * Mark it dirty so it will be logged and moved forward in the log as
+        * part of every commit.
+        */
+       xfs_trans_ijoin(tp, ip, 0);
+       xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+       /*
+        * Commit the transaction containing extent freeing and EFDs.
+        */
+       error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
+       if (error) {
+               ASSERT(XFS_FORCED_SHUTDOWN(mp));
+               goto error_unlock;
+       }
+
+       /*
+        * Remove the memory for extent descriptions (just bookkeeping).
+        */
+       if (ip->i_df.if_bytes)
+               xfs_idata_realloc(ip, -ip->i_df.if_bytes, XFS_DATA_FORK);
+       ASSERT(ip->i_df.if_bytes == 0);
+
+       xfs_iunlock(ip, XFS_ILOCK_EXCL);
+       return 0;
+
+error_bmap_cancel:
+       xfs_bmap_cancel(&free_list);
+error_trans_cancel:
+       xfs_trans_cancel(tp, XFS_TRANS_RELEASE_LOG_RES | XFS_TRANS_ABORT);
+error_unlock:
+       xfs_iunlock(ip, XFS_ILOCK_EXCL);
+       return error;
+}
+
+/*
+ * xfs_inactive_symlink - free a symlink
+ */
+int
+xfs_inactive_symlink(
+       struct xfs_inode        *ip)
+{
+       struct xfs_mount        *mp = ip->i_mount;
+       int                     pathlen;
+
+       trace_xfs_inactive_symlink(ip);
+
+       if (XFS_FORCED_SHUTDOWN(mp))
+               return -EIO;
+
+       xfs_ilock(ip, XFS_ILOCK_EXCL);
+
+       /*
+        * Zero length symlinks _can_ exist.
+        */
+       pathlen = (int)ip->i_d.di_size;
+       if (!pathlen) {
+               xfs_iunlock(ip, XFS_ILOCK_EXCL);
+               return 0;
+       }
+
+       if (pathlen < 0 || pathlen > MAXPATHLEN) {
+               xfs_alert(mp, "%s: inode (0x%llx) bad symlink length (%d)",
+                        __func__, (unsigned long long)ip->i_ino, pathlen);
+               xfs_iunlock(ip, XFS_ILOCK_EXCL);
+               ASSERT(0);
+               return -EFSCORRUPTED;
+       }
+
+       if (ip->i_df.if_flags & XFS_IFINLINE) {
+               if (ip->i_df.if_bytes > 0) 
+                       xfs_idata_realloc(ip, -(ip->i_df.if_bytes),
+                                         XFS_DATA_FORK);
+               xfs_iunlock(ip, XFS_ILOCK_EXCL);
+               ASSERT(ip->i_df.if_bytes == 0);
+               return 0;
+       }
+
+       xfs_iunlock(ip, XFS_ILOCK_EXCL);
+
+       /* remove the remote symlink */
+       return xfs_inactive_symlink_rmt(ip);
+}