From f5f287738bddf38ec9ca79344f00dab675e1bbf5 Mon Sep 17 00:00:00 2001 From: Sonny Rao Date: Thu, 23 Jun 2005 16:57:56 -0500 Subject: JFS: performance patch Basically, we saw a large amount of time spent in the jfs_strfromUCS_le() function, mispredicting the branch inside the loop, so I just added some unlikely modifiers to the if statements to re-ordered the code. Again, these simple changes provided > 2 % on spec-sfs, so please consider it for inclusion. Signed-off-by: Dave Kleikamp --- fs/jfs/jfs_unicode.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/jfs/jfs_unicode.c b/fs/jfs/jfs_unicode.c index b32208aad550..f327decfb155 100644 --- a/fs/jfs/jfs_unicode.c +++ b/fs/jfs/jfs_unicode.c @@ -51,8 +51,9 @@ int jfs_strfromUCS_le(char *to, const __le16 * from, } } else { for (i = 0; (i < len) && from[i]; i++) { - if (le16_to_cpu(from[i]) & 0xff00) { - if (warn) { + if (unlikely(le16_to_cpu(from[i]) & 0xff00)) { + to[i] = '?'; + if (unlikely(warn)) { warn--; warn_again--; printk(KERN_ERR @@ -61,7 +62,7 @@ int jfs_strfromUCS_le(char *to, const __le16 * from, printk(KERN_ERR "mount with iocharset=utf8 to access\n"); } - to[i] = '?'; + } else to[i] = (char) (le16_to_cpu(from[i])); -- cgit v1.2.3 From b38a3ab3d1bb0dc3288f73903d4dc4672b5cd2d0 Mon Sep 17 00:00:00 2001 From: Dave Kleikamp Date: Mon, 27 Jun 2005 15:35:37 -0500 Subject: JFS: Code cleanup - getting rid of never-used debug code I'm finally getting around to cleaning out debug code that I've never used. There has always been code ifdef'ed out by _JFS_DEBUG_DMAP, _JFS_DEBUG_IMAP, _JFS_DEBUG_DTREE, and _JFS_DEBUG_XTREE, which I have personally never used, and I doubt that anyone has since the design stage back in OS/2. There is also a function, xtGather, that has never been used, and I don't know why it was ever there. Signed-off-by: Dave Kleikamp --- fs/jfs/jfs_dmap.c | 298 ++-------------------------------------------- fs/jfs/jfs_dtree.c | 199 ------------------------------- fs/jfs/jfs_dtree.h | 7 -- fs/jfs/jfs_imap.c | 105 ----------------- fs/jfs/jfs_xtree.c | 340 ----------------------------------------------------- fs/jfs/jfs_xtree.h | 6 - 6 files changed, 8 insertions(+), 947 deletions(-) (limited to 'fs') diff --git a/fs/jfs/jfs_dmap.c b/fs/jfs/jfs_dmap.c index cced2fed9d0f..0732f206ca60 100644 --- a/fs/jfs/jfs_dmap.c +++ b/fs/jfs/jfs_dmap.c @@ -25,36 +25,6 @@ #include "jfs_metapage.h" #include "jfs_debug.h" -/* - * Debug code for double-checking block map - */ -/* #define _JFS_DEBUG_DMAP 1 */ - -#ifdef _JFS_DEBUG_DMAP -#define DBINITMAP(size,ipbmap,results) \ - DBinitmap(size,ipbmap,results) -#define DBALLOC(dbmap,mapsize,blkno,nblocks) \ - DBAlloc(dbmap,mapsize,blkno,nblocks) -#define DBFREE(dbmap,mapsize,blkno,nblocks) \ - DBFree(dbmap,mapsize,blkno,nblocks) -#define DBALLOCCK(dbmap,mapsize,blkno,nblocks) \ - DBAllocCK(dbmap,mapsize,blkno,nblocks) -#define DBFREECK(dbmap,mapsize,blkno,nblocks) \ - DBFreeCK(dbmap,mapsize,blkno,nblocks) - -static void DBinitmap(s64, struct inode *, u32 **); -static void DBAlloc(uint *, s64, s64, s64); -static void DBFree(uint *, s64, s64, s64); -static void DBAllocCK(uint *, s64, s64, s64); -static void DBFreeCK(uint *, s64, s64, s64); -#else -#define DBINITMAP(size,ipbmap,results) -#define DBALLOC(dbmap, mapsize, blkno, nblocks) -#define DBFREE(dbmap, mapsize, blkno, nblocks) -#define DBALLOCCK(dbmap, mapsize, blkno, nblocks) -#define DBFREECK(dbmap, mapsize, blkno, nblocks) -#endif /* _JFS_DEBUG_DMAP */ - /* * SERIALIZATION of the Block Allocation Map. * @@ -242,7 +212,6 @@ int dbMount(struct inode *ipbmap) JFS_SBI(ipbmap->i_sb)->bmap = bmp; memset(bmp->db_active, 0, sizeof(bmp->db_active)); - DBINITMAP(bmp->db_mapsize, ipbmap, &bmp->db_DBmap); /* * allocate/initialize the bmap lock @@ -407,16 +376,12 @@ int dbFree(struct inode *ip, s64 blkno, s64 nblocks) */ nb = min(rem, BPERDMAP - (blkno & (BPERDMAP - 1))); - DBALLOCCK(bmp->db_DBmap, bmp->db_mapsize, blkno, nb); - /* free the blocks. */ if ((rc = dbFreeDmap(bmp, dp, blkno, nb))) { release_metapage(mp); IREAD_UNLOCK(ipbmap); return (rc); } - - DBFREE(bmp->db_DBmap, bmp->db_mapsize, blkno, nb); } /* write the last buffer. */ @@ -775,10 +740,6 @@ int dbAlloc(struct inode *ip, s64 hint, s64 nblocks, s64 * results) IWRITE_LOCK(ipbmap); rc = dbAllocAny(bmp, nblocks, l2nb, results); - if (rc == 0) { - DBALLOC(bmp->db_DBmap, bmp->db_mapsize, *results, - nblocks); - } goto write_unlock; } @@ -836,8 +797,6 @@ int dbAlloc(struct inode *ip, s64 hint, s64 nblocks, s64 * results) != -ENOSPC) { if (rc == 0) { *results = blkno; - DBALLOC(bmp->db_DBmap, bmp->db_mapsize, - *results, nblocks); mark_metapage_dirty(mp); } @@ -863,11 +822,8 @@ int dbAlloc(struct inode *ip, s64 hint, s64 nblocks, s64 * results) if ((rc = dbAllocNear(bmp, dp, blkno, (int) nblocks, l2nb, results)) != -ENOSPC) { - if (rc == 0) { - DBALLOC(bmp->db_DBmap, bmp->db_mapsize, - *results, nblocks); + if (rc == 0) mark_metapage_dirty(mp); - } release_metapage(mp); goto read_unlock; @@ -878,11 +834,8 @@ int dbAlloc(struct inode *ip, s64 hint, s64 nblocks, s64 * results) */ if ((rc = dbAllocDmapLev(bmp, dp, (int) nblocks, l2nb, results)) != -ENOSPC) { - if (rc == 0) { - DBALLOC(bmp->db_DBmap, bmp->db_mapsize, - *results, nblocks); + if (rc == 0) mark_metapage_dirty(mp); - } release_metapage(mp); goto read_unlock; @@ -896,13 +849,9 @@ int dbAlloc(struct inode *ip, s64 hint, s64 nblocks, s64 * results) * the same allocation group as the hint. */ IWRITE_LOCK(ipbmap); - if ((rc = dbAllocAG(bmp, agno, nblocks, l2nb, results)) - != -ENOSPC) { - if (rc == 0) - DBALLOC(bmp->db_DBmap, bmp->db_mapsize, - *results, nblocks); + if ((rc = dbAllocAG(bmp, agno, nblocks, l2nb, results)) != -ENOSPC) goto write_unlock; - } + IWRITE_UNLOCK(ipbmap); @@ -918,9 +867,6 @@ int dbAlloc(struct inode *ip, s64 hint, s64 nblocks, s64 * results) */ if ((rc = dbAllocAG(bmp, agno, nblocks, l2nb, results)) == -ENOSPC) rc = dbAllocAny(bmp, nblocks, l2nb, results); - if (rc == 0) { - DBALLOC(bmp->db_DBmap, bmp->db_mapsize, *results, nblocks); - } write_unlock: IWRITE_UNLOCK(ipbmap); @@ -992,10 +938,9 @@ int dbAllocExact(struct inode *ip, s64 blkno, int nblocks) IREAD_UNLOCK(ipbmap); - if (rc == 0) { - DBALLOC(bmp->db_DBmap, bmp->db_mapsize, blkno, nblocks); + if (rc == 0) mark_metapage_dirty(mp); - } + release_metapage(mp); return (rc); @@ -1144,7 +1089,6 @@ static int dbExtend(struct inode *ip, s64 blkno, s64 nblocks, s64 addnblocks) return -EIO; } - DBALLOCCK(bmp->db_DBmap, bmp->db_mapsize, blkno, nblocks); dp = (struct dmap *) mp->data; /* try to allocate the blocks immediately following the @@ -1155,11 +1099,9 @@ static int dbExtend(struct inode *ip, s64 blkno, s64 nblocks, s64 addnblocks) IREAD_UNLOCK(ipbmap); /* were we successful ? */ - if (rc == 0) { - DBALLOC(bmp->db_DBmap, bmp->db_mapsize, extblkno, - addnblocks); + if (rc == 0) write_metapage(mp); - } else + else /* we were not successful */ release_metapage(mp); @@ -3185,16 +3127,12 @@ int dbAllocBottomUp(struct inode *ip, s64 blkno, s64 nblocks) */ nb = min(rem, BPERDMAP - (blkno & (BPERDMAP - 1))); - DBFREECK(bmp->db_DBmap, bmp->db_mapsize, blkno, nb); - /* allocate the blocks. */ if ((rc = dbAllocDmapBU(bmp, dp, blkno, nb))) { release_metapage(mp); IREAD_UNLOCK(ipbmap); return (rc); } - - DBALLOC(bmp->db_DBmap, bmp->db_mapsize, blkno, nb); } /* write the last buffer. */ @@ -4041,223 +3979,3 @@ s64 dbMapFileSizeToMapSize(struct inode * ipbmap) return (nblocks); } - - -#ifdef _JFS_DEBUG_DMAP -/* - * DBinitmap() - */ -static void DBinitmap(s64 size, struct inode *ipbmap, u32 ** results) -{ - int npages; - u32 *dbmap, *d; - int n; - s64 lblkno, cur_block; - struct dmap *dp; - struct metapage *mp; - - npages = size / 32768; - npages += (size % 32768) ? 1 : 0; - - dbmap = (u32 *) xmalloc(npages * 4096, L2PSIZE, kernel_heap); - if (dbmap == NULL) - BUG(); /* Not robust since this is only unused debug code */ - - for (n = 0, d = dbmap; n < npages; n++, d += 1024) - bzero(d, 4096); - - /* Need to initialize from disk map pages - */ - for (d = dbmap, cur_block = 0; cur_block < size; - cur_block += BPERDMAP, d += LPERDMAP) { - lblkno = BLKTODMAP(cur_block, - JFS_SBI(ipbmap->i_sb)->bmap-> - db_l2nbperpage); - mp = read_metapage(ipbmap, lblkno, PSIZE, 0); - if (mp == NULL) { - jfs_error(ipbmap->i_sb, - "DBinitmap: could not read disk map page"); - continue; - } - dp = (struct dmap *) mp->data; - - for (n = 0; n < LPERDMAP; n++) - d[n] = le32_to_cpu(dp->wmap[n]); - - release_metapage(mp); - } - - *results = dbmap; -} - - -/* - * DBAlloc() - */ -void DBAlloc(uint * dbmap, s64 mapsize, s64 blkno, s64 nblocks) -{ - int word, nb, bitno; - u32 mask; - - assert(blkno > 0 && blkno < mapsize); - assert(nblocks > 0 && nblocks <= mapsize); - - assert(blkno + nblocks <= mapsize); - - dbmap += (blkno / 32); - while (nblocks > 0) { - bitno = blkno & (32 - 1); - nb = min(nblocks, 32 - bitno); - - mask = (0xffffffff << (32 - nb) >> bitno); - assert((mask & *dbmap) == 0); - *dbmap |= mask; - - dbmap++; - blkno += nb; - nblocks -= nb; - } -} - - -/* - * DBFree() - */ -static void DBFree(uint * dbmap, s64 mapsize, s64 blkno, s64 nblocks) -{ - int word, nb, bitno; - u32 mask; - - assert(blkno > 0 && blkno < mapsize); - assert(nblocks > 0 && nblocks <= mapsize); - - assert(blkno + nblocks <= mapsize); - - dbmap += (blkno / 32); - while (nblocks > 0) { - bitno = blkno & (32 - 1); - nb = min(nblocks, 32 - bitno); - - mask = (0xffffffff << (32 - nb) >> bitno); - assert((mask & *dbmap) == mask); - *dbmap &= ~mask; - - dbmap++; - blkno += nb; - nblocks -= nb; - } -} - - -/* - * DBAllocCK() - */ -static void DBAllocCK(uint * dbmap, s64 mapsize, s64 blkno, s64 nblocks) -{ - int word, nb, bitno; - u32 mask; - - assert(blkno > 0 && blkno < mapsize); - assert(nblocks > 0 && nblocks <= mapsize); - - assert(blkno + nblocks <= mapsize); - - dbmap += (blkno / 32); - while (nblocks > 0) { - bitno = blkno & (32 - 1); - nb = min(nblocks, 32 - bitno); - - mask = (0xffffffff << (32 - nb) >> bitno); - assert((mask & *dbmap) == mask); - - dbmap++; - blkno += nb; - nblocks -= nb; - } -} - - -/* - * DBFreeCK() - */ -static void DBFreeCK(uint * dbmap, s64 mapsize, s64 blkno, s64 nblocks) -{ - int word, nb, bitno; - u32 mask; - - assert(blkno > 0 && blkno < mapsize); - assert(nblocks > 0 && nblocks <= mapsize); - - assert(blkno + nblocks <= mapsize); - - dbmap += (blkno / 32); - while (nblocks > 0) { - bitno = blkno & (32 - 1); - nb = min(nblocks, 32 - bitno); - - mask = (0xffffffff << (32 - nb) >> bitno); - assert((mask & *dbmap) == 0); - - dbmap++; - blkno += nb; - nblocks -= nb; - } -} - - -/* - * dbPrtMap() - */ -static void dbPrtMap(struct bmap * bmp) -{ - printk(" mapsize: %d%d\n", bmp->db_mapsize); - printk(" nfree: %d%d\n", bmp->db_nfree); - printk(" numag: %d\n", bmp->db_numag); - printk(" agsize: %d%d\n", bmp->db_agsize); - printk(" agl2size: %d\n", bmp->db_agl2size); - printk(" agwidth: %d\n", bmp->db_agwidth); - printk(" agstart: %d\n", bmp->db_agstart); - printk(" agheigth: %d\n", bmp->db_agheigth); - printk(" aglevel: %d\n", bmp->db_aglevel); - printk(" maxlevel: %d\n", bmp->db_maxlevel); - printk(" maxag: %d\n", bmp->db_maxag); - printk(" agpref: %d\n", bmp->db_agpref); - printk(" l2nbppg: %d\n", bmp->db_l2nbperpage); -} - - -/* - * dbPrtCtl() - */ -static void dbPrtCtl(struct dmapctl * dcp) -{ - int i, j, n; - - printk(" height: %08x\n", le32_to_cpu(dcp->height)); - printk(" leafidx: %08x\n", le32_to_cpu(dcp->leafidx)); - printk(" budmin: %08x\n", dcp->budmin); - printk(" nleafs: %08x\n", le32_to_cpu(dcp->nleafs)); - printk(" l2nleafs: %08x\n", le32_to_cpu(dcp->l2nleafs)); - - printk("\n Tree:\n"); - for (i = 0; i < CTLLEAFIND; i += 8) { - n = min(8, CTLLEAFIND - i); - - for (j = 0; j < n; j++) - printf(" [%03x]: %02x", i + j, - (char) dcp->stree[i + j]); - printf("\n"); - } - - printk("\n Tree Leaves:\n"); - for (i = 0; i < LPERCTL; i += 8) { - n = min(8, LPERCTL - i); - - for (j = 0; j < n; j++) - printf(" [%03x]: %02x", - i + j, - (char) dcp->stree[i + j + CTLLEAFIND]); - printf("\n"); - } -} -#endif /* _JFS_DEBUG_DMAP */ diff --git a/fs/jfs/jfs_dtree.c b/fs/jfs/jfs_dtree.c index 8676aee3ae48..73b5fc7eda80 100644 --- a/fs/jfs/jfs_dtree.c +++ b/fs/jfs/jfs_dtree.c @@ -4554,202 +4554,3 @@ int dtModify(tid_t tid, struct inode *ip, return 0; } - -#ifdef _JFS_DEBUG_DTREE -/* - * dtDisplayTree() - * - * function: traverse forward - */ -int dtDisplayTree(struct inode *ip) -{ - int rc; - struct metapage *mp; - dtpage_t *p; - s64 bn, pbn; - int index, lastindex, v, h; - pxd_t *xd; - struct btstack btstack; - struct btframe *btsp; - struct btframe *parent; - u8 *stbl; - int psize = 256; - - printk("display B+-tree.\n"); - - /* clear stack */ - btsp = btstack.stack; - - /* - * start with root - * - * root resides in the inode - */ - bn = 0; - v = h = 0; - - /* - * first access of each page: - */ - newPage: - DT_GETPAGE(ip, bn, mp, psize, p, rc); - if (rc) - return rc; - - /* process entries forward from first index */ - index = 0; - lastindex = p->header.nextindex - 1; - - if (p->header.flag & BT_INTERNAL) { - /* - * first access of each internal page - */ - printf("internal page "); - dtDisplayPage(ip, bn, p); - - goto getChild; - } else { /* (p->header.flag & BT_LEAF) */ - - /* - * first access of each leaf page - */ - printf("leaf page "); - dtDisplayPage(ip, bn, p); - - /* - * process leaf page entries - * - for ( ; index <= lastindex; index++) - { - } - */ - - /* unpin the leaf page */ - DT_PUTPAGE(mp); - } - - /* - * go back up to the parent page - */ - getParent: - /* pop/restore parent entry for the current child page */ - if ((parent = (btsp == btstack.stack ? NULL : --btsp)) == NULL) - /* current page must have been root */ - return; - - /* - * parent page scan completed - */ - if ((index = parent->index) == (lastindex = parent->lastindex)) { - /* go back up to the parent page */ - goto getParent; - } - - /* - * parent page has entries remaining - */ - /* get back the parent page */ - bn = parent->bn; - /* v = parent->level; */ - DT_GETPAGE(ip, bn, mp, PSIZE, p, rc); - if (rc) - return rc; - - /* get next parent entry */ - index++; - - /* - * internal page: go down to child page of current entry - */ - getChild: - /* push/save current parent entry for the child page */ - btsp->bn = pbn = bn; - btsp->index = index; - btsp->lastindex = lastindex; - /* btsp->level = v; */ - /* btsp->node = h; */ - ++btsp; - - /* get current entry for the child page */ - stbl = DT_GETSTBL(p); - xd = (pxd_t *) & p->slot[stbl[index]]; - - /* - * first access of each internal entry: - */ - - /* get child page */ - bn = addressPXD(xd); - psize = lengthPXD(xd) << ip->i_ipmnt->i_l2bsize; - - printk("traverse down 0x%Lx[%d]->0x%Lx\n", pbn, index, bn); - v++; - h = index; - - /* release parent page */ - DT_PUTPAGE(mp); - - /* process the child page */ - goto newPage; -} - - -/* - * dtDisplayPage() - * - * function: display page - */ -int dtDisplayPage(struct inode *ip, s64 bn, dtpage_t * p) -{ - int rc; - struct metapage *mp; - struct ldtentry *lh; - struct idtentry *ih; - pxd_t *xd; - int i, j; - u8 *stbl; - wchar_t name[JFS_NAME_MAX + 1]; - struct component_name key = { 0, name }; - int freepage = 0; - - if (p == NULL) { - freepage = 1; - DT_GETPAGE(ip, bn, mp, PSIZE, p, rc); - if (rc) - return rc; - } - - /* display page control */ - printk("bn:0x%Lx flag:0x%08x nextindex:%d\n", - bn, p->header.flag, p->header.nextindex); - - /* display entries */ - stbl = DT_GETSTBL(p); - for (i = 0, j = 1; i < p->header.nextindex; i++, j++) { - dtGetKey(p, i, &key, JFS_SBI(ip->i_sb)->mntflag); - key.name[key.namlen] = '\0'; - if (p->header.flag & BT_LEAF) { - lh = (struct ldtentry *) & p->slot[stbl[i]]; - printf("\t[%d] %s:%d", i, key.name, - le32_to_cpu(lh->inumber)); - } else { - ih = (struct idtentry *) & p->slot[stbl[i]]; - xd = (pxd_t *) ih; - bn = addressPXD(xd); - printf("\t[%d] %s:0x%Lx", i, key.name, bn); - } - - if (j == 4) { - printf("\n"); - j = 0; - } - } - - printf("\n"); - - if (freepage) - DT_PUTPAGE(mp); - - return 0; -} -#endif /* _JFS_DEBUG_DTREE */ diff --git a/fs/jfs/jfs_dtree.h b/fs/jfs/jfs_dtree.h index 273a80130c9d..13e4fdf07724 100644 --- a/fs/jfs/jfs_dtree.h +++ b/fs/jfs/jfs_dtree.h @@ -269,11 +269,4 @@ extern int dtModify(tid_t tid, struct inode *ip, struct component_name * key, ino_t * orig_ino, ino_t new_ino, int flag); extern int jfs_readdir(struct file *filp, void *dirent, filldir_t filldir); - -#ifdef _JFS_DEBUG_DTREE -extern int dtDisplayTree(struct inode *ip); - -extern int dtDisplayPage(struct inode *ip, s64 bn, dtpage_t * p); -#endif /* _JFS_DEBUG_DTREE */ - #endif /* !_H_JFS_DTREE */ diff --git a/fs/jfs/jfs_imap.c b/fs/jfs/jfs_imap.c index 971af2977eff..4021d46da7e3 100644 --- a/fs/jfs/jfs_imap.c +++ b/fs/jfs/jfs_imap.c @@ -86,25 +86,6 @@ static int diIAGRead(struct inomap * imap, int, struct metapage **); static int copy_from_dinode(struct dinode *, struct inode *); static void copy_to_dinode(struct dinode *, struct inode *); -/* - * debug code for double-checking inode map - */ -/* #define _JFS_DEBUG_IMAP 1 */ - -#ifdef _JFS_DEBUG_IMAP -#define DBG_DIINIT(imap) DBGdiInit(imap) -#define DBG_DIALLOC(imap, ino) DBGdiAlloc(imap, ino) -#define DBG_DIFREE(imap, ino) DBGdiFree(imap, ino) - -static void *DBGdiInit(struct inomap * imap); -static void DBGdiAlloc(struct inomap * imap, ino_t ino); -static void DBGdiFree(struct inomap * imap, ino_t ino); -#else -#define DBG_DIINIT(imap) -#define DBG_DIALLOC(imap, ino) -#define DBG_DIFREE(imap, ino) -#endif /* _JFS_DEBUG_IMAP */ - /* * NAME: diMount() * @@ -188,8 +169,6 @@ int diMount(struct inode *ipimap) imap->im_ipimap = ipimap; JFS_IP(ipimap)->i_imap = imap; -// DBG_DIINIT(imap); - return (0); } @@ -1043,7 +1022,6 @@ int diFree(struct inode *ip) /* update the bitmap. */ iagp->wmap[extno] = cpu_to_le32(bitmap); - DBG_DIFREE(imap, inum); /* update the free inode counts at the iag, ag and * map level. @@ -1231,7 +1209,6 @@ int diFree(struct inode *ip) jfs_error(ip->i_sb, "diFree: the pmap does not show inode free"); } iagp->wmap[extno] = 0; - DBG_DIFREE(imap, inum); PXDlength(&iagp->inoext[extno], 0); PXDaddress(&iagp->inoext[extno], 0); @@ -1350,7 +1327,6 @@ diInitInode(struct inode *ip, int iagno, int ino, int extno, struct iag * iagp) struct jfs_inode_info *jfs_ip = JFS_IP(ip); ip->i_ino = (iagno << L2INOSPERIAG) + ino; - DBG_DIALLOC(JFS_IP(ipimap)->i_imap, ip->i_ino); jfs_ip->ixpxd = iagp->inoext[extno]; jfs_ip->agno = BLKTOAG(le64_to_cpu(iagp->agstart), sbi); jfs_ip->active_ag = -1; @@ -3185,84 +3161,3 @@ static void copy_to_dinode(struct dinode * dip, struct inode *ip) if (S_ISCHR(ip->i_mode) || S_ISBLK(ip->i_mode)) dip->di_rdev = cpu_to_le32(jfs_ip->dev); } - -#ifdef _JFS_DEBUG_IMAP -/* - * DBGdiInit() - */ -static void *DBGdiInit(struct inomap * imap) -{ - u32 *dimap; - int size; - size = 64 * 1024; - if ((dimap = (u32 *) xmalloc(size, L2PSIZE, kernel_heap)) == NULL) - assert(0); - bzero((void *) dimap, size); - imap->im_DBGdimap = dimap; -} - -/* - * DBGdiAlloc() - */ -static void DBGdiAlloc(struct inomap * imap, ino_t ino) -{ - u32 *dimap = imap->im_DBGdimap; - int w, b; - u32 m; - w = ino >> 5; - b = ino & 31; - m = 0x80000000 >> b; - assert(w < 64 * 256); - if (dimap[w] & m) { - printk("DEBUG diAlloc: duplicate alloc ino:0x%x\n", ino); - } - dimap[w] |= m; -} - -/* - * DBGdiFree() - */ -static void DBGdiFree(struct inomap * imap, ino_t ino) -{ - u32 *dimap = imap->im_DBGdimap; - int w, b; - u32 m; - w = ino >> 5; - b = ino & 31; - m = 0x80000000 >> b; - assert(w < 64 * 256); - if ((dimap[w] & m) == 0) { - printk("DEBUG diFree: duplicate free ino:0x%x\n", ino); - } - dimap[w] &= ~m; -} - -static void dump_cp(struct inomap * ipimap, char *function, int line) -{ - printk("\n* ********* *\nControl Page %s %d\n", function, line); - printk("FreeIAG %d\tNextIAG %d\n", ipimap->im_freeiag, - ipimap->im_nextiag); - printk("NumInos %d\tNumFree %d\n", - atomic_read(&ipimap->im_numinos), - atomic_read(&ipimap->im_numfree)); - printk("AG InoFree %d\tAG ExtFree %d\n", - ipimap->im_agctl[0].inofree, ipimap->im_agctl[0].extfree); - printk("AG NumInos %d\tAG NumFree %d\n", - ipimap->im_agctl[0].numinos, ipimap->im_agctl[0].numfree); -} - -static void dump_iag(struct iag * iag, char *function, int line) -{ - printk("\n* ********* *\nIAG %s %d\n", function, line); - printk("IagNum %d\tIAG Free %d\n", le32_to_cpu(iag->iagnum), - le32_to_cpu(iag->iagfree)); - printk("InoFreeFwd %d\tInoFreeBack %d\n", - le32_to_cpu(iag->inofreefwd), - le32_to_cpu(iag->inofreeback)); - printk("ExtFreeFwd %d\tExtFreeBack %d\n", - le32_to_cpu(iag->extfreefwd), - le32_to_cpu(iag->extfreeback)); - printk("NFreeInos %d\tNFreeExts %d\n", le32_to_cpu(iag->nfreeinos), - le32_to_cpu(iag->nfreeexts)); -} -#endif /* _JFS_DEBUG_IMAP */ diff --git a/fs/jfs/jfs_xtree.c b/fs/jfs/jfs_xtree.c index 31b34db4519e..a7fe2f2b969f 100644 --- a/fs/jfs/jfs_xtree.c +++ b/fs/jfs/jfs_xtree.c @@ -135,14 +135,6 @@ static int xtSearchNode(struct inode *ip, static int xtRelink(tid_t tid, struct inode *ip, xtpage_t * fp); #endif /* _STILL_TO_PORT */ -/* External references */ - -/* - * debug control - */ -/* #define _JFS_DEBUG_XTREE 1 */ - - /* * xtLookup() * @@ -4140,338 +4132,6 @@ s64 xtTruncate_pmap(tid_t tid, struct inode *ip, s64 committed_size) return 0; } - -#ifdef _JFS_DEBUG_XTREE -/* - * xtDisplayTree() - * - * function: traverse forward - */ -int xtDisplayTree(struct inode *ip) -{ - int rc = 0; - struct metapage *mp; - xtpage_t *p; - s64 bn, pbn; - int index, lastindex, v, h; - xad_t *xad; - struct btstack btstack; - struct btframe *btsp; - struct btframe *parent; - - printk("display B+-tree.\n"); - - /* clear stack */ - btsp = btstack.stack; - - /* - * start with root - * - * root resides in the inode - */ - bn = 0; - v = h = 0; - - /* - * first access of each page: - */ - getPage: - XT_GETPAGE(ip, bn, mp, PSIZE, p, rc); - if (rc) - return rc; - - /* process entries forward from first index */ - index = XTENTRYSTART; - lastindex = le16_to_cpu(p->header.nextindex) - 1; - - if (p->header.flag & BT_INTERNAL) { - /* - * first access of each internal page - */ - goto getChild; - } else { /* (p->header.flag & BT_LEAF) */ - - /* - * first access of each leaf page - */ - printf("leaf page "); - xtDisplayPage(ip, bn, p); - - /* unpin the leaf page */ - XT_PUTPAGE(mp); - } - - /* - * go back up to the parent page - */ - getParent: - /* pop/restore parent entry for the current child page */ - if ((parent = (btsp == btstack.stack ? NULL : --btsp)) == NULL) - /* current page must have been root */ - return; - - /* - * parent page scan completed - */ - if ((index = parent->index) == (lastindex = parent->lastindex)) { - /* go back up to the parent page */ - goto getParent; - } - - /* - * parent page has entries remaining - */ - /* get back the parent page */ - bn = parent->bn; - /* v = parent->level; */ - XT_GETPAGE(ip, bn, mp, PSIZE, p, rc); - if (rc) - return rc; - - /* get next parent entry */ - index++; - - /* - * internal page: go down to child page of current entry - */ - getChild: - /* push/save current parent entry for the child page */ - btsp->bn = pbn = bn; - btsp->index = index; - btsp->lastindex = lastindex; - /* btsp->level = v; */ - /* btsp->node = h; */ - ++btsp; - - /* get child page */ - xad = &p->xad[index]; - bn = addressXAD(xad); - - /* - * first access of each internal entry: - */ - /* release parent page */ - XT_PUTPAGE(mp); - - printk("traverse down 0x%lx[%d]->0x%lx\n", (ulong) pbn, index, - (ulong) bn); - v++; - h = index; - - /* process the child page */ - goto getPage; -} - - -/* - * xtDisplayPage() - * - * function: display page - */ -int xtDisplayPage(struct inode *ip, s64 bn, xtpage_t * p) -{ - int rc = 0; - xad_t *xad; - s64 xaddr, xoff; - int xlen, i, j; - - /* display page control */ - printf("bn:0x%lx flag:0x%x nextindex:%d\n", - (ulong) bn, p->header.flag, - le16_to_cpu(p->header.nextindex)); - - /* display entries */ - xad = &p->xad[XTENTRYSTART]; - for (i = XTENTRYSTART, j = 1; i < le16_to_cpu(p->header.nextindex); - i++, xad++, j++) { - xoff = offsetXAD(xad); - xaddr = addressXAD(xad); - xlen = lengthXAD(xad); - printf("\t[%d] 0x%lx:0x%lx(0x%x)", i, (ulong) xoff, - (ulong) xaddr, xlen); - - if (j == 4) { - printf("\n"); - j = 0; - } - } - - printf("\n"); -} -#endif /* _JFS_DEBUG_XTREE */ - - -#ifdef _JFS_WIP -/* - * xtGather() - * - * function: - * traverse for allocation acquiring tlock at commit time - * (vs at the time of update) logging backward top down - * - * note: - * problem - establishing that all new allocation have been - * processed both for append and random write in sparse file - * at the current entry at the current subtree root page - * - */ -int xtGather(btree_t *t) -{ - int rc = 0; - xtpage_t *p; - u64 bn; - int index; - btentry_t *e; - struct btstack btstack; - struct btsf *parent; - - /* clear stack */ - BT_CLR(&btstack); - - /* - * start with root - * - * root resides in the inode - */ - bn = 0; - XT_GETPAGE(ip, bn, mp, PSIZE, p, rc); - if (rc) - return rc; - - /* new root is NOT pointed by a new entry - if (p->header.flag & NEW) - allocate new page lock; - write a NEWPAGE log; - */ - - dopage: - /* - * first access of each page: - */ - /* process entries backward from last index */ - index = le16_to_cpu(p->header.nextindex) - 1; - - if (p->header.flag & BT_LEAF) { - /* - * first access of each leaf page - */ - /* process leaf page entries backward */ - for (; index >= XTENTRYSTART; index--) { - e = &p->xad[index]; - /* - * if newpage, log NEWPAGE. - * - if (e->flag & XAD_NEW) { - nfound =+ entry->length; - update current page lock for the entry; - newpage(entry); - * - * if moved, log move. - * - } else if (e->flag & XAD_MOVED) { - reset flag; - update current page lock for the entry; - } - */ - } - - /* unpin the leaf page */ - XT_PUTPAGE(mp); - - /* - * go back up to the parent page - */ - getParent: - /* restore parent entry for the current child page */ - if ((parent = BT_POP(&btstack)) == NULL) - /* current page must have been root */ - return 0; - - if ((index = parent->index) == XTENTRYSTART) { - /* - * parent page scan completed - */ - /* go back up to the parent page */ - goto getParent; - } else { - /* - * parent page has entries remaining - */ - /* get back the parent page */ - bn = parent->bn; - XT_GETPAGE(ip, bn, mp, PSIZE, p, rc); - if (rc) - return -EIO; - - /* first subroot page which - * covers all new allocated blocks - * itself not new/modified. - * (if modified from split of descendent, - * go down path of split page) - - if (nfound == nnew && - !(p->header.flag & (NEW | MOD))) - exit scan; - */ - - /* process parent page entries backward */ - index--; - } - } else { - /* - * first access of each internal page - */ - } - - /* - * internal page: go down to child page of current entry - */ - - /* save current parent entry for the child page */ - BT_PUSH(&btstack, bn, index); - - /* get current entry for the child page */ - e = &p->xad[index]; - - /* - * first access of each internal entry: - */ - /* - * if new entry, log btree_tnewentry. - * - if (e->flag & XAD_NEW) - update parent page lock for the entry; - */ - - /* release parent page */ - XT_PUTPAGE(mp); - - /* get child page */ - bn = e->bn; - XT_GETPAGE(ip, bn, mp, PSIZE, p, rc); - if (rc) - return rc; - - /* - * first access of each non-root page: - */ - /* - * if new, log btree_newpage. - * - if (p->header.flag & NEW) - allocate new page lock; - write a NEWPAGE log (next, prev); - */ - - /* process the child page */ - goto dopage; - - out: - return 0; -} -#endif /* _JFS_WIP */ - - #ifdef CONFIG_JFS_STATISTICS int jfs_xtstat_read(char *buffer, char **start, off_t offset, int length, int *eof, void *data) diff --git a/fs/jfs/jfs_xtree.h b/fs/jfs/jfs_xtree.h index a69784254fe7..af668a80b40f 100644 --- a/fs/jfs/jfs_xtree.h +++ b/fs/jfs/jfs_xtree.h @@ -131,10 +131,4 @@ extern int xtRelocate(tid_t tid, struct inode *ip, extern int xtAppend(tid_t tid, struct inode *ip, int xflag, s64 xoff, int maxblocks, int *xlenp, s64 * xaddrp, int flag); - -#ifdef _JFS_DEBUG_XTREE -extern int xtDisplayTree(struct inode *ip); -extern int xtDisplayPage(struct inode *ip, s64 bn, xtpage_t * p); -#endif /* _JFS_DEBUG_XTREE */ - #endif /* !_H_JFS_XTREE */ -- cgit v1.2.3 From 336d2ff7115bd4391108a16f476c739cb4a505b0 Mon Sep 17 00:00:00 2001 From: "Artem B. Bityuckiy" Date: Thu, 7 Jul 2005 16:45:32 +0100 Subject: [JFFS2] Avoid alloc/dealloc for zero sized nodes Signed-off-by: Artem B. Bityuckiy Signed-off-by: Thomas Gleixner --- fs/jffs2/readinode.c | 10 ++++------ 1 file changed, 4 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/jffs2/readinode.c b/fs/jffs2/readinode.c index 081656c1d49e..8f0be5557ae9 100644 --- a/fs/jffs2/readinode.c +++ b/fs/jffs2/readinode.c @@ -7,7 +7,7 @@ * * For licensing information, see the file 'LICENCE' in this directory. * - * $Id: readinode.c,v 1.120 2005/07/05 21:03:07 dwmw2 Exp $ + * $Id: readinode.c,v 1.124 2005/07/07 15:45:29 dedekind Exp $ * */ @@ -151,6 +151,9 @@ int jffs2_add_full_dnode_to_inode(struct jffs2_sb_info *c, struct jffs2_inode_in D1(printk(KERN_DEBUG "jffs2_add_full_dnode_to_inode(ino #%u, f %p, fn %p)\n", f->inocache->ino, f, fn)); + if (unlikely(!fn->size)) + return 0; + newfrag = jffs2_alloc_node_frag(); if (unlikely(!newfrag)) return -ENOMEM; @@ -158,11 +161,6 @@ int jffs2_add_full_dnode_to_inode(struct jffs2_sb_info *c, struct jffs2_inode_in D2(printk(KERN_DEBUG "adding node %04x-%04x @0x%08x on flash, newfrag *%p\n", fn->ofs, fn->ofs+fn->size, ref_offset(fn->raw), newfrag)); - if (unlikely(!fn->size)) { - jffs2_free_node_frag(newfrag); - return 0; - } - newfrag->ofs = fn->ofs; newfrag->size = fn->size; newfrag->node = fn; -- cgit v1.2.3 From 86ffc0d5f5f0c8a6fe70eccade3f7ae8e9630edf Mon Sep 17 00:00:00 2001 From: "Artem B. Bityuckiy" Date: Sun, 10 Jul 2005 14:13:58 +0100 Subject: [JFFS2] Remove needless variable initialization Signed-off-by: Artem B. Bityuckiy Signed-off-by: Thomas Gleixner --- fs/jffs2/readinode.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/jffs2/readinode.c b/fs/jffs2/readinode.c index 8f0be5557ae9..5b2a83599d73 100644 --- a/fs/jffs2/readinode.c +++ b/fs/jffs2/readinode.c @@ -7,7 +7,7 @@ * * For licensing information, see the file 'LICENCE' in this directory. * - * $Id: readinode.c,v 1.124 2005/07/07 15:45:29 dedekind Exp $ + * $Id: readinode.c,v 1.125 2005/07/10 13:13:55 dedekind Exp $ * */ @@ -558,7 +558,6 @@ static int jffs2_do_read_inode_internal(struct jffs2_sb_info *c, } next_tn: BUG_ON(rb->rb_left); - repl_rb = NULL; if (rb->rb_parent && rb->rb_parent->rb_left == rb) { /* We were then left-hand child of our parent. We need to move our own right-hand child into our place. */ -- cgit v1.2.3 From e4fef6618992bea48b7856ebbd236796b744964f Mon Sep 17 00:00:00 2001 From: "Artem B. Bityuckiy" Date: Sun, 10 Jul 2005 16:15:36 +0100 Subject: [JFFS2] Rename function and update comments We recently changed the method of collecting and sorting of tmp_dnode objects to use a temporary RB-tree instead of a temporary list. Rename function and update comments. Signed-off-by: Artem B. Bityuckiy Signed-off-by: Thomas Gleixner --- fs/jffs2/nodelist.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/jffs2/nodelist.c b/fs/jffs2/nodelist.c index c7bbdeec93a6..4991c348f6ec 100644 --- a/fs/jffs2/nodelist.c +++ b/fs/jffs2/nodelist.c @@ -7,7 +7,7 @@ * * For licensing information, see the file 'LICENCE' in this directory. * - * $Id: nodelist.c,v 1.97 2005/07/06 15:18:41 dwmw2 Exp $ + * $Id: nodelist.c,v 1.98 2005/07/10 15:15:32 dedekind Exp $ * */ @@ -55,11 +55,11 @@ void jffs2_add_fd_to_list(struct jffs2_sb_info *c, struct jffs2_full_dirent *new }); } -/* Put a new tmp_dnode_info into the list, keeping the list in - order of increasing version -*/ - -static void jffs2_add_tn_to_list(struct jffs2_tmp_dnode_info *tn, struct rb_root *list) +/* + * Put a new tmp_dnode_info into the temporaty RB-tree, keeping the list in + * order of increasing version. + */ +static void jffs2_add_tn_to_tree(struct jffs2_tmp_dnode_info *tn, struct rb_root *list) { struct rb_node **p = &list->rb_node; struct rb_node * parent = NULL; @@ -420,7 +420,7 @@ int jffs2_get_inode_nodes(struct jffs2_sb_info *c, struct jffs2_inode_info *f, D1(printk(KERN_DEBUG "dnode @%08x: ver %u, offset %04x, dsize %04x\n", ref_offset(ref), je32_to_cpu(node.i.version), je32_to_cpu(node.i.offset), je32_to_cpu(node.i.dsize))); - jffs2_add_tn_to_list(tn, &ret_tn); + jffs2_add_tn_to_tree(tn, &ret_tn); break; default: -- cgit v1.2.3 From b62205986a73a40e4fc174d2ef668e3895e925bc Mon Sep 17 00:00:00 2001 From: "Artem B. Bityuckiy" Date: Tue, 12 Jul 2005 17:37:12 +0100 Subject: [JFFS2] Init locks early during mount In case of a mount error locks might be uninitialized but accessed by the resulting call to jffs2_kill_sb(). Signed-off-by: Artem B. Bityuckiy Signed-off-by: Thomas Gleixner --- fs/jffs2/build.c | 9 +-------- fs/jffs2/super.c | 11 ++++++++++- 2 files changed, 11 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/jffs2/build.c b/fs/jffs2/build.c index 3dd5394921c9..97dc39796e2c 100644 --- a/fs/jffs2/build.c +++ b/fs/jffs2/build.c @@ -7,7 +7,7 @@ * * For licensing information, see the file 'LICENCE' in this directory. * - * $Id: build.c,v 1.70 2005/02/28 08:21:05 dedekind Exp $ + * $Id: build.c,v 1.71 2005/07/12 16:37:08 dedekind Exp $ * */ @@ -336,13 +336,6 @@ int jffs2_do_mount_fs(struct jffs2_sb_info *c) c->blocks[i].bad_count = 0; } - init_MUTEX(&c->alloc_sem); - init_MUTEX(&c->erase_free_sem); - init_waitqueue_head(&c->erase_wait); - init_waitqueue_head(&c->inocache_wq); - spin_lock_init(&c->erase_completion_lock); - spin_lock_init(&c->inocache_lock); - INIT_LIST_HEAD(&c->clean_list); INIT_LIST_HEAD(&c->very_dirty_list); INIT_LIST_HEAD(&c->dirty_list); diff --git a/fs/jffs2/super.c b/fs/jffs2/super.c index 2cf14cf8b35a..aaf9475cfb6a 100644 --- a/fs/jffs2/super.c +++ b/fs/jffs2/super.c @@ -7,7 +7,7 @@ * * For licensing information, see the file 'LICENCE' in this directory. * - * $Id: super.c,v 1.106 2005/05/18 11:37:25 dedekind Exp $ + * $Id: super.c,v 1.107 2005/07/12 16:37:08 dedekind Exp $ * */ @@ -140,6 +140,15 @@ static struct super_block *jffs2_get_sb_mtd(struct file_system_type *fs_type, D1(printk(KERN_DEBUG "jffs2_get_sb_mtd(): New superblock for device %d (\"%s\")\n", mtd->index, mtd->name)); + /* Initialize JFFS2 superblock locks, the further initialization will be + * done later */ + init_MUTEX(&c->alloc_sem); + init_MUTEX(&c->erase_free_sem); + init_waitqueue_head(&c->erase_wait); + init_waitqueue_head(&c->inocache_wq); + spin_lock_init(&c->erase_completion_lock); + spin_lock_init(&c->inocache_lock); + sb->s_op = &jffs2_super_operations; sb->s_flags = flags | MS_NOATIME; -- cgit v1.2.3 From 751382dd5cb2702368d281a50b55c2d6c4e8fbfc Mon Sep 17 00:00:00 2001 From: Todd Poynor Date: Tue, 12 Jul 2005 03:34:39 +0100 Subject: [JFFS2] Avoid compiler warnings when JFFS2_FS_WRITEBUFFER=n Signed-off-by: Todd Poynor Signed-off-by: Thomas Gleixner --- fs/jffs2/os-linux.h | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/jffs2/os-linux.h b/fs/jffs2/os-linux.h index 7bf72e012c94..d900c8929b09 100644 --- a/fs/jffs2/os-linux.h +++ b/fs/jffs2/os-linux.h @@ -7,7 +7,7 @@ * * For licensing information, see the file 'LICENCE' in this directory. * - * $Id: os-linux.h,v 1.57 2005/07/06 12:13:09 dwmw2 Exp $ + * $Id: os-linux.h,v 1.58 2005/07/12 02:34:35 tpoynor Exp $ * */ @@ -86,6 +86,8 @@ static inline void jffs2_init_inode_info(struct jffs2_inode_info *f) #define jffs2_dataflash(c) (0) #define jffs2_nor_ecc_flash_setup(c) (0) #define jffs2_nor_ecc_flash_cleanup(c) do {} while (0) +#define jffs2_dataflash_setup(c) (0) +#define jffs2_dataflash_cleanup(c) do {} while (0) #else /* NAND and/or ECC'd NOR support present */ -- cgit v1.2.3 From 6211502d7ee9e515e4458d0c0ebfbb70553dc7de Mon Sep 17 00:00:00 2001 From: Dave Kleikamp Date: Wed, 13 Jul 2005 09:07:53 -0500 Subject: JFS: Allow security.* xattrs to be set on symlinks All of the different xattr namespaces have different rules. user.* and ACL's are not allowed on symlinks, and since these were the first xattrs implemented, I assumed there was no need to support xattrs on symlinks. This one-line patch should fix it. Signed-off-by: Dave Kleikamp --- fs/jfs/xattr.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/jfs/xattr.c b/fs/jfs/xattr.c index ee438d429d45..fdd8f3f5a72f 100644 --- a/fs/jfs/xattr.c +++ b/fs/jfs/xattr.c @@ -781,7 +781,7 @@ static int can_set_xattr(struct inode *inode, const char *name, if (IS_RDONLY(inode)) return -EROFS; - if (IS_IMMUTABLE(inode) || IS_APPEND(inode) || S_ISLNK(inode->i_mode)) + if (IS_IMMUTABLE(inode) || IS_APPEND(inode)) return -EPERM; if(strncmp(name, XATTR_SYSTEM_PREFIX, XATTR_SYSTEM_PREFIX_LEN) == 0) -- cgit v1.2.3 From 59192ed9e7aa81b06a1803661419f7261afca1ad Mon Sep 17 00:00:00 2001 From: Ian Dall Date: Wed, 13 Jul 2005 09:15:18 -0500 Subject: JFS: Need to be root to create files with security context It turns out this is due to some inverted logic in xattr.c Signed-off-by: Dave Kleikamp --- fs/jfs/xattr.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/jfs/xattr.c b/fs/jfs/xattr.c index fdd8f3f5a72f..554ec739e49b 100644 --- a/fs/jfs/xattr.c +++ b/fs/jfs/xattr.c @@ -790,12 +790,12 @@ static int can_set_xattr(struct inode *inode, const char *name, */ return can_set_system_xattr(inode, name, value, value_len); - if(strncmp(name, XATTR_TRUSTED_PREFIX, XATTR_TRUSTED_PREFIX_LEN) != 0) + if(strncmp(name, XATTR_TRUSTED_PREFIX, XATTR_TRUSTED_PREFIX_LEN) == 0) return (capable(CAP_SYS_ADMIN) ? 0 : -EPERM); #ifdef CONFIG_JFS_SECURITY if (strncmp(name, XATTR_SECURITY_PREFIX, XATTR_SECURITY_PREFIX_LEN) - != 0) + == 0) return 0; /* Leave it to the security module */ #endif -- cgit v1.2.3 From 0399cb08c54708db231d616f106f64d920e0b723 Mon Sep 17 00:00:00 2001 From: Robert Love Date: Wed, 13 Jul 2005 12:38:18 -0400 Subject: [PATCH] inotify: move sysctl This moves the inotify sysctl knobs to "/proc/sys/fs/inotify" from "/proc/sys/fs". Also some related cleanup. Signed-off-by: Robert Love Signed-off-by: Linus Torvalds --- fs/inotify.c | 49 ++++++++++++++++++++++++++++++++++++++++++++---- include/linux/sysctl.h | 12 ++++++------ kernel/sysctl.c | 51 +++++++++++--------------------------------------- 3 files changed, 62 insertions(+), 50 deletions(-) (limited to 'fs') diff --git a/fs/inotify.c b/fs/inotify.c index e423bfe0c86f..fb4803131423 100644 --- a/fs/inotify.c +++ b/fs/inotify.c @@ -45,8 +45,8 @@ static kmem_cache_t *event_cachep; static struct vfsmount *inotify_mnt; -/* These are configurable via /proc/sys/inotify */ -int inotify_max_user_devices; +/* these are configurable via /proc/sys/fs/inotify/ */ +int inotify_max_user_instances; int inotify_max_user_watches; int inotify_max_queued_events; @@ -125,6 +125,47 @@ struct inotify_watch { u32 mask; /* event mask for this watch */ }; +#ifdef CONFIG_SYSCTL + +#include + +static int zero; + +ctl_table inotify_table[] = { + { + .ctl_name = INOTIFY_MAX_USER_INSTANCES, + .procname = "max_user_instances", + .data = &inotify_max_user_instances, + .maxlen = sizeof(int), + .mode = 0644, + .proc_handler = &proc_dointvec_minmax, + .strategy = &sysctl_intvec, + .extra1 = &zero, + }, + { + .ctl_name = INOTIFY_MAX_USER_WATCHES, + .procname = "max_user_watches", + .data = &inotify_max_user_watches, + .maxlen = sizeof(int), + .mode = 0644, + .proc_handler = &proc_dointvec_minmax, + .strategy = &sysctl_intvec, + .extra1 = &zero, + }, + { + .ctl_name = INOTIFY_MAX_QUEUED_EVENTS, + .procname = "max_queued_events", + .data = &inotify_max_queued_events, + .maxlen = sizeof(int), + .mode = 0644, + .proc_handler = &proc_dointvec_minmax, + .strategy = &sysctl_intvec, + .extra1 = &zero + }, + { .ctl_name = 0 } +}; +#endif /* CONFIG_SYSCTL */ + static inline void get_inotify_dev(struct inotify_device *dev) { atomic_inc(&dev->count); @@ -842,7 +883,7 @@ asmlinkage long sys_inotify_init(void) user = get_uid(current->user); - if (unlikely(atomic_read(&user->inotify_devs) >= inotify_max_user_devices)) { + if (unlikely(atomic_read(&user->inotify_devs) >= inotify_max_user_instances)) { ret = -EMFILE; goto out_err; } @@ -979,7 +1020,7 @@ static int __init inotify_init(void) inotify_mnt = kern_mount(&inotify_fs_type); inotify_max_queued_events = 8192; - inotify_max_user_devices = 128; + inotify_max_user_instances = 8; inotify_max_user_watches = 8192; atomic_set(&inotify_cookie, 0); diff --git a/include/linux/sysctl.h b/include/linux/sysctl.h index ce19a2aa0b21..bfbbe94b297d 100644 --- a/include/linux/sysctl.h +++ b/include/linux/sysctl.h @@ -61,8 +61,7 @@ enum CTL_DEV=7, /* Devices */ CTL_BUS=8, /* Busses */ CTL_ABI=9, /* Binary emulation */ - CTL_CPU=10, /* CPU stuff (speed scaling, etc) */ - CTL_INOTIFY=11 /* Inotify */ + CTL_CPU=10 /* CPU stuff (speed scaling, etc) */ }; /* CTL_BUS names: */ @@ -71,12 +70,12 @@ enum CTL_BUS_ISA=1 /* ISA */ }; -/* CTL_INOTIFY names: */ +/* /proc/sys/fs/inotify/ */ enum { - INOTIFY_MAX_USER_DEVICES=1, /* max number of inotify device instances per user */ - INOTIFY_MAX_USER_WATCHES=2, /* max number of inotify watches per user */ - INOTIFY_MAX_QUEUED_EVENTS=3 /* Max number of queued events per inotify device instance */ + INOTIFY_MAX_USER_INSTANCES=1, /* max instances per user */ + INOTIFY_MAX_USER_WATCHES=2, /* max watches per user */ + INOTIFY_MAX_QUEUED_EVENTS=3 /* max queued events per instance */ }; /* CTL_KERN names: */ @@ -685,6 +684,7 @@ enum FS_XFS=17, /* struct: control xfs parameters */ FS_AIO_NR=18, /* current system-wide number of aio requests */ FS_AIO_MAX_NR=19, /* system-wide maximum number of aio requests */ + FS_INOTIFY=20, /* inotify submenu */ }; /* /proc/sys/fs/quota/ */ diff --git a/kernel/sysctl.c b/kernel/sysctl.c index b240e2cb86fc..e60b9c36f1f0 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -67,12 +67,6 @@ extern int printk_ratelimit_jiffies; extern int printk_ratelimit_burst; extern int pid_max_min, pid_max_max; -#ifdef CONFIG_INOTIFY -extern int inotify_max_user_devices; -extern int inotify_max_user_watches; -extern int inotify_max_queued_events; -#endif - #if defined(CONFIG_X86_LOCAL_APIC) && defined(CONFIG_X86) int unknown_nmi_panic; extern int proc_unknown_nmi_panic(ctl_table *, int, struct file *, @@ -152,6 +146,9 @@ extern ctl_table random_table[]; #ifdef CONFIG_UNIX98_PTYS extern ctl_table pty_table[]; #endif +#ifdef CONFIG_INOTIFY +extern ctl_table inotify_table[]; +#endif #ifdef HAVE_ARCH_PICK_MMAP_LAYOUT int sysctl_legacy_va_layout; @@ -957,6 +954,14 @@ static ctl_table fs_table[] = { .mode = 0644, .proc_handler = &proc_dointvec, }, +#ifdef CONFIG_INOTIFY + { + .ctl_name = FS_INOTIFY, + .procname = "inotify", + .mode = 0555, + .child = inotify_table, + }, +#endif #endif { .ctl_name = KERN_SETUID_DUMPABLE, @@ -966,40 +971,6 @@ static ctl_table fs_table[] = { .mode = 0644, .proc_handler = &proc_dointvec, }, -#ifdef CONFIG_INOTIFY - { - .ctl_name = INOTIFY_MAX_USER_DEVICES, - .procname = "max_user_devices", - .data = &inotify_max_user_devices, - .maxlen = sizeof(int), - .mode = 0644, - .proc_handler = &proc_dointvec_minmax, - .strategy = &sysctl_intvec, - .extra1 = &zero, - }, - - { - .ctl_name = INOTIFY_MAX_USER_WATCHES, - .procname = "max_user_watches", - .data = &inotify_max_user_watches, - .maxlen = sizeof(int), - .mode = 0644, - .proc_handler = &proc_dointvec_minmax, - .strategy = &sysctl_intvec, - .extra1 = &zero, - }, - - { - .ctl_name = INOTIFY_MAX_QUEUED_EVENTS, - .procname = "max_queued_events", - .data = &inotify_max_queued_events, - .maxlen = sizeof(int), - .mode = 0644, - .proc_handler = &proc_dointvec_minmax, - .strategy = &sysctl_intvec, - .extra1 = &zero - }, -#endif { .ctl_name = 0 } }; -- cgit v1.2.3 From 9a556e89081b0c1c2f83cee915363b15a68a6f2d Mon Sep 17 00:00:00 2001 From: Robert Love Date: Wed, 13 Jul 2005 13:49:23 -0400 Subject: [PATCH] inotify: misc cleanup Really simple, basic cleanup. Signed-off-by: Robert Love Signed-off-by: Linus Torvalds --- fs/inotify.c | 9 +++------ 1 file changed, 3 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/inotify.c b/fs/inotify.c index fb4803131423..54757be888b6 100644 --- a/fs/inotify.c +++ b/fs/inotify.c @@ -29,8 +29,6 @@ #include #include #include -#include -#include #include #include #include @@ -934,7 +932,7 @@ asmlinkage long sys_inotify_add_watch(int fd, const char *path, u32 mask) dev = filp->private_data; - ret = find_inode ((const char __user*)path, &nd); + ret = find_inode((const char __user*) path, &nd); if (ret) goto fput_and_out; @@ -991,8 +989,9 @@ asmlinkage long sys_inotify_rm_watch(int fd, u32 wd) if (!filp) return -EBADF; dev = filp->private_data; - ret = inotify_ignore (dev, wd); + ret = inotify_ignore(dev, wd); fput(filp); + return ret; } @@ -1032,8 +1031,6 @@ static int __init inotify_init(void) sizeof(struct inotify_kernel_event), 0, SLAB_PANIC, NULL, NULL); - printk(KERN_INFO "inotify syscall\n"); - return 0; } -- cgit v1.2.3 From 88bd5121d635136e01369141367f315665534b3c Mon Sep 17 00:00:00 2001 From: Anton Altaparmakov Date: Wed, 13 Jul 2005 01:10:44 -0700 Subject: [PATCH] Fix soft lockup due to NTFS: VFS part and explanation Something has changed in the core kernel such that we now get concurrent inode write outs, one e.g via pdflush and one via sys_sync or whatever. This causes a nasty deadlock in ntfs. The only clean solution unfortunately requires a minor vfs api extension. First the deadlock analysis: Prerequisive knowledge: NTFS has a file $MFT (inode 0) loaded at mount time. The NTFS driver uses the page cache for storing the file contents as usual. More interestingly this file contains the table of on-disk inodes as a sequence of MFT_RECORDs. Thus NTFS driver accesses the on-disk inodes by accessing the MFT_RECORDs in the page cache pages of the loaded inode $MFT. The situation: VFS inode X on a mounted ntfs volume is dirty. For same inode X, the ntfs_inode is dirty and thus corresponding on-disk inode, which is as explained above in a dirty PAGE_CACHE_PAGE belonging to the table of inodes ($MFT, inode 0). What happens: Process 1: sys_sync()/umount()/whatever... calls __sync_single_inode() for $MFT -> do_writepages() -> write_page for the dirty page containing the on-disk inode X, the page is now locked -> ntfs_write_mst_block() which clears PageUptodate() on the page to prevent anyone else getting hold of it whilst it does the write out (this is necessary as the on-disk inode needs "fixups" applied before the write to disk which are removed again after the write and PageUptodate is then set again). It then analyses the page looking for dirty on-disk inodes and when it finds one it calls ntfs_may_write_mft_record() to see if it is safe to write this on-disk inode. This then calls ilookup5() to check if the corresponding VFS inode is in icache(). This in turn calls ifind() which waits on the inode lock via wait_on_inode whilst holding the global inode_lock. Process 2: pdflush results in a call to __sync_single_inode for the same VFS inode X on the ntfs volume. This locks the inode (I_LOCK) then calls write-inode -> ntfs_write_inode -> map_mft_record() -> read_cache_page() of the page (in page cache of table of inodes $MFT, inode 0) containing the on-disk inode. This page has PageUptodate() clear because of Process 1 (see above) so read_cache_page() blocks when tries to take the page lock for the page so it can call ntfs_read_page(). Thus Process 1 is holding the page lock on the page containing the on-disk inode X and it is waiting on the inode X to be unlocked in ifind() so it can write the page out and then unlock the page. And Process 2 is holding the inode lock on inode X and is waiting for the page to be unlocked so it can call ntfs_readpage() or discover that Process 1 set PageUptodate() again and use the page. Thus we have a deadlock due to ifind() waiting on the inode lock. The only sensible solution: NTFS does not care whether the VFS inode is locked or not when it calls ilookup5() (it doesn't use the VFS inode at all, it just uses it to find the corresponding ntfs_inode which is of course attached to the VFS inode (both are one single struct); and it uses the ntfs_inode which is subject to its own locking so I_LOCK is irrelevant) hence we want a modified ilookup5_nowait() which is the same as ilookup5() but it does not wait on the inode lock. Without such functionality I would have to keep my own ntfs_inode cache in the NTFS driver just so I can find ntfs_inodes independent of their VFS inodes which would be slow, memory and cpu cycle wasting, and incredibly stupid given the icache already exists in the VFS. Below is a patch that does the ilookup5_nowait() implementation in fs/inode.c and exports it. ilookup5_nowait.diff: Introduce ilookup5_nowait() which is basically the same as ilookup5() but it does not wait on the inode's lock (i.e. it omits the wait_on_inode() done in ifind()). This is needed to avoid a nasty deadlock in NTFS. Signed-off-by: Anton Altaparmakov Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/inode.c | 45 +++++++++++++++++++++++++++++++++++++++------ include/linux/fs.h | 3 +++ 2 files changed, 42 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/inode.c b/fs/inode.c index 96364fae0844..e57f1724db3e 100644 --- a/fs/inode.c +++ b/fs/inode.c @@ -757,6 +757,7 @@ EXPORT_SYMBOL(igrab); * @head: the head of the list to search * @test: callback used for comparisons between inodes * @data: opaque data pointer to pass to @test + * @wait: if true wait for the inode to be unlocked, if false do not * * ifind() searches for the inode specified by @data in the inode * cache. This is a generalized version of ifind_fast() for file systems where @@ -771,7 +772,7 @@ EXPORT_SYMBOL(igrab); */ static inline struct inode *ifind(struct super_block *sb, struct hlist_head *head, int (*test)(struct inode *, void *), - void *data) + void *data, const int wait) { struct inode *inode; @@ -780,7 +781,8 @@ static inline struct inode *ifind(struct super_block *sb, if (inode) { __iget(inode); spin_unlock(&inode_lock); - wait_on_inode(inode); + if (likely(wait)) + wait_on_inode(inode); return inode; } spin_unlock(&inode_lock); @@ -820,7 +822,7 @@ static inline struct inode *ifind_fast(struct super_block *sb, } /** - * ilookup5 - search for an inode in the inode cache + * ilookup5_nowait - search for an inode in the inode cache * @sb: super block of file system to search * @hashval: hash value (usually inode number) to search for * @test: callback used for comparisons between inodes @@ -832,7 +834,38 @@ static inline struct inode *ifind_fast(struct super_block *sb, * identification of an inode. * * If the inode is in the cache, the inode is returned with an incremented - * reference count. + * reference count. Note, the inode lock is not waited upon so you have to be + * very careful what you do with the returned inode. You probably should be + * using ilookup5() instead. + * + * Otherwise NULL is returned. + * + * Note, @test is called with the inode_lock held, so can't sleep. + */ +struct inode *ilookup5_nowait(struct super_block *sb, unsigned long hashval, + int (*test)(struct inode *, void *), void *data) +{ + struct hlist_head *head = inode_hashtable + hash(sb, hashval); + + return ifind(sb, head, test, data, 0); +} + +EXPORT_SYMBOL(ilookup5_nowait); + +/** + * ilookup5 - search for an inode in the inode cache + * @sb: super block of file system to search + * @hashval: hash value (usually inode number) to search for + * @test: callback used for comparisons between inodes + * @data: opaque data pointer to pass to @test + * + * ilookup5() uses ifind() to search for the inode specified by @hashval and + * @data in the inode cache. This is a generalized version of ilookup() for + * file systems where the inode number is not sufficient for unique + * identification of an inode. + * + * If the inode is in the cache, the inode lock is waited upon and the inode is + * returned with an incremented reference count. * * Otherwise NULL is returned. * @@ -843,7 +876,7 @@ struct inode *ilookup5(struct super_block *sb, unsigned long hashval, { struct hlist_head *head = inode_hashtable + hash(sb, hashval); - return ifind(sb, head, test, data); + return ifind(sb, head, test, data, 1); } EXPORT_SYMBOL(ilookup5); @@ -900,7 +933,7 @@ struct inode *iget5_locked(struct super_block *sb, unsigned long hashval, struct hlist_head *head = inode_hashtable + hash(sb, hashval); struct inode *inode; - inode = ifind(sb, head, test, data); + inode = ifind(sb, head, test, data, 1); if (inode) return inode; /* diff --git a/include/linux/fs.h b/include/linux/fs.h index c9bf3746a9fb..0f53e0124941 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1441,6 +1441,9 @@ extern int inode_needs_sync(struct inode *inode); extern void generic_delete_inode(struct inode *inode); extern void generic_drop_inode(struct inode *inode); +extern struct inode *ilookup5_nowait(struct super_block *sb, + unsigned long hashval, int (*test)(struct inode *, void *), + void *data); extern struct inode *ilookup5(struct super_block *sb, unsigned long hashval, int (*test)(struct inode *, void *), void *data); extern struct inode *ilookup(struct super_block *sb, unsigned long ino); -- cgit v1.2.3 From 7ee91ec14bd4def036b4084da29869382078e44b Mon Sep 17 00:00:00 2001 From: Steve Dickson Date: Wed, 13 Jul 2005 01:10:47 -0700 Subject: [PATCH] NFS: procfs/sysctl interfaces for lockd do not work on x86_64 Allow the setting of NLM timeouts and grace periods through the proc and sysclt interfaces on x86_64 architectures Signed-off-by: Steve Dickson Acked-by: Trond Myklebust Cc: Neil Brown Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/lockd/svc.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/lockd/svc.c b/fs/lockd/svc.c index 6e242556b903..12a857c29e25 100644 --- a/fs/lockd/svc.c +++ b/fs/lockd/svc.c @@ -331,7 +331,7 @@ static ctl_table nlm_sysctls[] = { .ctl_name = CTL_UNNUMBERED, .procname = "nlm_grace_period", .data = &nlm_grace_period, - .maxlen = sizeof(int), + .maxlen = sizeof(unsigned long), .mode = 0644, .proc_handler = &proc_doulongvec_minmax, .extra1 = (unsigned long *) &nlm_grace_period_min, @@ -341,7 +341,7 @@ static ctl_table nlm_sysctls[] = { .ctl_name = CTL_UNNUMBERED, .procname = "nlm_timeout", .data = &nlm_timeout, - .maxlen = sizeof(int), + .maxlen = sizeof(unsigned long), .mode = 0644, .proc_handler = &proc_doulongvec_minmax, .extra1 = (unsigned long *) &nlm_timeout_min, -- cgit v1.2.3 From 8e0a21812427313f8ee8521688e81bfcc6f207a2 Mon Sep 17 00:00:00 2001 From: Paolo 'Blaisorblade' Giarrusso Date: Thu, 14 Jul 2005 00:33:41 -0700 Subject: [PATCH] uml: fix hppfs error path Fix the error message to refer to the error code, i.e. err, not count, plus add some cosmetical fixes. Signed-off-by: Paolo 'Blaisorblade' Giarrusso Cc: Jeff Dike Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/hppfs/hppfs_kern.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/hppfs/hppfs_kern.c b/fs/hppfs/hppfs_kern.c index 6f553e17c375..ff150fedb981 100644 --- a/fs/hppfs/hppfs_kern.c +++ b/fs/hppfs/hppfs_kern.c @@ -233,7 +233,7 @@ static ssize_t read_proc(struct file *file, char *buf, ssize_t count, set_fs(USER_DS); if(ppos) *ppos = file->f_pos; - return(n); + return n; } static ssize_t hppfs_read_file(int fd, char *buf, ssize_t count) @@ -254,7 +254,7 @@ static ssize_t hppfs_read_file(int fd, char *buf, ssize_t count) err = os_read_file(fd, new_buf, cur); if(err < 0){ printk("hppfs_read : read failed, errno = %d\n", - count); + err); n = err; goto out_free; } @@ -271,7 +271,7 @@ static ssize_t hppfs_read_file(int fd, char *buf, ssize_t count) out_free: kfree(new_buf); out: - return(n); + return n; } static ssize_t hppfs_read(struct file *file, char *buf, size_t count, -- cgit v1.2.3 From a0d43df931870997467042931e8fb7657375d327 Mon Sep 17 00:00:00 2001 From: Paolo 'Blaisorblade' Giarrusso Date: Thu, 14 Jul 2005 00:33:43 -0700 Subject: [PATCH] uml: hostfs: unuse ROOT_DEV Minimal patch removing uses of ROOT_DEV; next patch unexports it. I've opposed this, but I've planned to reintroduce the functionality without using ROOT_DEV. Signed-off-by: Paolo 'Blaisorblade' Giarrusso Cc: Christoph Hellwig Cc: Jeff Dike Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/hostfs/hostfs_kern.c | 9 --------- 1 file changed, 9 deletions(-) (limited to 'fs') diff --git a/fs/hostfs/hostfs_kern.c b/fs/hostfs/hostfs_kern.c index 4bf43ea87c46..88e68caa3784 100644 --- a/fs/hostfs/hostfs_kern.c +++ b/fs/hostfs/hostfs_kern.c @@ -15,7 +15,6 @@ #include #include #include -#include #include #include #include @@ -160,8 +159,6 @@ static int read_name(struct inode *ino, char *name) ino->i_size = i_size; ino->i_blksize = i_blksize; ino->i_blocks = i_blocks; - if((ino->i_sb->s_dev == ROOT_DEV) && (ino->i_uid == getuid())) - ino->i_uid = 0; return(0); } @@ -841,16 +838,10 @@ int hostfs_setattr(struct dentry *dentry, struct iattr *attr) attrs.ia_mode = attr->ia_mode; } if(attr->ia_valid & ATTR_UID){ - if((dentry->d_inode->i_sb->s_dev == ROOT_DEV) && - (attr->ia_uid == 0)) - attr->ia_uid = getuid(); attrs.ia_valid |= HOSTFS_ATTR_UID; attrs.ia_uid = attr->ia_uid; } if(attr->ia_valid & ATTR_GID){ - if((dentry->d_inode->i_sb->s_dev == ROOT_DEV) && - (attr->ia_gid == 0)) - attr->ia_gid = getgid(); attrs.ia_valid |= HOSTFS_ATTR_GID; attrs.ia_gid = attr->ia_gid; } -- cgit v1.2.3 From afa597ba20e9ef55fc6283c1a564854b1c9f13c0 Mon Sep 17 00:00:00 2001 From: Carsten Otte Date: Fri, 15 Jul 2005 03:56:30 -0700 Subject: [PATCH] execute-in-place fixes This patch includes feedback from Andrew and Christoph. Thanks for taking time to review. Use of empty_zero_page was eliminated to fix compilation for architectures that don't have it. This patch removes setting pages up-to-date in ext2_get_xip_page and all bug checks to verify that the page is indeed up to date. Setting the page state on mapping to userland is bogus. None of the code patchs involved with these pages in mm cares about the page state. still on my ToDo list: identify a place outside second extended where __inode_direct_access should reside Signed-off-by: Carsten Otte Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext2/xip.c | 81 ++++++++++++++++++++++++++++++++------------------------ mm/filemap_xip.c | 23 ++++++---------- 2 files changed, 55 insertions(+), 49 deletions(-) (limited to 'fs') diff --git a/fs/ext2/xip.c b/fs/ext2/xip.c index d44431d1a338..0aa5ac159c09 100644 --- a/fs/ext2/xip.c +++ b/fs/ext2/xip.c @@ -15,66 +15,79 @@ #include "xip.h" static inline int -__inode_direct_access(struct inode *inode, sector_t sector, unsigned long *data) { +__inode_direct_access(struct inode *inode, sector_t sector, + unsigned long *data) +{ BUG_ON(!inode->i_sb->s_bdev->bd_disk->fops->direct_access); return inode->i_sb->s_bdev->bd_disk->fops ->direct_access(inode->i_sb->s_bdev,sector,data); } +static inline int +__ext2_get_sector(struct inode *inode, sector_t offset, int create, + sector_t *result) +{ + struct buffer_head tmp; + int rc; + + memset(&tmp, 0, sizeof(struct buffer_head)); + rc = ext2_get_block(inode, offset/ (PAGE_SIZE/512), &tmp, + create); + *result = tmp.b_blocknr; + + /* did we get a sparse block (hole in the file)? */ + if (!(*result)) { + BUG_ON(create); + rc = -ENODATA; + } + + return rc; +} + int -ext2_clear_xip_target(struct inode *inode, int block) { - sector_t sector = block*(PAGE_SIZE/512); +ext2_clear_xip_target(struct inode *inode, int block) +{ + sector_t sector = block * (PAGE_SIZE/512); unsigned long data; int rc; rc = __inode_direct_access(inode, sector, &data); - if (rc) - return rc; - clear_page((void*)data); - return 0; + if (!rc) + clear_page((void*)data); + return rc; } void ext2_xip_verify_sb(struct super_block *sb) { struct ext2_sb_info *sbi = EXT2_SB(sb); - if ((sbi->s_mount_opt & EXT2_MOUNT_XIP)) { - if ((sb->s_bdev == NULL) || - sb->s_bdev->bd_disk == NULL || - sb->s_bdev->bd_disk->fops == NULL || - sb->s_bdev->bd_disk->fops->direct_access == NULL) { - sbi->s_mount_opt &= (~EXT2_MOUNT_XIP); - ext2_warning(sb, __FUNCTION__, - "ignoring xip option - not supported by bdev"); - } + if ((sbi->s_mount_opt & EXT2_MOUNT_XIP) && + !sb->s_bdev->bd_disk->fops->direct_access) { + sbi->s_mount_opt &= (~EXT2_MOUNT_XIP); + ext2_warning(sb, __FUNCTION__, + "ignoring xip option - not supported by bdev"); } } -struct page* -ext2_get_xip_page(struct address_space *mapping, sector_t blockno, +struct page * +ext2_get_xip_page(struct address_space *mapping, sector_t offset, int create) { int rc; unsigned long data; - struct buffer_head tmp; + sector_t sector; - tmp.b_state = 0; - tmp.b_blocknr = 0; - rc = ext2_get_block(mapping->host, blockno/(PAGE_SIZE/512) , &tmp, - create); + /* first, retrieve the sector number */ + rc = __ext2_get_sector(mapping->host, offset, create, §or); if (rc) - return ERR_PTR(rc); - if (tmp.b_blocknr == 0) { - /* SPARSE block */ - BUG_ON(create); - return ERR_PTR(-ENODATA); - } + goto error; + /* retrieve address of the target data */ rc = __inode_direct_access - (mapping->host,tmp.b_blocknr*(PAGE_SIZE/512) ,&data); - if (rc) - return ERR_PTR(rc); + (mapping->host, sector * (PAGE_SIZE/512), &data); + if (!rc) + return virt_to_page(data); - SetPageUptodate(virt_to_page(data)); - return virt_to_page(data); + error: + return ERR_PTR(rc); } diff --git a/mm/filemap_xip.c b/mm/filemap_xip.c index 4553b2c5aab4..8c199f537732 100644 --- a/mm/filemap_xip.c +++ b/mm/filemap_xip.c @@ -68,13 +68,12 @@ do_xip_mapping_read(struct address_space *mapping, if (unlikely(IS_ERR(page))) { if (PTR_ERR(page) == -ENODATA) { /* sparse */ - page = virt_to_page(empty_zero_page); + page = ZERO_PAGE(0); } else { desc->error = PTR_ERR(page); goto out; } - } else - BUG_ON(!PageUptodate(page)); + } /* If users can be writing to this page using arbitrary * virtual addresses, take care about potential aliasing @@ -84,8 +83,7 @@ do_xip_mapping_read(struct address_space *mapping, flush_dcache_page(page); /* - * Ok, we have the page, and it's up-to-date, so - * now we can copy it to user space... + * Ok, we have the page, so now we can copy it to user space... * * The actor routine returns how many bytes were actually used.. * NOTE! This may not be the same as how much of a user buffer @@ -164,7 +162,7 @@ EXPORT_SYMBOL_GPL(xip_file_sendfile); * xip_write * * This function walks all vmas of the address_space and unmaps the - * empty_zero_page when found at pgoff. Should it go in rmap.c? + * ZERO_PAGE when found at pgoff. Should it go in rmap.c? */ static void __xip_unmap (struct address_space * mapping, @@ -187,7 +185,7 @@ __xip_unmap (struct address_space * mapping, * We need the page_table_lock to protect us from page faults, * munmap, fork, etc... */ - pte = page_check_address(virt_to_page(empty_zero_page), mm, + pte = page_check_address(ZERO_PAGE(address), mm, address); if (!IS_ERR(pte)) { /* Nuke the page table entry. */ @@ -230,7 +228,6 @@ xip_file_nopage(struct vm_area_struct * area, page = mapping->a_ops->get_xip_page(mapping, pgoff*(PAGE_SIZE/512), 0); if (!IS_ERR(page)) { - BUG_ON(!PageUptodate(page)); return page; } if (PTR_ERR(page) != -ENODATA) @@ -245,12 +242,11 @@ xip_file_nopage(struct vm_area_struct * area, pgoff*(PAGE_SIZE/512), 1); if (IS_ERR(page)) return NULL; - BUG_ON(!PageUptodate(page)); /* unmap page at pgoff from all other vmas */ __xip_unmap(mapping, pgoff); } else { - /* not shared and writable, use empty_zero_page */ - page = virt_to_page(empty_zero_page); + /* not shared and writable, use ZERO_PAGE() */ + page = ZERO_PAGE(address); } return page; @@ -319,8 +315,6 @@ __xip_file_write(struct file *filp, const char __user *buf, break; } - BUG_ON(!PageUptodate(page)); - copied = filemap_copy_from_user(page, offset, buf, bytes); flush_dcache_page(page); if (likely(copied > 0)) { @@ -435,8 +429,7 @@ xip_truncate_page(struct address_space *mapping, loff_t from) return 0; else return PTR_ERR(page); - } else - BUG_ON(!PageUptodate(page)); + } kaddr = kmap_atomic(page, KM_USER0); memset(kaddr + offset, 0, length); kunmap_atomic(kaddr, KM_USER0); -- cgit v1.2.3