Gitweb: http://git.fedorahosted.org/git/cluster.git?p=cluster.git;a=commitdiff;h=... Commit: 96c846d5aefd74b7a84078987ebb3610180f5925 Parent: d22e78f8fc996480d873585ad2939be826343b05 Author: Bob Peterson <rpeterso(a)redhat.com&gt; AuthorDate: Wed Aug 18 13:06:31 2010 -0500 Committer: Bob Peterson <rpeterso(a)redhat.com&gt; CommitterDate: Wed Aug 18 13:06:31 2010 -0500 fsck.gfs2 segfaults if journals are missing This patch makes it possible for fsck.gfs2 to detect missing and damaged journals and recreate them. Previous versions of fsck.gfs2 would segfault if journals were missing because it made assumptions about the jindex file pointing to a valid dinode. Also, I moved function ji_update which was exclusively used by fsck.gfs2 from libgfs2 to fsck. rhbz#622576 --- gfs2/convert/gfs2_convert.c | 12 +++-- gfs2/fsck/fs_recovery.c | 55 ++++++++++++++++++++++- gfs2/fsck/fs_recovery.h | 7 ++- gfs2/fsck/pass1.c | 104 +++++++++++++++++++++++++++--------------- gfs2/libgfs2/libgfs2.h | 5 +- gfs2/libgfs2/structures.c | 32 +++++++------ gfs2/libgfs2/super.c | 43 +---------------- 7 files changed, 154 insertions(+), 104 deletions(-) diff --git a/gfs2/convert/gfs2_convert.c b/gfs2/convert/gfs2_convert.c index 642d26a..6a754b1 100644 --- a/gfs2/convert/gfs2_convert.c +++ b/gfs2/convert/gfs2_convert.c @@ -1987,22 +1987,24 @@ static void conv_build_jindex(struct gfs2_sbd *sdp) sdp->md.jiinode = createi(sdp->master_dir, "jindex", S_IFDIR | 0700, GFS2_DIF_SYSTEM); + sdp->md.journal = malloc(sdp->md.journals * + sizeof(struct gfs2_inode *)); for (j = 0; j < sdp->md.journals; j++) { char name[256]; - struct gfs2_inode *ip; printf("Writing journal #%d...", j + 1); fflush(stdout); sprintf(name, "journal%u", j); - ip = createi(sdp->md.jiinode, name, S_IFREG | 0600, - GFS2_DIF_SYSTEM); - write_journal(sdp, ip, j, + sdp->md.journal[j] = createi(sdp->md.jiinode, name, S_IFREG | + 0600, GFS2_DIF_SYSTEM); + write_journal(sdp, j, sdp->jsize << 20 >> sdp->sd_sb.sb_bsize_shift); - inode_put(&ip); + inode_put(&sdp->md.journal[j]); printf("done.\n"); fflush(stdout); } + free(sdp->md.journal); if (sdp->debug) { printf("\nJindex:\n"); gfs2_dinode_print(&sdp->md.jiinode->i_di); diff --git a/gfs2/fsck/fs_recovery.c b/gfs2/fsck/fs_recovery.c index d908873..07c1552 100644 --- a/gfs2/fsck/fs_recovery.c +++ b/gfs2/fsck/fs_recovery.c @@ -12,6 +12,8 @@ #include "libgfs2.h" #include "util.h" +#define JOURNAL_NAME_SIZE 16 + unsigned int sd_found_jblocks = 0, sd_replayed_jblocks = 0; unsigned int sd_found_metablocks = 0, sd_replayed_metablocks = 0; unsigned int sd_found_revokes = 0; @@ -540,7 +542,7 @@ out: log_info( _("jid=%u: Failed\n"), j); reinit: if (query( _("Do you want to clear the journal instead? (y/n)"))) - error = write_journal(sdp, sdp->md.journal[j], j, + error = write_journal(sdp, j, sdp->md.journal[j]->i_di.di_size / sdp->sd_sb.sb_bsize); else @@ -569,6 +571,7 @@ int replay_journals(struct gfs2_sbd *sdp, int preen, int force_check, *clean_journals = 0; + sdp->jsize = GFS2_DEFAULT_JSIZE; /* Get master dinode */ gfs2_lookupi(sdp->master_dir, "jindex", 6, &sdp->md.jiinode); @@ -579,7 +582,19 @@ int replay_journals(struct gfs2_sbd *sdp, int preen, int force_check, } for(i = 0; i < sdp->md.journals; i++) { + if (!sdp->md.journal[i]) { + log_err(_("File system journal \"journal%d\" is " + "missing: pass1 will try to recreate it.\n"), + i); + continue; + } if (!error) { + uint64_t jsize = sdp->md.journal[i]->i_di.di_size / + (1024 * 1024); + + if (sdp->jsize == GFS2_DEFAULT_JSIZE && jsize && + jsize != sdp->jsize) + sdp->jsize = jsize; error = gfs2_recover_journal(sdp->md.journal[i], i, preen, force_check, &clean); @@ -600,3 +615,41 @@ int replay_journals(struct gfs2_sbd *sdp, int preen, int force_check, fsync(sdp->device_fd); return error; } + +/* + * ji_update - fill in journal info + * sdp: the incore superblock pointer + * + * Given the inode for the journal index, read in all + * the journal inodes. + * + * Returns: 0 on success, -1 on failure + */ +int ji_update(struct gfs2_sbd *sdp) +{ + struct gfs2_inode *jip, *ip = sdp->md.jiinode; + char journal_name[JOURNAL_NAME_SIZE]; + int i; + + if(!ip) { + log_crit("Journal index inode not found.\n"); + return -1; + } + + if(!(sdp->md.journal = calloc(ip->i_di.di_entries - 2, sizeof(struct gfs2_inode *)))) { + log_err("Unable to allocate journal index\n"); + return -1; + } + sdp->md.journals = 0; + memset(journal_name, 0, sizeof(*journal_name)); + for(i = 0; i < ip->i_di.di_entries - 2; i++) { + /* FIXME check snprintf return code */ + snprintf(journal_name, JOURNAL_NAME_SIZE, "journal%u", i); + gfs2_lookupi(sdp->md.jiinode, journal_name, strlen(journal_name), + &jip); + sdp->md.journal[i] = jip; + } + sdp->md.journals = ip->i_di.di_entries - 2; + return 0; + +} diff --git a/gfs2/fsck/fs_recovery.h b/gfs2/fsck/fs_recovery.h index a61884d..1a7aae3 100644 --- a/gfs2/fsck/fs_recovery.h +++ b/gfs2/fsck/fs_recovery.h @@ -3,9 +3,10 @@ #include "libgfs2.h" -int replay_journals(struct gfs2_sbd *sdp, int preen, int force_check, - int *clean_journals); -int preen_is_safe(struct gfs2_sbd *sdp, int preen, int force_check); +extern int replay_journals(struct gfs2_sbd *sdp, int preen, int force_check, + int *clean_journals); +extern int preen_is_safe(struct gfs2_sbd *sdp, int preen, int force_check); +extern int ji_update(struct gfs2_sbd *sdp); #endif /* __FS_RECOVERY_H__ */ diff --git a/gfs2/fsck/pass1.c b/gfs2/fsck/pass1.c index 4fb7abf..17b23df 100644 --- a/gfs2/fsck/pass1.c +++ b/gfs2/fsck/pass1.c @@ -1174,25 +1174,37 @@ static int handle_di(struct gfs2_sbd *sdp, struct gfs2_buffer_head *bh) /* Check system inode and verify it's marked "in use" in the bitmap: */ /* Should work for all system inodes: root, master, jindex, per_node, etc. */ -static int check_system_inode(struct gfs2_inode *sysinode, const char *filename, - int builder(struct gfs2_sbd *sbp), - enum gfs2_mark_block mark) +/* We have to pass the sysinode as ** because the pointer may change out from + under the reference by way of the builder() function. */ +static int check_system_inode(struct gfs2_sbd *sdp, + struct gfs2_inode **sysinode, + const char *filename, + int builder(struct gfs2_sbd *sbp), + enum gfs2_mark_block mark) { uint64_t iblock = 0; struct dir_status ds = {0}; int error; log_info( _("Checking system inode '%s'\n"), filename); - if (sysinode) { + if (*sysinode) { /* Read in the system inode, look at its dentries, and start * reading through them */ - iblock = sysinode->i_di.di_num.no_addr; + iblock = (*sysinode)->i_di.di_num.no_addr; log_info( _("System inode for '%s' is located at block %" PRIu64 " (0x%" PRIx64 ")\n"), filename, iblock, iblock); - - /* FIXME: check this block's validity */ - + if (gfs2_check_meta((*sysinode)->i_bh, GFS2_METATYPE_DI)) { + log_err( _("Found invalid system dinode at block #" + "%llu (0x%llx)\n"), + (unsigned long long)iblock, + (unsigned long long)iblock); + gfs2_blockmap_set(bl, iblock, gfs2_block_free); + check_n_fix_bitmap(sdp, iblock, gfs2_block_free); + inode_put(sysinode); + } + } + if (*sysinode) { ds.q = block_type(iblock); /* If the inode exists but the block is marked free, we might be recovering from a corrupt bitmap. In that case, don't @@ -1201,52 +1213,67 @@ static int check_system_inode(struct gfs2_inode *sysinode, const char *filename, if (ds.q == gfs2_block_free) { log_info( _("The inode exists but the block is not " "marked 'in use'; fixing it.\n")); - fsck_blockmap_set(sysinode, - sysinode->i_di.di_num.no_addr, + fsck_blockmap_set(*sysinode, + (*sysinode)->i_di.di_num.no_addr, filename, mark); ds.q = mark; if (mark == gfs2_inode_dir) - dirtree_insert(sysinode->i_di.di_num.no_addr); + dirtree_insert((*sysinode)->i_di.di_num.no_addr); } } else - log_info( _("System inode for '%s' is missing.\n"), filename); + log_info( _("System inode for '%s' is corrupt or missing.\n"), + filename); /* If there are errors with the inode here, we need to create a new inode and get it all setup - of course, everything will be in lost+found then, but we *need* our system inodes before we can do any of that. */ - if(!sysinode || ds.q != mark) { + if(!(*sysinode) || ds.q != mark) { log_err( _("Invalid or missing %s system inode (should be %d, " "is %d).\n"), filename, mark, ds.q); if (query(_("Create new %s system inode? (y/n) "), filename)) { - builder(sysinode->i_sbd); - fsck_blockmap_set(sysinode, - sysinode->i_di.di_num.no_addr, + log_err( _("Rebuilding system file \"%s\"\n"), + filename); + error = builder(sdp); + if (error) { + log_err( _("Error trying to rebuild system " + "file %s: Cannot continue\n"), + filename); + return error; + } + fsck_blockmap_set(*sysinode, + (*sysinode)->i_di.di_num.no_addr, filename, mark); ds.q = mark; if (mark == gfs2_inode_dir) - dirtree_insert(sysinode->i_di.di_num.no_addr); + dirtree_insert((*sysinode)->i_di.di_num.no_addr); } else { log_err( _("Cannot continue without valid %s inode\n"), filename); return -1; } } - if (S_ISDIR(sysinode->i_di.di_mode)) { + if (S_ISDIR((*sysinode)->i_di.di_mode)) { struct block_count bc = {0}; sysdir_fxns.private = &bc; - if (sysinode->i_di.di_flags & GFS2_DIF_EXHASH) - check_metatree(sysinode, &sysdir_fxns); + if ((*sysinode)->i_di.di_flags & GFS2_DIF_EXHASH) + check_metatree(*sysinode, &sysdir_fxns); else - check_linear_dir(sysinode, sysinode->i_bh, + check_linear_dir(*sysinode, (*sysinode)->i_bh, &sysdir_fxns); } - error = handle_ip(sysinode->i_sbd, sysinode); + error = handle_ip(sdp, *sysinode); return error; } static int build_a_journal(struct gfs2_sbd *sdp) { + char name[256]; + + /* First, try to delete the journal if it's in jindex */ + sprintf(name, "journal%u", sdp->md.journals); + gfs2_dirent_del(sdp->md.jiinode, name, strlen(name)); + /* Now rebuild it */ build_journal(sdp, sdp->md.journals, sdp->md.jiinode); return 0; } @@ -1258,43 +1285,43 @@ static int check_system_inodes(struct gfs2_sbd *sdp) /******************************************************************* ******* Check the system inode integrity ************* *******************************************************************/ - if (check_system_inode(sdp->master_dir, "master", build_master, + if (check_system_inode(sdp, &sdp->master_dir, "master", build_master, gfs2_inode_dir)) { stack; return -1; } - if (check_system_inode(sdp->md.rooti, "root", build_root, + if (check_system_inode(sdp, &sdp->md.rooti, "root", build_root, gfs2_inode_dir)) { stack; return -1; } - if (check_system_inode(sdp->md.inum, "inum", build_inum, + if (check_system_inode(sdp, &sdp->md.inum, "inum", build_inum, gfs2_inode_file)) { stack; return -1; } - if (check_system_inode(sdp->md.statfs, "statfs", build_statfs, + if (check_system_inode(sdp, &sdp->md.statfs, "statfs", build_statfs, gfs2_inode_file)) { stack; return -1; } - if (check_system_inode(sdp->md.jiinode, "jindex", build_jindex, + if (check_system_inode(sdp, &sdp->md.jiinode, "jindex", build_jindex, gfs2_inode_dir)) { stack; return -1; } - if (check_system_inode(sdp->md.riinode, "rindex", build_rindex, + if (check_system_inode(sdp, &sdp->md.riinode, "rindex", build_rindex, gfs2_inode_file)) { stack; return -1; } - if (check_system_inode(sdp->md.qinode, "quota", build_quota, + if (check_system_inode(sdp, &sdp->md.qinode, "quota", build_quota, gfs2_inode_file)) { stack; return -1; } - if (check_system_inode(sdp->md.pinode, "per_node", build_per_node, - gfs2_inode_dir)) { + if (check_system_inode(sdp, &sdp->md.pinode, "per_node", + build_per_node, gfs2_inode_dir)) { stack; return -1; } @@ -1306,7 +1333,7 @@ static int check_system_inodes(struct gfs2_sbd *sdp) char jname[16]; sprintf(jname, "journal%d", sdp->md.journals); - if (check_system_inode(sdp->md.journal[sdp->md.journals], + if (check_system_inode(sdp, &sdp->md.journal[sdp->md.journals], jname, build_a_journal, gfs2_inode_file)) { stack; @@ -1399,6 +1426,14 @@ int pass1(struct gfs2_sbd *sbp) skip_this_pass = FALSE; fflush(stdout); } + if (fsck_system_inode(sbp, block)) { + log_debug(_("Already processed system inode " + "%lld (0x%llx)\n"), + (unsigned long long)block, + (unsigned long long)block); + first = 0; + continue; + } bh = bread(sbp, block); /*log_debug( _("Checking metadata block #%" PRIu64 @@ -1417,11 +1452,6 @@ int pass1(struct gfs2_sbd *sbp) } check_n_fix_bitmap(sbp, block, gfs2_block_free); - } else if (fsck_system_inode(sbp, block)) { - log_debug(_("Already processed system inode " - "%lld (0x%llx)\n"), - (unsigned long long)block, - (unsigned long long)block); } else if (handle_di(sbp, bh) < 0) { stack; brelse(bh); diff --git a/gfs2/libgfs2/libgfs2.h b/gfs2/libgfs2/libgfs2.h index 8835a4f..c798bda 100644 --- a/gfs2/libgfs2/libgfs2.h +++ b/gfs2/libgfs2/libgfs2.h @@ -472,8 +472,8 @@ extern int gfs2_dirent_next(struct gfs2_inode *dip, struct gfs2_buffer_head *bh, extern void build_height(struct gfs2_inode *ip, int height); extern void unstuff_dinode(struct gfs2_inode *ip); extern unsigned int calc_tree_height(struct gfs2_inode *ip, uint64_t size); -extern int write_journal(struct gfs2_sbd *sdp, struct gfs2_inode *ip, - unsigned int j, unsigned int blocks); +extern int write_journal(struct gfs2_sbd *sdp, unsigned int j, + unsigned int blocks); /** * device_size - figure out a device's size @@ -671,7 +671,6 @@ extern int gfs2_next_rg_metatype(struct gfs2_sbd *sdp, struct rgrp_list *rgd, /* super.c */ extern int check_sb(struct gfs2_sb *sb); extern int read_sb(struct gfs2_sbd *sdp); -extern int ji_update(struct gfs2_sbd *sdp); extern int rindex_read(struct gfs2_sbd *sdp, int fd, int *count1, int *sane); extern int ri_update(struct gfs2_sbd *sdp, int fd, int *rgcount, int *sane); extern int write_sb(struct gfs2_sbd *sdp); diff --git a/gfs2/libgfs2/structures.c b/gfs2/libgfs2/structures.c index c288abd..66a5733 100644 --- a/gfs2/libgfs2/structures.c +++ b/gfs2/libgfs2/structures.c @@ -73,8 +73,7 @@ void build_sb(struct gfs2_sbd *sdp, const unsigned char *uuid) } } -int write_journal(struct gfs2_sbd *sdp, struct gfs2_inode *ip, unsigned int j, - unsigned int blocks) +int write_journal(struct gfs2_sbd *sdp, unsigned int j, unsigned int blocks) { struct gfs2_log_header lh; unsigned int x; @@ -84,8 +83,9 @@ int write_journal(struct gfs2_sbd *sdp, struct gfs2_inode *ip, unsigned int j, /* Build the height up so our journal blocks will be contiguous and */ /* not broken up by indirect block pages. */ - height = calc_tree_height(ip, (blocks + 1) * sdp->bsize); - build_height(ip, height); + height = calc_tree_height(sdp->md.journal[j], + (blocks + 1) * sdp->bsize); + build_height(sdp->md.journal[j], height); memset(&lh, 0, sizeof(struct gfs2_log_header)); lh.lh_header.mh_magic = GFS2_MAGIC; @@ -94,14 +94,16 @@ int write_journal(struct gfs2_sbd *sdp, struct gfs2_inode *ip, unsigned int j, lh.lh_flags = GFS2_LOG_HEAD_UNMOUNT; for (x = 0; x < blocks; x++) { - struct gfs2_buffer_head *bh = get_file_buf(ip, x, TRUE); + struct gfs2_buffer_head *bh = get_file_buf(sdp->md.journal[j], + x, TRUE); if (!bh) return -1; bmodified(bh); brelse(bh); } for (x = 0; x < blocks; x++) { - struct gfs2_buffer_head *bh = get_file_buf(ip, x, FALSE); + struct gfs2_buffer_head *bh = get_file_buf(sdp->md.journal[j], + x, FALSE); if (!bh) return -1; @@ -121,7 +123,7 @@ int write_journal(struct gfs2_sbd *sdp, struct gfs2_inode *ip, unsigned int j, if (sdp->debug) { printf("\nJournal %u:\n", j); - gfs2_dinode_print(&ip->i_di); + gfs2_dinode_print(&sdp->md.journal[j]->i_di); } return 0; } @@ -129,17 +131,14 @@ int write_journal(struct gfs2_sbd *sdp, struct gfs2_inode *ip, unsigned int j, int build_journal(struct gfs2_sbd *sdp, int j, struct gfs2_inode *jindex) { char name[256]; - struct gfs2_inode *ip; int ret; sprintf(name, "journal%u", j); - ip = createi(jindex, name, S_IFREG | 0600, GFS2_DIF_SYSTEM); - ret = write_journal(sdp, ip, j, + sdp->md.journal[j] = createi(jindex, name, S_IFREG | 0600, + GFS2_DIF_SYSTEM); + ret = write_journal(sdp, j, sdp->jsize << 20 >> sdp->sd_sb.sb_bsize_shift); - if (ret) - return ret; - inode_put(&ip); - return 0; + return ret; } int build_jindex(struct gfs2_sbd *sdp) @@ -150,17 +149,20 @@ int build_jindex(struct gfs2_sbd *sdp) jindex = createi(sdp->master_dir, "jindex", S_IFDIR | 0700, GFS2_DIF_SYSTEM); - + sdp->md.journal = malloc(sdp->md.journals * + sizeof(struct gfs2_inode *)); for (j = 0; j < sdp->md.journals; j++) { ret = build_journal(sdp, j, jindex); if (ret) return ret; + inode_put(&sdp->md.journal[j]); } if (sdp->debug) { printf("\nJindex:\n"); gfs2_dinode_print(&jindex->i_di); } + free(sdp->md.journal); inode_put(&jindex); return 0; } diff --git a/gfs2/libgfs2/super.c b/gfs2/libgfs2/super.c index c5b5800..5d232a6 100644 --- a/gfs2/libgfs2/super.c +++ b/gfs2/libgfs2/super.c @@ -75,6 +75,9 @@ int read_sb(struct gfs2_sbd *sdp) (sdp->sd_sb.sb_bsize-sizeof(struct gfs2_meta_header)) / sizeof(uint64_t); sdp->sd_jbsize = sdp->sd_sb.sb_bsize - sizeof(struct gfs2_meta_header); + sdp->sd_hash_bsize = sdp->bsize / 2; + sdp->sd_hash_bsize_shift = sdp->sd_sb.sb_bsize_shift - 1; + sdp->sd_hash_ptrs = sdp->sd_hash_bsize / sizeof(uint64_t); sdp->sd_heightsize[0] = sdp->sd_sb.sb_bsize - sizeof(struct gfs2_dinode); sdp->sd_heightsize[1] = sdp->sd_sb.sb_bsize * sdp->sd_diptrs; for (x = 2; ; x++){ @@ -113,46 +116,6 @@ int read_sb(struct gfs2_sbd *sdp) return error; } -#define JOURNAL_NAME_SIZE 16 - -/* - * ji_update - fill in journal info - * sdp: the incore superblock pointer - * - * Given the inode for the journal index, read in all - * the journal inodes. - * - * Returns: 0 on success, -1 on failure - */ -int ji_update(struct gfs2_sbd *sdp) -{ - struct gfs2_inode *jip, *ip = sdp->md.jiinode; - char journal_name[JOURNAL_NAME_SIZE]; - int i; - - if(!ip) { - log_crit("Journal index inode not found.\n"); - return -1; - } - - if(!(sdp->md.journal = calloc(ip->i_di.di_entries - 2, sizeof(struct gfs2_inode *)))) { - log_err("Unable to allocate journal index\n"); - return -1; - } - sdp->md.journals = 0; - memset(journal_name, 0, sizeof(*journal_name)); - for(i = 0; i < ip->i_di.di_entries - 2; i++) { - /* FIXME check snprintf return code */ - snprintf(journal_name, JOURNAL_NAME_SIZE, "journal%u", i); - gfs2_lookupi(sdp->md.jiinode, journal_name, strlen(journal_name), - &jip); - sdp->md.journal[i] = jip; - } - sdp->md.journals = ip->i_di.di_entries - 2; - return 0; - -} - /** * rindex_read - read in the rg index file * @sdp: the incore superblock pointer