X-Git-Url: https://git.whamcloud.com/?p=fs%2Flustre-release.git;a=blobdiff_plain;f=lustre%2Fllite%2Fdir.c;h=0233cce468be72971f4ff25d6665e0f9723c9faa;hp=ed8e1c7893ec0d104613c0b33492a62814ed91f4;hb=1ae726743b5a051bb4daa731d93496a3e7f7a93c;hpb=08aa217ce49aba1ded52e0f7adb8a607035123fd diff --git a/lustre/llite/dir.c b/lustre/llite/dir.c index ed8e1c7..0233cce 100644 --- a/lustre/llite/dir.c +++ b/lustre/llite/dir.c @@ -27,7 +27,7 @@ * Copyright (c) 2002, 2010, Oracle and/or its affiliates. All rights reserved. * Use is subject to license terms. * - * Copyright (c) 2011, 2012, Intel Corporation. + * Copyright (c) 2011, 2013, Intel Corporation. */ /* * This file is part of Lustre, http://www.lustre.org/ @@ -51,8 +51,8 @@ #include #include #include +#include #include -#include #include #include #include @@ -134,477 +134,187 @@ * a header lu_dirpage which describes the start/end hash, and whether this * page is empty (contains no dir entry) or hash collide with next page. * After client receives reply, several pages will be integrated into dir page - * in CFS_PAGE_SIZE (if CFS_PAGE_SIZE greater than LU_PAGE_SIZE), and the - * lu_dirpage for this integrated page will be adjusted. + * in PAGE_CACHE_SIZE (if PAGE_CACHE_SIZE greater than LU_PAGE_SIZE), and the + * lu_dirpage for this integrated page will be adjusted. See + * mdc_adjust_dirpages(). * */ - -/* returns the page unlocked, but with a reference */ -static int ll_dir_filler(void *_hash, struct page *page0) +struct page *ll_get_dir_page(struct inode *dir, struct md_op_data *op_data, + __u64 offset, struct ll_dir_chain *chain) { - struct inode *inode = page0->mapping->host; - int hash64 = ll_i2sbi(inode)->ll_flags & LL_SBI_64BIT_HASH; - struct obd_export *exp = ll_i2sbi(inode)->ll_md_exp; - struct ptlrpc_request *request; - struct mdt_body *body; - struct md_op_data *op_data; - __u64 hash = *((__u64 *)_hash); - struct page **page_pool; - struct page *page; -#ifndef HAVE_ADD_TO_PAGE_CACHE_LRU - struct pagevec lru_pvec; -#endif - struct lu_dirpage *dp; - int max_pages = ll_i2sbi(inode)->ll_md_brw_size >> CFS_PAGE_SHIFT; - int nrdpgs = 0; /* number of pages read actually */ - int npages; - int i; - int rc; - ENTRY; - - CDEBUG(D_VFSTRACE, "VFS Op:inode=%lu/%u(%p) hash "LPU64"\n", - inode->i_ino, inode->i_generation, inode, hash); - - LASSERT(max_pages > 0 && max_pages <= PTLRPC_MAX_BRW_PAGES); - - OBD_ALLOC(page_pool, sizeof(page) * max_pages); - if (page_pool != NULL) { - page_pool[0] = page0; - } else { - page_pool = &page0; - max_pages = 1; - } - for (npages = 1; npages < max_pages; npages++) { - page = page_cache_alloc_cold(inode->i_mapping); - if (!page) - break; - page_pool[npages] = page; - } - - op_data = ll_prep_md_op_data(NULL, inode, NULL, NULL, 0, 0, - LUSTRE_OPC_ANY, NULL); - op_data->op_npages = npages; - op_data->op_offset = hash; - rc = md_readpage(exp, op_data, page_pool, &request); - ll_finish_md_op_data(op_data); - if (rc == 0) { - body = req_capsule_server_get(&request->rq_pill, &RMF_MDT_BODY); - /* Checked by mdc_readpage() */ - LASSERT(body != NULL); - - if (body->valid & OBD_MD_FLSIZE) - cl_isize_write(inode, body->size); - - nrdpgs = (request->rq_bulk->bd_nob_transferred+CFS_PAGE_SIZE-1) - >> CFS_PAGE_SHIFT; - SetPageUptodate(page0); - } - unlock_page(page0); - ptlrpc_req_finished(request); - - CDEBUG(D_VFSTRACE, "read %d/%d pages\n", nrdpgs, npages); - - ll_pagevec_init(&lru_pvec, 0); - for (i = 1; i < npages; i++) { - unsigned long offset; - int ret; - - page = page_pool[i]; - - if (rc < 0 || i >= nrdpgs) { - page_cache_release(page); - continue; - } - - SetPageUptodate(page); - - dp = cfs_kmap(page); - hash = le64_to_cpu(dp->ldp_hash_start); - cfs_kunmap(page); - - offset = hash_x_index(hash, hash64); - - prefetchw(&page->flags); - ret = ll_add_to_page_cache_lru(page, inode->i_mapping, offset, - GFP_KERNEL); - if (ret == 0) { - unlock_page(page); - if (ll_pagevec_add(&lru_pvec, page) == 0) - ll_pagevec_lru_add_file(&lru_pvec); - } else { - CDEBUG(D_VFSTRACE, "page %lu add to page cache failed:" - " %d\n", offset, ret); - } - page_cache_release(page); - } - ll_pagevec_lru_add_file(&lru_pvec); + struct md_callback cb_op; + struct page *page; + int rc; - if (page_pool != &page0) - OBD_FREE(page_pool, sizeof(struct page *) * max_pages); - EXIT; - return rc; -} + cb_op.md_blocking_ast = ll_md_blocking_ast; + rc = md_read_page(ll_i2mdexp(dir), op_data, &cb_op, offset, &page); + if (rc != 0) + return ERR_PTR(rc); -static void ll_check_page(struct inode *dir, struct page *page) -{ - /* XXX: check page format later */ - SetPageChecked(page); + return page; } -void ll_release_page(struct page *page, int remove) +void ll_release_page(struct inode *inode, struct page *page, + bool remove) { - kunmap(page); - if (remove) { - lock_page(page); - if (likely(page->mapping != NULL)) - truncate_complete_page(page->mapping, page); - unlock_page(page); - } - page_cache_release(page); -} - -/* - * Find, kmap and return page that contains given hash. - */ -static struct page *ll_dir_page_locate(struct inode *dir, __u64 *hash, - __u64 *start, __u64 *end) -{ - int hash64 = ll_i2sbi(dir)->ll_flags & LL_SBI_64BIT_HASH; - struct address_space *mapping = dir->i_mapping; - /* - * Complement of hash is used as an index so that - * radix_tree_gang_lookup() can be used to find a page with starting - * hash _smaller_ than one we are looking for. - */ - unsigned long offset = hash_x_index(*hash, hash64); - struct page *page; - int found; - - TREE_READ_LOCK_IRQ(mapping); - found = radix_tree_gang_lookup(&mapping->page_tree, - (void **)&page, offset, 1); - if (found > 0) { - struct lu_dirpage *dp; - - page_cache_get(page); - TREE_READ_UNLOCK_IRQ(mapping); - /* - * In contrast to find_lock_page() we are sure that directory - * page cannot be truncated (while DLM lock is held) and, - * hence, can avoid restart. - * - * In fact, page cannot be locked here at all, because - * ll_dir_filler() does synchronous io. - */ - wait_on_page(page); - if (PageUptodate(page)) { - dp = cfs_kmap(page); - if (BITS_PER_LONG == 32 && hash64) { - *start = le64_to_cpu(dp->ldp_hash_start) >> 32; - *end = le64_to_cpu(dp->ldp_hash_end) >> 32; - *hash = *hash >> 32; - } else { - *start = le64_to_cpu(dp->ldp_hash_start); - *end = le64_to_cpu(dp->ldp_hash_end); - } - LASSERTF(*start <= *hash, "start = "LPX64",end = " - LPX64",hash = "LPX64"\n", *start, *end, *hash); - CDEBUG(D_VFSTRACE, "page %lu [%llu %llu], hash "LPU64"\n", - offset, *start, *end, *hash); - if (*hash > *end) { - ll_release_page(page, 0); - page = NULL; - } else if (*end != *start && *hash == *end) { - /* - * upon hash collision, remove this page, - * otherwise put page reference, and - * ll_get_dir_page() will issue RPC to fetch - * the page we want. - */ - ll_release_page(page, - le32_to_cpu(dp->ldp_flags) & LDF_COLLIDE); - page = NULL; - } - } else { - page_cache_release(page); - page = ERR_PTR(-EIO); - } + kunmap(page); + + /* Always remove the page for striped dir, because the page is + * built from temporarily in LMV layer */ + if (inode != NULL && S_ISDIR(inode->i_mode) && + ll_i2info(inode)->lli_lsm_md != NULL) { + __free_page(page); + return; + } - } else { - TREE_READ_UNLOCK_IRQ(mapping); - page = NULL; - } - return page; + if (remove) { + lock_page(page); + if (likely(page->mapping != NULL)) + truncate_complete_page(page->mapping, page); + unlock_page(page); + } + page_cache_release(page); } -struct page *ll_get_dir_page(struct inode *dir, __u64 hash, - struct ll_dir_chain *chain) +#ifdef HAVE_DIR_CONTEXT +int ll_dir_read(struct inode *inode, __u64 *ppos, struct md_op_data *op_data, + struct dir_context *ctx) { - ldlm_policy_data_t policy = {.l_inodebits = {MDS_INODELOCK_UPDATE} }; - struct address_space *mapping = dir->i_mapping; - struct lustre_handle lockh; - struct lu_dirpage *dp; - struct page *page; - ldlm_mode_t mode; - int rc; - __u64 start = 0; - __u64 end = 0; - __u64 lhash = hash; - struct ll_inode_info *lli = ll_i2info(dir); - int hash64 = ll_i2sbi(dir)->ll_flags & LL_SBI_64BIT_HASH; - - mode = LCK_PR; - rc = md_lock_match(ll_i2sbi(dir)->ll_md_exp, LDLM_FL_BLOCK_GRANTED, - ll_inode2fid(dir), LDLM_IBITS, &policy, mode, &lockh); - if (!rc) { - struct ldlm_enqueue_info einfo = { LDLM_IBITS, mode, - ll_md_blocking_ast, ldlm_completion_ast, - NULL, NULL, dir }; - struct lookup_intent it = { .it_op = IT_READDIR }; - struct ptlrpc_request *request; - struct md_op_data *op_data; - - op_data = ll_prep_md_op_data(NULL, dir, NULL, NULL, 0, 0, - LUSTRE_OPC_ANY, NULL); - if (IS_ERR(op_data)) - return (void *)op_data; - - rc = md_enqueue(ll_i2sbi(dir)->ll_md_exp, &einfo, &it, - op_data, &lockh, NULL, 0, NULL, 0); - - ll_finish_md_op_data(op_data); - - request = (struct ptlrpc_request *)it.d.lustre.it_data; - if (request) - ptlrpc_req_finished(request); - if (rc < 0) { - CERROR("lock enqueue: "DFID" at "LPU64": rc %d\n", - PFID(ll_inode2fid(dir)), hash, rc); - return ERR_PTR(rc); - } - } else { - /* for cross-ref object, l_ast_data of the lock may not be set, - * we reset it here */ - md_set_lock_data(ll_i2sbi(dir)->ll_md_exp, &lockh.cookie, - dir, NULL); - } - ldlm_lock_dump_handle(D_OTHER, &lockh); - - mutex_lock(&lli->lli_readdir_mutex); - page = ll_dir_page_locate(dir, &lhash, &start, &end); - if (IS_ERR(page)) { - CERROR("dir page locate: "DFID" at "LPU64": rc %ld\n", - PFID(ll_inode2fid(dir)), lhash, PTR_ERR(page)); - GOTO(out_unlock, page); - } else if (page != NULL) { - /* - * XXX nikita: not entirely correct handling of a corner case: - * suppose hash chain of entries with hash value HASH crosses - * border between pages P0 and P1. First both P0 and P1 are - * cached, seekdir() is called for some entry from the P0 part - * of the chain. Later P0 goes out of cache. telldir(HASH) - * happens and finds P1, as it starts with matching hash - * value. Remaining entries from P0 part of the chain are - * skipped. (Is that really a bug?) - * - * Possible solutions: 0. don't cache P1 is such case, handle - * it as an "overflow" page. 1. invalidate all pages at - * once. 2. use HASH|1 as an index for P1. - */ - GOTO(hash_collision, page); - } - - page = read_cache_page(mapping, hash_x_index(hash, hash64), - ll_dir_filler, &lhash); - if (IS_ERR(page)) { - CERROR("read cache page: "DFID" at "LPU64": rc %ld\n", - PFID(ll_inode2fid(dir)), hash, PTR_ERR(page)); - GOTO(out_unlock, page); - } - - wait_on_page(page); - (void)kmap(page); - if (!PageUptodate(page)) { - CERROR("page not updated: "DFID" at "LPU64": rc %d\n", - PFID(ll_inode2fid(dir)), hash, -5); - goto fail; - } - if (!PageChecked(page)) - ll_check_page(dir, page); - if (PageError(page)) { - CERROR("page error: "DFID" at "LPU64": rc %d\n", - PFID(ll_inode2fid(dir)), hash, -5); - goto fail; - } -hash_collision: - dp = page_address(page); - if (BITS_PER_LONG == 32 && hash64) { - start = le64_to_cpu(dp->ldp_hash_start) >> 32; - end = le64_to_cpu(dp->ldp_hash_end) >> 32; - lhash = hash >> 32; - } else { - start = le64_to_cpu(dp->ldp_hash_start); - end = le64_to_cpu(dp->ldp_hash_end); - lhash = hash; - } - if (end == start) { - LASSERT(start == lhash); - CWARN("Page-wide hash collision: "LPU64"\n", end); - if (BITS_PER_LONG == 32 && hash64) - CWARN("Real page-wide hash collision at ["LPU64" "LPU64 - "] with hash "LPU64"\n", - le64_to_cpu(dp->ldp_hash_start), - le64_to_cpu(dp->ldp_hash_end), hash); - /* - * Fetch whole overflow chain... - * - * XXX not yet. - */ - goto fail; - } -out_unlock: - mutex_unlock(&lli->lli_readdir_mutex); - ldlm_lock_decref(&lockh, mode); - return page; - -fail: - ll_release_page(page, 1); - page = ERR_PTR(-EIO); - goto out_unlock; -} - -int ll_dir_read(struct inode *inode, __u64 *_pos, void *cookie, - filldir_t filldir) +#else +int ll_dir_read(struct inode *inode, __u64 *ppos, struct md_op_data *op_data, + void *cookie, filldir_t filldir) { - struct ll_inode_info *info = ll_i2info(inode); - struct ll_sb_info *sbi = ll_i2sbi(inode); - __u64 pos = *_pos; - int api32 = ll_need_32bit_api(sbi); - int hash64 = sbi->ll_flags & LL_SBI_64BIT_HASH; - struct page *page; - struct ll_dir_chain chain; - int done = 0; +#endif + struct ll_sb_info *sbi = ll_i2sbi(inode); + __u64 pos = *ppos; + bool is_api32 = ll_need_32bit_api(sbi); + bool is_hash64 = sbi->ll_flags & LL_SBI_64BIT_HASH; + struct page *page; + struct ll_dir_chain chain; + bool done = false; int rc = 0; - ENTRY; - - ll_dir_chain_init(&chain); - - page = ll_get_dir_page(inode, pos, &chain); - - while (rc == 0 && !done) { - struct lu_dirpage *dp; - struct lu_dirent *ent; - - if (!IS_ERR(page)) { - /* - * If page is empty (end of directory is reached), - * use this value. - */ - __u64 hash = MDS_DIR_END_OFF; - __u64 next; - - dp = page_address(page); - for (ent = lu_dirent_start(dp); ent != NULL && !done; - ent = lu_dirent_next(ent)) { - __u16 type; - int namelen; - struct lu_fid fid; - __u64 lhash; - __u64 ino; - - /* - * XXX: implement correct swabbing here. - */ - - hash = le64_to_cpu(ent->lde_hash); - if (hash < pos) - /* - * Skip until we find target hash - * value. - */ - continue; - - namelen = le16_to_cpu(ent->lde_namelen); - if (namelen == 0) - /* - * Skip dummy record. - */ - continue; - - if (api32 && hash64) - lhash = hash >> 32; - else - lhash = hash; - fid_le_to_cpu(&fid, &ent->lde_fid); - ino = cl_fid_build_ino(&fid, api32); - type = ll_dirent_type_get(ent); - /* For 'll_nfs_get_name_filldir()', it will try - * to access the 'ent' through its 'lde_name', - * so the parameter 'name' for 'filldir()' must - * be part of the 'ent'. */ - done = filldir(cookie, ent->lde_name, namelen, - lhash, ino, type); - } - next = le64_to_cpu(dp->ldp_hash_end); - if (!done) { - pos = next; - if (pos == MDS_DIR_END_OFF) { - /* - * End of directory reached. - */ - done = 1; - ll_release_page(page, 0); - } else if (1 /* chain is exhausted*/) { - /* - * Normal case: continue to the next - * page. - */ - ll_release_page(page, - le32_to_cpu(dp->ldp_flags) & - LDF_COLLIDE); - next = pos; - page = ll_get_dir_page(inode, pos, - &chain); - } else { - /* - * go into overflow page. - */ - LASSERT(le32_to_cpu(dp->ldp_flags) & - LDF_COLLIDE); - ll_release_page(page, 1); - } - } else { - pos = hash; - ll_release_page(page, 0); - } - } else { - rc = PTR_ERR(page); - CERROR("error reading dir "DFID" at %lu: rc %d\n", - PFID(&info->lli_fid), (unsigned long)pos, rc); - } - } + ENTRY; - *_pos = pos; + ll_dir_chain_init(&chain); + + page = ll_get_dir_page(inode, op_data, pos, &chain); + + while (rc == 0 && !done) { + struct lu_dirpage *dp; + struct lu_dirent *ent; + __u64 hash; + __u64 next; + + if (IS_ERR(page)) { + rc = PTR_ERR(page); + break; + } + + hash = MDS_DIR_END_OFF; + dp = page_address(page); + for (ent = lu_dirent_start(dp); ent != NULL && !done; + ent = lu_dirent_next(ent)) { + __u16 type; + int namelen; + struct lu_fid fid; + __u64 lhash; + __u64 ino; + + hash = le64_to_cpu(ent->lde_hash); + if (hash < pos) + /* + * Skip until we find target hash + * value. + */ + continue; + + namelen = le16_to_cpu(ent->lde_namelen); + if (namelen == 0) + /* + * Skip dummy record. + */ + continue; + + if (is_api32 && is_hash64) + lhash = hash >> 32; + else + lhash = hash; + fid_le_to_cpu(&fid, &ent->lde_fid); + ino = cl_fid_build_ino(&fid, is_api32); + type = ll_dirent_type_get(ent); + /* For 'll_nfs_get_name_filldir()', it will try + * to access the 'ent' through its 'lde_name', + * so the parameter 'name' for 'filldir()' must + * be part of the 'ent'. */ +#ifdef HAVE_DIR_CONTEXT + ctx->pos = lhash; + done = !dir_emit(ctx, ent->lde_name, namelen, ino, + type); +#else + done = filldir(cookie, ent->lde_name, namelen, lhash, + ino, type); +#endif + } + + if (done) { + pos = hash; + ll_release_page(inode, page, false); + break; + } + + next = le64_to_cpu(dp->ldp_hash_end); + pos = next; + if (pos == MDS_DIR_END_OFF) { + /* + * End of directory reached. + */ + done = 1; + ll_release_page(inode, page, false); + } else { + /* + * Normal case: continue to the next + * page. + */ + ll_release_page(inode, page, + le32_to_cpu(dp->ldp_flags) & + LDF_COLLIDE); + next = pos; + page = ll_get_dir_page(inode, op_data, pos, + &chain); + } + } +#ifdef HAVE_DIR_CONTEXT + ctx->pos = pos; +#else + *ppos = pos; +#endif ll_dir_chain_fini(&chain); RETURN(rc); } +#ifdef HAVE_DIR_CONTEXT +static int ll_iterate(struct file *filp, struct dir_context *ctx) +#else static int ll_readdir(struct file *filp, void *cookie, filldir_t filldir) +#endif { struct inode *inode = filp->f_dentry->d_inode; struct ll_file_data *lfd = LUSTRE_FPRIVATE(filp); struct ll_sb_info *sbi = ll_i2sbi(inode); - __u64 pos = lfd->lfd_pos; int hash64 = sbi->ll_flags & LL_SBI_64BIT_HASH; int api32 = ll_need_32bit_api(sbi); + struct md_op_data *op_data; + __u64 pos; int rc; -#ifdef HAVE_TOUCH_ATIME_1ARG - struct path path; -#endif ENTRY; - CDEBUG(D_VFSTRACE, "VFS Op:inode=%lu/%u(%p) pos %lu/%llu " - " 32bit_api %d\n", inode->i_ino, inode->i_generation, + if (lfd != NULL) + pos = lfd->lfd_pos; + else + pos = 0; + + CDEBUG(D_VFSTRACE, "VFS Op:inode="DFID"(%p) pos/size" + "%lu/%llu 32bit_api %d\n", PFID(ll_inode2fid(inode)), inode, (unsigned long)pos, i_size_read(inode), api32); if (pos == MDS_DIR_END_OFF) @@ -613,27 +323,61 @@ static int ll_readdir(struct file *filp, void *cookie, filldir_t filldir) */ GOTO(out, rc = 0); - rc = ll_dir_read(inode, &pos, cookie, filldir); - lfd->lfd_pos = pos; - if (pos == MDS_DIR_END_OFF) { - if (api32) - filp->f_pos = LL_DIR_END_OFF_32BIT; - else - filp->f_pos = LL_DIR_END_OFF; - } else { - if (api32 && hash64) - filp->f_pos = pos >> 32; - else - filp->f_pos = pos; - } - filp->f_version = inode->i_version; -#ifdef HAVE_TOUCH_ATIME_1ARG - path.mnt = filp->f_vfsmnt; - path.dentry = filp->f_dentry; - touch_atime(&path); + op_data = ll_prep_md_op_data(NULL, inode, inode, NULL, 0, 0, + LUSTRE_OPC_ANY, inode); + if (IS_ERR(op_data)) + GOTO(out, rc = PTR_ERR(op_data)); + + if (unlikely(op_data->op_mea1 != NULL)) { + /* This is only needed for striped dir to fill .., + * see lmv_read_entry */ + if (filp->f_dentry->d_parent != NULL && + filp->f_dentry->d_parent->d_inode != NULL) { + __u64 ibits = MDS_INODELOCK_UPDATE; + struct inode *parent = + filp->f_dentry->d_parent->d_inode; + + if (ll_have_md_lock(parent, &ibits, LCK_MINMODE)) + op_data->op_fid3 = *ll_inode2fid(parent); + } + + /* If it can not find in cache, do lookup .. on the master + * object */ + if (fid_is_zero(&op_data->op_fid3)) { + rc = ll_dir_get_parent_fid(inode, &op_data->op_fid3); + if (rc != 0) { + ll_finish_md_op_data(op_data); + RETURN(rc); + } + } + } + op_data->op_max_pages = sbi->ll_md_brw_pages; +#ifdef HAVE_DIR_CONTEXT + ctx->pos = pos; + rc = ll_dir_read(inode, &pos, op_data, ctx); + pos = ctx->pos; +#else + rc = ll_dir_read(inode, &pos, op_data, cookie, filldir); +#endif + if (lfd != NULL) + lfd->lfd_pos = pos; + + if (pos == MDS_DIR_END_OFF) { + if (api32) + pos = LL_DIR_END_OFF_32BIT; + else + pos = LL_DIR_END_OFF; + } else { + if (api32 && hash64) + pos = pos >> 32; + } +#ifdef HAVE_DIR_CONTEXT + ctx->pos = pos; #else - touch_atime(filp->f_vfsmnt, filp->f_dentry); + filp->f_pos = pos; #endif + ll_finish_md_op_data(op_data); + filp->f_version = inode->i_version; out: if (!rc) @@ -642,7 +386,7 @@ out: RETURN(rc); } -int ll_send_mgc_param(struct obd_export *mgc, char *string) +static int ll_send_mgc_param(struct obd_export *mgc, char *string) { struct mgs_send_param *msp; int rc = 0; @@ -651,7 +395,7 @@ int ll_send_mgc_param(struct obd_export *mgc, char *string) if (!msp) return -ENOMEM; - strncpy(msp->mgs_param, string, MGS_PARAM_MAXLEN); + strlcpy(msp->mgs_param, string, sizeof(msp->mgs_param)); rc = obd_set_info_async(NULL, mgc, sizeof(KEY_SET_INFO), KEY_SET_INFO, sizeof(struct mgs_send_param), msp, NULL); if (rc) @@ -661,6 +405,59 @@ int ll_send_mgc_param(struct obd_export *mgc, char *string) return rc; } +/** + * Create striped directory with specified stripe(@lump) + * + * param[in]parent the parent of the directory. + * param[in]lump the specified stripes. + * param[in]dirname the name of the directory. + * param[in]mode the specified mode of the directory. + * + * retval =0 if striped directory is being created successfully. + * <0 if the creation is failed. + */ +static int ll_dir_setdirstripe(struct inode *parent, struct lmv_user_md *lump, + const char *dirname, umode_t mode) +{ + struct ptlrpc_request *request = NULL; + struct md_op_data *op_data; + struct ll_sb_info *sbi = ll_i2sbi(parent); + int err; + ENTRY; + + if (unlikely(lump->lum_magic != LMV_USER_MAGIC)) + RETURN(-EINVAL); + + CDEBUG(D_VFSTRACE, "VFS Op:inode="DFID"(%p) name %s" + "stripe_offset %d, stripe_count: %u\n", + PFID(ll_inode2fid(parent)), parent, dirname, + (int)lump->lum_stripe_offset, lump->lum_stripe_count); + + if (lump->lum_magic != cpu_to_le32(LMV_USER_MAGIC)) + lustre_swab_lmv_user_md(lump); + + if (!IS_POSIXACL(parent) || !exp_connect_umask(ll_i2mdexp(parent))) + mode &= ~current_umask(); + mode = (mode & (S_IRWXUGO | S_ISVTX)) | S_IFDIR; + op_data = ll_prep_md_op_data(NULL, parent, NULL, dirname, + strlen(dirname), mode, LUSTRE_OPC_MKDIR, + lump); + if (IS_ERR(op_data)) + GOTO(err_exit, err = PTR_ERR(op_data)); + + op_data->op_cli_flags |= CLI_SET_MEA; + err = md_create(sbi->ll_md_exp, op_data, lump, sizeof(*lump), mode, + from_kuid(&init_user_ns, current_fsuid()), + from_kgid(&init_user_ns, current_fsgid()), + cfs_curproc_cap_pack(), 0, &request); + ll_finish_md_op_data(op_data); + if (err) + GOTO(err_exit, err); +err_exit: + ptlrpc_req_finished(request); + return err; +} + int ll_dir_setstripe(struct inode *inode, struct lov_user_md *lump, int set_default) { @@ -685,22 +482,29 @@ int ll_dir_setstripe(struct inode *inode, struct lov_user_md *lump, lustre_swab_lov_user_md_v1(lump); lum_size = sizeof(struct lov_user_md_v1); break; - } + } case LOV_USER_MAGIC_V3: { if (lump->lmm_magic != cpu_to_le32(LOV_USER_MAGIC_V3)) lustre_swab_lov_user_md_v3( (struct lov_user_md_v3 *)lump); lum_size = sizeof(struct lov_user_md_v3); break; - } + } + case LMV_USER_MAGIC: { + if (lump->lmm_magic != cpu_to_le32(LMV_USER_MAGIC)) + lustre_swab_lmv_user_md( + (struct lmv_user_md *)lump); + lum_size = sizeof(struct lmv_user_md); + break; + } default: { CDEBUG(D_IOCTL, "bad userland LOV MAGIC:" " %#08x != %#08x nor %#08x\n", lump->lmm_magic, LOV_USER_MAGIC_V1, LOV_USER_MAGIC_V3); RETURN(-EINVAL); - } - } + } + } } else { lum_size = sizeof(struct lov_user_md_v1); } @@ -764,74 +568,112 @@ end: RETURN(rc); } -int ll_dir_getstripe(struct inode *inode, struct lov_mds_md **lmmp, - int *lmm_size, struct ptlrpc_request **request) +/** + * This function will be used to get default LOV/LMV/Default LMV + * @valid will be used to indicate which stripe it will retrieve + * OBD_MD_MEA LMV stripe EA + * OBD_MD_DEFAULT_MEA Default LMV stripe EA + * otherwise Default LOV EA. + * Each time, it can only retrieve 1 stripe EA + **/ +int ll_dir_getstripe(struct inode *inode, void **plmm, int *plmm_size, + struct ptlrpc_request **request, obd_valid valid) { - struct ll_sb_info *sbi = ll_i2sbi(inode); - struct mdt_body *body; - struct lov_mds_md *lmm = NULL; - struct ptlrpc_request *req = NULL; - int rc, lmmsize; - struct md_op_data *op_data; + struct ll_sb_info *sbi = ll_i2sbi(inode); + struct mdt_body *body; + struct lov_mds_md *lmm = NULL; + struct ptlrpc_request *req = NULL; + int rc, lmm_size; + struct md_op_data *op_data; + ENTRY; - rc = ll_get_max_mdsize(sbi, &lmmsize); - if (rc) - RETURN(rc); + rc = ll_get_default_mdsize(sbi, &lmm_size); + if (rc) + RETURN(rc); + + op_data = ll_prep_md_op_data(NULL, inode, NULL, NULL, + 0, lmm_size, LUSTRE_OPC_ANY, + NULL); + if (IS_ERR(op_data)) + RETURN(PTR_ERR(op_data)); + + op_data->op_valid = valid | OBD_MD_FLEASIZE | OBD_MD_FLDIREA; + rc = md_getattr(sbi->ll_md_exp, op_data, &req); + ll_finish_md_op_data(op_data); + if (rc < 0) { + CDEBUG(D_INFO, "md_getattr failed on inode " + DFID": rc %d\n", PFID(ll_inode2fid(inode)), rc); + GOTO(out, rc); + } - op_data = ll_prep_md_op_data(NULL, inode, NULL, NULL, - 0, lmmsize, LUSTRE_OPC_ANY, - NULL); - if (IS_ERR(op_data)) - RETURN(PTR_ERR(op_data)); + body = req_capsule_server_get(&req->rq_pill, &RMF_MDT_BODY); + LASSERT(body != NULL); - op_data->op_valid = OBD_MD_FLEASIZE | OBD_MD_FLDIREA; - rc = md_getattr(sbi->ll_md_exp, op_data, &req); - ll_finish_md_op_data(op_data); - if (rc < 0) { - CDEBUG(D_INFO, "md_getattr failed on inode " - "%lu/%u: rc %d\n", inode->i_ino, - inode->i_generation, rc); - GOTO(out, rc); - } + lmm_size = body->mbo_eadatasize; - body = req_capsule_server_get(&req->rq_pill, &RMF_MDT_BODY); - LASSERT(body != NULL); + if (!(body->mbo_valid & (OBD_MD_FLEASIZE | OBD_MD_FLDIREA)) || + lmm_size == 0) { + GOTO(out, rc = -ENODATA); + } - lmmsize = body->eadatasize; + lmm = req_capsule_server_sized_get(&req->rq_pill, + &RMF_MDT_MD, lmm_size); + LASSERT(lmm != NULL); + + /* + * This is coming from the MDS, so is probably in + * little endian. We convert it to host endian before + * passing it to userspace. + */ + /* We don't swab objects for directories */ + switch (le32_to_cpu(lmm->lmm_magic)) { + case LOV_MAGIC_V1: + if (LOV_MAGIC != cpu_to_le32(LOV_MAGIC)) + lustre_swab_lov_user_md_v1((struct lov_user_md_v1 *)lmm); + break; + case LOV_MAGIC_V3: + if (LOV_MAGIC != cpu_to_le32(LOV_MAGIC)) + lustre_swab_lov_user_md_v3((struct lov_user_md_v3 *)lmm); + break; + case LMV_MAGIC_V1: + if (LMV_MAGIC != cpu_to_le32(LMV_MAGIC)) + lustre_swab_lmv_mds_md((union lmv_mds_md *)lmm); + break; + case LMV_USER_MAGIC: + if (LMV_USER_MAGIC != cpu_to_le32(LMV_USER_MAGIC)) + lustre_swab_lmv_user_md((struct lmv_user_md *)lmm); + break; + default: + CERROR("unknown magic: %lX\n", (unsigned long)lmm->lmm_magic); + rc = -EPROTO; + } +out: + *plmm = lmm; + *plmm_size = lmm_size; + *request = req; + return rc; +} - if (!(body->valid & (OBD_MD_FLEASIZE | OBD_MD_FLDIREA)) || - lmmsize == 0) { - GOTO(out, rc = -ENODATA); - } +int ll_get_mdt_idx_by_fid(struct ll_sb_info *sbi, const struct lu_fid *fid) +{ + struct md_op_data *op_data; + int rc; + int mdt_index; + ENTRY; - lmm = req_capsule_server_sized_get(&req->rq_pill, - &RMF_MDT_MD, lmmsize); - LASSERT(lmm != NULL); - - /* - * This is coming from the MDS, so is probably in - * little endian. We convert it to host endian before - * passing it to userspace. - */ - /* We don't swab objects for directories */ - switch (le32_to_cpu(lmm->lmm_magic)) { - case LOV_MAGIC_V1: - if (LOV_MAGIC != cpu_to_le32(LOV_MAGIC)) - lustre_swab_lov_user_md_v1((struct lov_user_md_v1 *)lmm); - break; - case LOV_MAGIC_V3: - if (LOV_MAGIC != cpu_to_le32(LOV_MAGIC)) - lustre_swab_lov_user_md_v3((struct lov_user_md_v3 *)lmm); - break; - default: - CERROR("unknown magic: %lX\n", (unsigned long)lmm->lmm_magic); - rc = -EPROTO; - } -out: - *lmmp = lmm; - *lmm_size = lmmsize; - *request = req; - return rc; + OBD_ALLOC_PTR(op_data); + if (op_data == NULL) + RETURN(-ENOMEM); + + op_data->op_flags |= MF_GET_MDT_IDX; + op_data->op_fid1 = *fid; + rc = md_getattr(sbi->ll_md_exp, op_data, NULL); + mdt_index = op_data->op_mds; + OBD_FREE_PTR(op_data); + if (rc < 0) + RETURN(rc); + + RETURN(mdt_index); } /* @@ -839,42 +681,201 @@ out: */ int ll_get_mdt_idx(struct inode *inode) { - struct ll_sb_info *sbi = ll_i2sbi(inode); - struct md_op_data *op_data; - int rc, mdtidx; - ENTRY; + return ll_get_mdt_idx_by_fid(ll_i2sbi(inode), ll_inode2fid(inode)); +} - op_data = ll_prep_md_op_data(NULL, inode, NULL, NULL, 0, - 0, LUSTRE_OPC_ANY, NULL); - if (IS_ERR(op_data)) - RETURN(PTR_ERR(op_data)); +/** + * Generic handler to do any pre-copy work. + * + * It send a first hsm_progress (with extent length == 0) to coordinator as a + * first information for it that real work has started. + * + * Moreover, for a ARCHIVE request, it will sample the file data version and + * store it in \a copy. + * + * \return 0 on success. + */ +static int ll_ioc_copy_start(struct super_block *sb, struct hsm_copy *copy) +{ + struct ll_sb_info *sbi = ll_s2sbi(sb); + struct hsm_progress_kernel hpk; + int rc; + ENTRY; - op_data->op_flags |= MF_GET_MDT_IDX; - rc = md_getattr(sbi->ll_md_exp, op_data, NULL); - mdtidx = op_data->op_mds; - ll_finish_md_op_data(op_data); - if (rc < 0) { - CDEBUG(D_INFO, "md_getattr_name: %d\n", rc); - RETURN(rc); - } - return mdtidx; + /* Forge a hsm_progress based on data from copy. */ + hpk.hpk_fid = copy->hc_hai.hai_fid; + hpk.hpk_cookie = copy->hc_hai.hai_cookie; + hpk.hpk_extent.offset = copy->hc_hai.hai_extent.offset; + hpk.hpk_extent.length = 0; + hpk.hpk_flags = 0; + hpk.hpk_errval = 0; + hpk.hpk_data_version = 0; + + + /* For archive request, we need to read the current file version. */ + if (copy->hc_hai.hai_action == HSMA_ARCHIVE) { + struct inode *inode; + __u64 data_version = 0; + + /* Get inode for this fid */ + inode = search_inode_for_lustre(sb, ©->hc_hai.hai_fid); + if (IS_ERR(inode)) { + hpk.hpk_flags |= HP_FLAG_RETRY; + /* hpk_errval is >= 0 */ + hpk.hpk_errval = -PTR_ERR(inode); + GOTO(progress, rc = PTR_ERR(inode)); + } + + /* Read current file data version */ + rc = ll_data_version(inode, &data_version, LL_DV_RD_FLUSH); + iput(inode); + if (rc != 0) { + CDEBUG(D_HSM, "Could not read file data version of " + DFID" (rc = %d). Archive request (" + LPX64") could not be done.\n", + PFID(©->hc_hai.hai_fid), rc, + copy->hc_hai.hai_cookie); + hpk.hpk_flags |= HP_FLAG_RETRY; + /* hpk_errval must be >= 0 */ + hpk.hpk_errval = -rc; + GOTO(progress, rc); + } + + /* Store it the hsm_copy for later copytool use. + * Always modified even if no lsm. */ + copy->hc_data_version = data_version; + } + +progress: + /* On error, the request should be considered as completed */ + if (hpk.hpk_errval > 0) + hpk.hpk_flags |= HP_FLAG_COMPLETED; + rc = obd_iocontrol(LL_IOC_HSM_PROGRESS, sbi->ll_md_exp, sizeof(hpk), + &hpk, NULL); + + RETURN(rc); } -static int copy_and_ioctl(int cmd, struct obd_export *exp, void *data, int len) +/** + * Generic handler to do any post-copy work. + * + * It will send the last hsm_progress update to coordinator to inform it + * that copy is finished and whether it was successful or not. + * + * Moreover, + * - for ARCHIVE request, it will sample the file data version and compare it + * with the version saved in ll_ioc_copy_start(). If they do not match, copy + * will be considered as failed. + * - for RESTORE request, it will sample the file data version and send it to + * coordinator which is useful if the file was imported as 'released'. + * + * \return 0 on success. + */ +static int ll_ioc_copy_end(struct super_block *sb, struct hsm_copy *copy) { - void *ptr; - int rc; + struct ll_sb_info *sbi = ll_s2sbi(sb); + struct hsm_progress_kernel hpk; + int rc; + ENTRY; - OBD_ALLOC(ptr, len); - if (ptr == NULL) - return -ENOMEM; - if (cfs_copy_from_user(ptr, data, len)) { - OBD_FREE(ptr, len); - return -EFAULT; - } - rc = obd_iocontrol(cmd, exp, len, data, NULL); - OBD_FREE(ptr, len); - return rc; + /* If you modify the logic here, also check llapi_hsm_copy_end(). */ + /* Take care: copy->hc_hai.hai_action, len, gid and data are not + * initialized if copy_end was called with copy == NULL. + */ + + /* Forge a hsm_progress based on data from copy. */ + hpk.hpk_fid = copy->hc_hai.hai_fid; + hpk.hpk_cookie = copy->hc_hai.hai_cookie; + hpk.hpk_extent = copy->hc_hai.hai_extent; + hpk.hpk_flags = copy->hc_flags | HP_FLAG_COMPLETED; + hpk.hpk_errval = copy->hc_errval; + hpk.hpk_data_version = 0; + + /* For archive request, we need to check the file data was not changed. + * + * For restore request, we need to send the file data version, this is + * useful when the file was created using hsm_import. + */ + if (((copy->hc_hai.hai_action == HSMA_ARCHIVE) || + (copy->hc_hai.hai_action == HSMA_RESTORE)) && + (copy->hc_errval == 0)) { + struct inode *inode; + __u64 data_version = 0; + + /* Get lsm for this fid */ + inode = search_inode_for_lustre(sb, ©->hc_hai.hai_fid); + if (IS_ERR(inode)) { + hpk.hpk_flags |= HP_FLAG_RETRY; + /* hpk_errval must be >= 0 */ + hpk.hpk_errval = -PTR_ERR(inode); + GOTO(progress, rc = PTR_ERR(inode)); + } + + rc = ll_data_version(inode, &data_version, LL_DV_RD_FLUSH); + iput(inode); + if (rc) { + CDEBUG(D_HSM, "Could not read file data version. " + "Request could not be confirmed.\n"); + if (hpk.hpk_errval == 0) + hpk.hpk_errval = -rc; + GOTO(progress, rc); + } + + /* Store it the hsm_copy for later copytool use. + * Always modified even if no lsm. */ + hpk.hpk_data_version = data_version; + + /* File could have been stripped during archiving, so we need + * to check anyway. */ + if ((copy->hc_hai.hai_action == HSMA_ARCHIVE) && + (copy->hc_data_version != data_version)) { + CDEBUG(D_HSM, "File data version mismatched. " + "File content was changed during archiving. " + DFID", start:"LPX64" current:"LPX64"\n", + PFID(©->hc_hai.hai_fid), + copy->hc_data_version, data_version); + /* File was changed, send error to cdt. Do not ask for + * retry because if a file is modified frequently, + * the cdt will loop on retried archive requests. + * The policy engine will ask for a new archive later + * when the file will not be modified for some tunable + * time */ + /* we do not notify caller */ + hpk.hpk_flags &= ~HP_FLAG_RETRY; + /* hpk_errval must be >= 0 */ + hpk.hpk_errval = EBUSY; + } + + } + +progress: + rc = obd_iocontrol(LL_IOC_HSM_PROGRESS, sbi->ll_md_exp, sizeof(hpk), + &hpk, NULL); + + RETURN(rc); +} + + +static int copy_and_ioctl(int cmd, struct obd_export *exp, + const void __user *data, size_t size) +{ + void *copy; + int rc; + + OBD_ALLOC(copy, size); + if (copy == NULL) + return -ENOMEM; + + if (copy_from_user(copy, data, size)) { + rc = -EFAULT; + goto out; + } + + rc = obd_iocontrol(cmd, exp, size, copy, NULL); +out: + OBD_FREE(copy, size); + + return rc; } static int quotactl_ioctl(struct ll_sb_info *sbi, struct if_quotactl *qctl) @@ -897,12 +898,14 @@ static int quotactl_ioctl(struct ll_sb_info *sbi, struct if_quotactl *qctl) sbi->ll_flags & LL_SBI_RMT_CLIENT) RETURN(-EPERM); break; - case Q_GETQUOTA: - if (((type == USRQUOTA && cfs_curproc_euid() != id) || - (type == GRPQUOTA && !in_egroup_p(id))) && - (!cfs_capable(CFS_CAP_SYS_ADMIN) || - sbi->ll_flags & LL_SBI_RMT_CLIENT)) - RETURN(-EPERM); + case Q_GETQUOTA: + if (((type == USRQUOTA && + !uid_eq(current_euid(), make_kuid(&init_user_ns, id))) || + (type == GRPQUOTA && + !in_egroup_p(make_kgid(&init_user_ns, id)))) && + (!cfs_capable(CFS_CAP_SYS_ADMIN) || + sbi->ll_flags & LL_SBI_RMT_CLIENT)) + RETURN(-EPERM); break; case Q_GETINFO: break; @@ -1013,6 +1016,30 @@ out: RETURN(rc); } +static char * +ll_getname(const char __user *filename) +{ + int ret = 0, len; + char *tmp = __getname(); + + if (!tmp) + return ERR_PTR(-ENOMEM); + + len = strncpy_from_user(tmp, filename, PATH_MAX); + if (len == 0) + ret = -ENOENT; + else if (len > PATH_MAX) + ret = -ENAMETOOLONG; + + if (ret) { + __putname(tmp); + tmp = ERR_PTR(ret); + } + return tmp; +} + +#define ll_putname(filename) __putname(filename) + static long ll_dir_ioctl(struct file *file, unsigned int cmd, unsigned long arg) { struct inode *inode = file->f_dentry->d_inode; @@ -1021,8 +1048,8 @@ static long ll_dir_ioctl(struct file *file, unsigned int cmd, unsigned long arg) int rc = 0; ENTRY; - CDEBUG(D_VFSTRACE, "VFS Op:inode=%lu/%u(%p), cmd=%#x\n", - inode->i_ino, inode->i_generation, inode, cmd); + CDEBUG(D_VFSTRACE, "VFS Op:inode="DFID"(%p), cmd=%#x\n", + PFID(ll_inode2fid(inode)), inode, cmd); /* asm-ppc{,64} declares TCGETS, et. al. as type 't' not 'T' */ if (_IOC_TYPE(cmd) == 'T' || _IOC_TYPE(cmd) == 't') /* tty ioctls */ @@ -1055,43 +1082,96 @@ static long ll_dir_ioctl(struct file *file, unsigned int cmd, unsigned long arg) return 0; } case IOC_MDC_LOOKUP: { - struct ptlrpc_request *request = NULL; - int namelen, len = 0; - char *buf = NULL; - char *filename; - struct md_op_data *op_data; + int namelen, len = 0; + char *buf = NULL; + char *filename; + + rc = obd_ioctl_getdata(&buf, &len, (void *)arg); + if (rc != 0) + RETURN(rc); + data = (void *)buf; + + filename = data->ioc_inlbuf1; + namelen = strlen(filename); + if (namelen < 1) { + CDEBUG(D_INFO, "IOC_MDC_LOOKUP missing filename\n"); + GOTO(out_free, rc = -EINVAL); + } + + rc = ll_get_fid_by_name(inode, filename, namelen, NULL); + if (rc < 0) { + CERROR("%s: lookup %.*s failed: rc = %d\n", + ll_get_fsname(inode->i_sb, NULL, 0), namelen, + filename, rc); + GOTO(out_free, rc); + } +out_free: + obd_ioctl_freedata(buf, len); + return rc; + } + case LL_IOC_LMV_SETSTRIPE: { + struct lmv_user_md *lum; + char *buf = NULL; + char *filename; + int namelen = 0; + int lumlen = 0; + umode_t mode; + int len; + int rc; + + rc = obd_ioctl_getdata(&buf, &len, (void *)arg); + if (rc) + RETURN(rc); + + data = (void *)buf; + if (data->ioc_inlbuf1 == NULL || data->ioc_inlbuf2 == NULL || + data->ioc_inllen1 == 0 || data->ioc_inllen2 == 0) + GOTO(lmv_out_free, rc = -EINVAL); + + filename = data->ioc_inlbuf1; + namelen = data->ioc_inllen1; + + if (namelen < 1) { + CDEBUG(D_INFO, "IOC_MDC_LOOKUP missing filename\n"); + GOTO(lmv_out_free, rc = -EINVAL); + } + lum = (struct lmv_user_md *)data->ioc_inlbuf2; + lumlen = data->ioc_inllen2; + + if (lum->lum_magic != LMV_USER_MAGIC || + lumlen != sizeof(*lum)) { + CERROR("%s: wrong lum magic %x or size %d: rc = %d\n", + filename, lum->lum_magic, lumlen, -EFAULT); + GOTO(lmv_out_free, rc = -EINVAL); + } + +#if LUSTRE_VERSION_CODE < OBD_OCD_VERSION(2, 9, 50, 0) + mode = data->ioc_type != 0 ? data->ioc_type : S_IRWXUGO; +#else + mode = data->ioc_type; +#endif + rc = ll_dir_setdirstripe(inode, lum, filename, mode); +lmv_out_free: + obd_ioctl_freedata(buf, len); + RETURN(rc); - rc = obd_ioctl_getdata(&buf, &len, (void *)arg); - if (rc) - RETURN(rc); - data = (void *)buf; + } + case LL_IOC_LMV_SET_DEFAULT_STRIPE: { + struct lmv_user_md lum; + struct lmv_user_md __user *ulump = + (struct lmv_user_md __user *)arg; + int rc; - filename = data->ioc_inlbuf1; - namelen = strlen(filename); + if (copy_from_user(&lum, ulump, sizeof(lum))) + RETURN(-EFAULT); - if (namelen < 1) { - CDEBUG(D_INFO, "IOC_MDC_LOOKUP missing filename\n"); - GOTO(out_free, rc = -EINVAL); - } + if (lum.lum_magic != LMV_USER_MAGIC) + RETURN(-EINVAL); - op_data = ll_prep_md_op_data(NULL, inode, NULL, filename, namelen, - 0, LUSTRE_OPC_ANY, NULL); - if (IS_ERR(op_data)) - GOTO(out_free, rc = PTR_ERR(op_data)); + rc = ll_dir_setstripe(inode, (struct lov_user_md *)&lum, 0); - op_data->op_valid = OBD_MD_FLID; - rc = md_getattr_name(sbi->ll_md_exp, op_data, &request); - ll_finish_md_op_data(op_data); - if (rc < 0) { - CDEBUG(D_INFO, "md_getattr_name: %d\n", rc); - GOTO(out_free, rc); - } - ptlrpc_req_finished(request); - EXIT; -out_free: - obd_ioctl_freedata(buf, len); - return rc; - } + RETURN(rc); + } case LL_IOC_LOV_SETSTRIPE: { struct lov_user_md_v3 lumv3; struct lov_user_md_v1 *lumv1 = (struct lov_user_md_v1 *)&lumv3; @@ -1104,11 +1184,11 @@ out_free: LASSERT(sizeof(lumv3.lmm_objects[0]) == sizeof(lumv3p->lmm_objects[0])); /* first try with v1 which is smaller than v3 */ - if (cfs_copy_from_user(lumv1, lumv1p, sizeof(*lumv1))) + if (copy_from_user(lumv1, lumv1p, sizeof(*lumv1))) RETURN(-EFAULT); - if (lumv1->lmm_magic == LOV_USER_MAGIC_V3) { - if (cfs_copy_from_user(&lumv3, lumv3p, sizeof(lumv3))) + if ((lumv1->lmm_magic == LOV_USER_MAGIC_V3) ) { + if (copy_from_user(&lumv3, lumv3p, sizeof(lumv3))) RETURN(-EFAULT); } @@ -1120,8 +1200,119 @@ out_free: RETURN(rc); } - case LL_IOC_OBD_STATFS: - RETURN(ll_obd_statfs(inode, (void *)arg)); + case LL_IOC_LMV_GETSTRIPE: { + struct lmv_user_md __user *ulmv = + (struct lmv_user_md __user *)arg; + struct lmv_user_md lum; + struct ptlrpc_request *request = NULL; + union lmv_mds_md *lmm = NULL; + int lmmsize; + obd_valid valid = 0; + struct lmv_user_md *tmp = NULL; + int mdt_index; + int lum_size; + int stripe_count; + int i; + int rc; + + if (copy_from_user(&lum, ulmv, sizeof(*ulmv))) + RETURN(-EFAULT); + + /* lum_magic will indicate which stripe the ioctl will like + * to get, LMV_MAGIC_V1 is for normal LMV stripe, LMV_USER_MAGIC + * is for default LMV stripe */ + if (lum.lum_magic == LMV_MAGIC_V1) + valid |= OBD_MD_MEA; + else if (lum.lum_magic == LMV_USER_MAGIC) + valid |= OBD_MD_DEFAULT_MEA; + else + RETURN(-EINVAL); + + rc = ll_dir_getstripe(inode, (void **)&lmm, &lmmsize, &request, + valid); + if (rc != 0) + GOTO(finish_req, rc); + + /* Get default LMV EA */ + if (lum.lum_magic == LMV_USER_MAGIC) { + if (rc != 0) + GOTO(finish_req, rc); + + if (lmmsize > sizeof(*ulmv)) + GOTO(finish_req, rc = -EINVAL); + + if (copy_to_user(ulmv, lmm, lmmsize)) + GOTO(finish_req, rc = -EFAULT); + + GOTO(finish_req, rc); + } + + stripe_count = lmv_mds_md_stripe_count_get(lmm); + lum_size = lmv_user_md_size(stripe_count, LMV_MAGIC_V1); + OBD_ALLOC(tmp, lum_size); + if (tmp == NULL) + GOTO(finish_req, rc = -ENOMEM); + + mdt_index = ll_get_mdt_idx(inode); + if (mdt_index < 0) + GOTO(out_tmp, rc = -ENOMEM); + + tmp->lum_magic = LMV_MAGIC_V1; + tmp->lum_stripe_count = 0; + tmp->lum_stripe_offset = mdt_index; + for (i = 0; i < stripe_count; i++) { + struct lu_fid fid; + + fid_le_to_cpu(&fid, &lmm->lmv_md_v1.lmv_stripe_fids[i]); + mdt_index = ll_get_mdt_idx_by_fid(sbi, &fid); + if (mdt_index < 0) + GOTO(out_tmp, rc = mdt_index); + + tmp->lum_objects[i].lum_mds = mdt_index; + tmp->lum_objects[i].lum_fid = fid; + tmp->lum_stripe_count++; + } + + if (copy_to_user(ulmv, tmp, lum_size)) + GOTO(out_tmp, rc = -EFAULT); +out_tmp: + OBD_FREE(tmp, lum_size); +finish_req: + ptlrpc_req_finished(request); + return rc; + } + + case LL_IOC_REMOVE_ENTRY: { + char *filename = NULL; + int namelen = 0; + int rc; + + /* Here is a little hack to avoid sending REINT_RMENTRY to + * unsupported server, which might crash the server(LU-2730), + * Because both LVB_TYPE and REINT_RMENTRY will be supported + * on 2.4, we use OBD_CONNECT_LVB_TYPE to detect whether the + * server will support REINT_RMENTRY XXX*/ + if (!(exp_connect_flags(sbi->ll_md_exp) & OBD_CONNECT_LVB_TYPE)) + RETURN(-ENOTSUPP); + + filename = ll_getname((const char *)arg); + if (IS_ERR(filename)) + RETURN(PTR_ERR(filename)); + + namelen = strlen(filename); + if (namelen < 1) + GOTO(out_rmdir, rc = -EINVAL); + + rc = ll_rmdir_entry(inode, filename, namelen); +out_rmdir: + if (filename) + ll_putname(filename); + RETURN(rc); + } + case LL_IOC_LOV_SWAP_LAYOUTS: + RETURN(-EPERM); + case IOC_OBD_STATFS: + RETURN(ll_obd_statfs(inode, (void *)arg)); case LL_IOC_LOV_GETSTRIPE: case LL_IOC_MDC_GETINFO: case IOC_MDC_GETFILEINFO: @@ -1135,15 +1326,16 @@ out_free: if (cmd == IOC_MDC_GETFILEINFO || cmd == IOC_MDC_GETFILESTRIPE) { - filename = getname((const char *)arg); + filename = ll_getname((const char *)arg); if (IS_ERR(filename)) RETURN(PTR_ERR(filename)); rc = ll_lov_getstripe_ea_info(inode, filename, &lmm, &lmmsize, &request); - } else { - rc = ll_dir_getstripe(inode, &lmm, &lmmsize, &request); - } + } else { + rc = ll_dir_getstripe(inode, (void **)&lmm, &lmmsize, + &request, 0); + } if (request) { body = req_capsule_server_get(&request->rq_pill, @@ -1169,8 +1361,8 @@ out_free: lmdp = (struct lov_user_mds_data *)arg; lump = &lmdp->lmd_lmm; } - if (cfs_copy_to_user(lump, lmm, lmmsize)) { - if (cfs_copy_to_user(lump, lmm, sizeof(*lump))) + if (copy_to_user(lump, lmm, lmmsize)) { + if (copy_to_user(lump, lmm, sizeof(*lump))) GOTO(out_req, rc = -EFAULT); rc = -EOVERFLOW; } @@ -1179,22 +1371,22 @@ out_free: struct lov_user_mds_data *lmdp; lstat_t st = { 0 }; - st.st_dev = inode->i_sb->s_dev; - st.st_mode = body->mode; - st.st_nlink = body->nlink; - st.st_uid = body->uid; - st.st_gid = body->gid; - st.st_rdev = body->rdev; - st.st_size = body->size; - st.st_blksize = CFS_PAGE_SIZE; - st.st_blocks = body->blocks; - st.st_atime = body->atime; - st.st_mtime = body->mtime; - st.st_ctime = body->ctime; - st.st_ino = inode->i_ino; - - lmdp = (struct lov_user_mds_data *)arg; - if (cfs_copy_to_user(&lmdp->lmd_st, &st, sizeof(st))) + st.st_dev = inode->i_sb->s_dev; + st.st_mode = body->mbo_mode; + st.st_nlink = body->mbo_nlink; + st.st_uid = body->mbo_uid; + st.st_gid = body->mbo_gid; + st.st_rdev = body->mbo_rdev; + st.st_size = body->mbo_size; + st.st_blksize = PAGE_CACHE_SIZE; + st.st_blocks = body->mbo_blocks; + st.st_atime = body->mbo_atime; + st.st_mtime = body->mbo_mtime; + st.st_ctime = body->mbo_ctime; + st.st_ino = inode->i_ino; + + lmdp = (struct lov_user_mds_data *)arg; + if (copy_to_user(&lmdp->lmd_st, &st, sizeof(st))) GOTO(out_req, rc = -EFAULT); } @@ -1202,7 +1394,7 @@ out_free: out_req: ptlrpc_req_finished(request); if (filename) - putname(filename); + ll_putname(filename); return rc; } case IOC_LOV_GETINFO: { @@ -1220,9 +1412,12 @@ out_free: if (rc) RETURN(rc); - OBD_ALLOC_LARGE(lmm, lmmsize); - if (cfs_copy_from_user(lmm, lum, lmmsize)) - GOTO(free_lmm, rc = -EFAULT); + OBD_ALLOC_LARGE(lmm, lmmsize); + if (lmm == NULL) + RETURN(-ENOMEM); + + if (copy_from_user(lmm, lum, lmmsize)) + GOTO(free_lmm, rc = -EFAULT); switch (lmm->lmm_magic) { case LOV_USER_MAGIC_V1: @@ -1258,7 +1453,7 @@ out_free: if (rc) GOTO(free_lsm, rc); - if (cfs_copy_to_user(&lumd->lmd_st, &st, sizeof(st))) + if (copy_to_user(&lumd->lmd_st, &st, sizeof(st))) GOTO(free_lsm, rc = -EFAULT); EXIT; @@ -1268,10 +1463,7 @@ out_free: OBD_FREE_LARGE(lmm, lmmsize); return rc; } - case OBD_IOC_LLOG_CATINFO: { - RETURN(-EOPNOTSUPP); - } - case OBD_IOC_QUOTACHECK: { + case OBD_IOC_QUOTACHECK: { struct obd_quotactl *oqctl; int error = 0; @@ -1296,7 +1488,7 @@ out_free: OBD_FREE_PTR(oqctl); return error ?: rc; } - case OBD_IOC_POLL_QUOTACHECK: { + case OBD_IOC_POLL_QUOTACHECK: { struct if_quotacheck *check; if (!cfs_capable(CFS_CAP_SYS_ADMIN) || @@ -1311,9 +1503,9 @@ out_free: NULL); if (rc) { CDEBUG(D_QUOTA, "mdc ioctl %d failed: %d\n", cmd, rc); - if (cfs_copy_to_user((void *)arg, check, + if (copy_to_user((void *)arg, check, sizeof(*check))) - CDEBUG(D_QUOTA, "cfs_copy_to_user failed\n"); + CDEBUG(D_QUOTA, "copy_to_user failed\n"); GOTO(out_poll, rc); } @@ -1321,9 +1513,9 @@ out_free: NULL); if (rc) { CDEBUG(D_QUOTA, "osc ioctl %d failed: %d\n", cmd, rc); - if (cfs_copy_to_user((void *)arg, check, + if (copy_to_user((void *)arg, check, sizeof(*check))) - CDEBUG(D_QUOTA, "cfs_copy_to_user failed\n"); + CDEBUG(D_QUOTA, "copy_to_user failed\n"); GOTO(out_poll, rc); } out_poll: @@ -1345,7 +1537,7 @@ out_free: if (!qctl_20) GOTO(out_quotactl_18, rc = -ENOMEM); - if (cfs_copy_from_user(qctl_18, (void *)arg, sizeof(*qctl_18))) + if (copy_from_user(qctl_18, (void *)arg, sizeof(*qctl_18))) GOTO(out_quotactl_20, rc = -ENOMEM); QCTL_COPY(qctl_20, qctl_18); @@ -1370,7 +1562,7 @@ out_free: QCTL_COPY(qctl_18, qctl_20); qctl_18->obd_uuid = qctl_20->obd_uuid; - if (cfs_copy_to_user((void *)arg, qctl_18, + if (copy_to_user((void *)arg, qctl_18, sizeof(*qctl_18))) rc = -EFAULT; } @@ -1384,19 +1576,19 @@ out_free: #else #warning "remove old LL_IOC_QUOTACTL_18 compatibility code" #endif /* LUSTRE_VERSION_CODE < OBD_OCD_VERSION(2, 7, 50, 0) */ - case LL_IOC_QUOTACTL: { + case OBD_IOC_QUOTACTL: { struct if_quotactl *qctl; OBD_ALLOC_PTR(qctl); if (!qctl) RETURN(-ENOMEM); - if (cfs_copy_from_user(qctl, (void *)arg, sizeof(*qctl))) + if (copy_from_user(qctl, (void *)arg, sizeof(*qctl))) GOTO(out_quotactl, rc = -EFAULT); rc = quotactl_ioctl(sbi, qctl); - if (rc == 0 && cfs_copy_to_user((void *)arg,qctl,sizeof(*qctl))) + if (rc == 0 && copy_to_user((void *)arg, qctl, sizeof(*qctl))) rc = -EFAULT; out_quotactl: @@ -1414,11 +1606,11 @@ out_free: inode == inode->i_sb->s_root->d_inode) { struct ll_file_data *fd = LUSTRE_FPRIVATE(file); - LASSERT(fd != NULL); - rc = rct_add(&sbi->ll_rct, cfs_curproc_pid(), arg); - if (!rc) - fd->fd_flags |= LL_FILE_RMTACL; - RETURN(rc); + LASSERT(fd != NULL); + rc = rct_add(&sbi->ll_rct, current_pid(), arg); + if (!rc) + fd->fd_flags |= LL_FILE_RMTACL; + RETURN(rc); } else RETURN(0); } @@ -1427,7 +1619,7 @@ out_free: int count, vallen; struct obd_export *exp; - if (cfs_copy_from_user(&count, (int *)arg, sizeof(int))) + if (copy_from_user(&count, (int *)arg, sizeof(int))) RETURN(-EFAULT); /* get ost count when count is zero, get mdt count otherwise */ @@ -1440,13 +1632,13 @@ out_free: RETURN(rc); } - if (cfs_copy_to_user((int *)arg, &count, sizeof(int))) + if (copy_to_user((int *)arg, &count, sizeof(int))) RETURN(-EFAULT); RETURN(0); } case LL_IOC_PATH2FID: - if (cfs_copy_to_user((void *)arg, ll_inode2fid(inode), + if (copy_to_user((void *)arg, ll_inode2fid(inode), sizeof(struct lu_fid))) RETURN(-EFAULT); RETURN(0); @@ -1458,16 +1650,189 @@ out_free: rc = copy_and_ioctl(cmd, sbi->ll_md_exp, (void *)arg, sizeof(struct ioc_changelog)); RETURN(rc); - case OBD_IOC_FID2PATH: + case OBD_IOC_FID2PATH: RETURN(ll_fid2path(inode, (void *)arg)); - case LL_IOC_HSM_CT_START: - rc = copy_and_ioctl(cmd, sbi->ll_md_exp, (void *)arg, - sizeof(struct lustre_kernelcomm)); - RETURN(rc); - - default: - RETURN(obd_iocontrol(cmd, sbi->ll_dt_exp,0,NULL,(void *)arg)); - } + case LL_IOC_FID2MDTIDX: { + struct obd_export *exp = ll_i2mdexp(inode); + struct lu_fid fid; + __u32 index; + + if (copy_from_user(&fid, (const struct lu_fid __user *)arg, + sizeof(fid))) + RETURN(-EFAULT); + + /* Call mdc_iocontrol */ + rc = obd_iocontrol(LL_IOC_FID2MDTIDX, exp, sizeof(fid), &fid, + &index); + if (rc != 0) + RETURN(rc); + + RETURN(index); + } + case LL_IOC_HSM_REQUEST: { + struct hsm_user_request *hur; + ssize_t totalsize; + + OBD_ALLOC_PTR(hur); + if (hur == NULL) + RETURN(-ENOMEM); + + /* We don't know the true size yet; copy the fixed-size part */ + if (copy_from_user(hur, (void *)arg, sizeof(*hur))) { + OBD_FREE_PTR(hur); + RETURN(-EFAULT); + } + + /* Compute the whole struct size */ + totalsize = hur_len(hur); + OBD_FREE_PTR(hur); + if (totalsize < 0) + RETURN(-E2BIG); + + /* Final size will be more than double totalsize */ + if (totalsize >= MDS_MAXREQSIZE / 3) + RETURN(-E2BIG); + + OBD_ALLOC_LARGE(hur, totalsize); + if (hur == NULL) + RETURN(-ENOMEM); + + /* Copy the whole struct */ + if (copy_from_user(hur, (void *)arg, totalsize)) { + OBD_FREE_LARGE(hur, totalsize); + RETURN(-EFAULT); + } + + if (hur->hur_request.hr_action == HUA_RELEASE) { + const struct lu_fid *fid; + struct inode *f; + int i; + + for (i = 0; i < hur->hur_request.hr_itemcount; i++) { + fid = &hur->hur_user_item[i].hui_fid; + f = search_inode_for_lustre(inode->i_sb, fid); + if (IS_ERR(f)) { + rc = PTR_ERR(f); + break; + } + + rc = ll_hsm_release(f); + iput(f); + if (rc != 0) + break; + } + } else { + rc = obd_iocontrol(cmd, ll_i2mdexp(inode), totalsize, + hur, NULL); + } + + OBD_FREE_LARGE(hur, totalsize); + + RETURN(rc); + } + case LL_IOC_HSM_PROGRESS: { + struct hsm_progress_kernel hpk; + struct hsm_progress hp; + + if (copy_from_user(&hp, (void *)arg, sizeof(hp))) + RETURN(-EFAULT); + + hpk.hpk_fid = hp.hp_fid; + hpk.hpk_cookie = hp.hp_cookie; + hpk.hpk_extent = hp.hp_extent; + hpk.hpk_flags = hp.hp_flags; + hpk.hpk_errval = hp.hp_errval; + hpk.hpk_data_version = 0; + + /* File may not exist in Lustre; all progress + * reported to Lustre root */ + rc = obd_iocontrol(cmd, sbi->ll_md_exp, sizeof(hpk), &hpk, + NULL); + RETURN(rc); + } + case LL_IOC_HSM_CT_START: + if (!cfs_capable(CFS_CAP_SYS_ADMIN)) + RETURN(-EPERM); + + rc = copy_and_ioctl(cmd, sbi->ll_md_exp, (void *)arg, + sizeof(struct lustre_kernelcomm)); + RETURN(rc); + + case LL_IOC_HSM_COPY_START: { + struct hsm_copy *copy; + int rc; + + OBD_ALLOC_PTR(copy); + if (copy == NULL) + RETURN(-ENOMEM); + if (copy_from_user(copy, (char *)arg, sizeof(*copy))) { + OBD_FREE_PTR(copy); + RETURN(-EFAULT); + } + + rc = ll_ioc_copy_start(inode->i_sb, copy); + if (copy_to_user((char *)arg, copy, sizeof(*copy))) + rc = -EFAULT; + + OBD_FREE_PTR(copy); + RETURN(rc); + } + case LL_IOC_HSM_COPY_END: { + struct hsm_copy *copy; + int rc; + + OBD_ALLOC_PTR(copy); + if (copy == NULL) + RETURN(-ENOMEM); + if (copy_from_user(copy, (char *)arg, sizeof(*copy))) { + OBD_FREE_PTR(copy); + RETURN(-EFAULT); + } + + rc = ll_ioc_copy_end(inode->i_sb, copy); + if (copy_to_user((char *)arg, copy, sizeof(*copy))) + rc = -EFAULT; + + OBD_FREE_PTR(copy); + RETURN(rc); + } + case LL_IOC_MIGRATE: { + char *buf = NULL; + const char *filename; + int namelen = 0; + int len; + int rc; + int mdtidx; + + rc = obd_ioctl_getdata(&buf, &len, (void __user *)arg); + if (rc < 0) + RETURN(rc); + + data = (struct obd_ioctl_data *)buf; + if (data->ioc_inlbuf1 == NULL || data->ioc_inlbuf2 == NULL || + data->ioc_inllen1 == 0 || data->ioc_inllen2 == 0) + GOTO(migrate_free, rc = -EINVAL); + + filename = data->ioc_inlbuf1; + namelen = data->ioc_inllen1; + /* \0 is packed at the end of filename */ + if (namelen < 1 || namelen != strlen(filename) + 1) + GOTO(migrate_free, rc = -EINVAL); + + if (data->ioc_inllen2 != sizeof(mdtidx)) + GOTO(migrate_free, rc = -EINVAL); + mdtidx = *(int *)data->ioc_inlbuf2; + + rc = ll_migrate(inode, file, mdtidx, filename, namelen - 1); +migrate_free: + obd_ioctl_freedata(buf, len); + + RETURN(rc); + } + default: + RETURN(obd_iocontrol(cmd, sbi->ll_dt_exp, 0, NULL, + (void *)arg)); + } } static loff_t ll_dir_seek(struct file *file, loff_t offset, int origin) @@ -1521,24 +1886,28 @@ out: return ret; } -int ll_dir_open(struct inode *inode, struct file *file) +static int ll_dir_open(struct inode *inode, struct file *file) { ENTRY; RETURN(ll_file_open(inode, file)); } -int ll_dir_release(struct inode *inode, struct file *file) +static int ll_dir_release(struct inode *inode, struct file *file) { ENTRY; RETURN(ll_file_release(inode, file)); } -struct file_operations ll_dir_operations = { - .llseek = ll_dir_seek, - .open = ll_dir_open, - .release = ll_dir_release, - .read = generic_read_dir, - .readdir = ll_readdir, - .unlocked_ioctl = ll_dir_ioctl, - .fsync = ll_fsync, +const struct file_operations ll_dir_operations = { + .llseek = ll_dir_seek, + .open = ll_dir_open, + .release = ll_dir_release, + .read = generic_read_dir, +#ifdef HAVE_DIR_CONTEXT + .iterate = ll_iterate, +#else + .readdir = ll_readdir, +#endif + .unlocked_ioctl = ll_dir_ioctl, + .fsync = ll_fsync, };