X-Git-Url: https://git.whamcloud.com/?p=fs%2Flustre-release.git;a=blobdiff_plain;f=lustre%2Fllite%2Fdir.c;h=2021897e2664a5720e7d98e80b43365944e093ca;hp=ed8e1c7893ec0d104613c0b33492a62814ed91f4;hb=869790802c78a41440533512549b96ad7d478428;hpb=08aa217ce49aba1ded52e0f7adb8a607035123fd diff --git a/lustre/llite/dir.c b/lustre/llite/dir.c index ed8e1c7..2021897 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/ @@ -134,115 +134,111 @@ * 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 + * lmv_adjust_dirpages(). * */ /* returns the page unlocked, but with a reference */ static int ll_dir_filler(void *_hash, struct page *page0) { - 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; + 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; + struct page **page_pool; + struct page *page; + struct lu_dirpage *dp; + int max_pages = ll_i2sbi(inode)->ll_md_brw_size >> PAGE_CACHE_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); + CDEBUG(D_VFSTRACE, "VFS Op:inode="DFID"(%p) hash "LPU64"\n", + PFID(ll_inode2fid(inode)), inode, hash); - LASSERT(max_pages > 0 && max_pages <= PTLRPC_MAX_BRW_PAGES); + LASSERT(max_pages > 0 && max_pages <= MD_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); + if (IS_ERR(op_data)) + RETURN(PTR_ERR(op_data)); - 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); + 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; + } - if (body->valid & OBD_MD_FLSIZE) - cl_isize_write(inode, body->size); + 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 + + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT; + SetPageUptodate(page0); + } + unlock_page(page0); + ptlrpc_req_finished(request); - 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); - CDEBUG(D_VFSTRACE, "read %d/%d pages\n", nrdpgs, npages); + for (i = 1; i < npages; i++) { + unsigned long offset; + int ret; - ll_pagevec_init(&lru_pvec, 0); - for (i = 1; i < npages; i++) { - unsigned long offset; - int ret; + page = page_pool[i]; - page = page_pool[i]; + if (rc < 0 || i >= nrdpgs) { + page_cache_release(page); + continue; + } - if (rc < 0 || i >= nrdpgs) { - page_cache_release(page); - continue; - } + SetPageUptodate(page); - SetPageUptodate(page); + dp = kmap(page); + hash = le64_to_cpu(dp->ldp_hash_start); + kunmap(page); - dp = cfs_kmap(page); - hash = le64_to_cpu(dp->ldp_hash_start); - cfs_kunmap(page); + offset = hash_x_index(hash, hash64); - offset = hash_x_index(hash, hash64); + prefetchw(&page->flags); + ret = add_to_page_cache_lru(page, inode->i_mapping, offset, + GFP_KERNEL); + if (ret == 0) + unlock_page(page); + else + CDEBUG(D_VFSTRACE, "page %lu add to page cache failed:" + " %d\n", offset, ret); + page_cache_release(page); + } - 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); + if (page_pool != &page0) + OBD_FREE(page_pool, sizeof(struct page *) * max_pages); - if (page_pool != &page0) - OBD_FREE(page_pool, sizeof(struct page *) * max_pages); - EXIT; - return rc; + RETURN(rc); } static void ll_check_page(struct inode *dir, struct page *page) @@ -280,14 +276,14 @@ static struct page *ll_dir_page_locate(struct inode *dir, __u64 *hash, struct page *page; int found; - TREE_READ_LOCK_IRQ(mapping); + spin_lock_irq(&mapping->tree_lock); 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); + spin_unlock_irq(&mapping->tree_lock); /* * In contrast to find_lock_page() we are sure that directory * page cannot be truncated (while DLM lock is held) and, @@ -296,9 +292,9 @@ static struct page *ll_dir_page_locate(struct inode *dir, __u64 *hash, * 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); + wait_on_page_locked(page); + if (PageUptodate(page)) { + dp = 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; @@ -331,7 +327,7 @@ static struct page *ll_dir_page_locate(struct inode *dir, __u64 *hash, } } else { - TREE_READ_UNLOCK_IRQ(mapping); + spin_unlock_irq(&mapping->tree_lock); page = NULL; } return page; @@ -356,32 +352,40 @@ struct page *ll_get_dir_page(struct inode *dir, __u64 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); - } + if (!rc) { + struct ldlm_enqueue_info einfo = { + .ei_type = LDLM_IBITS, + .ei_mode = mode, + .ei_cb_bl = ll_md_blocking_ast, + .ei_cb_cp = ldlm_completion_ast, + }; + 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, dir, 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); + } + + CDEBUG(D_INODE, "setting lr_lvb_inode to inode "DFID"(%p)\n", + PFID(ll_inode2fid(dir)), dir); + md_set_lock_data(ll_i2sbi(dir)->ll_md_exp, + &it.d.lustre.it_lock_handle, dir, NULL); } else { /* for cross-ref object, l_ast_data of the lock may not be set, * we reset it here */ @@ -422,13 +426,13 @@ struct page *ll_get_dir_page(struct inode *dir, __u64 hash, 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; - } + wait_on_page_locked(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)) { @@ -594,7 +598,7 @@ static int ll_readdir(struct file *filp, void *cookie, filldir_t filldir) 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; + __u64 pos; int hash64 = sbi->ll_flags & LL_SBI_64BIT_HASH; int api32 = ll_need_32bit_api(sbi); int rc; @@ -603,8 +607,13 @@ static int ll_readdir(struct file *filp, void *cookie, filldir_t filldir) #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 %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) @@ -614,7 +623,8 @@ 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 (lfd != NULL) + lfd->lfd_pos = pos; if (pos == MDS_DIR_END_OFF) { if (api32) filp->f_pos = LL_DIR_END_OFF_32BIT; @@ -628,7 +638,11 @@ static int ll_readdir(struct file *filp, void *cookie, filldir_t filldir) } filp->f_version = inode->i_version; #ifdef HAVE_TOUCH_ATIME_1ARG +#ifdef HAVE_F_PATH_MNT + path.mnt = filp->f_path.mnt; +#else path.mnt = filp->f_vfsmnt; +#endif path.dentry = filp->f_dentry; touch_atime(&path); #else @@ -661,6 +675,36 @@ int ll_send_mgc_param(struct obd_export *mgc, char *string) return rc; } +int ll_dir_setdirstripe(struct inode *dir, struct lmv_user_md *lump, + char *filename) +{ + struct ptlrpc_request *request = NULL; + struct md_op_data *op_data; + struct ll_sb_info *sbi = ll_i2sbi(dir); + int mode; + int err; + + ENTRY; + + mode = (0755 & (S_IRWXUGO|S_ISVTX) & ~current->fs->umask) | S_IFDIR; + op_data = ll_prep_md_op_data(NULL, dir, NULL, filename, + strlen(filename), 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, + current_fsuid(), 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 +729,22 @@ 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; - } + } 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); } @@ -710,6 +754,9 @@ int ll_dir_setstripe(struct inode *inode, struct lov_user_md *lump, if (IS_ERR(op_data)) RETURN(PTR_ERR(op_data)); + if (lump != NULL && lump->lmm_magic == cpu_to_le32(LMV_USER_MAGIC)) + op_data->op_cli_flags |= CLI_SET_MEA; + /* swabbing is done in lov_setstripe() on server side */ rc = md_setattr(sbi->ll_md_exp, op_data, lump, lum_size, NULL, 0, &req, NULL); @@ -788,9 +835,8 @@ int ll_dir_getstripe(struct inode *inode, struct lov_mds_md **lmmp, 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); + CDEBUG(D_INFO, "md_getattr failed on inode " + DFID": rc %d\n", PFID(ll_inode2fid(inode)), rc); GOTO(out, rc); } @@ -860,21 +906,198 @@ int ll_get_mdt_idx(struct inode *inode) return mdtidx; } -static int copy_and_ioctl(int cmd, struct obd_export *exp, void *data, int len) +/** + * 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) { - 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; + /* 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); +} + +/** + * 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) +{ + struct ll_sb_info *sbi = ll_s2sbi(sb); + struct hsm_progress_kernel hpk; + int rc; + ENTRY; + + /* 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 +1120,12 @@ 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 && current_euid() != id) || + (type == GRPQUOTA && !in_egroup_p(id))) && + (!cfs_capable(CFS_CAP_SYS_ADMIN) || + sbi->ll_flags & LL_SBI_RMT_CLIENT)) + RETURN(-EPERM); break; case Q_GETINFO: break; @@ -1013,6 +1236,30 @@ 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 +1268,8 @@ 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 */ @@ -1092,6 +1339,51 @@ 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; + 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); + } + + /** + * ll_dir_setdirstripe will be used to set dir stripe + * mdc_create--->mdt_reint_create (with dirstripe) + */ + rc = ll_dir_setdirstripe(inode, lum, filename); +lmv_out_free: + obd_ioctl_freedata(buf, len); + 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 +1396,11 @@ 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); } @@ -1120,6 +1412,72 @@ out_free: RETURN(rc); } + case LL_IOC_LMV_GETSTRIPE: { + struct lmv_user_md *lump = (struct lmv_user_md *)arg; + struct lmv_user_md lum; + struct lmv_user_md *tmp; + int lum_size; + int rc = 0; + int mdtindex; + + if (copy_from_user(&lum, lump, sizeof(struct lmv_user_md))) + RETURN(-EFAULT); + + if (lum.lum_magic != LMV_MAGIC_V1) + RETURN(-EINVAL); + + lum_size = lmv_user_md_size(1, LMV_MAGIC_V1); + OBD_ALLOC(tmp, lum_size); + if (tmp == NULL) + GOTO(free_lmv, rc = -ENOMEM); + + memcpy(tmp, &lum, sizeof(lum)); + tmp->lum_type = LMV_STRIPE_TYPE; + tmp->lum_stripe_count = 1; + mdtindex = ll_get_mdt_idx(inode); + if (mdtindex < 0) + GOTO(free_lmv, rc = -ENOMEM); + + tmp->lum_stripe_offset = mdtindex; + tmp->lum_objects[0].lum_mds = mdtindex; + memcpy(&tmp->lum_objects[0].lum_fid, ll_inode2fid(inode), + sizeof(struct lu_fid)); + if (copy_to_user((void *)arg, tmp, lum_size)) + GOTO(free_lmv, rc = -EFAULT); +free_lmv: + if (tmp) + OBD_FREE(tmp, lum_size); + 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 LL_IOC_OBD_STATFS: RETURN(ll_obd_statfs(inode, (void *)arg)); case LL_IOC_LOV_GETSTRIPE: @@ -1135,7 +1493,7 @@ 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)); @@ -1169,8 +1527,8 @@ 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; } @@ -1186,7 +1544,7 @@ out_free: st.st_gid = body->gid; st.st_rdev = body->rdev; st.st_size = body->size; - st.st_blksize = CFS_PAGE_SIZE; + st.st_blksize = PAGE_CACHE_SIZE; st.st_blocks = body->blocks; st.st_atime = body->atime; st.st_mtime = body->mtime; @@ -1194,7 +1552,7 @@ out_free: st.st_ino = inode->i_ino; lmdp = (struct lov_user_mds_data *)arg; - if (cfs_copy_to_user(&lmdp->lmd_st, &st, sizeof(st))) + if (copy_to_user(&lmdp->lmd_st, &st, sizeof(st))) GOTO(out_req, rc = -EFAULT); } @@ -1202,7 +1560,7 @@ out_free: out_req: ptlrpc_req_finished(request); if (filename) - putname(filename); + ll_putname(filename); return rc; } case IOC_LOV_GETINFO: { @@ -1220,9 +1578,12 @@ 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 +1619,7 @@ 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; @@ -1311,9 +1672,9 @@ 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 +1682,9 @@ 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 +1706,7 @@ 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 +1731,7 @@ 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; } @@ -1391,12 +1752,12 @@ out_free: 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 +1775,11 @@ 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 +1788,7 @@ 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 +1801,13 @@ 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); @@ -1460,14 +1821,135 @@ out_free: RETURN(rc); 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_HSM_REQUEST: { + struct hsm_user_request *hur; + int 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); + + /* Make sure the size is reasonable */ + if (totalsize >= MDS_MAXREQSIZE) + 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); + } + 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)