From patchwork Thu Apr 15 04:02:38 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: James Simmons X-Patchwork-Id: 12204289 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-16.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS,URIBL_BLOCKED,USER_AGENT_GIT autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id E4F59C433B4 for ; Thu, 15 Apr 2021 04:05:03 +0000 (UTC) Received: from pdx1-mailman02.dreamhost.com (pdx1-mailman02.dreamhost.com [64.90.62.194]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id AA963610CB for ; Thu, 15 Apr 2021 04:05:03 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org AA963610CB Authentication-Results: mail.kernel.org; dmarc=none (p=none dis=none) header.from=infradead.org Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=lustre-devel-bounces@lists.lustre.org Received: from pdx1-mailman02.dreamhost.com (localhost [IPv6:::1]) by pdx1-mailman02.dreamhost.com (Postfix) with ESMTP id 7AC3532DD7E; Wed, 14 Apr 2021 21:03:59 -0700 (PDT) Received: from smtp4.ccs.ornl.gov (smtp4.ccs.ornl.gov [160.91.203.40]) by pdx1-mailman02.dreamhost.com (Postfix) with ESMTP id EBFA132F53E for ; Wed, 14 Apr 2021 21:02:59 -0700 (PDT) Received: from star.ccs.ornl.gov (star.ccs.ornl.gov [160.91.202.134]) by smtp4.ccs.ornl.gov (Postfix) with ESMTP id CFC43100F376; Thu, 15 Apr 2021 00:02:45 -0400 (EDT) Received: by star.ccs.ornl.gov (Postfix, from userid 2004) id CE6F76DF70; Thu, 15 Apr 2021 00:02:45 -0400 (EDT) From: James Simmons To: Andreas Dilger , Oleg Drokin , NeilBrown Date: Thu, 15 Apr 2021 00:02:38 -0400 Message-Id: <1618459361-17909-47-git-send-email-jsimmons@infradead.org> X-Mailer: git-send-email 1.8.3.1 In-Reply-To: <1618459361-17909-1-git-send-email-jsimmons@infradead.org> References: <1618459361-17909-1-git-send-email-jsimmons@infradead.org> Subject: [lustre-devel] [PATCH 46/49] lustre: lov: return valid stripe_count/size for PFL files X-BeenThere: lustre-devel@lists.lustre.org X-Mailman-Version: 2.1.23 Precedence: list List-Id: "For discussing Lustre software development." List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: Lustre Development List MIME-Version: 1.0 Errors-To: lustre-devel-bounces@lists.lustre.org Sender: "lustre-devel" From: Emoly Liu Dump struct lov_comp_md_v1 in function ll_lov_getstripe_ea_info() correctly to avoid stripe_count=0 or stripe_size=0 returned by old interface llapi_file_get_stripe(), which will cause divide-by-zero for older userspace that calls this ioctl, e.g. lustre ADIO driver. The rule is: - if stripe_count=0, return stripe_count=1; - if stripe_size=0, -- for DoM files, return the stripe size of the second component, since the first component of DoM file data is placed on the MDT for faster access; -- else, return the stripe size of the last component. WC-bug-id: https://jira.whamcloud.com/browse/LU-14337 Lustre-commit: abf04e7ea356e8b ("LU-14337 lov: return valid stripe_count/size for PFL files") Signed-off-by: Emoly Liu Reviewed-on: https://review.whamcloud.com/41803 Reviewed-by: Andreas Dilger Reviewed-by: Bobi Jam Reviewed-by: Oleg Drokin Signed-off-by: James Simmons --- fs/lustre/llite/file.c | 74 ++++++++++++++++++++++++++++++++++++++---------- fs/lustre/lov/lov_pack.c | 7 ----- 2 files changed, 59 insertions(+), 22 deletions(-) diff --git a/fs/lustre/llite/file.c b/fs/lustre/llite/file.c index bbb2ff9..2558a60 100644 --- a/fs/lustre/llite/file.c +++ b/fs/lustre/llite/file.c @@ -2059,6 +2059,7 @@ int ll_lov_getstripe_ea_info(struct inode *inode, const char *filename, } body = req_capsule_server_get(&req->rq_pill, &RMF_MDT_BODY); + LASSERT(body); /* checked by mdc_getattr_name */ lmmsize = body->mbo_eadatasize; @@ -2069,6 +2070,7 @@ int ll_lov_getstripe_ea_info(struct inode *inode, const char *filename, } lmm = req_capsule_server_sized_get(&req->rq_pill, &RMF_MDT_MD, lmmsize); + LASSERT(lmm); if (lmm->lmm_magic != cpu_to_le32(LOV_MAGIC_V1) && lmm->lmm_magic != cpu_to_le32(LOV_MAGIC_V3) && @@ -2083,8 +2085,7 @@ int ll_lov_getstripe_ea_info(struct inode *inode, const char *filename, * little endian. We convert it to host endian before * passing it to userspace. */ - if ((lmm->lmm_magic & __swab32(LOV_MAGIC_MAGIC)) == - __swab32(LOV_MAGIC_MAGIC)) { + if (cpu_to_le32(LOV_MAGIC) != LOV_MAGIC) { int stripe_count = 0; if (lmm->lmm_magic == cpu_to_le32(LOV_MAGIC_V1) || @@ -2093,24 +2094,67 @@ int ll_lov_getstripe_ea_info(struct inode *inode, const char *filename, if (le32_to_cpu(lmm->lmm_pattern) & LOV_PATTERN_F_RELEASED) stripe_count = 0; + + lustre_swab_lov_user_md((struct lov_user_md *)lmm, 0); + + /* if function called for directory - we should + * avoid swab not existent lsm objects + */ + if (lmm->lmm_magic == LOV_MAGIC_V1 && + S_ISREG(body->mbo_mode)) + lustre_swab_lov_user_md_objects(((struct lov_user_md_v1 *)lmm)->lmm_objects, + stripe_count); + else if (lmm->lmm_magic == LOV_MAGIC_V3 && + S_ISREG(body->mbo_mode)) + lustre_swab_lov_user_md_objects(((struct lov_user_md_v3 *)lmm)->lmm_objects, + stripe_count); + } else if (lmm->lmm_magic == cpu_to_le32(LOV_MAGIC_COMP_V1)) { + lustre_swab_lov_comp_md_v1((struct lov_comp_md_v1 *)lmm); } + } - lustre_swab_lov_user_md((struct lov_user_md *)lmm, 0); + if (lmm->lmm_magic == LOV_MAGIC_COMP_V1) { + struct lov_comp_md_v1 *comp_v1 = NULL; + struct lov_comp_md_entry_v1 *ent; + struct lov_user_md_v1 *v1; + u32 off; + int i = 0; + + comp_v1 = (struct lov_comp_md_v1 *)lmm; + /* Dump the striping information */ + for (; i < comp_v1->lcm_entry_count; i++) { + ent = &comp_v1->lcm_entries[i]; + off = ent->lcme_offset; + v1 = (struct lov_user_md_v1 *)((char *)lmm + off); + CDEBUG(D_INFO, + "comp[%d]: stripe_count=%u, stripe_size=%u\n", + i, v1->lmm_stripe_count, v1->lmm_stripe_size); + } - /* if function called for directory - we should - * avoid swab not existent lsm objects + /** + * Return valid stripe_count and stripe_size instead of 0 for + * DoM files to avoid divide-by-zero for older userspace that + * calls this ioctl, e.g. lustre ADIO driver. */ - if (lmm->lmm_magic == LOV_MAGIC_V1 && S_ISREG(body->mbo_mode)) - lustre_swab_lov_user_md_objects( - ((struct lov_user_md_v1 *)lmm)->lmm_objects, - stripe_count); - else if (lmm->lmm_magic == LOV_MAGIC_V3 && - S_ISREG(body->mbo_mode)) - lustre_swab_lov_user_md_objects( - ((struct lov_user_md_v3 *)lmm)->lmm_objects, - stripe_count); + if (lmm->lmm_stripe_count == 0) + lmm->lmm_stripe_count = 1; + if (lmm->lmm_stripe_size == 0) { + /* Since the first component of the file data is placed + * on the MDT for faster access, the stripe_size of the + * second one is always that applications which are + * doing large IOs. + */ + if (lmm->lmm_pattern == LOV_PATTERN_MDT) + i = comp_v1->lcm_entry_count > 1 ? 1 : 0; + else + i = comp_v1->lcm_entry_count > 1 ? + comp_v1->lcm_entry_count - 1 : 0; + ent = &comp_v1->lcm_entries[i]; + off = ent->lcme_offset; + v1 = (struct lov_user_md_v1 *)((char *)lmm + off); + lmm->lmm_stripe_size = v1->lmm_stripe_size; + } } - out: *lmmp = lmm; *lmm_size = lmmsize; diff --git a/fs/lustre/lov/lov_pack.c b/fs/lustre/lov/lov_pack.c index 1962472..c97093e 100644 --- a/fs/lustre/lov/lov_pack.c +++ b/fs/lustre/lov/lov_pack.c @@ -450,13 +450,6 @@ int lov_getstripe(const struct lu_env *env, struct lov_object *obj, } /** - * Return stripe_count=1 instead of 0 for DoM files to avoid - * divide-by-zero for older userspace that calls this ioctl, - * e.g. lustre ADIO driver. - */ - if ((lum.lmm_stripe_count == 0) && (lum.lmm_pattern & LOV_PATTERN_MDT)) - lum.lmm_stripe_count = 1; - /** * User specified limited buffer size, usually the buffer is * from ll_lov_setstripe(), and the buffer can only hold basic * layout template info.