Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
From:   James Simmons <jsimmons@infradead.org>
To:     Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        devel@driverdev.osuosl.org,
        Andreas Dilger <andreas.dilger@intel.com>,
        Oleg Drokin <oleg.drokin@intel.com>, NeilBrown <neilb@suse.com>
Cc:     Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        Lustre Development List <lustre-devel@lists.lustre.org>,
        "John L. Hammond" <john.hammond@intel.com>,
        James Simmons <jsimmons@infradead.org>
Subject: [PATCH 5/5] staging: lustre: mdc: use large xattr buffers for old servers
Date:   Mon, 14 May 2018 22:17:03 -0400
Message-Id: <1526350623-4616-6-git-send-email-jsimmons@infradead.org>
In-Reply-To: <1526350623-4616-1-git-send-email-jsimmons@infradead.org>
References: <1526350623-4616-1-git-send-email-jsimmons@infradead.org>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

From: "John L. Hammond" <john.hammond@intel.com>

Pre 2.10.1 MDTs will crash when they receive a listxattr (MDS_GETXATTR
with OBD_MD_FLXATTRLS) RPC for an orphan or dead object. So for
clients connected to these older MDTs, try to avoid sending listxattr
RPCs by making the bulk getxattr (MDS_GETXATTR with OBD_MD_FLXATTRALL)
more likely to succeed and thereby reducing the chances of falling
back to listxattr.

Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-10912
Reviewed-on: https://review.whamcloud.com/31990
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
 drivers/staging/lustre/lustre/mdc/mdc_locks.c | 31 +++++++++++++++++++++------
 1 file changed, 24 insertions(+), 7 deletions(-)

diff --git a/drivers/staging/lustre/lustre/mdc/mdc_locks.c b/drivers/staging/lustre/lustre/mdc/mdc_locks.c
index a8aa0fa..b991c6f 100644
--- a/drivers/staging/lustre/lustre/mdc/mdc_locks.c
+++ b/drivers/staging/lustre/lustre/mdc/mdc_locks.c
@@ -326,8 +326,10 @@ static void mdc_realloc_openmsg(struct ptlrpc_request *req,
 {
 	struct ptlrpc_request	*req;
 	struct ldlm_intent	*lit;
+	u32 min_buf_size = 0;
 	int rc, count = 0;
 	LIST_HEAD(cancels);
+	u32 buf_size = 0;
 
 	req = ptlrpc_request_alloc(class_exp2cliimp(exp),
 				   &RQF_LDLM_INTENT_GETXATTR);
@@ -344,18 +346,33 @@ static void mdc_realloc_openmsg(struct ptlrpc_request *req,
 	lit = req_capsule_client_get(&req->rq_pill, &RMF_LDLM_INTENT);
 	lit->opc = IT_GETXATTR;
 
+#if LUSTRE_VERSION_CODE < OBD_OCD_VERSION(3, 0, 53, 0)
+	/* If the supplied buffer is too small then the server will
+	 * return -ERANGE and llite will fallback to using non cached
+	 * xattr operations. On servers before 2.10.1 a (non-cached)
+	 * listxattr RPC for an orphan or dead file causes an oops. So
+	 * let's try to avoid sending too small a buffer to too old a
+	 * server. This is effectively undoing the memory conservation
+	 * of LU-9417 when it would be *more* likely to crash the
+	 * server. See LU-9856.
+	 */
+	if (exp->exp_connect_data.ocd_version < OBD_OCD_VERSION(2, 10, 1, 0))
+		min_buf_size = exp->exp_connect_data.ocd_max_easize;
+#endif
+	buf_size = max_t(u32, min_buf_size,
+			 GA_DEFAULT_EA_NAME_LEN * GA_DEFAULT_EA_NUM);
+
 	/* pack the intended request */
-	mdc_pack_body(req, &op_data->op_fid1, op_data->op_valid,
-		      GA_DEFAULT_EA_NAME_LEN * GA_DEFAULT_EA_NUM, -1, 0);
+	mdc_pack_body(req, &op_data->op_fid1, op_data->op_valid, buf_size,
+		      -1, 0);
 
-	req_capsule_set_size(&req->rq_pill, &RMF_EADATA, RCL_SERVER,
-			     GA_DEFAULT_EA_NAME_LEN * GA_DEFAULT_EA_NUM);
+	req_capsule_set_size(&req->rq_pill, &RMF_EADATA, RCL_SERVER, buf_size);
 
-	req_capsule_set_size(&req->rq_pill, &RMF_EAVALS, RCL_SERVER,
-			     GA_DEFAULT_EA_NAME_LEN * GA_DEFAULT_EA_NUM);
+	req_capsule_set_size(&req->rq_pill, &RMF_EAVALS, RCL_SERVER, buf_size);
 
 	req_capsule_set_size(&req->rq_pill, &RMF_EAVALS_LENS, RCL_SERVER,
-			     sizeof(u32) * GA_DEFAULT_EA_NUM);
+			     max_t(u32, min_buf_size,
+				   sizeof(u32) * GA_DEFAULT_EA_NUM));
 
 	req_capsule_set_size(&req->rq_pill, &RMF_ACL, RCL_SERVER, 0);
 
-- 
1.8.3.1