Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1758864AbcCVLRB (ORCPT ); Tue, 22 Mar 2016 07:17:01 -0400 Received: from youngberry.canonical.com ([91.189.89.112]:56541 "EHLO youngberry.canonical.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1758717AbcCVKmC (ORCPT ); Tue, 22 Mar 2016 06:42:02 -0400 From: Luis Henriques To: linux-kernel@vger.kernel.org, stable@vger.kernel.org, kernel-team@lists.ubuntu.com Cc: Ilya Dryomov , Luis Henriques Subject: [PATCH 3.16.y-ckt 043/142] libceph: don't bail early from try_read() when skipping a message Date: Tue, 22 Mar 2016 10:39:32 +0000 Message-Id: <1458643271-4227-44-git-send-email-luis.henriques@canonical.com> In-Reply-To: <1458643271-4227-1-git-send-email-luis.henriques@canonical.com> References: <1458643271-4227-1-git-send-email-luis.henriques@canonical.com> X-Extended-Stable: 3.16 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1792 Lines: 47 3.16.7-ckt26 -stable review patch. If anyone has any objections, please let me know. ---8<------------------------------------------------------------ From: Ilya Dryomov commit e7a88e82fe380459b864e05b372638aeacb0f52d upstream. The contract between try_read() and try_write() is that when called each processes as much data as possible. When instructed by osd_client to skip a message, try_read() is violating this contract by returning after receiving and discarding a single message instead of checking for more. try_write() then gets a chance to write out more requests, generating more replies/skips for try_read() to handle, forcing the messenger into a starvation loop. Reported-by: Varada Kari Signed-off-by: Ilya Dryomov Tested-by: Varada Kari Reviewed-by: Alex Elder Signed-off-by: Luis Henriques --- net/ceph/messenger.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/net/ceph/messenger.c b/net/ceph/messenger.c index 6b15183a68c8..f2ea1a093800 100644 --- a/net/ceph/messenger.c +++ b/net/ceph/messenger.c @@ -2285,7 +2285,7 @@ static int read_partial_message(struct ceph_connection *con) con->in_base_pos = -front_len - middle_len - data_len - sizeof(m->footer); con->in_tag = CEPH_MSGR_TAG_READY; - return 0; + return 1; } else if ((s64)seq - (s64)con->in_seq > 1) { pr_err("read_partial_message bad seq %lld expected %lld\n", seq, con->in_seq + 1); @@ -2318,7 +2318,7 @@ static int read_partial_message(struct ceph_connection *con) sizeof(m->footer); con->in_tag = CEPH_MSGR_TAG_READY; con->in_seq++; - return 0; + return 1; } BUG_ON(!con->in_msg);