From patchwork Tue Jan 29 15:47:28 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Ilya Dryomov X-Patchwork-Id: 10786519 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 7FFE01390 for ; Tue, 29 Jan 2019 15:48:00 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 5E5F42CADC for ; Tue, 29 Jan 2019 15:48:00 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id 4F3EE2CEED; Tue, 29 Jan 2019 15:48:00 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-8.0 required=2.0 tests=BAYES_00,DKIM_SIGNED, DKIM_VALID,DKIM_VALID_AU,FREEMAIL_FROM,MAILING_LIST_MULTI,RCVD_IN_DNSWL_HI autolearn=ham version=3.3.1 Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id AA99A2CADC for ; Tue, 29 Jan 2019 15:47:59 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1727512AbfA2Pr6 (ORCPT ); Tue, 29 Jan 2019 10:47:58 -0500 Received: from mail-wm1-f66.google.com ([209.85.128.66]:36835 "EHLO mail-wm1-f66.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1725788AbfA2Pr6 (ORCPT ); Tue, 29 Jan 2019 10:47:58 -0500 Received: by mail-wm1-f66.google.com with SMTP id p6so18293191wmc.1 for ; Tue, 29 Jan 2019 07:47:56 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=from:to:subject:date:message-id:in-reply-to:references; bh=q58NgTMXPV4o+9oNQwc0lQ3/x/i9CYytxUSpD9gDhJ8=; b=ebPeel2Yiv4ylictqbjaCg8F0UM6YqC9ng5VJ2rOubN1ea1MUbfnNjEuK3jfjD1eic XPPuzzxn/bby9h7fmZ9N6IncaOGIWbzVqlJloM2otKgp+6I3BdxFKjFZ2c8YbkfMrLPL BnQHVKOJ9hH4QIRjDvYE4gz/aMs5MRunyB+th9zJmkgZVbtOK8IJ9M+uaO5IMEHYK5Ab n6PrS65lBnmxsHNcAyXrNvgbRuUz785M06HXRmqEaBY8AZvZLd3GgfSw8c1EqE6ZSjSE DEohNuQib1gSbXGJZOu5HFPLlfoUhgAhcuzAqE5kZCB8CrL2Xr6Kl02m7lHPBgm7299c +/+Q== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:from:to:subject:date:message-id:in-reply-to :references; bh=q58NgTMXPV4o+9oNQwc0lQ3/x/i9CYytxUSpD9gDhJ8=; b=Gdo0jMvT0BJOwW0OZvNM7DMdpwCTaUzckVeW10YLrE1HEsNvBR2fF2oiFCvgktLbF+ srDFS4k/zMUArr1QaeEdGd2IuEobATTAsfjD9J/brxdaYgy9ZX/bd++E3FT5cufJwhcp DZm626XbVRZ5V5on9H4xNKZQKVQr2jONjCa0jhhpc5x3HvNRB00J2KyQVGwnvMSd/JaF mTOQ9mZMLfG7Lgw2aykmyL5nJEGXMkVrUK8Sh9v5AuGzhp+VQQnjCcw9V2XyhZ6E+jM9 Yksnu4ZjT86ofYQ/h0duDVtiDFfgtKLxNqvk2/1unAIGnVBWM7y8xlHz54bSsS/oQi+2 1DSw== X-Gm-Message-State: AJcUukeRgaigOpyKxEaYsHvYJz++wI0mjB9VDBzVUd3deqorxm+POrYJ pn03FeLtQiHwk9Wwi/hBTYaEDywG X-Google-Smtp-Source: ALg8bN4NB0o4acjaOF7he43gWZ4vDD8z+fH7CD+ewTp0WZtyLAEQp72t8Ym44aLWTR8h8Ck5q2gGTQ== X-Received: by 2002:a1c:c181:: with SMTP id r123mr21698416wmf.8.1548776875396; Tue, 29 Jan 2019 07:47:55 -0800 (PST) Received: from orange.brq.redhat.com (nat-pool-brq-t.redhat.com. [213.175.37.10]) by smtp.gmail.com with ESMTPSA id g9sm3129423wmg.44.2019.01.29.07.47.54 for (version=TLS1_2 cipher=ECDHE-RSA-CHACHA20-POLY1305 bits=256/256); Tue, 29 Jan 2019 07:47:54 -0800 (PST) From: Ilya Dryomov To: ceph-devel@vger.kernel.org Subject: [PATCH 2/3] rbd: handle DISCARD and WRITE_ZEROES separately Date: Tue, 29 Jan 2019 16:47:28 +0100 Message-Id: <20190129154729.1031-3-idryomov@gmail.com> X-Mailer: git-send-email 2.14.4 In-Reply-To: <20190129154729.1031-1-idryomov@gmail.com> References: <20190129154729.1031-1-idryomov@gmail.com> Sender: ceph-devel-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: ceph-devel@vger.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP With discard_zeroes_data gone in commit 48920ff2a5a9 ("block: remove the discard_zeroes_data flag"), continuing to provide this guarantee is pointless: applications can't query it and discards can only be used for deallocating. Add OBJ_OP_ZEROOUT and move the existing logic under it. As the first step to divorcing OBJ_OP_DISCARD, stop doing copyups and worrying about whether parent blocks are still seen after the discard. Signed-off-by: Ilya Dryomov --- drivers/block/rbd.c | 54 +++++++++++++++++++++++++++++++++++++++++++---------- 1 file changed, 44 insertions(+), 10 deletions(-) diff --git a/drivers/block/rbd.c b/drivers/block/rbd.c index 7f78b814d57f..87f30bf49f1f 100644 --- a/drivers/block/rbd.c +++ b/drivers/block/rbd.c @@ -214,6 +214,7 @@ enum obj_operation_type { OBJ_OP_READ = 1, OBJ_OP_WRITE, OBJ_OP_DISCARD, + OBJ_OP_ZEROOUT, }; /* @@ -857,6 +858,8 @@ static char* obj_op_name(enum obj_operation_type op_type) return "write"; case OBJ_OP_DISCARD: return "discard"; + case OBJ_OP_ZEROOUT: + return "zeroout"; default: return "???"; } @@ -1419,6 +1422,7 @@ static bool rbd_img_is_write(struct rbd_img_request *img_req) return false; case OBJ_OP_WRITE: case OBJ_OP_DISCARD: + case OBJ_OP_ZEROOUT: return true; default: BUG(); @@ -1841,7 +1845,33 @@ static int rbd_obj_setup_write(struct rbd_obj_request *obj_req) return 0; } -static void __rbd_obj_setup_discard(struct rbd_obj_request *obj_req, +static u16 truncate_or_zero_opcode(struct rbd_obj_request *obj_req) +{ + return rbd_obj_is_tail(obj_req) ? CEPH_OSD_OP_TRUNCATE : + CEPH_OSD_OP_ZERO; +} + +static int rbd_obj_setup_discard(struct rbd_obj_request *obj_req) +{ + obj_req->osd_req = rbd_osd_req_create(obj_req, 1); + if (!obj_req->osd_req) + return -ENOMEM; + + if (rbd_obj_is_entire(obj_req)) { + osd_req_op_init(obj_req->osd_req, 0, CEPH_OSD_OP_DELETE, 0); + } else { + osd_req_op_extent_init(obj_req->osd_req, 0, + truncate_or_zero_opcode(obj_req), + obj_req->ex.oe_off, obj_req->ex.oe_len, + 0, 0); + } + + obj_req->write_state = RBD_OBJ_WRITE_FLAT; + rbd_osd_req_format_write(obj_req); + return 0; +} + +static void __rbd_obj_setup_zeroout(struct rbd_obj_request *obj_req, unsigned int which) { u16 opcode; @@ -1856,10 +1886,8 @@ static void __rbd_obj_setup_discard(struct rbd_obj_request *obj_req, CEPH_OSD_OP_DELETE, 0); opcode = 0; } - } else if (rbd_obj_is_tail(obj_req)) { - opcode = CEPH_OSD_OP_TRUNCATE; } else { - opcode = CEPH_OSD_OP_ZERO; + opcode = truncate_or_zero_opcode(obj_req); } if (opcode) @@ -1871,7 +1899,7 @@ static void __rbd_obj_setup_discard(struct rbd_obj_request *obj_req, rbd_osd_req_format_write(obj_req); } -static int rbd_obj_setup_discard(struct rbd_obj_request *obj_req) +static int rbd_obj_setup_zeroout(struct rbd_obj_request *obj_req) { unsigned int num_osd_ops, which = 0; int ret; @@ -1907,7 +1935,7 @@ static int rbd_obj_setup_discard(struct rbd_obj_request *obj_req) return ret; } - __rbd_obj_setup_discard(obj_req, which); + __rbd_obj_setup_zeroout(obj_req, which); return 0; } @@ -1932,6 +1960,9 @@ static int __rbd_img_fill_request(struct rbd_img_request *img_req) case OBJ_OP_DISCARD: ret = rbd_obj_setup_discard(obj_req); break; + case OBJ_OP_ZEROOUT: + ret = rbd_obj_setup_zeroout(obj_req); + break; default: rbd_assert(0); } @@ -2392,9 +2423,9 @@ static int rbd_obj_issue_copyup(struct rbd_obj_request *obj_req, u32 bytes) case OBJ_OP_WRITE: __rbd_obj_setup_write(obj_req, 1); break; - case OBJ_OP_DISCARD: + case OBJ_OP_ZEROOUT: rbd_assert(!rbd_obj_is_entire(obj_req)); - __rbd_obj_setup_discard(obj_req, 1); + __rbd_obj_setup_zeroout(obj_req, 1); break; default: rbd_assert(0); @@ -2524,6 +2555,7 @@ static bool __rbd_obj_handle_request(struct rbd_obj_request *obj_req) case OBJ_OP_WRITE: return rbd_obj_handle_write(obj_req); case OBJ_OP_DISCARD: + case OBJ_OP_ZEROOUT: if (rbd_obj_handle_write(obj_req)) { /* * Hide -ENOENT from delete/truncate/zero -- discarding @@ -3636,9 +3668,11 @@ static void rbd_queue_workfn(struct work_struct *work) switch (req_op(rq)) { case REQ_OP_DISCARD: - case REQ_OP_WRITE_ZEROES: op_type = OBJ_OP_DISCARD; break; + case REQ_OP_WRITE_ZEROES: + op_type = OBJ_OP_ZEROOUT; + break; case REQ_OP_WRITE: op_type = OBJ_OP_WRITE; break; @@ -3718,7 +3752,7 @@ static void rbd_queue_workfn(struct work_struct *work) img_request->rq = rq; snapc = NULL; /* img_request consumes a ref */ - if (op_type == OBJ_OP_DISCARD) + if (op_type == OBJ_OP_DISCARD || op_type == OBJ_OP_ZEROOUT) result = rbd_img_fill_nodata(img_request, offset, length); else result = rbd_img_fill_from_bio(img_request, offset, length,