From patchwork Mon Apr 19 15:11:42 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Greg Kurz X-Patchwork-Id: 1467981 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@bilbo.ozlabs.org Authentication-Results: ozlabs.org; spf=pass (sender SPF authorized) smtp.mailfrom=nongnu.org (client-ip=209.51.188.17; helo=lists.gnu.org; envelope-from=qemu-devel-bounces+incoming=patchwork.ozlabs.org@nongnu.org; receiver=) Received: from lists.gnu.org (lists.gnu.org [209.51.188.17]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by ozlabs.org (Postfix) with ESMTPS id 4FP9QN3ztLz9vFB for ; Tue, 20 Apr 2021 01:16:44 +1000 (AEST) Received: from localhost ([::1]:57270 helo=lists1p.gnu.org) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1lYVdi-0005v0-Hm for incoming@patchwork.ozlabs.org; Mon, 19 Apr 2021 11:16:42 -0400 Received: from eggs.gnu.org ([2001:470:142:3::10]:48712) by lists.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1lYVZX-0000p1-Vo for qemu-devel@nongnu.org; Mon, 19 Apr 2021 11:12:24 -0400 Received: from us-smtp-delivery-44.mimecast.com ([207.211.30.44]:45780) by eggs.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1lYVZW-0008DN-9z for qemu-devel@nongnu.org; Mon, 19 Apr 2021 11:12:23 -0400 Received: from mimecast-mx01.redhat.com (mimecast-mx01.redhat.com [209.132.183.4]) (Using TLS) by relay.mimecast.com with ESMTP id us-mta-541-Lc1nq3W_OmW9HJXOJbF-ow-1; Mon, 19 Apr 2021 11:12:08 -0400 X-MC-Unique: Lc1nq3W_OmW9HJXOJbF-ow-1 Received: from smtp.corp.redhat.com (int-mx06.intmail.prod.int.phx2.redhat.com [10.5.11.16]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mimecast-mx01.redhat.com (Postfix) with ESMTPS id E2BD0100A5F1; Mon, 19 Apr 2021 15:12:06 +0000 (UTC) Received: from bahia.redhat.com (ovpn-112-134.ams2.redhat.com [10.36.112.134]) by smtp.corp.redhat.com (Postfix) with ESMTP id 61FB75C1CF; Mon, 19 Apr 2021 15:12:05 +0000 (UTC) From: Greg Kurz To: qemu-devel@nongnu.org Subject: [for-6.1 2/2] virtiofsd: Add support for FUSE_SYNCFS request Date: Mon, 19 Apr 2021 17:11:42 +0200 Message-Id: <20210419151142.276439-3-groug@kaod.org> In-Reply-To: <20210419151142.276439-1-groug@kaod.org> References: <20210419151142.276439-1-groug@kaod.org> MIME-Version: 1.0 X-Scanned-By: MIMEDefang 2.79 on 10.5.11.16 X-Mimecast-Spam-Score: 0 X-Mimecast-Originator: kaod.org Received-SPF: softfail client-ip=207.211.30.44; envelope-from=groug@kaod.org; helo=us-smtp-delivery-44.mimecast.com X-Spam_score_int: -18 X-Spam_score: -1.9 X-Spam_bar: - X-Spam_report: (-1.9 / 5.0 requ) BAYES_00=-1.9, RCVD_IN_DNSWL_LOW=-0.7, SPF_HELO_NONE=0.001, SPF_SOFTFAIL=0.665 autolearn=no autolearn_force=no X-Spam_action: no action X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.23 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: Miklos Szeredi , Greg Kurz , "Dr. David Alan Gilbert" , virtio-fs@redhat.com, Stefan Hajnoczi , Vivek Goyal Errors-To: qemu-devel-bounces+incoming=patchwork.ozlabs.org@nongnu.org Sender: "Qemu-devel" Honor the expected behavior of syncfs() to synchronously flush all data and metadata on linux systems. Like the ->sync_fs() superblock operation in the linux kernel, FUSE_SYNCFS has a 'wait' argument that tells whether the server should wait for outstanding I/Os to complete before replying to the client. Anything virtiofsd can do to flush the caches implies blocking syscalls, so nothing is done if waiting isn't requested. Flushing is done with syncfs(). This is suboptimal as it will also flush writes performed by any other process on the same file system, and thus add an unbounded time penalty to syncfs(). This may be optimized in the future, but enforce correctness first. Signed-off-by: Greg Kurz --- tools/virtiofsd/fuse_lowlevel.c | 19 ++++++++++++++++++ tools/virtiofsd/fuse_lowlevel.h | 13 ++++++++++++ tools/virtiofsd/passthrough_ll.c | 29 +++++++++++++++++++++++++++ tools/virtiofsd/passthrough_seccomp.c | 1 + 4 files changed, 62 insertions(+) diff --git a/tools/virtiofsd/fuse_lowlevel.c b/tools/virtiofsd/fuse_lowlevel.c index 58e32fc96369..2d0c47a7a60e 100644 --- a/tools/virtiofsd/fuse_lowlevel.c +++ b/tools/virtiofsd/fuse_lowlevel.c @@ -1870,6 +1870,24 @@ static void do_lseek(fuse_req_t req, fuse_ino_t nodeid, } } +static void do_syncfs(fuse_req_t req, fuse_ino_t nodeid, + struct fuse_mbuf_iter *iter) +{ + struct fuse_syncfs_in *arg; + + arg = fuse_mbuf_iter_advance(iter, sizeof(*arg)); + if (!arg) { + fuse_reply_err(req, EINVAL); + return; + } + + if (req->se->op.syncfs) { + req->se->op.syncfs(req, arg->wait); + } else { + fuse_reply_err(req, ENOSYS); + } +} + static void do_init(fuse_req_t req, fuse_ino_t nodeid, struct fuse_mbuf_iter *iter) { @@ -2267,6 +2285,7 @@ static struct { [FUSE_RENAME2] = { do_rename2, "RENAME2" }, [FUSE_COPY_FILE_RANGE] = { do_copy_file_range, "COPY_FILE_RANGE" }, [FUSE_LSEEK] = { do_lseek, "LSEEK" }, + [FUSE_SYNCFS] = { do_syncfs, "SYNCFS" }, }; #define FUSE_MAXOP (sizeof(fuse_ll_ops) / sizeof(fuse_ll_ops[0])) diff --git a/tools/virtiofsd/fuse_lowlevel.h b/tools/virtiofsd/fuse_lowlevel.h index 3bf786b03485..b5ac42c31799 100644 --- a/tools/virtiofsd/fuse_lowlevel.h +++ b/tools/virtiofsd/fuse_lowlevel.h @@ -1225,6 +1225,19 @@ struct fuse_lowlevel_ops { */ void (*lseek)(fuse_req_t req, fuse_ino_t ino, off_t off, int whence, struct fuse_file_info *fi); + + /** + * Synchronize file system content + * + * If this request is answered with an error code of ENOSYS, + * this is treated as success and future calls to syncfs() will + * succeed automatically without being sent to the filesystem + * process. + * + * @param req request handle + * @param wait whether to wait for outstanding I/Os to complete + */ + void (*syncfs)(fuse_req_t req, int wait); }; /** diff --git a/tools/virtiofsd/passthrough_ll.c b/tools/virtiofsd/passthrough_ll.c index 1553d2ef454f..6b66f3208be0 100644 --- a/tools/virtiofsd/passthrough_ll.c +++ b/tools/virtiofsd/passthrough_ll.c @@ -3124,6 +3124,34 @@ static void lo_lseek(fuse_req_t req, fuse_ino_t ino, off_t off, int whence, } } +static void lo_syncfs(fuse_req_t req, int wait) +{ + if (wait) { + struct lo_data *lo = lo_data(req); + int fd, ret; + + fd = lo_inode_open(lo, &lo->root, O_RDONLY); + if (fd < 0) { + fuse_reply_err(req, errno); + return; + } + + /* + * FIXME: this is suboptimal because it will also flush unrelated + * writes not coming from the client. This can dramatically + * increase the time spent in syncfs() if some process is + * writing lots of data on the same filesystem as virtiofsd. + */ + ret = syncfs(fd); + /* syncfs() never fails on a valid fd */ + assert(ret == 0); + + close(fd); + } + + fuse_reply_err(req, 0); +} + static void lo_destroy(void *userdata) { struct lo_data *lo = (struct lo_data *)userdata; @@ -3184,6 +3212,7 @@ static struct fuse_lowlevel_ops lo_oper = { .copy_file_range = lo_copy_file_range, #endif .lseek = lo_lseek, + .syncfs = lo_syncfs, .destroy = lo_destroy, }; diff --git a/tools/virtiofsd/passthrough_seccomp.c b/tools/virtiofsd/passthrough_seccomp.c index 62441cfcdb95..343188447901 100644 --- a/tools/virtiofsd/passthrough_seccomp.c +++ b/tools/virtiofsd/passthrough_seccomp.c @@ -107,6 +107,7 @@ static const int syscall_allowlist[] = { SCMP_SYS(set_robust_list), SCMP_SYS(setxattr), SCMP_SYS(symlinkat), + SCMP_SYS(syncfs), SCMP_SYS(time), /* Rarely needed, except on static builds */ SCMP_SYS(tgkill), SCMP_SYS(unlinkat),