From patchwork Mon Mar 7 19:25:14 2016 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Markus Armbruster X-Patchwork-Id: 8521871 Return-Path: X-Original-To: patchwork-qemu-devel@patchwork.kernel.org Delivered-To: patchwork-parsemail@patchwork1.web.kernel.org Received: from mail.kernel.org (mail.kernel.org [198.145.29.136]) by patchwork1.web.kernel.org (Postfix) with ESMTP id CECAC9F46A for ; Mon, 7 Mar 2016 19:32:13 +0000 (UTC) Received: from mail.kernel.org (localhost [127.0.0.1]) by mail.kernel.org (Postfix) with ESMTP id 29EFE202C8 for ; Mon, 7 Mar 2016 19:32:13 +0000 (UTC) Received: from lists.gnu.org (lists.gnu.org [208.118.235.17]) (using TLSv1 with cipher AES256-SHA (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id 1F9A620295 for ; Mon, 7 Mar 2016 19:32:12 +0000 (UTC) Received: from localhost ([::1]:57944 helo=lists.gnu.org) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1ad0t1-0006rl-H9 for patchwork-qemu-devel@patchwork.kernel.org; Mon, 07 Mar 2016 14:32:11 -0500 Received: from eggs.gnu.org ([2001:4830:134:3::10]:44629) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1ad0n6-00043M-Rw for qemu-devel@nongnu.org; Mon, 07 Mar 2016 14:26:07 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1ad0n2-0003se-7y for qemu-devel@nongnu.org; Mon, 07 Mar 2016 14:26:04 -0500 Received: from mx1.redhat.com ([209.132.183.28]:36238) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1ad0n2-0003s2-0l for qemu-devel@nongnu.org; Mon, 07 Mar 2016 14:26:00 -0500 Received: from int-mx13.intmail.prod.int.phx2.redhat.com (int-mx13.intmail.prod.int.phx2.redhat.com [10.5.11.26]) by mx1.redhat.com (Postfix) with ESMTPS id B4A1B155DC; Mon, 7 Mar 2016 19:25:57 +0000 (UTC) Received: from blackfin.pond.sub.org (ovpn-116-61.ams2.redhat.com [10.36.116.61]) by int-mx13.intmail.prod.int.phx2.redhat.com (8.14.4/8.14.4) with ESMTP id u27JPtvF005218 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-GCM-SHA384 bits=256 verify=NO); Mon, 7 Mar 2016 14:25:57 -0500 Received: by blackfin.pond.sub.org (Postfix, from userid 1000) id 17681303E59A; Mon, 7 Mar 2016 20:25:55 +0100 (CET) From: Markus Armbruster To: qemu-devel@nongnu.org Date: Mon, 7 Mar 2016 20:25:14 +0100 Message-Id: <1457378754-21649-3-git-send-email-armbru@redhat.com> In-Reply-To: <1457378754-21649-1-git-send-email-armbru@redhat.com> References: <1457378754-21649-1-git-send-email-armbru@redhat.com> X-Scanned-By: MIMEDefang 2.68 on 10.5.11.26 X-detected-operating-system: by eggs.gnu.org: GNU/Linux 3.x X-Received-From: 209.132.183.28 Cc: claudio.fontana@huawei.com, cam@cs.ualberta.ca, mlureau@redhat.com, david.marchand@6wind.com, pbonzini@redhat.com Subject: [Qemu-devel] [PATCH v2 02/42] exec: Fix memory allocation when memory path isn't on hugetlbfs X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.14 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: qemu-devel-bounces+patchwork-qemu-devel=patchwork.kernel.org@nongnu.org Sender: qemu-devel-bounces+patchwork-qemu-devel=patchwork.kernel.org@nongnu.org X-Spam-Status: No, score=-6.9 required=5.0 tests=BAYES_00, RCVD_IN_DNSWL_HI, UNPARSEABLE_RELAY autolearn=unavailable version=3.3.1 X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on mail.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP gethugepagesize() works reliably only when its argument is on hugetlbfs. When it's not, it returns the filesystem's "optimal transfer block size", which may or may not be the actual page size you'll get when you mmap(). If the value is too small or not a power of two, we fail qemu_ram_mmap()'s assertions. These were added in commit 794e8f3 (v2.5.0). The bug's impact before that is currently unknown. Seems fairly unlikely at least when the normal page size is 4KiB. Else, if the value is too large, we align more strictly than necessary. gethugepagesize() goes back to commit c902760 (v0.13). That commit clearly intended gethugepagesize() to be used on hugetlbfs only. Not only was it named accordingly, it also printed a warning when used on anything else. However, the commit neglected to spell out the restriction in user documentation of -mem-path. Commit bfc2a1a (v2.5.0) dropped the warning as bogus "because QEMU functions perfectly well with the path on a regular tmpfs filesystem". It sure does when you're sufficiently lucky. In my testing, I was lucky, too. Fix by switching to qemu_fd_getpagesize(). Rename the variable holding its result from hpagesize to page_size. Cc: Paolo Bonzini Signed-off-by: Markus Armbruster --- exec.c | 40 +++++++--------------------------------- 1 file changed, 7 insertions(+), 33 deletions(-) diff --git a/exec.c b/exec.c index 5275ff4..d41194e 100644 --- a/exec.c +++ b/exec.c @@ -1207,27 +1207,6 @@ void qemu_mutex_unlock_ramlist(void) } #ifdef __linux__ - -#include - -#define HUGETLBFS_MAGIC 0x958458f6 - -static long gethugepagesize(int fd) -{ - struct statfs fs; - int ret; - - do { - ret = fstatfs(fd, &fs); - } while (ret != 0 && errno == EINTR); - - if (ret != 0) { - return -1; - } - - return fs.f_bsize; -} - static void *file_ram_alloc(RAMBlock *block, ram_addr_t memory, const char *path, @@ -1239,7 +1218,7 @@ static void *file_ram_alloc(RAMBlock *block, char *c; void *area; int fd; - int64_t hpagesize; + int64_t page_size; if (kvm_enabled() && !kvm_has_sync_mmu()) { error_setg(errp, @@ -1294,22 +1273,17 @@ static void *file_ram_alloc(RAMBlock *block, */ } - hpagesize = gethugepagesize(fd); - if (hpagesize < 0) { - error_setg_errno(errp, errno, "can't get page size for %s", - path); - goto error; - } - block->mr->align = hpagesize; + page_size = qemu_fd_getpagesize(fd); + block->mr->align = page_size; - if (memory < hpagesize) { + if (memory < page_size) { error_setg(errp, "memory size 0x" RAM_ADDR_FMT " must be equal to " "or larger than page size 0x%" PRIx64, - memory, hpagesize); + memory, page_size); goto error; } - memory = ROUND_UP(memory, hpagesize); + memory = ROUND_UP(memory, page_size); /* * ftruncate is not supported by hugetlbfs in older @@ -1321,7 +1295,7 @@ static void *file_ram_alloc(RAMBlock *block, perror("ftruncate"); } - area = qemu_ram_mmap(fd, memory, hpagesize, block->flags & RAM_SHARED); + area = qemu_ram_mmap(fd, memory, page_size, block->flags & RAM_SHARED); if (area == MAP_FAILED) { error_setg_errno(errp, errno, "unable to map backing store for guest RAM");