From patchwork Wed Oct 30 04:13:53 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Ruan Shiyang X-Patchwork-Id: 11218979 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 514B8912 for ; Wed, 30 Oct 2019 04:14:23 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 38EC4208E3 for ; Wed, 30 Oct 2019 04:14:23 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1727189AbfJ3EOW (ORCPT ); Wed, 30 Oct 2019 00:14:22 -0400 Received: from mail.cn.fujitsu.com ([183.91.158.132]:4003 "EHLO heian.cn.fujitsu.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1727183AbfJ3EOW (ORCPT ); Wed, 30 Oct 2019 00:14:22 -0400 X-IronPort-AV: E=Sophos;i="5.68,245,1569254400"; d="scan'208";a="77665115" Received: from unknown (HELO cn.fujitsu.com) ([10.167.33.5]) by heian.cn.fujitsu.com with ESMTP; 30 Oct 2019 12:14:20 +0800 Received: from G08CNEXCHPEKD01.g08.fujitsu.local (unknown [10.167.33.80]) by cn.fujitsu.com (Postfix) with ESMTP id 5D4BF4B6EC86; Wed, 30 Oct 2019 12:06:20 +0800 (CST) Received: from localhost.localdomain (10.167.225.140) by G08CNEXCHPEKD01.g08.fujitsu.local (10.167.33.89) with Microsoft SMTP Server (TLS) id 14.3.439.0; Wed, 30 Oct 2019 12:14:28 +0800 From: Shiyang Ruan To: , , , , , CC: , , , , Subject: [RFC PATCH v2 2/7] dax: copy data before write. Date: Wed, 30 Oct 2019 12:13:53 +0800 Message-ID: <20191030041358.14450-3-ruansy.fnst@cn.fujitsu.com> X-Mailer: git-send-email 2.17.0 In-Reply-To: <20191030041358.14450-1-ruansy.fnst@cn.fujitsu.com> References: <20191030041358.14450-1-ruansy.fnst@cn.fujitsu.com> MIME-Version: 1.0 X-Originating-IP: [10.167.225.140] X-yoursite-MailScanner-ID: 5D4BF4B6EC86.A86AA X-yoursite-MailScanner: Found to be clean X-yoursite-MailScanner-From: ruansy.fnst@cn.fujitsu.com X-Spam-Status: No Sender: linux-xfs-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-xfs@vger.kernel.org Add dax_copy_edges() into each dax actor functions to perform copy on write. Signed-off-by: Shiyang Ruan --- fs/dax.c | 36 +++++++++++++++++++++++++++++++++--- 1 file changed, 33 insertions(+), 3 deletions(-) diff --git a/fs/dax.c b/fs/dax.c index e1f4493ce56a..949a40cf1fe7 100644 --- a/fs/dax.c +++ b/fs/dax.c @@ -1159,7 +1159,7 @@ dax_iomap_actor(struct inode *inode, loff_t pos, loff_t length, void *data, return iov_iter_zero(min(length, end - pos), iter); } - if (WARN_ON_ONCE(iomap->type != IOMAP_MAPPED)) + if (WARN_ON_ONCE(iomap->type != IOMAP_MAPPED && iomap == srcmap)) return -EIO; /* @@ -1198,6 +1198,12 @@ dax_iomap_actor(struct inode *inode, loff_t pos, loff_t length, void *data, break; } + if (iomap != srcmap) { + ret = dax_copy_edges(pos, length, srcmap, kaddr, false); + if (ret) + break; + } + map_len = PFN_PHYS(map_len); kaddr += offset; map_len -= offset; @@ -1306,6 +1312,7 @@ static vm_fault_t dax_iomap_pte_fault(struct vm_fault *vmf, pfn_t *pfnp, vm_fault_t ret = 0; void *entry; pfn_t pfn; + void *kaddr; trace_dax_pte_fault(inode, vmf, ret); /* @@ -1387,19 +1394,27 @@ static vm_fault_t dax_iomap_pte_fault(struct vm_fault *vmf, pfn_t *pfnp, switch (iomap.type) { case IOMAP_MAPPED: +cow: if (iomap.flags & IOMAP_F_NEW) { count_vm_event(PGMAJFAULT); count_memcg_event_mm(vma->vm_mm, PGMAJFAULT); major = VM_FAULT_MAJOR; } error = dax_iomap_direct_access(&iomap, pos, PAGE_SIZE, &pfn, - NULL); + &kaddr); if (error < 0) goto error_finish_iomap; entry = dax_insert_entry(&xas, mapping, vmf, entry, pfn, 0, write && !sync); + if (srcmap.type != IOMAP_HOLE) { + error = dax_copy_edges(pos, PAGE_SIZE, &srcmap, kaddr, + false); + if (error) + goto error_finish_iomap; + } + /* * If we are doing synchronous page fault and inode needs fsync, * we can insert PTE into page tables only after that happens. @@ -1423,6 +1438,9 @@ static vm_fault_t dax_iomap_pte_fault(struct vm_fault *vmf, pfn_t *pfnp, goto finish_iomap; case IOMAP_UNWRITTEN: + if (srcmap.type != IOMAP_HOLE) + goto cow; + /*FALLTHRU*/ case IOMAP_HOLE: if (!write) { ret = dax_load_hole(&xas, mapping, &entry, vmf); @@ -1530,6 +1548,7 @@ static vm_fault_t dax_iomap_pmd_fault(struct vm_fault *vmf, pfn_t *pfnp, loff_t pos; int error; pfn_t pfn; + void *kaddr; /* * Check whether offset isn't beyond end of file now. Caller is @@ -1610,15 +1629,23 @@ static vm_fault_t dax_iomap_pmd_fault(struct vm_fault *vmf, pfn_t *pfnp, sync = dax_fault_is_synchronous(iomap_flags, vma, &iomap); switch (iomap.type) { +cow: case IOMAP_MAPPED: error = dax_iomap_direct_access(&iomap, pos, PMD_SIZE, &pfn, - NULL); + &kaddr); if (error < 0) goto finish_iomap; entry = dax_insert_entry(&xas, mapping, vmf, entry, pfn, DAX_PMD, write && !sync); + if (srcmap.type != IOMAP_HOLE) { + error = dax_copy_edges(pos, PMD_SIZE, &srcmap, kaddr, + true); + if (error) + goto unlock_entry; + } + /* * If we are doing synchronous page fault and inode needs fsync, * we can insert PMD into page tables only after that happens. @@ -1637,6 +1664,9 @@ static vm_fault_t dax_iomap_pmd_fault(struct vm_fault *vmf, pfn_t *pfnp, result = vmf_insert_pfn_pmd(vmf, pfn, write); break; case IOMAP_UNWRITTEN: + if (srcmap.type != IOMAP_HOLE) + goto cow; + /*FALLTHRU*/ case IOMAP_HOLE: if (WARN_ON_ONCE(write)) break;