From patchwork Wed Oct 30 04:13:54 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Ruan Shiyang X-Patchwork-Id: 11218987 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 6EA99912 for ; Wed, 30 Oct 2019 04:14:28 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 4C96B21479 for ; Wed, 30 Oct 2019 04:14:28 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1727283AbfJ3EOX (ORCPT ); Wed, 30 Oct 2019 00:14:23 -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 S1727222AbfJ3EOX (ORCPT ); Wed, 30 Oct 2019 00:14:23 -0400 X-IronPort-AV: E=Sophos;i="5.68,245,1569254400"; d="scan'208";a="77665116" 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 786AD4CE1510; Wed, 30 Oct 2019 12:06:22 +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 12:14:30 +0800 From: Shiyang Ruan To: , , , , , CC: , , , , , Goldwyn Rodrigues Subject: [RFC PATCH v2 3/7] dax: replace mmap entry in case of CoW Date: Wed, 30 Oct 2019 12:13:54 +0800 Message-ID: <20191030041358.14450-4-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: 786AD4CE1510.A420B 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 From: Goldwyn Rodrigues We replace the existing entry to the newly allocated one in case of CoW. Also, we mark the entry as PAGECACHE_TAG_TOWRITE so writeback marks this entry as writeprotected. This helps us snapshots so new write pagefaults after snapshots trigger a CoW. btrfs does not support hugepages so we don't handle PMD. Signed-off-by: Goldwyn Rodrigues --- fs/dax.c | 40 ++++++++++++++++++++++++++++++---------- 1 file changed, 30 insertions(+), 10 deletions(-) diff --git a/fs/dax.c b/fs/dax.c index 949a40cf1fe7..e6174c0c669a 100644 --- a/fs/dax.c +++ b/fs/dax.c @@ -705,6 +705,9 @@ static int copy_user_dax(struct block_device *bdev, struct dax_device *dax_dev, return 0; } +#define DAX_IF_DIRTY (1ULL << 0) +#define DAX_IF_COW (1ULL << 1) + /* * By this point grab_mapping_entry() has ensured that we have a locked entry * of the appropriate size so we don't have to worry about downgrading PMDs to @@ -714,14 +717,17 @@ static int copy_user_dax(struct block_device *bdev, struct dax_device *dax_dev, */ static void *dax_insert_entry(struct xa_state *xas, struct address_space *mapping, struct vm_fault *vmf, - void *entry, pfn_t pfn, unsigned long flags, bool dirty) + void *entry, pfn_t pfn, unsigned long flags, + unsigned int insert_flags) { void *new_entry = dax_make_entry(pfn, flags); + bool dirty = insert_flags & DAX_IF_DIRTY; + bool cow = insert_flags & DAX_IF_COW; if (dirty) __mark_inode_dirty(mapping->host, I_DIRTY_PAGES); - if (dax_is_zero_entry(entry) && !(flags & DAX_ZERO_PAGE)) { + if (cow || (dax_is_zero_entry(entry) && !(flags & DAX_ZERO_PAGE))) { unsigned long index = xas->xa_index; /* we are replacing a zero page with block mapping */ if (dax_is_pmd_entry(entry)) @@ -733,7 +739,7 @@ static void *dax_insert_entry(struct xa_state *xas, xas_reset(xas); xas_lock_irq(xas); - if (dax_is_zero_entry(entry) || dax_is_empty_entry(entry)) { + if (cow || dax_is_zero_entry(entry) || dax_is_empty_entry(entry)) { void *old; dax_disassociate_entry(entry, mapping, false); @@ -757,6 +763,9 @@ static void *dax_insert_entry(struct xa_state *xas, if (dirty) xas_set_mark(xas, PAGECACHE_TAG_DIRTY); + if (cow) + xas_set_mark(xas, PAGECACHE_TAG_TOWRITE); + xas_unlock_irq(xas); return entry; } @@ -1044,7 +1053,7 @@ static vm_fault_t dax_load_hole(struct xa_state *xas, vm_fault_t ret; *entry = dax_insert_entry(xas, mapping, vmf, *entry, pfn, - DAX_ZERO_PAGE, false); + DAX_ZERO_PAGE, 0); ret = vmf_insert_mixed(vmf->vma, vaddr, pfn); trace_dax_load_hole(inode, vmf, ret); @@ -1313,6 +1322,7 @@ static vm_fault_t dax_iomap_pte_fault(struct vm_fault *vmf, pfn_t *pfnp, void *entry; pfn_t pfn; void *kaddr; + unsigned long insert_flags = 0; trace_dax_pte_fault(inode, vmf, ret); /* @@ -1405,8 +1415,10 @@ static vm_fault_t dax_iomap_pte_fault(struct vm_fault *vmf, pfn_t *pfnp, if (error < 0) goto error_finish_iomap; - entry = dax_insert_entry(&xas, mapping, vmf, entry, pfn, - 0, write && !sync); + if (write && !sync) + insert_flags |= DAX_IF_DIRTY; + entry = dax_insert_entry(&xas, mapping, vmf, entry, pfn, 0, + insert_flags); if (srcmap.type != IOMAP_HOLE) { error = dax_copy_edges(pos, PAGE_SIZE, &srcmap, kaddr, @@ -1438,8 +1450,10 @@ 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) + if (srcmap.type != IOMAP_HOLE) { + insert_flags |= DAX_IF_COW; goto cow; + } /*FALLTHRU*/ case IOMAP_HOLE: if (!write) { @@ -1497,7 +1511,7 @@ static vm_fault_t dax_pmd_load_hole(struct xa_state *xas, struct vm_fault *vmf, pfn = page_to_pfn_t(zero_page); *entry = dax_insert_entry(xas, mapping, vmf, *entry, pfn, - DAX_PMD | DAX_ZERO_PAGE, false); + DAX_PMD | DAX_ZERO_PAGE, 0); if (arch_needs_pgtable_deposit()) { pgtable = pte_alloc_one(vma->vm_mm); @@ -1549,6 +1563,7 @@ static vm_fault_t dax_iomap_pmd_fault(struct vm_fault *vmf, pfn_t *pfnp, int error; pfn_t pfn; void *kaddr; + unsigned long insert_flags = 0; /* * Check whether offset isn't beyond end of file now. Caller is @@ -1636,8 +1651,11 @@ static vm_fault_t dax_iomap_pmd_fault(struct vm_fault *vmf, pfn_t *pfnp, if (error < 0) goto finish_iomap; + if (write && !sync) + insert_flags |= DAX_IF_DIRTY; + entry = dax_insert_entry(&xas, mapping, vmf, entry, pfn, - DAX_PMD, write && !sync); + DAX_PMD, insert_flags); if (srcmap.type != IOMAP_HOLE) { error = dax_copy_edges(pos, PMD_SIZE, &srcmap, kaddr, @@ -1664,8 +1682,10 @@ 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) + if (srcmap.type != IOMAP_HOLE) { + insert_flags |= DAX_IF_COW; goto cow; + } /*FALLTHRU*/ case IOMAP_HOLE: if (WARN_ON_ONCE(write))