diff options
author | 2015-08-04 12:17:53 -0700 | |
---|---|---|
committer | 2015-08-04 15:44:42 -0700 | |
commit | 9ca8dbcc65cfc63d6f5ef3312a33184e1d726e00 (patch) | |
tree | 1c9cafbcd35f783a87880a10f85d1a060db1a563 /kernel/fs/ocfs2/mmap.c | |
parent | 98260f3884f4a202f9ca5eabed40b1354c489b29 (diff) |
Add the rt linux 4.1.3-rt3 as base
Import the rt linux 4.1.3-rt3 as OPNFV kvm base.
It's from git://git.kernel.org/pub/scm/linux/kernel/git/rt/linux-rt-devel.git linux-4.1.y-rt and
the base is:
commit 0917f823c59692d751951bf5ea699a2d1e2f26a2
Author: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Date: Sat Jul 25 12:13:34 2015 +0200
Prepare v4.1.3-rt3
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
We lose all the git history this way and it's not good. We
should apply another opnfv project repo in future.
Change-Id: I87543d81c9df70d99c5001fbdf646b202c19f423
Signed-off-by: Yunhong Jiang <yunhong.jiang@intel.com>
Diffstat (limited to 'kernel/fs/ocfs2/mmap.c')
-rw-r--r-- | kernel/fs/ocfs2/mmap.c | 193 |
1 files changed, 193 insertions, 0 deletions
diff --git a/kernel/fs/ocfs2/mmap.c b/kernel/fs/ocfs2/mmap.c new file mode 100644 index 000000000..9581d190f --- /dev/null +++ b/kernel/fs/ocfs2/mmap.c @@ -0,0 +1,193 @@ +/* -*- mode: c; c-basic-offset: 8; -*- + * vim: noexpandtab sw=8 ts=8 sts=0: + * + * mmap.c + * + * Code to deal with the mess that is clustered mmap. + * + * Copyright (C) 2002, 2004 Oracle. All rights reserved. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public + * License as published by the Free Software Foundation; either + * version 2 of the License, or (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + * + * You should have received a copy of the GNU General Public + * License along with this program; if not, write to the + * Free Software Foundation, Inc., 59 Temple Place - Suite 330, + * Boston, MA 021110-1307, USA. + */ + +#include <linux/fs.h> +#include <linux/types.h> +#include <linux/highmem.h> +#include <linux/pagemap.h> +#include <linux/uio.h> +#include <linux/signal.h> +#include <linux/rbtree.h> + +#include <cluster/masklog.h> + +#include "ocfs2.h" + +#include "aops.h" +#include "dlmglue.h" +#include "file.h" +#include "inode.h" +#include "mmap.h" +#include "super.h" +#include "ocfs2_trace.h" + + +static int ocfs2_fault(struct vm_area_struct *area, struct vm_fault *vmf) +{ + sigset_t oldset; + int ret; + + ocfs2_block_signals(&oldset); + ret = filemap_fault(area, vmf); + ocfs2_unblock_signals(&oldset); + + trace_ocfs2_fault(OCFS2_I(area->vm_file->f_mapping->host)->ip_blkno, + area, vmf->page, vmf->pgoff); + return ret; +} + +static int __ocfs2_page_mkwrite(struct file *file, struct buffer_head *di_bh, + struct page *page) +{ + int ret = VM_FAULT_NOPAGE; + struct inode *inode = file_inode(file); + struct address_space *mapping = inode->i_mapping; + loff_t pos = page_offset(page); + unsigned int len = PAGE_CACHE_SIZE; + pgoff_t last_index; + struct page *locked_page = NULL; + void *fsdata; + loff_t size = i_size_read(inode); + + last_index = (size - 1) >> PAGE_CACHE_SHIFT; + + /* + * There are cases that lead to the page no longer bebongs to the + * mapping. + * 1) pagecache truncates locally due to memory pressure. + * 2) pagecache truncates when another is taking EX lock against + * inode lock. see ocfs2_data_convert_worker. + * + * The i_size check doesn't catch the case where nodes truncated and + * then re-extended the file. We'll re-check the page mapping after + * taking the page lock inside of ocfs2_write_begin_nolock(). + * + * Let VM retry with these cases. + */ + if ((page->mapping != inode->i_mapping) || + (!PageUptodate(page)) || + (page_offset(page) >= size)) + goto out; + + /* + * Call ocfs2_write_begin() and ocfs2_write_end() to take + * advantage of the allocation code there. We pass a write + * length of the whole page (chopped to i_size) to make sure + * the whole thing is allocated. + * + * Since we know the page is up to date, we don't have to + * worry about ocfs2_write_begin() skipping some buffer reads + * because the "write" would invalidate their data. + */ + if (page->index == last_index) + len = ((size - 1) & ~PAGE_CACHE_MASK) + 1; + + ret = ocfs2_write_begin_nolock(file, mapping, pos, len, 0, &locked_page, + &fsdata, di_bh, page); + if (ret) { + if (ret != -ENOSPC) + mlog_errno(ret); + if (ret == -ENOMEM) + ret = VM_FAULT_OOM; + else + ret = VM_FAULT_SIGBUS; + goto out; + } + + if (!locked_page) { + ret = VM_FAULT_NOPAGE; + goto out; + } + ret = ocfs2_write_end_nolock(mapping, pos, len, len, locked_page, + fsdata); + BUG_ON(ret != len); + ret = VM_FAULT_LOCKED; +out: + return ret; +} + +static int ocfs2_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf) +{ + struct page *page = vmf->page; + struct inode *inode = file_inode(vma->vm_file); + struct buffer_head *di_bh = NULL; + sigset_t oldset; + int ret; + + sb_start_pagefault(inode->i_sb); + ocfs2_block_signals(&oldset); + + /* + * The cluster locks taken will block a truncate from another + * node. Taking the data lock will also ensure that we don't + * attempt page truncation as part of a downconvert. + */ + ret = ocfs2_inode_lock(inode, &di_bh, 1); + if (ret < 0) { + mlog_errno(ret); + goto out; + } + + /* + * The alloc sem should be enough to serialize with + * ocfs2_truncate_file() changing i_size as well as any thread + * modifying the inode btree. + */ + down_write(&OCFS2_I(inode)->ip_alloc_sem); + + ret = __ocfs2_page_mkwrite(vma->vm_file, di_bh, page); + + up_write(&OCFS2_I(inode)->ip_alloc_sem); + + brelse(di_bh); + ocfs2_inode_unlock(inode, 1); + +out: + ocfs2_unblock_signals(&oldset); + sb_end_pagefault(inode->i_sb); + return ret; +} + +static const struct vm_operations_struct ocfs2_file_vm_ops = { + .fault = ocfs2_fault, + .page_mkwrite = ocfs2_page_mkwrite, +}; + +int ocfs2_mmap(struct file *file, struct vm_area_struct *vma) +{ + int ret = 0, lock_level = 0; + + ret = ocfs2_inode_lock_atime(file_inode(file), + file->f_path.mnt, &lock_level); + if (ret < 0) { + mlog_errno(ret); + goto out; + } + ocfs2_inode_unlock(file_inode(file), lock_level); +out: + vma->vm_ops = &ocfs2_file_vm_ops; + return 0; +} + |