From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755916AbbHDUAi (ORCPT ); Tue, 4 Aug 2015 16:00:38 -0400 Received: from mga11.intel.com ([192.55.52.93]:13229 "EHLO mga11.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754035AbbHDT6P (ORCPT ); Tue, 4 Aug 2015 15:58:15 -0400 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.15,611,1432623600"; d="scan'208";a="761919177" From: Matthew Wilcox To: Andrew Morton , linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org, linux-mm@kvack.org Cc: Matthew Wilcox Subject: [PATCH 05/11] ext4: Start transaction before calling into DAX Date: Tue, 4 Aug 2015 15:57:59 -0400 Message-Id: <1438718285-21168-6-git-send-email-matthew.r.wilcox@intel.com> X-Mailer: git-send-email 2.4.6 In-Reply-To: <1438718285-21168-1-git-send-email-matthew.r.wilcox@intel.com> References: <1438718285-21168-1-git-send-email-matthew.r.wilcox@intel.com> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org From: Matthew Wilcox Jan Kara pointed out that in the case where we are writing to a hole, we can end up with a lock inversion between the page lock and the journal lock. We can avoid this by starting the transaction in ext4 before calling into DAX. The journal lock nests inside the superblock pagefault lock, so we have to duplicate that code from dax_fault, like XFS does. Signed-off-by: Matthew Wilcox --- fs/ext4/file.c | 55 ++++++++++++++++++++++++++++++++++++++++++++++++++++--- 1 file changed, 52 insertions(+), 3 deletions(-) diff --git a/fs/ext4/file.c b/fs/ext4/file.c index d5219e4..113837e 100644 --- a/fs/ext4/file.c +++ b/fs/ext4/file.c @@ -207,14 +207,63 @@ static void ext4_end_io_unwritten(struct buffer_head *bh, int uptodate) static int ext4_dax_fault(struct vm_area_struct *vma, struct vm_fault *vmf) { - return dax_fault(vma, vmf, ext4_get_block_dax, ext4_end_io_unwritten); + int result; + handle_t *handle = NULL; + struct super_block *sb = file_inode(vma->vm_file)->i_sb; + bool write = vmf->flags & FAULT_FLAG_WRITE; + + if (write) { + sb_start_pagefault(sb); + file_update_time(vma->vm_file); + handle = ext4_journal_start_sb(sb, EXT4_HT_WRITE_PAGE, + EXT4_DATA_TRANS_BLOCKS(sb)); + } + + if (IS_ERR(handle)) + result = VM_FAULT_SIGBUS; + else + result = __dax_fault(vma, vmf, ext4_get_block_dax, + ext4_end_io_unwritten); + + if (write) { + if (!IS_ERR(handle)) + ext4_journal_stop(handle); + sb_end_pagefault(sb); + } + + return result; } static int ext4_dax_pmd_fault(struct vm_area_struct *vma, unsigned long addr, pmd_t *pmd, unsigned int flags) { - return dax_pmd_fault(vma, addr, pmd, flags, ext4_get_block_dax, - ext4_end_io_unwritten); + int result; + handle_t *handle = NULL; + struct inode *inode = file_inode(vma->vm_file); + struct super_block *sb = inode->i_sb; + bool write = flags & FAULT_FLAG_WRITE; + + if (write) { + sb_start_pagefault(sb); + file_update_time(vma->vm_file); + handle = ext4_journal_start_sb(sb, EXT4_HT_WRITE_PAGE, + ext4_chunk_trans_blocks(inode, + PMD_SIZE / PAGE_SIZE)); + } + + if (IS_ERR(handle)) + result = VM_FAULT_SIGBUS; + else + result = __dax_pmd_fault(vma, addr, pmd, flags, + ext4_get_block_dax, ext4_end_io_unwritten); + + if (write) { + if (!IS_ERR(handle)) + ext4_journal_stop(handle); + sb_end_pagefault(sb); + } + + return result; } static int ext4_dax_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf) -- 2.1.4