All of lore.kernel.org
 help / color / mirror / Atom feed
From: Matthew Wilcox <willy@linux.intel.com>
To: Jan Kara <jack@suse.cz>
Cc: Matthew Wilcox <matthew.r.wilcox@intel.com>,
	linux-fsdevel@vger.kernel.org, linux-mm@kvack.org,
	linux-kernel@vger.kernel.org
Subject: Re: [PATCH v7 07/22] Replace the XIP page fault handler with the DAX page fault handler
Date: Wed, 30 Jul 2014 17:02:40 -0400	[thread overview]
Message-ID: <20140730210239.GS6754@linux.intel.com> (raw)
In-Reply-To: <20140730095229.GA19205@quack.suse.cz>

On Wed, Jul 30, 2014 at 11:52:29AM +0200, Jan Kara wrote:
>   I see the problem now. How about an attached patch? Do you see other
> lockdep warnings with it?

This patch fixes the problem, thanks!  Regardless of DAX, I think this
patch should be applied in order to avoid creating a dependency between
i_mmap_mutex and jbd2_handle.

I've now run into a different problem with COW pages ... more later.

> >From c01c905cf3c4c6304a5ea9836389d9cf0d575884 Mon Sep 17 00:00:00 2001
> From: Jan Kara <jack@suse.cz>
> Date: Wed, 30 Jul 2014 11:49:07 +0200
> Subject: [PATCH] ext4: Avoid lock inversion between i_mmap_mutex and
>  transaction start
> 
> When DAX is enabled, it uses i_mmap_mutex as a protection against
> truncate during page fault. This inevitably forces i_mmap_mutex to rank
> outside of a transaction start and thus we have to avoid calling
> pagecache purging operations when transaction is started.
> 
> Signed-off-by: Jan Kara <jack@suse.cz>
> ---
>  fs/ext4/inode.c | 14 ++++++++++----
>  1 file changed, 10 insertions(+), 4 deletions(-)
> 
> diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
> index 8a064734e6eb..494a8645d63e 100644
> --- a/fs/ext4/inode.c
> +++ b/fs/ext4/inode.c
> @@ -3631,13 +3631,19 @@ int ext4_punch_hole(struct inode *inode, loff_t offset, loff_t length)
>  	if (IS_SYNC(inode))
>  		ext4_handle_sync(handle);
>  
> -	/* Now release the pages again to reduce race window */
> +	inode->i_mtime = inode->i_ctime = ext4_current_time(inode);
> +	ext4_mark_inode_dirty(handle, inode);
> +	ext4_journal_stop(handle);
> +
> +	/*
> +	 * Now release the pages again to reduce race window. This has to happen
> +	 * outside of a transaction to avoid lock inversion on i_mmap_mutex
> +	 * when DAX is enabled.
> +	 */
>  	if (last_block_offset > first_block_offset)
>  		truncate_pagecache_range(inode, first_block_offset,
>  					 last_block_offset);
> -
> -	inode->i_mtime = inode->i_ctime = ext4_current_time(inode);
> -	ext4_mark_inode_dirty(handle, inode);
> +	goto out_dio;
>  out_stop:
>  	ext4_journal_stop(handle);
>  out_dio:
> -- 
> 1.8.1.4
> 

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

WARNING: multiple messages have this Message-ID (diff)
From: Matthew Wilcox <willy@linux.intel.com>
To: Jan Kara <jack@suse.cz>
Cc: Matthew Wilcox <matthew.r.wilcox@intel.com>,
	linux-fsdevel@vger.kernel.org, linux-mm@kvack.org,
	linux-kernel@vger.kernel.org
Subject: Re: [PATCH v7 07/22] Replace the XIP page fault handler with the DAX page fault handler
Date: Wed, 30 Jul 2014 17:02:40 -0400	[thread overview]
Message-ID: <20140730210239.GS6754@linux.intel.com> (raw)
In-Reply-To: <20140730095229.GA19205@quack.suse.cz>

On Wed, Jul 30, 2014 at 11:52:29AM +0200, Jan Kara wrote:
>   I see the problem now. How about an attached patch? Do you see other
> lockdep warnings with it?

This patch fixes the problem, thanks!  Regardless of DAX, I think this
patch should be applied in order to avoid creating a dependency between
i_mmap_mutex and jbd2_handle.

I've now run into a different problem with COW pages ... more later.

> >From c01c905cf3c4c6304a5ea9836389d9cf0d575884 Mon Sep 17 00:00:00 2001
> From: Jan Kara <jack@suse.cz>
> Date: Wed, 30 Jul 2014 11:49:07 +0200
> Subject: [PATCH] ext4: Avoid lock inversion between i_mmap_mutex and
>  transaction start
> 
> When DAX is enabled, it uses i_mmap_mutex as a protection against
> truncate during page fault. This inevitably forces i_mmap_mutex to rank
> outside of a transaction start and thus we have to avoid calling
> pagecache purging operations when transaction is started.
> 
> Signed-off-by: Jan Kara <jack@suse.cz>
> ---
>  fs/ext4/inode.c | 14 ++++++++++----
>  1 file changed, 10 insertions(+), 4 deletions(-)
> 
> diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
> index 8a064734e6eb..494a8645d63e 100644
> --- a/fs/ext4/inode.c
> +++ b/fs/ext4/inode.c
> @@ -3631,13 +3631,19 @@ int ext4_punch_hole(struct inode *inode, loff_t offset, loff_t length)
>  	if (IS_SYNC(inode))
>  		ext4_handle_sync(handle);
>  
> -	/* Now release the pages again to reduce race window */
> +	inode->i_mtime = inode->i_ctime = ext4_current_time(inode);
> +	ext4_mark_inode_dirty(handle, inode);
> +	ext4_journal_stop(handle);
> +
> +	/*
> +	 * Now release the pages again to reduce race window. This has to happen
> +	 * outside of a transaction to avoid lock inversion on i_mmap_mutex
> +	 * when DAX is enabled.
> +	 */
>  	if (last_block_offset > first_block_offset)
>  		truncate_pagecache_range(inode, first_block_offset,
>  					 last_block_offset);
> -
> -	inode->i_mtime = inode->i_ctime = ext4_current_time(inode);
> -	ext4_mark_inode_dirty(handle, inode);
> +	goto out_dio;
>  out_stop:
>  	ext4_journal_stop(handle);
>  out_dio:
> -- 
> 1.8.1.4
> 


  reply	other threads:[~2014-07-30 21:02 UTC|newest]

Thread overview: 180+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-03-23 19:08 [PATCH v7 00/22] Support ext4 on NV-DIMMs Matthew Wilcox
2014-03-23 19:08 ` Matthew Wilcox
2014-03-23 19:08 ` [PATCH v7 01/22] Fix XIP fault vs truncate race Matthew Wilcox
2014-03-23 19:08   ` Matthew Wilcox
2014-03-29 15:57   ` Jan Kara
2014-03-29 15:57     ` Jan Kara
2014-03-23 19:08 ` [PATCH v7 02/22] Allow page fault handlers to perform the COW Matthew Wilcox
2014-03-23 19:08   ` Matthew Wilcox
2014-04-08 16:34   ` Jan Kara
2014-04-08 16:34     ` Jan Kara
2014-03-23 19:08 ` [PATCH v7 03/22] axonram: Fix bug in direct_access Matthew Wilcox
2014-03-23 19:08   ` Matthew Wilcox
2014-03-29 16:22   ` Jan Kara
2014-03-29 16:22     ` Jan Kara
2014-04-02 19:24     ` Matthew Wilcox
2014-04-02 19:24       ` Matthew Wilcox
2014-03-23 19:08 ` [PATCH v7 04/22] Change direct_access calling convention Matthew Wilcox
2014-03-23 19:08   ` Matthew Wilcox
2014-03-29 16:30   ` Jan Kara
2014-03-29 16:30     ` Jan Kara
2014-04-02 19:27     ` Matthew Wilcox
2014-04-02 19:27       ` Matthew Wilcox
2014-03-23 19:08 ` [PATCH v7 05/22] Introduce IS_DAX(inode) Matthew Wilcox
2014-03-23 19:08   ` Matthew Wilcox
2014-04-08 15:32   ` Jan Kara
2014-04-08 15:32     ` Jan Kara
2014-03-23 19:08 ` [PATCH v7 06/22] Replace XIP read and write with DAX I/O Matthew Wilcox
2014-03-23 19:08   ` Matthew Wilcox
2014-04-08 17:56   ` Jan Kara
2014-04-08 17:56     ` Jan Kara
2014-04-08 20:21     ` Matthew Wilcox
2014-04-08 20:21       ` Matthew Wilcox
2014-04-09  9:14       ` Jan Kara
2014-04-09  9:14         ` Jan Kara
2014-04-09 15:19         ` Matthew Wilcox
2014-04-09 15:19           ` Matthew Wilcox
2014-04-09 20:55           ` Jan Kara
2014-04-09 20:55             ` Jan Kara
2014-04-13 18:05             ` Matthew Wilcox
2014-04-13 18:05               ` Matthew Wilcox
2014-04-09 12:04   ` Jan Kara
2014-04-09 12:04     ` Jan Kara
2014-03-23 19:08 ` [PATCH v7 07/22] Replace the XIP page fault handler with the DAX page fault handler Matthew Wilcox
2014-03-23 19:08   ` Matthew Wilcox
2014-04-08 22:05   ` Jan Kara
2014-04-08 22:05     ` Jan Kara
2014-04-09 20:48     ` Matthew Wilcox
2014-04-09 20:48       ` Matthew Wilcox
2014-04-09 21:12       ` Jan Kara
2014-04-09 21:12         ` Jan Kara
2014-04-13 11:21         ` Matthew Wilcox
2014-04-13 11:21           ` Matthew Wilcox
2014-04-14 16:04           ` Jan Kara
2014-04-14 16:04             ` Jan Kara
2014-04-09 10:27   ` Jan Kara
2014-04-09 10:27     ` Jan Kara
2014-04-09 20:51     ` Matthew Wilcox
2014-04-09 20:51       ` Matthew Wilcox
2014-04-09 21:43       ` Jan Kara
2014-04-09 21:43         ` Jan Kara
2014-04-13 18:03         ` Matthew Wilcox
2014-04-13 18:03           ` Matthew Wilcox
2014-07-29 12:12         ` Matthew Wilcox
2014-07-29 12:12           ` Matthew Wilcox
2014-07-29 21:04           ` Jan Kara
2014-07-29 21:04             ` Jan Kara
2014-07-29 21:23             ` Matthew Wilcox
2014-07-29 21:23               ` Matthew Wilcox
2014-07-30  9:52               ` Jan Kara
2014-07-30  9:52                 ` Jan Kara
2014-07-30 21:02                 ` Matthew Wilcox [this message]
2014-07-30 21:02                   ` Matthew Wilcox
2014-08-09 11:00                 ` Matthew Wilcox
2014-08-09 11:00                   ` Matthew Wilcox
2014-08-11  8:51                   ` Jan Kara
2014-08-11  8:51                     ` Jan Kara
2014-08-11 14:13                     ` Matthew Wilcox
2014-08-11 14:13                       ` Matthew Wilcox
2014-08-11 14:35                       ` Jan Kara
2014-08-11 14:35                         ` Jan Kara
2014-08-11 15:02                         ` Matthew Wilcox
2014-08-11 15:02                           ` Matthew Wilcox
2014-08-11 15:25                           ` Jan Kara
2014-08-11 15:25                             ` Jan Kara
2014-05-21 20:35   ` Toshi Kani
2014-05-21 20:35     ` Toshi Kani
2014-06-05 22:38     ` Toshi Kani
2014-06-05 22:38       ` Toshi Kani
2014-03-23 19:08 ` [PATCH v7 08/22] Replace xip_truncate_page with dax_truncate_page Matthew Wilcox
2014-03-23 19:08   ` Matthew Wilcox
2014-04-08 22:17   ` Jan Kara
2014-04-08 22:17     ` Jan Kara
2014-04-09  9:26     ` Jan Kara
2014-04-09  9:26       ` Jan Kara
2014-04-13 19:07       ` Matthew Wilcox
2014-04-13 19:07         ` Matthew Wilcox
2014-03-23 19:08 ` [PATCH v7 09/22] Remove mm/filemap_xip.c Matthew Wilcox
2014-03-23 19:08   ` Matthew Wilcox
2014-04-08 18:21   ` Jan Kara
2014-04-08 18:21     ` Jan Kara
2014-03-23 19:08 ` [PATCH v7 10/22] Remove get_xip_mem Matthew Wilcox
2014-03-23 19:08   ` Matthew Wilcox
2014-04-08 18:20   ` Jan Kara
2014-04-08 18:20     ` Jan Kara
2014-03-23 19:08 ` [PATCH v7 11/22] Replace ext2_clear_xip_target with dax_clear_blocks Matthew Wilcox
2014-03-23 19:08   ` Matthew Wilcox
2014-04-09  9:46   ` Jan Kara
2014-04-09  9:46     ` Jan Kara
2014-04-10 14:16     ` Matthew Wilcox
2014-04-10 14:16       ` Matthew Wilcox
2014-04-10 18:31       ` Jan Kara
2014-04-10 18:31         ` Jan Kara
2014-03-23 19:08 ` [PATCH v7 12/22] ext2: Remove ext2_xip_verify_sb() Matthew Wilcox
2014-03-23 19:08   ` Matthew Wilcox
2014-04-09  9:52   ` Jan Kara
2014-04-09  9:52     ` Jan Kara
2014-04-10 14:22     ` Matthew Wilcox
2014-04-10 14:22       ` Matthew Wilcox
2014-04-10 18:35       ` Jan Kara
2014-04-10 18:35         ` Jan Kara
2014-03-23 19:08 ` [PATCH v7 13/22] ext2: Remove ext2_use_xip Matthew Wilcox
2014-03-23 19:08   ` Matthew Wilcox
2014-04-09  9:55   ` Jan Kara
2014-04-09  9:55     ` Jan Kara
2014-03-23 19:08 ` [PATCH v7 14/22] ext2: Remove xip.c and xip.h Matthew Wilcox
2014-03-23 19:08   ` Matthew Wilcox
2014-04-09  9:59   ` Jan Kara
2014-04-09  9:59     ` Jan Kara
2014-03-23 19:08 ` [PATCH v7 15/22] Remove CONFIG_EXT2_FS_XIP and rename CONFIG_FS_XIP to CONFIG_FS_DAX Matthew Wilcox
2014-03-23 19:08   ` Matthew Wilcox
2014-04-09  9:59   ` Jan Kara
2014-04-09  9:59     ` Jan Kara
2014-04-10 14:23     ` Matthew Wilcox
2014-04-10 14:23       ` Matthew Wilcox
2014-03-23 19:08 ` [PATCH v7 16/22] ext2: Remove ext2_aops_xip Matthew Wilcox
2014-03-23 19:08   ` Matthew Wilcox
2014-04-09 10:02   ` Jan Kara
2014-04-09 10:02     ` Jan Kara
2014-03-23 19:08 ` [PATCH v7 17/22] Get rid of most mentions of XIP in ext2 Matthew Wilcox
2014-03-23 19:08   ` Matthew Wilcox
2014-04-09 10:04   ` Jan Kara
2014-04-09 10:04     ` Jan Kara
2014-04-10 14:26     ` Matthew Wilcox
2014-04-10 14:26       ` Matthew Wilcox
2014-04-10 18:40       ` Jan Kara
2014-04-10 18:40         ` Jan Kara
2014-03-23 19:08 ` [PATCH v7 18/22] xip: Add xip_zero_page_range Matthew Wilcox
2014-03-23 19:08   ` Matthew Wilcox
2014-04-09 10:15   ` Jan Kara
2014-04-09 10:15     ` Jan Kara
2014-04-10 14:27     ` Matthew Wilcox
2014-04-10 14:27       ` Matthew Wilcox
2014-04-10 18:43       ` Jan Kara
2014-04-10 18:43         ` Jan Kara
2014-03-23 19:08 ` [PATCH v7 19/22] ext4: Make ext4_block_zero_page_range static Matthew Wilcox
2014-03-23 19:08   ` Matthew Wilcox
2014-03-24 19:11   ` tytso
2014-03-24 19:11     ` tytso
2014-03-23 19:08 ` [PATCH v7 20/22] ext4: Add DAX functionality Matthew Wilcox
2014-03-23 19:08   ` Matthew Wilcox
2014-04-09 12:17   ` Jan Kara
2014-04-09 12:17     ` Jan Kara
2014-03-23 19:08 ` [PATCH v7 21/22] ext4: Fix typos Matthew Wilcox
2014-03-23 19:08   ` Matthew Wilcox
2014-03-24 19:16   ` tytso
2014-03-24 19:16     ` tytso
2014-03-23 19:08 ` [PATCH v7 22/22] brd: Rename XIP to DAX Matthew Wilcox
2014-03-23 19:08   ` Matthew Wilcox
2014-04-09 10:07   ` Jan Kara
2014-04-09 10:07     ` Jan Kara
2014-05-18 14:58 ` [PATCH v7 00/22] Support ext4 on NV-DIMMs Boaz Harrosh
2014-05-18 14:58   ` Boaz Harrosh
2014-05-18 23:24   ` Matthew Wilcox
2014-05-18 23:24     ` Matthew Wilcox
2014-06-17 18:11 ` Boaz Harrosh
2014-06-17 18:11   ` Boaz Harrosh
2014-06-17 18:19   ` Matthew Wilcox
2014-06-17 18:19     ` Matthew Wilcox
2014-06-17 18:39     ` Boaz Harrosh
2014-06-17 18:39       ` Boaz Harrosh

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20140730210239.GS6754@linux.intel.com \
    --to=willy@linux.intel.com \
    --cc=jack@suse.cz \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=matthew.r.wilcox@intel.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.