| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Date: Tue, 20 Feb 2018 14:07:44 +0100
From: Jan Kara <jack@...e.cz>
To: Nikolay Borisov <nborisov@...e.com>
Cc: linux-ext4@...r.kernel.org, tytso@....edu, jack@...e.cz
Subject: Re: [PATCH] ext4: Remove EXT4_STATE_DIOREAD_LOCK flag
On Tue 20-02-18 12:20:05, Nikolay Borisov wrote:
> Commit 16c54688592c ("ext4: Allow parallel DIO reads") reworked the way
> locking happens around parallel dio reads. This resulted in obviating
> the need for EXT4_STATE_DIOREAD_LOCK flag and accompanying logic.
> Currently this amounts to dead code so let's remove it. No functional
> changes
>
> Signed-off-by: Nikolay Borisov <nborisov@...e.com>
Very good point. The patch looks good. You can add:
Reviewed-by: Jan Kara <jack@...e.cz>
Honza
> ---
> fs/ext4/ext4.h | 17 -----------------
> fs/ext4/extents.c | 19 +++++--------------
> fs/ext4/inode.c | 8 --------
> fs/ext4/ioctl.c | 4 ----
> fs/ext4/move_extent.c | 4 ----
> fs/ext4/super.c | 12 +++++-------
> 6 files changed, 10 insertions(+), 54 deletions(-)
>
> diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
> index 3241475a1733..a42e71203e53 100644
> --- a/fs/ext4/ext4.h
> +++ b/fs/ext4/ext4.h
> @@ -1522,8 +1522,6 @@ enum {
> EXT4_STATE_EXT_MIGRATE, /* Inode is migrating */
> EXT4_STATE_DIO_UNWRITTEN, /* need convert on dio done*/
> EXT4_STATE_NEWENTRY, /* File just added to dir */
> - EXT4_STATE_DIOREAD_LOCK, /* Disable support for dio read
> - nolocking */
> EXT4_STATE_MAY_INLINE_DATA, /* may have in-inode data */
> EXT4_STATE_EXT_PRECACHED, /* extents have been precached */
> EXT4_STATE_LUSTRE_EA_INODE, /* Lustre-style ea_inode */
> @@ -3181,21 +3179,6 @@ static inline void set_bitmap_uptodate(struct buffer_head *bh)
> set_bit(BH_BITMAP_UPTODATE, &(bh)->b_state);
> }
>
> -/*
> - * Disable DIO read nolock optimization, so new dioreaders will be forced
> - * to grab i_mutex
> - */
> -static inline void ext4_inode_block_unlocked_dio(struct inode *inode)
> -{
> - ext4_set_inode_state(inode, EXT4_STATE_DIOREAD_LOCK);
> - smp_mb();
> -}
> -static inline void ext4_inode_resume_unlocked_dio(struct inode *inode)
> -{
> - smp_mb();
> - ext4_clear_inode_state(inode, EXT4_STATE_DIOREAD_LOCK);
> -}
> -
> #define in_range(b, first, len) ((b) >= (first) && (b) <= (first) + (len) - 1)
>
> /* For ioend & aio unwritten conversion wait queues */
> diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c
> index 054416e9d827..837f0a6357c7 100644
> --- a/fs/ext4/extents.c
> +++ b/fs/ext4/extents.c
> @@ -4796,7 +4796,6 @@ static long ext4_zero_range(struct file *file, loff_t offset,
> flags |= EXT4_GET_BLOCKS_KEEP_SIZE;
>
> /* Wait all existing dio workers, newcomers will block on i_mutex */
> - ext4_inode_block_unlocked_dio(inode);
> inode_dio_wait(inode);
>
> /* Preallocate the range including the unaligned edges */
> @@ -4807,7 +4806,7 @@ static long ext4_zero_range(struct file *file, loff_t offset,
> round_down(offset, 1 << blkbits)) >> blkbits,
> new_size, flags);
> if (ret)
> - goto out_dio;
> + goto out_mutex;
>
> }
>
> @@ -4824,7 +4823,7 @@ static long ext4_zero_range(struct file *file, loff_t offset,
> ret = ext4_update_disksize_before_punch(inode, offset, len);
> if (ret) {
> up_write(&EXT4_I(inode)->i_mmap_sem);
> - goto out_dio;
> + goto out_mutex;
> }
> /* Now release the pages and zero block aligned part of pages */
> truncate_pagecache_range(inode, start, end - 1);
> @@ -4834,10 +4833,10 @@ static long ext4_zero_range(struct file *file, loff_t offset,
> flags);
> up_write(&EXT4_I(inode)->i_mmap_sem);
> if (ret)
> - goto out_dio;
> + goto out_mutex;
> }
> if (!partial_begin && !partial_end)
> - goto out_dio;
> + goto out_mutex;
>
> /*
> * In worst case we have to writeout two nonadjacent unwritten
> @@ -4850,7 +4849,7 @@ static long ext4_zero_range(struct file *file, loff_t offset,
> if (IS_ERR(handle)) {
> ret = PTR_ERR(handle);
> ext4_std_error(inode->i_sb, ret);
> - goto out_dio;
> + goto out_mutex;
> }
>
> inode->i_mtime = inode->i_ctime = current_time(inode);
> @@ -4875,8 +4874,6 @@ static long ext4_zero_range(struct file *file, loff_t offset,
> ext4_handle_sync(handle);
>
> ext4_journal_stop(handle);
> -out_dio:
> - ext4_inode_resume_unlocked_dio(inode);
> out_mutex:
> inode_unlock(inode);
> return ret;
> @@ -4964,11 +4961,9 @@ long ext4_fallocate(struct file *file, int mode, loff_t offset, loff_t len)
> }
>
> /* Wait all existing dio workers, newcomers will block on i_mutex */
> - ext4_inode_block_unlocked_dio(inode);
> inode_dio_wait(inode);
>
> ret = ext4_alloc_file_blocks(file, lblk, max_blocks, new_size, flags);
> - ext4_inode_resume_unlocked_dio(inode);
> if (ret)
> goto out;
>
> @@ -5485,7 +5480,6 @@ int ext4_collapse_range(struct inode *inode, loff_t offset, loff_t len)
> }
>
> /* Wait for existing dio to complete */
> - ext4_inode_block_unlocked_dio(inode);
> inode_dio_wait(inode);
>
> /*
> @@ -5562,7 +5556,6 @@ int ext4_collapse_range(struct inode *inode, loff_t offset, loff_t len)
> ext4_journal_stop(handle);
> out_mmap:
> up_write(&EXT4_I(inode)->i_mmap_sem);
> - ext4_inode_resume_unlocked_dio(inode);
> out_mutex:
> inode_unlock(inode);
> return ret;
> @@ -5635,7 +5628,6 @@ int ext4_insert_range(struct inode *inode, loff_t offset, loff_t len)
> }
>
> /* Wait for existing dio to complete */
> - ext4_inode_block_unlocked_dio(inode);
> inode_dio_wait(inode);
>
> /*
> @@ -5737,7 +5729,6 @@ int ext4_insert_range(struct inode *inode, loff_t offset, loff_t len)
> ext4_journal_stop(handle);
> out_mmap:
> up_write(&EXT4_I(inode)->i_mmap_sem);
> - ext4_inode_resume_unlocked_dio(inode);
> out_mutex:
> inode_unlock(inode);
> return ret;
> diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
> index c94780075b04..7d285e79fd9f 100644
> --- a/fs/ext4/inode.c
> +++ b/fs/ext4/inode.c
> @@ -4251,7 +4251,6 @@ int ext4_punch_hole(struct inode *inode, loff_t offset, loff_t length)
> }
>
> /* Wait all existing dio workers, newcomers will block on i_mutex */
> - ext4_inode_block_unlocked_dio(inode);
> inode_dio_wait(inode);
>
> /*
> @@ -4324,7 +4323,6 @@ int ext4_punch_hole(struct inode *inode, loff_t offset, loff_t length)
> ext4_journal_stop(handle);
> out_dio:
> up_write(&EXT4_I(inode)->i_mmap_sem);
> - ext4_inode_resume_unlocked_dio(inode);
> out_mutex:
> inode_unlock(inode);
> return ret;
> @@ -5506,9 +5504,7 @@ int ext4_setattr(struct dentry *dentry, struct iattr *attr)
> */
> if (orphan) {
> if (!ext4_should_journal_data(inode)) {
> - ext4_inode_block_unlocked_dio(inode);
> inode_dio_wait(inode);
> - ext4_inode_resume_unlocked_dio(inode);
> } else
> ext4_wait_for_tail_page_commit(inode);
> }
> @@ -5999,7 +5995,6 @@ int ext4_change_inode_journal_flag(struct inode *inode, int val)
> return -EROFS;
>
> /* Wait for all existing dio workers */
> - ext4_inode_block_unlocked_dio(inode);
> inode_dio_wait(inode);
>
> /*
> @@ -6015,7 +6010,6 @@ int ext4_change_inode_journal_flag(struct inode *inode, int val)
> err = filemap_write_and_wait(inode->i_mapping);
> if (err < 0) {
> up_write(&EXT4_I(inode)->i_mmap_sem);
> - ext4_inode_resume_unlocked_dio(inode);
> return err;
> }
> }
> @@ -6038,7 +6032,6 @@ int ext4_change_inode_journal_flag(struct inode *inode, int val)
> if (err < 0) {
> jbd2_journal_unlock_updates(journal);
> percpu_up_write(&sbi->s_journal_flag_rwsem);
> - ext4_inode_resume_unlocked_dio(inode);
> return err;
> }
> ext4_clear_inode_flag(inode, EXT4_INODE_JOURNAL_DATA);
> @@ -6050,7 +6043,6 @@ int ext4_change_inode_journal_flag(struct inode *inode, int val)
>
> if (val)
> up_write(&EXT4_I(inode)->i_mmap_sem);
> - ext4_inode_resume_unlocked_dio(inode);
>
> /* Finally we can mark the inode as dirty. */
>
> diff --git a/fs/ext4/ioctl.c b/fs/ext4/ioctl.c
> index 7e99ad02f1ba..3644e31ba347 100644
> --- a/fs/ext4/ioctl.c
> +++ b/fs/ext4/ioctl.c
> @@ -124,8 +124,6 @@ static long swap_inode_boot_loader(struct super_block *sb,
> truncate_inode_pages(&inode_bl->i_data, 0);
>
> /* Wait for all existing dio workers */
> - ext4_inode_block_unlocked_dio(inode);
> - ext4_inode_block_unlocked_dio(inode_bl);
> inode_dio_wait(inode);
> inode_dio_wait(inode_bl);
>
> @@ -186,8 +184,6 @@ static long swap_inode_boot_loader(struct super_block *sb,
> ext4_double_up_write_data_sem(inode, inode_bl);
>
> journal_err_out:
> - ext4_inode_resume_unlocked_dio(inode);
> - ext4_inode_resume_unlocked_dio(inode_bl);
> unlock_two_nondirectories(inode, inode_bl);
> iput(inode_bl);
> return err;
> diff --git a/fs/ext4/move_extent.c b/fs/ext4/move_extent.c
> index b96e4bd3b3ec..8e17efdcbf11 100644
> --- a/fs/ext4/move_extent.c
> +++ b/fs/ext4/move_extent.c
> @@ -601,8 +601,6 @@ ext4_move_extents(struct file *o_filp, struct file *d_filp, __u64 orig_blk,
> lock_two_nondirectories(orig_inode, donor_inode);
>
> /* Wait for all existing dio workers */
> - ext4_inode_block_unlocked_dio(orig_inode);
> - ext4_inode_block_unlocked_dio(donor_inode);
> inode_dio_wait(orig_inode);
> inode_dio_wait(donor_inode);
>
> @@ -693,8 +691,6 @@ ext4_move_extents(struct file *o_filp, struct file *d_filp, __u64 orig_blk,
> ext4_ext_drop_refs(path);
> kfree(path);
> ext4_double_up_write_data_sem(orig_inode, donor_inode);
> - ext4_inode_resume_unlocked_dio(orig_inode);
> - ext4_inode_resume_unlocked_dio(donor_inode);
> unlock_two_nondirectories(orig_inode, donor_inode);
>
> return ret;
> diff --git a/fs/ext4/super.c b/fs/ext4/super.c
> index 39bf464c35f1..8781f8afbe50 100644
> --- a/fs/ext4/super.c
> +++ b/fs/ext4/super.c
> @@ -101,15 +101,13 @@ static struct inode *ext4_get_journal_inode(struct super_block *sb,
> * i_data_sem (rw)
> *
> * truncate:
> - * sb_start_write -> i_mutex -> EXT4_STATE_DIOREAD_LOCK (w) -> i_mmap_sem (w) ->
> - * i_mmap_rwsem (w) -> page lock
> - * sb_start_write -> i_mutex -> EXT4_STATE_DIOREAD_LOCK (w) -> i_mmap_sem (w) ->
> - * transaction start -> i_data_sem (rw)
> + * sb_start_write -> i_mutex -> i_mmap_sem (w) -> i_mmap_rwsem (w) -> page lock
> + * sb_start_write -> i_mutex -> i_mmap_sem (w) -> transaction start ->
> + * i_data_sem (rw)
> *
> * direct IO:
> - * sb_start_write -> i_mutex -> EXT4_STATE_DIOREAD_LOCK (r) -> mmap_sem
> - * sb_start_write -> i_mutex -> EXT4_STATE_DIOREAD_LOCK (r) ->
> - * transaction start -> i_data_sem (rw)
> + * sb_start_write -> i_mutex -> mmap_sem
> + * sb_start_write -> i_mutex -> transaction start -> i_data_sem (rw)
> *
> * writepages:
> * transaction start -> page lock(s) -> i_data_sem (rw)
> --
> 2.7.4
>
--
Jan Kara <jack@...e.com>
SUSE Labs, CR
Powered by blists - more mailing lists