John Garry <john.g.garry@xxxxxxxxxx> writes:
> On 01/11/2024 06:50, Ritesh Harjani (IBM) wrote:
>> This patch adds base support for atomic writes via statx getattr.
>> On bs < ps systems, we can create FS with say bs of 16k. That means
>> both atomic write min and max unit can be set to 16k for supporting
>> atomic writes.
>>
>> Co-developed-by: Ojaswin Mujoo <ojaswin@xxxxxxxxxxxxx>
>> Signed-off-by: Ojaswin Mujoo <ojaswin@xxxxxxxxxxxxx>
>> Signed-off-by: Ritesh Harjani (IBM) <ritesh.list@xxxxxxxxx>
>
> Regardless of nitpicks:
>
> Reviewed-by: John Garry <john.g.garry@xxxxxxxxxx>
>
Thanks John for the review!
Since as you too mentioned the remaining points are minor and not
critical review comments, I will address them next time in the
multi-fsblock variant. With all other aspects now finalized in this v4
version, this looks ready to be picked up for the merge window.
-ritesh
>> ---
>> fs/ext4/ext4.h | 10 ++++++++++
>> fs/ext4/inode.c | 12 ++++++++++++
>> fs/ext4/super.c | 31 +++++++++++++++++++++++++++++++
>> 3 files changed, 53 insertions(+)
>>
>> diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
>> index 44b0d418143c..494d443e9fc9 100644
>> --- a/fs/ext4/ext4.h
>> +++ b/fs/ext4/ext4.h
>> @@ -1729,6 +1729,10 @@ struct ext4_sb_info {
>> */
>> struct work_struct s_sb_upd_work;
>>
>> + /* Atomic write unit values in bytes */
>> + unsigned int s_awu_min;
>> + unsigned int s_awu_max;
>> +
>> /* Ext4 fast commit sub transaction ID */
>> atomic_t s_fc_subtid;
>>
>> @@ -3855,6 +3859,12 @@ static inline int ext4_buffer_uptodate(struct buffer_head *bh)
>> return buffer_uptodate(bh);
>> }
>>
>> +static inline bool ext4_inode_can_atomic_write(struct inode *inode)
>> +{
>> +
>
> nit: superfluous blank line
>
Sure.
>> + return S_ISREG(inode->i_mode) && EXT4_SB(inode->i_sb)->s_awu_min > 0;
>
> I am not sure if the S_ISREG() check is required. Other callers also do
> the check (like ext4_getattr() for when calling
> ext4_inode_can_atomic_write()) or don't need it (ext4_file_open()). I
> say ext4_file_open() doesn't need it as ext4_file_open() is only ever
> called for regular files, right?
>
Yes. However I believe we might end up using this from other places when
we add support of extsize. So we might need S_ISREG check.
But sure let me re-think on that during the multi-fsblock variant time.
>> +}
>> +
>> extern int ext4_block_write_begin(handle_t *handle, struct folio *folio,
>> loff_t pos, unsigned len,
>> get_block_t *get_block);
>> diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
>> index 54bdd4884fe6..3e827cfa762e 100644
>> --- a/fs/ext4/inode.c
>> +++ b/fs/ext4/inode.c
>> @@ -5578,6 +5578,18 @@ int ext4_getattr(struct mnt_idmap *idmap, const struct path *path,
>> }
>> }
>>
>> + if ((request_mask & STATX_WRITE_ATOMIC) && S_ISREG(inode->i_mode)) {
>
> nit: maybe you could have factored out the S_ISREG() check with
> STATX_DIOALIGN
>
Sure.
>> + struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
>> + unsigned int awu_min = 0, awu_max = 0;
>> +
>> + if (ext4_inode_can_atomic_write(inode)) {
>> + awu_min = sbi->s_awu_min;
>> + awu_max = sbi->s_awu_max;
>> + }
>> +
>> + generic_fill_statx_atomic_writes(stat, awu_min, awu_max);
>> + }
>> +
>> flags = ei->i_flags & EXT4_FL_USER_VISIBLE;
>> if (flags & EXT4_APPEND_FL)
>> stat->attributes |= STATX_ATTR_APPEND;
>> diff --git a/fs/ext4/super.c b/fs/ext4/super.c
>> index 16a4ce704460..ebe1660bd840 100644
>> --- a/fs/ext4/super.c
>> +++ b/fs/ext4/super.c
>> @@ -4425,6 +4425,36 @@ static int ext4_handle_clustersize(struct super_block *sb)
>> return 0;
>> }
>>
>> +/*
>> + * ext4_atomic_write_init: Initializes filesystem min & max atomic write units.
>> + * @sb: super block
>> + * TODO: Later add support for bigalloc
>> + */
>> +static void ext4_atomic_write_init(struct super_block *sb)
>> +{
>> + struct ext4_sb_info *sbi = EXT4_SB(sb);
>> + struct block_device *bdev = sb->s_bdev;
>> +
>> + if (!bdev_can_atomic_write(bdev))
>> + return;
>> +
>> + if (!ext4_has_feature_extents(sb))
>> + return;
>> +
>> + sbi->s_awu_min = max(sb->s_blocksize,
>> + bdev_atomic_write_unit_min_bytes(bdev));
>> + sbi->s_awu_max = min(sb->s_blocksize,
>> + bdev_atomic_write_unit_max_bytes(bdev));
>> + if (sbi->s_awu_min && sbi->s_awu_max &&
>> + sbi->s_awu_min <= sbi->s_awu_max) {
>> + ext4_msg(sb, KERN_NOTICE, "Supports (experimental) DIO atomic writes awu_min: %u, awu_max: %u",
>> + sbi->s_awu_min, sbi->s_awu_max);
>> + } else {
>> + sbi->s_awu_min = 0;
>> + sbi->s_awu_max = 0;
>> + }
>> +}
>> +
>> static void ext4_fast_commit_init(struct super_block *sb)
>> {
>> struct ext4_sb_info *sbi = EXT4_SB(sb);
>> @@ -5336,6 +5366,7 @@ static int __ext4_fill_super(struct fs_context *fc, struct super_block *sb)
>>
>> spin_lock_init(&sbi->s_bdev_wb_lock);
>>
>> + ext4_atomic_write_init(sb);
>> ext4_fast_commit_init(sb);
>>
>> sb->s_root = NULL;
