On Tue, Apr 19, 2016 at 11:39 PM, Mateusz Guzik <mguzik@xxxxxxxxxx> wrote:
> On Tue, Apr 19, 2016 at 10:48:41PM +0200, Florian Margaine wrote:
>> The behavior was removed in 18e9e5104fcd9a973ffe3eed3816c87f2a1b6cd2
>> noting that this was a better idea than using a counter. However, this
>> behavior is actually wanted if multiple applications want to freeze
>> concurrently while remaining non-racy.
>>
>> This patch reintroduces this feature by using a counter.
>>
>
> This patch is wrong.
>
> It uses non-atomic ops to modify the counter and no locks are
> held to protect it.
>
> I would argue the code should track that freezing has started and
> additional freezers must only return when the state is
> SB_FREEZE_COMPLETE.
Does it all additional freezers must return in order, or all
additional freezers can return asap when SB_FREEZE_COMPLETE is set?
I'm not sure the former is easily doable.
>
>> ---
>> fs/super.c | 15 +++++++++++----
>> include/linux/fs.h | 1 +
>> 2 files changed, 12 insertions(+), 4 deletions(-)
>>
>> diff --git a/fs/super.c b/fs/super.c
>> index 74914b1..9fa8ca1 100644
>> --- a/fs/super.c
>> +++ b/fs/super.c
>> @@ -231,6 +231,7 @@ static struct super_block *alloc_super(struct
>> file_system_type *type, int flags)
>> */
>> down_write_nested(&s->s_umount, SINGLE_DEPTH_NESTING);
>> s->s_count = 1;
>> + s->s_freezers = 0;
>> atomic_set(&s->s_active, 1);
>> mutex_init(&s->s_vfs_rename_mutex);
>> lockdep_set_class(&s->s_vfs_rename_mutex, &type-
>> >s_vfs_rename_key);
>> @@ -1275,12 +1276,12 @@ int freeze_super(struct super_block *sb)
>> {
>> int ret;
>>
>> + sb->s_freezers++;
>> + if (sb->s_writers.frozen != SB_UNFROZEN)
>> + return 0;
>> +
>> atomic_inc(&sb->s_active);
>> down_write(&sb->s_umount);
>> - if (sb->s_writers.frozen != SB_UNFROZEN) {
>> - deactivate_locked_super(sb);
>> - return -EBUSY;
>> - }
>>
>> if (!(sb->s_flags & MS_BORN)) {
>> up_write(&sb->s_umount);
>> @@ -1338,14 +1339,20 @@ EXPORT_SYMBOL(freeze_super);
>> * @sb: the super to thaw
>> *
>> * Unlocks the filesystem and marks it writeable again after
>> freeze_super().
>> + * Since nesting freezes is allowed, only the last freeze actually
>> unlocks.
>> */
>> int thaw_super(struct super_block *sb)
>> {
>> int error;
>>
>> + sb->s_freezers--;
>> + if (sb->s_freezers > 0)
>> + return 0;
>> +
>> down_write(&sb->s_umount);
>> if (sb->s_writers.frozen == SB_UNFROZEN) {
>> up_write(&sb->s_umount);
>> + sb->s_freezers++;
>> return -EINVAL;
>> }
>>
>> diff --git a/include/linux/fs.h b/include/linux/fs.h
>> index e514f76..c045e2a 100644
>> --- a/include/linux/fs.h
>> +++ b/include/linux/fs.h
>> @@ -1333,6 +1333,7 @@ struct super_block {
>> struct quota_info s_dquot; /* Diskquota specific
>> options */
>>
>> struct sb_writers s_writers;
>> + int s_freezers;
>>
>> char s_id[32]; /* Informational
>> name */
>> u8 s_uuid[16]; /* UUID */
>> --
>> 2.8.0
>>
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
>> the body of a message to majordomo@xxxxxxxxxxxxxxx
>> More majordomo info at http://vger.kernel.org/majordomo-info.html
>
> --
> Mateusz Guzik
--
Florian Margaine
--
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
