Re: [PATCH 01/12] fs/locks: rename some lists and pointers.

From: NeilBrown
Date: Thu Nov 08 2018 - 19:32:41 EST

Next message: NeilBrown: "Re: [PATCH 10/12] fs/locks: create a tree of dependent requests."
Previous message: Matthew Wilcox: "Re: [PATCH v5 04/27] x86/fpu/xstate: Add XSAVES system states for shadow stack"
In reply to: J. Bruce Fields: "Re: [PATCH 01/12] fs/locks: rename some lists and pointers."
Next in thread: J. Bruce Fields: "Re: [PATCH 01/12] fs/locks: rename some lists and pointers."
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

On Thu, Nov 08 2018, J. Bruce Fields wrote:

> On Mon, Nov 05, 2018 at 12:30:47PM +1100, NeilBrown wrote:
>> struct file lock contains an 'fl_next' pointer which
>> is used to point to the lock that this request is blocked
>> waiting for. So rename it to fl_blocker.
>>
>> The fl_blocked list_head in an active lock is the head of a list of
>> blocked requests. In a request it is a node in that list.
>> These are two distinct uses, so replace with two list_heads
>> with different names.
>> fl_blocked is the head of a list of blocked requests
>> fl_block is a node on that list.
>
> Reading these, I have a lot of trouble keeping fl_blocked, fl_block, and
> fl_blocker straight. Is it just me?

"Naming is hard" - but that is no excuse.
I suspect it isn't just you.

I particularly like "fl_blocker".

error = wait_event_interruptible(fl->fl_wait, !fl->fl_blocker);

reads well to me - wait until this lock has a no blocker - i.e. until
nothing blocks it.

fl_blocked could be fl_blockees (the things that I block), but I doubt
that is an improvement.

>
> I guess they form a tree, so fl_children, fl_siblings, and fl_parent
> might be easier for me to keep straight.

This requires one to know a priori that the tree records which locks
block which requests, which is obvious to us now, but might not be so
obvious in 5 years time when we look at this code again.

An I have never really liked the "siblings" naming. 'struct dentry' uses
"d_child", which is possibly ever more confusing.
I would like it to be obvious that this is a list-member, not a
list-head. Rusty once posted patches to allow the list head to be a
different type to the members, but that fell on deaf ears.
So
fl_blocked_member
might be an improvement - this is a member of the fl_blocked list.
It would be easier to search for than fl_block - which needs
fl_block[^a-z] to avoid false positives.

I'd be quite happy to change fl_block is any two people can agree on a
better name. I'm less inclined to change the others without a really
good proposal.

Hmmm. what is the inverse of "Block"? If I block you then you .... I
know, you are a usurper.
So
fl_blocker points to the "parent"
fl_usurpers is a list of "children"
fl_usurpers_member is my linkage in that list.
or not.

Thanks,
NeilBrown

>
> --b.
>
>>
>> The two different list_heads are never used at the same time, but that
>> will change in a future patch.
>>
>> Note that a tracepoint is changed to report fl_blocker instead
>> of fl_next.
>>
>> Signed-off-by: NeilBrown <neilb@xxxxxxxx>
>> ---
>> fs/cifs/file.c | 2 +-
>> fs/locks.c | 38 ++++++++++++++++++++------------------
>> include/linux/fs.h | 7 +++++--
>> include/trace/events/filelock.h | 16 ++++++++--------
>> 4 files changed, 34 insertions(+), 29 deletions(-)
>>
>> diff --git a/fs/cifs/file.c b/fs/cifs/file.c
>> index 74c33d5fafc8..d7ed895e05d1 100644
>> --- a/fs/cifs/file.c
>> +++ b/fs/cifs/file.c
>> @@ -1103,7 +1103,7 @@ cifs_posix_lock_set(struct file *file, struct file_lock *flock)
>> rc = posix_lock_file(file, flock, NULL);
>> up_write(&cinode->lock_sem);
>> if (rc == FILE_LOCK_DEFERRED) {
>> - rc = wait_event_interruptible(flock->fl_wait, !flock->fl_next);
>> + rc = wait_event_interruptible(flock->fl_wait, !flock->fl_blocker);
>> if (!rc)
>> goto try_again;
>> posix_unblock_lock(flock);
>> diff --git a/fs/locks.c b/fs/locks.c
>> index 2ecb4db8c840..a6c6d601286c 100644
>> --- a/fs/locks.c
>> +++ b/fs/locks.c
>> @@ -189,7 +189,7 @@ static DEFINE_HASHTABLE(blocked_hash, BLOCKED_HASH_BITS);
>> * This lock protects the blocked_hash. Generally, if you're accessing it, you
>> * want to be holding this lock.
>> *
>> - * In addition, it also protects the fl->fl_block list, and the fl->fl_next
>> + * In addition, it also protects the fl->fl_blocked list, and the fl->fl_blocker
>> * pointer for file_lock structures that are acting as lock requests (in
>> * contrast to those that are acting as records of acquired locks).
>> *
>> @@ -293,6 +293,7 @@ static void locks_init_lock_heads(struct file_lock *fl)
>> {
>> INIT_HLIST_NODE(&fl->fl_link);
>> INIT_LIST_HEAD(&fl->fl_list);
>> + INIT_LIST_HEAD(&fl->fl_blocked);
>> INIT_LIST_HEAD(&fl->fl_block);
>> init_waitqueue_head(&fl->fl_wait);
>> }
>> @@ -332,6 +333,7 @@ void locks_free_lock(struct file_lock *fl)
>> {
>> BUG_ON(waitqueue_active(&fl->fl_wait));
>> BUG_ON(!list_empty(&fl->fl_list));
>> + BUG_ON(!list_empty(&fl->fl_blocked));
>> BUG_ON(!list_empty(&fl->fl_block));
>> BUG_ON(!hlist_unhashed(&fl->fl_link));
>>
>> @@ -667,7 +669,7 @@ static void __locks_delete_block(struct file_lock *waiter)
>> {
>> locks_delete_global_blocked(waiter);
>> list_del_init(&waiter->fl_block);
>> - waiter->fl_next = NULL;
>> + waiter->fl_blocker = NULL;
>> }
>>
>> static void locks_delete_block(struct file_lock *waiter)
>> @@ -683,16 +685,16 @@ static void locks_delete_block(struct file_lock *waiter)
>> * it seems like the reasonable thing to do.
>> *
>> * Must be called with both the flc_lock and blocked_lock_lock held. The
>> - * fl_block list itself is protected by the blocked_lock_lock, but by ensuring
>> + * fl_blocked list itself is protected by the blocked_lock_lock, but by ensuring
>> * that the flc_lock is also held on insertions we can avoid taking the
>> - * blocked_lock_lock in some cases when we see that the fl_block list is empty.
>> + * blocked_lock_lock in some cases when we see that the fl_blocked list is empty.
>> */
>> static void __locks_insert_block(struct file_lock *blocker,
>> struct file_lock *waiter)
>> {
>> BUG_ON(!list_empty(&waiter->fl_block));
>> - waiter->fl_next = blocker;
>> - list_add_tail(&waiter->fl_block, &blocker->fl_block);
>> + waiter->fl_blocker = blocker;
>> + list_add_tail(&waiter->fl_block, &blocker->fl_blocked);
>> if (IS_POSIX(blocker) && !IS_OFDLCK(blocker))
>> locks_insert_global_blocked(waiter);
>> }
>> @@ -716,18 +718,18 @@ static void locks_wake_up_blocks(struct file_lock *blocker)
>> /*
>> * Avoid taking global lock if list is empty. This is safe since new
>> * blocked requests are only added to the list under the flc_lock, and
>> - * the flc_lock is always held here. Note that removal from the fl_block
>> + * the flc_lock is always held here. Note that removal from the fl_blocked
>> * list does not require the flc_lock, so we must recheck list_empty()
>> * after acquiring the blocked_lock_lock.
>> */
>> - if (list_empty(&blocker->fl_block))
>> + if (list_empty(&blocker->fl_blocked))
>> return;
>>
>> spin_lock(&blocked_lock_lock);
>> - while (!list_empty(&blocker->fl_block)) {
>> + while (!list_empty(&blocker->fl_blocked)) {
>> struct file_lock *waiter;
>>
>> - waiter = list_first_entry(&blocker->fl_block,
>> + waiter = list_first_entry(&blocker->fl_blocked,
>> struct file_lock, fl_block);
>> __locks_delete_block(waiter);
>> if (waiter->fl_lmops && waiter->fl_lmops->lm_notify)
>> @@ -878,7 +880,7 @@ static struct file_lock *what_owner_is_waiting_for(struct file_lock *block_fl)
>>
>> hash_for_each_possible(blocked_hash, fl, fl_link, posix_owner_key(block_fl)) {
>> if (posix_same_owner(fl, block_fl))
>> - return fl->fl_next;
>> + return fl->fl_blocker;
>> }
>> return NULL;
>> }
>> @@ -1237,7 +1239,7 @@ static int posix_lock_inode_wait(struct inode *inode, struct file_lock *fl)
>> error = posix_lock_inode(inode, fl, NULL);
>> if (error != FILE_LOCK_DEFERRED)
>> break;
>> - error = wait_event_interruptible(fl->fl_wait, !fl->fl_next);
>> + error = wait_event_interruptible(fl->fl_wait, !fl->fl_blocker);
>> if (!error)
>> continue;
>>
>> @@ -1324,7 +1326,7 @@ int locks_mandatory_area(struct inode *inode, struct file *filp, loff_t start,
>> error = posix_lock_inode(inode, &fl, NULL);
>> if (error != FILE_LOCK_DEFERRED)
>> break;
>> - error = wait_event_interruptible(fl.fl_wait, !fl.fl_next);
>> + error = wait_event_interruptible(fl.fl_wait, !fl.fl_blocker);
>> if (!error) {
>> /*
>> * If we've been sleeping someone might have
>> @@ -1518,7 +1520,7 @@ int __break_lease(struct inode *inode, unsigned int mode, unsigned int type)
>>
>> locks_dispose_list(&dispose);
>> error = wait_event_interruptible_timeout(new_fl->fl_wait,
>> - !new_fl->fl_next, break_time);
>> + !new_fl->fl_blocker, break_time);
>>
>> percpu_down_read_preempt_disable(&file_rwsem);
>> spin_lock(&ctx->flc_lock);
>> @@ -1931,7 +1933,7 @@ static int flock_lock_inode_wait(struct inode *inode, struct file_lock *fl)
>> error = flock_lock_inode(inode, fl);
>> if (error != FILE_LOCK_DEFERRED)
>> break;
>> - error = wait_event_interruptible(fl->fl_wait, !fl->fl_next);
>> + error = wait_event_interruptible(fl->fl_wait, !fl->fl_blocker);
>> if (!error)
>> continue;
>>
>> @@ -2210,7 +2212,7 @@ static int do_lock_file_wait(struct file *filp, unsigned int cmd,
>> error = vfs_lock_file(filp, cmd, fl, NULL);
>> if (error != FILE_LOCK_DEFERRED)
>> break;
>> - error = wait_event_interruptible(fl->fl_wait, !fl->fl_next);
>> + error = wait_event_interruptible(fl->fl_wait, !fl->fl_blocker);
>> if (!error)
>> continue;
>>
>> @@ -2581,7 +2583,7 @@ posix_unblock_lock(struct file_lock *waiter)
>> int status = 0;
>>
>> spin_lock(&blocked_lock_lock);
>> - if (waiter->fl_next)
>> + if (waiter->fl_blocker)
>> __locks_delete_block(waiter);
>> else
>> status = -ENOENT;
>> @@ -2707,7 +2709,7 @@ static int locks_show(struct seq_file *f, void *v)
>>
>> lock_get_status(f, fl, iter->li_pos, "");
>>
>> - list_for_each_entry(bfl, &fl->fl_block, fl_block)
>> + list_for_each_entry(bfl, &fl->fl_blocked, fl_block)
>> lock_get_status(f, bfl, iter->li_pos, " ->");
>>
>> return 0;
>> diff --git a/include/linux/fs.h b/include/linux/fs.h
>> index c95c0807471f..a161bcdca8a2 100644
>> --- a/include/linux/fs.h
>> +++ b/include/linux/fs.h
>> @@ -1044,10 +1044,13 @@ bool opens_in_grace(struct net *);
>> * Obviously, the last two criteria only matter for POSIX locks.
>> */
>> struct file_lock {
>> - struct file_lock *fl_next; /* singly linked list for this inode */
>> + struct file_lock *fl_blocker; /* The lock, that is blocking us */
>> struct list_head fl_list; /* link into file_lock_context */
>> struct hlist_node fl_link; /* node in global lists */
>> - struct list_head fl_block; /* circular list of blocked processes */
>> + struct list_head fl_blocked; /* list of requests with
>> + * ->fl_blocker pointing here */
>> + struct list_head fl_block; /* node in
>> + * ->fl_blocker->fl_blocked */
>> fl_owner_t fl_owner;
>> unsigned int fl_flags;
>> unsigned char fl_type;
>> diff --git a/include/trace/events/filelock.h b/include/trace/events/filelock.h
>> index 68b17c116907..fad7befa612d 100644
>> --- a/include/trace/events/filelock.h
>> +++ b/include/trace/events/filelock.h
>> @@ -68,7 +68,7 @@ DECLARE_EVENT_CLASS(filelock_lock,
>> __field(struct file_lock *, fl)
>> __field(unsigned long, i_ino)
>> __field(dev_t, s_dev)
>> - __field(struct file_lock *, fl_next)
>> + __field(struct file_lock *, fl_blocker)
>> __field(fl_owner_t, fl_owner)
>> __field(unsigned int, fl_pid)
>> __field(unsigned int, fl_flags)
>> @@ -82,7 +82,7 @@ DECLARE_EVENT_CLASS(filelock_lock,
>> __entry->fl = fl ? fl : NULL;
>> __entry->s_dev = inode->i_sb->s_dev;
>> __entry->i_ino = inode->i_ino;
>> - __entry->fl_next = fl ? fl->fl_next : NULL;
>> + __entry->fl_blocker = fl ? fl->fl_blocker : NULL;
>> __entry->fl_owner = fl ? fl->fl_owner : NULL;
>> __entry->fl_pid = fl ? fl->fl_pid : 0;
>> __entry->fl_flags = fl ? fl->fl_flags : 0;
>> @@ -92,9 +92,9 @@ DECLARE_EVENT_CLASS(filelock_lock,
>> __entry->ret = ret;
>> ),
>>
>> - TP_printk("fl=0x%p dev=0x%x:0x%x ino=0x%lx fl_next=0x%p fl_owner=0x%p fl_pid=%u fl_flags=%s fl_type=%s fl_start=%lld fl_end=%lld ret=%d",
>> + TP_printk("fl=0x%p dev=0x%x:0x%x ino=0x%lx fl_blocker=0x%p fl_owner=0x%p fl_pid=%u fl_flags=%s fl_type=%s fl_start=%lld fl_end=%lld ret=%d",
>> __entry->fl, MAJOR(__entry->s_dev), MINOR(__entry->s_dev),
>> - __entry->i_ino, __entry->fl_next, __entry->fl_owner,
>> + __entry->i_ino, __entry->fl_blocker, __entry->fl_owner,
>> __entry->fl_pid, show_fl_flags(__entry->fl_flags),
>> show_fl_type(__entry->fl_type),
>> __entry->fl_start, __entry->fl_end, __entry->ret)
>> @@ -125,7 +125,7 @@ DECLARE_EVENT_CLASS(filelock_lease,
>> __field(struct file_lock *, fl)
>> __field(unsigned long, i_ino)
>> __field(dev_t, s_dev)
>> - __field(struct file_lock *, fl_next)
>> + __field(struct file_lock *, fl_blocker)
>> __field(fl_owner_t, fl_owner)
>> __field(unsigned int, fl_flags)
>> __field(unsigned char, fl_type)
>> @@ -137,7 +137,7 @@ DECLARE_EVENT_CLASS(filelock_lease,
>> __entry->fl = fl ? fl : NULL;
>> __entry->s_dev = inode->i_sb->s_dev;
>> __entry->i_ino = inode->i_ino;
>> - __entry->fl_next = fl ? fl->fl_next : NULL;
>> + __entry->fl_blocker = fl ? fl->fl_blocker : NULL;
>> __entry->fl_owner = fl ? fl->fl_owner : NULL;
>> __entry->fl_flags = fl ? fl->fl_flags : 0;
>> __entry->fl_type = fl ? fl->fl_type : 0;
>> @@ -145,9 +145,9 @@ DECLARE_EVENT_CLASS(filelock_lease,
>> __entry->fl_downgrade_time = fl ? fl->fl_downgrade_time : 0;
>> ),
>>
>> - TP_printk("fl=0x%p dev=0x%x:0x%x ino=0x%lx fl_next=0x%p fl_owner=0x%p fl_flags=%s fl_type=%s fl_break_time=%lu fl_downgrade_time=%lu",
>> + TP_printk("fl=0x%p dev=0x%x:0x%x ino=0x%lx fl_blocker=0x%p fl_owner=0x%p fl_flags=%s fl_type=%s fl_break_time=%lu fl_downgrade_time=%lu",
>> __entry->fl, MAJOR(__entry->s_dev), MINOR(__entry->s_dev),
>> - __entry->i_ino, __entry->fl_next, __entry->fl_owner,
>> + __entry->i_ino, __entry->fl_blocker, __entry->fl_owner,
>> show_fl_flags(__entry->fl_flags),
>> show_fl_type(__entry->fl_type),
>> __entry->fl_break_time, __entry->fl_downgrade_time)
>>

Attachment: signature.asc
Description: PGP signature

Next message: NeilBrown: "Re: [PATCH 10/12] fs/locks: create a tree of dependent requests."
Previous message: Matthew Wilcox: "Re: [PATCH v5 04/27] x86/fpu/xstate: Add XSAVES system states for shadow stack"
In reply to: J. Bruce Fields: "Re: [PATCH 01/12] fs/locks: rename some lists and pointers."
Next in thread: J. Bruce Fields: "Re: [PATCH 01/12] fs/locks: rename some lists and pointers."
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]