Skip to contentSkip to conversation

locks:Remove spinlock in unshare_files

RFC PATCH In review [email protected] 3 participants 4 messages Mar 13, 2020 – Mar 16, 2020

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

RFC PATCH 1/1 locks:Remove spinlock in unshare_files

From: Ma Ling <[email protected]>

Processor support atomic operation for long/int/short/char type, we use the feature to avoid spinlock, which cost hundreds cycles.

Appreciate your comments
Ling

---
 kernel/fork.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)
kernel/fork.c +1 -3
index 60a1295..fe54600 100644
@@ -3041,9 +3041,7 @@ int unshare_files(struct files_struct **displaced)
 		return error;
 	}
 	*displaced = task->files;
-	task_lock(task);
-	task->files = copy;
-	task_unlock(task);
+	WRITE_ONCE(task->files, copy);
 	return 0;
 }
 
Signature · 1 line
1.8.3.1
author
#1 lore raw

Any comments ?

Thanks
Ling

Quoted 32 lines · <[email protected]> 于2020年3月13日周五 上午11:09写道:

From: Ma Ling <[email protected]>

Processor support atomic operation for long/int/short/char type, we use the feature to avoid spinlock, which cost hundreds cycles.

Appreciate your comments
Ling

---
 kernel/fork.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)
kernel/fork.c +1 -4
index 60a1295..fe54600 100644
@@ -3041,9 +3041,7 @@ int unshare_files(struct files_struct **displaced)
                return error;
        }
        *displaced = task->files;
-       task_lock(task);
-       task->files = copy;
-       task_unlock(task);
+       WRITE_ONCE(task->files, copy);
        return 0;
 }
 
--

1.8.3.1

maintainer
#2 lore raw
Quoted 1 line · Ling Ma

Any comments ?

A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?
A: Top-posting.
Q: What is the most annoying thing in e-mail?

Also, it probably helps to Cc the right people.

Quoted 15 lines
Quoted 14 lines · <[email protected]> 于2020年3月13日周五 上午11:09写道:

From: Ma Ling <[email protected]>

Processor support atomic operation for long/int/short/char type, we use the feature to avoid spinlock, which cost hundreds cycles.

Appreciate your comments
Ling

---
 kernel/fork.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)
Quoted 14 lines
Quoted 5 lines
kernel/fork.c +0 -0
index 60a1295..fe54600 100644

kernel/fork.c int unshare_files(struct files_struct **displaced) · line 3046

Quoted 9 lines
kernel/fork.c +1 -3
@@ -3041,9 +3041,7 @@ int unshare_files(struct files_struct **displaced)
                return error;
        }
        *displaced = task->files;
-       task_lock(task);
-       task->files = copy;
-       task_unlock(task);
+       WRITE_ONCE(task->files, copy);
        return 0;
 }

AFAICT this is completely and utterly buggered.

IFF task->files was lockless, like say RCU, then you'd still need smp_store_release(). But if we look at fs/file.c then everything uses task_lock() and removing it like the above is actively broken.

#3 lore raw
Quoted 17 lines · Peter Zijlstra
Quoted 6 lines
Quoted 6 lines
kernel/fork.c +0 -0
index 60a1295..fe54600 100644
@@ -3041,9 +3041,7 @@ int unshare_files(struct files_struct **displaced)
                return error;
Quoted 8 lines

kernel/fork.c int unshare_files(struct files_struct **displaced) · line 3046

Quoted 8 lines
kernel/fork.c +1 -3
        }
        *displaced = task->files;
-       task_lock(task);
-       task->files = copy;
-       task_unlock(task);
+       WRITE_ONCE(task->files, copy);
        return 0;
 }

AFAICT this is completely and utterly buggered.

Quoted 3 lines

IFF task->files was lockless, like say RCU, then you'd still need smp_store_release(). But if we look at fs/file.c then everything uses task_lock() and removing it like the above is actively broken.

The problem is not fs/file.c; it's the code that does (read-only) access to *other* threads' ->files. procfs, SAK, some cgroup shite (pardon the redundancy)... All of those rely upon task_lock.

FWIW, having just grepped around, I'm worried about the crap io_uring is pulling off - interplay with unshare(2) could be unpleasant.

In any case - task_lock in the code that assigns to ->files (and it's not just unshare_files()) serves to protect the 3rd-party readers (including get_files_struct()) from having the fucker taken apart under them. It's not just freeing the thing - it's the entire close_files().

And no, we do *NOT* want to convert everything to get_files_struct() + being clever in it. I would rather have get_files_struct() taken out and shot, TBH - the only real reason it hadn't been killed years ago is the loop in proc_readfd_common()...

I'd prefer to have 3rd-party readers indicate their interest in a way that would be distinguishable from normal references, with close_files() waiting until all of those are gone. One way to do that would be
* secondary counter in files_struct
* rcu-delayed freeing of actual structure (not a problem)
* rcu_read_lock in 3rd-party readers (among other things
it means that proc_readfd_common() would need to be rearchitected a bit)
* close_files() starting with subtraction of large constant
from the secondary counter and then spinning until it gets to -<large constant>
* 3rd-party readers (under rcu_read_lock()) fetching task->files,
bumping the secondary counter unless it's negative, doing their thing, then decrementing the counter.