[lttng-dev] Do updaters of a cds_lfht need to be synchronized?

[lttng-dev] Do updaters of a cds_lfht need to be synchronized?

[Charles Jacobsen]

The documentation for the lock-free hash table does not mention anything about synchronizing updaters (e.g., with a lock):


https://lwn.net/Articles/573431/

<https://lwn.net/Articles/573431/>https://lwn.net/Articles/573432/


The source for rculfhash (== cds_lfht?) also mentions in the comments that add and remove are lock free:


https://github.com/urcu/userspace-rcu/blob/master/src/rculfhash.c


Is it safe to assume that cds_lfht_add, cds_lfht_add_unique, and so on, *only* require a surrounding rcu read lock/unlock?


"lock free" is part of the name, so I would assume the answer is "yes".


Thank you.

Disclaimer

The information contained in this communication from the sender is confidential. It is intended solely for use by the recipient and others authorized to receive it. If you are not the recipient, you are hereby notified that any disclosure, copying, distribution or taking action in relation of the contents of this information is strictly prohibited and may be unlawful.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.lttng.org/pipermail/lttng-dev/attachments/20161028/8260b328/attachment.html>

[lttng-dev] Do updaters of a cds_lfht need to be synchronized?

[Mathieu Desnoyers]

----- On Oct 28, 2016, at 11:40 AM, Charles Jacobsen <charles.jacobsen at primarydata.com> wrote: 

> The documentation for the lock-free hash table does not mention anything about
> synchronizing updaters (e.g., with a lock):

> [ https://lwn.net/Articles/573431/ | https://lwn.net/Articles/573431/ ]

> [ https://lwn.net/Articles/573432/ | https://lwn.net/Articles/573432/ ]

> The source for rculfhash (== cds_lfht?) also mentions in the comments that add
> and remove are lock free:

> [ https://github.com/urcu/userspace-rcu/blob/master/src/rculfhash.c |
> https://github.com/urcu/userspace-rcu/blob/master/src/rculfhash.c ]

> Is it safe to assume that cds_lfht_add, cds_lfht_add_unique, and so on, *only*
> require a surrounding rcu read lock/unlock?

> "lock free" is part of the name, so I would assume the answer is "yes".

Yes, this assumption is correct. 

Thanks, 

Mathieu 

> Thank you.

> Disclaimer

> The information contained in this communication from the sender is confidential.
> It is intended solely for use by the recipient and others authorized to receive
> it. If you are not the recipient, you are hereby notified that any disclosure,
> copying, distribution or taking action in relation of the contents of this
> information is strictly prohibited and may be unlawful.

> _______________________________________________
> lttng-dev mailing list
> lttng-dev at lists.lttng.org
> https://lists.lttng.org/cgi-bin/mailman/listinfo/lttng-dev

-- 
Mathieu Desnoyers 
EfficiOS Inc. 
http://www.efficios.com 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.lttng.org/pipermail/lttng-dev/attachments/20161031/8b0a2774/attachment.html>

[lttng-dev] Do updaters of a cds_lfht need to be synchronized?

[Charles Jacobsen]

fyi, see below.


I hope I posed the question correctly.


________________________________
From: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Sent: Monday, October 31, 2016 9:50 AM
To: Charles Jacobsen
Cc: lttng-dev
Subject: Re: [lttng-dev] Do updaters of a cds_lfht need to be synchronized?


----- On Oct 28, 2016, at 11:40 AM, Charles Jacobsen <charles.jacobsen at primarydata.com> wrote:

The documentation for the lock-free hash table does not mention anything about synchronizing updaters (e.g., with a lock):


https://lwn.net/Articles/573431/

https://lwn.net/Articles/573432/


The source for rculfhash (== cds_lfht?) also mentions in the comments that add and remove are lock free:


https://github.com/urcu/userspace-rcu/blob/master/src/rculfhash.c


Is it safe to assume that cds_lfht_add, cds_lfht_add_unique, and so on, *only* require a surrounding rcu read lock/unlock?


"lock free" is part of the name, so I would assume the answer is "yes".

Yes, this assumption is correct.

Thanks,

Mathieu





Thank you.



Disclaimer

The information contained in this communication from the sender is confidential. It is intended solely for use by the recipient and others authorized to receive it. If you are not the recipient, you are hereby notified that any disclosure, copying, distribution or taking action in relation of the contents of this information is strictly prohibited and may be unlawful.

_______________________________________________
lttng-dev mailing list
lttng-dev at lists.lttng.org
https://lists.lttng.org/cgi-bin/mailman/listinfo/lttng-dev

--
Mathieu Desnoyers
EfficiOS Inc.
http://www.efficios.com

Disclaimer

The information contained in this communication from the sender is confidential. It is intended solely for use by the recipient and others authorized to receive it. If you are not the recipient, you are hereby notified that any disclosure, copying, distribution or taking action in relation of the contents of this information is strictly prohibited and may be unlawful.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.lttng.org/pipermail/lttng-dev/attachments/20161031/89f001af/attachment-0001.html>