Re: [PATCH 0/3] 64-bit futexes: Intro


* Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx> wrote:

IOW, I'm faking it, but I'm making a point. Namely that you can
efficiently do read-write lock using *only* 32-bit ops, and without
any real kind of limitations on the number of readers and writers.

So here goes the explanation and the pseudo-code.

- have two levels of locking: the contended case, and the uncontended
case

i suspect _any_ abstract locking functionality around a data structure
can be implemented via atomic control over just a single user-space bit.

That bit can be used as a lock and if all access to the state of that
atomic variable uses it, arbitrary higher-order atomic state transitions
can be derived from it. The cost would be a bit more instructions in the
fastpath, but there would still only be a single atomic op (the acquire
op), as the unlock would be a natural barrier (on x86 at least).

Concurrency (and scheduling) of that lock would still be exactly the
same as with genuine 64-bit (or even larger) atomic ops, and the
fastpath would be very close as well.

Ingo
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Relevant Pages

  • Re: using clustered index to optimize inserts ...
    ... I will try to explain locking in terms of Sybase docs... ... Allpages Locking: Allpages locking locks both data pages and index ... the data page is locked with an exclusive lock. ... Clustered Index: The datarows will be arranged as per the clustered ...
    (comp.databases.sybase)
  • Re: CSingleLock - known behaviour?
    ... It is better to design code that doesn't require locking. ... If you don't need the resource, don't lock it. ... magnitude less efficient, than locking once. ...
    (microsoft.public.vc.mfc)
  • Re: Strange multi-user timing phenomenon
    ... table record with pessimistic locking and both edit the same record. ... the recordset will be sufficient to lock out other users. ... update the recordset (if you had to edit data) and then close it. ...
    (microsoft.public.access.formscoding)
  • Re: CMultiLock example
    ... Seriously, though, locking is *mandatory* if two or more threads are accessing non-scalar ... lists ever be modified while the threads are running. ... without a lock. ... You must prevent that one piece of data is modified from multiple ...
    (microsoft.public.vc.mfc)
  • Re: Strange multi-user timing phenomenon
    ... I tried your locking method using a perssimistic recordset and it appears to ... user places his lock and grabs the same record. ... I would try opening a fixed single-row recordset with pessimistic locking, ...
    (microsoft.public.access.formscoding)