Re: the idea for improving the performance of b-tree search

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



>> > 3) Do delayed inode deletion. ÂSee above, but change the word insert
>> > with delete ;)
>>
>> And the reserved-for-the-future-but-not-used-yet flags field in the
>> ioctl#21 control structure gets a use: bits to indicate for the
>> completion of which of the potential delayed things the caller is
>> waiting.
>
> This is similar to the delayed allocation we already do for file data
> extents. ÂIt is supposed to be transparent to userland, and just a way
> to more efficiently poke the on disk structures.
>
> -chris

What I mean is, if we want to give a userland way to see when any
deferred anything is done, using one wait queue that wakes all waiters
on all completion events will be simpler than having a separate wait
queue structure for every class of event, at the cost of unnecessary
tests when a waiter is woken on the wrong event. That is what is
traded off when deciding how many wait queues -- just like locks -- to
have; it's a granularity question.

Now that there is rmdir of empty subvolumes, though, ioctl#21 might be
redundant; I suppose that's your call.

Dave
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html


[Index of Archives]     [Linux Filesystem Development]     [Linux NFS]     [Linux NILFS]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux