Re: magic sysrq keys functionality

From: Matthias Andree <ma_at_dt.e-technik.uni-dortmund.de>
Date: Wed, 28 Jul 2004 11:44:43 +0200
Garrett Wollman <wollman_at_khavrinen.lcs.mit.edu> writes:

> <<On Mon, 26 Jul 2004 16:06:33 -0600 (MDT), Scott Long <scottl_at_FreeBSD.ORG> said:
>
>> This works right now because we assume that disks will commit blocks
>> in order, and that assumption generally hasn't been broken.
>
> I don't think soft updates cares about what order blocks are
> committed, because it will not in general consider a dependency
> resolved until it is notified that the buffer has been written.  What
> we do assume is that the disk (or driver) doesn't lie to us and claim
> that a block was written when it really wasn't.

Makes me wonder about efficiency (write latency).

I admit I'm not familiar with how the buffers are scheduled in
particular, if there are "write batches" or something.

If however softdep needs to wait for individual blocks, real tagged
queueing (with ordered tags in the right places and such) might be
faster because the drive can then decide for itself in which order the
blocks are written to the disks fastest, without violating any of the
ordering assumptions softupdates code relies on.

-- 
Matthias Andree

Encrypted mail welcome: my GnuPG key ID is 0x052E7D95 (PGP/MIME preferred)
Received on Wed Jul 28 2004 - 07:44:47 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:38:03 UTC