Project

General

Profile

Bug #5175

implement dmu_read_uio_dbuf() to improve cached read performance

Added by Matthew Ahrens almost 5 years ago. Updated almost 5 years ago.

Status:
Closed
Priority:
Normal
Category:
zfs - Zettabyte File System
Start date:
2014-09-17
Due date:
% Done:

100%

Estimated time:
Difficulty:
Medium
Tags:
needs-triage

Description

On a synthetic benchmark of cached reads of 8k blocks, lots of time is spent
holding the dnode. We should implement and use a dmu_read_uio_dbuf() that
takes the dbuf rather than the object ID, analogous to dmu_write_uio_dbuf().

Tests show that this results in a ~10% performance improvement.

History

#1

Updated by Electric Monk almost 5 years ago

  • % Done changed from 0 to 100
  • Status changed from New to Closed

git commit f8554bb92be6be88a684bea58293b41a6168a538

commit  f8554bb92be6be88a684bea58293b41a6168a538
Author: Matthew Ahrens <mahrens@delphix.com>
Date:   2014-10-07T18:10:05.000Z

    5175 implement dmu_read_uio_dbuf() to improve cached read performance
    Reviewed by: Adam Leventhal <ahl@delphix.com>
    Reviewed by: Alex Reece <alex.reece@delphix.com>
    Reviewed by: George Wilson <george@delphix.com>
    Reviewed by: Richard Elling <richard.elling@gmail.com>
    Approved by: Robert Mustacchi <rm@joyent.com>

Also available in: Atom PDF