spinlock contention in ext4_getattr()

Solution Verified - Updated -

Issue

  • spinlock contention in ext4_getattr()
  • If ls is issued from multiple vsftpd daemon processes on a directory with a large number of files in it (like 80,000) they all contend on the i_block_reservation_lock spin lock as they all attempt to get details of the files in the directory. This effectively single threads all of the processes and produces a high contention rate on each inodes spinlock in series.

perf result

80.07%         vsftpd  [kernel.kallsyms]     [k] _spin_lock                 <==
 5.03%         vsftpd  ^E                     [.] 0x0000000000089745
 1.06%         vsftpd  [kernel.kallsyms]     [k] _atomic_dec_and_lock
 1.00%         :15904  [kernel.kallsyms]     [k] _spin_lock
 1.00%         :15918  [kernel.kallsyms]     [k] _spin_lock

Environment

  • Red Hat Enterprise Linux 6
    • kernel before kernel-2.6.32-642.4.2.el6

Subscriber exclusive content

A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.

Current Customers and Partners

Log in for full access

Log In

New to Red Hat?

Learn more about Red Hat subscriptions

Using a Red Hat product through a public cloud?

How to access this content