Re: port-arm/52666: ARM spinlocks are suboptimal

[Matt Thomas <matt%3am-software.com@localhost>]

The following reply was made to PR port-arm/52666; it has been noted by GNATS.

From: Matt Thomas <matt%3am-software.com@localhost>
To: gnats-bugs%NetBSD.org@localhost
Cc: port-arm-maintainer%netbsd.org@localhost,
 gnats-admin%netbsd.org@localhost,
 netbsd-bugs%netbsd.org@localhost
Subject: Re: port-arm/52666: ARM spinlocks are suboptimal
Date: Tue, 31 Oct 2017 16:10:03 -0700

 > On Oct 28, 2017, at 11:40 AM, coypu%SDF.ORG@localhost wrote:
 >=20
 >> Number:         52666
 >> Category:       port-arm
 >> Synopsis:       ARM spinlocks are suboptimal
 >> Confidential:   no
 >> Severity:       serious
 >> Priority:       medium
 >> Responsible:    port-arm-maintainer
 >> State:          open
 >> Class:          sw-bug
 >> Submitter-Id:   net
 >> Arrival-Date:   Sat Oct 28 18:40:00 +0000 2017
 >> Originator:     coypu
 >> Release:        NetBSD 8.99.5
 >> Organization:
 >> Environment:
 > NetBSD 8.99.5
 >> Description:
 > newer ARM has an instruction 'wfe' essentially made for spinlocks. we =
 should use it, but I don't see a relevant use of it, only an unused =
 macro.
 >=20
 > (In fact only x86 has a spinlock backoff function).
 >> How-To-Repeat:
 >=20
 >> Fix:
 >=20
 
 
 Note that the sev 'send event' instruction needs to be used in parallel. =
  When I experimented earlier I found that wfe/sev would many spurious =
 wakeups negating advantages.=