Re: kern/52700: Testbed VMs hang on wd1 access since recent ATA changes

[Jaromír Doleček <jaromir.dolecek%gmail.com@localhost>]

There was a fix for wdbiorestart(), but that should not affect normal operation.

I had a brief look into wd.c, seems after the dksubrification, the wddone() no longer calls ata_channel_start(), which likely will cause I/O to stall on machines with port multiplier. This should be fixed of course, but I don't think it causes this issue.

Jaromir

2017-11-05 14:40 GMT+01:00 Martin Husemann <martin%duskware.de@localhost>:

The following reply was made to PR kern/52700; it has been noted by GNATS.

From: Martin Husemann <martin%duskware.de@localhost>
To: gnats-bugs%NetBSD.org@localhost
Cc:
Subject: Re: kern/52700: Testbed VMs hang on wd1 access since recent ATA
 changes
Date: Sun, 5 Nov 2017 14:36:10 +0100

 On Sun, Nov 05, 2017 at 01:20:01PM +0000, Andreas Gustafsson wrote:
 > >Number:         52700
 > >Category:       kern
 > >Synopsis:       Testbed VMs hang on wd1 access since recent ATA changes
 > >Confidential:   no
 > >Severity:       critical
 > >Priority:       high
 > >Responsible:    kern-bug-people
 > >State:          open
 > >Class:          sw-bug
 > >Submitter-Id:   net
 > >Arrival-Date:   Sun Nov 05 13:20:00 +0000 2017
 > >Originator:     Andreas Gustafsson
 > >Release:        NetBSD-current, source date >= 2017.11.01.19.34.46
 > >Organization:
 >
 > >Environment:
 > System: NetBSD
 > Architecture: i386
 > Machine: i386
 > >Description:
 >
 > The test results at
 >
 >   http://releng.netbsd.org/b5reports/i386/commits-2017.11.html#2017.11.01.19.15.31
 >
 > include "ATF output: raw, xml, html" links as expected up to
 > and including source date 2017.11.01.19.15.31:
 >
 >   build: OK with 483383 lines of log, install: OK, tests: 6004 passed, 173 skipped, 123 expected_failure, 8 failed, ATF output: raw, xml, html
 >
 > but they are missing from the next run, of source date
 > 2017.11.01.21.13.26, and subsequent ones:
 >
 >   build: OK with 484342 lines of log, install: OK, tests: 6004 passed, 173 skipped, 123 expected_failure, 8 failed
 >
 > Since the ATF output files are exported from the test VM using a
 > secondary virtual hard disk (wd1), it seems likely that this is
 > related to the ATA changes committed between these two dates:
 >
 >   commit 2017.11.01.19.34.45 mlelstv src/sys/dev/files.dev 1.4

 There have been follow-up fixes on 2017.11.03, didn't those make this
 work again?

 Martin