sentex tinderbox build cluster failures

[Jeremy Chadwick]

Mike,

For many days now (almost a week?) there have been tinderbox build mails
coming from the sentex.ca build cluster, specific to powerpc, mips, and
sparc64 platforms.

The errors in question are strange and may indicate some kind of
filesystem corruption or equivalent -- I'm really not sure.

Breaking them down:

powerpc/powerpc, RELENG_8_2 --

/src/sys/powerpc/booke/platform_bare.c:82: error: expected '}' before ';' token

This file hasn't been touched in over 2 months:
http://www.freebsd.org/cgi/cvsweb.cgi/src/sys/powerpc/booke/platform_bare.c

mips/mips, RELENG_8_2 --

/src/sys/mips/alchemy/obio.c:513: warning: pointer type mismatch in conditional expression

This file hasn't been touched in over 2 months:
http://www.freebsd.org/cgi/cvsweb.cgi/src/sys/mips/alchemy/obio.c

mips/mips, RELENG_8_1 --

In file included from /src/sys/mips/sentry5/siba_cc.c:58:
/src/sys/dev/siba/sibavar.h:487: error: field 'sd_id' has incomplete type
/src/sys/dev/siba/sibavar.h:526: error: 'SIBA_MAX_CORES' undeclared here (not in a function)

src/sys/mips/sentry5/siba_cc.c has been Attic'd, and the last
time it was touched was 14 months ago:
http://www.freebsd.org/cgi/cvsweb.cgi/src/sys/mips/sentry5/Attic/siba_cc.c

src/sys/dev/siba/sibavar.h hasn't been touched in over 2 months:
http://www.freebsd.org/cgi/cvsweb.cgi/src/sys/dev/siba/sibavar.h

All recent tinderbox failures for RELENG_8 (for mips, ia64, and sparc64)
appear to be related to a commit Jack did ~7 hours ago, and are almost
certainly temporary/transient errors (e.g. csup/cvs pulldown didn't get
the full commit). So we can ignore those.

The others need to be investigated though.

Users on the freebsd-stable lists have been complaining about these
continual errors, so I was wondering if they're being investigated.
I also got one off-list mail from someone personally asking me if I knew
what could be going on or had any insights to this (not sure why that
person thought that -- I have nothing to do with tinderbox :-) ).

Let us know if you could. Thanks!

--
| Jeremy Chadwick j...@parodius.com |
| Parodius Networking http://www.parodius.com/ |
| UNIX Systems Administrator Mountain View, CA, US |
| Making life hard for others since 1977. PGP 4BD6C0CB |

_______________________________________________
freebsd...@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stabl...@freebsd.org"

[Mike Tancsa]

On 2/3/2012 3:48 AM, Jeremy Chadwick wrote:
> Mike,
>
> For many days now (almost a week?) there have been tinderbox build mails
> coming from the sentex.ca build cluster, specific to powerpc, mips, and
> sparc64 platforms.

Hi,
I will let des@freebsd speak to the errors. We host the boxes here,
but he is in charge of them and better understands the issue

---Mike

-------------------
Mike Tancsa, tel +1 519 651 3400
Sentex Communications, mi...@sentex.net
Providing Internet services since 1994 www.sentex.net
Cambridge, Ontario Canada http://www.tancsa.com/

[Dag-Erling Smørgrav]

Jeremy Chadwick <fre...@jdc.parodius.com> writes:
> For many days now (almost a week?) there have been tinderbox build mails
> coming from the sentex.ca build cluster, specific to powerpc, mips, and
> sparc64 platforms.
>
> The errors in question are strange and may indicate some kind of
> filesystem corruption or equivalent -- I'm really not sure.

You could have taken the time to check the code before jumping to
conclusions about "filesystem corruption or equivalent". These are
genuine bugs. The tinderbox didn't warn about them earlier because it
didn't build other kernels than LINT and GENERIC, and these bugs are in
machine-dependent code that isn't included in LINT and GENERIC. I
enabled additional kernels after the arm / mips people complained that
the arm / mips build was sometimes broken for weeks without anybody
noticing.

DES
--
Dag-Erling Smørgrav - d...@des.no

[John Baldwin]

The periodically failing ones are old bugs that were uncovered recently
when the tinderbox was fixed to compile kernels for these branches. Most
of these bugs are trivial to fix, but require committing a change to release
branches which requires a whole bunch of hoopla and red tape (otherwise I
would have merged the fixes by now).

--
John Baldwin