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 jdc@parodius.com | | Parodius Networking http://www.parodius.com/ | | UNIX Systems Administrator Mountain View, CA, US | | Making life hard for others since 1977. PGP 4BD6C0CB |
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> > 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! >-- ------------------- Mike Tancsa, tel +1 519 651 3400 Sentex Communications, mike@sentex.net Providing Internet services since 1994 www.sentex.net Cambridge, Ontario Canada http://www.tancsa.com/
Jeremy Chadwick <freebsd@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 Sm?rgrav - des@des.no
On Friday, February 03, 2012 3:48:17 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. > > 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 inconditional 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!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