5.0 Stable (amd64) build appears broken.

classic Classic list List threaded Threaded
22 messages Options
12
Reply | Threaded
Open this post in threaded view
|

5.0 Stable (amd64) build appears broken.

Duncan Patton a Campbell
arrrgh. hit the send b4 that was finished, anyways


I install 5.0 AMD64 base..
# uname -a
OpenBSD jimg.indx.ca 5.0 GENERIC.MP#63 amd64

then fetch the patchbranch source..

export CVSROOT=[hidden email]:/cvs
cvs -d$CVSROOT checkout -rOPENBSD_5_0 -P src ports xenocara

then do the kernel make
# cd /usr/src/sys/arch/amd64/conf
# config GENERIC
config: symlink(machine -> ../../../../arch/amd64/include): File exists
config: symlink(amd64 -> machine): File exists
*** Stop.
(this is where we start to go sideways, but mebbe file extance is ok..)

# cd ../compile/GENERIC          
# make clean                      
# make
cc  -Werror -Wall -Wstrict-prototypes -Wmissing-prototypes  -Wno-main -Wno-uninitialized -Wno-format  -Wstack-larger-than-2047 -mcmodel=kernel -mno-red-zone -mno-sse2 -mno-sse -mno-3dnow  -mno-mmx -msoft-float -fno-omit-frame-pointer -fno-builtin-printf -fno-builtin-snprintf  -fno-builtin-vsnprintf -fno-builtin-log  -fno-builtin-log2 -fno-builtin-malloc -O2 -pipe -nostdinc -I. -I../../../.. -I../../../../arch -DDDB -DDIAGNOSTIC -DKTRACE -DACCOUNTING -DKMEMSTATS -DPTRACE -DCRYPTO -DSYSVMSG -DSYSVSEM -DSYSVSHM -DUVM_SWAP_ENCRYPT -DCOMPAT_43 -DCOMPAT_O48 -DLKM -DFFS -DFFS2 -DFFS_SOFTUPDATES -DUFS_DIRHASH -DQUOTA -DEXT2FS -DMFS -DNFSCLIENT -DNFSSERVER -DCD9660 -DUDF -DMSDOSFS -DFIFO -DSOCKET_SPLICE -DTCP_SACK -DTCP_ECN -DTCP_SIGNATURE -DINET -DALTQ -DINET6 -DIPSEC -DPPP_BSDCOMP -DPPP_DEFLATE -DPIPEX -DMROUTING -DMPLS -DBOOT_CONFIG -DUSER_PCICONF -DAPERTURE -DMTRR -DNTFS -DPCIVERBOSE -DUSBVERBOSE -DWSDISPLAY_COMPAT_USL -DWSDISPLAY_COMPAT_RAWKBD -DWSDISPLAY_DEFAULTSCREENS="6" -DWS!
 DISPLAY_COMPAT_PCVT -DX86EMU -DONEWIREVERBOSE -DMAXUSERS=80 -D_KERNEL -MD -MP  -c ioconf.c
cc1: warnings being treated as errors
ioconf.c:803: warning: excess elements in struct initializer
ioconf.c:803: warning: (near initialization for 'cfdata[0]')
ioconf.c:805: warning: excess elements in struct initializer
....
ioconf.c:1521: warning: (near initialization for 'cfdata[359]')
ioconf.c:1523: warning: excess elements in struct initializer
ioconf.c:1523: warning: (near initialization for 'cfdata[360]')
*** Error code 1

Stop in /usr/src/sys/arch/amd64/compile/GENERIC (line 735 of Makefile).
#

And thats all.  Same thing happens trying to build 5.0stable on a 4.9 system...

Any help would be appreciated,

Thanks,

Dhu


--
Duncan Patton a Campbell is Dhu >>> Ne obliviscaris, vix ea nostro voco.

Reply | Threaded
Open this post in threaded view
|

Re: 5.0 Stable (amd64) build appears broken.

Ted Unangst-6
you are running config from a different version than the source you
are trying to compile.

On Mon, Feb 20, 2012, Duncan Patton a Campbell wrote:

> arrrgh. hit the send b4 that was finished, anyways
>
>
> I install 5.0 AMD64 base..
> # uname -a
> OpenBSD jimg.indx.ca 5.0 GENERIC.MP#63 amd64
>
> then fetch the patchbranch source..
>
> export CVSROOT=[hidden email]:/cvs
> cvs -d$CVSROOT checkout -rOPENBSD_5_0 -P src ports xenocara
>
> then do the kernel make
> # cd /usr/src/sys/arch/amd64/conf
> # config GENERIC
> config: symlink(machine -> ../../../../arch/amd64/include): File exists
> config: symlink(amd64 -> machine): File exists
> *** Stop.
> (this is where we start to go sideways, but mebbe file extance is ok..)
>
> # cd ../compile/GENERIC
> # make clean
> # make
> cc  -Werror -Wall -Wstrict-prototypes -Wmissing-prototypes  -Wno-main
> -Wno-uninitialized -Wno-format  -Wstack-larger-than-2047 -mcmodel=kernel
> -mno-red-zone -mno-sse2 -mno-sse -mno-3dnow  -mno-mmx -msoft-float
> -fno-omit-frame-pointer -fno-builtin-printf -fno-builtin-snprintf
> -fno-builtin-vsnprintf -fno-builtin-log  -fno-builtin-log2
> -fno-builtin-malloc -O2 -pipe -nostdinc -I. -I../../../..
> -I../../../../arch -DDDB -DDIAGNOSTIC -DKTRACE -DACCOUNTING -DKMEMSTATS
> -DPTRACE -DCRYPTO -DSYSVMSG -DSYSVSEM -DSYSVSHM -DUVM_SWAP_ENCRYPT
> -DCOMPAT_43 -DCOMPAT_O48 -DLKM -DFFS -DFFS2 -DFFS_SOFTUPDATES
> -DUFS_DIRHASH -DQUOTA -DEXT2FS -DMFS -DNFSCLIENT -DNFSSERVER -DCD9660
> -DUDF -DMSDOSFS -DFIFO -DSOCKET_SPLICE -DTCP_SACK -DTCP_ECN
> -DTCP_SIGNATURE -DINET -DALTQ -DINET6 -DIPSEC -DPPP_BSDCOMP -DPPP_DEFLATE
> -DPIPEX -DMROUTING -DMPLS -DBOOT_CONFIG -DUSER_PCICONF -DAPERTURE -DMTRR
> -DNTFS -DPCIVERBOSE -DUSBVERBOSE -DWSDISPLAY_COMPAT_USL
> -DWSDISPLAY_COMPAT_RAWKBD -DWSDISPLAY_DEFAULTSCREENS="6" -DWS!
> DISPLAY_COMPAT_PCVT -DX86EMU -DONEWIREVERBOSE -DMAXUSERS=80 -D_KERNEL -MD
> -MP  -c ioconf.c
> cc1: warnings being treated as errors
> ioconf.c:803: warning: excess elements in struct initializer
> ioconf.c:803: warning: (near initialization for 'cfdata[0]')
> ioconf.c:805: warning: excess elements in struct initializer
> ....
> ioconf.c:1521: warning: (near initialization for 'cfdata[359]')
> ioconf.c:1523: warning: excess elements in struct initializer
> ioconf.c:1523: warning: (near initialization for 'cfdata[360]')
> *** Error code 1
>
> Stop in /usr/src/sys/arch/amd64/compile/GENERIC (line 735 of Makefile).
> #
>
> And thats all.  Same thing happens trying to build 5.0stable on a 4.9
> system...
>
> Any help would be appreciated,
>
> Thanks,
>
> Dhu
>
>
> --
> Duncan Patton a Campbell is Dhu >>> Ne obliviscaris, vix ea nostro voco.

Reply | Threaded
Open this post in threaded view
|

Re: 5.0 Stable (amd64) build appears broken.

Duncan Patton a Campbell
On Mon, 20 Feb 2012 10:05:19 -0500
Ted Unangst <[hidden email]> wrote:

> you are running config from a different version than the source you
> are trying to compile.

Ok, but if I've started with 5.0 Release, how do I get 5.0 Stable's config without
building it?  

d

>
> On Mon, Feb 20, 2012, Duncan Patton a Campbell wrote:
> > arrrgh. hit the send b4 that was finished, anyways
> >
> >
> > I install 5.0 AMD64 base..
> > # uname -a
> > OpenBSD jimg.indx.ca 5.0 GENERIC.MP#63 amd64
> >
> > then fetch the patchbranch source..
> >
> > export CVSROOT=[hidden email]:/cvs
> > cvs -d$CVSROOT checkout -rOPENBSD_5_0 -P src ports xenocara
> >
> > then do the kernel make
> > # cd /usr/src/sys/arch/amd64/conf
> > # config GENERIC
> > config: symlink(machine -> ../../../../arch/amd64/include): File exists
> > config: symlink(amd64 -> machine): File exists
> > *** Stop.
> > (this is where we start to go sideways, but mebbe file extance is ok..)
> >
> > # cd ../compile/GENERIC
> > # make clean
> > # make
> > cc  -Werror -Wall -Wstrict-prototypes -Wmissing-prototypes  -Wno-main
> > -Wno-uninitialized -Wno-format  -Wstack-larger-than-2047 -mcmodel=kernel
> > -mno-red-zone -mno-sse2 -mno-sse -mno-3dnow  -mno-mmx -msoft-float
> > -fno-omit-frame-pointer -fno-builtin-printf -fno-builtin-snprintf
> > -fno-builtin-vsnprintf -fno-builtin-log  -fno-builtin-log2
> > -fno-builtin-malloc -O2 -pipe -nostdinc -I. -I../../../..
> > -I../../../../arch -DDDB -DDIAGNOSTIC -DKTRACE -DACCOUNTING -DKMEMSTATS
> > -DPTRACE -DCRYPTO -DSYSVMSG -DSYSVSEM -DSYSVSHM -DUVM_SWAP_ENCRYPT
> > -DCOMPAT_43 -DCOMPAT_O48 -DLKM -DFFS -DFFS2 -DFFS_SOFTUPDATES
> > -DUFS_DIRHASH -DQUOTA -DEXT2FS -DMFS -DNFSCLIENT -DNFSSERVER -DCD9660
> > -DUDF -DMSDOSFS -DFIFO -DSOCKET_SPLICE -DTCP_SACK -DTCP_ECN
> > -DTCP_SIGNATURE -DINET -DALTQ -DINET6 -DIPSEC -DPPP_BSDCOMP -DPPP_DEFLATE
> > -DPIPEX -DMROUTING -DMPLS -DBOOT_CONFIG -DUSER_PCICONF -DAPERTURE -DMTRR
> > -DNTFS -DPCIVERBOSE -DUSBVERBOSE -DWSDISPLAY_COMPAT_USL
> > -DWSDISPLAY_COMPAT_RAWKBD -DWSDISPLAY_DEFAULTSCREENS="6" -DWS!
> > DISPLAY_COMPAT_PCVT -DX86EMU -DONEWIREVERBOSE -DMAXUSERS=80 -D_KERNEL -MD
> > -MP  -c ioconf.c
> > cc1: warnings being treated as errors
> > ioconf.c:803: warning: excess elements in struct initializer
> > ioconf.c:803: warning: (near initialization for 'cfdata[0]')
> > ioconf.c:805: warning: excess elements in struct initializer
> > ....
> > ioconf.c:1521: warning: (near initialization for 'cfdata[359]')
> > ioconf.c:1523: warning: excess elements in struct initializer
> > ioconf.c:1523: warning: (near initialization for 'cfdata[360]')
> > *** Error code 1
> >
> > Stop in /usr/src/sys/arch/amd64/compile/GENERIC (line 735 of Makefile).
> > #
> >
> > And thats all.  Same thing happens trying to build 5.0stable on a 4.9
> > system...
> >
> > Any help would be appreciated,
> >
> > Thanks,
> >
> > Dhu
> >
> >
> > --
> > Duncan Patton a Campbell is Dhu >>> Ne obliviscaris, vix ea nostro voco.
>


--
Duncan Patton a Campbell is Dhu >>> Ne obliviscaris, vix ea nostro voco.

Reply | Threaded
Open this post in threaded view
|

Re: 5.0 Stable (amd64) build appears broken.

Ted Unangst-6
In reply to this post by Duncan Patton a Campbell
On Mon, Feb 20, 2012, Duncan Patton a Campbell wrote:
> On Mon, 20 Feb 2012 10:05:19 -0500
> Ted Unangst <[hidden email]> wrote:
>
>> you are running config from a different version than the source you
>> are trying to compile.
>
> Ok, but if I've started with 5.0 Release, how do I get 5.0 Stable's config
> without
> building it?

I would revisit the assumption you are building 5.0 on 5.0.  You are
reporting an error nobody else is reporting, that just happens to be
exactly the error one sees when building with a mismatched config and
src, so.... You see where I'm coming from?

Reply | Threaded
Open this post in threaded view
|

Re: 5.0 Stable (amd64) build appears broken.

Duncan Patton a Campbell
On Mon, 20 Feb 2012 23:24:56 -0500
Ted Unangst <[hidden email]> wrote:

> On Mon, Feb 20, 2012, Duncan Patton a Campbell wrote:
> > On Mon, 20 Feb 2012 10:05:19 -0500
> > Ted Unangst <[hidden email]> wrote:
> >
> >> you are running config from a different version than the source you
> >> are trying to compile.
> >
> > Ok, but if I've started with 5.0 Release, how do I get 5.0 Stable's config
> > without
> > building it?
>
> I would revisit the assumption you are building 5.0 on 5.0.  You are
> reporting an error nobody else is reporting, that just happens to be
> exactly the error one sees when building with a mismatched config and
> src, so.... You see where I'm coming from?
>

Yes.  I do.  I can only assume I've buggerd up the src tree somehow.

uname -svmpr
OpenBSD 5.0 GENERIC.MP#63 amd64 AMD A6-3670 APU with Radeon(tm) HD Graphics

thanks,

D

--
Duncan Patton a Campbell is Dhu >>> Ne obliviscaris, vix ea nostro voco.

Reply | Threaded
Open this post in threaded view
|

Re: 5.0 Stable (amd64) build appears broken.

Philip Guenther-2
On Tue, Feb 21, 2012 at 8:16 AM, Duncan Patton a Campbell
<[hidden email]> wrote:

> On Mon, 20 Feb 2012 23:24:56 -0500
> Ted Unangst <[hidden email]> wrote:
>
>> On Mon, Feb 20, 2012, Duncan Patton a Campbell wrote:
>> > On Mon, 20 Feb 2012 10:05:19 -0500
>> > Ted Unangst <[hidden email]> wrote:
>> >
>> >> you are running config from a different version than the source you
>> >> are trying to compile.
>> >
>> > Ok, but if I've started with 5.0 Release, how do I get 5.0 Stable's
config

>> > without
>> > building it?
>>
>> I would revisit the assumption you are building 5.0 on 5.0.  You are
>> reporting an error nobody else is reporting, that just happens to be
>> exactly the error one sees when building with a mismatched config and
>> src, so.... You see where I'm coming from?
>>
>
> Yes.  I do.  I can only assume I've buggerd up the src tree somehow.
>
> uname -svmpr
> OpenBSD 5.0 GENERIC.MP#63 amd64 AMD A6-3670 APU with Radeon(tm) HD Graphics

It's most likely that you're currently running a 5.0 kernel with 4.9
binaries.  To upgrade from 4.9 to 5.0 by building from source you
would have needed to follow the steps in the "following -current" page
of the FAQ during the run-up to the 5.0 release.  Once 5.0 was
released, the presumption became that you'll install 5.0 sets and work
from there.

Probably the simplest way forward is to just follow the steps outlined here:
   http://www.openbsd.org/faq/upgrade50.html

That even documents the other configuration changes you'll need to
consider as part of the upgrade, something which building from source
does *not* automatically do for you.


Philip Guenther

Reply | Threaded
Open this post in threaded view
|

Re: 5.0 Stable (amd64) build appears broken.

Duncan Patton a Campbell
In reply to this post by Duncan Patton a Campbell
On Tue, 21 Feb 2012 09:16:34 -0700
Duncan Patton a Campbell <[hidden email]> wrote:

> On Mon, 20 Feb 2012 23:24:56 -0500
> Ted Unangst <[hidden email]> wrote:
>
> > On Mon, Feb 20, 2012, Duncan Patton a Campbell wrote:
> > > On Mon, 20 Feb 2012 10:05:19 -0500
> > > Ted Unangst <[hidden email]> wrote:
> > >
> > >> you are running config from a different version than the source you
> > >> are trying to compile.
> > >
> > > Ok, but if I've started with 5.0 Release, how do I get 5.0 Stable's config
> > > without
> > > building it?
> >
> > I would revisit the assumption you are building 5.0 on 5.0.  You are
> > reporting an error nobody else is reporting, that just happens to be
> > exactly the error one sees when building with a mismatched config and
> > src, so.... You see where I'm coming from?
> >
>
> Yes.  I do.  I can only assume I've buggerd up the src tree somehow.
>
> uname -svmpr
> OpenBSD 5.0 GENERIC.MP#63 amd64 AMD A6-3670 APU with Radeon(tm) HD Graphics
>
> thanks,
>

So, the problem appears to be that you cannot check out a 5.0-Stable source tree
directly on a 5.0-Release system, but instead must "prime" the tree with the
Release code and then checkout the Stable stuff on top.

thanks,

Dhu

> D
>
> --
> Duncan Patton a Campbell is Dhu >>> Ne obliviscaris, vix ea nostro voco.
>


--
Duncan Patton a Campbell is Dhu >>> Ne obliviscaris, vix ea nostro voco.

Reply | Threaded
Open this post in threaded view
|

Re: 5.0 Stable (amd64) build appears broken.

Duncan Patton a Campbell
In reply to this post by Philip Guenther-2
On Tue, 21 Feb 2012 09:53:48 -0800
Philip Guenther <[hidden email]> wrote:

> On Tue, Feb 21, 2012 at 8:16 AM, Duncan Patton a Campbell
> <[hidden email]> wrote:
> > On Mon, 20 Feb 2012 23:24:56 -0500
> > Ted Unangst <[hidden email]> wrote:
> >
> >> On Mon, Feb 20, 2012, Duncan Patton a Campbell wrote:
> >> > On Mon, 20 Feb 2012 10:05:19 -0500
> >> > Ted Unangst <[hidden email]> wrote:
> >> >
> >> >> you are running config from a different version than the source you
> >> >> are trying to compile.
> >> >
> >> > Ok, but if I've started with 5.0 Release, how do I get 5.0 Stable's
> config
> >> > without
> >> > building it?
> >>
> >> I would revisit the assumption you are building 5.0 on 5.0.  You are
> >> reporting an error nobody else is reporting, that just happens to be
> >> exactly the error one sees when building with a mismatched config and
> >> src, so.... You see where I'm coming from?
> >>
> >
> > Yes.  I do.  I can only assume I've buggerd up the src tree somehow.
> >
> > uname -svmpr
> > OpenBSD 5.0 GENERIC.MP#63 amd64 AMD A6-3670 APU with Radeon(tm) HD Graphics
>
> It's most likely that you're currently running a 5.0 kernel with 4.9
> binaries.  To upgrade from 4.9 to 5.0 by building from source you
> would have needed to follow the steps in the "following -current" page
> of the FAQ during the run-up to the 5.0 release.  Once 5.0 was
> released, the presumption became that you'll install 5.0 sets and work
> from there.
>
> Probably the simplest way forward is to just follow the steps outlined here:
>    http://www.openbsd.org/faq/upgrade50.html

There's a minor error in http://www.openbsd.org/faq/faq5.html such that
it specifies that to "follow stable" you can pull the whole tree with cvs
and run as such.  This is not the case.  The 5.0-Stable tree will not
build without the tree being first primed with the 5.0-Release code.

Dhu
>
> That even documents the other configuration changes you'll need to
> consider as part of the upgrade, something which building from source
> does *not* automatically do for you.
>
>
> Philip Guenther
>


--
Duncan Patton a Campbell is Dhu >>> Ne obliviscaris, vix ea nostro voco.

Reply | Threaded
Open this post in threaded view
|

Re: 5.0 Stable (amd64) build appears broken.

Miod Vallat
> There's a minor error in http://www.openbsd.org/faq/faq5.html such that
> it specifies that to "follow stable" you can pull the whole tree with cvs
> and run as such.  This is not the case.  The 5.0-Stable tree will not
> build without the tree being first primed with the 5.0-Release code.

Of course it will, BUT you need to build from a 5.0 userland, which
means that you should not update by extracting the `base' set only.

Reply | Threaded
Open this post in threaded view
|

Re: 5.0 Stable (amd64) build appears broken.

Bryan Irvine
In reply to this post by Duncan Patton a Campbell
On Tue, Feb 21, 2012 at 10:28 AM, Duncan Patton a Campbell
<[hidden email]> wrote:

> On Tue, 21 Feb 2012 09:53:48 -0800
> Philip Guenther <[hidden email]> wrote:
>
>> On Tue, Feb 21, 2012 at 8:16 AM, Duncan Patton a Campbell
>> <[hidden email]> wrote:
>> > On Mon, 20 Feb 2012 23:24:56 -0500
>> > Ted Unangst <[hidden email]> wrote:
>> >
>> >> On Mon, Feb 20, 2012, Duncan Patton a Campbell wrote:
>> >> > On Mon, 20 Feb 2012 10:05:19 -0500
>> >> > Ted Unangst <[hidden email]> wrote:
>> >> >
>> >> >> you are running config from a different version than the source you
>> >> >> are trying to compile.
>> >> >
>> >> > Ok, but if I've started with 5.0 Release, how do I get 5.0 Stable's
>> config
>> >> > without
>> >> > building it?
>> >>
>> >> I would revisit the assumption you are building 5.0 on 5.0.  You are
>> >> reporting an error nobody else is reporting, that just happens to be
>> >> exactly the error one sees when building with a mismatched config and
>> >> src, so.... You see where I'm coming from?
>> >>
>> >
>> > Yes.  I do.  I can only assume I've buggerd up the src tree somehow.
>> >
>> > uname -svmpr
>> > OpenBSD 5.0 GENERIC.MP#63 amd64 AMD A6-3670 APU with Radeon(tm) HD
Graphics
>>
>> It's most likely that you're currently running a 5.0 kernel with 4.9
>> binaries.  To upgrade from 4.9 to 5.0 by building from source you
>> would have needed to follow the steps in the "following -current" page
>> of the FAQ during the run-up to the 5.0 release.  Once 5.0 was
>> released, the presumption became that you'll install 5.0 sets and work
>> from there.
>>
>> Probably the simplest way forward is to just follow the steps outlined
here:
>>    http://www.openbsd.org/faq/upgrade50.html
>
> There's a minor error in http://www.openbsd.org/faq/faq5.html such that
> it specifies that to "follow stable" you can pull the whole tree with cvs
> and run as such.  This is not the case.  The 5.0-Stable tree will not
> build without the tree being first primed with the 5.0-Release code.

I think you are doing something wrong. I haven't 'primed' it before a
build for at least 5 years now.  I suspect you might be trying to
build -current (5.1-beta) on a 5.0 -release system.

-B

Reply | Threaded
Open this post in threaded view
|

Re: 5.0 Stable (amd64) build appears broken.

Duncan Patton a Campbell
In reply to this post by Miod Vallat
On Tue, 21 Feb 2012 18:38:25 +0000
Miod Vallat <[hidden email]> wrote:

> > There's a minor error in http://www.openbsd.org/faq/faq5.html such that
> > it specifies that to "follow stable" you can pull the whole tree with cvs
> > and run as such.  This is not the case.  The 5.0-Stable tree will not
> > build without the tree being first primed with the 5.0-Release code.
>
> Of course it will, BUT you need to build from a 5.0 userland, which
> means that you should not update by extracting the `base' set only.
>

The 5.0-Release userland appears insufficient to this: this was a novel
installation from a 5.0Rel disk image:


previously, after a cd install of the rel image and direct cvs pull
of the src tree using this kernel:

uname -svmpr
OpenBSD 5.0 GENERIC.MP#63 amd64 AMD A6-3670 APU with Radeon(tm) HD Graphics
compilation of the new kernel resulted in the mixed error.  

after priming the /usr/src tree from sys.tar.gz and src.tar.gz, and
then updating with cvs, this same kernel compiled correctly.  Basically
it was Ted's obs that tipped me off because most systems he speaks of
would be incrementally built.

uname -svmpr
OpenBSD 5.0 GENERIC.MP#0 amd64 AMD A6-3670 APU with Radeon(tm) HD Graphics
is now compiling userland

Dhu

--
Duncan Patton a Campbell is Dhu >>> Ne obliviscaris, vix ea nostro voco.

Reply | Threaded
Open this post in threaded view
|

Re: 5.0 Stable (amd64) build appears broken.

Nick Holland
In reply to this post by Duncan Patton a Campbell
On 02/21/2012 01:28 PM, Duncan Patton a Campbell wrote:
...
> There's a minor error in http://www.openbsd.org/faq/faq5.html such that
> it specifies that to "follow stable" you can pull the whole tree with cvs
> and run as such.  This is not the case.  The 5.0-Stable tree will not
> build without the tree being first primed with the 5.0-Release code.
>
> Dhu

BULLSHIT.

Complete and total bullshit.

Nick.

Reply | Threaded
Open this post in threaded view
|

Re: 5.0 Stable (amd64) build appears broken.

Philip Guenther-2
In reply to this post by Duncan Patton a Campbell
Going back to the first message in the thread...

On Mon, Feb 20, 2012 at 5:58 AM, Duncan Patton a Campbell
<[hidden email]> wrote:
...
> then do the kernel make
> # cd /usr/src/sys/arch/amd64/conf
> # config GENERIC
> config: symlink(machine -> ../../../../arch/amd64/include): File exists
> config: symlink(amd64 -> machine): File exists
> *** Stop.
> (this is where we start to go sideways, but mebbe file extance is ok..)

config failed.  If you don't understand why it failed, then you should
assume the resulting compile directory cannot be used, as it might not
have written the correct files.

So, go back and figure out what happened there.  Read-only mount?
Corrupt filesystem?  Aliens?


Philip Guenther

Reply | Threaded
Open this post in threaded view
|

Re: 5.0 Stable (amd64) build appears broken.

Duncan Patton a Campbell
In reply to this post by Miod Vallat
On Tue, 21 Feb 2012 18:38:25 +0000
Miod Vallat <[hidden email]> wrote:

> > There's a minor error in http://www.openbsd.org/faq/faq5.html such that
> > it specifies that to "follow stable" you can pull the whole tree with cvs
> > and run as such.  This is not the case.  The 5.0-Stable tree will not
> > build without the tree being first primed with the 5.0-Release code.
>
> Of course it will, BUT you need to build from a 5.0 userland, which
> means that you should not update by extracting the `base' set only.

This appears to be essentially correct.  The problem was this:
(from the machine)

ls -l /usr/
total 96
drwxr-xr-x   7 root  wheel   512 Aug 11  2011 X11R6
drwxr-xr-x   2 root  wheel  6144 Feb 20 00:25 bin
drwxr-xr-x   2 root  wheel  1024 Feb 20 00:25 games
drwxr-xr-x  34 root  bin    3072 Feb 21 11:35 include
drwxr-xr-x   6 root  wheel  5120 Feb 21 11:55 lib
drwxr-xr-x   6 root  wheel   512 Feb 20 00:25 libdata
drwxr-xr-x   8 root  wheel  1536 Feb 20 00:25 libexec
drwxr-xr-x   2 root  wheel   512 Aug 17  2011 lkm
drwxr-xr-x  11 root  wheel   512 Feb 20 00:31 local
drwxr-xr-x   2 root  wheel   512 Aug 17  2011 mdec
lrwxr-xr-x   1 root  wheel     9 Feb 21 11:33 obj -> /dmnt/obj
drwxrwxr-x   2 root  wsrc    512 Aug 17  2011 obj.X1
lrwxr-xr-x   1 root  wheel    11 Feb 21 10:00 ports -> /dmnt/ports
drwxr-xr-x  46 root  wheel  1536 Feb 20 06:47 ports.X1
drwxr-xr-x   3 root  wheel   512 Feb 20 07:30 pub
drwxr-xr-x   2 root  wheel  4096 Feb 20 00:25 sbin
drwxr-xr-x  19 root  wheel   512 Aug 17  2011 share
lrwxr-xr-x   1 root  wheel     9 Feb 21 09:57 src -> /dmnt/src
drwxr-xr-x  18 root  wheel   512 Feb 20 06:47 src.X1
lrwxr-xr-x   1 root  wheel    14 Feb 21 10:00 xenocara -> /dmnt/xenocara
drwxr-xr-x  11 root  wheel   512 Feb 20 06:45 xenocara.X1
lrwxr-xr-x   1 root  wheel    10 Feb 21 12:06 xobj -> /dmnt/xobj
drwxrwxr-x   2 root  wsrc    512 Aug 17  2011 xobj.X1

Rather than update the original sources (I wanted to keep them around)
I sidlined them and linked to a 1T partition where I would build
using the newly pulled CVS (I'm pretty close to the site, netwise)
sources.  

Dhu



--
Duncan Patton a Campbell is Dhu >>> Ne obliviscaris, vix ea nostro voco.

Reply | Threaded
Open this post in threaded view
|

Re: 5.0 Stable (amd64) build appears broken.

Carson Chittom-3
In reply to this post by Duncan Patton a Campbell
Duncan Patton a Campbell <[hidden email]> writes:

> On Tue, 21 Feb 2012 09:16:34 -0700
> Duncan Patton a Campbell <[hidden email]> wrote:
>
>> On Mon, 20 Feb 2012 23:24:56 -0500
>> Ted Unangst <[hidden email]> wrote:
>>
>> > On Mon, Feb 20, 2012, Duncan Patton a Campbell wrote:
>> > > On Mon, 20 Feb 2012 10:05:19 -0500
>> > > Ted Unangst <[hidden email]> wrote:
>> > >
>> > >> you are running config from a different version than the source you
>> > >> are trying to compile.
>> > >
>> > > Ok, but if I've started with 5.0 Release, how do I get 5.0 Stable's config
>> > > without
>> > > building it?
>> >
>> > I would revisit the assumption you are building 5.0 on 5.0.  You are
>> > reporting an error nobody else is reporting, that just happens to be
>> > exactly the error one sees when building with a mismatched config and
>> > src, so.... You see where I'm coming from?
>> >
>>
>> Yes.  I do.  I can only assume I've buggerd up the src tree somehow.
>>
>> uname -svmpr
>> OpenBSD 5.0 GENERIC.MP#63 amd64 AMD A6-3670 APU with Radeon(tm) HD Graphics
>>
>> thanks,
>>
>
> So, the problem appears to be that you cannot check out a 5.0-Stable source tree
> directly on a 5.0-Release system, but instead must "prime" the tree with the
> Release code and then checkout the Stable stuff on top.

No.  You're wrong.  At least, I did exactly what you said I "cannot"
do.  I installed (from CD) 5.0-release, and then cvs co'd -stable;
and then installed per the usual directions.  From dmesg:

OpenBSD 5.0-stable (GENERIC.MP) #0: Mon Jan 30 13:21:14 CST 2012
    [hidden email]:/usr/src/sys/arch/amd64/compile/GENERIC.MP

So unless something very strange has happened since January 30, you
appear to be Doing Something Wrong.

Reply | Threaded
Open this post in threaded view
|

Re: 5.0 Stable (amd64) build appears broken.

Duncan Patton a Campbell
On Tue, 21 Feb 2012 13:52:27 -0600
Carson Chittom <[hidden email]> wrote:

> Duncan Patton a Campbell <[hidden email]> writes:
>
> > On Tue, 21 Feb 2012 09:16:34 -0700
> > Duncan Patton a Campbell <[hidden email]> wrote:
> >
> >> On Mon, 20 Feb 2012 23:24:56 -0500
> >> Ted Unangst <[hidden email]> wrote:
> >>
> >> > On Mon, Feb 20, 2012, Duncan Patton a Campbell wrote:
> >> > > On Mon, 20 Feb 2012 10:05:19 -0500
> >> > > Ted Unangst <[hidden email]> wrote:
> >> > >
> >> > >> you are running config from a different version than the source you
> >> > >> are trying to compile.
> >> > >
> >> > > Ok, but if I've started with 5.0 Release, how do I get 5.0 Stable's config
> >> > > without
> >> > > building it?
> >> >
> >> > I would revisit the assumption you are building 5.0 on 5.0.  You are
> >> > reporting an error nobody else is reporting, that just happens to be
> >> > exactly the error one sees when building with a mismatched config and
> >> > src, so.... You see where I'm coming from?
> >> >
> >>
> >> Yes.  I do.  I can only assume I've buggerd up the src tree somehow.
> >>
> >> uname -svmpr
> >> OpenBSD 5.0 GENERIC.MP#63 amd64 AMD A6-3670 APU with Radeon(tm) HD Graphics
> >>
> >> thanks,
> >>
> >
> > So, the problem appears to be that you cannot check out a 5.0-Stable source tree
> > directly on a 5.0-Release system, but instead must "prime" the tree with the
> > Release code and then checkout the Stable stuff on top.
>
> No.  You're wrong.  At least, I did exactly what you said I "cannot"
> do.  I installed (from CD) 5.0-release, and then cvs co'd -stable;
> and then installed per the usual directions.  From dmesg:
>

read and weap.  i did.  when you do a cd install, it puts
src (sys), and xenocara in /usr.  that "primes" the src/sys
tree.  if you then _move_ those trees out of the way entirely,
and do a cvs checkout of the whole tree, well that what *I* saw
anyways.

Dhu

> OpenBSD 5.0-stable (GENERIC.MP) #0: Mon Jan 30 13:21:14 CST 2012
>     [hidden email]:/usr/src/sys/arch/amd64/compile/GENERIC.MP
>
> So unless something very strange has happened since January 30, you
> appear to be Doing Something Wrong.
>


--
Duncan Patton a Campbell is Dhu >>> Ne obliviscaris, vix ea nostro voco.

Reply | Threaded
Open this post in threaded view
|

Re: 5.0 Stable (amd64) build appears broken.

patrick keshishian
On Tue, Feb 21, 2012 at 2:37 PM, Duncan Patton a Campbell
<[hidden email]> wrote:

> On Tue, 21 Feb 2012 13:52:27 -0600
> Carson Chittom <[hidden email]> wrote:
>
>> Duncan Patton a Campbell <[hidden email]> writes:
>>
>> > On Tue, 21 Feb 2012 09:16:34 -0700
>> > Duncan Patton a Campbell <[hidden email]> wrote:
>> >
>> >> On Mon, 20 Feb 2012 23:24:56 -0500
>> >> Ted Unangst <[hidden email]> wrote:
>> >>
>> >> > On Mon, Feb 20, 2012, Duncan Patton a Campbell wrote:
>> >> > > On Mon, 20 Feb 2012 10:05:19 -0500
>> >> > > Ted Unangst <[hidden email]> wrote:
>> >> > >
>> >> > >> you are running config from a different version than the source
you
>> >> > >> are trying to compile.
>> >> > >
>> >> > > Ok, but if I've started with 5.0 Release, how do I get 5.0 Stable's
config

>> >> > > without
>> >> > > building it?
>> >> >
>> >> > I would revisit the assumption you are building 5.0 on 5.0.  You are
>> >> > reporting an error nobody else is reporting, that just happens to be
>> >> > exactly the error one sees when building with a mismatched config and
>> >> > src, so.... You see where I'm coming from?
>> >> >
>> >>
>> >> Yes.  I do.  I can only assume I've buggerd up the src tree somehow.
>> >>
>> >> uname -svmpr
>> >> OpenBSD 5.0 GENERIC.MP#63 amd64 AMD A6-3670 APU with Radeon(tm) HD
Graphics
>> >>
>> >> thanks,
>> >>
>> >
>> > So, the problem appears to be that you cannot check out a 5.0-Stable
source tree
>> > directly on a 5.0-Release system, but instead must "prime" the tree with
the

>> > Release code and then checkout the Stable stuff on top.
>>
>> No.  You're wrong.  At least, I did exactly what you said I "cannot"
>> do.  I installed (from CD) 5.0-release, and then cvs co'd -stable;
>> and then installed per the usual directions.  From dmesg:
>>
>
> read and weap.  i did.  when you do a cd install, it puts
> src (sys), and xenocara in /usr.  that "primes" the src/sys
> tree.  if you then _move_ those trees out of the way entirely,
> and do a cvs checkout of the whole tree, well that what *I* saw
> anyways.

It depends on the exact cvs checkout options you use.

--patrick

Reply | Threaded
Open this post in threaded view
|

Re: 5.0 Stable (amd64) build appears broken.

Josh Grosse
In reply to this post by Duncan Patton a Campbell
My apologies if the redactions cause improper attributions, but
this has gotten very deep, and Dhu replied to himself.  For readability
I've tried to wrap lines, too.

On Tue, Feb 21, 2012 at 03:37:25PM -0700, Duncan Patton a Campbell wrote:
> On Tue, 21 Feb 2012 13:52:27 -0600
> Carson Chittom <[hidden email]> wrote:
>
> > Duncan Patton a Campbell <[hidden email]> writes:
> >
> > > On Tue, 21 Feb 2012 09:16:34 -0700
> > > Duncan Patton a Campbell <[hidden email]> wrote:
...[snip]...

> > > So, the problem appears to be that you cannot check out a 5.0-Stable
> > > source tree
> > > directly on a 5.0-Release system, but instead must "prime" the
> > > tree with the
> > > Release code and then checkout the Stable stuff on top.
> >
> > No.  You're wrong.  At least, I did exactly what you said I "cannot"
> > do.  I installed (from CD) 5.0-release, and then cvs co'd -stable;
> > and then installed per the usual directions.  From dmesg:
> >
>
> read and weap.  i did.  when you do a cd install, it puts
> src (sys), and xenocara in /usr.  that "primes" the src/sys
> tree.  if you then _move_ those trees out of the way entirely,
> and do a cvs checkout of the whole tree, well that what *I* saw

Many people, for many years, have been able to checkout a tagged working
directory from CVS.  I believe this is a PEBKAC, but not the one described.

Because the error appeared to be an out-of-sync issue, I would have liked
to see the output of:  

ls -l `which config`
cat /usr/src/sys/CVS/Tag

That would have at least confirmed if the kernel source tree and the config
were of appropriate vintages.

Dhu, an additional suggestion: instead of uname -anything, please use
sysctl kern.version.  It produces far more useful information about your
specific kernel. I sent you a note earlier today via Email, apparently it's
still in your MTA queue.

Reply | Threaded
Open this post in threaded view
|

Re: 5.0 Stable (amd64) build appears broken.

Duncan Patton a Campbell
On Tue, 21 Feb 2012 18:00:20 -0500
Josh Grosse <[hidden email]> wrote:

> My apologies if the redactions cause improper attributions, but
> this has gotten very deep, and Dhu replied to himself.  For readability
> I've tried to wrap lines, too.
>
> On Tue, Feb 21, 2012 at 03:37:25PM -0700, Duncan Patton a Campbell wrote:
> > On Tue, 21 Feb 2012 13:52:27 -0600
> > Carson Chittom <[hidden email]> wrote:
> >
> > > Duncan Patton a Campbell <[hidden email]> writes:
> > >
> > > > On Tue, 21 Feb 2012 09:16:34 -0700
> > > > Duncan Patton a Campbell <[hidden email]> wrote:
> ...[snip]...
>
> > > > So, the problem appears to be that you cannot check out a 5.0-Stable
> > > > source tree
> > > > directly on a 5.0-Release system, but instead must "prime" the
> > > > tree with the
> > > > Release code and then checkout the Stable stuff on top.
> > >
> > > No.  You're wrong.  At least, I did exactly what you said I "cannot"
> > > do.  I installed (from CD) 5.0-release, and then cvs co'd -stable;
> > > and then installed per the usual directions.  From dmesg:
> > >
> >
> > read and weap.  i did.  when you do a cd install, it puts
> > src (sys), and xenocara in /usr.  that "primes" the src/sys
> > tree.  if you then _move_ those trees out of the way entirely,
> > and do a cvs checkout of the whole tree, well that what *I* saw
>
> Many people, for many years, have been able to checkout a tagged working
> directory from CVS.  I believe this is a PEBKAC, but not the one described.
>
> Because the error appeared to be an out-of-sync issue, I would have liked
> to see the output of:  
>
> ls -l `which config`
> cat /usr/src/sys/CVS/Tag
>

# ls -l `which config`
-r-xr-xr-x  1 root  bin  97464 Feb 21 12:29 /usr/sbin/config
# cat /usr/src/sys/CVS/Tag
TOPENBSD_5_0


> That would have at least confirmed if the kernel source tree and the config
> were of appropriate vintages.
>
> Dhu, an additional suggestion: instead of uname -anything, please use
> sysctl kern.version.  It produces far more useful information about your
> specific kernel. I sent you a note earlier today via Email, apparently it's
> still in your MTA queue.

Yes.  I have it... now ;/

>


--
Duncan Patton a Campbell is Dhu >>> Ne obliviscaris, vix ea nostro voco.

Reply | Threaded
Open this post in threaded view
|

Re: 5.0 Stable (amd64) build appears broken.

Josh Grosse
On Tue, Feb 21, 2012 at 04:28:11PM -0700, Duncan Patton a Campbell wrote:

> On Tue, 21 Feb 2012 18:00:20 -0500
> Josh Grosse <[hidden email]> wrote:
>
> > Because the error appeared to be an out-of-sync issue, I would have liked
> > to see the output of:  
> >
> > ls -l `which config`
> > cat /usr/src/sys/CVS/Tag
> >
>
> # ls -l `which config`
> -r-xr-xr-x  1 root  bin  97464 Feb 21 12:29 /usr/sbin/config

That executable was just built... so it is not the executable from
5.0-release/amd64.  That should have an August 17 date stamp, if memory
serves from earlier today.

> # cat /usr/src/sys/CVS/Tag
> TOPENBSD_5_0

That's the right tag for 5.0-stable source.

12