Re: Building a new kernel

1999-08-04 Thread David Scheidt

On Wed, 4 Aug 1999, Warner Losh wrote:

 
 I have a freebsd-stable system.  I can't build a kernel for
 freebsd-current on that system unless I upgrade my compiler to egcs.
 Will this cause problems for our upgrade proceedure?
 
 gcc 2.7.2.3 doesn't like i386/include/atomic.h.  It complains about
 bad assmbler contraints.

I upgraded a -STABLE system to -CURRENT using source a month or two 
ago.  The first step is to build the new toolchain, so you shouldn't 
ever be compiling a new kernel with an old compiler.  

David



To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message



Re: Building a new kernel

1999-08-04 Thread Osokin Sergey




On Wed, 4 Aug 1999, Warner Losh wrote:

 
 I have a freebsd-stable system.  I can't build a kernel for
 freebsd-current on that system unless I upgrade my compiler to egcs.
 Will this cause problems for our upgrade proceedure?
 
 gcc 2.7.2.3 doesn't like i386/include/atomic.h.  It complains about
 bad assmbler contraints.
try to cvsup your source tree to 4.0, then rebuild your system
with simply make world procedure.

Rgdz,
Sergey Osokin aka oZZ,
[EMAIL PROTECTED]



To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message



Re: Building a new kernel

1999-08-04 Thread Warner Losh

In message [EMAIL PROTECTED] Osokin Sergey 
writes:
: try to cvsup your source tree to 4.0, then rebuild your system
: with simply make world procedure.

I can't do that.  This system *MUST* be a 3.2-stable system.  I was
building the kernel to test to see if a nasty NFS bug I've found in
-stable is present in the exact same environment but with a -current
kernel rather than a -stable one.

I was just noting this for people in the future.  It is something that
I've tradtionally been able to do and I couldn't do it in this case.
Others less worldly surely will hit the problem over time...  Likely
near the time we release 4.0 if the compiler situation remains the
same.

Warner


To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message



Re: Building a new kernel

1999-08-04 Thread Warner Losh

In message
[EMAIL PROTECTED] David
Scheidt writes: 
: Read the docs?  Who me?  It sounds like the 3.X to 4.0-RELEASE documentation
: should say not to do this.  Unless, of course, gcc-2.95 is imported before
: t hen.  

Give me a F*ing break.  No such documetation exists and the more that
we change in how things traditionally the more problem's we'll have.

gcc 2.95 is the same thing as egcs, so that wouldn't matter...

Warner


To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message



Re: Building a new kernel

1999-08-04 Thread David Scheidt
On Wed, 4 Aug 1999, Warner Losh wrote:

 
 I have a freebsd-stable system.  I can't build a kernel for
 freebsd-current on that system unless I upgrade my compiler to egcs.
 Will this cause problems for our upgrade proceedure?
 
 gcc 2.7.2.3 doesn't like i386/include/atomic.h.  It complains about
 bad assmbler contraints.

I upgraded a -STABLE system to -CURRENT using source a month or two 
ago.  The first step is to build the new toolchain, so you shouldn't 
ever be compiling a new kernel with an old compiler.  

David



To Unsubscribe: send mail to majord...@freebsd.org
with unsubscribe freebsd-hackers in the body of the message



Re: Building a new kernel

1999-08-04 Thread Osokin Sergey



On Wed, 4 Aug 1999, Warner Losh wrote:

 
 I have a freebsd-stable system.  I can't build a kernel for
 freebsd-current on that system unless I upgrade my compiler to egcs.
 Will this cause problems for our upgrade proceedure?
 
 gcc 2.7.2.3 doesn't like i386/include/atomic.h.  It complains about
 bad assmbler contraints.
try to cvsup your source tree to 4.0, then rebuild your system
with simply make world procedure.

Rgdz,
Sergey Osokin aka oZZ,
o...@etrust.ru



To Unsubscribe: send mail to majord...@freebsd.org
with unsubscribe freebsd-hackers in the body of the message



Re: Building a new kernel

1999-08-04 Thread Warner Losh
In message
pine.neb.3.96.990804145111.73456a-100...@shell-3.enteract.com David
Scheidt writes: 
: I upgraded a -STABLE system to -CURRENT using source a month or two 
: ago.  The first step is to build the new toolchain, so you shouldn't 
: ever be compiling a new kernel with an old compiler.  

In the past, we've given advise to build a new kernel, then reboot and
do a make upgrade or make world (depending on if you were branch
jumping or not).  Also, as part of the aout-to-elf target, a kernel is
built...

Warner


To Unsubscribe: send mail to majord...@freebsd.org
with unsubscribe freebsd-hackers in the body of the message



Re: Building a new kernel

1999-08-04 Thread Warner Losh
In message pine.bsf.4.10.9908042352560.1550-100...@ozz.etrust.ru Osokin 
Sergey writes:
: try to cvsup your source tree to 4.0, then rebuild your system
: with simply make world procedure.

I can't do that.  This system *MUST* be a 3.2-stable system.  I was
building the kernel to test to see if a nasty NFS bug I've found in
-stable is present in the exact same environment but with a -current
kernel rather than a -stable one.

I was just noting this for people in the future.  It is something that
I've tradtionally been able to do and I couldn't do it in this case.
Others less worldly surely will hit the problem over time...  Likely
near the time we release 4.0 if the compiler situation remains the
same.

Warner


To Unsubscribe: send mail to majord...@freebsd.org
with unsubscribe freebsd-hackers in the body of the message



Re: Building a new kernel

1999-08-04 Thread David Scheidt
On Wed, 4 Aug 1999, Warner Losh wrote:

 In message
 pine.neb.3.96.990804145111.73456a-100...@shell-3.enteract.com David
 Scheidt writes: 
 : I upgraded a -STABLE system to -CURRENT using source a month or two 
 : ago.  The first step is to build the new toolchain, so you shouldn't 
 : ever be compiling a new kernel with an old compiler.  
 
 In the past, we've given advise to build a new kernel, then reboot and
 do a make upgrade or make world (depending on if you were branch
 jumping or not).  Also, as part of the aout-to-elf target, a kernel is
 built...

Read the docs?  Who me?  It sounds like the 3.X to 4.0-RELEASE documentation
should say not to do this.  Unless, of course, gcc-2.95 is imported before
t hen.  

David



To Unsubscribe: send mail to majord...@freebsd.org
with unsubscribe freebsd-hackers in the body of the message



Re: Building a new kernel

1999-08-04 Thread Warner Losh
In message
pine.neb.3.96.990804161056.80097a-100...@shell-3.enteract.com David
Scheidt writes: 
: Read the docs?  Who me?  It sounds like the 3.X to 4.0-RELEASE documentation
: should say not to do this.  Unless, of course, gcc-2.95 is imported before
: t hen.  

Give me a F*ing break.  No such documetation exists and the more that
we change in how things traditionally the more problem's we'll have.

gcc 2.95 is the same thing as egcs, so that wouldn't matter...

Warner


To Unsubscribe: send mail to majord...@freebsd.org
with unsubscribe freebsd-hackers in the body of the message