2.6.10-rc1 initramfs busted

2.6.10-rc1 initramfs busted

[Russell King]

A build using O= does this:

  HOSTCC  usr/gen_init_cpio
  GEN_INITRAMFS_LIST usr/initramfs_list
Using shipped usr/initramfs_list
  CPIO    usr/initramfs_data.cpio
ERROR: unable to open 'usr/initramfs_list': No such file or directory

Usage:
        ./usr/gen_init_cpio <cpio_list>

<cpio_list> is a file containing newline separated entries that
describe the files to be included in the initramfs archive:

The source tree contains this in usr:

-rw-r--r--  1 rmk rmk 1657 Oct 23 11:41 Makefile
-rw-r--r--  1 rmk rmk 8335 Oct 23 11:41 gen_init_cpio.c
-rw-r--r--  1 rmk rmk 1024 Aug  1  2003 initramfs_data.S
-rw-r--r--  1 rmk rmk  146 Oct 23 11:41 initramfs_list

and the build tree usr contains:

-rwxr-xr-x  1 rmk rmk 10834 Oct 23 13:29 gen_init_cpio
-rw-r--r--  1 rmk rmk     0 Oct 23 13:29 initramfs_data.cpio

Running with V=1 shows:

make -f /home/rmk/build/linux-v2.6-local/scripts/Makefile.build obj=usr
  echo Using shipped usr/initramfs_list
Using shipped usr/initramfs_list
  ./usr/gen_init_cpio usr/initramfs_list > usr/initramfs_data.cpio
ERROR: unable to open 'usr/initramfs_list': No such file or directory

so it's referencing the wrong directory.

-- 
Russell King
 Linux kernel    2.6 ARM Linux   - http://www.arm.linux.org.uk/
 maintainer of:  2.6 PCMCIA      - http://pcmcia.arm.linux.org.uk/
                 2.6 Serial core

Re: 2.6.10-rc1 initramfs busted [u]

[Martin Schlemmer [c]]

[-- Attachment #1: Type: text/plain, Size: 1657 bytes --]

On Sat, 2004-10-23 at 13:31 +0100, Russell King wrote:
> A build using O= does this:
> 
>   HOSTCC  usr/gen_init_cpio
>   GEN_INITRAMFS_LIST usr/initramfs_list
> Using shipped usr/initramfs_list
>   CPIO    usr/initramfs_data.cpio
> ERROR: unable to open 'usr/initramfs_list': No such file or directory
> 
> Usage:
>         ./usr/gen_init_cpio <cpio_list>
> 
> <cpio_list> is a file containing newline separated entries that
> describe the files to be included in the initramfs archive:
> 
> The source tree contains this in usr:
> 
> -rw-r--r--  1 rmk rmk 1657 Oct 23 11:41 Makefile
> -rw-r--r--  1 rmk rmk 8335 Oct 23 11:41 gen_init_cpio.c
> -rw-r--r--  1 rmk rmk 1024 Aug  1  2003 initramfs_data.S
> -rw-r--r--  1 rmk rmk  146 Oct 23 11:41 initramfs_list
> 
> and the build tree usr contains:
> 
> -rwxr-xr-x  1 rmk rmk 10834 Oct 23 13:29 gen_init_cpio
> -rw-r--r--  1 rmk rmk     0 Oct 23 13:29 initramfs_data.cpio
> 
> Running with V=1 shows:
> 
> make -f /home/rmk/build/linux-v2.6-local/scripts/Makefile.build obj=usr
>   echo Using shipped usr/initramfs_list
> Using shipped usr/initramfs_list
>   ./usr/gen_init_cpio usr/initramfs_list > usr/initramfs_data.cpio
> ERROR: unable to open 'usr/initramfs_list': No such file or directory
> 
> so it's referencing the wrong directory.
> 

Yep.  Please see the

 [PATCH 2.6.9-bk7] Select cpio_list or source directory for initramfs image updates

thread for latest patch to fix this and other issues.  There are some
other patches floating around that is slightly smaller, but I would
appreciate testing and feedback.


Cheers,

-- 
Martin Schlemmer


[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: 2.6.10-rc1 initramfs busted [u]

[Russell King]

On Sat, Oct 23, 2004 at 02:42:08PM +0200, Martin Schlemmer [c] wrote:
> On Sat, 2004-10-23 at 13:31 +0100, Russell King wrote:
> > A build using O= does this:
> > 
> >   HOSTCC  usr/gen_init_cpio
> >   GEN_INITRAMFS_LIST usr/initramfs_list
> > Using shipped usr/initramfs_list
> >   CPIO    usr/initramfs_data.cpio
> > ERROR: unable to open 'usr/initramfs_list': No such file or directory
> > 
> > Usage:
> >         ./usr/gen_init_cpio <cpio_list>
> > 
> > <cpio_list> is a file containing newline separated entries that
> > describe the files to be included in the initramfs archive:
> > 
> > The source tree contains this in usr:
> > 
> > -rw-r--r--  1 rmk rmk 1657 Oct 23 11:41 Makefile
> > -rw-r--r--  1 rmk rmk 8335 Oct 23 11:41 gen_init_cpio.c
> > -rw-r--r--  1 rmk rmk 1024 Aug  1  2003 initramfs_data.S
> > -rw-r--r--  1 rmk rmk  146 Oct 23 11:41 initramfs_list
> > 
> > and the build tree usr contains:
> > 
> > -rwxr-xr-x  1 rmk rmk 10834 Oct 23 13:29 gen_init_cpio
> > -rw-r--r--  1 rmk rmk     0 Oct 23 13:29 initramfs_data.cpio
> > 
> > Running with V=1 shows:
> > 
> > make -f /home/rmk/build/linux-v2.6-local/scripts/Makefile.build obj=usr
> >   echo Using shipped usr/initramfs_list
> > Using shipped usr/initramfs_list
> >   ./usr/gen_init_cpio usr/initramfs_list > usr/initramfs_data.cpio
> > ERROR: unable to open 'usr/initramfs_list': No such file or directory
> > 
> > so it's referencing the wrong directory.
> > 
> 
> Yep.  Please see the
> 
>  [PATCH 2.6.9-bk7] Select cpio_list or source directory for initramfs image updates
> 
> thread for latest patch to fix this and other issues.  There are some
> other patches floating around that is slightly smaller, but I would
> appreciate testing and feedback.

There's this obvious error in the patch (missing 'i'):

+$(obj)/$(initramfs_list): FORCE
+	$(if $(nitramfs_list_state_uptodate),, \

though I am wondering why this looks rather complicated.

-- 
Russell King
 Linux kernel    2.6 ARM Linux   - http://www.arm.linux.org.uk/
 maintainer of:  2.6 PCMCIA      - http://pcmcia.arm.linux.org.uk/
                 2.6 Serial core

Re: 2.6.10-rc1 initramfs busted [u]

[Martin Schlemmer [c]]

[-- Attachment #1: Type: text/plain, Size: 2503 bytes --]

On Sat, 2004-10-23 at 21:36 +0100, Russell King wrote:
> On Sat, Oct 23, 2004 at 02:42:08PM +0200, Martin Schlemmer [c] wrote:
> > On Sat, 2004-10-23 at 13:31 +0100, Russell King wrote:
> > > A build using O= does this:
> > > 
> > >   HOSTCC  usr/gen_init_cpio
> > >   GEN_INITRAMFS_LIST usr/initramfs_list
> > > Using shipped usr/initramfs_list
> > >   CPIO    usr/initramfs_data.cpio
> > > ERROR: unable to open 'usr/initramfs_list': No such file or directory
> > > 
> > > Usage:
> > >         ./usr/gen_init_cpio <cpio_list>
> > > 
> > > <cpio_list> is a file containing newline separated entries that
> > > describe the files to be included in the initramfs archive:
> > > 
> > > The source tree contains this in usr:
> > > 
> > > -rw-r--r--  1 rmk rmk 1657 Oct 23 11:41 Makefile
> > > -rw-r--r--  1 rmk rmk 8335 Oct 23 11:41 gen_init_cpio.c
> > > -rw-r--r--  1 rmk rmk 1024 Aug  1  2003 initramfs_data.S
> > > -rw-r--r--  1 rmk rmk  146 Oct 23 11:41 initramfs_list
> > > 
> > > and the build tree usr contains:
> > > 
> > > -rwxr-xr-x  1 rmk rmk 10834 Oct 23 13:29 gen_init_cpio
> > > -rw-r--r--  1 rmk rmk     0 Oct 23 13:29 initramfs_data.cpio
> > > 
> > > Running with V=1 shows:
> > > 
> > > make -f /home/rmk/build/linux-v2.6-local/scripts/Makefile.build obj=usr
> > >   echo Using shipped usr/initramfs_list
> > > Using shipped usr/initramfs_list
> > >   ./usr/gen_init_cpio usr/initramfs_list > usr/initramfs_data.cpio
> > > ERROR: unable to open 'usr/initramfs_list': No such file or directory
> > > 
> > > so it's referencing the wrong directory.
> > > 
> > 
> > Yep.  Please see the
> > 
> >  [PATCH 2.6.9-bk7] Select cpio_list or source directory for initramfs image updates
> > 
> > thread for latest patch to fix this and other issues.  There are some
> > other patches floating around that is slightly smaller, but I would
> > appreciate testing and feedback.
> 
> There's this obvious error in the patch (missing 'i'):
> 
> +$(obj)/$(initramfs_list): FORCE
> +	$(if $(nitramfs_list_state_uptodate),, \
> 

Ah, thanks!.  Should not matter though, as we do nothing with
uptodate currently.

> though I am wondering why this looks rather complicated.
> 

I guess I can drop the  uptodate $(if, as we only really need to check
outofdate and missing, but I was not sure if anybody wanted to print or
do something on uptodate ...

Or is it not the nested $(if's you are referring to?


thanks,

-- 
Martin Schlemmer


[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: 2.6.10-rc1 initramfs busted

[Matthias Urlichs]

Hi, Russell King wrote:

> A build using O= does this:

I already posted a small patch for this, three days ago.

-- 
Matthias Urlichs   |   {M:U} IT Design @ m-u-it.de   |  smurf@smurf.noris.de