All of lore.kernel.org
 help / color / mirror / Atom feed
* Re: S98configure role
@ 2010-10-14 14:45 l.bolognini
  0 siblings, 0 replies; 6+ messages in thread
From: l.bolognini @ 2010-10-14 14:45 UTC (permalink / raw)
  To: openembedded-devel

Thank you Chris et al. for your reply.
Even if I told you in my last 
mail I had no chanche to investigate further, I took some time to burn 
some SD and tried to understand the topic.
Well, in my build machine, 4 
packages remain unpacked:
openssh-sshd
net-snmp-server
udev
update-
modules

But, since openssh-sshd and update-modules are intentionally 
left unpacked (if test "x$D" != "x"; then exit 1 - in the .postinst 
script), it seems that for udev and net-snmp-server something goes 
wrong.
Looking in depth, I saw that in udev_141.bb, inside 
pkg_postinst_udev_append(), the usual test
if test "x$D" != "x"; then

        OPT="-r $D"
else
        OPT="-s"
fi
isn't written before 
update-rc.d command:
update-rc.d $OPT udev-cache start 12 S .
Simply, 
this piece of code is not available.
Result: update.rc.d fails in the 
build machine because OPT is empty.
Then we have not S03udev and 
S12udev-cache init links and S98configure can repair the situation but 
only from 2nd boot on.
To solve the problem I had to add that usual 
test inside pkg_postinst_udev_append() in udev_141.bb, then bitbake 
udev again.

Similar thing happens for net-snmp-server, where the 
postinst script has the same piece of code written twice, resulting in 
an error and then in the flag unpacked for the package.
Commenting out 
the first occurrance I got good results.

Here net-snmp configuration 
by means of S98configure is less dramatical because net-snmp agent 
starts after S98configure and then everything is fine also in 1st boot.


I hope this investigation could help someone involved in the same 
problems,
bye,
Luca

----Messaggio originale----
Da: clarson@kergoth.
com
Data: 14/10/2010 16.01
A: <openembedded-devel@lists.openembedded.
org>
Ogg: Re: [oe] S98configure role

On Thu, Oct 14, 2010 at 2:32 AM, 
l.bolognini@tiscali.it <
l.bolognini@tiscali.it> wrote:

> I'm 
wondering what is the role of this script, placed in /etc/rcS.d.
>
> 
What I guess is that, for some reasons, some packages get installed but

> not configured (then unpacked) since their postinstall script returns

> non zero value.
>

Yes, this is correct.  Some postinsts do things 
which can be done at image
creation time on the build machine, but not 
all.  Those which must *run* the
binaries in the filesystem, for 
example, cannot be run at build time, so
must be run at first boot time 
on the target.
-- 
Christopher Larson
clarson at kergoth dot com

Founder - BitBake, OpenEmbedded, OpenZaurus
Maintainer - Tslib
Senior 
Software Engineer, Mentor Graphics

_______________________________________________
Openembedded-devel 
mailing list
Openembedded-devel@lists.openembedded.org
http://lists.
linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel





Supera i limiti: raddoppia la velocità da 10 a 20 Mega!   Risparmia con Tutto Incluso: telefono + adsl 20 mega a soli 26,60 € al mese per un anno!  SCONTO DI 160 EURO!  http://abbonati.tiscali.it/telefono-adsl/prodotti/tc/tuttoincluso/?WT.mc_id=01fw



^ permalink raw reply	[flat|nested] 6+ messages in thread
* S98configure role
@ 2010-10-14  9:32 l.bolognini
  2010-10-14 14:01 ` Chris Larson
  0 siblings, 1 reply; 6+ messages in thread
From: l.bolognini @ 2010-10-14  9:32 UTC (permalink / raw)
  To: Open Embedded Mailing List

I'm wondering what is the role of this script, placed in /etc/rcS.d.

What I guess is that, for some reasons, some packages get installed but 
not configured (then unpacked) since their postinstall script returns 
non zero value.
Here it is a small extract of rootfs_ipk.bbclass:

for 
i in ${IMAGE_ROOTFS}${libdir}/opkg/info/*.postinst; do
                
if [ -f $i ] && ! sh -e $i configure; then
                        opkg-
cl ${IPKG_ARGS} flag unpacked `basename $i .postinst`
                
fi
done

For example, in my beagleboard, I get udev unpacked because 
udev-cache can't be found in /etc/init.d (sounds strange, it's there!).

Then udev remain unpacked but the magic S98configure heals udev, 
adjusting all stuff.
But, here is my question, what about the very 
first boot? Do I have a first boot without udev-cache?
Unfortunately at 
the moment I haven't the chance to investigate and burn a SD card and 
can't go further, however I'm very interested to listen to your 
opinions.

Thank you,
bye,
Luca



Supera i limiti: raddoppia la velocità da 10 a 20 Mega!   Risparmia con Tutto Incluso: telefono + adsl 20 mega a soli 26,60 € al mese per un anno!  SCONTO DI 160 EURO!  http://abbonati.tiscali.it/telefono-adsl/prodotti/tc/tuttoincluso/?WT.mc_id=01fw



^ permalink raw reply	[flat|nested] 6+ messages in thread
end of thread, other threads:[~2010-10-14 14:48 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2010-10-14 14:45 S98configure role l.bolognini
  -- strict thread matches above, loose matches on Subject: below --
2010-10-14  9:32 l.bolognini
2010-10-14 14:01 ` Chris Larson
2010-10-14 14:33   ` Hauser, Wolfgang (external)
2010-10-14 14:36     ` Chris Larson
2010-10-14 14:47       ` Hauser, Wolfgang (external)

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.