After spending some time figuring out why my system failed to reboot,
checking config options, kernel version, modules installed. I figured
out that the patch in http://patchwork.kernel.org/patch/26483/ broke
the system boot on the latest mmotm. Reverting it helped my system
call depmod and hence boot.
The patch is really unhelpful, since the changelog is completely
empty. Sigh!
--
Balbir
On Wed, Jun 03, 2009 at 10:39:44PM +0800, Balbir Singh wrote:
> After spending some time figuring out why my system failed to reboot,
> checking config options, kernel version, modules installed. I figured
> out that the patch in http://patchwork.kernel.org/patch/26483/ broke
> the system boot on the latest mmotm. Reverting it helped my system
> call depmod and hence boot.
>
> The patch is really unhelpful, since the changelog is completely
> empty. Sigh!
Yeah, sorry for not nak'ing it earlier. Has been busy.
Andrew removed it from -mm and I notified the author about the
problems with the patch (in private mail it seems - I just replied to what I had).
Sam
* Sam Ravnborg <[email protected]> [2009-06-03 21:45:26]:
> On Wed, Jun 03, 2009 at 10:39:44PM +0800, Balbir Singh wrote:
> > After spending some time figuring out why my system failed to reboot,
> > checking config options, kernel version, modules installed. I figured
> > out that the patch in http://patchwork.kernel.org/patch/26483/ broke
> > the system boot on the latest mmotm. Reverting it helped my system
> > call depmod and hence boot.
> >
> > The patch is really unhelpful, since the changelog is completely
> > empty. Sigh!
>
> Yeah, sorry for not nak'ing it earlier. Has been busy.
> Andrew removed it from -mm and I notified the author about the
> problems with the patch (in private mail it seems - I just replied to what I had).
Cool! Thanks for the update. I'll get the latest mmotm and test as
well.
--
Balbir
Hi Balbir and all,
Sorry for the trouble and the missing changelog. I'm still learning;)
Can you please elaborate a bit more what exactly is broken?
Are you cross- or native compiling?
Here is what I tried to fix:
In my cross build environment (ELDK4.2) the following is defined:
export | grep DEPMOD
declare -x DEPMOD="/opt/eldk-4.2-ppc_4xx/usr/bin/depmod.pl"
without my patch it breaks like this:
make V=1 ARCH=powerpc CROSS_COMPILE=ppc_4xx-
INSTALL_MOD_PATH=$ELDK_PREFIX/eldk-4.2-ppc_4xx/ppc_4xx modules_install
Warning: you may need to install module-init-tools
See http://www.codemonkey.org.uk/docs/post-halloween-2.6.txt
make -f
/opt/eldk-4.2-ppc_4xx/usr/src/linux-v2.6.29.1/scripts/Makefile.modinst
mkdir -p
/opt/eldk-4.2-ppc_4xx/ppc_4xx/lib/modules/2.6.30-rc5-01317-g982dd86/kernel/drivers/scsi;
cp drivers/scsi/scsi_wait_scan.ko
/opt/eldk-4.2-ppc_4xx/ppc_4xx/lib/modules/2.6.30-rc5-01317-g982dd86/kernel/drivers/scsi
; true
/opt/eldk-4.2-ppc_4xx/ppc_4xx/lib/modules/2.6.30-rc5-01317-g982dd86/kernel/drivers/scsi/scsi_wait_scan.ko
make -f
/opt/eldk-4.2-ppc_4xx/usr/src/linux-v2.6.29.1/scripts/Makefile.fwinst
obj=firmware __fw_modinst
if [ -r System.map -a -x /opt/eldk-4.2-ppc_4xx/usr/bin/depmod.pl ]; then
/opt/eldk-4.2-ppc_4xx/usr/bin/depmod.pl -ae -F System.map -b
/opt/eldk-4.2-ppc_4xx/ppc_4xx 2.6.30-rc5-01317-g982dd86; fi
make: *** [_modinst_post] Error 1
with my patch it is more happy:
make V=1 ARCH=powerpc CROSS_COMPILE=ppc_4xx-
INSTALL_MOD_PATH=$ELDK_PREFIX/eldk-4.2-ppc_4xx/ppc_4xx modules_install
Warning: you may need to install module-init-tools
See http://www.codemonkey.org.uk/docs/post-halloween-2.6.txt
make -f
/opt/eldk-4.2-ppc_4xx/usr/src/linux-v2.6.29.1/scripts/Makefile.modinst
mkdir -p
/opt/eldk-4.2-ppc_4xx/ppc_4xx/lib/modules/2.6.30-rc5-01317-g982dd86/kernel/drivers/scsi;
cp drivers/scsi/scsi_wait_scan.ko
/opt/eldk-4.2-ppc_4xx/ppc_4xx/lib/modules/2.6.30-rc5-01317-g982dd86/kernel/drivers/scsi
; true
/opt/eldk-4.2-ppc_4xx/ppc_4xx/lib/modules/2.6.30-rc5-01317-g982dd86/kernel/drivers/scsi/scsi_wait_scan.ko
make -f
/opt/eldk-4.2-ppc_4xx/usr/src/linux-v2.6.29.1/scripts/Makefile.fwinst
obj=firmware __fw_modinst
I could most likely just undefine DEPMOD in the environment.
(here DEPMOD is not defined):
make v=1 ARCH=powerpc CROSS_COMPILE=ppc_4xx-
INSTALL_MOD_PATH=$ELDK_PREFIX/eldk-4.2-ppc_4xx/ppc_4xx modules_install
INSTALL drivers/scsi/scsi_wait_scan.ko
DEPMOD 2.6.30-rc5-01317-g982dd86
This seems to be working fine as well.
I did not try the native compilation. Maybe the problem is there?
Regards,
Robert
On Thu, 4 Jun 2009 14:31:44 +0800, Balbir Singh <[email protected]>
wrote:
> * Sam Ravnborg <[email protected]> [2009-06-03 21:45:26]:
>
>> On Wed, Jun 03, 2009 at 10:39:44PM +0800, Balbir Singh wrote:
>> > After spending some time figuring out why my system failed to reboot,
>> > checking config options, kernel version, modules installed. I figured
>> > out that the patch in http://patchwork.kernel.org/patch/26483/ broke
>> > the system boot on the latest mmotm. Reverting it helped my system
>> > call depmod and hence boot.
>> >
>> > The patch is really unhelpful, since the changelog is completely
>> > empty. Sigh!
>>
>> Yeah, sorry for not nak'ing it earlier. Has been busy.
>> Andrew removed it from -mm and I notified the author about the
>> problems with the patch (in private mail it seems - I just replied to
>> what I had).
>
> Cool! Thanks for the update. I'll get the latest mmotm and test as
> well.
* [email protected] <[email protected]> [2009-06-04 01:47:18]:
> Hi Balbir and all,
>
> Sorry for the trouble and the missing changelog. I'm still learning;)
>
> Can you please elaborate a bit more what exactly is broken?
> Are you cross- or native compiling?
>
I am doing a native compile.
--
Balbir
Hi Balbir,
Even when I do a native compile things look good to me.
make v=1 INSTALL_MOD_PATH=/tmp modules_install
INSTALL arch/x86/crypto/aes-i586.ko
INSTALL arch/x86/crypto/salsa20-i586.ko
INSTALL arch/x86/crypto/twofish-i586.ko
INSTALL arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.ko
... and so on until the end.
Is the depmod in your default path /sbin/depmod or something else?
Do you have the environment variable DEPMOD defined?
Do you see any errors when you do make modules_install with my patch?
Regards,
Robert
On Thu, 4 Jun 2009 15:32:40 +0800, Balbir Singh <[email protected]>
wrote:
> * [email protected] <[email protected]>
> [2009-06-04 01:47:18]:
>
>> Hi Balbir and all,
>>
>> Sorry for the trouble and the missing changelog. I'm still learning;)
>>
>> Can you please elaborate a bit more what exactly is broken?
>> Are you cross- or native compiling?
>>
>
> I am doing a native compile.
On Thu, 2009-06-04 at 14:31 +0800, Balbir Singh wrote:
> * Sam Ravnborg <[email protected]> [2009-06-03 21:45:26]:
>
> > On Wed, Jun 03, 2009 at 10:39:44PM +0800, Balbir Singh wrote:
> > > After spending some time figuring out why my system failed to reboot,
> > > checking config options, kernel version, modules installed. I figured
> > > out that the patch in http://patchwork.kernel.org/patch/26483/ broke
> > > the system boot on the latest mmotm. Reverting it helped my system
> > > call depmod and hence boot.
> > >
> > > The patch is really unhelpful, since the changelog is completely
> > > empty. Sigh!
> >
> > Yeah, sorry for not nak'ing it earlier. Has been busy.
> > Andrew removed it from -mm and I notified the author about the
> > problems with the patch (in private mail it seems - I just replied to what I had).
>
> Cool! Thanks for the update. I'll get the latest mmotm and test as
> well.
The general idea of using the cross-depmod is good though :) We have had
a few people recently run their newer host depmod and wonder why their
older embedded device can't read the new binary index files.
Jon.
On Fri, Jun 05, 2009 at 03:17:12AM -0400, Jon Masters wrote:
> On Thu, 2009-06-04 at 14:31 +0800, Balbir Singh wrote:
> > * Sam Ravnborg <[email protected]> [2009-06-03 21:45:26]:
> >
> > > On Wed, Jun 03, 2009 at 10:39:44PM +0800, Balbir Singh wrote:
> > > > After spending some time figuring out why my system failed to reboot,
> > > > checking config options, kernel version, modules installed. I figured
> > > > out that the patch in http://patchwork.kernel.org/patch/26483/ broke
> > > > the system boot on the latest mmotm. Reverting it helped my system
> > > > call depmod and hence boot.
> > > >
> > > > The patch is really unhelpful, since the changelog is completely
> > > > empty. Sigh!
> > >
> > > Yeah, sorry for not nak'ing it earlier. Has been busy.
> > > Andrew removed it from -mm and I notified the author about the
> > > problems with the patch (in private mail it seems - I just replied to what I had).
> >
> > Cool! Thanks for the update. I'll get the latest mmotm and test as
> > well.
>
> The general idea of using the cross-depmod is good though :) We have had
> a few people recently run their newer host depmod and wonder why their
> older embedded device can't read the new binary index files.
But then the patch would lok like this - no?
index 610d1c3..05c5744 100644
--- a/Makefile
+++ b/Makefile
@@ -325,7 +325,8 @@ OBJCOPY = $(CROSS_COMPILE)objcopy
OBJDUMP = $(CROSS_COMPILE)objdump
AWK = awk
GENKSYMS = scripts/genksyms/genksyms
-DEPMOD = /sbin/depmod
+DEPMOD = $(if $(shell which $(CROSS_COMPILE)depmod),
+ $(CROSS_COMPILE)depmod,/sbin/depmod)
KALLSYMS = scripts/kallsyms
PERL = perl
CHECK = sparse
Sam