Patchwork [U-Boot] pull request for u-boot-tegra/master

login
register
mail settings
Submitter Allen Martin
Date Aug. 6, 2012, 10:46 p.m.
Message ID <20120806224622.GF7791@nvidia.com>
Download mbox | patch
Permalink /patch/175470/
State Superseded
Delegated to: Albert ARIBAUD
Headers show

Comments

Allen Martin - Aug. 6, 2012, 10:46 p.m.
> > -----Original Message-----
> > From: Albert ARIBAUD [mailto:albert.u.boot@aribaud.net]
> > Sent: Sunday, August 05, 2012 1:01 PM
> > To: Albert ARIBAUD
> > Cc: Tom Warren; Simon@theia.denx.de; u-boot@lists.denx.de; Stephen Warren;
> > Tom Warren
> > Subject: Re: [U-Boot] pull request for u-boot-tegra/master
> > 
> > Hi Albert,
> > 
> > On Sun, 5 Aug 2012 21:46:21 +0200, Albert ARIBAUD
> > <albert.u.boot@aribaud.net> wrote:
> > > Hi Tom,
> > >
> > > On Fri, 3 Aug 2012 16:35:40 -0700, Tom Warren <twarren.nvidia@gmail.com>
> > wrote:
> > > > Albert,
> > > >
> > > > Has this been pulled into ARM master? I don't think I saw a response
> > > > from you, but I might have missed it.
> > >
> > > Sorry Tom, but I have been unavailable for some time and especially on
> > > jul 30 and 31, so no, it has not been pulled in. :(
> > >
> > > I am pulling it in right now, but I have sparse access to Internet right
> > now.
> > 
> > Seems like f414d013c5e129e66c04ff699bd60b34894d3d38 breaks a lot of builds
> > with the stock Ubuntu linaro toolchain, e.g.
> > 
> > ----------------------------------------------------------------------------
> > ----------
> > 
> > $ ./MAKEALL jornada
> > Configuring for jornada board...
> > make: *** [jornada_config] Error 1
> > arm-linux-gnueabi-size: './u-boot': No such file
> > make: *** No rule to make target `include/autoconf.mk', needed by `depend'.
> > Stop.
> > make: *** Waiting for unfinished jobs....
> > 

I think I found the problem.  For some reason I don't fully understand
when a board doesn't have ${soc} defined the following code in mkconfig:

( echo "ARCH   = ${arch}"
    if [ ! -z "$spl_cpu" ] ; then
        echo 'ifeq ($(CONFIG_SPL_BUILD),y)'
        echo "CPU    = ${spl_cpu}"
        echo "else"
        echo "CPU    = ${cpu}"
        echo "endif"
    else
        echo "CPU    = ${cpu}"
    fi
    echo "BOARD  = ${board}"

    [ "${vendor}" ] && echo "VENDOR = ${vendor}"
    [ "${soc}"    ] && echo "SOC    = ${soc}" ) > config.mk

causes the subshell to return a nonzero exit status.  I think the
behavior may change depending on what shell you have /bin/sh pointing
to.  Can you try out the following fix to verify?  I'll put together a
proper patch in the mean time


-----------------------------------------------------------------------------------
This email message is for the sole use of the intended recipient(s) and may contain
confidential information.  Any unauthorized review, use, disclosure or distribution
is prohibited.  If you are not the intended recipient, please contact the sender by
reply email and destroy all copies of the original message.
-----------------------------------------------------------------------------------
Albert ARIBAUD - Aug. 7, 2012, 6:53 a.m.
Hi Allen,

On Mon, 6 Aug 2012 15:46:22 -0700, Allen Martin <amartin@nvidia.com> wrote:
> > > -----Original Message-----
> > > From: Albert ARIBAUD [mailto:albert.u.boot@aribaud.net]
> > > Sent: Sunday, August 05, 2012 1:01 PM
> > > To: Albert ARIBAUD
> > > Cc: Tom Warren; Simon@theia.denx.de; u-boot@lists.denx.de; Stephen Warren;
> > > Tom Warren
> > > Subject: Re: [U-Boot] pull request for u-boot-tegra/master
> > > 
> > > Hi Albert,
> > > 
> > > On Sun, 5 Aug 2012 21:46:21 +0200, Albert ARIBAUD
> > > <albert.u.boot@aribaud.net> wrote:
> > > > Hi Tom,
> > > >
> > > > On Fri, 3 Aug 2012 16:35:40 -0700, Tom Warren <twarren.nvidia@gmail.com>
> > > wrote:
> > > > > Albert,
> > > > >
> > > > > Has this been pulled into ARM master? I don't think I saw a response
> > > > > from you, but I might have missed it.
> > > >
> > > > Sorry Tom, but I have been unavailable for some time and especially on
> > > > jul 30 and 31, so no, it has not been pulled in. :(
> > > >
> > > > I am pulling it in right now, but I have sparse access to Internet right
> > > now.
> > > 
> > > Seems like f414d013c5e129e66c04ff699bd60b34894d3d38 breaks a lot of builds
> > > with the stock Ubuntu linaro toolchain, e.g.
> > > 
> > > ----------------------------------------------------------------------------
> > > ----------
> > > 
> > > $ ./MAKEALL jornada
> > > Configuring for jornada board...
> > > make: *** [jornada_config] Error 1
> > > arm-linux-gnueabi-size: './u-boot': No such file
> > > make: *** No rule to make target `include/autoconf.mk', needed by `depend'.
> > > Stop.
> > > make: *** Waiting for unfinished jobs....
> > > 
> 
> I think I found the problem.  For some reason I don't fully understand
> when a board doesn't have ${soc} defined the following code in mkconfig:
> 
> ( echo "ARCH   = ${arch}"
>     if [ ! -z "$spl_cpu" ] ; then
>         echo 'ifeq ($(CONFIG_SPL_BUILD),y)'
>         echo "CPU    = ${spl_cpu}"
>         echo "else"
>         echo "CPU    = ${cpu}"
>         echo "endif"
>     else
>         echo "CPU    = ${cpu}"
>     fi
>     echo "BOARD  = ${board}"
> 
>     [ "${vendor}" ] && echo "VENDOR = ${vendor}"
>     [ "${soc}"    ] && echo "SOC    = ${soc}" ) > config.mk
> 
> causes the subshell to return a nonzero exit status.  I think the
> behavior may change depending on what shell you have /bin/sh pointing
> to.  Can you try out the following fix to verify?  I'll put together a
> proper patch in the mean time
> 
> diff --git a/mkconfig b/mkconfig
> index 07031c6..d3363c6 100755
> --- a/mkconfig
> +++ b/mkconfig
> @@ -144,7 +144,8 @@ fi
>      echo "BOARD  = ${board}"
> 
>      [ "${vendor}" ] && echo "VENDOR = ${vendor}"
> -    [ "${soc}"    ] && echo "SOC    = ${soc}" ) > config.mk
> +    [ "${soc}"    ] && echo "SOC    = ${soc}"
> +    exit 0 ) > config.mk
> 
>  # Assign board directory to BOARDIR variable
>  if [ -z "${vendor}" ] ; then

Seems to work with this fix applied on top of the current tegra/master.

Tom, can you fix the branch so that the original commit (f414d013...) contains
the fix above, rather than just adding the fix on top of the branch? That'll
help any bisecting through this branch's commits.

Also, Allen:
 
> -----------------------------------------------------------------------------------
> This email message is for the sole use of the intended recipient(s) and may contain
> confidential information.  Any unauthorized review, use, disclosure or distribution
> is prohibited.  If you are not the intended recipient, please contact the sender by
> reply email and destroy all copies of the original message.
> -----------------------------------------------------------------------------------

Can this be removed from future e-mails to the list?

Amicalement,
Allen Martin - Aug. 7, 2012, 5:15 p.m.
On Mon, Aug 06, 2012 at 11:53:03PM -0700, Albert ARIBAUD wrote:
> Hi Allen,
> 
> Also, Allen:
>  
> > -----------------------------------------------------------------------------------
> > This email message is for the sole use of the intended recipient(s) and may contain
> > confidential information.  Any unauthorized review, use, disclosure or distribution
> > is prohibited.  If you are not the intended recipient, please contact the sender by
> > reply email and destroy all copies of the original message.
> > -----------------------------------------------------------------------------------
> 
> Can this be removed from future e-mails to the list?

My bad, the "nvpublic" in my .signature prevents our mail server from
adding the legal crap, but I must have accidentaly snipped it.

Thank you for finding the regression also, sorry I didn't find it
myself.

-Allen

Patch

diff --git a/mkconfig b/mkconfig
index 07031c6..d3363c6 100755
--- a/mkconfig
+++ b/mkconfig
@@ -144,7 +144,8 @@  fi
     echo "BOARD  = ${board}"

     [ "${vendor}" ] && echo "VENDOR = ${vendor}"
-    [ "${soc}"    ] && echo "SOC    = ${soc}" ) > config.mk
+    [ "${soc}"    ] && echo "SOC    = ${soc}"
+    exit 0 ) > config.mk

 # Assign board directory to BOARDIR variable
 if [ -z "${vendor}" ] ; then