astlinux-devel Mailing List for AstLinux (Page 2)

Philip,

Thanks for reviewing this...
Feel free to apply your clean-up skills to this commit.

1) I might suggest keeping the grep / sed lines as is, then a non- 
Philip can read and understand the code. :-)  Efficiency is not an  
issue here.

2) When this script is called by the altweb System tab, I redirect  
stderr to /dev/null, so it may be a good idea to enable stderr output  
from the script as you suggest, then terminal users would get a better  
error experience.

3) I kind of like keeping the `` in the tests, it is important I don't  
get any extra stdout output so my error strings are as expected.

4) Your last comment below...
> No color code?

Yes, no color code since that only applies to terminal shell use.

Thanks,
Lonnie

On Nov 10, 2008, at 2:27 PM, Philip Prindeville wrote:

> Some code review comments inline...
>
>
> dha...@us... wrote:
>> Revision: 2059
>>         http://astlinux.svn.sourceforge.net/astlinux/?rev=2059&view=rev
>> Author:   dhartman
>> Date:     2008-11-10 18:58:46 +0000 (Mon, 10 Nov 2008)
>>
>> Log Message:
>> -----------
>> add script to manage astlinux image upgrades from Lonnie's web  
>> interface
>>
>> Added Paths:
>> -----------
>>   trunk/target/generic/target_skeleton/usr/sbin/upgrade-run-image
>>
>> Added: trunk/target/generic/target_skeleton/usr/sbin/upgrade-run- 
>> image
>> ===================================================================
>> --- trunk/target/generic/target_skeleton/usr/sbin/upgrade-run- 
>> image	                        (rev 0)
>> +++ trunk/target/generic/target_skeleton/usr/sbin/upgrade-run-image	 
>> 2008-11-10 18:58:46 UTC (rev 2059)
>> @@ -0,0 +1,325 @@
>> +#!/bin/sh
>> +# red color error message for web interface
>> +RED=1
>> +# orange color error message for web interface
>> +ORANGE=2
>> +# green color message for web interface
>> +GREEN=0
>> +
>> +mount_oldroot_cdrom() {
>> +  if `mount | grep -q "/oldroot/cdrom"`; then
>> +    BASE=/oldroot/cdrom
>> +  else
>> +    if [ -x /sbin/findfs ]; then
>> +      LABEL=`/sbin/findfs LABEL=RUNNIX 2>/dev/null`
>> +      if [ -n "$LABEL" ]; then
>> +        mount -t vfat "$LABEL" /oldroot/cdrom
>> +      fi
>> +    fi
>> +    if `mount | grep -q "/oldroot/cdrom"`; then
>> +      BASE=/oldroot/cdrom
>> +    else
>> +      return 1
>> +    fi
>> +  fi
>> +  return 0
>> +}
>> +
>> +unmount_oldroot_cdrom() {
>> +  mount -o ro,remount /oldroot/cdrom >/dev/null 2>&1
>>
>
> On commands that you're not expecting to ever fail, like this one,  
> you really do want to see stderr when it does fail.
>
> So I'd recommend using:
>
> mount -o ro,remount /oldroot/cdrom >/dev/null
>
> and keep any messages to stderr on console.  Leaving the `` out  
> means "don't capture the output... just pay attention to the exit  
> value".
>
>
>> +}
>> +
>> +mount_tmpfs_up() {
>> +  mkdir /tmp/up >/dev/null 2>&1
>>
>
> Ditto.
>
>> +
>> +  if `mount -t tmpfs -o size=56m none /tmp/up >/dev/null 2>&1`; then
>>
>
> Ditto.
>
> Also, you can use $? to get the exit code, etc. i.e.
>
> if ! mount -t tmpfs -o size=56m none /tmp/up >/dev/null; then
>
> or:
>
> mount -t tmpfs -o size=56m none /tmp/up >/dev/null
> if [ $? -ne 0 ]; then
> rmdir /tmp/up
> return 1
> fi
>
>
>> +    :
>> +  else
>> +    rmdir /tmp/up
>> +    return 1
>> +  fi
>> +  return 0
>> +}
>> +
>> +unmount_tmpfs_up() {
>> +  cd /tmp
>> +  umount /tmp/up
>> +  rmdir /tmp/up
>> +}
>> +
>> +version_from_astimg() {
>> +  RUN_VER=""
>> +  for i in `cat /proc/cmdline`; do
>> +    if `echo $i | grep -q "astimg="`; then
>> +      RUN_VER=`echo $i | grep "astimg=" | sed -e 's/astimg=//' -e  
>> 's/.run$//'`
>>
>
> ... sed -e 's/astimg=\(.*\)\.run$/\1/'
>
> means find the string that starts with "astimg=" and ends with "\.run 
> $" and give me just the middle part.
>
> I don't think we need to grep for astimg= twice, by the way.  The  
> first one (in the "if") makes the second one in the RUN_VER=  
> superfluous.
>
>
>> +    fi
>> +  done
>> +}
>> +
>> +after_reboot() {
>> +  AFTER_REBOOT=""
>> +  if [ -n "$RUN_VER" ]; then
>> +    REBOOT_VER=$RUN_VER
>> +    if [ -f $BASE/os/ver ]; then
>> +      REBOOT_VER=`cat $BASE/os/ver`
>> +    fi
>> +    if [ "$REBOOT_VER" != "$RUN_VER" ]; then
>> +      AFTER_REBOOT=" [after reboot]"
>> +    fi
>> +  fi
>> +}
>> +
>> +check() {
>> +  for i in `cat /proc/cmdline`; do
>> +    if `echo $i | grep -q "astlinux="`; then
>> +      ARCH=`echo $i | grep "astlinux=" | sed -e 's/astlinux=//'`
>>
>
> ditto on the double-grep.
>
>
>> +    elif `echo $i | grep -q "console="`; then
>> +      SERIAL=`echo $i | grep "console=" | sed -e 's/console=//'`
>>
>
> ditto.
>
>
>> +    fi
>> +  done
>> +
>> +  if [ -z "$ARCH" ]; then
>> +    echo "Unknown run image architecture type."
>> +    exit $RED
>> +  fi
>> +
>> +  if [ "$ARCH" = "via" -o "$ARCH" = "viac7" ]; then
>> +    if [ -n "$SERIAL" ]; then
>> +      ARCH="${ARCH}-serial"
>> +    fi
>> +  fi
>> +
>> +  URL="${SRC_URL}/${ARCH}"
>> +
>> +  version_from_astimg
>> +
>> +  mount_oldroot_cdrom
>> +  if [ "$?" -ne 0 ]; then
>>
>
> you never need quotes around $? ... it will always be numeric and  
> non-empty.
>
>> +    echo "Unable to find Runnix partition."
>> +    exit $RED
>> +  fi
>> +
>> +  if `wget -q -c -O /tmp/up_ver $URL/ver >/dev/null 2>&1`; then
>>
>
> if wget -q -c -O /tmp/up_ver $URL/ver >/dev/null; then
>
>
>> +    VER=`cat /tmp/up_ver`
>> +    rm -f /tmp/up_ver
>> +  else
>> +    rm -f /tmp/up_ver
>> +    echo "No version available."
>> +    exit $RED
>> +  fi
>> +
>> +  if [ -f $BASE/os/ver ]; then
>> +    OVER=`cat $BASE/os/ver`
>> +  else
>> +    OVER=$RUN_VER
>> +  fi
>> +
>> +  if [ -z "$OVER" ]; then
>> +    echo "Unknown current version."
>> +    exit $RED
>> +  fi
>> +
>> +  if [ -f $BASE/os/Xver ]; then
>> +    XVER=`cat $BASE/os/Xver`
>> +  else
>> +    XVER=""
>> +  fi
>> +  +  after_reboot
>> +
>> +  if [ "$VER" = "$OVER" ]; then
>> +    echo "You are running the latest available version: ${VER}$ 
>> {AFTER_REBOOT}"
>> +    exit $ORANGE
>> +  fi
>> +}
>> +
>> +upgrade() {
>> +  if `mount -o rw,remount /oldroot/cdrom >/dev/null 2>&1`; then
>>
>
> as above...
>
>
>> +    :
>> +  else
>> +    echo "Unable to write to Runnix partition."
>> +    exit $RED
>> +  fi
>> +
>> +  mount_tmpfs_up
>> +  if [ "$?" -ne 0 ]; then
>> +    unmount_oldroot_cdrom
>> +    echo "Error creating temporary filesystem."
>> +    exit $RED
>> +  fi
>> +
>> +  if `wget -q -c -O /tmp/up/$VER.tar.gz $URL/$VER.tar.gz >/dev/ 
>> null 2>&1`; then
>>
>
> as above.  And use ! to skip the "then" portion and move the "else"  
> portion into the "then".
>
>> +    :
>> +  else
>> +    unmount_tmpfs_up
>> +    unmount_oldroot_cdrom
>> +    echo "Firmware download failed."
>> +    exit $RED
>> +  fi
>> +
>> +  if `wget -q -c -O /tmp/up/$VER.tar.gz.sha1 $URL/$VER.tar.gz.sha1  
>> >/dev/null 2>&1`; then
>>
>
> Ditto.
>
>> +    :
>> +  else
>> +    unmount_tmpfs_up
>> +    unmount_oldroot_cdrom
>> +    echo "Firmware sha1sum download failed."
>> +    exit $RED
>> +  fi
>> +
>> +  mkdir $BASE/tmp
>> +  cd /tmp/up
>> +
>> +  if `sha1sum -cs $VER.tar.gz.sha1`; then
>>
>
> Can we just use the exit value?
>
>
>> +    if [ -n "$XVER" ]; then
>> +      rm $BASE/os/$XVER.* >/dev/null 2>&1
>> +      rm $BASE/os/Xver >/dev/null 2>&1
>>
>
> If you're worried about these not existing, then use -f and don't  
> throw away the stderr...
>
> Most unix commands don't generate output on success anyway.
>
>
>> +    fi
>> +    tar -xzf /tmp/up/$VER.tar.gz -C $BASE/tmp/
>> +    sync
>> +  else
>> +    unmount_tmpfs_up
>> +    rmdir $BASE/tmp
>> +    unmount_oldroot_cdrom
>> +    echo "Firmware verification failed."
>> +    exit $RED
>> +  fi
>> +
>> +  cd $BASE/tmp/$VER
>> +
>> +  if `sha1sum -cs $VER.run.sha1`; then
>>
>
> Ditto.
>
>
>> +    mv $BASE/tmp/$VER/* $BASE/os/
>> +    echo "$OVER" > $BASE/os/Xver
>> +    echo "$VER" > $BASE/os/ver
>> +    rm -r $BASE/tmp/
>> +  else
>> +    unmount_tmpfs_up
>> +    rm -r $BASE/tmp/
>> +    unmount_oldroot_cdrom
>> +    echo "Firmware verification failed."
>> +    exit $RED
>> +  fi
>> +  +  after_reboot
>> +
>> +  unmount_tmpfs_up
>> +  unmount_oldroot_cdrom
>> +}
>> +
>> +show() {
>> +  version_from_astimg
>> +
>> +  mount_oldroot_cdrom
>> +  if [ "$?" -ne 0 ]; then
>>
>
> Ditto on the quotes.  You can also write this as:
>
> if ! mount_oldroot_cdrom; then
> ...
>
>
>> +    echo "Unable to find Runnix partition."
>> +    exit $RED
>> +  fi
>> +
>> +  if [ -f $BASE/os/ver ]; then
>> +    VER=`cat $BASE/os/ver`
>> +  else
>> +    VER=$RUN_VER
>> +  fi
>> +
>> +  if [ -z "$VER" ]; then
>> +    echo "Unknown current version."
>>
>
> Pipe to stderr via >&2 on non-zero exit.
>
>
>> +    exit $RED
>> +  fi
>> +
>> +  after_reboot
>> +
>> +  if [ -f $BASE/os/Xver ]; then
>> +    XVER=`cat $BASE/os/Xver`
>> +  else
>> +    echo "Current version is: ${VER}${AFTER_REBOOT}, no previous  
>> saved version."
>>
>
> Ditto.  Oh, wait... GREEN is 0...  never mind here.
>
>
>
>> +    exit $GREEN
>> +  fi
>> +}
>> +
>> +revert() {
>> +  version_from_astimg
>> +
>> +  mount_oldroot_cdrom
>> +  if [ "$?" -ne 0 ]; then
>>
>
> Ditto.
>
>
>> +    echo "Unable to find Runnix partition."
>>
>
> Ditto, and elsewhere.
>
>
>> +    exit $RED
>> +  fi
>> +
>> +  if [ -f $BASE/os/Xver ]; then
>> +    XVER=`cat $BASE/os/Xver`
>> +  else
>> +    echo "Revert failed, there is no previous saved version."
>> +    exit $RED
>> +  fi
>> +
>> +  if [ -f $BASE/os/ver ]; then
>> +    VER=`cat $BASE/os/ver`
>> +  else
>> +    echo "Unable to find the current version."
>> +    exit $RED
>> +  fi
>> +
>> +  if `mount -o rw,remount /oldroot/cdrom >/dev/null 2>&1`; then
>>
>
> As above.
>
>
>> +    echo $VER > $BASE/os/Xver
>> +    echo $XVER > $BASE/os/ver
>> +    mount -o ro,remount /oldroot/cdrom >/dev/null 2>&1
>>
>
> As above.
>
>
>> +  else
>> +    echo "Revert failed, unable to change version files."
>> +    exit $RED
>> +  fi
>> +  +  after_reboot
>> +}
>> +
>> +# main
>> +
>> +if [ -d /tmp/up ]; then
>> +  echo "Firmware Upgrade in Progress, please wait..."
>> +  exit $ORANGE
>> +fi
>> +
>> +case $1 in
>> +
>> +check)
>> +  if [ -n "$2" ]; then
>> +    SRC_URL=$2
>> +  else
>> +    echo "Usage: upgrade-run-image check firmware_repository_url"
>>
>
> as above.
>
>
>> +    exit 1
>> +  fi
>> +  check
>> +  echo "Current version is: ${OVER}${AFTER_REBOOT},  Latest  
>> available version is: ${VER}"
>> +  exit $GREEN
>> +  ;;
>> +  +upgrade)
>> +  if [ -n "$2" ]; then
>> +    SRC_URL=$2
>> +  else
>> +    echo "Usage: upgrade-run-image upgrade firmware_repository_url"
>> +    exit 1
>>
>
> No color code?
>
>
>> +  fi
>> +  check
>> +  upgrade
>> +  echo "Successful upgrade to: ${VER}${AFTER_REBOOT}"
>> +  exit $GREEN
>> +  ;;
>> +  +show)
>> +  show
>> +  echo "Current version is: ${VER}${AFTER_REBOOT}, Previous saved  
>> version is: ${XVER}"
>> +  exit $GREEN
>> +  ;;
>> +  +revert)
>> +  revert
>> +  echo "Current version is now: ${XVER}${AFTER_REBOOT}"
>> +  exit $GREEN
>> +  ;;
>> +  +*)
>> +  echo "Usage: upgrade-run-image check|upgrade|show|revert  
>> firmware_repository_url"
>> +  exit 1
>> +  ;;
>> +  +esac
>> +
>>
>
>

Some code review comments inline...


dha...@us... wrote:
> Revision: 2059
>           http://astlinux.svn.sourceforge.net/astlinux/?rev=2059&view=rev
> Author:   dhartman
> Date:     2008-11-10 18:58:46 +0000 (Mon, 10 Nov 2008)
>
> Log Message:
> -----------
> add script to manage astlinux image upgrades from Lonnie's web interface
>
> Added Paths:
> -----------
>     trunk/target/generic/target_skeleton/usr/sbin/upgrade-run-image
>
> Added: trunk/target/generic/target_skeleton/usr/sbin/upgrade-run-image
> ===================================================================
> --- trunk/target/generic/target_skeleton/usr/sbin/upgrade-run-image	                        (rev 0)
> +++ trunk/target/generic/target_skeleton/usr/sbin/upgrade-run-image	2008-11-10 18:58:46 UTC (rev 2059)
> @@ -0,0 +1,325 @@
> +#!/bin/sh
> +# red color error message for web interface
> +RED=1
> +# orange color error message for web interface
> +ORANGE=2
> +# green color message for web interface
> +GREEN=0
> +
> +mount_oldroot_cdrom() {
> +  if `mount | grep -q "/oldroot/cdrom"`; then
> +    BASE=/oldroot/cdrom
> +  else
> +    if [ -x /sbin/findfs ]; then
> +      LABEL=`/sbin/findfs LABEL=RUNNIX 2>/dev/null`
> +      if [ -n "$LABEL" ]; then
> +        mount -t vfat "$LABEL" /oldroot/cdrom
> +      fi
> +    fi
> +    if `mount | grep -q "/oldroot/cdrom"`; then
> +      BASE=/oldroot/cdrom
> +    else
> +      return 1
> +    fi
> +  fi
> +  return 0
> +}
> +
> +unmount_oldroot_cdrom() {
> +  mount -o ro,remount /oldroot/cdrom >/dev/null 2>&1
>   

On commands that you're not expecting to ever fail, like this one, you 
really do want to see stderr when it does fail.

So I'd recommend using:

mount -o ro,remount /oldroot/cdrom >/dev/null

and keep any messages to stderr on console.  Leaving the `` out means 
"don't capture the output... just pay attention to the exit value".


> +}
> +
> +mount_tmpfs_up() {
> +  mkdir /tmp/up >/dev/null 2>&1
>   

Ditto.

> +
> +  if `mount -t tmpfs -o size=56m none /tmp/up >/dev/null 2>&1`; then
>   

Ditto.

Also, you can use $? to get the exit code, etc. i.e.

if ! mount -t tmpfs -o size=56m none /tmp/up >/dev/null; then

or:

mount -t tmpfs -o size=56m none /tmp/up >/dev/null
if [ $? -ne 0 ]; then
  rmdir /tmp/up
  return 1
fi


> +    :
> +  else
> +    rmdir /tmp/up
> +    return 1
> +  fi
> +  return 0
> +}
> +
> +unmount_tmpfs_up() {
> +  cd /tmp
> +  umount /tmp/up
> +  rmdir /tmp/up
> +}
> +
> +version_from_astimg() {
> +  RUN_VER=""
> +  for i in `cat /proc/cmdline`; do
> +    if `echo $i | grep -q "astimg="`; then
> +      RUN_VER=`echo $i | grep "astimg=" | sed -e 's/astimg=//' -e 's/.run$//'`
>   

... sed -e 's/astimg=\(.*\)\.run$/\1/'

means find the string that starts with "astimg=" and ends with "\.run$" 
and give me just the middle part.

I don't think we need to grep for astimg= twice, by the way.  The first 
one (in the "if") makes the second one in the RUN_VER= superfluous.


> +    fi
> +  done
> +}
> +
> +after_reboot() {
> +  AFTER_REBOOT=""
> +  if [ -n "$RUN_VER" ]; then
> +    REBOOT_VER=$RUN_VER
> +    if [ -f $BASE/os/ver ]; then
> +      REBOOT_VER=`cat $BASE/os/ver`
> +    fi
> +    if [ "$REBOOT_VER" != "$RUN_VER" ]; then
> +      AFTER_REBOOT=" [after reboot]"
> +    fi
> +  fi
> +}
> +
> +check() {
> +  for i in `cat /proc/cmdline`; do
> +    if `echo $i | grep -q "astlinux="`; then
> +      ARCH=`echo $i | grep "astlinux=" | sed -e 's/astlinux=//'`
>   

ditto on the double-grep.


> +    elif `echo $i | grep -q "console="`; then
> +      SERIAL=`echo $i | grep "console=" | sed -e 's/console=//'`
>   

ditto.


> +    fi
> +  done
> +
> +  if [ -z "$ARCH" ]; then
> +    echo "Unknown run image architecture type."
> +    exit $RED
> +  fi
> +
> +  if [ "$ARCH" = "via" -o "$ARCH" = "viac7" ]; then
> +    if [ -n "$SERIAL" ]; then
> +      ARCH="${ARCH}-serial"
> +    fi
> +  fi
> +
> +  URL="${SRC_URL}/${ARCH}"
> +
> +  version_from_astimg
> +
> +  mount_oldroot_cdrom
> +  if [ "$?" -ne 0 ]; then
>   

you never need quotes around $? ... it will always be numeric and non-empty.

> +    echo "Unable to find Runnix partition."
> +    exit $RED
> +  fi
> +
> +  if `wget -q -c -O /tmp/up_ver $URL/ver >/dev/null 2>&1`; then
>   

if wget -q -c -O /tmp/up_ver $URL/ver >/dev/null; then


> +    VER=`cat /tmp/up_ver`
> +    rm -f /tmp/up_ver
> +  else
> +    rm -f /tmp/up_ver
> +    echo "No version available."
> +    exit $RED
> +  fi
> +
> +  if [ -f $BASE/os/ver ]; then
> +    OVER=`cat $BASE/os/ver`
> +  else
> +    OVER=$RUN_VER
> +  fi
> +
> +  if [ -z "$OVER" ]; then
> +    echo "Unknown current version."
> +    exit $RED
> +  fi
> +
> +  if [ -f $BASE/os/Xver ]; then
> +    XVER=`cat $BASE/os/Xver`
> +  else
> +    XVER=""
> +  fi
> +  
> +  after_reboot
> +
> +  if [ "$VER" = "$OVER" ]; then
> +    echo "You are running the latest available version: ${VER}${AFTER_REBOOT}"
> +    exit $ORANGE
> +  fi
> +}
> +
> +upgrade() {
> +  if `mount -o rw,remount /oldroot/cdrom >/dev/null 2>&1`; then
>   

as above...


> +    :
> +  else
> +    echo "Unable to write to Runnix partition."
> +    exit $RED
> +  fi
> +
> +  mount_tmpfs_up
> +  if [ "$?" -ne 0 ]; then
> +    unmount_oldroot_cdrom
> +    echo "Error creating temporary filesystem."
> +    exit $RED
> +  fi
> +
> +  if `wget -q -c -O /tmp/up/$VER.tar.gz $URL/$VER.tar.gz >/dev/null 2>&1`; then
>   

as above.  And use ! to skip the "then" portion and move the "else" 
portion into the "then".

> +    :
> +  else
> +    unmount_tmpfs_up
> +    unmount_oldroot_cdrom
> +    echo "Firmware download failed."
> +    exit $RED
> +  fi
> +
> +  if `wget -q -c -O /tmp/up/$VER.tar.gz.sha1 $URL/$VER.tar.gz.sha1 >/dev/null 2>&1`; then
>   

Ditto.

> +    :
> +  else
> +    unmount_tmpfs_up
> +    unmount_oldroot_cdrom
> +    echo "Firmware sha1sum download failed."
> +    exit $RED
> +  fi
> +
> +  mkdir $BASE/tmp
> +  cd /tmp/up
> +
> +  if `sha1sum -cs $VER.tar.gz.sha1`; then
>   

Can we just use the exit value?


> +    if [ -n "$XVER" ]; then
> +      rm $BASE/os/$XVER.* >/dev/null 2>&1
> +      rm $BASE/os/Xver >/dev/null 2>&1
>   

If you're worried about these not existing, then use -f and don't throw 
away the stderr...

Most unix commands don't generate output on success anyway.


> +    fi
> +    tar -xzf /tmp/up/$VER.tar.gz -C $BASE/tmp/
> +    sync
> +  else
> +    unmount_tmpfs_up
> +    rmdir $BASE/tmp
> +    unmount_oldroot_cdrom
> +    echo "Firmware verification failed."
> +    exit $RED
> +  fi
> +
> +  cd $BASE/tmp/$VER
> +
> +  if `sha1sum -cs $VER.run.sha1`; then
>   

Ditto.


> +    mv $BASE/tmp/$VER/* $BASE/os/
> +    echo "$OVER" > $BASE/os/Xver
> +    echo "$VER" > $BASE/os/ver
> +    rm -r $BASE/tmp/
> +  else
> +    unmount_tmpfs_up
> +    rm -r $BASE/tmp/
> +    unmount_oldroot_cdrom
> +    echo "Firmware verification failed."
> +    exit $RED
> +  fi
> +  
> +  after_reboot
> +
> +  unmount_tmpfs_up
> +  unmount_oldroot_cdrom
> +}
> +
> +show() {
> +  version_from_astimg
> +
> +  mount_oldroot_cdrom
> +  if [ "$?" -ne 0 ]; then
>   

Ditto on the quotes.  You can also write this as:

if ! mount_oldroot_cdrom; then
...


> +    echo "Unable to find Runnix partition."
> +    exit $RED
> +  fi
> +
> +  if [ -f $BASE/os/ver ]; then
> +    VER=`cat $BASE/os/ver`
> +  else
> +    VER=$RUN_VER
> +  fi
> +
> +  if [ -z "$VER" ]; then
> +    echo "Unknown current version."
>   

Pipe to stderr via >&2 on non-zero exit.


> +    exit $RED
> +  fi
> +
> +  after_reboot
> +
> +  if [ -f $BASE/os/Xver ]; then
> +    XVER=`cat $BASE/os/Xver`
> +  else
> +    echo "Current version is: ${VER}${AFTER_REBOOT}, no previous saved version."
>   

Ditto.  Oh, wait... GREEN is 0...  never mind here.



> +    exit $GREEN
> +  fi
> +}
> +
> +revert() {
> +  version_from_astimg
> +
> +  mount_oldroot_cdrom
> +  if [ "$?" -ne 0 ]; then
>   

Ditto.


> +    echo "Unable to find Runnix partition."
>   

Ditto, and elsewhere.


> +    exit $RED
> +  fi
> +
> +  if [ -f $BASE/os/Xver ]; then
> +    XVER=`cat $BASE/os/Xver`
> +  else
> +    echo "Revert failed, there is no previous saved version."
> +    exit $RED
> +  fi
> +
> +  if [ -f $BASE/os/ver ]; then
> +    VER=`cat $BASE/os/ver`
> +  else
> +    echo "Unable to find the current version."
> +    exit $RED
> +  fi
> +
> +  if `mount -o rw,remount /oldroot/cdrom >/dev/null 2>&1`; then
>   

As above.


> +    echo $VER > $BASE/os/Xver
> +    echo $XVER > $BASE/os/ver
> +    mount -o ro,remount /oldroot/cdrom >/dev/null 2>&1
>   

As above.


> +  else
> +    echo "Revert failed, unable to change version files."
> +    exit $RED
> +  fi
> +  
> +  after_reboot
> +}
> +
> +# main
> +
> +if [ -d /tmp/up ]; then
> +  echo "Firmware Upgrade in Progress, please wait..."
> +  exit $ORANGE
> +fi
> +
> +case $1 in
> +
> +check)
> +  if [ -n "$2" ]; then
> +    SRC_URL=$2
> +  else
> +    echo "Usage: upgrade-run-image check firmware_repository_url"
>   

as above.


> +    exit 1
> +  fi
> +  check
> +  echo "Current version is: ${OVER}${AFTER_REBOOT},  Latest available version is: ${VER}"
> +  exit $GREEN
>   
> +  ;;
> +  
> +upgrade)
> +  if [ -n "$2" ]; then
> +    SRC_URL=$2
> +  else
> +    echo "Usage: upgrade-run-image upgrade firmware_repository_url"
> +    exit 1
>   

No color code?


> +  fi
> +  check
> +  upgrade
> +  echo "Successful upgrade to: ${VER}${AFTER_REBOOT}"
> +  exit $GREEN
> +  ;;
> +  
> +show)
> +  show
> +  echo "Current version is: ${VER}${AFTER_REBOOT}, Previous saved version is: ${XVER}"
> +  exit $GREEN
> +  ;;
> +  
> +revert)
> +  revert
> +  echo "Current version is now: ${XVER}${AFTER_REBOOT}"
> +  exit $GREEN
> +  ;;
> +  
> +*)
> +  echo "Usage: upgrade-run-image check|upgrade|show|revert firmware_repository_url"
> +  exit 1
> +  ;;
> +  
> +esac
> +
>   

Darrick Hartman wrote:
> Kristian Kielhofner wrote:
>   
>> On Fri, Nov 7, 2008 at 4:54 PM, Darrick Hartman
>> <dha...@dj...> wrote:
>>     
>>> Kristian,
>>>
>>> There are so many changes that it's not practical to use the same kernel
>>> config that was previously in trunk and create a diff.  Your best bet is
>>>  to unpack the kernel then cd into the build_ix86/linux directory and
>>> run make menuconfig, then copy the resulting .config back to linux.config
>>>
>>> Philip chose the net5501 as the first target because it's something that
>>> we both can readily test with.
>>>
>>> Honestly, those config files are pretty easy to take care of.  We're
>>> stuck on an include issue with nistnet and rhino.  I think if we solve
>>> one, we solve both so that would be an awesome place to get another set
>>> of eyes.
>>>
>>> Darrick
>>>
>>>       
>> Darrick,
>>
>>   I'm talking about taking the old net5501 config (for 2.6.20) and the
>> new net5501 config (for 2.6.25) and creating a diff for the other
>> devices to include the new parameters in 2.6.25 (taken from net5501 -
>> the only completed target).  Someone (Philip) had to configure those
>> by hand with make oldconfig or something.  Whatever those more or less
>> generic 2.6.25 config options are, they have to be consistent across
>> devices.  For instance, you can't have a netfilter module build on one
>> device and not the other.  These are the kinds of changes I'm looking
>> for.
>>     
>
> Kristian,
>
> I understand what you're saying.
>
>   
>>   BTW, the best way to generate a new kernel config from an old config
>> is with "make oldconfig" - just like AstLinux.  They're both KConfig,
>> after all...
>>     
>
> Yes, and after doing a make oldconfig, going in and using make 
> menuconfig to select the correct options.
>
>   
>>   Also, as a matter of best practice we should start kernel upgrades
>> with the geni586 target device.  It's the most easy to move new
>> configuration options to other kernels.  It is, after all, the most
>> generic.
>>     
>
> I agree with this, however, Philip started with the net5501 because it 
> was easy for him to test.  I agree that it would have been easier to go 
> with the more general one first then move to the more specific.  We can 
> do that with the rest of them. (move to the geni586, then the via 
> targets...)  Do we want to remove the gumstix and 64bit targets?  They 
> are not maintained and may confuse people if they are left in place.
>
>   
>>   I don't think it is that easy - it's mostly a manual process to make
>> sure each new config is done properly.
>>     
>
> It is manual, but it is easier to do that than to find the include thing 
> that I've been beating my head over for the last few days.  Rhino and 
> nistnet won't build.  My preference was to find out what is not working 
> right with those two first before changing 50 bazillion other things.  I 
> can do the kernel config changes, but I'm not going to do that until the 
> kernel include stuff is figured out (which is what's preventing rhino 
> and zaptel from compiling properly).  Something changed between the 
> 2.6.20.x kernel and 2.6.25.x kernel with respect to the include files.
>
> Darrick
>   

I concur with Darrick:  until we have a reliably building image that we 
can run and verify correctness with (hopefully this won't take but a few 
more days), then it's not worth moving our changes into the other 
platforms because we don't know what we'd be propagating is damaged or not.

-Philip

Kristian Kielhofner wrote:
> On Fri, Nov 7, 2008 at 4:54 PM, Darrick Hartman
> <dha...@dj...> wrote:
>> Kristian,
>>
>> There are so many changes that it's not practical to use the same kernel
>> config that was previously in trunk and create a diff.  Your best bet is
>>  to unpack the kernel then cd into the build_ix86/linux directory and
>> run make menuconfig, then copy the resulting .config back to linux.config
>>
>> Philip chose the net5501 as the first target because it's something that
>> we both can readily test with.
>>
>> Honestly, those config files are pretty easy to take care of.  We're
>> stuck on an include issue with nistnet and rhino.  I think if we solve
>> one, we solve both so that would be an awesome place to get another set
>> of eyes.
>>
>> Darrick
>>
> 
> Darrick,
> 
>   I'm talking about taking the old net5501 config (for 2.6.20) and the
> new net5501 config (for 2.6.25) and creating a diff for the other
> devices to include the new parameters in 2.6.25 (taken from net5501 -
> the only completed target).  Someone (Philip) had to configure those
> by hand with make oldconfig or something.  Whatever those more or less
> generic 2.6.25 config options are, they have to be consistent across
> devices.  For instance, you can't have a netfilter module build on one
> device and not the other.  These are the kinds of changes I'm looking
> for.

Kristian,

I understand what you're saying.

>   BTW, the best way to generate a new kernel config from an old config
> is with "make oldconfig" - just like AstLinux.  They're both KConfig,
> after all...

Yes, and after doing a make oldconfig, going in and using make 
menuconfig to select the correct options.

>   Also, as a matter of best practice we should start kernel upgrades
> with the geni586 target device.  It's the most easy to move new
> configuration options to other kernels.  It is, after all, the most
> generic.

I agree with this, however, Philip started with the net5501 because it 
was easy for him to test.  I agree that it would have been easier to go 
with the more general one first then move to the more specific.  We can 
do that with the rest of them. (move to the geni586, then the via 
targets...)  Do we want to remove the gumstix and 64bit targets?  They 
are not maintained and may confuse people if they are left in place.

>   I don't think it is that easy - it's mostly a manual process to make
> sure each new config is done properly.

It is manual, but it is easier to do that than to find the include thing 
that I've been beating my head over for the last few days.  Rhino and 
nistnet won't build.  My preference was to find out what is not working 
right with those two first before changing 50 bazillion other things.  I 
can do the kernel config changes, but I'm not going to do that until the 
kernel include stuff is figured out (which is what's preventing rhino 
and zaptel from compiling properly).  Something changed between the 
2.6.20.x kernel and 2.6.25.x kernel with respect to the include files.

Darrick

Awesome!  That's the sort of cleanup I love to see.

-Philip


kr...@us... wrote:
> Revision: 2053
>           http://astlinux.svn.sourceforge.net/astlinux/?rev=2053&view=rev
> Author:   krisk84
> Date:     2008-11-07 21:33:46 +0000 (Fri, 07 Nov 2008)
>
> Log Message:
> -----------
> large update to move padlock support and init of the engine into openssl itself - not each application
>
> Modified Paths:
> --------------
>     trunk/package/minihttpd/minihttpd.mk
>     trunk/package/msmtp/msmtp.mk
>     trunk/package/openssh/openssh.mk
>     trunk/package/openssl/padlock.patch
>     trunk/package/openvpn/openvpn.mk
>
> Removed Paths:
> -------------
>     trunk/package/minihttpd/padlock.patch
>     trunk/package/msmtp/padlock.patch
>     trunk/package/openssh/padlock.patch
>     trunk/package/openvpn/padlock.patch
> [snip]
>   

That's right.

We've got a fair amount of churn on some files (like the OCF patches and 
the net5501/linux.config file) and what that to stabilize before we 
figure out what really got changed.

Not that many people are using trunk, and I think those that are are 
mostly using net5501's...

I think the sanest way to fix the configs is the following:

* pick up the changes to linux.mk for the net5501 and munge them into 
the other platform;
* "make linux-source" for that platform (make sure to tweak your root 
.config file's BR2_TARGET_xxx, of course);
* run "make linux-config" and set the additional options by hand 
(haven't figure out how to loop back and turn off options we shouldn't 
have or that don't apply... but these are minimal);
* copy the resulting build_i586/linux-*/.config into 
target/device/XXXX/linux.config
* commit

Darrick:  is that what you did?

-Philip


Kristian Kielhofner wrote:
> So...  It appears that the generic, apply to any kernel patches in
> target/device/kernel-patches have been updated but the Makefiles and
> the configs have not been updated for any target device *except*
> net5501.
>
> I'd love to update the configs for the other devices but I don't know
> what we've done with the new kernel options.  I suppose I could look
> at a diff and try to strip out net5501 specific options but does
> anyone have a baseline for the config changes made in the net5501
> kernel config?
>
>   

Kristian Kielhofner wrote:
> On Fri, Nov 7, 2008 at 4:27 PM, Philip Prindeville
> <phi...@re...> wrote:
>   
>> Thanks, Chris.  BTW:  can we get the version number out of the rules?
>>
>> It should appear in OPENSSL_VER and nowhere else.
>>
>> You can get:
>>
>> OPENSSL_VER_NUM:=$(shell echo $(OPENSSL_VER) | tr -d '[:lower:]')
>>
>> $ make var.OPENSSL_VER_NUM
>> OPENSSL_VER_NUM="0.9.8"
>> $
>>
>> Thanks.
>>
>>     
>
> I know...  We really should.
>
> On another note, I can't get 0.9.8 to build over here.  I just made
> another commit to update VIA padlock support for 0.9.8.  One of the
> nice things about 0.9.8 is the ability to initialize the padlock
> engine with the SSL engine.  Any app that uses openssl can benefit
> from padlock without any specific patches.
>
> I committed what I think *should* work.  However, I can't really build
> 0.9.8 anyway (with or without my padlock patch).  It builds properly
> but fails in install.
>
>   

Is there a reason that PADLOCK_NO_RNG isn't passed in via MAKE as 
$CFLAGS or whatever?

-Philip

On Fri, Nov 7, 2008 at 4:54 PM, Darrick Hartman
<dha...@dj...> wrote:
> Kristian,
>
> There are so many changes that it's not practical to use the same kernel
> config that was previously in trunk and create a diff.  Your best bet is
>  to unpack the kernel then cd into the build_ix86/linux directory and
> run make menuconfig, then copy the resulting .config back to linux.config
>
> Philip chose the net5501 as the first target because it's something that
> we both can readily test with.
>
> Honestly, those config files are pretty easy to take care of.  We're
> stuck on an include issue with nistnet and rhino.  I think if we solve
> one, we solve both so that would be an awesome place to get another set
> of eyes.
>
> Darrick
>

Darrick,

  I'm talking about taking the old net5501 config (for 2.6.20) and the
new net5501 config (for 2.6.25) and creating a diff for the other
devices to include the new parameters in 2.6.25 (taken from net5501 -
the only completed target).  Someone (Philip) had to configure those
by hand with make oldconfig or something.  Whatever those more or less
generic 2.6.25 config options are, they have to be consistent across
devices.  For instance, you can't have a netfilter module build on one
device and not the other.  These are the kinds of changes I'm looking
for.

  BTW, the best way to generate a new kernel config from an old config
is with "make oldconfig" - just like AstLinux.  They're both KConfig,
after all...

  Also, as a matter of best practice we should start kernel upgrades
with the geni586 target device.  It's the most easy to move new
configuration options to other kernels.  It is, after all, the most
generic.

  I don't think it is that easy - it's mostly a manual process to make
sure each new config is done properly.

-- 
Kristian Kielhofner
http://blog.krisk.org
http://www.submityoursip.com
http://www.astlinux.org
http://www.star2star.com

Kristian,

There are so many changes that it's not practical to use the same kernel 
config that was previously in trunk and create a diff.  Your best bet is 
  to unpack the kernel then cd into the build_ix86/linux directory and 
run make menuconfig, then copy the resulting .config back to linux.config

Philip chose the net5501 as the first target because it's something that 
we both can readily test with.

Honestly, those config files are pretty easy to take care of.  We're 
stuck on an include issue with nistnet and rhino.  I think if we solve 
one, we solve both so that would be an awesome place to get another set 
of eyes.

Darrick

Kristian Kielhofner wrote:
> So...  It appears that the generic, apply to any kernel patches in
> target/device/kernel-patches have been updated but the Makefiles and
> the configs have not been updated for any target device *except*
> net5501.
> 
> I'd love to update the configs for the other devices but I don't know
> what we've done with the new kernel options.  I suppose I could look
> at a diff and try to strip out net5501 specific options but does
> anyone have a baseline for the config changes made in the net5501
> kernel config?
> 

On Fri, Nov 7, 2008 at 4:27 PM, Philip Prindeville
<phi...@re...> wrote:
> Thanks, Chris.  BTW:  can we get the version number out of the rules?
>
> It should appear in OPENSSL_VER and nowhere else.
>
> You can get:
>
> OPENSSL_VER_NUM:=$(shell echo $(OPENSSL_VER) | tr -d '[:lower:]')
>
> $ make var.OPENSSL_VER_NUM
> OPENSSL_VER_NUM="0.9.8"
> $
>
> Thanks.
>

I know...  We really should.

On another note, I can't get 0.9.8 to build over here.  I just made
another commit to update VIA padlock support for 0.9.8.  One of the
nice things about 0.9.8 is the ability to initialize the padlock
engine with the SSL engine.  Any app that uses openssl can benefit
from padlock without any specific patches.

I committed what I think *should* work.  However, I can't really build
0.9.8 anyway (with or without my padlock patch).  It builds properly
but fails in install.

-- 
Kristian Kielhofner
http://blog.krisk.org
http://www.submityoursip.com
http://www.astlinux.org
http://www.star2star.com

So...  It appears that the generic, apply to any kernel patches in
target/device/kernel-patches have been updated but the Makefiles and
the configs have not been updated for any target device *except*
net5501.

I'd love to update the configs for the other devices but I don't know
what we've done with the new kernel options.  I suppose I could look
at a diff and try to strip out net5501 specific options but does
anyone have a baseline for the config changes made in the net5501
kernel config?

-- 
Kristian Kielhofner
http://blog.krisk.org
http://www.submityoursip.com
http://www.astlinux.org
http://www.star2star.com

Thanks, Chris.  BTW:  can we get the version number out of the rules?

It should appear in OPENSSL_VER and nowhere else.

You can get:

OPENSSL_VER_NUM:=$(shell echo $(OPENSSL_VER) | tr -d '[:lower:]')

$ make var.OPENSSL_VER_NUM
OPENSSL_VER_NUM="0.9.8"
$

Thanks.


kr...@us... wrote:
> Revision: 2050
>           http://astlinux.svn.sourceforge.net/astlinux/?rev=2050&view=rev
> Author:   krisk84
> Date:     2008-11-07 20:52:36 +0000 (Fri, 07 Nov 2008)
>
> Log Message:
> -----------
> update for 0.9.8
>
> Modified Paths:
> --------------
>     trunk/package/openssl/openssl.mk
>
> Modified: trunk/package/openssl/openssl.mk
> ===================================================================
> --- trunk/package/openssl/openssl.mk	2008-11-07 20:27:37 UTC (rev 2049)
> +++ trunk/package/openssl/openssl.mk	2008-11-07 20:52:36 UTC (rev 2050)
> @@ -67,21 +67,21 @@
>  $(STAGING_DIR)/lib/libcrypto.a: $(OPENSSL_DIR)/apps/openssl
>  	$(MAKE) CC=$(TARGET_CC) INSTALL_PREFIX=$(STAGING_DIR) -C $(OPENSSL_DIR) all install_sw
>  	cp -fa $(OPENSSL_DIR)/libcrypto.so* $(STAGING_DIR)/lib/
> -	chmod a-x $(STAGING_DIR)/lib/libcrypto.so.0.9.7
> -	(cd $(STAGING_DIR)/lib; ln -fs libcrypto.so.0.9.7 libcrypto.so)
> -	(cd $(STAGING_DIR)/lib; ln -fs libcrypto.so.0.9.7 libcrypto.so.0)
> +	chmod a-x $(STAGING_DIR)/lib/libcrypto.so.0.9.8
> +	(cd $(STAGING_DIR)/lib; ln -fs libcrypto.so.0.9.8 libcrypto.so)
> +	(cd $(STAGING_DIR)/lib; ln -fs libcrypto.so.0.9.8 libcrypto.so.0)
>  	cp -fa $(OPENSSL_DIR)/libssl.so* $(STAGING_DIR)/lib/
> -	chmod a-x $(STAGING_DIR)/lib/libssl.so.0.9.7
> -	(cd $(STAGING_DIR)/lib; ln -fs libssl.so.0.9.7 libssl.so)
> -	(cd $(STAGING_DIR)/lib; ln -fs libssl.so.0.9.7 libssl.so.0)
> +	chmod a-x $(STAGING_DIR)/lib/libssl.so.0.9.8
> +	(cd $(STAGING_DIR)/lib; ln -fs libssl.so.0.9.8 libssl.so)
> +	(cd $(STAGING_DIR)/lib; ln -fs libssl.so.0.9.8 libssl.so.0)
>  
> -$(TARGET_DIR)/usr/lib/libcrypto.so.0.9.7: $(STAGING_DIR)/lib/libcrypto.a
> +$(TARGET_DIR)/usr/lib/libcrypto.so.0.9.8: $(STAGING_DIR)/lib/libcrypto.a
>  	mkdir -p $(TARGET_DIR)/usr/lib
>  	cp -fa $(STAGING_DIR)/lib/libcrypto.so* $(TARGET_DIR)/usr/lib/
>  	cp -fa $(STAGING_DIR)/lib/libssl.so* $(TARGET_DIR)/usr/lib/
>  	#cp -fa $(STAGING_DIR)/bin/openssl  $(TARGET_DIR)/bin/
> -	-$(STRIP) $(TARGET_DIR)/usr/lib/libssl.so.0.9.7
> -	-$(STRIP) $(TARGET_DIR)/usr/lib/libcrypto.so.0.9.7
> +	-$(STRIP) $(TARGET_DIR)/usr/lib/libssl.so.0.9.8
> +	-$(STRIP) $(TARGET_DIR)/usr/lib/libcrypto.so.0.9.8
>  
>  $(TARGET_DIR)/usr/lib/libssl.a: $(STAGING_DIR)/lib/libcrypto.a
>  	mkdir -p $(TARGET_DIR)/usr/include 
> @@ -92,7 +92,7 @@
>  
>  openssl-headers: $(TARGET_DIR)/usr/lib/libssl.a
>  
> -openssl: uclibc $(TARGET_DIR)/usr/lib/libcrypto.so.0.9.7
> +openssl: uclibc $(TARGET_DIR)/usr/lib/libcrypto.so.0.9.8
>  
>  openssl-source: $(OPENSSL_DIR)/.unpacked
>  
>   

Heads up...  That last commit of openssl is known to not compile.

I stared at it for 2 hours last night and couldn't tell why.  If there's 
a meaningful message, it's lost in the noise.

If you need to build trunk to test other things, do the following:

(cd package ; svn update -r 2048 openssl)

This should restore openssl to the previous version.  You can sync it 
back to the latest later.

-Philip


Darrick Hartman wrote:
> Won't compile if enabled.
>
> Philip Prindeville wrote:
>   
>> Good lord, I'm babbling...  "Won't build, or won't run"???
>>
>>
>> Philip Prindeville wrote:
>>     
>>> Broken how?  Won't build, or won't compile?
>>>
>>> What new net5501 capabilities have been unlocked?
>>>
>>> -Philip
>>>
>>>
>>> Darrick Hartman wrote:
>>>   
>>>       
>>>> OCF is still broken with the latest patches that Philip applied.
>>>>
>>>> mISDN now builds.
>>>>
>>>> I updated the kernel config to properly reflect the hardware options 
>>>> with the net5501 device.
>>>>
>>>> rhino and nistnet also won't build (related to the kernel include files).
>>>>
>>>> Progress is being made though.
>>>>
>>>> Darrick
>>>>   
>>>>     
>>>>         
>> -------------------------------------------------------------------------
>> This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
>> Build the coolest Linux based applications with Moblin SDK & win great prizes
>> Grand prize is a trip for two to an Open Source event anywhere in the world
>> http://moblin-contest.org/redirect.php?banner_id=100&url=/
>> _______________________________________________
>> Astlinux-devel mailing list
>> Ast...@li...
>> https://lists.sourceforge.net/lists/listinfo/astlinux-devel
>>
>> Donations to support AstLinux are graciously accepted via PayPal to pa...@kr....
>>     
>
>
> -------------------------------------------------------------------------
> This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
> Build the coolest Linux based applications with Moblin SDK & win great prizes
> Grand prize is a trip for two to an Open Source event anywhere in the world
> http://moblin-contest.org/redirect.php?banner_id=100&url=/
> _______________________________________________
> Astlinux-devel mailing list
> Ast...@li...
> https://lists.sourceforge.net/lists/listinfo/astlinux-devel
>
> Donations to support AstLinux are graciously accepted via PayPal to pa...@kr....
>   

Won't compile if enabled.

Philip Prindeville wrote:
> Good lord, I'm babbling...  "Won't build, or won't run"???
> 
> 
> Philip Prindeville wrote:
>> Broken how?  Won't build, or won't compile?
>>
>> What new net5501 capabilities have been unlocked?
>>
>> -Philip
>>
>>
>> Darrick Hartman wrote:
>>   
>>> OCF is still broken with the latest patches that Philip applied.
>>>
>>> mISDN now builds.
>>>
>>> I updated the kernel config to properly reflect the hardware options 
>>> with the net5501 device.
>>>
>>> rhino and nistnet also won't build (related to the kernel include files).
>>>
>>> Progress is being made though.
>>>
>>> Darrick
>>>   
>>>     
> 
> 
> -------------------------------------------------------------------------
> This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
> Build the coolest Linux based applications with Moblin SDK & win great prizes
> Grand prize is a trip for two to an Open Source event anywhere in the world
> http://moblin-contest.org/redirect.php?banner_id=100&url=/
> _______________________________________________
> Astlinux-devel mailing list
> Ast...@li...
> https://lists.sourceforge.net/lists/listinfo/astlinux-devel
> 
> Donations to support AstLinux are graciously accepted via PayPal to pa...@kr....

Good lord, I'm babbling...  "Won't build, or won't run"???


Philip Prindeville wrote:
> Broken how?  Won't build, or won't compile?
>
> What new net5501 capabilities have been unlocked?
>
> -Philip
>
>
> Darrick Hartman wrote:
>   
>> OCF is still broken with the latest patches that Philip applied.
>>
>> mISDN now builds.
>>
>> I updated the kernel config to properly reflect the hardware options 
>> with the net5501 device.
>>
>> rhino and nistnet also won't build (related to the kernel include files).
>>
>> Progress is being made though.
>>
>> Darrick
>>   
>>     

Broken how?  Won't build, or won't compile?

What new net5501 capabilities have been unlocked?

-Philip


Darrick Hartman wrote:
> OCF is still broken with the latest patches that Philip applied.
>
> mISDN now builds.
>
> I updated the kernel config to properly reflect the hardware options 
> with the net5501 device.
>
> rhino and nistnet also won't build (related to the kernel include files).
>
> Progress is being made though.
>
> Darrick
>   

OCF is still broken with the latest patches that Philip applied.

mISDN now builds.

I updated the kernel config to properly reflect the hardware options 
with the net5501 device.

rhino and nistnet also won't build (related to the kernel include files).

Progress is being made though.

Darrick

I suspect it's your environment.  That doesn't happen to me.

What version of autoconf are you running?   I'm using FC8 with autoconf 
2.62.


Ingmar Schraub wrote:
> package/flex/flex.mk
>
> When you run ./configure --help in build_ixxx/flex-2.5.4, you see the
> following output below.
>
> flex.mk specifies however:
>
>                 --bindir=/usr/bin \
>                 --sbindir=/usr/sbin \
>                 --libexecdir=/usr/lib \
>                 --sysconfdir=/etc \
>                 --datadir=/usr/share \
>                 --localstatedir=/var \
>                 --mandir=/usr/man \
>                 --infodir=/usr/info \
>
> My build process stopped because these are un-recognized options. Is is
> my build environment? Haven't had this problem in the past.
>
> ingmar@nexoc:~/dev/astlinux-newtrunk/build_i586/flex-2.5.4> ./configure
> --help
> Usage: configure [options] [host]
> Options: [defaults in brackets after descriptions]
> Configuration:
>   --cache-file=FILE       cache test results in FILE
>   --help                  print this message
>   --no-create             do not create output files
>   --quiet, --silent       do not print `checking...' messages
>   --version               print the version of autoconf that created
> configure
> Directory and file names:
>   --prefix=PREFIX         install architecture-independent files in PREFIX
>                           [/usr/local]
>   --exec-prefix=PREFIX    install architecture-dependent files in PREFIX
>                           [same as prefix]
>   --srcdir=DIR            find the sources in DIR [configure dir or ..]
>   --program-prefix=PREFIX prepend PREFIX to installed program names
>   --program-suffix=SUFFIX append SUFFIX to installed program names
>   --program-transform-name=PROGRAM run sed PROGRAM on installed program
> names
> Host type:
>   --build=BUILD           configure for building on BUILD [BUILD=HOST]
>   --host=HOST             configure for HOST [guessed]
>   --target=TARGET         configure for TARGET [TARGET=HOST]
> Features and packages:
>   --disable-FEATURE       do not include FEATURE (same as
> --enable-FEATURE=no)
>   --enable-FEATURE[=ARG]  include FEATURE [ARG=yes]
>   --with-PACKAGE[=ARG]    use PACKAGE [ARG=yes]
>   --without-PACKAGE       do not use PACKAGE (same as --with-PACKAGE=no)
>   --x-includes=DIR        X include files are in DIR
>   --x-libraries=DIR       X library files are in DIR
> --enable and --with options recognized:
>
>
> Philip Prindeville wrote:
>   
>> I built everything from scratch.
>>
>> Which flex are you talking about?
>>
>> At least one package has it's own copy of flex...
>>
>> -Philip
>>
>> Ingmar Schraub wrote:
>>     
>>> Hello Philip,
>>>
>>> did you build recently "flex" from scratch? Apparently flex 2.5.4 has
>>> less 'configure' options..  Needs a fix.
>>>
>>> regards,
>>>
>>> Ingmar
>>>
>>> Philip Prindeville wrote:
>>>   
>>>       
>>>> Following the movement to Linux kernel 2.6.25, I've had to change a few 
>>>> packages to update their versions and/or change their build instructions 
>>>> because of header file changes, things installing into different places, 
>>>> and all sorts of other kruft.
>>>>     
>>>>         

package/flex/flex.mk

When you run ./configure --help in build_ixxx/flex-2.5.4, you see the
following output below.

flex.mk specifies however:

                --bindir=/usr/bin \
                --sbindir=/usr/sbin \
                --libexecdir=/usr/lib \
                --sysconfdir=/etc \
                --datadir=/usr/share \
                --localstatedir=/var \
                --mandir=/usr/man \
                --infodir=/usr/info \

My build process stopped because these are un-recognized options. Is is
my build environment? Haven't had this problem in the past.

ingmar@nexoc:~/dev/astlinux-newtrunk/build_i586/flex-2.5.4> ./configure
--help
Usage: configure [options] [host]
Options: [defaults in brackets after descriptions]
Configuration:
  --cache-file=FILE       cache test results in FILE
  --help                  print this message
  --no-create             do not create output files
  --quiet, --silent       do not print `checking...' messages
  --version               print the version of autoconf that created
configure
Directory and file names:
  --prefix=PREFIX         install architecture-independent files in PREFIX
                          [/usr/local]
  --exec-prefix=PREFIX    install architecture-dependent files in PREFIX
                          [same as prefix]
  --srcdir=DIR            find the sources in DIR [configure dir or ..]
  --program-prefix=PREFIX prepend PREFIX to installed program names
  --program-suffix=SUFFIX append SUFFIX to installed program names
  --program-transform-name=PROGRAM run sed PROGRAM on installed program
names
Host type:
  --build=BUILD           configure for building on BUILD [BUILD=HOST]
  --host=HOST             configure for HOST [guessed]
  --target=TARGET         configure for TARGET [TARGET=HOST]
Features and packages:
  --disable-FEATURE       do not include FEATURE (same as
--enable-FEATURE=no)
  --enable-FEATURE[=ARG]  include FEATURE [ARG=yes]
  --with-PACKAGE[=ARG]    use PACKAGE [ARG=yes]
  --without-PACKAGE       do not use PACKAGE (same as --with-PACKAGE=no)
  --x-includes=DIR        X include files are in DIR
  --x-libraries=DIR       X library files are in DIR
--enable and --with options recognized:


Philip Prindeville wrote:
> I built everything from scratch.
> 
> Which flex are you talking about?
> 
> At least one package has it's own copy of flex...
> 
> -Philip
> 
> Ingmar Schraub wrote:
>> Hello Philip,
>>
>> did you build recently "flex" from scratch? Apparently flex 2.5.4 has
>> less 'configure' options..  Needs a fix.
>>
>> regards,
>>
>> Ingmar
>>
>> Philip Prindeville wrote:
>>   
>>> Following the movement to Linux kernel 2.6.25, I've had to change a few 
>>> packages to update their versions and/or change their build instructions 
>>> because of header file changes, things installing into different places, 
>>> and all sorts of other kruft.
>>>     
> 
> 
> 
> -------------------------------------------------------------------------
> This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
> Build the coolest Linux based applications with Moblin SDK & win great prizes
> Grand prize is a trip for two to an Open Source event anywhere in the world
> http://moblin-contest.org/redirect.php?banner_id=100&url=/
> _______________________________________________
> Astlinux-devel mailing list
> Ast...@li...
> https://lists.sourceforge.net/lists/listinfo/astlinux-devel
> 
> Donations to support AstLinux are graciously accepted via PayPal to pa...@kr....

I've diff'd (original) target/device/net5501/linux.config against the 
build_i586/linux-2.6.25.19-astlinux/.config file.

New capabilities being added in as "+" aren't very suspect.  These 
either default on (+CONFIG_) or default off (+# CONFIG_xxx is not set).  
There might be some handy new capabilities that I don't recognize, in 
which case call them out and we'll enable them...

Capabilities going away (-CONFIG_xxx) are more important.  Some of them 
are legacy stuff (like ISA bus retirement).

Things like CONFIG_HDLC_PPP don't even appear as "# ... is not set" -- 
presumably they've gone away.

Not sure what's going on with CONFIG_OCF_xxx ...  I've turned this on, 
but... it keeps going away.

Can someone have a look at the attachment and tell me how it looks?

A lot of new capabilities have sprung up.

Also, this isn't the trunk version of 
target/device/net5501/linux.config... it's slightly modified:

--- target/device/net5501/linux.config  (revision 2043)
+++ target/device/net5501/linux.config  (working copy)
@@ -366,7 +366,7 @@
 CONFIG_INET6_AH=m
 CONFIG_INET6_ESP=m
 CONFIG_INET6_IPCOMP=m
-CONFIG_IPV6_MIP6=m
+CONFIG_IPV6_MIP6=y
 CONFIG_INET6_XFRM_TUNNEL=m
 CONFIG_INET6_TUNNEL=m
 CONFIG_INET6_XFRM_MODE_TRANSPORT=m
@@ -1061,8 +1061,8 @@
 #
 # Wireless LAN
 #
-# CONFIG_WLAN_PRE80211 is not set
-# CONFIG_WLAN_80211 is not set
+CONFIG_WLAN_PRE80211=y
+CONFIG_WLAN_80211=y
 
 #
 # USB Network Adapters
@@ -2106,6 +2106,22 @@
 # CONFIG_VIRTIO_BALLOON is not set
 
 #
+# OCF Configuration
+#
+CONFIG_OCF_OCF=m
+CONFIG_OCF_RANDOMHARVEST=m
+CONFIG_OCF_FIPS=m
+CONFIG_OCF_CRYPTODEV=m
+CONFIG_OCF_CRYPTOSOFT=m
+CONFIG_OCF_SAFE=m
+# CONFIG_OCF_IXP4XX is not set
+CONFIG_OCF_HIFN=m
+CONFIG_OCF_HIFNHIPP=m
+# CONFIG_OCF_TALITOS is not set
+CONFIG_OCF_OCFNULL=m
+CONFIG_OCF_BENCH=m
+
+#
 # Library routines
 #
 CONFIG_BITREVERSE=y


Again, notice that the OCF stuff doesn't seem to "stick".  Not sure why.

-Philip

I built everything from scratch.

Which flex are you talking about?

At least one package has it's own copy of flex...

-Philip

Ingmar Schraub wrote:
> Hello Philip,
>
> did you build recently "flex" from scratch? Apparently flex 2.5.4 has
> less 'configure' options..  Needs a fix.
>
> regards,
>
> Ingmar
>
> Philip Prindeville wrote:
>   
>> Following the movement to Linux kernel 2.6.25, I've had to change a few 
>> packages to update their versions and/or change their build instructions 
>> because of header file changes, things installing into different places, 
>> and all sorts of other kruft.
>>     

Hello Philip,

did you build recently "flex" from scratch? Apparently flex 2.5.4 has
less 'configure' options..  Needs a fix.

regards,

Ingmar

Philip Prindeville wrote:
> Following the movement to Linux kernel 2.6.25, I've had to change a few 
> packages to update their versions and/or change their build instructions 
> because of header file changes, things installing into different places, 
> and all sorts of other kruft.

wanrouter crashes severely (in kernel) when unloading...

And Arno's firewall uses --dport which iptables 1.4.1.1 don't seem to 
support.

I'll dig into Arno's firewall and figure out what needs to be fixed.

-Philip

So, at this point, I still need to do 4 things for a proof-of-concept.

(1) find the multi-route patch for 2.6.25
(2) find the pf_ring version for 2.6.25
(3) find the real-time patches for 2.6.25
(4) update the .config file to include all of the new flags, and verify 
the old ones

and I can run an image on the net5501.

Anyone got some time to poke around for those first 3, or look at the 
last (which will need to be done after the patches, in case any more 
CONFIG_xxx values get introduced in the patches)?

Thanks,

-Philip

Yow.  This is with 2.6.25.19 with all the patches I could find (posted 
last night).

The number of new kernel options is huge.

Anyone notice any for the net5501 that aren't on, but should be, or vice 
versa?

I noticed that CONFIG_GENERIC_GPIO isn't on but should be.  I'll fix that.

Not sure why CONFIG_OCF_xxx went away...  The patch:

target/device/kernel-patches/linux-2.6.x-ocf.patch

definitely contains a Kconfig.in file for it.

Also what's strange is when I do a build now, the tools don't get 
unpacked before Linux does...