• Home
  • Help
  • Search
  • Login
  • Register
  Show Posts
Pages: 1 ... 9 10 [11] 12
151  Hardware and U-Boot firmware / U-Boot stuff / USB trap on: May 03, 2011, 06:52:36 PM
(Using 3.4.27+pingtoo patch)

While upgrading a kernel on a USB drive, I could not figure why I still had the old kernel after the upgrade. The trap was my uBoot parameters for booting from USB. This uses the reset after the bootm to handle the intermitent case when the ext2load fails.

Code:
usb start;ext2load usb 0:1 0x8000000 /uImage; bootm 0x8000000;reset

The catch is that without initializing the memory the bootm command will load what was last in memory, I was suprised to find out that 'reset' does not initialize. I have changed my USB boot to:

Code:
mw 0x8000000 0 1000;usb start;ext2load usb 0:1 0x8000000 /uImage; bootm 0x8000000;reset
152  Linux Stuff / Kernel / Re: 2.6.38.4 new kernel available on: May 02, 2011, 06:41:55 AM
Reading the patch list 2.6.35.13 has a significant UBIFS patch
153  General Category / General Discussion / Re: rootfs size on: May 01, 2011, 03:38:37 PM
du reports the size directory contents so -

Code:
# du -sh /var |grep M

should nail it
154  General Category / General Discussion / Re: rootfs size on: April 30, 2011, 07:41:35 PM
Experiment. Backup before you try anything.

Code:
du -h --max-depth 1 /var

1. log

Refine your /etc/rsyslog.conf and /etc/logrotate.conf setups to minimize the size of your logs. You can delete logfiles that are not in use.

2.  apt

Try

Code:
# apt-get autoremove
# apt-get autoclean

The largest directory generated by apt is usually  /var/cache/apt

3.   Doc.

If you do not want man pages and docs you can delete files in the following directories
/usr/share/doc
/usr/share/man
/usr/share/doc-base


You might also check out localepurge
155  General Category / General Discussion / Re: rootfs size on: April 29, 2011, 04:27:44 PM
Start with

Code:
# du -sh /var
156  Linux Stuff / Kernel / Re: iwconfig kernel oops on: April 22, 2011, 03:58:27 PM
The crash looks very similar to the one that I was experiencing with a SheevaPlug and USB wifi until 2.6.38.2 cured the problem.

Same path to a crash.

2.6.38.2 has been very stable for me.
157  Linux Stuff / General Linux questions / Flashybrid on: April 06, 2011, 06:48:13 PM
Sheevaplug, Debian Wheezy sheeva-with-linux 2.6.38.2

I am posting this to summarize my Flashybrid experiences (bump Flashybrid in the topic list).

I have used Flashybrid on both SD and NAND. It seems to work OK. This incurs minimal flash writes so this should lead to a very long flash life cycle. The running Debian system is a readonly root. For the systems that I am running the main issue has always been /var/log. Flashybrid makes for a very inactive SD card.

The positive experience here is that Flashybrid on a Sheevaplug results in minimal flash activity.

See:
 http://plugcomputer.org/plugforum/index.php?topic=1074.msg11987#msg11987
http://plugcomputer.org/plugforum/index.php?topic=1074.msg8720#msg8720

I have a remaining question regarding the ramtmp. It would seem to me that the ramtmp specified in  the above links is redundant when one uses an fstab like the one below. I guess I do not see the reason for using ramtmp in place of fstab. Comments of the following fstab would be appreciated, anyone?

fstab
Code:
# NAND root
/dev/root       /       ubifs   defaults,noatime,rw     0 0
# SDHC root
#/dev/root      /       ext2    rw,relatime,errors=continue 0 0
tmpfs   /lib/init/rw    tmpfs   size=16K,rw,nosuid,relatime,mode=0755 0 0
tmpfs   /var/run        tmpfs   size=1M,rw,nosuid,mode=0755     0 0
tmpfs   /var/lock       tmpfs   size=512K,rw,noexec,nosuid,nodev,mode=1777 0 0
tmpfs   /var/tmp        tmpfs   size=512K,rw,noexec,nosuid,nodev,mode=1777 0 0
tmpfs   /tmp    tmpfs   defaults,nosuid,nodev,mode=1777 0 0

ramtmp
Code:
/var/lib/dhcp3
/var/lib/php5
/var/lib/misc
/var/lib/ntp
/var/lib/urandom

ramstore
Code:
/etc
/var/backups
/var/lib/dbus
/var/lib/logrotate
/var/log
/root
/var/spool/cron

The patch for /etc/init.d/flashybrid is:

Code:
--- /xtc/init.d/flashybrid      2006-10-30 00:46:50.000000000 +1100
+++ /etc/init.d/flashybrid      2011-04-05 10:59:50.000000000 +1000
@@ -12,12 +12,14 @@

 ### BEGIN INIT INFO
 # Provides:          flashybrid
-# Required-Start:    $all
-# Required-Stop:     $syslog
+# Required-Start:    $local_fs
+# Required-Stop:     $local_fs
 # Should-Start:      $local_fs
 # Should-Stop:       $local_fs
-# Default-Start:     2 3 4 5
-# Default-Stop:      0 1 6
+# X-Start-Before:    $network
+# X-Stop-After:      $network
+# Default-Start:     S
+# Default-Stop:      0 6
 # Short-Description: automates use of a flash disk as the root filesystem
 # Description:       Flashybrid is a system to help in setting up and managing hybrid
 #                    flash/disk/ram based Debian systems which can run most of the time

Followed by:
Code:
rm /etc/rc{0,1,2,3,4,5,6}.d/*flashybrid
insserv -v flashybrid
158  Linux Stuff / Kernel / Issues with 2.6.38.2 on: April 03, 2011, 07:00:05 PM
I have encountered a few issues with 2.6.38.2, I am not sure when they first appeared.
Also if they have been addressed somewhere else on the forum, please respond with a link.
The uBoot (3.4.27+pingtoo) setup is fixed to boot only from the USB drive.
Using Debian Wheezy.
SD card(s):  Integral 4GB class 6, 8GB class 6 and 8GB class 10.

Issue 1) When booting from USB (HDD via a HUB), if I have a SD card inserted when the plug boots (as distinct from inserting the SD after I have booted) then the plug never sees the SD. e.g. fdisk -l or dmesg both pretend that the SD does not exist. If I remove the SD and re-insert it after I have booted then the SD starts to work.


The only mention of mmc in dmesg is:

Code:
sdhci: Secure Digital Host Controller Interface driver
sdhci: Copyright(c) Pierre Ossman
mmc0: mvsdio driver initialized, lacking card detect (fall back to polling)




Issue 2) I have a script which is cloned from the Sheevaplug installer rcS which reads

Code:
        echo "**** Removing MMC partitions"
        echo -e "o\nw\n" | fdisk /dev/mmcblk0
        if [ $? != 0 ]; then
                echo "**** ERROR - Could not find or remove MMC partitions"
                echo "**** ERROR - Will try to continue anyway"
        fi

        echo "**** Creating /dev/mmcblk0p1 & /dev/mmcblk0p2 partitions"
        echo -e "n\np\n1\n\n+4M\nn\np\n2\n\n\nw\n" | fdisk /dev/mmcblk0

        echo "**** Making EXT2 partition on /dev/mmcblk0p1 for uImage"
        mkfs.ext2 -L shvplug-uImage /dev/mmcblk0p1
        if [ $? != 0 ]; then
                echo "**** ERROR - mkfs.ext2 /dev/mmcblk0p1"
                exit 1
        fi


The mkfs fails with

Code:
Could not stat /dev/mmcblk0p1 --- No such file or directory

If I add a <tt>sleep 1</tt> before the mkfs it works fine.

This did not happen with the 2.6.35.* kernels.
159  Linux Stuff / Kernel / Re: Sheevaplug - wireless access point on: March 31, 2011, 04:10:02 PM
So far (fingers crossed) it would appear that 2.6.38.2 does not have the problem.
2.6.37.* and 2.6.35.11 definitely fail
160  Linux Stuff / Kernel / Sheevaplug - wireless access point on: March 30, 2011, 05:42:44 PM
Using Debian Wheezy and  USB wireless I have a wireless access point working.
My current kernel is sheeva-with-linux 2.6.38.2

The USB wireless is a very cheap (~$20) Planex GW-USBMini2N. This would seem to be a handy solution for a wireless AP.

I am using the rt2800usb module driver (generic), hostapd and dnsmasq. (I can detail the setup of these if someone wants it. )
The result is good, low CPU utilzations, high speed ... except I get the occasional kernel panic and always when I am not watching it.

Modules linked in: iptable_filter ipt_MASQUERADE iptable_nat nf_nat nf_conntrack_ipv4 nf_conntrack nf_defrag_ipv4 ip_tables x_tables dm_crypt dm_mod rt2800usb rt2800lib rt2x00usb rt2x00lib mac80211 cfg80211 rfkill mv_cesa

I started with the 2.6.35.11 kernel and have shifted to the 2.6.38.2 hoping to have the problem go away. 

After using Sheevaplugs for some time now, this is my first kernel crash. What is the usual diagnostic path? It would seem that I have to wait for a day to generate the problem.

Anyone have any suggestions?
161  Hardware and U-Boot firmware / Hardware / Re: dreamplug questions on: March 18, 2011, 11:26:49 PM
I believe later versions of uboot have a 'sleep' command. Any chance?
162  Hardware and U-Boot firmware / U-Boot stuff / Re: Upgrading uboot on Globalscale Sheevaplug on: March 10, 2011, 02:57:38 AM
OK.
I was trying to point out that you are using ext2load from a fat device.
Either make it ext2 or use fatload
163  Hardware and U-Boot firmware / U-Boot stuff / Re: Upgrading uboot on Globalscale Sheevaplug on: March 09, 2011, 05:30:42 PM
Quote
3 Fat partitions

setenv bootcmd_usb 'usb start; ext2load usb 0:1 0x0800000 /uImage; ext2load usb 0:1 0x01100000 /uInitrd'

(Re)start USB...
USB:   scanning bus for devices... 1 USB Device(s) found
       scanning bus for storage devices... 0 Storage Device(s) found
** Bad partition 1 **
** Bad partition 1 **
## Booting image at 00800000 ...
Bad Magic Number
 

Try fatload instead?
164  Linux Stuff / Kernel / Re: 2.6.37.2 new kernel available on: March 02, 2011, 07:49:08 PM
Can someone point me at a message or documentation reference that tells me which uBoot(s) to use for these kernels on a GuruPlug?
165  Linux Stuff / Kernel / Re: new 2.6.36.4, new 2.6.37.1 and updated 2.6.35.11 kernels available on: February 19, 2011, 04:05:15 PM
I have spent some and effort trying to get an ASUS-NB13 USB wifi card working on debian 2.6.35.11(2) (including compiling an ASUS supplied 2870 module that has missing symbols). Do you know if there is anything in the later kernels that will stop me from crushing the NB13?
Pages: 1 ... 9 10 [11] 12