2.6.30 new release

[cbxbiker61]

2.6.30 is now available.

Kernel and modules are available from the following locations:

IPV4: http://sheeva.with-linux.com/sheeva/

IPV6: http://sheeva6.with-linux.com/sheeva/

Features IPV6, CIFS, NFS4, EXT3, EXT4, JFS, XFS, FUSE(for ntfs-3g), UBIFS, usb-serial, uvcvideo, iptables, appletalk, bluetooth and v4l.

Those of you who have requested feature please test functionality.

[Rabeeh Khoury]

Thanks a lot for the post.

BTW - Which mvsdio patches are you using as part of the build?

[cbxbiker61]

Rabeeh, patches are in /sheeva/2.6.30/patches.

[Pandemonium]

Could someone who has this working post a printenv, or maybe give me some hints what I have wrong?  I'm getting Kernel panic - not syncing: No init found after flashing with this.

Here's my printenv:

Code:

baudrate=115200
loads_echo=0
ipaddr=10.4.50.165
serverip=10.4.50.5
rootpath=/mnt/ARM_FS/
netmask=255.255.255.0
CASset=min
MALLOC_len=1
ethprime=egiga0
bootargs_root=root=/dev/nfs rw
bootargs_end=:::DB88FXX81:eth0:none
image_name=uImage
standalone=fsload 0x2000000 $(image_name);setenv bootargs $(console) root=/dev/mtdblock0 rw ip=$(ipaddr):$(serverip)$(bootargs_end) $(mvPhoneConfig); bootm 0x2000000;
ethaddr=00:50:43:56:23:2e
ethmtu=1500
mvPhoneConfig=mv_phone_config=dev0:fxs,dev1:fxs
mvNetConfig=mv_net_config=(00:11:88:0f:62:81,0:1:2:3),mtu=1500
usb0Mode=host
yuk_ethaddr=00:00:00:EE:51:81
nandEcc=1bit
netretry=no
rcvrip=169.254.100.100
autoload=no
ethact=egiga0
loadaddr=0x2000000
mtd1Size=0x300000
filesize=0x44ad000
arcNumber=2097
bootcmd=nand read.e 0x800000 0x100000 0x400000; bootm 0x800000
bootargs=rootfstype=jffs2 console=ttyS0,115200 mtdparts=orion_nand:0x400000@0x100000(uImage),0x1fb00000@0x500000(rootfs) rw root=/dev/mtdblock1
stdin=serial
stdout=serial
stderr=serial
console=console=ttyS0,115200 mtdparts=nand_mtd:0x100000@0(uboot)ro,0x300000@0x100000(uImage),0x1fc00000@0x400000(rootfs)rw
mainlineLinux=yes
enaMonExt=no
enaCpuStream=no
enaWrAllo=no
pexMode=RC
disL2Cache=no
setL2CacheWT=yes
disL2Prefetch=yes
enaICPref=yes
enaDCPref=yes
sata_dma_mode=yes
netbsd_en=no
vxworks_en=no
bootdelay=3
disaMvPnp=no
enaAutoRecovery=yes

Environment size: 1350/131068 bytes

[cbxbiker61]

Did you change your boot partition from 3M to 4M? or is it still at 3M?

It looks like you copied the bootargs from the README verbatim, which wouldn't be correct if you have a 3M uboot partition.  In that case you should be able to replace the partition size/locations in the "bootargs" with those I see in your "console" variable.

[Pandemonium]

Ya know, it's been long enough since I messed with it that I don't remember.  Looks like you're right:

Code:

-sh-3.2# cat /proc/mtd
dev:    size   erasesize  name
mtd0: 00100000 00020000 "uboot"
mtd1: 00300000 00020000 "uImage"
mtd2: 1fc00000 00020000 "rootfs"

I took my best shot at changing the bootargs, but I'm still getting the same thing.

Code:

Marvell>> printenv
baudrate=115200
loads_echo=0
ipaddr=10.4.50.165
serverip=10.4.50.5
rootpath=/mnt/ARM_FS/
netmask=255.255.255.0
CASset=min
MALLOC_len=1
ethprime=egiga0
bootargs_root=root=/dev/nfs rw
bootargs_end=:::DB88FXX81:eth0:none
image_name=uImage
standalone=fsload 0x2000000 $(image_name);setenv bootargs $(console) root=/dev/mtdblock0 rw ip=$(ipaddr):$(serverip)$(bootargs_end) $(mvPhoneConfig); bootm 0x2000000;
ethaddr=00:50:43:56:23:2e
ethmtu=1500
mvPhoneConfig=mv_phone_config=dev0:fxs,dev1:fxs
mvNetConfig=mv_net_config=(00:11:88:0f:62:81,0:1:2:3),mtu=1500
usb0Mode=host
yuk_ethaddr=00:00:00:EE:51:81
nandEcc=1bit
netretry=no
rcvrip=169.254.100.100
autoload=no
ethact=egiga0
loadaddr=0x2000000
mtd1Size=0x300000
filesize=0x44ad000
arcNumber=2097
bootargs=rootfstype=jffs2 console=ttyS0,115200 mtdparts=orion_nand:0x300000@0x100000(uImage),0x1fb00000@0x500000(rootfs) rw root=/dev/mtdblock1
bootcmd=nand read.e 0x800000 0x100000 0x300000; bootm 0x800000’
stdin=serial
stdout=serial
stderr=serial
console=console=ttyS0,115200 mtdparts=nand_mtd:0x100000@0(uboot)ro,0x300000@0x100000(uImage),0x1fc00000@0x400000(rootfs)rw
mainlineLinux=yes
enaMonExt=no
enaCpuStream=no
enaWrAllo=no
pexMode=RC
disL2Cache=no
setL2CacheWT=yes
disL2Prefetch=yes
enaICPref=yes
enaDCPref=yes
sata_dma_mode=yes
netbsd_en=no
vxworks_en=no
bootdelay=3
disaMvPnp=no
enaAutoRecovery=yes

[Pandemonium]

Woops, I think the problem was with the jffs2 image I was using.  I just tried another one, and I booted in fine.

Though, I did get four message during boot like these:

Code:

FATAL: Could not load /lib/modules/2.6.30/modules.dep: No such file or directory

Are those anything to worry about?

[kilowatt]

try executing:

Code:

depmod -a

[CharlesWGreenJr]

MythTV works with the HVR-1950 on this kernel / module set.  Just confirmed NTSC recording and playback with video + audio (I haven't tried QAM or ATSC yet, left as an exercise for others at least until I get the time to bend my 'if it ain't broke' rule and alter my working configuration).

Thanks yet again, cbxbiker61!

-Charles Green

[levrin]

These kernel images and flashing scripts are so convenient! Could I make a request for bridge support? I was thinking of trying to run a VPN server on my plug, but 'brctl addbr br0' spat out 'add bridge failed: Package not installed'.

[cbxbiker61]

levrin, bridging support has just been added.

[Rabeeh Khoury]

I'm worried about how much performance loss we will get on TCP/IP based applications when adding bridging, iptables etc...

Any thoughts about this?

[cbxbiker61]

Regarding performance.

I ran iperf to the sheeva and an x86 machine.  The raw data rate to the sheeva is considerably lower to the sheeva compared to the x86 machine.  The cpu utilization also peaked at about 98% to the sheeva whereas the x86 machine peaked at about 25%.  Whether this would adversely affect a particular usage scenario would depend a lot upon the type of load.  Serving files coming from a USB attached device would already be limited by USB speeds to approximately 30MB/sec (240 MBits/sec).

A server process that is purely in memory could hit the data rate ceiling.  Off the top of my head I can't think of a memory only server process (besides firewalling, vpn, etc., which require the extension modules in question) that would benefit greatly from an increase in the raw data rate.

I would consider adding a stripped kernel in addition to the existing kernel if someone has a real world workload that:
A. Isn't limited by USB speeds
B. Doesn't require iptables/bridging
C. Would benefit from an increase raw throughput

***************
iperf -c sheeva
------------------------------------------------------------
Client connecting to sheeva, TCP port 5001
TCP window size: 16.0 KByte (default)
------------------------------------------------------------
[  3] local 192.168.1.4 port 58031 connected with 192.168.1.9 port 5001
[ ID] Interval       Transfer     Bandwidth
[  3]  0.0-10.0 sec    580 MBytes    486 Mbits/sec

**************
iperf -c x86pc
------------------------------------------------------------
Client connecting to cheetah, TCP port 5001
TCP window size: 16.0 KByte (default)
------------------------------------------------------------
[  3] local 192.168.1.4 port 59802 connected with 192.168.1.2 port 5001
[ ID] Interval       Transfer     Bandwidth
[  3]  0.0-10.0 sec  1021 MBytes    856 Mbits/sec

[Pandemonium]

I assume those iperf numbers are from this 2.6.30 kernel...?
Do we know how they compare to the stock kernel, or at least one without iptables, bridging, and other applicable additions?

[cbxbiker61]

Yeah, the iperf are for 2.6.30.  I seem to recall that the RC's that I tested had similar results.

I haven't tested with iptables, etc. stripped out.  I probably won't do that unless someone can come up with a use case that would benefit from such a change.