2xUSB BB arrived...
ernstoud
First impressions:
- larger but still surprisingly small
- the IO board has been prefitted in the housing when delivered;
getting it out needed some clever thinking; the best way I found was
using the tip of the power supply fitted into the power supply
connector and pushing gently inwards and prying the PCB upwards using
the pin and tearing with a finger on the ethernet port ridge; the
board then flips out (the power supply connector and USB ports
protrude into the plastic ridge of the housing, so the board cannot be
lifted out normally)
- I doubt whether the case can be opened again when the top lid is put
on; the 4 clamps that hold the lid appear to be very sturdy (hard to
press inward)
- the lid comes with a screw, there is a hole in the IO board for it
to secure it into the housing but the first bullet above suggests it
is't really needed
- the pre-fitted 0,1" pins for serial and jtag are very very nice to
have
- for an electronic engineer as myself the floating CPU board held in
place *only* by its 0,05" connector pins is a kludge, using the box in
a moving environment puts all the stress on the connector soldering;
not the way to do it if you ask me, but then again, the box has not
been designed to use in a car for instance
On to some software tests... will be back soon...
biff...@yahoo.co.uk
> After (only) 4 days it arrived (UK - NL).
Cool - so mail is *starting* to get back to normal at last.
> First impressions:
>
> - larger but still surprisingly small
>
> - the IO board has been prefitted in the housing when delivered;
> getting it out needed some clever thinking; the best way I found was
> using the tip of the power supply fitted into the power supply
> connector and pushing gently inwards and prying the PCB upwards using
> the pin and tearing with a finger on the ethernet port ridge; the
> board then flips out (the power supply connector and USB ports
> protrude into the plastic ridge of the housing, so the board cannot be
> lifted out normally)
Ah yes, we can probably supply these outside the case if this is
needed, but using the case in this way makes the shipping much more
secure. I'm not sure if they will survive the trip without a sturdy
box round them. You are right, and I should have mentioned this - the
boards are put into the boxes by locating the 2 USB ports first, then
snapping the PSU jack into place. It's a strange arrangement, but
makes it quite secure.
> - I doubt whether the case can be opened again when the top lid is put
> on; the 4 clamps that hold the lid appear to be very sturdy (hard to
> press inward)
Indeed, I've already broken one case like this, although I was in a
hurry and there may be a way with some thought. Perhaps some holes
would have to be drilled to push in the clamps, or maybe the clamps
have to be filed down a bit to make them weaker. The problem is not
so much the clamps, but the sealing ridge - they work together to make
it difficult.
> - the lid comes with a screw, there is a hole in the IO board for it
> to secure it into the housing but the first bullet above suggests it
> is't really needed
Yes, the screw is certainly not needed if you don't use the SDIO card
slot, maybe needed if you're pushing the card in and out a lot.
> - the pre-fitted 0,1" pins for serial and jtag are very very nice to
> have
Cool.
> - for an electronic engineer as myself the floating CPU board held in
> place *only* by its 0,05" connector pins is a kludge, using the box in
> a moving environment puts all the stress on the connector soldering;
> not the way to do it if you ask me, but then again, the box has not
> been designed to use in a car for instance
Agreed. This is one of the reasons the units are supplied with the
box. They just don't hold together as nicely as the original BB, so
it's recommended to keep them in the box. Perhaps an alternative
support mechanism can be devised to make it all sturdy. It's one of
the downsides of using 0.05 pitch connectors.
> On to some software tests... will be back soon...
Looking forward to it!
cheers,
Biff.
ernstoud
- The board comes with an 2.6.30.5 linux and minimal rootfs installed.
Nice; but it is really very minimal. Also it runs the udhpcd daemon
in discovery mode so without a network connection it doesn't boot into
its minimal (hush) shell but continues to spout 'sending discovery'
messages which I couldnt't abort. Also this minimal rootfs can't mount
anything else. It would have been nice if it would mount /dev/sda1 and
would contain the kexec in /boot so that it could function as a
primary boot loader.
- I ended up flashing 2.6.30.9 into it and it runs fine. (Took some
time to find a way to flash the 8 Mb chip, BiffJTAG won't do it and I
don't have a native linux that runs python so I ended up using the
original Windows XP RDC JTAG tool with my fingers very, very crossed).
- My 8GB SDHC card is recognized and mounted, but... df doesn't show
it for some reason as mounted although it is (at /mnt/sdb1) and I
can't start mke2fs on the BB2 (generates a kernel error, although it
works fine on my 1 Mb. BB1 running the same kernel - weird) so I can't
create a proper filesystem on it. I used fdisk, made a partition and
changed it to type 83 (linux) and tried copying (using tar) the whole
2.6.30.9 rootfs (using vmware on my windows notebook) which worked
fine, no I/O errors whatsoever. Which is weird since without mke2fs
there is no ext2 filesystem on it (it was previously formatted as
FAT32). The result appears to contain a good copy of the rootfs but
the BB2 wouldn't boot from it; probably since it doesn't contain a
proper ext2 filesystem.
And without a working mke2fs I can't create it... I am stuck here
since all my other SD reader's don't recognize SDHC cards. This will
have to wait until I get my SDHC USB reader from China.
So far so good.
On Jan 22, 1:17 pm, "biffe...@yahoo.co.uk" <biffe...@yahoo.co.uk>
wrote:
biff...@yahoo.co.uk
> - The board comes with an 2.6.30.5 linux and minimal rootfs installed.
> Nice; but it is really very minimal. Also it runs the udhpcd daemon
> in discovery mode so without a network connection it doesn't boot into
> its minimal (hush) shell but continues to spout 'sending discovery'
> messages which I couldnt't abort. Also this minimal rootfs can't mount
> anything else. It would have been nice if it would mount /dev/sda1 and
> would contain the kexec in /boot so that it could function as a
> primary boot loader.
That's just a test kernel I'm afraid, I expected it to be replaced
with Slackware/Debian or whatever by the user.
> - I ended up flashing 2.6.30.9 into it and it runs fine. (Took some
> time to find a way to flash the 8 Mb chip, BiffJTAG won't do it and I
> don't have a native linux that runs python so I ended up using the
> original Windows XP RDC JTAG tool with my fingers very, very crossed).
You have a Linux machine that can run BiffJTAG, but you don't use my
upload scripts, because you don't want/can't run Python? So as a
result you JTAG kernel + Bootloader together every time you want a new
kernel? Wow - that is really twisted! :).
Incidentally, the 8Mb version of the ether upload script is here:
https://bifferboard.svn.sourceforge.net/svnroot/bifferboard/slack/files/boot/
I made a serial version of this and tested it, but since it was a one-
line change I seem to have forgotten to check it in. I'll do that
tonight (even if it's only MP who uses it...).
> - My 8GB SDHC card is recognized and mounted, but... df doesn't show
> it for some reason as mounted although it is (at /mnt/sdb1) and I
> can't start mke2fs on the BB2 (generates a kernel error, although it
> works fine on my 1 Mb. BB1 running the same kernel - weird) so I can't
> create a proper filesystem on it. I used fdisk, made a partition and
> changed it to type 83 (linux) and tried copying (using tar) the whole
> 2.6.30.9 rootfs (using vmware on my windows notebook) which worked
> fine, no I/O errors whatsoever. Which is weird since without mke2fs
> there is no ext2 filesystem on it (it was previously formatted as
> FAT32). The result appears to contain a good copy of the rootfs but
> the BB2 wouldn't boot from it; probably since it doesn't contain a
> proper ext2 filesystem.
It's prob. still FAT, Linux doesn't care what the partition type is
and just uses it as a hint, e.g. to find swap partitions etc....
> And without a working mke2fs I can't create it... I am stuck here
> since all my other SD reader's don't recognize SDHC cards. This will
> have to wait until I get my SDHC USB reader from China.
>
> So far so good.
You can always try the Slackware distro, I'm sure that has mke2fs :).
regards,
Biff.
ernstoud
sector when the data in it are not FF, so it is reasonably fast.)
In the meantime I got the SD card fdisk'ed and mke2fs'd properly, i.e.
it has a primary partition table, its type is linux, it is bootable
and it contains a functioning ext2 filesystem. I can work with it
fine, e2fsck does not complain, but the BB2 does not boot from it; the
bootlog seems fine but in the end it complains it can't find /etc/
preinit.... weird:
Output of [fdisk /mnt/sdb] show partition info:
Disk /dev/sdb: 7969 MB, 7969177600 bytes
221 heads, 20 sectors/track, 3521 cylinders
Units = cylinders of 4420 * 512 = 2263040 bytes
Disk identifier: 0x00000000
Device Boot Start End Blocks Id System
/dev/sdb1 * 1 3521 7781400 83 Linux
Directory /mnt/sdb1:
root@OpenWrt:/#ls /mnt/sdb1
/bin /home /lost+found /rom /sys /
var
/dev /jffs /mnt /root /
tmp /www
/etc /lib /proc /sbin /usr
But the bootlog says (snippet):
[ 1.061358] Waiting for root device /dev/sda1...
[ 6.035231] scsi 0:0:0:0: Direct-Access Generic STORAGE
DEVICE 9451 PQ: 0 ANSI: 0
[ 6.188243] sd 0:0:0:0: [sda] 15564800 512-byte hardware sectors
(7969 MB)
[ 6.191478] sd 0:0:0:0: [sda] Write Protect is off
[ 6.192727] sd 0:0:0:0: [sda] Assuming drive cache: write through
[ 6.199022] sd 0:0:0:0: [sda] 15564800 512-byte hardware sectors
(7969 MB)
[ 6.202347] sd 0:0:0:0: [sda] Write Protect is off
[ 6.203607] sd 0:0:0:0: [sda] Assuming drive cache: write through
[ 6.205049] sda: sda1
[ 6.213015] sd 0:0:0:0: [sda] Attached SCSI removable disk
[ 6.479590] VFS: Mounted root (ext2 filesystem).
[ 6.482096] Freeing unused kernel memory: 112k freed
[ 6.486101] Warning: unable to open an initial console.
[ 6.487516] Failed to execute /etc/preinit. Attempting defaults...
[ 6.489137] Kernel panic - not syncing: No init found. Try passing
init= option to kernel.
I give up.
ernstoud
behold; it boots from SDHC. Hurrah.
Sigh.
---
ernstoud
The latest e2fsprogs don't work with uclibc 0.9.30.1 (they all
generate segfaults).
(There is a bug report pending on this.)
The version 1.40.11-1 (available at http://downloads.openwrt.org/kamikaze/8.09/rdc/packages/)
does work with 0.9.30.1.
That was the reason I couldn't make the file system on the SDHC card.
Andrew Scheller
>> Nice; but it is really very minimal. Also it runs the udhpcd daemon
>> in discovery mode so without a network connection it doesn't boot into
>> its minimal (hush) shell but continues to spout 'sending discovery'
>> messages which I couldnt't abort. Also this minimal rootfs can't mount
>> anything else. It would have been nice if it would mount /dev/sda1 and
>> would contain the kexec in /boot so that it could function as a
>> primary boot loader.
>
> That's just a test kernel I'm afraid, I expected it to be replaced
> with Slackware/Debian or whatever by the user.
Even if it's only a test kernel, any chance you could bung it up on
http://bifferos.bizhat.com/firmware/ for people wanting to "reset to
factory default" ?
Lurch
Andrew Scheller
> - the pre-fitted 0,1" pins for serial and jtag are very very nice to
> have
Be nicer if they were normal length (see other thread).
> - for an electronic engineer as myself the floating CPU board held in
> place *only* by its 0,05" connector pins is a kludge, using the box in
Yeah, I find that when (trying to!) putting a connector on the header
pins, I have to press down on the corner of the top board nearest the
SD slot & USB ports to stop it going wonky.
> a moving environment puts all the stress on the connector soldering;
> not the way to do it if you ask me, but then again, the box has not
> been designed to use in a car for instance
I guess if vibration (e.g. in a car) is a concern (and you aren't
worried about getting the boards apart again), you could put a blob of
hot glue on top of each of the USB ports and on top of the capacitor
with the 'pad', before attaching the CPU board, which might give a bit
more mechanical stability?
Lurch