Debian on HP T5325 Thin Client

LeggoMyEggo Wrote:
-------------------------------------------------------
> bodhi Wrote:
> --------------------------------------------------
> -----
> > This is a pretty good box if you can get it for
> <=
> > 30 USD used.

yes you can - i got 3 for less than £25 delivered

> 10/100 Ethernet -1

Agreed, but we cant have everything

> IMO, HP T5740 thin client is best cheapest low
> power box with expansion options. Not Marvell
> though.....

Had a few of those and other HP variants, the footprint is considerably larger tho, and it will be a nice exercise / to the make this available hardware workable for those of us that like to tinker with cheap arm hardware :)

Gravelrash,

> yes you can - i got 3 for less than £25
> delivered
>

Just saw eBay has a bundle of 6 or 7, works out to something like 17 USD each! there are single ones in range 22-29.

> Had a few of those and other HP variants, the
> footprint is considerably larger tho, and it will
> be a nice exercise / to the make this available
> hardware workable for those of us that like to
> tinker with cheap arm hardware :)

Making Debian kernel/rootfs available was easy, we already have it :). Some effort is needed with u-boot, but I will have a test u-boot sometime this weekend.

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)

@bodhi - i volunteer to mangle one of my units..

i mean test the uboot on one of my devices... I have two left.. I had three but already one has been "liberated".

I wonder if Joey wants one for Easter? ;)

my serial cable should arrive before end of the week.

> > This is a pretty good box if you can get it for
> <=
> > 30 USD used.

I've got a couple units +shipping for 30 euros total. They can be had cheep. Probably because being ARM (and HP) they are the low of the low end.


> 10/100 Ethernet -1

I'm dreaming of hacking it away and exposing its pcie lines (hoping that HP engineers used the pcie for that and not a usb port) so I can hack in a gigabit minipcie card or something.

But I probably won't have the time to do so.


> IMO, HP T5740 thin client is best cheapest low
> power box with expansion options.

you mean like this? http://www.ebay.com/itm/HP-T5740-Thin-Client-16GB-F-2GB-R-w-Expansion-32bit-PCI-Riser-PS-Lot-Avail-/201531654105?hash=item2eec38ffd9:g:wjAAAOSwyQtV48TV

Thanks, but no thanks. :)

bobafetthotmail Wrote:
-------------------------------------------------------

> I'm dreaming of hacking it away and exposing its
> pcie lines (hoping that HP engineers used the pcie
> for that and not a usb port) so I can hack in a
> gigabit minipcie card or something.

Thats what they used... however to access it you will need to remove the graphics chip....

i dont know what i have done to this unit of mine...... somehow i cant seem to break into uboot to stop it then issue the command and a sandisk usb thumb drive i have is exhibiting the same errors... aarrrgghh :headbang

well back to it and im going to compare my uboot environment settings to CV's and see where i went wrong.

Great T5325 reference page with uboot info and links to SATA and GBE mods:

http://www.parkytowers.me.uk/thin/hp/t5325/index.shtml

ahhh parkytowers - loved that site. it started me off a few years ago with my love affair for low power computing

@Gravelrash: for me this works best
- totally shutdown the t5325 (unplug PSU!)
- start kwboot
- plugin PSU, sometimes then it directly starts, sometimes i need to press power on

Another thing wich helped me, but probably has nothing to do with kwboot as that interacts earlier with (uboot or a loader before that? ) was to set bootwait to 5.
1 was to short a time for me to get a keypress recognized.

And I had (and still have) problems with my PL2303 USB serial on Mac and Windows - for me it works reliably only under Linux.

@Bodhi:
>> Some effort is needed with u-boot, but I will have a test u-boot sometime this weekend.

Great, can't nearly await!
(took my breather :-) )

> - plugin PSU, sometimes then it directly starts,
> sometimes i need to press power on
>

Should press power on. Any box with power button, either:

- start kwboot and press power on, or
- power on and start kwboot immediately within 1 second

> Another thing wich helped me, but probably has
> nothing to do with kwboot as that interacts
> earlier with (uboot or a loader before that? )
> was to set bootwait to 5.

It is not applicable. The bootdelay is only relevant after u-boot started.

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)

> Thats what they used... however to access it you will need to remove the graphics chip....

?? no wait. there are 2 pcie controllers in the Kirkwood SoC, one goes to the GPU and that is what the hacker linked above used to get a pcie card in.

I was hoping that HP engineers used the second pcie for the ethernet controller (since because they are insane and could not just use one of the TWO integrated gigabit controllers already there in the SoC, it would also make sense for them to waste a pcie to add a fast ethernet controller).

So i could do a similar hack on the much more disposable ethernet controller.

>i dont know what i have done to this unit of mine

you sure that electrical connections are ok? many times I had to solder cables instead of using cheep chinese rainbow header cables.

>and a sandisk usb thumb drive i have is exhibiting the same errors

I think it could be related to a dumb u-boot. with nsa-325 I had plenty of issues with stock uboot as it could not recognize many usb drives that were too slow to initialize for it.

post in error



Edited 1 time(s). Last edit at 03/03/2016 03:16AM by Gravelrash.

bobafetthotmail,

> I was hoping that HP engineers used the second
> pcie for the ethernet controller (since because
> they are insane and could not just use one of the
> TWO integrated gigabit controllers already there
> in the SoC, it would also make sense for them to
> waste a pcie to add a fast ethernet controller).

Not insane! It's called product tiers. They hoped to sell you a more expensive T5xxx thin client box with gbit ethernet and an eSATA port later :))

Probably the same reason that the Zyxel NSA310 use a Realtek ethernet chip instead of the SoC gbit ethernet (perhaps they got a deal going with Realtek).

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)

@bdhi and CV

to install the test uboot in intend to mangle around with the HP recovery disk and see if it cant be installed that way.

thoughts and comments appreciated

@Gravelrash:

>>to install the test uboot in intend to mangle around with the HP recovery disk and see if it cant be installed that way.

Well, For me kwbooting the new uboot looks easier and safer than flashing and reflashing possibly multiple times...
and really, kwboot works for me each and every time when i do it as i decribed:
i first start kwboot, then cold boot t5325 and it works...

does your serial cable really work correctly?
Can you get to uboot prompt?

edit: I just tried again, i can watch the output through many reboot cycles: uboot, kernel booting, i login and shutdown, see kernel shutdown log, then again uboot starting and so on... all having one serial sessin open (screen /dev/tty.usbserial 115200).



Edited 1 time(s). Last edit at 03/03/2016 01:03PM by CV.

@all

Im going to take myself outside and have a word with myself in the alleyway... i changed my ttl adaptor and things are looking up - managed to break the uboot sequence.

*has stern word with self about keeping more spares to hand*

EDIT: 1 as i go along i will update this last post,
also grabbed a different sandisk drive to boot from this time a cruzer fit and the beast boots... whoop whoop.... now to try the kwboot

EDIT: 2 kwboot is a no go for me - ah well
however, i can happily report that the 3.18.5 kernel with jessie rootfs boots.

as an aside to all this
when i do a printenv from the console after breaking the boot sequence, the following additional information is available

stdin=serial
stdout=serial
stderr=serial
mainlineLinux=no
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=0
disaMvPnp=no
enaAutoRecovery=yes

Environment size: 2226/4092 bytes

[PUE variables]
manufacturer=Hewlett-Packard
product_name=hp t5325
vendor=mitac
serial=9c140c
mac=f4:ce:46:23:27:a0
uuid=11ff7a90-ee38-11de-a140-f4ce462327a0

Environment size: 142/4092 bytes

[HPE variables]
AssetNumber=0
AssetLocation=0
InitialNumlock=On
ClockSpeed=1200
WOLEnabled=Yes
GotoWOL=No
UbootVersion=0x03041916
UbootReleaseDate=01082010

Environment size: 140/4092 bytes
HP>>

so as i cant seem to use kwboot from my x64 system, im more than happy to live dangerously and burn the box if the new uboot fails.



Edited 4 time(s). Last edit at 03/03/2016 03:08PM by Gravelrash.

Gravelrash,

> *has stern word with self about keeping more
> spares to hand*

:) it happpens!

> so as i cant seem to use kwboot from my x64
> system, im more than happy to live dangerously and
> burn the box if the new uboot fails.

I think that it would work if you try the ARM kwboot from your other Kikrwood box, such as Pogo V4/Mobile.

And please get this ouput at stock u-boot serial console:

flinfo
ping 8.8.8.8

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)

@Gravelrash: your and my stock uboot version are the same, look at my attached putty.log somewhere higher up.

I really do not understand why kwboot does not work for you - I took the kwboot from bodhis github and compiled it myself for 32bit - maybe you should also get the source and compile?

Edit: and maybe try on a 32bit linux, a Virtual Box VM will work, I tried, USB attaching the serial through to the VM works.
(Just a wild guess, I often had problems with stuff not working on 32 or 64bit)



Edited 1 time(s). Last edit at 03/04/2016 12:49AM by CV.

@bodhi - i will do that this evening (UK time)

@CV - you are probably right that i should compile the kwboot from github, i tried to save a few moments and used the precompiled one. if it still doesnt work for me - i am going to put it down to the batch of TTL adaptors i am using.

did some googling around and i think i am running a set of "fake/clone" Prolifics. my adaptor works with most other devices i use and the E02 are fabulous with it.

quote from another site

"The USB adaptors are not reliable enough, especially if you
 are linking to old hardware, you will find that they do not reproduce the signal
 timings accurately, which will cause problems with some devices......
 Having said that, the PL2303 is one of the better chips, and should work with
 most devices."

this is what i'm using - output below. i cant expect the world for £2 a piece :)

Bus 002 Device 003: ID 067b:2303 Prolific Technology, Inc. PL2303 Serial Port
Device Descriptor:
  bLength                18
  bDescriptorType         1
  bcdUSB               1.10
  bDeviceClass            0 (Defined at Interface level)
  bDeviceSubClass         0 
  bDeviceProtocol         0 
  bMaxPacketSize0        64
  idVendor           0x067b Prolific Technology, Inc.
  idProduct          0x2303 PL2303 Serial Port
  bcdDevice            3.00
  iManufacturer           1 
  iProduct                2 
  iSerial                 0 
  bNumConfigurations      1
  Configuration Descriptor:
    bLength                 9
    bDescriptorType         2
    wTotalLength           39
    bNumInterfaces          1
    bConfigurationValue     1
    iConfiguration          0 
    bmAttributes         0x80
      (Bus Powered)
    MaxPower              100mA
    Interface Descriptor:
      bLength                 9
      bDescriptorType         4
      bInterfaceNumber        0
      bAlternateSetting       0
      bNumEndpoints           3
      bInterfaceClass       255 Vendor Specific Class
      bInterfaceSubClass      0 
      bInterfaceProtocol      0 
      iInterface              0 
      Endpoint Descriptor:
        bLength                 7
        bDescriptorType         5
        bEndpointAddress     0x81  EP 1 IN
        bmAttributes            3
          Transfer Type            Interrupt
          Synch Type               None
          Usage Type               Data
        wMaxPacketSize     0x000a  1x 10 bytes
        bInterval               1
      Endpoint Descriptor:
        bLength                 7
        bDescriptorType         5
        bEndpointAddress     0x02  EP 2 OUT
        bmAttributes            2
          Transfer Type            Bulk
          Synch Type               None
          Usage Type               Data
        wMaxPacketSize     0x0040  1x 64 bytes
        bInterval               0
      Endpoint Descriptor:
        bLength                 7
        bDescriptorType         5
        bEndpointAddress     0x83  EP 3 IN
        bmAttributes            2
          Transfer Type            Bulk
          Synch Type               None
          Usage Type               Data
        wMaxPacketSize     0x0040  1x 64 bytes
        bInterval               0

Edited 1 time(s). Last edit at 03/04/2016 11:56AM by Gravelrash.

bodhi Wrote:

flinfo
ping 8.8.8.8

Outputs as below

USB 0: host mode
PEX 0: PCI Express Root Complex Interface
PEX interface detected Link X1
Net:   egiga0 [PRIME], egiga1
Hit any key to stop autoboot:  0 
HP>>         flinfo

Bank # 1: 
Flash Base Address  : 0xf8000000
Flash Model         : SST 25VF080B
Manufacturer ID     : 0xbf
Device Id           : 0x258e
Sector Size         : 64K
Number of sectors   : 16
Page Size           : 256
Write Protection    : All

reset the device then ran the following after modifying my variables to match my network

USB 0: host mode
PEX 0: PCI Express Root Complex Interface
PEX interface detected Link X1
Net:   egiga0 [PRIME], egiga1
Hit any key to stop autoboot:  0 
HP>>          
Unknown command '<NULL>' - try 'help'
HP>> ping 8.8.8.8
Using egiga0 device
host 8.8.8.8 is alive
HP>>

Edited 1 time(s). Last edit at 03/04/2016 12:14PM by Gravelrash.

@CV,

Please get output of flinfo in serial console too:

flinfo

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)

All,

I need help with finding u-boot GPL for this box. Here is the ftp server that suppose to have eveything stock T5325:

ftp://ftp.hp.com/pub/tcdebian


UPDATE:

Looks like it is the right one.

ftp://ftp.hp.com/pub/tcdebian//pool/t5325/source/u-boot-1.1.4.tar.bz2

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)



Edited 2 time(s). Last edit at 03/04/2016 07:25PM by bodhi.

Here is a test version. Please try with kwboot.

./kwboot  -t -B 115200 /dev/ttyUSB0 -b uboot.2015.10-tld-2-rc1.t5325.mtd0.kwb   -p

Interrupt serial console after u-boot was loaded. And I'd like to see what these 2 features will behave: Ethernet, and SPI flash.

setenv netmask x.x.x.x
setenv gatewayip x.x.x.x
setenv ipaddr  x.x.x.x
setenv serverip x.x.x.x
ping 8.8.8.8

printenv

And it probably will not boot USB or SATA. So just do reset after testing so that it boots normally back to stock u-boot.

reset

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)

@Bodhi:

>> Please get output of flinfo in serial console too:

HP>> flinfo

Bank # 1:
Flash Base Address  : 0xf8000000
Flash Model         : Numonyx-25P80
Manufacturer ID     : 0x20
Device Id           : 0x2014
Sector Size         : 64K
Number of sectors   : 16
Page Size           : 256
Write Protection    : All

HP>>

So I have another type of flash chip than Gravelrash.

Ping:

HP>> dhcp
BOOTP broadcast 1
*** Unhandled DHCP Option in OFFER/ACK: 28
*** Unhandled DHCP Option in OFFER/ACK: 42
*** Unhandled DHCP Option in OFFER/ACK: 28
*** Unhandled DHCP Option in OFFER/ACK: 42
DHCP client bound to address 192.168.178.143
HP>> ping 8.8.8.8
Using egiga0 device
host 8.8.8.8 is alive
HP>>

Yes, I'm aware of that from the GPL source. This box uses at least 2 different types of SPI flash chips. So try kwboot the test version to see if it recognizes your SPI flash.

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)

To make the u-boot envs area writeable. In serial console:

setenv bootargs_usb 'console=ttyS0,115200 root=LABEL=rootfs rootdelay=10 mtdparts=spi_flash:0x80000@0(uboot),0x80000@0x80000(root)'
saveenv

From now on, fw_setenv will commit the changes to u-boot envs.

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)

@both
as i cant get kwboot to work with my equipment but i do have a tftp server accessible on my lan.

is there a possibility i can assist with testing this new uboot from tftp?

The last thing you can say about me is that i am skilled with uboot env settings, however, im itching to test this if we can do it from tftp.

HAD A GO MYSELF AT BOOTING FROM TFTP USING THE FOLLOWING

setenv serverip 192.168.0.10 # IP of your TFTP server
setenv ipaddr 192.168.0.6
tftpboot

no joy, however i did learn that it is looking for a file called

C0A80006.img
HP>>     tftpboot
*** Warning: no boot file name; using 'C0A80006.img'
Using egiga0 device
TFTP from server 192.168.0.10; our IP address is 192.168.0.6
Filename 'C0A80006.img'.
Load address: 0x2000000
Loading: T T T T T T T T T T 
Retry count exceeded; starting again

armed wit this informatin i changed the name of the above u-boot file
to match and still no joy.

so i changed the load address to something kirkwoody

HP>> tftpboot 0x0800000 u-boot.kwb
Using egiga0 device
TFTP from server 192.168.0.10; our IP address is 192.168.0.6
Filename 'u-boot.kwb'.
Load address: 0x800000
Loading: T T T T T T T T T T 
Retry count exceeded; starting again


HP>> tftpboot 0x0800000 C0A80006.img
Using egiga0 device
TFTP from server 192.168.0.10; our IP address is 192.168.0.6
Filename 'C0A80006.img'.
Load address: 0x800000
Loading: T T T T T T T T T T 
Retry count exceeded; starting again
HP>>

Ok so i spannered the config settings on my tftp server and forgot to amend them a while ago.

Using egiga0 device
host 192.168.0.10 is alive
HP>> tftpboot 0x0800000 u-boot.kwb
Using egiga0 device
TFTP from server 192.168.0.10; our IP address is 192.168.0.6
Filename 'u-boot.kwb'.
Load address: 0x800000
Loading: #################################################################
	 ################
done
Bytes transferred = 411124 (645f4 hex)
HP>>

Happy in the knowledge that my tftp server was working, i thought i would take a chance on there being some closeness in the uboot envs and registers.... no such luck, i didnt imagine it would but it was worth a shot. It just sits at the following. oh well.... now im stuck and looking for any advice on moving forward

HP>> ping 192.168.0.10
Using egiga0 device
host 192.168.0.10 is alive
HP>> tftpboot 0x0800000 u-boot.kwb
Using egiga0 device
TFTP from server 192.168.0.10; our IP address is 192.168.0.6
Filename 'u-boot.kwb'.
Load address: 0x800000
Loading: #################################################################
	 ################
done
Bytes transferred = 411124 (645f4 hex)
HP>> go 0x800200
## Starting application at 0x00800200 ...

Right now im off to wrestle with the graphical output,....



Edited 4 time(s). Last edit at 03/05/2016 08:47AM by Gravelrash.

Gravelrash,

The image name does not matter.

> Ok so i spannered the config settings on my tftp
> server and forgot to amend them a while ago.

> Happy in the knowledge that my tftp server was
> working, i thought i would take a chance on there
> being some closeness in the uboot envs and
> registers.... no such luck, i didnt imagine it
> would but it was worth a shot. It just sits at the
> following. oh well.... now im stuck and looking
> for any advice on moving forward

> HP>> tftpboot 0x0800000 u-boot.kwb
> HP>> go 0x800200

Yes, you've confirmed that it will not work tftp or ex2load the new u-boot image (ext2load would be the same as tftp). The current u-boot is too old. But we probably can chainboot from the new u-boot back to the old one (this is a boot-to-stock setup that I like to try in the future after the new u-boot works).

I think there is another way you can test the new u-boot. But it will render your box unusable for booting to Debian for a while. Let me think about it a little more.

Since you have 3, have you tried connecting serial converter to one T5325 box and kwboot the other using my built kwboot binary? I think it is very odd that you can't kwboot. The bootROM is highly unlikely to be different between yours and CV's. So I think it is the binary that you used causing problem with kwboot.

Have you tried warm start kwboot? (start kwboot and then "shutdown -r now" the test box in Debian). Cold start is to prove that you can recover from bad flash. But warm start is just as good for testing.

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)



Edited 1 time(s). Last edit at 03/05/2016 04:23PM by bodhi.

bodhi Wrote:
-------------------------------------------------------

> But we probably can chainboot from the new u-boot
> back to the old one (this is a boot-to-stock setup
> that I like to try in the future after the new
> u-boot works).

very happy to "guinea pig" this for you - im down to 2 boxes as of early last week. 1 was "liberated"

> I think there is another way you can test the new
> u-boot. But it will render your box unusable for
> booting to Debian for a while. Let me think about
> it a little more.

Happy to test this out!

>
> Since you have 3, have you tried connecting serial
> converter to one T5325 box and kwboot the other
> using my built kwboot binary? I think it is very
> odd that you can't kwboot. The bootROM is highly
> unlikely to be different between yours and CV's.
> So I think it is the binary that you used causing
> problem with kwboot.

can you post me a link to a precompiled arm binary for this platform - i have arm5 / 6 / 7 units i can load it on to use. i would love to prove either way if its my hardware or the binary i used - ignore that i found it

>
> Have you tried warm start kwboot? (start kwboot
> and then "shutdown -r now" the test box in
> Debian).

Done all the permutations i could think of.



Edited 1 time(s). Last edit at 03/05/2016 05:09PM by Gravelrash.