Debian on Dell Kace M300

//note, I posted twice in a row, see my last post post on the previous page for the fail I had testing M300 #5.

I need to fix the uImage for the bodhi kernel I'm using, apparently I made it with the 0x10008000 value I noted above:

## Booting image at 00100000 ...
   Image Name:   Linux-5.5.1-kirkwood-tld-1
   Created:      2020-04-10  15:36:18 UTC
   Image Type:   ARM Linux Kernel Image (uncompressed)
   Data Size:    5213369 Bytes =  5 MB
   Load Address: 10008000
   Entry Point:  10008000
   Verifying Checksum ... OK
OK
## Loading Ramdisk Image at 01100000 ...
   Image Name:   uInitrd
   Created:      2020-02-06  22:11:51 UTC
   Image Type:   ARM Linux RAMDisk Image (uncompressed)
   Data Size:    100 Bytes =  0.1 kB
   Load Address: 00000000
   Entry Point:  00000000
   Verifying Checksum ... OK

Starting kernel ...

Uncompressing Linux... done, booting the kernel.
[    0.000000][    T0] Booting Linux on physical CPU 0x0
[    0.000000][    T0] Linux version 5.5.1-kirkwood-tld-1 (root@tldDebian) (gcc version 8.3.0 (Debian 8.3.0-6)) #1.0 PREEMPT Sat Feb 1 22:28:36 PST 2020
[    0.000000][    T0] CPU: Feroceon 88FR131 [56251311] revision 1 (ARMv5TE), cr=0005397f
[    0.000000][    T0] CPU: VIVT data cache, VIVT instruction cache
[    0.000000][    T0] OF: fdt: Machine model: Dell KACE M300
[    0.000000][    T0] OF: fdt: Ignoring memory range 0x0 - 0x10000000
[    0.000000][    T0] printk: bootconsole [earlycon0] enabled
[    0.000000][    T0] Memory policy: Data cache writeback
[    0.000000][    T0] INITRD: 0x01100000+0x00001000 is not a memory region - disabling initrd
[    0.000000][    T0] Built 1 zonelists, mobility grouping on.  Total pages: 455168
[    0.000000][    T0] Kernel command line: console=ttyS0,115200 mtdparts=spi_flash:0x7f000@0(uboot),0x1000@0x7f000(u-boot-env) earlyprintk root=/dev/sda2
[    0.000000][    T0] Dentry cache hash table entries: 262144 (order: 8, 1048576 bytes, linear)
[    0.000000][    T0] Inode-cache hash table entries: 131072 (order: 7, 524288 bytes, linear)
[    0.000000][    T0] mem auto-init: stack:off, heap alloc:on, heap free:off
[    0.000000][    T0] Memory: 1802436K/1835008K available (9216K kernel code, 926K rwdata, 3860K rodata, 1024K init, 296K bss, 32572K reserved, 0K cma-reserved)

Note that the line that says "Ignoring memory range 0x0 - 0x10000000" I think that goes away when I make the uImage for 0x00008000, again, not sure if this is really a big deal??



Edited 1 time(s). Last edit at 04/13/2020 04:05PM by sodface.

sodface,

> Note that the line that says "Ignoring memory
> range 0x0 - 0x10000000" I think that goes away
> when I make the uImage for 0x00008000, again, not
> sure if this is really a big deal??

You do need to redo uImage, or copy a good one over.

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

I'll fix the uImage bodhi.

I've been running off the internal sata using the factory load since yesterday and have run the transfer test script multiple times without a problem so far.

sodface,

> I've been running off the internal sata using the
> factory load since yesterday and have run the
> transfer test script multiple times without a
> problem so far.

Exactly my experience, too. I run one of my M300 boxes using internal SATA with the latest Debian rootfs.

The reason I suspect this is a power problem is: The internal SATA is just like an SD card, which needs to draw even less juice than USB.

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



Edited 1 time(s). Last edit at 04/14/2020 05:55PM by bodhi.

bodhi Wrote:
-------------------------------------------------------
> sodface,
>
> > I've been running off the internal sata using
> the
> > factory load since yesterday and have run the
> > transfer test script multiple times without a
> > problem so far.
>
> Exactly my experience, too. I run one of my M300
> boxes using internal SATA with the latest Debian
> rootfs.
>
> The reason I suspect this is a power problem is:
> The internal SATA is just like an SD card, which
> needs to draw even less juice than USB.

Finally got the chance to remove the factory SATA drive from my problem M300 and boot from USB running 5.6.5 with no hard drive installed. Ran stress-ng all night and even compiled some drivers and I experienced no issues or kernel panics. Looks like in my case it's the hard drive.



Edited 1 time(s). Last edit at 05/28/2020 09:17AM by LeggoMyEggo.

LeggoMyEggo Wrote:
-------------------------------------------------------
> Finally got the chance to remove the factory SATA
> drive from my problem M300 and boot from USB
> running 5.6.5 with no hard drive installed. Ran
> stress-ng all night and even compiled some drivers
> and I experienced no issues or kernel panics.
> Looks like in my case it's the hard drive.


I'm confused. When you say the "factory SATA", are you referring to the SSD? You think that's the culprit or did you have a mechanical hard drive? I don't understand how a SSD could cause a power problem, but I do understand how one could be bad.

renojim Wrote:
-------------------------------------------------------
> LeggoMyEggo Wrote:
> -------------------------------------------------------
> > Finally got the chance to remove the factory
> SATA
> > drive from my problem M300 and boot from USB
> > running 5.6.5 with no hard drive installed.
> Ran
> > stress-ng all night and even compiled some
> drivers
> > and I experienced no issues or kernel panics.
> > Looks like in my case it's the hard drive.
>
>
> I'm confused. When you say the "factory SATA",
> are you referring to the SSD? You think that's
> the culprit or did you have a mechanical hard
> drive? I don't understand how a SSD could cause a
> power problem, but I do understand how one could
> be bad.

I should have been more exact, I meant the factory SSD module. As soon as I booted without the factory SSD module, my problems went away.

LeggoMyEggo,

> I should have been more exact, I meant the factory
> SSD module. As soon as I booted without the
> factory SSD module, my problems went away.

That's a little worse than I thought. I think this stock u-boot has wome weird bug that the stock kernel works around it.

Because the internal SSD is really like an SD card. There is not much power used by it. The USB seems to draw more power (from info I get at stock u-boot prompt). And usually, SD cards use very little power.

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

bodhi Wrote:
-------------------------------------------------------
> There is not much power used by it. The USB
> seems to draw more power (from info I get at stock
> u-boot prompt). And usually, SD cards use very
> little power.

You may be on to something about the power supply. The power connector on the motherboard is really loose on this unit and a couple of the others I own. Even when fully inserted, it's loose. I just shut down the M300 to move it and when I reconnected power, it failed to show up on my IP checker. Turns out fiddling with the connection to the unit made it start booting. Wonder if this is causing issues. I trust the power supply itself as it is of good build quality, not a typical wall wart. But that barrel connector on the motherboard, it doesn't inspire confidence.

Wanted to post some pictures of the 128GB Microcenter SSD I mounted in the M300. I used a 12 inch 22 pin combo SATA / power extension cable to allow mounting with Velcro to the bottom cover of the case. Works great and puts the SSD at the bottom of the case further away from the heat of the SOC. Nice mod to try that doesn't require cutting or other risky activities.

Leggo,

Very nice mods! You always have a really architectural-driven mods :)

This box is a nice NAS, and could be a basic 2G/2G desktop with a USB VGA adapter too. A VESA mounted desktop. Or a portable NAS.

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

> Wanted to post some pictures of the 128GB
> Microcenter SSD I mounted in the M300. I used a
> 12 inch 22 pin combo SATA / power extension cable
> to allow mounting with Velcro to the bottom cover
> of the case. Works great and puts the SSD at the
> bottom of the case further away from the heat of
> the SOC. Nice mod to try that doesn't require
> cutting or other risky activities.

Up in the Wiki thread

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

bodhi Wrote:
-------------------------------------------------------
> Leggo,
>
> Very nice mods! You always have a really
> architectural-driven mods :)
>
> This box is a nice NAS, and could be a basic
> 2G/2G desktop with a USB VGA adapter too. A VESA
> mounted desktop. Or a portable NAS.

Bodhi,

Still working on my ultimate mod for this device, an internal ESATA port for my external port multiplier 4 bay case. Just havent figured out a way to cut the square hole needed for the SATA passthrough. Don't want to butcher this beautiful case. One of these days :)

LeggoMyEggo,

> Still working on my ultimate mod for this device,
> an internal ESATA port for my external port
> multiplier 4 bay case. Just havent figured out a
> way to cut the square hole needed for the SATA
> passthrough. Don't want to butcher this
> beautiful case. One of these days :)

If even a hardware-challenge software modder like me can learn how to drill a hole on the plastic case of the Pogo to thread the serial console wires through :)) you the master modder will figure it out :)

BTW, I did not know there is a drill bit for plastic until I decided on a rainly day to make all my Pogoplugs look somewhat decent :) Better modders like our user syong or you, would make an audio-jack connector for make serial console that looks awesome!

Long ago syong sent me his picture of a vertical stack of 10 HP T5325's (iirc), all with audio-jack connector for serial console. That was from a PM so I could not post here.

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



Edited 3 time(s). Last edit at 09/23/2020 12:18AM by bodhi.

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

> Long ago syong sent me his picture of a vertical stack of 10 HP T5325's (iirc), all with audio-jack connector for serial console. That was from a PM so I could not post here.

I remember that mod and at one point was going to implement it in all my plugs. Had the parts list and everything. But as I am fickle, found new plug devices to tinker with so the serial 3.5mm mod went to the back burner. Still, very clever mod and important if you are doing serious experimentation with kernel and u-boot.

LeggoMyEggo Wrote:
-------------------------------------------------------
> bodhi Wrote:
> -------------------------------------------------------
>
> I remember that mod and at one point was going to
> implement it in all my plugs. Had the parts list
> and everything. But as I am fickle, found new
> plug devices to tinker with so the serial 3.5mm
> mod went to the back burner. Still, very clever
> mod and important if you are doing serious
> experimentation with kernel and u-boot.

Ha, Ha, I resemble this remark. Recently ran across the bag of 3.5mm audio jacks that were intended for this use. Perhaps I'll get around to it some day!

I am just now ordering some Serial ATA Data Power Combo Extension Cables and velcro as I have a case of M300 to upgrade!

I just picked up a pair of M300's still in the factory boxes. Is there anything I can do to verify the operation of the devices before trying to replace the OS? I take it that out of the box these things are web page driven. Is it a good idea to not expose it to the internet (and letting it phone home)? I have a spare router so I guess I could make a non-internet-connected LAN and see if the device appears.

dougdeep Wrote:
-------------------------------------------------------
> I just picked up a pair of M300's still in the
> factory boxes. Is there anything I can do to
> verify the operation of the devices before trying
> to replace the OS? I take it that out of the box
> these things are web page driven. Is it a good
> idea to not expose it to the internet (and letting
> it phone home)? I have a spare router so I guess
> I could make a non-internet-connected LAN and see
> if the device appears.

Not aware of any issues with the M300's phoning home and somehow preventing loading bodhi's Debian on the box. U-boot is not locked but you really need a serial console adapter to modify the u-boot env's. I use a cheap CP2102 clone to modify the env's. If you are concerned about preserving the factory load, use Win32DiskImager (or similar) to read the entire internal SSD via a USB to 22 pin SATA adapter to a img file on your workstation. That gives you a backup to return to stock although I can't think of any reason you would do this as the service this box was associated with is long deprecated (this device came out in 2011!). Also retain a copy of the stock u-boot env's for completeness.

Thank you for the response. I just wanted to make sure there was no funny business going on like with the Sonos brick debacle. That, plus this has Dell's name plastered all over it and I'm old enough to remember Dell's blacklisting BIOS updates aimed at keeping customers from upgrading thier Optiplexes from Pentium 2 to Pentium 3.

I have two of these Kace M300 units and they work great when booted off a USB stick. The trouble I’m having is getting them to boot off the internal mSATA drive. After I got the units booting Debian and the 2Gb kernel, I tried copying the system to mSATA drive by creating a tar from the USB on another machine (as su with permissions and such also copied) and extracting it on the Kace drive after rebooting with the USB stick. I used the setenv lines from Bodhi’s post of 9/24/19 (remembering to do “protect off all” first and “saveenv” after) and labeled the mSATA as “sata_rootfs”. It starts to boot but gets stuck here:

Starting version 241
[    8.345623][  T164] input: gpio_keys as /devices/platform/gpio_keys/input/input0
Begin: Loading essential drivers ... done.
Begin: Running /scripts/init-premount ... done.
Begin: Mounting root file system ... Begin: Running /scripts/local-top ... done.
Begin: Running /scripts/local-premount ... done.
Segmentation fault
Segmentation fault

The last three lines continue over and over.

To verify that it wasn’t a faulty mSATA, I also tried it with a small conventional SATA drive and had it quit at the same point. I’m thinking that there needs to be some additional change made to the “sata_rootfs” drive besides the env setting. Any ideas would be helpful.

dougdeep,

For problem like this, you should post a full serial console bootlog (from u-boot prompt until it stops booting).

> Segmentation fault
> The last three lines continue over and over.

u-boot envs don't play any role in this type of errors. It is most likely the SATA rootfs was not created properly.

If you'd like, you could post the session when you create a tarball and restore it to the SATA drive to create the rootfs.

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

Dougdeep,

Don't forget to label the sata drive with e2label to "sata_rootfs" and you need to change the first line in /etc/fstab (on the sata drive) to sata_rootfs from rootfs. This assumes you already adjusted the uboot environment per bodhis directions.

I've made this oversight half a dozen times so I know from experience.

LME



Edited 2 time(s). Last edit at 11/20/2020 11:54AM by LeggoMyEggo.

Doug:

Deja vu all over again for me. You can look up thread to earlier this year and see my similar troubles. For fun, try plugging any old USB drive into the box and then see if it boots from mSata. That's what I had to do to get a reliable boot using the original 16GB sata card. On the other hand, another M300 boots 95% first time from a 128GB internal sata SSD.

All,

Yup, Mike is right, I forgot about this M300 problem. Wish I have a bit more time to finish building the new u-boot. It can be used for recovery purpose now (I've uploaded it here), but it is not ready (it's on my TODO list, like, forever :-).

Dell is the worst GPL abuser that I've seen.The stock u-boot apparently did some tricks to activate the USB power and seems to disable it before handing off to the kernel. On their forum, they claim that they did not change anything to GPL source, so they are not obligate to release the GPL source or the build script.

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

For fun, try plugging any old USB drive into the box and then see if it boots from mSata

Thank you, Mikeh49. Attaching a USB memory stick did indeed result in a successful startup. I know that's a workaround and not a fix but it's alright for now.

I also tried some non-memory USB sticks like a WIFI and an audio adapter and those worked as well. It seems like there needs to be a USB "something" plugged in but it doesn't care what it is. I'll try it out on my USB mood light next.

dougdeep,


> I also tried some non-memory USB sticks like a
> WIFI and an audio adapter and those worked as
> well. It seems like there needs to be a USB
> "something" plugged in but it doesn't care what it
> is. I'll try it out on my USB mood light next.

Cool! that's good info. It means there is no need to draw much voltage from the USB port, but it is a trigger of some sort.

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



Edited 1 time(s). Last edit at 11/22/2020 12:16AM by bodhi.

Glad it worked! USB mood light, perfect. A USB audio adapter could make it a very good network media player. I'll try it on mine.

Update

Booted OK with 3 different USB audio cards and with an unpowered cheap USB 2.0 hub.



Edited 2 time(s). Last edit at 11/22/2020 01:55PM by mikeh49.

Hmm... I thought I tried plugging "something" into the USB port, but it was quite a while ago and at least one of my M300s just started booting reliably from SSD on its own. Out of curiosity, if you remove the USB device does it remain stable?

For me, yes, but only limited use. My production box with an internal 128GB SSD is very reliable. I've only restarted it a few times in a couple of years, no problem. There must be something about the 16GB OEM memory card?

Has anyone connected a SATA hard drive? I tried two different ones expecting to get a huge speed increase over a USB connected drive, but it's barely faster if at all. Do I just have cheap slow drives or is there something I'm missing? I've been wanting to take my venerable Dockstar out of service and replace it with this, but so far it's been very disappointing as far as getting any kind of speed increase for anything I've tried from disk speed to processing speed.