TinyCore RedPill Loader Build Support Tool ( M-Shell )

[ansh]

Hi

DELL PowerEdge 2950 - 6 disks.

Used tinycore-redpill.v095.m-shell bootloader.

DSM 7.2 DS3622xs+ shows disks with disk2.

 

 

root@syn-dmz-beta:/dev# parted -l 2> /dev/null |grep "^Disk /dev/s"
Disk /dev/sdb: 299GB
Disk /dev/sdc: 299GB
Disk /dev/sdd: 299GB
Disk /dev/sde: 299GB
Disk /dev/sdf: 299GB
Disk /dev/sdg: 299GB
root@syn-dmz-beta:/dev#

 

Probably Disk1 is a USB flash drive with a bootloader.
How to start disk numbering from /dev/sda?

dmesg.txt

Edited November 17, 2023 by ansh

[maxime]

17 minutes ago, Peter Suh said:

 

You did not explain the purpose of updating M-SHELL.
For TCRP, TTYD is not required during normal update process.

Thanks Peter, I'm sorry if I did not well explain. For example, I would know how to remotly upgrade the loader for a new DSM version, without connect display and keyboard to my XPE server

[Peter Suh]

13 minutes ago, maxime said:

Thanks Peter, I'm sorry if I did not well explain. For example, I would know how to remotly upgrade the loader for a new DSM version, without connect display and keyboard to my XPE server

 

Task Scheduler includes an addon called RebootToTcrp.
If you run this scheduler manually, the order of the Grub boot menu will be changed, so when rebooting, you will enter the TCRP loader build menu instead of booting to FRIEND.

This Tinycore Linux comes with an SSH daemon already running.
You can remotely access ssh using the tc / P@ssw0rd account and password.
(If possible, it is recommended to fix the MAC address on the router so that it can receive a fixed IP from DHCP.)
Now you can run ./menu.sh and use the same menu that appears in the upper left corner that you normally use.
Rebuild the loader with the desired DSM revision and select Reboot.
There should not be a single mistake, such as shutting down, when selecting the reboot menu.

When DSM of a new revision is booted through FRIEND, you will see the update screen for the corresponding revision when connected through port 5000.
Just proceed as instructed.

[Bose321]

Just tried to update from 69057 to .1 version but it doesn't seem to be booting anymore.

 

It is looping with these messages:

Quote

Running "install.sh" for dbgutils->on_jrExit->jrExit
mount: mounting /dev/synoboot1 on /mnt failed: Invalid argument
umount: can't unmount /mnt: Invalid argument
Ran "install.sh" for dbgutils->on_jrExit->jrExit - exit=1
ERROR: Some of the scripts failed! See above for any with exit != 0
:: Executing "on_jrExit" custom scripts ... [  OK  ]
Exit on error [99] jrExit ext init exec failure...
Mon Nov 20 12:55:03 UTC 2023
:: Loading kernel modules from extensions ...
:: Loading kernel modules from extensions ... [  OK  ]
:: Executing "on_jrExit" "jrExit" custom scripts ...

I tried to boot into the loader and run my.sh postupdate but it didn't do anything. What else can I try?

 

Just rebuilt the loader. Seems to work again. I thought this loader eliminated the need of rebuilding loaders the whole time and having to mess around?

Edited November 20, 2023 by Bose321

[Peter Suh]

15 hours ago, Bose321 said:

Just tried to update from 69057 to .1 version but it doesn't seem to be booting anymore.

 

It is looping with these messages:

I tried to boot into the loader and run my.sh postupdate but it didn't do anything. What else can I try?

 

Just rebuilt the loader. Seems to work again. I thought this loader eliminated the need of rebuilding loaders the whole time and having to mess around?

 

Because of related issues
There was an adjustment made 3 weeks ago to prevent the jrExit() event from working.
Your rebuild may have excluded jrExit as it applies a new script.

 

https://github.com/PeterSuh-Q3/redpill-load/commit/a1a3aecf614b218ef40ce4be30cfa9cd6c027b6b

 

https://github.com/PeterSuh-Q3/tcrp-addons/commit/9ff535cb637d285f9589b7edd452bc477b0c1518

 

In these two scripts
The time when the first script was distributed and
There is a difference of approximately 11 minutes between the distribution times of the second script.

Perhaps you were unlucky enough to receive and use only the first script within these 11 minutes.

 

The jrExit error occurred from the moment the first script was deployed.
The second script again prevented the error.

Edited November 21, 2023 by Peter Suh

[mealan]

On 9/4/2023 at 8:03 PM, Peter Suh said:

 

Please let me know for reference which Synology model you are trying and which DSM version is targeted.

Is version 7.0 still the target?

 

The vmw_pvscsi module used by SAS controller 15add000007c0 seems to have a different TCRP version than the version used in ARC.

{
     "name": "vmw_pvscsi",
     "alias": "pci:v000015ADd000007C0sv*sd*bc*sc*i*"
},

 

In fact, this module was compiled yesterday as well.

 

https://github.com/PeterSuh-Q3/arpl-modules

 

What I am trying now is to make the HBA operate in a Device-Tree based model.
The compilation target is actually the mpt3sas module, but this module is being compiled incidentally.

 

This is an attempt to find the disk on the SAS controller.
mpt3sas is still being tested. Please wait a moment.

Trying on vm hardware version 20, VMware ESXi, 8.0.1, 22088125, still has no drives connected. ARC is working but your loader is nicer. What can I do to help? It detects them here, but doesn't show green or red. Thank you!

[Peter Suh]

15 minutes ago, mealan said:

Trying on vm hardware version 20, VMware ESXi, 8.0.1, 22088125, still has no drives connected. ARC is working but your loader is nicer. What can I do to help? It detects them here, but doesn't show green or red. Thank you!

 

Can you take a screenshot of the overall settings of your esxi?
I hope you can check all the details related to SATA / SCSI / SAS controllers.

Are the physical disks only mounted on the SATA controller?

[mealan]

Just to validate things I made multiple scsi controllers of all different types and another sata controller. sadly none of them appear "green" each hard disk is attached a different controller just to try to get one of them to show up. What's odd is that they show up in the drives as above. but after the mapping and building the loader get the "no drives detected" section

[mealan]

This seems to be a VM hardware version issues, hardware version 17 has sata in green, however pvscsi in grey with no drive connected, and compiling / booting ends up with no drives connected in dsm. So i do believe it's got to do with the sata / scsi mapping. With adding any scsi controller (pvscsi would be preferred), even with no disks attached, everything is grey again. Remove the controller and boot back so sata only and  things are back to green and happy

[Peter Suh]

2 hours ago, mealan said:

This seems to be a VM hardware version issues, hardware version 17 has sata in green, however pvscsi in grey with no drive connected, and compiling / booting ends up with no drives connected in dsm. So i do believe it's got to do with the sata / scsi mapping. With adding any scsi controller (pvscsi would be preferred), even with no disks attached, everything is grey again. Remove the controller and boot back so sata only and  things are back to green and happy

 

 

ARC may be different, but TCRP is not a commercial program.
There is no obligation to operate flexibly in environments where various and harsh controllers are used.

As far as I know, in Linux, the port number per controller must not exceed 26, which is the number of alphabets, to be recognized.
The environment I use is Esxi 7.0.2, hardware version 19.

As you can see in my capture, the 4 disks subordinate to one SATA controller are located within the 26 alphabets without any problems and are all recognized normally.
However, when I extend and add controllers like you, I experience the same symptoms.
Focus on the one controller you actually want to use and subordinate the disks to it.

 

 

 

 

 

[Nox71]

Hi all, after many days of trying and failing now i need some help from the community. I have a custom built server with an Asus WS motherboard with an Intel Core i7-13700K CPU. The motherboard come with 2x Intel 2.5 GB network card. I have tried building a loader with the DS3622xs+ model and when booting the loader everything seem to go fine but i can't see or ping the server after that. During booting i have access to the web interface on port 7681 and after the DSM start booting i am loosing the access and that't it i can't find the server with the Synology online tools, i can't ping it and i can't acces the web interface on port 5000.

 

Here is a screenshot of the web interface during booting of the loader:

 

 And also a screenshot of the server:

 

The motherboard is an Asus WS W680 ACE IPMI with an Intel Core i7 13700K cpu with 128 GB of RAM. The motherboard come with 2 x Intel 2.5 GB network card.

 

Any help will be appreciated. Thanks.

 

[Peter Suh]

49 minutes ago, Nox71 said:

Hi all, after many days of trying and failing now i need some help from the community. I have a custom built server with an Asus WS motherboard with an Intel Core i7-13700K CPU. The motherboard come with 2x Intel 2.5 GB network card. I have tried building a loader with the DS3622xs+ model and when booting the loader everything seem to go fine but i can't see or ping the server after that. During booting i have access to the web interface on port 7681 and after the DSM start booting i am loosing the access and that't it i can't find the server with the Synology online tools, i can't ping it and i can't acces the web interface on port 5000.

 

Here is a screenshot of the web interface during booting of the loader:

 

 And also a screenshot of the server:

 

The motherboard is an Asus WS W680 ACE IPMI with an Intel Core i7 13700K cpu with 128 GB of RAM. The motherboard come with 2 x Intel 2.5 GB network card.

 

Any help will be appreciated. Thanks.

 

 

If UART is enabled in BIOS, disable it.
If the problem persists, go back to the TCRP Loader Build menu and reattach the log shown using the menu below.
It contains logs that can help diagnose problems.

 

"Show Error Log of Running Loader"

 

And usually errors are shown at the bottom of this log.
Use the down arrow key to scroll to the last content.

Edited November 23, 2023 by Peter Suh

[mealan]

15 hours ago, Peter Suh said:

 

 

ARC may be different, but TCRP is not a commercial program.
There is no obligation to operate flexibly in environments where various and harsh controllers are used.

As far as I know, in Linux, the port number per controller must not exceed 26, which is the number of alphabets, to be recognized.
The environment I use is Esxi 7.0.2, hardware version 19.

As you can see in my capture, the 4 disks subordinate to one SATA controller are located within the 26 alphabets without any problems and are all recognized normally.
However, when I extend and add controllers like you, I experience the same symptoms.
Focus on the one controller you actually want to use and subordinate the disks to it.

 

 

 

 

 

 

 

 

 

More than fair, I thought there was a requirement to have a boot device on a seperate first sata controller but that time must have past. I will use sata and stress test and validate it's good for me. Was merely trying to keep things optimized by using the pv scsi controller. 

 

THANK YOU for everything you're doing, it's great work and very useful. 

[Nox71]

3 hours ago, Peter Suh said:

 

If UART is enabled in BIOS, disable it.
If the problem persists, go back to the TCRP Loader Build menu and reattach the log shown using the menu below.
It contains logs that can help diagnose problems.

 

"Show Error Log of Running Loader"

 

And usually errors are shown at the bottom of this log.
Use the down arrow key to scroll to the last content.

Ok i have checked and didn't find anything related to UART in the bios.

Is there a way to load drivers for the Intel 2.5 gb mic into TCRP because everytime i need to do a change i have to plusg a USB Ethernet adapter to get internet access ?

[Peter Suh]

35 minutes ago, Nox71 said:

Ok i have checked and didn't find anything related to UART in the bios.

Is there a way to load drivers for the Intel 2.5 gb mic into TCRP because everytime i need to do a change i have to plusg a USB Ethernet adapter to get internet access ?

 

I think your 2.5G Intel NIC chipset is i225 or i226, but there is an issue where it does not work perfectly in tinycore linux 12. After the bootloader is built, the igc module, which operates as a loader extension module, requires active feedback and verification. This is because recently, there was a request for improvement from a Korean i226 user who was experiencing the same phenomenon as you, and there was a module version upgrade related to this. I imported the latest igc module from rr. I have mobos that use i225, but they pass without any issues, so it is difficult to reproduce the situation. (This means that there is no environment for debugging.)

 

 

By the way, did you not obtain the log I requested? This might be a starting point for the debugging mentioned above.

 

[Nox71]

54 minutes ago, Peter Suh said:

 

I think your 2.5G Intel NIC chipset is i225 or i226, but there is an issue where it does not work perfectly in tinycore linux 12. After the bootloader is built, the igc module, which operates as a loader extension module, requires active feedback and verification. This is because recently, there was a request for improvement from a Korean i226 user who was experiencing the same phenomenon as you, and there was a module version upgrade related to this. I imported the latest igc module from rr. I have mobos that use i225, but they pass without any issues, so it is difficult to reproduce the situation. (This means that there is no environment for debugging.)

 

 

By the way, did you not obtain the log I requested? This might be a starting point for the debugging mentioned above.

 

I also have some Intel dual 10 Gigabit network cards i can test if you think that there is better chance of success, those were removed from and old Dell server. I will send the logs later today since i am at work now.

Thanks.

[Nox71]

4 hours ago, Peter Suh said:

 

I think your 2.5G Intel NIC chipset is i225 or i226, but there is an issue where it does not work perfectly in tinycore linux 12. After the bootloader is built, the igc module, which operates as a loader extension module, requires active feedback and verification. This is because recently, there was a request for improvement from a Korean i226 user who was experiencing the same phenomenon as you, and there was a module version upgrade related to this. I imported the latest igc module from rr. I have mobos that use i225, but they pass without any issues, so it is difficult to reproduce the situation. (This means that there is no environment for debugging.)

 

 

By the way, did you not obtain the log I requested? This might be a starting point for the debugging mentioned above.

 

I get a file not found error when selecting this opton in the menu.

[Nox71]

22 hours ago, Peter Suh said:

 

I think your 2.5G Intel NIC chipset is i225 or i226, but there is an issue where it does not work perfectly in tinycore linux 12. After the bootloader is built, the igc module, which operates as a loader extension module, requires active feedback and verification. This is because recently, there was a request for improvement from a Korean i226 user who was experiencing the same phenomenon as you, and there was a module version upgrade related to this. I imported the latest igc module from rr. I have mobos that use i225, but they pass without any issues, so it is difficult to reproduce the situation. (This means that there is no environment for debugging.)

 

 

By the way, did you not obtain the log I requested? This might be a starting point for the debugging mentioned above.

 

Om i am getting somewhere, just redone everything on another USB key and plug it on another USB port on the server, now i can see two devices since it get two different ip address on the two network cards but as you can see there is no MAC on Serial number on each details of both devices. If i try to configure one i get a time out in CHrome:

 

If i reboot into the build menu i now have logs:

 

[Peter Suh]

45 minutes ago, Nox71 said:

Om i am getting somewhere, just redone everything on another USB key and plug it on another USB port on the server, now i can see two devices since it get two different ip address on the two network cards but as you can see there is no MAC on Serial number on each details of both devices. If i try to configure one i get a time out in CHrome:

 

If i reboot into the build menu i now have logs:

 

 

 

The process of finding and loading two ixgbe modules related to the 10g nic appears in the log. We need to find a reason why the rest of the NICs can't even try. How many physical NIC ports are there? And the RAID controller and SATA controller seem to be special. I can't find the related expansion module. What models do you know of these two controllers? Please let me know the results of running the command below in tinycore linux. lspci -tvnnq

[Nox71]

22 minutes ago, Peter Suh said:

 

 

The process of finding and loading two ixgbe modules related to the 10g nic appears in the log. We need to find a reason why the rest of the NICs can't even try. How many physical NIC ports are there? And the RAID controller and SATA controller seem to be special. I can't find the related expansion module. What models do you know of these two controllers? Please let me know the results of running the command below in tinycore linux. lspci -tvnnq

Here is the result of the command:

[Peter Suh]

11 hours ago, Nox71 said:

Here is the result of the command:

 

 

After comparing the error log you provided with the information from lspci,
It seems that the currently used DDSML mode does not cover the VID+PID of the 12th/13th generation devices.
I will improve the device compatibility of DDSML by reinforcing the modules.alias.4.json file.

I suggest switching to EUDEV, which uses dynamic mapping, instead of DDSML, which uses static module mapping.
It exists as the first menu in the TCRP loader build menu.
After switching to DDSML -> EUDEV, rebuild the loader and check again whether the NIC or SATA controller is recognized.

If the problem still repeats, we will guide you again on how to check the error log. The case of EUDEV may be slightly different.

Edited November 25, 2023 by Peter Suh

[BartAcid]

On my HP Z240 DVA1622 7.2.1 update 3 applied smoothly.

[Captainfingerbang]

12 hours ago, Peter Suh said:

 

 

After comparing the error log you provided with the information from lspci,
It seems that the currently used DDSML mode does not cover the VID+PID of the 12th/13th generation devices.
I will improve the device compatibility of DDSML by reinforcing the modules.alias.4.json file.

I suggest switching to EUDEV, which uses dynamic mapping, instead of DDSML, which uses static module mapping.
It exists as the first menu in the TCRP loader build menu.
After switching to DDSML -> EUDEV, rebuild the loader and check again whether the NIC or SATA controller is recognized.

If the problem still repeats, we will guide you again on how to check the error log. The case of EUDEV may be slightly different.

 

 

Hi peter, i came back here specifically to ask you these questions OR ANYONE WHO CAN HELP because im running your software TCRP FRIEND.

I have a problem, that my cpu is much slower than i prefer.. I bought Xeon E5 2680v4 Broadwell. Which is fast , BUT because i am using 918+ because its my only real serial/mac there is an 8 core limit i believe, still. So in bios i have to disable cores and hyperthreading because i have a total of 28 cores/threads.. Total waste of the Xeon.

 

I have an extra motherboard and cpu i could swap.

What would be faster? An I3-8100 + Gigabyte B360 HD3 OR  E5 2680 v4 in X99 but only using 8 slow cores?

I3-8100 has         4 total threads at 3.60 GHz

My current Xpen is E5 2680 v4 has 28 total threads but can only use 8 cores on 918+ and only 1.2Ghz from what dsm says 

 

Which would be faster?

 

 

Lastly: Can anyone Please tell me if i am using m-shell by looking at my main screen below? 

SHOULD i be using mshell? Im currently using TCRP FRIEND LATEST I THINK

 

 

Edited November 25, 2023 by Captainfingerbang

[Peter Suh]

 
 
Hi peter, i came back here specifically to ask you these questions OR ANYONE WHO CAN HELP because im running your software TCRP FRIEND.
I have a problem, that my cpu is much slower than i prefer.. I bought Xeon E5 2680v4 Broadwell. Which is fast , BUT because i am using 918+ because its my only real serial/mac there is an 8 core limit i believe, still. So in bios i have to disable cores and hyperthreading because i have a total of 28 cores/threads.. Total waste of the Xeon.
 
I have an extra motherboard and cpu i could swap.
What would be faster? An I3-8100 + Gigabyte B360 HD3 OR  E5 2680 v4 in X99 but only using 8 slow cores?
I3-8100 has         4 total threads at 3.60 GHz
My current Xpen is E5 2680 v4 has 28 total threads but can only use 8 cores on 918+ and only 1.2Ghz from what dsm says 
 
Which would be faster?
 
 
Lastly: Can anyone Please tell me if i am using m-shell by looking at my main screen below? 
SHOULD i be using mshell? Im currently using TCRP FRIEND LATEST I THINK
 

 

The reason why the CPU speed appears low is because the CPU frequency scalable add-on is running. This add-on lowers the frequency during idle timing to save power. In the opposite case, when the CPU starts operating, the frequency increases and goes up to the maximum allowable value. A quick search of the forums will find a q/a on this issue. However, one bug is that, as you captured, only the minimum speed at idle is displayed and the changing increased speed cannot be displayed. Did you run into problems if you didn't set thread limits on your Xeon? If there are no particular problems, the performance of Xeon, which can utilize all threads, is likely to be superior.


Sent from my iPhone using Tapatalk

[Peter Suh]

[NOTICE]

TCRP-MSHELL SA6400 i915 transcoding stabilization notice

 

A few days ago, when updating SA6400-7.2.1 U3, there was a fatal error when automatically patching the RAM disk, so I fixed it.

There was one more error left.

When automatic updates occur, not only ramdisk patches but also kernel patches occur.

 

The kernel patch script for kernel 5 was in the friend tools folder developed by @pocopico.

This script was not complete. So there was a problem that resulted in a kernel panic.

 

Stabilize the same by referring to ARPL's patch script.

Today I upgraded the FRIEND kernel to version 0.0.9i.

 

 

Currently, this version is verified by @MoetaYuko for Intel 4th, 11th, and 13th generations (a stable version that replaced the existing @jimmmmm version)

It has been replaced with the i915.ko file.

You can apply it by rebuilding the TCRP-mshell loader.

 

It seems that all processors from the 4th generation of Intel that were previously capable of transcoding will be capable of transcoding.

I think users should verify it one by one and give feedback.

 

Status of automatic update up to SA6400-7.2.1 U3 and presence of /dev/dri folder

This confirms the stable loading status of the i915 module.

 

 

 

I haven't been able to check the HW transcoding message because I don't have a plex pass yet, but if you have a pass, (HW) will likely appear.

I confirmed that even with SW transcoding, it played well while maintaining a CPU load of 7%.

 

 

No additional activation processing for transcoding, such as installing AME, etc. was done.

The serial number is also a virtual serial number randomly generated by TCRP.

Edited November 26, 2023 by Peter Suh