PM9A3 Firmware / ERRORMOD related to fw version

Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.

joerambo

New Member
Aug 30, 2023
15
1
3
Look in the Repository, i found it yesterday on a Website from H3C... jxtc404q
From mjshark is too alot of the files, many thanks.
WOW, this forum and its users are incredible value. Thanks for nothing Samsung!

The steps i took to update FW were:
downloaded latest Samsung toolkit 3 Linux version as file samsung-tool

chmod +x samsung-tool
./ samsung-tool -L

| 0 | /dev/sda | SAMSUNG MZ7L37T6HBLA-00A07 | XXXXXXXXXXX | JXTC304Q | N/A | 7153 GB | GOOD | 0.00 TB | N/A |

Your drives will be listed and then execute dor desired disk number:

./samsung-tool --disk 0 -F -p JXTC404Q_PM893.bin

to flash drive wanted.

Worked perfectly, but as always, have backup at hand as Your mileage might vary.
 
  • Like
Reactions: Ezy

mindio

New Member
Feb 28, 2023
7
3
3
You have a „not Samsung OEM“ branded PM983.
Thus you will not have success using DC-toolkit.
You might try NVME-cli tools on Linux but I think the firmware will not fit.
-00FB suffix is probably Facebook flavor. Only -0007 suffix is „Vanilla Samsung OEM“ and only those accepting firmware flashing with DC toolkit.
Getting back to trying resurrect my bricked SAMSUNG MZ1LB1T9HALS-000FB. Attempting:


sudo nvme fw-download -f EDA5900Q_nonSED.bin /dev/nvme0
Firmware download success

sudo nvme fw-commit -s 2 -a 0 /dev/nvme0
NVMe status: INVALID_FIELD: A reserved coded value or an unsupported value in a defined field(0x2002)
Multiple Update Detected (MUD) Value: 46
Detected an overlapping firmware/boot partition image update command sequence due to processing a command from an Admin SQ on a controller

So probably low chance to change this custom FW to any version of vanilla one?
Are there any ideas to at least unbrick current FW from ERRORMOD with nvme-cli?

Thanks!
 

mindio

New Member
Feb 28, 2023
7
3
3
Getting back to trying resurrect my bricked SAMSUNG MZ1LB1T9HALS-000FB. Attempting:

sudo nvme fw-download -f EDA5900Q_nonSED.bin /dev/nvme0
Firmware download success

sudo nvme fw-commit -s 2 -a 0 /dev/nvme0
NVMe status: INVALID_FIELD: A reserved coded value or an unsupported value in a defined field(0x2002)
Multiple Update Detected (MUD) Value: 46
Detected an overlapping firmware/boot partition image update command sequence due to processing a command from an Admin SQ on a controller

So probably low chance to change this custom FW to any version of vanilla one?
Are there any ideas to at least unbrick current FW from ERRORMOD with nvme-cli?

Thanks!
It seems I am trying to upload U.2 firmware on m.2 PM983 drive. Does anyone know where to get any version of MZ1LB1T9HALS-00007 firmware?
 

DarkServant

Member
Apr 5, 2022
53
58
18
It seems I am trying to upload U.2 firmware on m.2 PM983 drive. Does anyone know where to get any version of MZ1LB1T9HALS-00007 firmware?
Added Firmware for m.2 pm983 drives to the the repository.
Look always for the first four digits, which represents the hardware revision/type, after this comes the f/w revision or version from 0 to F, next the vendor code, then probably some capabilities like SED/non-SED, lastly Q for ...Q?
An EDA5**** is wrong for an EDA7**** or EDB5****. But i think it will not accept it anyway because of the vendor-lock.
In the SSD Manager software can be an option to get drives out of errormod, i think i saw something last time i used it with parameter /? or -help.
 

DarkServant

Member
Apr 5, 2022
53
58
18
Cool, thanks a lot!

Edit: can’t see the PM983 folder?
Edit edit: need new glasses…
Easy to overlook by all this model names "pm983, pm883a, pm893, pm893a, pm983a, etc.".
Yeah even the pm893/897 got an "a" upgrade, there is some kind of a new v6 Prime V-NAND out with 133-layers. I'm surprised that the v7 176-layer and 236-layer v8 has not been seen in any enterprise/datacenter SSD yet.... ;)

BTW: i put my glasses away long time ago and i still see too much of this world.
 

mindio

New Member
Feb 28, 2023
7
3
3
Added Firmware for m.2 pm983 drives to the the repository.
Look always for the first four digits, which represents the hardware revision/type, after this comes the f/w revision or version from 0 to F, next the vendor code, then probably some capabilities like SED/non-SED, lastly Q for ...Q?
An EDA5**** is wrong for an EDA7**** or EDB5****. But i think it will not accept it anyway because of the vendor-lock.
In the SSD Manager software can be an option to get drives out of errormod, i think i saw something last time i used it with parameter /? or -help.
Thanks for the FW! But it didn't work, same error... :/
Which exactly SSD Manager software do you mean?
 

DarkServant

Member
Apr 5, 2022
53
58
18
Thanks for the FW! But it didn't work, same error... :/
Which exactly SSD Manager software do you mean?
There i found some output with Magician DC v2.0 parameter -H:

-V [ --vendor-utility ] Enables the user to execute Vendor Utility command on specified disk
Example:
magician --disk 1 --vendor-utility --FAlog-dump [or] magician -d 1 -V -fa
magician --disk 1 --vendor-utility --PLP-log [or] magician -d 1 -V -p
magician --disk 1 --vendor-utility --Escape [or] magician -d 1 -V -e
magician --disk 1 --vendor-utility --DSLR [or] magician -d 1 -V -dslr

-d [ --disk ] Disk-Number of the disk to execute Vendor Utility command.
-fa [ --FAlog-dump ] Enables the user to extract the log data from a coreview block of the SSD.
-p [ --PLP-log ] Enables the user to extract the PLP log data from coreview block of the SSD.
-e [ --Escape ] Attempts to escape the device from the ERROR MODE state.
-dslr [ --DSLR ] Extracts DSLR from the SSD.

BUT this tool seems to be for ATA only, and i don't know if the -V -e works on the other NVMe Tool...
Tried to issue an format command to the drive with nvme-cli?
 

Joel

Active Member
Jan 30, 2015
864
209
43
43
There i found some output with Magician DC v2.0 parameter -H:

-V [ --vendor-utility ] Enables the user to execute Vendor Utility command on specified disk
Example:
magician --disk 1 --vendor-utility --FAlog-dump [or] magician -d 1 -V -fa
magician --disk 1 --vendor-utility --PLP-log [or] magician -d 1 -V -p
magician --disk 1 --vendor-utility --Escape [or] magician -d 1 -V -e
magician --disk 1 --vendor-utility --DSLR [or] magician -d 1 -V -dslr

-d [ --disk ] Disk-Number of the disk to execute Vendor Utility command.
-fa [ --FAlog-dump ] Enables the user to extract the log data from a coreview block of the SSD.
-p [ --PLP-log ] Enables the user to extract the PLP log data from coreview block of the SSD.
-e [ --Escape ] Attempts to escape the device from the ERROR MODE state.
-dslr [ --DSLR ] Extracts DSLR from the SSD.

BUT this tool seems to be for ATA only, and i don't know if the -V -e works on the other NVMe Tool...
Tried to issue an format command to the drive with nvme-cli?
Which software are you using? I'm trying to unbrick an ERRORMOD PM983 using Samsung DC Toolkit version 2.1 (version 3 fails on my machine with a SEGFAULT). Vendor utils do not show an escape option for me.


Code:
 -V [ --vendor-utility           ] Execute Vendor Unique command on specified disk

    -d [ --disk ]
    Disk-Number of the disk to execute Vendor Utility command.

    -fa [ --FAlog-dump ]
    Extract the log data from a coreview block of the SSD.

    -plp [ --PLP-log ]
    Enables the user to extract the PLP log data from a coreview block of the SSD.

    -dslr [ --DSLR ]
    Extracts DSLR from the SSD.

    -p [ --path ]
    Output path to make output file
 

mindio

New Member
Feb 28, 2023
7
3
3
BUT this tool seems to be for ATA only, and i don't know if the -V -e works on the other NVMe Tool...
Tried to issue an format command to the drive with nvme-cli?
I managed to unbrick the disk with simple: nvme reset command! :)
vanilla firmwares wouldn't still upgrade, but the disk is at least usable again and showing correct capacity.
 

Joel

Active Member
Jan 30, 2015
864
209
43
43
Thank you! Worked like a charm. I was able to update the firmware using the Samsung tool afterwards, but AFAICT my drive wasn't vendor-specific.
 

DarkServant

Member
Apr 5, 2022
53
58
18
nice that it worked, another drive which doesn't land in the trash and instead do it's intended job. The samsung tool is anyway based on nvme-cli... which can too be used to set the sector size to native 4k.
 

foureight84

Well-Known Member
Jun 26, 2018
281
259
63
Hmm. I just noticed one of my PM983 is running at degraded x2 Gen 1. Looking at lspci -vv, it starts out at x4 gen 3 but then degrades to x2 gen1. Faulty drive?

1715019384645.png

Played around with swapping PCIe slots and it's now running at PCIe Gen 3 x2 but still not full speed. I've tried just a single m.2 slot to PCIE but that doesn't seem to make a difference since it still degrades to PCIe Gen 1 x2. Really odd stuff. It was already on the newest firmware.
 
Last edited:

dante4

Member
Jul 8, 2021
60
10
8
Hmm. I just noticed one of my PM983 is running at degraded x2 Gen 1. Looking at lspci -vv, it starts out at x4 gen 3 but then degrades to x2 gen1. Faulty drive?

Played around with swapping PCIe slots and it's now running at PCIe Gen 3 x2 but still not full speed. I've tried just a single m.2 slot to PCIE but that doesn't seem to make a difference since it still degrades to PCIe Gen 1 x2. Really odd stuff. It was already on the newest firmware.
How many pcie drives are in your system? What motherboard is used?
If it downgrades to pcie 1.0 x2, that means training fails to complete in time, it's M.2 version?
 

foureight84

Well-Known Member
Jun 26, 2018
281
259
63
How many pcie drives are in your system? What motherboard is used?
If it downgrades to pcie 1.0 x2, that means training fails to complete in time, it's M.2 version?
I have 6 total. But it shouldn't be an issue. It's in an x299 WS Sage with an I-9920x and the drives are in a HighPoint SSD7103 raid card sitting in a x16 slot that's connected to a PLX chip. These larger drives replaced smaller drives that were in the same HighPoint SSD raid card as before. Those other drives ran at full speed without issues. I've also tried installing this drive by itself in an m 2 slot and also an m.2 pcie riser and both yielded the same degraded speed. Ultimately, I saw the drive managing Pcie 3 x 2 in the same HighPoint Raid card toward the end of my test, but never reached it's original pcie 3 x4 speed.
 

dante4

Member
Jul 8, 2021
60
10
8
I have 6 total. But it shouldn't be an issue. It's in an x299 WS Sage with an I-9920x and the drives are in a HighPoint SSD7103 raid card sitting in a x16 slot that's connected to a PLX chip. These larger drives replaced smaller drives that were in the same HighPoint SSD raid card as before. Those other drives ran at full speed without issues. I've also tried installing this drive by itself in an m 2 slot and also an m.2 pcie riser and both yielded the same degraded speed. Ultimately, I saw the drive managing Pcie 3 x 2 in the same HighPoint Raid card toward the end of my test, but never reached it's original pcie 3 x4 speed.
I would test it on different hardware system. I had similar problem when I tried to connect some NVMe U.2 drives to Rock 5B via m.2 --> pcie adapter
 

Joel

Active Member
Jan 30, 2015
864
209
43
43
Well here I am again... silly me didn't write down which drive had the issue before but my pool is down a drive AGAIN with the same issue.

Code:
sudo nvme reset /dev/ng3n1 
Reset: Inappropriate ioctl for device
So this time around a simple nvme reset didn't work, do I need to offline the drive somehow?

Note to future me: This time it's 5829.