Free P3700 in Exchange for Firmware

AJXCR

Active Member
Jan 20, 2017
565
95
28
32
As the title states, I have a free Intel 400GB P3700 (HPE) for the person that finds firmware which gets my PM963's functioning properly.

More information about the issue/firmware needed here:

2.5" 960GB PM963 Firmware
 

Evan

Well-Known Member
Jan 6, 2016
3,144
530
113
AFAIK for the PM963 you are able to use Samsung DC Magician 2.0 which should contain the firmware.
Does not contain the firmware files as far as I know, it’s just the software that can apply them, I tried to help out but by OEM contacts also don’t have the firmware, vendor in question seems to be a specific version put on their drives from Samsung and that’s it.
 

AJXCR

Active Member
Jan 20, 2017
565
95
28
32
Does not contain the firmware files as far as I know, it’s just the software that can apply them, I tried to help out but by OEM contacts also don’t have the firmware, vendor in question seems to be a specific version put on their drives from Samsung and that’s it.
Right, unless I've just completely missed the boat, DC can flash the firmware but does not include it.


I've been traveling all weekend and figured that this thread would have received some traffic. If a free P3700 doesn't get the job done, I suppose I'm screwed.. :(
 

AJXCR

Active Member
Jan 20, 2017
565
95
28
32
Can you extract the firmware? Then you could just buy another one;) Or do you actually need someone to have it from a previous update run?

Samsung PM963 2.5" 3.84TB SSD (PCIe Gen3) | eBay
larger drive, same firmware (at least originally if seller never updated)
I looked into this extensively when I still had a drive with the firmware I needed (before I mistakenly wiped it), but could never make it work.. I found a thread on another forum detailing a process to extract firmware off SSD's using software Dell had had floating around at one point, but was never able to make it work; it seemed like users were only able to make it work for a specific group of drives.

If anyone has any experience with extracting firmware i'm all ears, but the more I read, the more it seemed like this was next to impossible.

I can't believe that finding run of the mill original firmware for the drives is such a complicated feat...! Is this not the same stuff that almost all PM963's shipped with?
 

Rand__

Well-Known Member
Mar 6, 2014
4,620
918
113
looks like it - but if you can't extract it where to get it from if its not included in the software...
 

AJXCR

Active Member
Jan 20, 2017
565
95
28
32
looks like it - but if you can't extract it where to get it from if its not included in the software...
Generally I think you'd get it from either Samsung or the company that packaged the drives in a system. In this case where Samsung basically told me they are OEM drives and then refused to communicate further, and the drives don't have any non-Samsung references on the labels, I'm at a complete loss.
 

AJXCR

Active Member
Jan 20, 2017
565
95
28
32
ah darn

Considered playing around with sg_read_buffer to extract firmware?
Don't know anything at all about sg_read_buffer.. Let me do some reading I suppose.

Firmware Enterprise SSDs - SM1625, PM853T, and others
I actually found this thread and sent the original poster a message, but was never able to get a response..

SAS2 expanders $60 (IBM, LSI chip, Intel alternative)
I've tried the drives on both the Intel board/expanders/backplane as well as the SM board/expanders/backplane with no change :/
 

AJXCR

Active Member
Jan 20, 2017
565
95
28
32
sg_read_buffer • help
Usage: sg_read_buffer [--16] [--help] [--hex] [--id=ID] [--length=LEN]
Man Page Help Output
Note: help and version output are generated by a naïve script which tries a few variants of <command> --help, <command> -h etc. to find the command's help and version info. Sometimes it gets lucky, sometimes it doesn't; if the output below looks wrong, it probably is.
sg_read_buffer --version (return code: 0)
version: 1.15 20160131
sg_read_buffer --help (return code: 0)
Usage: sg_read_buffer [--16] [--help] [--hex] [--id=ID] [--length=LEN] [--long] [--mode=MO] [--offset=OFF] [--raw] [--readonly] [--specific=MS] [--verbose] [--version]
DEVICE

where:
--16|-L issue READ BUFFER(16) (def: 10)
--help|-h print out usage message
--hex|-H print output in hex
--id=ID|-i ID buffer identifier (0 (default) to 255)
--length=LEN|-l LEN length in bytes to read (def: 4)
--long|-L issue READ BUFFER(16) (def: 10)
--mode=MO|-m MO read buffer mode, MO is number or acronym (def: 0)
--offset=OFF|-o OFF buffer offset (unit: bytes, def: 0)
--raw|-r output response to stdout
--specific=MS|-S MS mode specific value; 3 bit field (0 to 7)
--readonly|-R open DEVICE read-only (def: read-write)
--verbose|-v increase verbosity
--version|-V print version string and exit

Performs a SCSI READ BUFFER (10 or 16) command. Use '-m xxx' to list available modes. Numbers given in options are decimal unless they have a hex indicator (e.g. a leading '0x').


Do you think this will fly on the PM963's considering that they are NVMe?
 

Rand__

Well-Known Member
Mar 6, 2014
4,620
918
113
sigh, you are right of course, rather unlikely. Testing would be easy enough of course, there are list tools in the sg_utils library.


Revisiting the idea of getting a lenovo firmware - maybe you can get the same level from a lenovo package for the 1.92/3.84 disk? They seem to use the same FW release, i dont think that size would be hardcoded
 

Terry Kennedy

Well-Known Member
Jun 25, 2015
1,068
508
113
New York City
www.glaver.org
sigh, you are right of course, rather unlikely. Testing would be easy enough of course, there are list tools in the sg_utils library.
Most drives don't let you read the firmware back. This is for both technical reasons (all the way back to the Seagate 2GB Barracuda, the drive split the firmware and part was in flash and part was on the disk - the flash was just enough to spin the disk up, do a seek and read of the main firmware - the drive knew how to split a firmware image [.LOD file] into the flash and disk parts of it, but the firmware didn't implement reading it back) and for marketing reasons (the companies like HP, Oracle, etc. that sold drives with huge price mark-ups due to being "certified" don't want you converting generic drives).

For example, the Dell firmware images will only apply using the Dell Nautilus tool - they're not in industry-standard format. And the tool checks to see if it is a Dell drive. Even if you override that check, the drive will probably reject the firmware as incompatible. On a "classic" WD drive (not one actually built by HGST) the OEM info is encoded into the part after the dash - like WD2003FYYS-02W0B0. The drive will reject any firmware not matching 02. The rest (W0B0) describes manufacturing variants which may also have different and incompatible firmware - for example some drives changed from 5-platter to 4-platter during production. It also describes mechanical changes unrelated to firmware, for example whether the drive has FDB bearings or not.
 

AJXCR

Active Member
Jan 20, 2017
565
95
28
32
Most drives don't let you read the firmware back. This is for both technical reasons (all the way back to the Seagate 2GB Barracuda, the drive split the firmware and part was in flash and part was on the disk - the flash was just enough to spin the disk up, do a seek and read of the main firmware - the drive knew how to split a firmware image [.LOD file] into the flash and disk parts of it, but the firmware didn't implement reading it back) and for marketing reasons (the companies like HP, Oracle, etc. that sold drives with huge price mark-ups due to being "certified" don't want you converting generic drives).

For example, the Dell firmware images will only apply using the Dell Nautilus tool - they're not in industry-standard format. And the tool checks to see if it is a Dell drive. Even if you override that check, the drive will probably reject the firmware as incompatible. On a "classic" WD drive (not one actually built by HGST) the OEM info is encoded into the part after the dash - like WD2003FYYS-02W0B0. The drive will reject any firmware not matching 02. The rest (W0B0) describes manufacturing variants which may also have different and incompatible firmware - for example some drives changed from 5-platter to 4-platter during production. It also describes mechanical changes unrelated to firmware, for example whether the drive has FDB bearings or not.
This is sort of the direction I wound up in last time I tried to get on the "firmware extraction" path..

I think at the moment it would be highly beneficial to simply determine what OEM these drives belong to.. Any thoughts on how one might go about finding this?
 

Terry Kennedy

Well-Known Member
Jun 25, 2015
1,068
508
113
New York City
www.glaver.org
I think at the moment it would be highly beneficial to simply determine what OEM these drives belong to.. Any thoughts on how one might go about finding this?
Other than searching for the firmware version string that the OEM used, probably not a lot to go on. You said the drive label was identical to the generic drive, like some small OEM had installed custom firmware themselves, right? Can you ask the seller you bought them from what sort of system they came out of? Weren't these the drives that worked fine when mounted on a NoName PCIe adapter, but had weird reboot issues in a hot-swap backplane? If so, try a different backplane, at least as a test.
 

AJXCR

Active Member
Jan 20, 2017
565
95
28
32
Other than searching for the firmware version string that the OEM used, probably not a lot to go on. You said the drive label was identical to the generic drive, like some small OEM had installed custom firmware themselves, right? Can you ask the seller you bought them from what sort of system they came out of? Weren't these the drives that worked fine when mounted on a NoName PCIe adapter, but had weird reboot issues in a hot-swap backplane? If so, try a different backplane, at least as a test.
Terry,

Unfortunately the seller is NLA (I've had the drives for quite some time at this point). The drives seem to work fine on any (single drive) adapter regardless of motherboard/system.

Regarding a different backplane, I've tried them in two separate systems (motherboards/backplanes/chassis) with exactly the same result. Unfortunately I only have two systems which include NVMe capable backplanes..