Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

firmware update not working on windows 10? #130

Open
111alan opened this issue Mar 18, 2020 · 15 comments
Open

firmware update not working on windows 10? #130

111alan opened this issue Mar 18, 2020 · 15 comments
Labels
question Further information is requested

Comments

@111alan
Copy link

111alan commented Mar 18, 2020

Tried to download firmware to the 2 dcpms using ipmctl in windows 10 enterprise 1809, it seems the progress stuck here. I waited about 20min but there's no update at all.

Also tried the latest version of the firmware, still no luck.

stuck

@StevenPontsler
Copy link
Contributor

try running the command ipmctl show -d ARSStatus -dimm before trying to load the new firmware and see if anything is in progress. This will greatly slow down uploads though I would think that it would complete in 20 minutes for 2 modules.

Try running the command with -v option (I think it supports it) and that will give more output.

@kellycouch kellycouch added the question Further information is requested label Mar 18, 2020
@111alan
Copy link
Author

111alan commented Mar 18, 2020

I got this, but still no luck when updating the firmware.
C:\WINDOWS\system32>ipmctl show -d ARSStatus -dimm ---DimmID=0x0000--- ARSStatus=Aborted ---DimmID=0x0100--- ARSStatus=Aborted

Also, -v is not supported by any of these commands.
Syntax Error: Invalid or unexpected token -v.

@kellycouch
Copy link
Contributor

options are position dependent...
ipmctl load -v -source ...

@111alan
Copy link
Author

111alan commented Mar 19, 2020

Thank you, now the -v works. It seems that the progress is extremely slow, estimated more than an hour...
`NVM_DBG_LOGGER NVDIMM-DBG:Dimm.c::FwCmdUpdateFw:3193: BytesToCopy: 64 9344 / 266240. TT: 0x1

NVM_DBG_LOGGER NVDIMM-DBG:Dimm.c::PassThru:7323: Calling 0x9:0x0 over ddrt sp on DCPMM 0x100`

@111alan
Copy link
Author

111alan commented Mar 19, 2020

After about 3 hours the update of 2 512GB DCPMs was finally completed. Thank you for the guidance. I still think the efficiency needs to be higher for a simple firmware update.

The command I used is here, are there any ways to make the process faster next time?
ipmctl load -v -source Z:\Programs\Hardware_Tools\Storage\Intel\fw_ekvb0_1.2.0.5417_rel.bin -examine -dimm

Thanks

@StevenPontsler
Copy link
Contributor

That is much longer than I have heard. My experience has been around 12 seconds per dimm. If ARS runs at the same time it is a couple of minutes.

Can you capture the entire (or at least more) of the output with -v or is it gone?

@nolanhergert
Copy link
Contributor

Thanks for submitting this issue Alan. I'm trying to reproduce it internally. A few more things that you could check that would assist us in root-causing:

  • When you run "-v", does it take about one second between the "BytesToCopy" messages?
  • What firmware version is on the PMem module currently?
  • Did you update BIOS recently, and if so what version?

@nolanhergert
Copy link
Contributor

Also, please file an IPS with your findings. If you don't know how to do this, I can file an equivalent internally for you. Thanks!

@111alan
Copy link
Author

111alan commented Mar 19, 2020

  1. yes, almost exactly 1 second between messages
  2. The original version is 1.2.0.5310, the updated version is 1.2.0.5417
  3. I've been running the latest P2.30 bios for EPC621D6U-2T16R for some times now, the board manufacturer rarely update bios for this board.

I can redo the whole process again a little later if you need the full log.

Thank you.

@nolanhergert
Copy link
Contributor

I think that's all the information we need for now. No need to collect logs anymore. We'll let you know here if we have more info.

@nolanhergert
Copy link
Contributor

We're going to keep this issue open to address the backwards compatibility issue you see, but I do observe our CR 1.0 build of ipmctl working correctly: https://github.com/intel/ipmctl/releases/tag/v01.00.00.3474

FYI: Our goal for our "master" builds is to ensure backwards compatibility with previous hardware generations, but we are not done yet with validating that internally. Additionally, our "master" builds are considered Beta software until the hardware generation under development is released officially.

@111alan
Copy link
Author

111alan commented Mar 20, 2020

OK, thanks, I'll try 1.0 on another platform.

Also could you give me a bit of information, about what else were changed other than frequency in Barlow Pass comparing to Apache Pass? And does these platforms cross-compatible(for example, use Apache Pass on Icelake or use Barlow Pass on Cascadelake)?

Thank you.

@nolanhergert
Copy link
Contributor

Apache Pass is supported on Cascade Lake and Barlow Pass is supported on Cooper Lake and Ice Lake.

@111alan
Copy link
Author

111alan commented Mar 24, 2020

Apache Pass is supported on Cascade Lake and Barlow Pass is supported on Cooper Lake and Ice Lake.

Does this mean Apache pass is not supported and can't be recognized on Cooper and Ice lake?

@nolanhergert
Copy link
Contributor

nolanhergert commented Dec 29, 2021

I don't have an answer for your specific question, other than we don't do any testing of Apache Pass on Cooper or Ice Lake.

Also, what is your board/BIOS that you are using? I think the BIOS implementation is running our payload transactions in SMM mode for some reason, which is subject to throttling. If you force it to use large payload mailbox, you'll still have that 1 second penalty but you'll get in a lot more data per 1s and it should complete much faster.

ipmctl load -ddrt -lpmb -v -source <fw.bin> -dimm

Let me know what you find out. We didn't default to this behavior because our reference BIOS didn't throttle ddrt small payload transactions, so they completed a few seconds faster than using large payload.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
question Further information is requested
Projects
None yet
Development

No branches or pull requests

4 participants