Overclock.net banner

561 - 580 of 614 Posts

·
Registered
Joined
·
2,286 Posts
-snip-

Given that you are willing to rely a Phison controller with your data.
Wait, are Phison controllers of dubious quality/reliability?
 

·
ٴٴٴ╲⎝⧹˙͜>˙⧸⎠╱
Joined
·
6,320 Posts
I mean in theory if you use PSPTool and know where to look for, you maaybe can get that to work
But the theory would require a fix inside SMU firmware very likely
Then ported back AMD CBS flags in order to support the PCIe 4.0 switch

And yes, PSP firmware (on the chip) needs to be able to talk with valid commands to SMU
theoretically i could work, if AMD didn't/won't change the encryption method before 1005AGESA
And updating to a new one will invalidate the public key, and update PSP Firmware
Yes TK421, maybe :wheee:

EDIT:
As stamped several revisions of ryzens exist, one UG one UF and one without a stamp
it might be possible that inside PSP firmware this sector is permanently locked on the new stamped batch too :p
The PSP firmwares are indeed part of AGESA, but thats quite an understatement really.

On these platforms, the PSP starts up first and only then are the rest of the IPs / hardware initialised.
There is exactly zero control over anything prior the PSP and basically until it has completed initialising the rest of the hardware, you are running a non-x86 system that is completely isolated.

What almost certainly happens is, that the PSP checks the fused DeviceID of the chipset, to determine if the present variant is allowed to run PCIe Gen. 4.
Probably the easiest way around this would be physically replacing the existing Promontory / Promontory LP chipsets with the variant, that is white listed for Gen. 4 enablement in the FW (Promontory LP; B550A, B550).

There is some indication that some kind of anti roll-back measures for the PSP FW have been implemented, but I'm not sure how tripping into such measures would manifest itself.
From the prior experience I know that the cross-compability of the different AGESA modules is very limited and with PSP modules it is virtually non-existent. Essentially, until the PSP has been satisfied, you're dealing with a brick.

Even if it would be somehow possible to downgrade to a PSP FW which allows Gen. 4 operation on older board designs, it probably wouldn't be wise to use such old version. Afterall, PSP stands for "Platform Security Processor" at this
point (more than 6 months after the change), there have been tons of bug fixes and other kinds of improvements.

Also personally, I fail to see the point in enabling the Gen. 4 PCIe on Promontory chipset (300 & 400-series) based boards.
Since the chipset itself doesn't support Gen. 4, there is no bandwidth advantage for the chipset connected devices, that rely on the x4 link between the CPU and the chipset and because of that are potentially bandwidth starved at Gen. 3 bandwidth.
Currently the only M.2 Gen. 4 drives have Phison controllers on them (...) and the only GPUs that support Gen. 4 are Navi based Radeons (...). So since Radeons most certainly don't need Gen. 4, the only potential use would be M.2 storage. Given that you are willing to rely a Phison controller with your data.





does psptool even exist, I know tools for intel ME does but unaware of amd stuff being accessible








2080ti is already showing perf loss on P3x8, and samsung is going to release their pcie4 nvme drives soon


most consumers will have 1 gpu and 1 nvme connected to the cpu
 

·
Premium Member
Joined
·
2,738 Posts
Discussion Starter #563
Wait, are Phison controllers of dubious quality/reliability?
Personally, I'm not touching them even with a ten foot pole due to the horrible experiences on both, their USB and SATA flash controllers.
Granted things might have improved since then, but personally I won't touch anything made by Phison ever again.
 

·
Premium Member
Joined
·
2,738 Posts
Discussion Starter #564
does psptool even exist, I know tools for intel ME does but unaware of amd stuff being accessible








2080ti is already showing perf loss on P3x8, and samsung is going to release their pcie4 nvme drives soon


most consumers will have 1 gpu and 1 nvme connected to the cpu
Based on the very brief look I had at the tool, to me it appears the "PSPTool" is only able to manipulate the different components of the PSP stack (insert, extract, delete or replace) not touch the code itself.
The firmwares themselves are extremely well protected and any kind of modifications to the actual code would require the access to the key that only AMD has.

I could be wrong, but thats how I understood it based on the quick look.
 

·
Registered
Joined
·
2,286 Posts
Personally, I'm not touching them even with a ten foot pole due to the horrible experiences on both, their USB and SATA flash controllers.
Granted things might have improved since then, but personally I won't touch anything made by Phison ever again.
Good to know. Thank you.

Is AGESA 1.0.0.5 anything to look forward to?
 

·
Old school
Joined
·
38 Posts

·
Registered
Joined
·
26 Posts
Hello anyone here using asus b450i? Im using 3004 buos and i believe this is the 1.0.0.4.

I tried overclocking and i need 1.37volts for it to reach 4.2ghz so someone suggested that i can try 1.0.0.3

However when i checked the asus website. I dont know if i can diwngrade to 2901 or 2801 without affecting my m2 slots since i am using ryzen 5 3600 and m2s at the same time.

Any inputs will be appreciated


Sent from my SM-N960F using Tapatalk
 

·
Registered
Joined
·
25 Posts
@The Stilt

Hi mate, sorry for my request, but would you be so kind to compile latest flashrom 1.2 with -internal support under pure DOS environment ?

I've downloaded source from the github repository but my attempts to compile from myself have failed hundred times and I'm very tired.

If you could make a 1.2 version of this wonderful tool it would be a great and useful thing.

Thanks anyway for your help and keep up the good work. ;)
 

·
Premium Member
Joined
·
2,738 Posts
Discussion Starter #571
@The Stilt

Hi mate, sorry for my request, but would you be so kind to compile latest flashrom 1.2 with -internal support under pure DOS environment ?

I've downloaded source from the github repository but my attempts to compile from myself have failed hundred times and I'm very tired.

If you could make a 1.2 version of this wonderful tool it would be a great and useful thing.

Thanks anyway for your help and keep up the good work. ;)
Unfortunately it isn't possible, since I'm now using 3rd gen. Threadripper as a workstation and these aren't compatible with Linux.
 

·
Premium Member
Joined
·
2,738 Posts
Discussion Starter #573
Don't worry, thanks anyway. ;)
Actually, I took interest in the reasons why I couldn't get TRX40 platform to work under Linux, despite all of the know workarounds I tried.
Apparently it has something to do with the GPU UEFI bios, which caused the system to halt at initilization.

I compiled few different flavors, since I'm not sure how stable the latest commits are...
The one marked 1.2 is the 1.2 release from February, the rest are different commits. flashrom_cc71eb5 is based on the newest commit (7a7fee1 is an intermediate version), but I had to patch it before it could be compiled for DOS.
The maintainers have forgotten to disable support for Raiden for DOS builds, so without disabling Raiden (libusb dependency) build it could not be compiled.

The "patch" is extremely simple, open Makefile and add after line 213:

Code:
ifeq ($(CONFIG_RAIDEN), yes)
UNSUPPORTED_FEATURES += CONFIG_RAIDEN=yes
else
override CONFIG_RAIDEN = no
endif
https://1drv.ms/u/s!Ag6oE4SOsCmDh1h7BCmrdrMkbTIT

Note that these builds DO NOT include support for Ryzen CPUs.
 

·
Registered
Joined
·
368 Posts
Flashrom works fine with my ASRock X570 Pro4 mainboard

@The Stilt

Thank you very much for your phantastic work.
I can confirm, that the Flashrom tool is able to flawlessly dump the 32MB sized BIOS Region of my ASRock X570 Pro4 and to successfully flash the updated/modded BACKUP.ROM into the SPI Chip of the board.

Since I was not able to get any UBU tool modded BIOS properly flashed into my ASRock X570 Pro4 mainboard by using the in-box Instant Flash tool, it was your friend aGeoM, who encouraged me to try the Flashrom tool and helped me to find the exact BIOS flashing procedure. Meanwhile I have written a short guide into the start post of >this< Win-RAID Forum thread.

Today I wanted to find out, whether it is possible to do it as well by using FREEDOS instead of MS DOS. After having removed the diskcopy.dll file from my Win10 system drive, I created the bootable FREEDOS image by using the original (unpatched) Rufus version 3.10. To my surprise I succeeded with the creation of the BIOS Region dump.

Question to you: Do you think, that it is possible as well to get a modded BIOS properly flashed by using the same FREEDOS USB Stick?
Thanks in advance for your reply.

Kind regards
Dieter (alias Fernando)
 

Attachments

·
Premium Member
Joined
·
2,738 Posts
Discussion Starter #576
@The Stilt

Thank you very much for your phantastic work.
I can confirm, that the Flashrom tool is able to flawlessly dump the 32MB sized BIOS Region of my ASRock X570 Pro4 and to successfully flash the updated/modded BACKUP.ROM into the SPI Chip of the board.

Since I was not able to get any UBU tool modded BIOS properly flashed into my ASRock X570 Pro4 mainboard by using the in-box Instant Flash tool, it was your friend aGeoM, who encouraged me to try the Flashrom tool and helped me to find the exact BIOS flashing procedure. Meanwhile I have written a short guide into the start post of >this< Win-RAID Forum thread.

Today I wanted to find out, whether it is possible to do it as well by using FREEDOS instead of MS DOS. After having removed the diskcopy.dll file from my Win10 system drive, I created the bootable FREEDOS image by using the original (unpatched) Rufus version 3.10. To my surprise I succeeded with the creation of the BIOS Region dump.

Question to you: Do you think, that it is possible as well to get a modded BIOS properly flashed by using the same FREEDOS USB Stick?
Thanks in advance for your reply.

Kind regards
Dieter (alias Fernando)
Hi Dieter,

There is no reason why FreeDOS shouldn't work with Flashrom.

The thing is just that I haven't personally used Flashrom with anything else besides MS-DOS, and there are at least few cases where the Flashrom has failed
when it was used under FreeDOS (allegedly). That being said it is impossible to determine whenever the failures have actually had anything to do with FreeDOS itself, or did they occur due to
user actions (error) or hardware instability for instance. In most cases the failures happen due to user error however, since I have no experience on FreeDOS I really have no opinion about the subject.
 

·
Registered
Joined
·
368 Posts
@The Stilt
Thanks for your quick reply and for your statement.
I will report here, if I should run into problems with Flashrom due to the usage of a FREEDOS instead of an MS DOS image.
 

·
Registered
Joined
·
112 Posts
[MENTION=348110] ...it was your friend aGeoM...
:) I never said he is my "friend", in terms of knowing him or ever interact with him on any forum, but I truly respect his knowledge as his work and must of all his good will to help the community all over this years asking nothing and offering everything, in that sense, yes... he is my friend and friend of all of us, like many others that I respect silently, but honoring them helping others.
About Flashrom1.1 for Ryzen, I did flash my ASUS STRIX X570-E with FREEDOS bootable USB stick successfully, I also did try flashrom 1.2 even knowing that it have no Ryzen support but I had to try anyway. :)
 
  • Rep+
Reactions: Fernando 1

·
Overclock the World
Joined
·
1,984 Posts
@Fernando 1 when you ever get your hands on an SPI flasher
And please take this with untested grain of salt
Can you try to check if the security capsules on current X570 ASRock boards, still align with my findings and attempt
https://twitter.com/VeiiTM/status/1248852503132884992 <- tiny guide

The old UBU Patch method was filling out Capsule:
GUID 414D94AD-998D-47D2-BFCD-4E882241DE32
Subtype GUID: 5A88641B-BBB9-4AA6-80F7-498AE407C31F
With FF, that way purposely disabling verification check after bios-flashing, after post
~ i forgot the user who found it, but it was from the winraid forum :)

Tho this method is flawed sadly and not what we should aim for,
Once UBU breaks their verification capsule, you have to force flash it via afuefi or flashrom
Although at least it doesn't error out as "unverified bios" and boots up

I found on the "recent" 6.20A bios about 2 months ago a better method but people couldn't verify it till now
8B A6 3C 4A 23 77 FB 48 80 3D 57 8C C1 FE C4 4D for the location lookup anchor and has to stay there
and is found inside:
Capsule, iFlashDXE, SecSMIFLash, Recovery / at least on the 1 partition 16mb bios
Likely it's twice on the X570 Asrock

Would be great if you could doublecheck when UBU pushes the bios and patches, it if the verification capsule exist twice or there are new links to it
Yes overall, it would be nice if you can check and continue research on that part till the end ~ or forward bit of the research for other ASRock users
I sadly lost resources and the possibility to finish the whitelist since then, although i think the try here should work:
https://www.overclock.net/forum/11-...chi-overclocking-thread-746.html#post28403358

You can inspect the modules in MODTC to see my changes vs what UBU does patch by the old method ~ or visible via twitter
The only part i needed to finish and explore, was the size of the verification "key"
How much needs to be 00 data and how much needs to be a spaceholder
Well who knows, it might've worked but no one confirmed so far
Overall if you are willing to continue research - resources are linked :)
I'll finish it someday when i get back on my feet and RL settles
Mentioned you, as i've seen your name over on winraid often enough ~ thought you might be interested as ASRock user :whistle:
 

·
Registered
Joined
·
164 Posts
May I interrupt the Flow just to say something?

It is amazing to see old Win Raid forum guys over here in amd stuff!! It also really seems to symbolize how Intel is behaving in a while.

On the topic, i have a Crosshair VI with flashback and willing to flash anything on it in order to achieve ryzen 4000 boot (I don't care at all of all the boost technologies and PSP securities) or Pciex 4.0.

Thank you guys!

Inviato dal mio SM-G925F utilizzando Tapatalk
 
561 - 580 of 614 Posts
Top