I did report it through feedback hub. But there is no “storage device” option under “device and driver”. I’m afraid my reporting will land on a wrong team, and hopefully they can re-assign the issue to the correct team who is in charge of HMB implementation.
2 Likes
I have tried setting HostMemoryBufferBytes to 64 MB as described and it works.
Event logger now shows:
Attemped to Allocate: 0x4000000 bytes
Actually Allocated: 0x4000000 bytes
Device Minimum: 0x337000 bytes
Device Preferred: 0x4000000 bytes
Policy Maximum: 0x1FA56000 bytes
SSD is SN770 2 TB, if you have this drive you can just copy paste from nissel’s post “VEN, DEV and REV” are the same.
1 Like
Could you please share a link so we can chime in?
Unfortunately 3 days later, I still have not received any response from WD.
That is very disappointing!
Feedback hub link
I did use the Feedback Hub to report some Hardware Independent Flip mode issues way back in the Windows 10 21H2 days… They tend to relay these reports to their respective teams relatively fast, so I doubt there will be problems in reaching proper devs.
I am not allowed to open that feedback. Anyway, I hope that the problem gets addressed soon. This is a huge problem for lots of users.
I started new thread about this issue in Microsoft community forum. if Microsoft is not aware of the problem they will not release a fix. Post there too.
Windows 24H2 + WD SN580/SN770 = blue screens
2 Likes
Can confirm this method of setting the HMB size works, the error has disappeared from Event Viewer and there are no more crashes!
1 Like
Could you please show me where can I find these values in the Event logger? Thanks
That is not from the event log it’s just 64MiB in bytes, 1048576 x 64.
I mean these:
Attemped to Allocate: 0x4000000 bytes
Actually Allocated: 0x4000000 bytes
Device Minimum: 0x337000 bytes
Device Preferred: 0x4000000 bytes
Policy Maximum: 0x1FA56000 bytes
I would like to check, but I don’t know where can I find it
With a reboot one can open Windows Event Viewer, then navigate to Applications & Services Logs, Microsoft, Windows, StorPort, Microsoft-Windows-Storage-Storport/Operational. Sort by Event IDs and look for IDs with #: 513. This will give you the StorPort events that report the allocation type as above plus the amount, minimum and maximum HMB your controller supports and whether allocation was successful. It’ll look something like this (taken from the SM11 with HMBAllocationPolicy set to 2, or maximum):
Host Memory Buffer Allocation for \Device\RaidPort1 succeeded.
Failure Reason: None
Allocation Policy: Maximum
Attempted to Allocate: 0x1000000 bytes
Actually Allocated: 0x1000000 bytes
Device Minimum: 0x1000000 bytes
Device Preferred: 0x1000000 bytes
The question is why there are no problems with the disks of other manufacturers, but only with those of Western Digital? Maybe Windows update will mask the problem, but WD must release firmware update like they did for SN850X drives
Thanks, I found it!
Yeah i can only guess it’s a firmware bug that only WD made in a couple of their products.
SN850X and SN750 are different. They do have DRAM cache on board, so they don’t using HMB.
For those who has SN770, do you have also this error in event list?
I’m about to try the solution that nissell described, setting the 64MB manually: “VEN_15B7&DEV_5017&REV_01 67108864” but on the other hand when I tried to manually install 64bit Micron NVMe driver v2.1.19.0 WHQL and stornvme disappeared I still had some crashes, much less, but they were there.
I bought this new pc and if I totally refund it I will probably lose my OEM windows key that I used already with this hardware.
My plan is try this last solution that you guys posted for two days and check if the error is completely fixed, but I’m afraid about that volmgr error and the other crashes.
You’ll need to be using the standard microsoft NVME driver for the registry settings to have any effect.
For me when I tried the micron driver it just disabled HMB.
It’s possible your remaining crashes are from a different cause. For me my system is very stable now.
SN850X were affected too. DRAMless or not, only WD drives are affected. Maybe the problem is something else? Something is wrong with WD drives for sure.
Anyway, I will wait few weeks and if they not fixed it I will send for warranty replacement.
I’m not sure what problem you are talking about. Let’s get it clear.
-
BSOD with error “critical process died” is related to HMB. Disabling HMB or set maximum HMB to 64MB solve the problem.
-
Event viewer error “The driver detected a controller error on \Device\RaidPort0”. I’m not sure what is the cause. But it seems WD support saying it is not affect performance (source). And it seems WD has released a firmware update to SN850X model. (source)
-
Event viewer error from volmgr. I saw this error on the machine with 23H2 + SN770, on which the HMB size is 64MB and drives performs no problem. No system freeze and no BSOD. So I guess this is a separate issue. However, on 24H2 + SN770, after disabling HMB, this error is gone.
