Slow Pre-Session Init Boot Phase (WPR Log Attached)

zinou said:

I think you're focusing on the wrong culprit.
At some point the storport driver make a call to the samsung nvme driver (secnvme.sys). This driver is 2 years old.



Is there any new version for this driver.

Can you check the Windows events log and see if there are events related to failed disk operations?
The trace show that the system is taking too much time mounting a volume !

I installed the Samsung NVMe driver in response to the problem. Before it was just using the Windows generic driver. I saw no difference with either driver.

There's no apparent problem with any disk other than the bizarre mounting time. All shadow volumes suffer the same problem as the primary drive, where they'll take 20+ seconds to mount.
Attachment 352183Attachment 352184Attachment 352185Attachment 352186

Beyond the mounting itself, the storport driver appears to be abnormal. In the documentation for storport https://docs.microsoft.com/en-us/win.../ddi/storport/ there is a lot of references to functions with SCSI, and barely anything related to RAID. However my storport has a huge amount of RAID references and barely any SCSI references, and most of those are functions for converting SCSI requests to RAID:

I'm convinced it's not a drive hardware failure, but a bug with my Windows and perhaps Mobo. I'll keep investigating why my Windows is configured for RAID.

- - - Updated - - -

Checking a known good systems storport.sys revealed the same pattern of RAID functions. I guess that's normal for storport for now...

- - - Updated - - -

zinou said:

I think you're focusing on the wrong culprit.
At some point the storport driver make a call to the samsung nvme driver (secnvme.sys). This driver is 2 years old.



Is there any new version for this driver.

Can you check the Windows events log and see if there are events related to failed disk operations?
The trace show that the system is taking too much time mounting a volume !

Just to add to that, the secnvme.sys driver is only called in the first 7 seconds, and the storahci.sys driver seems held up by the delay as well, whereas the storport.sys is wailing away at RAID functions:

So I'm inclined to believe neither NVMe or AHCI are the source of my problems.

In fact, the storport DPC routine lines up perfectly with the delay, with the rest of storport appearing to be waiting for the DPC routines:

- - - Updated - - -

I just noticed something peculiar. It seems my drive is somehow set to IDE mode:



And somehow I never noticed in Device Manager, but it was there too:



Despite the loaded controllers being AHCI, and the BIOS setting being AHCI, somehow the drive is set to IDE. The BIOS I use doesn't even have an IDE setting, so perhaps it is from my previous mobo. I will look into a way to get that changed to AHCI.

- - - Updated - - -

It appears there is a mismatch between what Windows thinks the drive is and what it is loaded as:

I think I may have gotten derailed on what could be causing the slow boot...

Revisiting the WPR file, I noticed a dramatic difference between the Ntfs.sys!NtfsAddCachedRun of a good system compared to my system.

Good System:



My System:



The Ntfs.sys!NtfsAddCachedRun on a good system is only called once. On my system it is called 19,294 times!!!

Trying to search up information on what Ntfs.sys!NtfsAddCachedRun does yields next to nothing. Is anyone familiar why it would be repeatedly calling this function?