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
BSOD on cleanup #372
Comments
Hi @sdecugis , This seems to be the same as #344 So it does not offen happen ? |
Hi @Liryna , indeed the crash happens within FsRtlUninitializeOplock in both case, but from a different call path (in #344 it seems to be at the time of file creation while in my case it is when the file is cleanup). It could be the same root cause still. Since it is kernel memory being corrupted, I really have no access to that memory so I am not sure I could do anything in my callbacks to improve the situation. It looks a kernel driver issue to me, do you agree ? The issue happens very often with some of our sequences of using the virtual filesystem (almost never runs to completion), while some other relatively similar sequence never hit it... I was not able to clearly qualify what makes it happen more or less often. We have one workaround at the moment of stopping MsMpEng.exe but that s far from ideal... I'll write to you separately concerning your dream :D |
@sdecugis I agree it is a kernel issue that has to be fixed ofc. I am just wondering, can you test with the RC4 to see if it also happen ? |
I ll try to get this tested tomorrow as well, not sure when I can get feedback... |
Hi, A couple of additional information:
|
A fixed has been proposed here https://github.com/dokan-dev/dokany/tree/back-out-oplock and waiting test result. |
Hi guys, seems that the BSOD described here is similar to my reported one(#344). Since we did fix it(unintentionally unfortunately) I can speculate on what caused the BSOD - in my opinion our application was causing it by holding the CreateFile or FindFilesWithPattern method calls too long, and it seems that using multiple drive instances exaggerated the BSOD occurrence. If you need any further logs I can try to retrieve them, unfortunately it is difficult because the BSOD corrupts the WinDbg program as well as any debug info it collected. |
Hi @velislav87 , Humm what you describe is different from what we have been to detect during our test. Maybe your BSOD was different in your case ? Only a dump analyse can answer to this 😢 |
Hi @Liryna , do you have any idea on how to go about analyzing the issue further? I can collect, analyze and upload the minidump file from the crash, however any debug info gathered by WinDbg is lost after the BSOD. Will the information in the minidump be enough for the investigation? |
You can analyse the minidump on the machin after the bsod/reboot and post the report. |
We still need some feedback of people facing the issue here before merging the fix . |
IRP MJ CREATE cancel is now implemented ea6c3fd |
Environment
Check List
Description
With the newer Win10 and Windows Defender from this version, we are seeing a BSOD that happens randomly but frequently. We narrowed the sequence to the following (just highlighting the relevant data, the log spreads above 200k lines and contains confidential information):
In user-space library, the log shows that the last callback that was involved was Cleanup.
Note:
Logs
Minidump generated after the BSOD attached.
102016-74718-01.zip
The text was updated successfully, but these errors were encountered: