High Latency and DPC issues storport.sys

[grit782]

Hello Friends

I have been having this issue for over a year, and today it got much much worse. It's extremely frustrating, and it is hindering my ability to study. Please assist me, I would be immensely grateful.

What I've tried so far

• reset Windows (choosing keeping my files option) - lags still present after this step.
• upgraded RAM (performance improved but jitters and lags are still present)
• updated BIOS to the latest version
• Updated NVIDIA Drivers
• installed audio and network drivers from motherboard manufacturer's web page
• Upgraded Chipset drivers from AMD website, and upgraded my SSD's firmware
• Enabled the optimal performance Power mode in the Nvidia control panel
• Enabled Ryzen Balanced Power plan in Windows Power settings

Link to requested files: Desktop.zip - Google Drive - I hope that this will be okay, as the file was not uploading via the normal sysnative.com uploader.

Speccy Link: http://speccy.piriform.com/results/AVkTGFHYr2Hq5YzTjisz0jl

Latency Monitor Screenshot:
:

• Desktop?
• Windows 10,
• 64bit
• All hardware is 15 months old besides GPU (23 months) and RAM (3 Months)
• What was original installed OS on system? Windows
• Have you re-installed the OS? Yes
• CPU: Ryzen 1600x
• RAM G.Skill F4-2666C15S-8GVR Ripjaws V 8GB
• Video Card: EVGA GTX 1060 6GB
• MotherBoard - Biostar B350 GTN
• PSU: Super Flower SF-550F14MT Leadex 550W 80 Plus Silver
• Is driver verifier enabled or disabled? Enabled
• What security software are you using? Windows Built in antivirus
• Are you using proxy, vpn, ipfilters or similar software? Occasionally use ExpressVPN, but only got that recently and the problem was occuring before then.
• Are you using Disk Image tools? (like daemon tools, alcohol 52% or 120%, virtual CloneDrive, roxio software) - I am not.
• Are you currently under/overclocking? Are there overclocking software installed on your system? I am not.

 

[grit782]

Please can someone help me, I'm truly desperate. I rebooted into safe mode with networking and the issue was still present.

 

[cwsink]

Hi grit782,

The LatencyMon screenshot suggests a problem while performing high speed disk I/O and your pagefile is on D. Have you checked the health of your 3TB Toshiba drive?

 

[grit782]

Hi grit782,


Hi cwsink

Thank you so much for the reply.



The LatencyMon screenshot suggests a problem while performing high speed disk I/O and your pagefile is on D. Have you checked the health of your 3TB Toshiba drive?

I am waiting for someone to bring a flash drive, so that I can test it with SeaTools. In the mean time, I removed both the power cable, as well as the SATA cable. The issue continues to persist, and now my Latency Mon shows a different process as being the source of the problem.

 

[cwsink]

LatencyMon can be useful but I wouldn't consider what it says as the determining factor of whether or not you're having ISR/DPC latency issues. The reading you were getting for storport.sys certainly indicated a problem but the latest screenshot doesn't really suggest you're having a problem; it's not great but shouldn't be causing media playback issues like audio pops and dropped video frames. Are you experiencing those symptoms or other symptoms which make you suspect ISR/DPC latency problems?

As far as "jitters and lags", is that while playing music/video or playing games or ?

 

[grit782]

LatencyMon can be useful but I wouldn't consider what it says as the determining factor of whether or not you're having ISR/DPC latency issues. The reading you were getting for storport.sys certainly indicated a problem but the latest screenshot doesn't really suggest you're having a problem; it's not great but shouldn't be causing media playback issues like audio pops and dropped video frames. Are you experiencing those symptoms or other symptoms which make you suspect ISR/DPC latency problems?

As far as "jitters and lags", is that while playing music/video or playing games or ?

Hi cwsink

Thank you for your help. The problem seems to be slightly better, but I have still experienced some video and audio jitters and lags.

 

[grit782]

LatencyMon can be useful but I wouldn't consider what it says as the determining factor of whether or not you're having ISR/DPC latency issues. The reading you were getting for storport.sys certainly indicated a problem but the latest screenshot doesn't really suggest you're having a problem; it's not great but shouldn't be causing media playback issues like audio pops and dropped video frames. Are you experiencing those symptoms or other symptoms which make you suspect ISR/DPC latency problems?

As far as "jitters and lags", is that while playing music/video or playing games or ?

This is my latest Latency Monitor report - my Toshiba 3TB hard drive is unplugged from its SATA and Power cables at the moment. Could this indicate my SSD is faulty?

Regards

 

[cwsink]

It could mean the SSD is faulty but not necessarily; it could just mean a DPC associated with storport.sys is waiting on something else. I think we probably should look at a trace to hopefully get an idea of what's going on.

If you want to try capturing a trace, download the Windows 10 SDK installer from here and just select the Windows Performance Toolkit option for installation. Once it's installed you'd need to start an xperf trace, use the system as you normally would to capture instances of the problem happening, stop the trace, and make it available in a reply. Probably the easiest way to do so would be to use the batch file in the attached zip. So, extract the zip, right-click the TraceCPU.cmd file and select Run as administrator. Press a key to start the trace, use the computer in a way that will cause the glitches, and then press a key again to stop the trace. It should create a file called trace.etl on the Desktop. Assuming that works, please make the trace available for download in a reply. You'll likely need to upload it to a cloud drive or file sharing service and make a download link available rather than uploading it to the Sysnative server.

View attachment TraceCPU.zip

 

[grit782]

It could mean the SSD is faulty but not necessarily; it could just mean a DPC associated with storport.sys is waiting on something else. I think we probably should look at a trace to hopefully get an idea of what's going on.

If you want to try capturing a trace, download the Windows 10 SDK installer from here and just select the Windows Performance Toolkit option for installation. Once it's installed you'd need to start an xperf trace, use the system as you normally would to capture instances of the problem happening, stop the trace, and make it available in a reply. Probably the easiest way to do so would be to use the batch file in the attached zip. So, extract the zip, right-click the TraceCPU.cmd file and select Run as administrator. Press a key to start the trace, use the computer in a way that will cause the glitches, and then press a key again to stop the trace. It should create a file called trace.etl on the Desktop. Assuming that works, please make the trace available for download in a reply. You'll likely need to upload it to a cloud drive or file sharing service and make a download link available rather than uploading it to the Sysnative server.

View attachment 43208

Hi

I posted a trace file, along with a dxdiag and msinfo 32 file in the original post. Here is the link: Desktop.zip - Google Drive

I tried to perform a trace with the TraceCPU file you sent me, and I get an error saying that xperf is not recognised as an internal or external command.

 

[grit782]

It could mean the SSD is faulty but not necessarily; it could just mean a DPC associated with storport.sys is waiting on something else. I think we probably should look at a trace to hopefully get an idea of what's going on.

If you want to try capturing a trace, download the Windows 10 SDK installer from here and just select the Windows Performance Toolkit option for installation. Once it's installed you'd need to start an xperf trace, use the system as you normally would to capture instances of the problem happening, stop the trace, and make it available in a reply. Probably the easiest way to do so would be to use the batch file in the attached zip. So, extract the zip, right-click the TraceCPU.cmd file and select Run as administrator. Press a key to start the trace, use the computer in a way that will cause the glitches, and then press a key again to stop the trace. It should create a file called trace.etl on the Desktop. Assuming that works, please make the trace available for download in a reply. You'll likely need to upload it to a cloud drive or file sharing service and make a download link available rather than uploading it to the Sysnative server.

View attachment 43208

This is my latest Latency Monitor report.

 

[cwsink]

Thank you for pointing out the trace - I missed it, somehow. The script should run if xperf is in the Path . Did you use xperf to get the first trace? If so, since it worked before I'm guessing xperf might have been installed on D so won't be found if D isn't connected. My path includes the directory C:\Program Files (x86)\Windows Kits\10\Windows Performance Toolkit\ which is where xperf was installed on my system, for example.

As far as the trace, it looks like almost every core has at least 1 instance of a DPC taking over 100 ms to run which is a long time. Those seem to be while calling either ntoskrnl.exe!ExpGetPoolTagInfoTarget or storport.sys!RaidpAdapterTimerDpcRoutine so apparently while working with pool memory or a disk I/O timer DPC, respectively.

I'm curious to see what the memory is doing while the problems are occurring but I'll need to see what parameters should be used in the script. In the meantime, please see if you can get xperf installed and included in the PATH environment variable to try and get a trace using the batch.

 

[grit782]

Thank you for pointing out the trace - I missed it, somehow. The script should run if xperf is in the Path . Did you use xperf to get the first trace? If so, since it worked before I'm guessing xperf might have been installed on D so won't be found if D isn't connected. My path includes the directory C:\Program Files (x86)\Windows Kits\10\Windows Performance Toolkit\ which is where xperf was installed on my system, for example.

As far as the trace, it looks like almost every core has at least 1 instance of a DPC taking over 100 ms to run which is a long time. Those seem to be while calling either ntoskrnl.exe!ExpGetPoolTagInfoTarget or storport.sys!RaidpAdapterTimerDpcRoutine so apparently while working with pool memory or a disk I/O timer DPC, respectively.

I'm curious to see what the memory is doing while the problems are occurring but I'll need to see what parameters should be used in the script. In the meantime, please see if you can get xperf installed and included in the PATH environment variable to try and get a trace using the batch.

Hi

The problem has improved somewhat - it has only occured thrice in around 8 hours. All 3 times Latency Monitor shows the issue being with storport.sys - the first time it occured, I was installing Office onto the SSD. The second, I switched the PC on after hibernation, and the issue occured immediately. The third time I was browsing Firefox. The first two times the issue resolved itself without the need for a restart, but this third time as I am typing, the issue is present and has not went away.

What I did since my last post.
1. I formatted my SSD totally, and clean installed Windows - I also updated all my drivers to their latest versions.
2. I inserted my SSD and Hard Drive's SATA cables into different SATA ports, to check if the original ports were faulty.
3. Someone on a forum suggested to someone that they disable Legacy USB 3.0 support, and set Legacy USB to auto in their BIOS. I did the same.


Attached is the trace you requested: trace.etl - Google Drive

Thank you for your assistance, and I truly hope you can come up with a solution. I am wondering if my SSD is faulty - Sandisk's software tool found no errors however.

 

[grit782]

Thank you for pointing out the trace - I missed it, somehow. The script should run if xperf is in the Path . Did you use xperf to get the first trace? If so, since it worked before I'm guessing xperf might have been installed on D so won't be found if D isn't connected. My path includes the directory C:\Program Files (x86)\Windows Kits\10\Windows Performance Toolkit\ which is where xperf was installed on my system, for example.

As far as the trace, it looks like almost every core has at least 1 instance of a DPC taking over 100 ms to run which is a long time. Those seem to be while calling either ntoskrnl.exe!ExpGetPoolTagInfoTarget or storport.sys!RaidpAdapterTimerDpcRoutine so apparently while working with pool memory or a disk I/O timer DPC, respectively.

I'm curious to see what the memory is doing while the problems are occurring but I'll need to see what parameters should be used in the script. In the meantime, please see if you can get xperf installed and included in the PATH environment variable to try and get a trace using the batch.

Latency Monitor is also showing me this occasionally:

 

[cwsink]

The latest trace seems to be blaming the same culprits (pool memory related functions and storport.sys) and has a very consistent pattern. Every 200 ms or so the system seems to be performing a clock interrupt and shortly afterwards a storport.sys related DPC starts executing and takes around 100 ms. Interestingly, in this trace, the DPCs are all running on core 10 which I wouldn't think would cause problems for audio playing back on core 0 unless that DPC is getting data needed by a media app running on core 0.

It still suggests a problem with drive I/O but I'm not sure why it would be intermittent. It makes me wonder if the system is experiencing memory pressure; 8GB of RAM is not hard to max out these days. Have you looked at the memory information in Task Manager while the problems are happening? Paged and non-paged pool in particular. A screenshot of the memory information of the Performance tab in Task Manager might be useful to see.

 

[cwsink]

The latest trace seems to be blaming the same culprits (pool memory related functions and storport.sys) and has a very consistent pattern. Every 200 ms or so the system seems to be performing a clock interrupt and shortly afterwards a storport.sys related DPC starts executing and takes around 100 ms. Interestingly, in this trace, the DPCs are all running on core 10 which I wouldn't think would cause problems for audio playing back on core 0 unless that DPC is getting data needed by a media app running on core 0.

It still suggests a problem with drive I/O but I'm not sure why it would be intermittent. It makes me wonder if the system is experiencing memory pressure; 8GB of RAM is not hard to max out these days. Have you looked at the memory information in Task Manager while the problems are happening? Paged and non-paged pool in particular. A screenshot of the memory information of the Performance tab in Task Manager while the problem is occurring might be useful to see.

 

[grit782]

The latest trace seems to be blaming the same culprits (pool memory related functions and storport.sys) and has a very consistent pattern. Every 200 ms or so the system seems to be performing a clock interrupt and shortly afterwards a storport.sys related DPC starts executing and takes around 100 ms. Interestingly, in this trace, the DPCs are all running on core 10 which I wouldn't think would cause problems for audio playing back on core 0 unless that DPC is getting data needed by a media app running on core 0.

It still suggests a problem with drive I/O but I'm not sure why it would be intermittent. It makes me wonder if the system is experiencing memory pressure; 8GB of RAM is not hard to max out these days. Have you looked at the memory information in Task Manager while the problems are happening? Paged and non-paged pool in particular. A screenshot of the memory information of the Performance tab in Task Manager might be useful to see.

The issue occured right now - interestingly, video is stuttering on YouTube, but the audio is smooth. What do you mean by a problem with drive I/O?

Here is the screenshot you requested

 

[cwsink]

I mean the same driver (storport.sys) and function calls showing odd values as described in an earlier reply. In the latest trace it's showing core 10 as trying to handle the long running DPCs involving storport.sys while the older trace shows all cores other than core 0 attempting to do so. Core 0 is likely playing back media and looks like it should be able to do so without a problem if you only look at core 0. But I suspect the 100 ms runtime of the DPCs on the other cores is too long and ends up starving core 0 of data to playback rather than core 0 being overwhelmed. The system seems to be having trouble providing data for media playback to core 0 in a timely manner. The spikes in this graph are core 10 maxing out the CPU while running the function storport!RaidpAdapterTimerDpcRoutine. That's not normal.

What is to blame I'm not sure. I'd expect you to be getting crashes if it was a hardware problem but not necessarily. If it takes a while for the problem to appear after a system restart I would suspect memory pressure causing excessive pagefile activity but your pool memory sizes look fine and you have 1.5GB of physical memory available according to the screenshot.

The Ethernet looks busy. Are you playing videos over the Ethernet connection and having playback issues? If so, is it over the Internet or a network attached storage device? Do you have problems playing back media on local storage devices?

 

[grit782]

I mean the same driver (storport.sys) and function calls showing odd values as described in an earlier reply. In the latest trace it's showing core 10 as trying to handle the long running DPCs involving storport.sys while the older trace shows all cores other than core 0 attempting to do so. Core 0 is likely playing back media and looks like it should be able to do so without a problem if you only look at core 0. But I suspect the 100 ms runtime of the DPCs on the other cores is too long and ends up starving core 0 of data to playback rather than core 0 being overwhelmed. The system seems to be having trouble providing data for media playback to core 0 in a timely manner. The spikes in this graph are core 10 maxing out the CPU while running the function storport!RaidpAdapterTimerDpcRoutine. That's not normal.

What is to blame I'm not sure. I'd expect you to be getting crashes if it was a hardware problem but not necessarily. If it takes a while for the problem to appear after a system restart I would suspect memory pressure causing excessive pagefile activity but your pool memory sizes look fine and you have 1.5GB of physical memory available according to the screenshot.

The Ethernet looks busy. Are you playing videos over the Ethernet connection and having playback issues? If so, is it over the Internet or a network attached storage device? Do you have problems playing back media on local storage devices?

Hi

My ethernet was busy as I was currently downloading some things. The videos stutter whenever there DPC latency issue occurs. Do you have any possible solutions? I have created a partition on my hard drive, and installed Windows on it. I will unplug my SSD to test if the issues still appear.


I am not sure what else could be causing this. Do you think any of my graphics card, CPU or Motherboard could be faulty?

Thanks once again for your assistance.

 

[grit782]

I mean the same driver (storport.sys) and function calls showing odd values as described in an earlier reply. In the latest trace it's showing core 10 as trying to handle the long running DPCs involving storport.sys while the older trace shows all cores other than core 0 attempting to do so. Core 0 is likely playing back media and looks like it should be able to do so without a problem if you only look at core 0. But I suspect the 100 ms runtime of the DPCs on the other cores is too long and ends up starving core 0 of data to playback rather than core 0 being overwhelmed. The system seems to be having trouble providing data for media playback to core 0 in a timely manner. The spikes in this graph are core 10 maxing out the CPU while running the function storport!RaidpAdapterTimerDpcRoutine. That's not normal.

What is to blame I'm not sure. I'd expect you to be getting crashes if it was a hardware problem but not necessarily. If it takes a while for the problem to appear after a system restart I would suspect memory pressure causing excessive pagefile activity but your pool memory sizes look fine and you have 1.5GB of physical memory available according to the screenshot.

The Ethernet looks busy. Are you playing videos over the Ethernet connection and having playback issues? If so, is it over the Internet or a network attached storage device? Do you have problems playing back media on local storage devices?

I bought a new SSD, clean installed the latest Windows build onto it, and updated all my drivers. Since then, I am not getting any audio stuttering or pops/dropouts. However, there is video lag, but somehow it seems different to the video lag I used to get before. I'm not sure how to explain it. Latency Monitor is still showing issues, and blaming ntoskrnl now, and not storport.sys like before.

Attached is a screenshot:

 

[cwsink]

It's not clear to me what you might be referring to as far as video lag. Not responding to mouse input as quickly as you're used to or ?

As far as LatencyMon, I wouldn't be too worried if you're no longer getting the audio pops and dropouts. I recently installed a Windows 10 Insider build and ran a trace of my system. It was showing quite high DPC processing times for DPCs running on cores other than 0 and sure enough it was associated with ntoskrnl.exe. One took as long as 17 milliseconds which would normally be considered terrible and yet my system was running better than it has since the Spectre/Meltdown updates last year. So I think Microsoft have done some tweaks regarding ISR/DPC scheduling to make better use of multiple processors in the Insider builds and I doubt LatencyMon is taking those changes into account yet - if ever they do. When I reverted back to build 1809 all my DPCs went back to sub-millisecond levels so if you're still using an Insider build it might be worth trying the 1809 build to see if your DPCs also become more reasonable.