Problem with intel i217LM network connection

Hi everyone,

I have installed x86 OpenWrt 19.07.3 (r11063-85e04e9f46) on my hp 800 g1 mini PC. It has intel i217LM network connection.

The problem is when the network speed above 500mbps, this intel i217LM card will stop working, after a few second, the network card will come back.
I found this problem when I test my network speed. I tested many times, the intel i217LM will stop work around 450-500mbps. The detailed log shows below: I start speed test around 586s

[  567.591357] e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: Rx/Tx
[  588.069173] e1000e 0000:00:19.0 eth0: Detected Hardware Unit Hang:
[  588.069173]   TDH                  <59>
[  588.069173]   TDT                  <91>
[  588.069173]   next_to_use          <91>
[  588.069173]   next_to_clean        <55>
[  588.069173] buffer_info[next_to_clean]:
[  588.069173]   time_stamp           <100007013>
[  588.069173]   next_to_watch        <59>
[  588.069173]   jiffies              <100007088>
[  588.069173]   next_to_watch.status <0>
[  588.069173] MAC Status             <40080083>
[  588.069173] PHY Status             <796d>
[  588.069173] PHY 1000BASE-T Status  <3800>
[  588.069173] PHY Extended Status    <3000>
[  588.069173] PCI Status             <10>
[  590.069200] e1000e 0000:00:19.0 eth0: Detected Hardware Unit Hang:
[  590.069200]   TDH                  <59>
[  590.069200]   TDT                  <91>
[  590.069200]   next_to_use          <91>
[  590.069200]   next_to_clean        <55>
[  590.069200] buffer_info[next_to_clean]:
[  590.069200]   time_stamp           <100007013>
[  590.069200]   next_to_watch        <59>
[  590.069200]   jiffies              <100007150>
[  590.069200]   next_to_watch.status <0>
[  590.069200] MAC Status             <40080083>
[  590.069200] PHY Status             <796d>
[  590.069200] PHY 1000BASE-T Status  <3800>
[  590.069200] PHY Extended Status    <3000>
[  590.069200] PCI Status             <10>
[  592.069168] e1000e 0000:00:19.0 eth0: Detected Hardware Unit Hang:
[  592.069168]   TDH                  <59>
[  592.069168]   TDT                  <91>
[  592.069168]   next_to_use          <91>
[  592.069168]   next_to_clean        <55>
[  592.069168] buffer_info[next_to_clean]:
[  592.069168]   time_stamp           <100007013>
[  592.069168]   next_to_watch        <59>
[  592.069168]   jiffies              <100007218>
[  592.069168]   next_to_watch.status <0>
[  592.069168] MAC Status             <40080083>
[  592.069168] PHY Status             <796d>
[  592.069168] PHY 1000BASE-T Status  <3800>
[  592.069168] PHY Extended Status    <3000>
[  592.069168] PCI Status             <10>
[  593.108818] e1000e 0000:00:19.0 eth0: Reset adapter unexpectedly
[  597.091216] e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: Rx/Tx

The intel i217LM card works well in windows 10, which could reach around 900mbps.
So, is there any solution about this problem.
Thank you very much!