pci: ethernet no longer receives packets

[jessicah]

Virtual Box 7.0.4. UEFI boot. Intel PRO/1000 MT Desktop nic.

Ran git bisect to test, resulted in a23ac85fa9e39c68ce3ff2ef6e0d78cf705bec4e, as suspected.

Builds prior, ping, works fine; after, ping and all other networking fails to receive any data. ifconfig receive is 0.

Virtual Box config:

<?xml version="1.0"?>
<!--
** DO NOT EDIT THIS FILE.
** If you make changes to this file while any VirtualBox related application
** is running, your changes will be overwritten later, without taking effect.
** Use VBoxManage or the VirtualBox Manager GUI to make changes.
-->
<VirtualBox xmlns="http://www.virtualbox.org/" version="1.19-windows">
  <Machine uuid="{be093f2d-27dc-4fd8-86b7-41a9a8aa1a94}" name="Haiku" OSType="Other_64" snapshotFolder="Snapshots" lastStateChange="2023-04-30T04:55:26Z">
    <MediaRegistry>
      <HardDisks>
        <HardDisk uuid="{a523aca5-3fb7-4b3c-b8ad-2a57f25ab102}" location="Haiku.vdi" format="VDI" type="Normal"/>
        <HardDisk uuid="{4be0e633-8047-4ee3-9003-4a73ec6eeb0b}" location="O:/Haiku.vdi" format="VDI" type="Normal"/>
        <HardDisk uuid="{bcf65d0b-8650-44e1-be98-dff19fdc2385}" location="HaikuEFI.vdi" format="VDI" type="Normal"/>
      </HardDisks>
      <DVDImages>
      </DVDImages>
    </MediaRegistry>
    <ExtraData>
    </ExtraData>
    <Hardware>
      <CPU count="4">
        <PAE enabled="false"/>
        <LongMode enabled="true"/>
        <X2APIC enabled="true"/>
        <HardwareVirtExLargePages enabled="true"/>
      </CPU>
      <Memory RAMSize="12288"/>
      <Firmware type="EFI"/>
      <HID Pointing="USBMultiTouch"/>
      <Chipset type="ICH9"/>
      <Boot>
        <Order position="1" device="HardDisk"/>
        <Order position="2" device="None"/>
        <Order position="3" device="None"/>
        <Order position="4" device="None"/>
      </Boot>
      <Display VRAMSize="32" monitorCount="2"/>
      <BIOS>
        <IOAPIC enabled="true"/>
      </BIOS>
      <USB>
        <Controllers>
          <Controller name="OHCI" type="OHCI"/>
        </Controllers>
        <DeviceFilters>
        </DeviceFilters>
      </USB>
      <Network>
        <Adapter slot="0" enabled="true" MACAddress="080027515663" type="82540EM">
          <DisabledModes>
            <NAT localhost-reachable="true"/>
            <InternalNetwork name="intnet"/>
            <NATNetwork name="NatNetwork"/>
          </DisabledModes>
          <BridgedInterface name="Intel(R) I210 Gigabit Network Connection"/>
        </Adapter>
      </Network>
      <AudioAdapter driver="DirectSound" enabled="true" enabledOut="true"/>
      <Clipboard/>
      <GuestProperties>
        <GuestProperty name="/VirtualBox/HostInfo/GUI/LanguageID" value="en_US" timestamp="1673135705105451100" flags="RDONLYGUEST"/>
      </GuestProperties>
      <StorageControllers>
        <StorageController name="AHCI" type="AHCI" PortCount="3" useHostIOCache="false" Bootable="true" IDE0MasterEmulationPort="0" IDE0SlaveEmulationPort="1" IDE1MasterEmulationPort="2" IDE1SlaveEmulationPort="3">
          <AttachedDevice type="HardDisk" hotpluggable="false" port="0" device="0">
            <Image uuid="{a523aca5-3fb7-4b3c-b8ad-2a57f25ab102}"/>
          </AttachedDevice>
          <AttachedDevice type="HardDisk" hotpluggable="false" port="1" device="0">
            <Image uuid="{bcf65d0b-8650-44e1-be98-dff19fdc2385}"/>
          </AttachedDevice>
        </StorageController>
        <StorageController name="NVMe 2" type="NVMe" PortCount="1" Instance="1" useHostIOCache="false" Bootable="false">
          <AttachedDevice type="HardDisk" hotpluggable="false" port="0" device="0">
            <Image uuid="{4be0e633-8047-4ee3-9003-4a73ec6eeb0b}"/>
          </AttachedDevice>
        </StorageController>
      </StorageControllers>
    </Hardware>
  </Machine>
</VirtualBox>

[jmairboeck]

I had the same problem, but weirdly, it now works again.

I also noticed that the mouse integration (using "USB Tablet") didn't work any more.

I checked the syslog, noticed some USB (xhci) messages recurring over and over. Then I noticed that the Extension Pack was out of date (I have VBox 7.0.6 but had Extension pack 7.0.4), so I shut down the VM, updated the Extension Pack, booted again and then everything worked.

[X512]

<Subscribe>

[waddlesplash]

a23ac85fa9e39c68ce3ff2ef6e0d78cf705bec4e

It's not the MSI refactor commit, so that's actually notable, I think.

[jessicah]

I should also note that this only happens with ICH9 as the configured chipset. Noticed as when switching chipset for my x86_gcc2-hybrid VM, I no longer got an IP address; switching back to legacy PIIX3, and networking works fine.

[korli]

could it be that the network device changed because of switching to legacy PIIX3?

[jessicah]

No, it's the same networking device.

[pulkomandy]

It is possibly depending on the (emulated) chipset supporting message-signaled interrupts or not. It is possibly related to #18177 but that was already a problem before the PCI refactor.

[BlueSky]

Not sure if this is related but my Haiku x86_64 VM running under VirtualBox 6.1 lost network connectivity after updating to hrev 56974. The configured network adapter is Intel Pro/1000 MT Desktop (82540EM) and uses the ipro1000 driver. The adapter is detected but DHCP discovery doesn't work, also no network connection possible with manually configured IPv4 settings. Last known working hrev is 56935. Syslog is attached

[BlueSky]

I can confirm that the network access works after switching the chipset to legacy PIIX3, as mentioned above

[waddlesplash]

Running on QEMU with -machine q35 -nic model=e1000 (i.e. ICH9), ping works fine here, there are packets being received. e1000e is broken as in #18177 but packets are still received nonetheless.

[waddlesplash]

Ran git bisect to test, resulted in a23ac85fa9e39c68ce3ff2ef6e0d78cf705bec4e, as suspected.

This commit actually does not change anything with respect to MSI interrupt handling, as far as I can tell; it's the commit before that which refactored MSI. If you test with the parent commit (e942269a1f114a4b90f7d254a26a47d3cda53959), does everything still work?

[waddlesplash]

Please retest after hrev57034, it may make a difference here, given that the MSI refactor commit is not the one that broke things.

[jessicah]

Retested, still didn't resolve the issue. Haven't tested mentioned parent commit explicitly yet, although I did use git bisect, so I expect it to have identified the faulty commit correctly.

[waddlesplash]

Please retest after hrev57045.

[jessicah]

Confirmed working now.