PCI | Programming Collective Intelligence Code | Collaboration library

We did a test to confirm that if a 64-bit PCI BAR would work on a 32-bit system.

We created a 32-bit Virtual Machine on a 64-bit system having a 64-bit PCI BAR device attached and did pass through of the PCI function (virtual function, which is also 64-bit) onto the VM. When using the lspci command on the VM, we saw 64-bit BAR mapping of the passed through device on the 32-bit VM. We also sent packets (testing if the pass through is working on the VM), which worked as they normally should.

Following is the result of the lspci command on the 32-bit VM: lspci Output

Nvidia RTX 3080 cards are based on the Ampere architecture for which compatible CUDA version start with 11.x.

Up gradation of tensorflow from 2.3 to 2.4 or 2.5 will solve above issue. For more details you can refer here.

I think the best way is to look for the PCI Express capability, which is in the regular capability space, not the extended space. The presence of this capability indicates a PCIe device. The capability ID is 0x10.

Intel Xeon E5-2620 is Sandy Bridge CPU which officially supports AVX and not AVX2.

DPDK 20.11 meson build, ninja -C build will generate code with AVX instructions and not AVX2. But (Based on the live debug) PKTGEN forces the compiler to add AVX2 to be inserted, thus causing illegal instruction.

Solution: edit meson.build in line 22

from

Connecting to SQL Server with an OAuth token requires use of a pre-connection attribute (basically a pointer to the token string). There is an open feature request at the odbc Github repo for this. I encourage you to upvote it, hopefully if it's popular enough it will get implemented.

A QEMU image running the 'user-mode' networking (as in your command line example) already has access to the host machine. It can access it either via any (non-loopback) IP address the host has, or by using the special 'gateway' IP address. If you're using the default 10.0.2.0/24 network setting then the 'gateway' is 10.0.2.2. I haven't confirmed but suspect that for a non-default net setting it will still be on .2, so in this example 192.168.76.2.

You cannot literally make 'localhost' in the guest point to the host PC, because 'localhost' for the guest is the guest itself, and having it point somewhere else would likely confuse software running in the guest.

The hostfwd option is for forwarding connections from the outside world to a server which is running on the guest. "hostfwd=tcp::HOSTPORT-:GUESTPORT" says "QEMU should listen on the host on port HOSTPORT; whenever a connection arrives there, it should forward it to the guest's port GUESTPORT (which hopefully has a server listening there)".

You seem to be running a server on the host. You can't have more than one thing listening on a particular port on one machine, so either the python3 server program can listen on port 28009 and respond to connections there, or QEMU can listen on port 28009 to respond to connections there (forwarding them to the guest), but not both at once. Whichever is started second will complain that something's already using the port.

If you want to run a server on the host and connect to it from the guest, you don't need any QEMU options at all. QEMU's 'usermode' networking will allow guest programs to make connections outwards to any IP address (including the wider internet but also directly to the host), so if you are trying to run a client on the guest and a server on the host that should just work. You can tell the guest client to connect either to the host's real IP address or you can use the special 'gateway' IP address 10.0.2.2 which is how the host machine appears on the fake network that the guest sees.

After reboot everything works ! It seems that after installing cuda & tensorflow I had to restart the PC.

[EDIT: update to reflect only using OVS-DPDK and iperf3 on container]

There are multiple ways one can connect 2 dockers to talk directly with each other using to run iperf3.

1. Virtual Interface like TAP-1|MAC-VETH-1 from Docker-1 is connected to TAP-2| MAC-VETH-2 via Linux Bridge.
2. Virtual port-1 (TAP|memif) from OVS-DPDK to Docker-1 and virtual port-2 (tap|memif) to Docker-2 via DPDK-OVS

For scenario 2 one needs to add TAP interface to OVS. because end application iperf3 is using Kernel Stack for TCP|UDP termination. One can use the below settings (modified based on OVS-DPDK version) to achieve the result.