RHEL Engineering is moving the tracking of its product development work on RHEL 6 through RHEL 9 to Red Hat Jira (issues.redhat.com). If you're a Red Hat customer, please continue to file support cases via the Red Hat customer portal. If you're not, please head to the "RHEL project" in Red Hat Jira and file new tickets here. Individual Bugzilla bugs in the statuses "NEW", "ASSIGNED", and "POST" are being migrated throughout September 2023. Bugs of Red Hat partners with an assigned Engineering Partner Manager (EPM) are migrated in late September as per pre-agreed dates. Bugs against components "kernel", "kernel-rt", and "kpatch" are only migrated if still in "NEW" or "ASSIGNED". If you cannot log in to RH Jira, please consult article #7032570. That failing, please send an e-mail to the RH Jira admins at rh-issues@redhat.com to troubleshoot your issue as a user management inquiry. The email creates a ServiceNow ticket with Red Hat. Individual Bugzilla bugs that are migrated will be moved to status "CLOSED", resolution "MIGRATED", and set with "MigratedToJIRA" in "Keywords". The link to the successor Jira issue will be found under "Links", have a little "two-footprint" icon next to it, and direct you to the "RHEL project" in Red Hat Jira (issue links are of type "https://issues.redhat.com/browse/RHEL-XXXX", where "X" is a digit). This same link will be available in a blue banner at the top of the page informing you that that bug has been migrated.
Bug 1275949 - [Doc] Edit SR-IOV content
Summary: [Doc] Edit SR-IOV content
Keywords:
Status: CLOSED CURRENTRELEASE
Alias: None
Product: Red Hat Enterprise Linux 7
Classification: Red Hat
Component: doc-Virtualization_Deployment_and_Administration_Guide
Version: 7.2
Hardware: Unspecified
OS: Unspecified
unspecified
unspecified
Target Milestone: rc
: ---
Assignee: Yehuda Zimmerman
QA Contact:
URL:
Whiteboard:
Depends On:
Blocks: 1201058
TreeView+ depends on / blocked
 
Reported: 2015-10-28 07:27 UTC by Dayle Parker
Modified: 2019-03-06 01:17 UTC (History)
7 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Environment:
Last Closed: 2017-03-29 18:03:51 UTC
Target Upstream Version:
Embargoed:


Attachments (Terms of Use)

Comment 6 Laine Stump 2016-12-12 15:13:37 UTC
Troubleshoot.xml:

The comment about coredump is inaccurate - while a normal coredump isn't possible on a guest with an assigned PCI device, it *can* be done if you specify the option "--memory-only" then you will be able to get a coredump of a guest that has PCI devices assigned.


PCI_Assignment.xml:

* According to Alex, the blanket statement that "Virtual machines that use the Xeon E3-1200 series chip set, do not support SR-IOV" is inaccurate. Due to "quirk" entries in the kernel, device assignment can usually be supported with that CPU (as well as with i3/i5/i7 CPUs that aren't mentioned in this document, but *are* mentioned in Alex's blog post that you reference in this doc). I'll leave it up to him to provide proper text for that paragraph.

* In several places, this file refers to "mapping" a VF to a virtual machine, e.g.:

  The hypervisor can map one or more Virtual Functions to a virtual machine

That should instead use the word "assign" (there are other uses of the word "map" that probably *are* correct):

  The hypervisor can assign one or more Virtual Functions to a virtual machine


(In general the description of SRIOV seems more complicated than necessary, but I don't have the presence of mind to provide an alternative :-))

* Under "Advantages of SRIOV" you can add this:

"When an SRIOV VF is assigned to a virtual machine, it can be configured to (transparently to the virtual machine) place all network traffic going out that VF onto a oarticular VLAN; the virtual machine is unable to detect that its traffic is being tagged for a VLAN, and will be unable to change or eliminate this tagging."

*In "Using SR-IOV" section, I don't think it's necessary to say "libvirt-0.9.10 and newer", since libvirt has been beyond that version since RHEL 6.4 or so and this documentation is for RHEL7.

* The section "Start the SR-IOV kernel modules" should be unnecessary - I have never seen a system where this didn't happen automatically at boot when the card was installed (Alex, do you agree?)

* the section "Activate Virtual Functions" should just tell them to run:

    echo ${num_vfs} > /sys/class/net/${pf_name}/device/sriov_numvfs

(i.e. the same command that they're told to add to /etc/rd.d/rc.local in the following section, "Make the Virtual Functions persistent".)

* the short note after the sample output of "virsh nodedev-list" calls the names of the devices "serial numbers". Thos are not serial numbers, they are PCI addresses.

*in "get device details" section - don't call it "advanced output", call it device details".

* In the example that starts with "This example adds the Virtual Function <systemitem>pci_0000_0b_10_0</systemitem> to the virtual machine in Step 9", it would probably make things less confusing if:

1) you made an example that used one of the VFs listed in the previous "nodedev-dumpxml" example, rather than a device at some other random address. So, e.g. use the VF device pci_0000_03_10_2, which has this address:

     <address domain='0x0000' bus='0x03' slot='0x10' function='0x2'/>

(be sure to include the leading 0x on each number, as they are hexadecimal (in the eaxmple youo have now, youo convert the numbers to decimal, but don't explain that. It's best to just keep them as hex). Also, the description of the first example describes a vlan tag of 42, but that isn't shown until the 2nd example.

Comment 7 Laine Stump 2016-12-12 15:16:20 UTC
Alex - what is the proper thing to be said about ACS at the top of the diff for PCI_Assignment.xml? This is what is said currently:

"Virtual machines that use the Xeon E3-1200 series chip set, do not support SR-IOV. More information can be found on Intel's website or in this <ulink url="http://vfio.blogspot.com/2014/08/iommu-groups-inside-and-out.html">article</ulink>."

(You had told me in IRC that this isn't accurate)


Note You need to log in before you can comment on or make changes to this bug.