The branch main has been updated by kbowling: URL: https://cgit.FreeBSD.org/src/commit/?id=f589d1fc36cebcc298eb708bf5eae57bd1d491d7
commit f589d1fc36cebcc298eb708bf5eae57bd1d491d7 Author: Kevin Bowling <[email protected]> AuthorDate: 2026-09-14 23:51:01 +0000 Commit: Kevin Bowling <[email protected]> CommitDate: 2026-09-15 02:07:11 +0000 ixv: Recover E610 Hyper-V carrier without the PCI link query Current Windows PF drivers do not emulate the E610 link-status query at PCI configuration offset 0x209. The installed Intel ixw 1.8.54.0 driver returns zero even with the physical port connected, leaving the VF at no carrier after an otherwise successful hn handoff. When the query returns zero, use VFLINKS for carrier only. Its speed field is not the negotiated E610 line rate: the tested 1 Gb/s port reports the default 10 Gb/s encoding. Publish an unknown speed instead, including in the verbose link-up diagnostic, and document this fallback. Keep nonzero PCI results authoritative and reject all-ones reads. Other Hyper-V families and the native E610 mailbox path are unchanged. A zero result cannot distinguish an unsupported query from a down link; in either case this fallback relies on the legacy carrier indication. The connected and disconnected ports reported VFLINKS.UP set and clear, respectively, while both returned zero from the PCI query. MFC after: 2 weeks Sponsored by: BBOX.io --- share/man/man4/ixv.4 | 5 ++++- sys/dev/ixgbe/if_ixv.c | 10 +++++++--- sys/dev/ixgbe/ixgbe_hv_vf.c | 17 +++++++++++++++++ 3 files changed, 28 insertions(+), 4 deletions(-) diff --git a/share/man/man4/ixv.4 b/share/man/man4/ixv.4 index 907ff30103ec..f49d3eb88aeb 100644 --- a/share/man/man4/ixv.4 +++ b/share/man/man4/ixv.4 @@ -28,7 +28,7 @@ .\" NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS .\" SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. .\" -.Dd September 6, 2026 +.Dd September 14, 2026 .Dt IXV 4 .Os .Sh NAME @@ -76,6 +76,9 @@ PCI configuration space instead of the Intel PF/VF mailbox. The driver uses one queue pair in this environment. MAC address, VLAN, multicast filter, and receive mode policies remain under host control. +On E610, when the host does not support the PCI link-status query, +the driver uses the virtual-function carrier indication and reports an +unknown link speed. .Pp The virtual function cannot configure the physical media. Only automatic media selection is supported, and reported link state and diff --git a/sys/dev/ixgbe/if_ixv.c b/sys/dev/ixgbe/if_ixv.c index a5e8a9b39de1..c3d56e2ae913 100644 --- a/sys/dev/ixgbe/if_ixv.c +++ b/sys/dev/ixgbe/if_ixv.c @@ -1370,9 +1370,13 @@ ixv_if_update_admin_status(if_ctx_t ctx) if (bootverbose) { baudrate = ixgbe_link_speed_to_baudrate( sc->link_speed); - device_printf(dev, - "Link is up %ju Mbps Full Duplex\n", - (uintmax_t)(baudrate / IF_Mbps(1))); + if (baudrate != 0) + device_printf(dev, + "Link is up %ju Mbps Full Duplex\n", + (uintmax_t)(baudrate / IF_Mbps(1))); + else + device_printf(dev, + "Link is up (speed unknown)\n"); } sc->link_active = true; iflib_link_state_change(ctx, LINK_STATE_UP, diff --git a/sys/dev/ixgbe/ixgbe_hv_vf.c b/sys/dev/ixgbe/ixgbe_hv_vf.c index 6ce994ff06f1..b65a2438adee 100644 --- a/sys/dev/ixgbe/ixgbe_hv_vf.c +++ b/sys/dev/ixgbe/ixgbe_hv_vf.c @@ -148,6 +148,23 @@ ixgbevf_hv_check_mac_link_vf(struct ixgbe_hw *hw, ixgbe_link_speed *speed, goto out; links_reg = ixgbevf_hv_read_links(hw); + if (hw->mac.type == ixgbe_mac_E610_vf) { + *speed = IXGBE_LINK_SPEED_UNKNOWN; + if (links_reg == UINT32_MAX) + goto out; + if (links_reg == 0) { + /* + * Some hosts do not emulate the PCI link-status query. + * VFLINKS still reports carrier, but its speed is not the + * negotiated line rate. Do not report its default 10 Gb/s. + */ + links_reg = IXGBE_READ_REG(hw, IXGBE_VFLINKS); + if (links_reg != UINT32_MAX && + (links_reg & IXGBE_LINKS_UP) != 0) + mac->get_link_status = false; + goto out; + } + } if (!(links_reg & IXGBE_LINKS_UP)) goto out;
