The branch main has been updated by tuexen:

URL: 
https://cgit.FreeBSD.org/src/commit/?id=f343f036c780cfd8e7f243b7708d96b006a257d7

commit f343f036c780cfd8e7f243b7708d96b006a257d7
Author:     Timo Völker <[email protected]>
AuthorDate: 2026-08-16 18:31:33 +0000
Commit:     Michael Tuexen <[email protected]>
CommitDate: 2026-08-16 18:38:36 +0000

    epair: allow disabling receive checksum offloading
    
    Allow disabling RXCSUM and RXCSUM6 on an epair interface. If disabled,
    epair unsets the mbuf flags that indicate a valid checksum when
    transferring a packet from one epair end to the other. This gives a
    user in a jail the power to control whether the user wants to use the
    result of a previous validation (by a physical interface) or not.
    
    Reviewed by:            kp, tuexen
    MFC after:              1 month
    MFC to:                 stable/15
    Differential Revision:  https://reviews.freebsd.org/D58786
---
 share/man/man4/epair.4 | 10 +++++++---
 sys/net/if_epair.c     | 24 ++++++++++++++++--------
 2 files changed, 23 insertions(+), 11 deletions(-)

diff --git a/share/man/man4/epair.4 b/share/man/man4/epair.4
index ba42106d45f8..f28f68bdfc19 100644
--- a/share/man/man4/epair.4
+++ b/share/man/man4/epair.4
@@ -25,7 +25,7 @@
 .\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 .\" SUCH DAMAGE.
 .\"
-.Dd January 30, 2026
+.Dd August 16, 2026
 .Dt EPAIR 4
 .Os
 .Sh NAME
@@ -111,8 +111,12 @@ configured on top of it.
 .Pp
 The
 .Nm
-has RXCSUM and RXCSUM6 enabled because it may receive a packet where the
-checksum has already been validated by a physical interface.
+supports RXCSUM and RXCSUM6 but does not validate a checksum.
+In case an interface previously validated the checksum,
+.Nm
+leaves the flags set that indicate a valid checksum if RXCSUM and RXCSUM6 are
+enabled or unsets these flags if RXCSUM and RXCSUM6 are disabled.
+Note that RXCSUM and RXCSUM6 can only be enabled or disabled as a pair.
 .Pp
 The
 .Nm
diff --git a/sys/net/if_epair.c b/sys/net/if_epair.c
index 5addd43b6190..c3354bc76e64 100644
--- a/sys/net/if_epair.c
+++ b/sys/net/if_epair.c
@@ -242,6 +242,9 @@ epair_prepare_mbuf(struct mbuf *m, struct ifnet *src_ifp)
        epair_clear_mbuf(m);
        if_setrcvif(m, src_ifp);
        M_SETFIB(m, src_ifp->if_fib);
+       if ((if_getcapenable(src_ifp) & IFCAP_RXCSUM) == 0)
+               m->m_pkthdr.csum_flags &= ~(CSUM_L3_CALC | CSUM_L3_VALID |
+                   CSUM_L4_CALC | CSUM_L4_VALID);
 
        MPASS(m->m_nextpkt == NULL);
        MPASS((m->m_pkthdr.csum_flags & CSUM_SND_TAG) == 0);
@@ -455,7 +458,7 @@ epair_ioctl(struct ifnet *ifp, u_long cmd, caddr_t data)
 {
        struct epair_softc *sc;
        struct ifreq *ifr;
-       int error;
+       int error, cap;
 
        ifr = (struct ifreq *)data;
        switch (cmd) {
@@ -484,14 +487,19 @@ epair_ioctl(struct ifnet *ifp, u_long cmd, caddr_t data)
                break;
        case SIOCSIFCAP:
                /*
-                * Enable/disable capabilities as requested, besides
-                * IFCAP_RXCSUM(_IPV6), which always remain enabled.
-                * Incoming packets may have the mbuf flag CSUM_DATA_VALID set.
-                * Without IFCAP_RXCSUM(_IPV6), this flag would have to be
-                * removed, which does not seem helpful.
+                * Enable/disable capabilities as requested, but treat
+                * IFCAP_RXCSUM and IFCAP_RXCSUM_IPV6 special as they can only
+                * be set or unset as pair.
                 */
-               ifp->if_capenable = ifr->ifr_reqcap | IFCAP_RXCSUM |
-                   IFCAP_RXCSUM_IPV6;
+               cap = ifr->ifr_reqcap;
+               if (((cap & IFCAP_RXCSUM) == 0) !=
+                   ((cap & IFCAP_RXCSUM_IPV6) == 0)) {
+                       if ((ifp->if_capenable & IFCAP_RXCSUM) == 0)
+                               cap |= (IFCAP_RXCSUM | IFCAP_RXCSUM_IPV6);
+                       else
+                               cap &= ~(IFCAP_RXCSUM | IFCAP_RXCSUM_IPV6);
+               }
+               ifp->if_capenable = cap;
                epair_caps_changed(ifp);
                VLAN_CAPABILITIES(ifp);
                /*

Reply via email to