diff options
| author | James Tucker <james@tailscale.com> | 2024-08-06 14:40:17 -0700 |
|---|---|---|
| committer | James Tucker <james@tailscale.com> | 2024-08-06 14:40:17 -0700 |
| commit | ab9d3474d400590c53863f6d69f6e2bd98fe2d90 (patch) | |
| tree | fd55b0edff38c39894461a2dfe2a571d7c7bdc43 /go.mod | |
| parent | a93dc6cdb10bcc87742011d3746660f06cf8cd1e (diff) | |
| download | tailscale-raggi/linux6644.tar.xz tailscale-raggi/linux6644.zip | |
go.mod: bump wireguard-go to 6.6.44 mitigation testraggi/linux6644
Linux 6.6.44 introduced a regression in GSO handling that spuriously
returns EINVAL for tail packets shorter than gso_size. The referenced
wireguard-go patch works around this by not including packets matching
this shape in the offload, which should be rare, but are dependent on
traffic patterns.
This patch tests out fine in my lab, creating a branch for original
reporters to also test with.
Updates #13041
Signed-off-by: James Tucker <james@tailscale.com>
Diffstat (limited to 'go.mod')
| -rw-r--r-- | go.mod | 2 |
1 files changed, 1 insertions, 1 deletions
@@ -80,7 +80,7 @@ require ( github.com/tailscale/peercred v0.0.0-20240214030740-b535050b2aa4 github.com/tailscale/web-client-prebuilt v0.0.0-20240226180453-5db17b287bf1 github.com/tailscale/wf v0.0.0-20240214030419-6fbb0a674ee6 - github.com/tailscale/wireguard-go v0.0.0-20240731203015-71393c576b98 + github.com/tailscale/wireguard-go v0.0.0-20240806212730-80ee55ffe40c github.com/tailscale/xnet v0.0.0-20240729143630-8497ac4dab2e github.com/tc-hib/winres v0.2.1 github.com/tcnksm/go-httpstat v0.2.0 |