mbox series

[net,v2,0/2] Silence bad offload warning when sending UDP GSO with IPv6 extension headers

Message ID 20240801-udp-gso-egress-from-tunnel-v2-0-9a2af2f15d8d@cloudflare.com (mailing list archive)
Headers show
Series Silence bad offload warning when sending UDP GSO with IPv6 extension headers | expand

Message

Jakub Sitnicki Aug. 1, 2024, 1:52 p.m. UTC
This series addresses a recent regression report from syzbot [1].
Please see patch 1 description for details.

[1] https://lore.kernel.org/all/000000000000e1609a061d5330ce@google.com/

Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
---
Changes in v2:
- Contain the fix inside the GSO stack after discussing with Willem
- Rework tests after realizing the regression has nothing to do with tunnels
- Link to v1: https://lore.kernel.org/r/20240725-udp-gso-egress-from-tunnel-v1-0-5e5530ead524@cloudflare.com

---
Jakub Sitnicki (2):
      gso: Skip bad offload detection when device supports requested GSO
      selftests/net: Add coverage for UDP GSO with IPv6 extension headers

 net/core/gso.c                       |  6 ++++--
 tools/testing/selftests/net/udpgso.c | 25 ++++++++++++++++++++++++-
 2 files changed, 28 insertions(+), 3 deletions(-)

Comments

Jakub Kicinski Aug. 2, 2024, 1:36 a.m. UTC | #1
On Thu, 01 Aug 2024 15:52:52 +0200 Jakub Sitnicki wrote:
> This series addresses a recent regression report from syzbot [1].
> Please see patch 1 description for details.

The test doesn't seem super happy in netdev CI:

https://netdev-3.bots.linux.dev/vmksft-net-dbg/results/709100/79-udpgso-sh/stdout
https://netdev-3.bots.linux.dev/vmksft-net/results/709101/78-udpgso-sh/stdout
https://netdev-3.bots.linux.dev/vmksft-net-dbg/results/708921/75-udpgso-sh/stdout
https://netdev-3.bots.linux.dev/vmksft-net/results/708921/78-udpgso-sh/stdout
Jakub Sitnicki Aug. 2, 2024, 11:20 a.m. UTC | #2
On Thu, Aug 01, 2024 at 06:36 PM -07, Jakub Kicinski wrote:
> On Thu, 01 Aug 2024 15:52:52 +0200 Jakub Sitnicki wrote:
>> This series addresses a recent regression report from syzbot [1].
>> Please see patch 1 description for details.
>
> The test doesn't seem super happy in netdev CI:
>
> https://netdev-3.bots.linux.dev/vmksft-net-dbg/results/709100/79-udpgso-sh/stdout
> https://netdev-3.bots.linux.dev/vmksft-net/results/709101/78-udpgso-sh/stdout
> https://netdev-3.bots.linux.dev/vmksft-net-dbg/results/708921/75-udpgso-sh/stdout
> https://netdev-3.bots.linux.dev/vmksft-net/results/708921/78-udpgso-sh/stdout

Embarassing. I must have not recompiled the tests after tweaking it.
Sorry for the oversight. I will deal with it.

Reproduces for me locally:

[pid   507] setsockopt(4, SOL_IPV6, IPV6_HOPOPTS, NULL, 0) = 0
[pid   507] recvfrom(3, 0x55fcf63813a0, 65535, 0, NULL, NULL) = -1 EAGAIN (Resource temporarily unavailable)