[net,v4,0/3] vsock/bpf: Handle races between sockmap update and connect() disconnecting

Message

Michal Luczaj March 17, 2025, 9:52 a.m. UTC

Signal delivery during connect() may disconnect an already established
socket. Problem is that such socket might have been placed in a sockmap
before the connection was closed.

PATCH 1 ensures this race won't lead to an unconnected vsock staying in the
sockmap. PATCH 2 selftests it. 

PATCH 3 fixes a related race. Note that selftest in PATCH 2 does test this
code as well, but winning this race variant may take more than 2 seconds,
so I'm not advertising it.

Signed-off-by: Michal Luczaj <[e-mail]>
---
Changes in v4:
- Selftest: send signal to only our own process
- Link to v3: https://lore.kernel.org/r/[e-mail]

Changes in v3:
- Selftest: drop unnecessary variable initialization and reorder the calls
- Link to v2: https://lore.kernel.org/r/[e-mail]

Changes in v2:
- Handle one more path of tripping the warning
- Add a selftest
- Collect R-b [Stefano]
- Link to v1: https://lore.kernel.org/r/[e-mail]

---
Michal Luczaj (3):
      vsock/bpf: Fix EINTR connect() racing sockmap update
      selftest/bpf: Add test for AF_VSOCK connect() racing sockmap update
      vsock/bpf: Fix bpf recvmsg() racing transport reassignment

 net/vmw_vsock/af_vsock.c                           | 10 ++-
 net/vmw_vsock/vsock_bpf.c                          | 24 ++++--
 .../selftests/bpf/prog_tests/sockmap_basic.c       | 99 ++++++++++++++++++++++
 3 files changed, 124 insertions(+), 9 deletions(-)
---
base-commit: da9e8efe7ee10e8425dc356a9fc593502c8e3933
change-id: 20250305-vsock-trans-signal-race-d62f7718d099

Best regards,

Comments

Michael S. Tsirkin March 19, 2025, 9:21 a.m. UTC | #1

On Mon, Mar 17, 2025 at 10:52:22AM +0100, Michal Luczaj wrote:
> Signal delivery during connect() may disconnect an already established
> socket. Problem is that such socket might have been placed in a sockmap
> before the connection was closed.
> 
> PATCH 1 ensures this race won't lead to an unconnected vsock staying in the
> sockmap. PATCH 2 selftests it. 
> 
> PATCH 3 fixes a related race. Note that selftest in PATCH 2 does test this
> code as well, but winning this race variant may take more than 2 seconds,
> so I'm not advertising it.
> 
> Signed-off-by: Michal Luczaj <[e-mail]>

vsock things:

Acked-by: Michael S. Tsirkin <[e-mail]>

> ---
> Changes in v4:
> - Selftest: send signal to only our own process
> - Link to v3: https://lore.kernel.org/r/[e-mail]
> 
> Changes in v3:
> - Selftest: drop unnecessary variable initialization and reorder the calls
> - Link to v2: https://lore.kernel.org/r/[e-mail]
> 
> Changes in v2:
> - Handle one more path of tripping the warning
> - Add a selftest
> - Collect R-b [Stefano]
> - Link to v1: https://lore.kernel.org/r/[e-mail]
> 
> ---
> Michal Luczaj (3):
>       vsock/bpf: Fix EINTR connect() racing sockmap update
>       selftest/bpf: Add test for AF_VSOCK connect() racing sockmap update
>       vsock/bpf: Fix bpf recvmsg() racing transport reassignment
> 
>  net/vmw_vsock/af_vsock.c                           | 10 ++-
>  net/vmw_vsock/vsock_bpf.c                          | 24 ++++--
>  .../selftests/bpf/prog_tests/sockmap_basic.c       | 99 ++++++++++++++++++++++
>  3 files changed, 124 insertions(+), 9 deletions(-)
> ---
> base-commit: da9e8efe7ee10e8425dc356a9fc593502c8e3933
> change-id: 20250305-vsock-trans-signal-race-d62f7718d099
> 
> Best regards,
> -- 
> Michal Luczaj <[e-mail]>