Re: [PATCH net 1/1] tcp: check skb is non-NULL in tcp_rto_delta_us()

From: Josh Hunt
Date: Fri Aug 23 2024 - 15:43:59 EST




On 8/22/24 11:55 PM, Eric Dumazet wrote:
!-------------------------------------------------------------------|
This Message Is From an External Sender
This message came from outside your organization.
|-------------------------------------------------------------------!

On Fri, Aug 23, 2024 at 5:34 AM Josh Hunt <johunt@xxxxxxxxxx> wrote:

There have been multiple occassions where we have crashed in this path
because packets_out suggested there were packets on the write or retransmit
queues, but in fact there weren't leading to a NULL skb being dereferenced.
While we should fix that root cause we should also just make sure the skb
is not NULL before dereferencing it. Also add a warn once here to capture
some information if/when the problem case is hit again.

Signed-off-by: Josh Hunt <johunt@xxxxxxxxxx>

Hi Josh

We do not want a patch series of one patch, with the stack trace in
the cover letter.
Please send a standalone patch, with all the information in its changelog.

1) Add Neal Cardwell in the CC list.

2) Are you using TCP_REPAIR by any chance ?

3) Please double check your kernel has these fixes.

commit 1f85e6267caca44b30c54711652b0726fadbb131 tcp: do not send
empty skb from tcp_write_xmit()
commit 0c175da7b0378445f5ef53904247cfbfb87e0b78 tcp: prohibit
TCP_REPAIR_OPTIONS if data was already sent


Thanks Eric. I will resend and also check the commits you mentioned. I didn't include the writeup in the patch submission b/c it was rather long and detailed, but will include it in a v2.

Josh