Re: [PATCH net v2 3/3] net: skbuff: don't touch shared zerocopy state in skb_tx_error()

Hi Norbert,

I independently tested this patch as part of the v2 series using a
separate, non-destructive page-cache canary reproducer.

The test exercised the cloned OVS path while the original skb continued
through the remaining actions:

  clone(userspace(pid=1)),2

Before the fix, the root-owned read-only page-cache canary changed. With
v2 applied, the canary remained unchanged. No kernel warning, oops, or
panic was observed.

Tested-by: Jongmin Jang <[email protected]>

Best regards,
Jongmin Jang