Re: [bpf, xdp] headroom - was: Re: Question about to KMSAN: uninit-value in can_receive

[Jakub Kicinski <kuba@kernel.org>]

On Sat, 3 Jan 2026 13:20:34 +0100 Oliver Hartkopp wrote:
> Hello Jakub,
> 
> thanks for stepping in!
> 
> On 02.01.26 21:04, Jakub Kicinski wrote:
> 
> > You're asking the wrong person, IIUC Andrii is tangentially involved
> > in XDP (via bpf links?):
> >   
> (..)
> > 
> > Without looking too deeply - XDP has historically left the new space
> > uninitialized after push, expecting programs to immediately write the
> > headers in that space. syzbot had run into this in the past but I can't
> > find any references to past threads quickly :(  
> 
> To identify Andrii I mainly looked into the code with 'git blame' that 
> led to this problematic call chain:
> 
>    pskb_expand_head+0x226/0x1a60 net/core/skbuff.c:2275
>    netif_skb_check_for_xdp net/core/dev.c:5081 [inline]
>    netif_receive_generic_xdp net/core/dev.c:5112 [inline]
>    do_xdp_generic+0x9e3/0x15a0 net/core/dev.c:5180
> 
> Having in mind that the syzkaller refers to 
> 6.13.0-rc7-syzkaller-00039-gc3812b15000c I wonder if we can leave this 
> report as-is, as the problem might be solved in the meantime??
> 
> In any case I wonder, if we should add some code to re-check if the 
> headroom of the CAN-related skbs is still consistent and not changed in 
> size by other players. And maybe add some WARN_ON_ONCE() before dropping 
> the skb then.
> 
> When the skb headroom is not safe to be used we need to be able to 
> identify and solve it.

Ugh, I should have looked at the report. The struct can_skb_priv
business is highly unconventional for the networking stack.
Would it be possible to kmalloc() this info and pass it to the socket
via shinfo->destructor_arg?