Re: NULL pointer deref inside xdp_do_flush due to bpf_net_ctx_get_all_used_flush_lists

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



"Yury Vostrikov" <mon@xxxxxxxxxxx> writes:

> Hi,
>
> I stumbled upon a NULL pointer derefence inside BPF code. The triggering condition is 
> message from OOM killer + netconsole. The crash happens at 
>
> 	u32 kern_flags = bpf_net_ctx->ri.kern_flags;
>
> line of bpf_net_ctx_get_all_used_flush_lists() function. bpf_net_ctx is NULL here. With trivial fix
>
> diff --git a/include/linux/filter.h b/include/linux/filter.h
> index 7d7578a8eac1..cba16bf307f7 100644
> --- a/include/linux/filter.h
> +++ b/include/linux/filter.h
> @@ -844,6 +844,9 @@ static inline void bpf_net_ctx_get_all_used_flush_lists(struct list_head **lh_ma
>                                                         struct list_head **lh_xsk)
>  {
>         struct bpf_net_context *bpf_net_ctx = bpf_net_ctx_get();
> +       WARN_ON(bpf_net_ctx == NULL);
> +       if (bpf_net_ctx == NULL)
> +               return;
>         u32 kern_flags = bpf_net_ctx->ri.kern_flags;
>         struct list_head *lh;
>  
> I get the following backtrace instead of crash:

[...]

> [  177.216474]  efx_poll+0x178/0x380 [sfc_siena]

Looks like the sfc driver is missing the context setup stuff entirely...

-Toke




[Index of Archives]     [Linux Samsung SoC]     [Linux Rockchip SoC]     [Linux Actions SoC]     [Linux for Synopsys ARC Processors]     [Linux NFS]     [Linux NILFS]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]


  Powered by Linux