netdev - Re: [PATCH v5 bpf-next 4/6] bpf, x64: rework pro/epilogue and tailcall handling in JIT

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <38407b6c-12a2-671f-9929-2ef39be62542@iogearbox.net>
Date:   Tue, 28 Jul 2020 23:33:12 +0200
From:   Daniel Borkmann <daniel@...earbox.net>
To:     Maciej Fijalkowski <maciej.fijalkowski@...el.com>, ast@...nel.org
Cc:     bpf@...r.kernel.org, netdev@...r.kernel.org, bjorn.topel@...el.com,
        magnus.karlsson@...el.com
Subject: Re: [PATCH v5 bpf-next 4/6] bpf, x64: rework pro/epilogue and
 tailcall handling in JIT

On 7/24/20 7:35 PM, Maciej Fijalkowski wrote:
> This commit serves two things:
> 1) it optimizes BPF prologue/epilogue generation
> 2) it makes possible to have tailcalls within BPF subprogram
[...]
[...]
> diff --git a/kernel/bpf/core.c b/kernel/bpf/core.c
> index 7be02e555ab9..d86a35474d7b 100644
> --- a/kernel/bpf/core.c
> +++ b/kernel/bpf/core.c
> @@ -773,7 +773,8 @@ int bpf_jit_add_poke_descriptor(struct bpf_prog *prog,
>   
>   	if (size > poke_tab_max)
>   		return -ENOSPC;
> -	if (poke->ip || poke->ip_stable || poke->adj_off)
> +	if (poke->tailcall_target || poke->tailcall_target_stable ||
> +	    poke->tailcall_bypass || poke->adj_off || poke->bypass_addr)
>   		return -EINVAL;

Hmm, I thought we've been through this from prior review rounds, but these sort of changes
break bisectability. You've already renamed the whole thing in patch 3/6 (poke->ip and the
poke->ip_stable). So if you've applied up to patch 3, then build breaks right here.

>   	switch (poke->reason) {
>