[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAM1=_QSnRvN=rW3W79TCoNe38pgQ4-5Dmu4uRWCV5hqX4nwE_Q@mail.gmail.com>
Date: Thu, 29 Jul 2021 23:30:42 +0200
From: Johan Almbladh <johan.almbladh@...finetworks.com>
To: Yonghong Song <yhs@...com>
Cc: Alexei Starovoitov <ast@...nel.org>,
Daniel Borkmann <daniel@...earbox.net>,
Andrii Nakryiko <andrii@...nel.org>,
Martin KaFai Lau <kafai@...com>,
Song Liu <songliubraving@...com>,
John Fastabend <john.fastabend@...il.com>,
KP Singh <kpsingh@...nel.org>,
Tony Ambardar <Tony.Ambardar@...il.com>,
Networking <netdev@...r.kernel.org>, bpf <bpf@...r.kernel.org>
Subject: Re: [PATCH 01/14] bpf/tests: Add BPF_JMP32 test cases
On Thu, Jul 29, 2021 at 12:31 AM Yonghong Song <yhs@...com> wrote:
> > + /* BPF_JMP32 | BPF_JGT | BPF_X */
> > + {
> > + "JMP32_JGT_X",
> > + .u.insns_int = {
> > + BPF_ALU32_IMM(BPF_MOV, R0, 0xfffffffe),
> > + BPF_ALU32_IMM(BPF_MOV, R1, 0xffffffff),
> > + BPF_JMP32_REG(BPF_JGT, R0, R1, 1),
>
> Maybe change the offset from 1 to 2? Otherwise, this may jump to
> BPF_JMP32_REG(BPF_JGT, R0, R1, 1)
> which will just do the same comparison and jump to BTT_EXIT_INSN()
> which will also have R0 = 0xfffffffe at the end.
You are right. All BPF_X versions should have the first jump offset
incremented by one to account for the extra MOV that is not present in
the BPF_K version of the test. I'll correct it.
Powered by blists - more mailing lists