[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20190204.091508.799387970409692312.davem@davemloft.net>
Date: Mon, 04 Feb 2019 09:15:08 -0800 (PST)
From: David Miller <davem@...emloft.net>
To: heiko.carstens@...ibm.com
Cc: ast@...nel.org, schwidefsky@...ibm.com, jiong.wang@...ronome.com,
linux-s390@...r.kernel.org, netdev@...r.kernel.org
Subject: Re: [PATCH] s390: bpf: fix JMP32 code-gen
From: Heiko Carstens <heiko.carstens@...ibm.com>
Date: Mon, 4 Feb 2019 16:44:55 +0100
> Commit 626a5f66da0d19 ("s390: bpf: implement jitting of JMP32") added
> JMP32 code-gen support for s390. However it triggers the warning below
> due to some unusual gotos in the original s390 bpf jit code.
>
> Add a couple of additional "is_jmp32" initializations to fix this.
> Also fix the wrong opcode for the "llilf" instruction that was
> introduced with the same commit.
>
> arch/s390/net/bpf_jit_comp.c: In function 'bpf_jit_insn':
> arch/s390/net/bpf_jit_comp.c:248:55: warning: 'is_jmp32' may be used uninitialized in this function [-Wmaybe-uninitialized]
> _EMIT6(op1 | reg(b1, b2) << 16 | (rel & 0xffff), op2 | mask); \
> ^
> arch/s390/net/bpf_jit_comp.c:1211:8: note: 'is_jmp32' was declared here
> bool is_jmp32 = BPF_CLASS(insn->code) == BPF_JMP32;
>
> Fixes: 626a5f66da0d19 ("s390: bpf: implement jitting of JMP32")
> Cc: Jiong Wang <jiong.wang@...ronome.com>
> Cc: Martin Schwidefsky <schwidefsky@...ibm.com>
> Signed-off-by: Heiko Carstens <heiko.carstens@...ibm.com>
Acked-by: David S. Miller <davem@...emloft.net>
Powered by blists - more mailing lists