Date: Fri, 1 Apr 2016 15:54:22 +0530
From: "Naveen N. Rao" <naveen.n.rao@linux.vnet.ibm.com>
To: linux-kernel@vger.kernel.org, linuxppc-dev@lists.ozlabs.org
Cc: oss@buserror.net, Matt Evans <matt@ozlabs.org>,
        Michael Ellerman <mpe@ellerman.id.au>,
        Paul Mackerras <paulus@samba.org>, Alexei Starovoitov <ast@fb.com>,
        "David S. Miller" <davem@davemloft.net>,
        Ananth N Mavinakayanahalli <ananth@in.ibm.com>
Subject: Re: [RFC PATCH 0/6] eBPF JIT for PPC64
Message-ID: <20160401102422.GE17907@naverao1-tp.ibm.com>
References: <cover.1459504222.git.naveen.n.rao@linux.vnet.ibm.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <cover.1459504222.git.naveen.n.rao@linux.vnet.ibm.com>
User-Agent: Mutt/1.5.24 (2015-08-30)
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1075
Lines: 30

On 2016/04/01 03:28PM, Naveen N Rao wrote:
> Implement extended BPF JIT for ppc64. We retain the classic BPF JIT for
> ppc32 and move ppc64 BE/LE to use the new JIT. Classic BPF filters will
> be converted to extended BPF (see convert_filter()) and JIT'ed with the
> new compiler.
> 
> Most of the existing macros are retained and fixed/enhanced where
> appropriate. Patches 1-4 are geared towards this.
> 
> Patch 5 breaks out the classic BPF JIT specifics into a separate
> bpf_jit32.h header file, while retaining all the generic instruction
> macros in bpf_jit.h. Most of these macros can potentially be generalized
> and moved to more common code (tagged with a TODO in patch 6).
> 
> Patch 6 implements eBPF JIT for ppc64.

As a comparison, here are the test results with the BPF test suite 
kernel module:

With the classic BPF JIT:
test_bpf: Summary: 291 PASSED, 0 FAILED, [85/283 JIT'ed]

and with the extended BPF JIT:
test_bpf: Summary: 291 PASSED, 0 FAILED, [234/283 JIT'ed]

As noted in patch 6, there are still a few more instructions to be 
JIT'ed.


- Naveen