From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:50311) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1ZNFqS-0000VP-2B for qemu-devel@nongnu.org; Thu, 06 Aug 2015 03:44:09 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1ZNFqN-0000SD-ES for qemu-devel@nongnu.org; Thu, 06 Aug 2015 03:44:08 -0400 Received: from mx1.redhat.com ([209.132.183.28]:56873) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1ZNFqN-0000Rt-6B for qemu-devel@nongnu.org; Thu, 06 Aug 2015 03:44:03 -0400 Message-ID: <55C3103A.9030605@redhat.com> Date: Thu, 06 Aug 2015 15:43:54 +0800 From: Jason Wang MIME-Version: 1.0 References: <1438677044-13030-1-git-send-email-yanghy@cn.fujitsu.com> <1438677044-13030-3-git-send-email-yanghy@cn.fujitsu.com> <55C307B5.8060201@redhat.com> <55C30B1D.1070304@cn.fujitsu.com> <55C30CBD.9080705@redhat.com> <55C30E3C.3010106@cn.fujitsu.com> In-Reply-To: <55C30E3C.3010106@cn.fujitsu.com> Content-Type: text/plain; charset=windows-1252 Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] [PATCH v4 02/11] init/cleanup of netfilter object List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Yang Hongyang , qemu-devel@nongnu.org Cc: thuth@redhat.com, zhang.zhanghailiang@huawei.com, lizhijian@cn.fujitsu.com, eddie.dong@intel.com, mrhines@linux.vnet.ibm.com, stefanha@redhat.com On 08/06/2015 03:35 PM, Yang Hongyang wrote: > On 08/06/2015 03:29 PM, Jason Wang wrote: >> >> >> On 08/06/2015 03:22 PM, Yang Hongyang wrote: >>> On 08/06/2015 03:07 PM, Jason Wang wrote: >>>> >>>> >>>> On 08/04/2015 04:30 PM, Yang Hongyang wrote: >>>>> QTAILQ_ENTRY global_list but used by filter layer, so that we can >>>>> manage all filters together. >>>>> QTAILQ_ENTRY next used by netdev, filter belongs to the specific >>>>> netdev is >>>>> in this queue. >>>>> This is mostly the same with init/cleanup of netdev object. >>>>> >>>>> Signed-off-by: Yang Hongyang >>>>> --- >>>>> include/net/filter.h | 39 +++++++++++++ >>>>> include/net/net.h | 1 + >>>>> include/qemu/typedefs.h | 1 + >>>>> net/filter.c | 147 >>>>> ++++++++++++++++++++++++++++++++++++++++++++++++ >>>>> net/net.c | 1 + >>>>> qapi-schema.json | 37 ++++++++++++ >>>>> 6 files changed, 226 insertions(+) >>>>> >>>>> diff --git a/include/net/filter.h b/include/net/filter.h >>>>> index 4242ded..9aafe08 100644 >>>>> --- a/include/net/filter.h >>>>> +++ b/include/net/filter.h >>>>> @@ -9,7 +9,46 @@ >>>>> #define QEMU_NET_FILTER_H >>>>> >>>>> #include "qemu-common.h" >>>>> +#include "qemu/typedefs.h" >>>>> + >>>>> +/* the netfilter chain */ >>>>> +enum { >>>>> + NET_FILTER_IN, >>>>> + NET_FILTER_OUT, >>>>> + NET_FILTER_ALL, >>>>> +}; >>>>> + >>>>> +typedef void (FilterCleanup) (NetFilterState *); >>>>> +/* >>>>> + * Return: >>>>> + * 0: finished handling the packet, we should continue >>>>> + * size: filter stolen this packet, we stop pass this packet >>>>> further >>>>> + */ >>>>> +typedef ssize_t (FilterReceiveIOV)(NetFilterState *, NetClientState >>>>> *sender, >>>>> + unsigned flags, const struct >>>>> iovec *, int); >>>> >>>> Please name all parameters. >>> >>> Ok, thanks. >>> >>>> >>>>> + >>>>> +typedef struct NetFilterInfo { >>>>> + NetFilterOptionsKind type; >>>>> + size_t size; >>>>> + FilterCleanup *cleanup; >>>>> + FilterReceiveIOV *receive_iov; >>>>> +} NetFilterInfo; >>>>> + >>>>> +struct NetFilterState { >>>>> + NetFilterInfo *info; >>>>> + char *model; >>>> >>>> Looks like model is never used? >>> >>> It can be used when we want to find filters by model. For example, >>> when we need to find all "buffer" filter, and release all buffered >>> packets. >> >> But this is not implemented in this series. > > true. > >> And I don't get why you need >> a such command. Management should keep track of this. > > There's a patch in v1 which introduce an API use this, and I dropped that > because there's no user currently, but that API I suppose to be used > by MC or COLO which can do periodic checkpoint, to release packets. > the API is filter_buffer_release_all() IIRC... I see, let's start with simple case to make review and merging easier. You can re-introduce them when it has real users.