Re: [PATCH 1/4] nfs: use-after-free in svc_process_common()

From: Vasily Averin <vvs@virtuozzo.com>
To: "bfields@fieldses.org" <bfields@fieldses.org>
Cc: Trond Myklebust <trondmy@hammerspace.com>,
	"anna.schumaker@netapp.com" <anna.schumaker@netapp.com>,
	"khorenko@virtuozzo.com" <khorenko@virtuozzo.com>,
	"linux-nfs@vger.kernel.org" <linux-nfs@vger.kernel.org>,
	"eshatokhin@virtuozzo.com" <eshatokhin@virtuozzo.com>,
	"chuck.lever@oracle.com" <chuck.lever@oracle.com>,
	"jlayton@kernel.org" <jlayton@kernel.org>
Subject: Re: [PATCH 1/4] nfs: use-after-free in svc_process_common()
Date: Mon, 24 Dec 2018 00:03:25 +0300	[thread overview]
Message-ID: <e1877b8b-1472-b6ae-83e1-8605dfc5e589@virtuozzo.com> (raw)
In-Reply-To: <20181223205252.GA2613@fieldses.org>

On 12/23/18 11:52 PM, bfields@fieldses.org wrote:
> On Sat, Dec 22, 2018 at 08:46:55PM +0300, Vasily Averin wrote:
>> On 12/21/18 4:00 AM, bfields@fieldses.org wrote:
>>> On Tue, Dec 18, 2018 at 02:55:15PM +0000, Trond Myklebust wrote:
>>>> No. We don't care about xpt_flags for the back channel because there is
>>>> no "server transport". The actual transport is stored in the 'struct
>>>> rpc_rqst', and is the struct rpc_xprt corresponding to the client
>>>> socket or RDMA channel.
>>>>
>>>> IOW: All we really need in svc_process_common() is to be able to run
>>>> rqstp->rq_xprt->xpt_ops->xpo_prep_reply_hdr(), and that can be passed
>>>> either as a pointer to the struct svc_xprt_ops itself.
>>>
>>> For what it's worth, I'd rather get rid of that op--it's an awfully
>>> roundabout way just to do "svc_putnl(resv, 0);" in the tcp case.
>>
>> Do you mean that svc_create_xprt(serv, "tcp-bc", ...) was used ONLY to call 
>> svc_tcp_prep_reply_hdr() in svc_process_common() ?
>> And according call for rdma-bc does nothing useful at all? 
> 
> Right, in the rdma case it's:
> 
> 	void svc_rdma_prep_reply_hdr(struct svc_rqst *rqstp)
> 	{
> 	}
> 
>> I've just tried to remove svc_create_xprt() from xs_tcp_bc_up() and
>> just provide pointer to svc_tcp_prep_reply_hdr() in
>> svc_process_common() via per-netns sunrpc_net -- and seems it was
>> enough, my testcase worked correctly.
>>
>> Am I missed something probably?  Should we really remove
>> svc_create_xprt( "tcp/rdma-bc"...) related stuff? ?
> 
> Haven't looked carefully, but off the top of my head I can't see why
> that wouldn't work.

I've prepared new patch version removed svc_create_xprt( "tcp/rdma-bc"...)
as far as I see it works correctly.
I'm going to submit it tomorrow morning.

> I also tried some patches that replace that op by a flag bit (doesn't
> address the original problem here, just seemed like a simplification):
> 
> 	git://linux-nfs.org/~bfields/linux-topics.git
> 
> but I don't if that's compatible with what you've done.
> 
> --b.
> 