linux-kernel - Re: [V9fs-developer] [PATCH 2/2] fs/9p: detecting invalid options as much as possible

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-Id: <923F6291-1E42-42DF-9312-5720050037E3@gmx.com>
Date:   Mon, 16 Apr 2018 19:37:32 +0800
From:   cgxu519@....com
To:     Dominique Martinet <asmadeus@...ewreck.org>
Cc:     "cgxu519@....com" <cgxu519@....com>, ericvh@...il.com,
        rminnich@...dia.gov, lucho@...kov.net,
        v9fs-developer@...ts.sourceforge.net, linux-kernel@...r.kernel.org
Subject: Re: [V9fs-developer] [PATCH 2/2] fs/9p: detecting invalid options as
 much as possible

Hi Dominique,

Thanks for your quick reply and comment.


在 2018年4月16日，下午3:56，Dominique Martinet <asmadeus@...ewreck.org> 写道：
> 
> Chengguang Xu wrote on Mon, Apr 16, 2018:
>> 		default:
>> +			p9_debug(P9_DEBUG_ERROR,
>> +				"unrecognized mount option \"%s\" or missing value\n",
>> +					p);
>> 			continue;
> 
> The problem with that is that the same options are passed to the vfs,
> net and transport init later on - none of which know the full subset of
> valid options to tell what option has been recognized or not.
> 
> Unless we do some backward-incompatible breakage of passing all the
> net/transport options in its own option (e.g. net=foo:bar:moo) there
> unfortunately is no nice way of fixing this now.

Yes, I agree with you.


> 
> 
> (I don't mind the rest of the patches, although I'd say if we hit ENOMEM
> there is likely trouble going on so I'm not so sure about hiding it if
> there also is an EINVAL, but I don't really care)

The initial motivation of hiding ENOMEM here is when reproducing the error，
the error code may change by system condition(more accurately memory condition),
after this patch the error code will be persistent. However, as you pointed out,
when we hit ENOMEM there would be a serious trouble, so in real life maybe we
cannot benefit from it. What do you think? 

Thanks,
Chengguang.