linux-kernel - Re: [PATCH] NFSD: Fix the share reservation conflict to courteous server logic in nfs4_upgrade

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]

Date:   Sun, 30 Oct 2022 09:32:50 -0700
From:   dai.ngo@...cle.com
To:     Chuck Lever III <chuck.lever@...cle.com>,
        Christophe JAILLET <christophe.jaillet@...adoo.fr>
Cc:     Jeff Layton <jlayton@...nel.org>,
        Bruce Fields <bfields@...ldses.org>,
        LKML <linux-kernel@...r.kernel.org>,
        "kernel-janitors@...r.kernel.org" <kernel-janitors@...r.kernel.org>,
        Linux NFS Mailing List <linux-nfs@...r.kernel.org>
Subject: Re: [PATCH] NFSD: Fix the share reservation conflict to courteous
 server logic in nfs4_upgrade_open()

On 10/30/22 8:26 AM, Chuck Lever III wrote:
>
>> On Oct 30, 2022, at 3:26 AM, Christophe JAILLET <christophe.jaillet@...adoo.fr> wrote:
>>
>> 'status != nfserr_share_denied' is known to be true because we test
>> 'status == nfs_ok' the line just above.
>>
>> So nfs4_resolve_deny_conflicts_locked() can never be called.
>>
>> Fix the logic and avoid the dead code.
>>
>> Fixes: 3d6942715180 ("NFSD: add support for share reservation conflict to courteous server")
>> Signed-off-by: Christophe JAILLET <christophe.jaillet@...adoo.fr>
>> ---
>> This patch is speculative.
>> It is compile tested only.
>>
>> REVIEW WITH CARE.
>> ---
>> fs/nfsd/nfs4state.c | 14 ++++++--------
>> 1 file changed, 6 insertions(+), 8 deletions(-)
>>
>> diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c
>> index 1ded89235111..de0565e9485c 100644
>> --- a/fs/nfsd/nfs4state.c
>> +++ b/fs/nfsd/nfs4state.c
>> @@ -5260,15 +5260,13 @@ nfs4_upgrade_open(struct svc_rqst *rqstp, struct nfs4_file *fp,
>> 	spin_lock(&fp->fi_lock);
>> 	status = nfs4_file_check_deny(fp, open->op_share_deny);
> I agree there's dead code here. I believe the bug is the first check is
> supposed to be "if (status != nfs_ok)". I will let Dai have a look at
> this to confirm.

Yes, it's actually a bug when nfs4_file_check_deny returns
nfserr_share_denied we won't try to resolve the conflict at all.

Thanks for catching this!
-Dai

>
> But, in the fix, let's replace this logic with "switch (status) { }".
>
>
>> 	if (status == nfs_ok) {
>> -		if (status != nfserr_share_denied) {
>> -			set_deny(open->op_share_deny, stp);
>> -			fp->fi_share_deny |=
>> +		set_deny(open->op_share_deny, stp);
>> +		fp->fi_share_deny |=
>> 				(open->op_share_deny & NFS4_SHARE_DENY_BOTH);
>> -		} else {
>> -			if (nfs4_resolve_deny_conflicts_locked(fp, false,
>> -					stp, open->op_share_deny, false))
>> -				status = nfserr_jukebox;
>> -		}
>> +	} else if (status == nfserr_share_denied) {
>> +		if (nfs4_resolve_deny_conflicts_locked(fp, false, stp,
>> +				open->op_share_deny, false))
>> +			status = nfserr_jukebox;
>> 	}
>> 	spin_unlock(&fp->fi_lock);
>>
>> -- 
>> 2.34.1
>>
> --
> Chuck Lever
>
>
>