Re: [PATCH] nfsd: prevent NULL ptr derefs on fault injection

[Bryan Schumaker]

On 11/27/2012 12:01 PM, J. Bruce Fields wrote:
> On Tue, Nov 27, 2012 at 11:31:11AM -0500, Sasha Levin wrote:
>> A recent patch series has moved hashtable initialization to when the net
>> struct is initialized.
>>
>> When injecting faults, we tried accessing the hashtables even if the struct
>> wasn't really initialized (nfsd wasn't in use) - this caused a NULL ptr
>> deref.
> 
> Thanks, adding Bryan to cc.--b.

I was just looking over this :).  Unfortunately, this patch changes code that my most recent patch set removes so my patches will need to change again if this goes in first.  I'm looking for the best place to put this check to avoid having to change each of the forget_something() functions.

- Bryan
> 
>>
>> A simple test would be:
>>
>> 	echo 1 > /sys/kernel/debug/nfsd/forget_locks
>>
>> Signed-off-by: Sasha Levin <sasha.levin@xxxxxxxxxx>
>> ---
>>  fs/nfsd/netns.h     | 3 +++
>>  fs/nfsd/nfs4state.c | 9 +++++++++
>>  2 files changed, 12 insertions(+)
>>
>> diff --git a/fs/nfsd/netns.h b/fs/nfsd/netns.h
>> index 227b93e..c5806a57 100644
>> --- a/fs/nfsd/netns.h
>> +++ b/fs/nfsd/netns.h
>> @@ -83,5 +83,8 @@ struct nfsd_net {
>>  	struct delayed_work laundromat_work;
>>  };
>>  
>> +/* Simple check to find out if a given net was properly initialized */
>> +#define nfsd_netns_ready(nn) ((nn)->sessionid_hashtbl)
>> +
>>  extern int nfsd_net_id;
>>  #endif /* __NFSD_NETNS_H__ */
>> diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c
>> index e75872f..0e7428c 100644
>> --- a/fs/nfsd/nfs4state.c
>> +++ b/fs/nfsd/nfs4state.c
>> @@ -4598,6 +4598,9 @@ void nfsd_forget_clients(u64 num)
>>  	int count = 0;
>>  	struct nfsd_net *nn = net_generic(current->nsproxy->net_ns, nfsd_net_id);
>>  
>> +	if (!nfsd_netns_ready(nn))
>> +		return;
>> +
>>  	nfs4_lock_state();
>>  	list_for_each_entry_safe(clp, next, &nn->client_lru, cl_lru) {
>>  		expire_client(clp);
>> @@ -4643,6 +4646,9 @@ void nfsd_forget_locks(u64 num)
>>  	int count;
>>  	struct nfsd_net *nn = net_generic(&init_net, nfsd_net_id);
>>  
>> +	if (!nfsd_netns_ready(nn))
>> +		return;
>> +
>>  	nfs4_lock_state();
>>  	count = nfsd_release_n_owners(num, false, release_lockowner_sop, nn);
>>  	nfs4_unlock_state();
>> @@ -4655,6 +4661,9 @@ void nfsd_forget_openowners(u64 num)
>>  	int count;
>>  	struct nfsd_net *nn = net_generic(&init_net, nfsd_net_id);
>>  
>> +	if (!nfsd_netns_ready(nn))
>> +		return;
>> +
>>  	nfs4_lock_state();
>>  	count = nfsd_release_n_owners(num, true, release_openowner_sop, nn);
>>  	nfs4_unlock_state();
>> -- 
>> 1.8.0
>>

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/