> On Nov 18, 2016, at 00:01, NeilBrown <neilb@xxxxxxxx> wrote:
>
>
> Ping?
> No sign of this in linux-next, and no replies….
I’d like some confirmation from the NetApp folks on this. They are the main stakeholders for the pNFS files implementation. I don’t think we need an equivalent for flexfiles.
>
> Thanks,
> NeilBrown
>
>
> On Fri, Oct 21 2016, NeilBrown wrote:
>
>>
>> Various places assume that if nfs4_fl_prepare_ds() turns a non-NULL
>> 'ds', then ds->ds_clp will also be non-NULL.
>>
>> This is not necessasrily true in the case when the process received a
>> fatal signal while nfs4_pnfs_ds_connect is waiting in
>> nfs4_wait_ds_connect(). In that case ->ds_clp may not be set, and the
>> devid may not recently have been marked unavailable.
>>
>> So add a test for ->ds_clp == NULL and return NULL in that case.
>>
>> Fixes: c23266d532b4 ("NFS4.1 Fix data server connection race")
>> Signed-off-by: NeilBrown <neilb@xxxxxxxx>
>> ---
>> fs/nfs/filelayout/filelayoutdev.c | 3 ++-
>> 1 file changed, 2 insertions(+), 1 deletion(-)
>>
>> diff --git a/fs/nfs/filelayout/filelayoutdev.c b/fs/nfs/filelayout/filelayoutdev.c
>> index 4946ef40ba87..85ef38f9765f 100644
>> --- a/fs/nfs/filelayout/filelayoutdev.c
>> +++ b/fs/nfs/filelayout/filelayoutdev.c
>> @@ -283,7 +283,8 @@ nfs4_fl_prepare_ds(struct pnfs_layout_segment *lseg, u32 ds_idx)
>> s->nfs_client->cl_rpcclient->cl_auth->au_flavor);
>>
>> out_test_devid:
>> - if (filelayout_test_devid_unavailable(devid))
>> + if (ret->ds_clp == NULL ||
>> + filelayout_test_devid_unavailable(devid))
>> ret = NULL;
>> out:
>> return ret;
>> --
>> 2.10.1
��.n��������+%������w��{.n�����{��w��� jg��� �����ݢj����G���� ���j:+v���w�m������w�������h�����٥
