On Tue, Nov 26, 2013 at 10:25:20PM +0800, Kinglong Mee wrote:
> commit 5b6feee9608dce7afd2646f457c93e612526d1d8 forgot
> recording the back channel attrs in nfsd4_session.
>
> nfsd just check the back channel attars by check_backchannel_attrs,
> but do not record it in nfsd4_session in the latest kernel.
Whoops. So the current server is returning random uninitialized data in
those fields, or all zeros?
It would be easy enough to write a pynfs test for this, I think.
--b.
>
> Signed-off-by: Kinglong Mee <[email protected]>
> ---
> fs/nfsd/nfs4state.c | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c
> index 105d6fa..1aed9be 100644
> --- a/fs/nfsd/nfs4state.c
> +++ b/fs/nfsd/nfs4state.c
> @@ -999,6 +999,8 @@ static void init_session(struct svc_rqst *rqstp, struct nfsd4_session *new, stru
> spin_unlock(&nn->client_lock);
> memcpy(&new->se_fchannel, &cses->fore_channel,
> sizeof(struct nfsd4_channel_attrs));
> + memcpy(&new->se_bchannel, &cses->back_channel,
> + sizeof(struct nfsd4_channel_attrs));
> if (cses->flags & SESSION4_BACK_CHAN) {
> struct sockaddr *sa = svc_addr(rqstp);
> /*
> --
> 1.8.1.4
>
2013/12/7 J. Bruce Fields <[email protected]>:
> On Tue, Dec 03, 2013 at 10:08:59AM +0800, Kinglong Mee wrote:
>> On 12/03/2013 12:55 AM, J. Bruce Fields wrote:
>> > On Tue, Nov 26, 2013 at 10:25:20PM +0800, Kinglong Mee wrote:
>> >> commit 5b6feee9608dce7afd2646f457c93e612526d1d8 forgot
>> >> recording the back channel attrs in nfsd4_session.
>> >>
>> >> nfsd just check the back channel attars by check_backchannel_attrs,
>> >> but do not record it in nfsd4_session in the latest kernel.
>> >
>> > Whoops. So the current server is returning random uninitialized data in
>> > those fields, or all zeros?
>> >
>> > It would be easy enough to write a pynfs test for this, I think.
>>
>> No, server returns right data for nfsd4_encode_create_session encoding
>> back-channel attrs from struct nfsd4_create_session, not from nfsd4_session.
>>
>> Without this patch, se_bchannel in struct nfsd4_session contains data with all zeros.
>
> Oh, I see.
>
> But actually we never use se_bchannel for anything. Maybe the right
> thing is just to remove that field from nfsd4_session.
I'd like keep it in nfsd4_session.
Maybe nfsd will use it after a while.
thanks,
Kinglong Mee
>>
>> >>
>> >> Signed-off-by: Kinglong Mee <[email protected]>
>> >> ---
>> >> fs/nfsd/nfs4state.c | 2 ++
>> >> 1 file changed, 2 insertions(+)
>> >>
>> >> diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c
>> >> index 105d6fa..1aed9be 100644
>> >> --- a/fs/nfsd/nfs4state.c
>> >> +++ b/fs/nfsd/nfs4state.c
>> >> @@ -999,6 +999,8 @@ static void init_session(struct svc_rqst *rqstp, struct nfsd4_session *new, stru
>> >> spin_unlock(&nn->client_lock);
>> >> memcpy(&new->se_fchannel, &cses->fore_channel,
>> >> sizeof(struct nfsd4_channel_attrs));
>> >> + memcpy(&new->se_bchannel, &cses->back_channel,
>> >> + sizeof(struct nfsd4_channel_attrs));
>> >> if (cses->flags & SESSION4_BACK_CHAN) {
>> >> struct sockaddr *sa = svc_addr(rqstp);
>> >> /*
>> >> --
>> >> 1.8.1.4
>> >>
>> >
>>
On 12/03/2013 12:55 AM, J. Bruce Fields wrote:
> On Tue, Nov 26, 2013 at 10:25:20PM +0800, Kinglong Mee wrote:
>> commit 5b6feee9608dce7afd2646f457c93e612526d1d8 forgot
>> recording the back channel attrs in nfsd4_session.
>>
>> nfsd just check the back channel attars by check_backchannel_attrs,
>> but do not record it in nfsd4_session in the latest kernel.
>
> Whoops. So the current server is returning random uninitialized data in
> those fields, or all zeros?
>
> It would be easy enough to write a pynfs test for this, I think.
No, server returns right data for nfsd4_encode_create_session encoding
back-channel attrs from struct nfsd4_create_session, not from nfsd4_session.
Without this patch, se_bchannel in struct nfsd4_session contains data with all zeros.
thanks,
Kinglong Mee
>>
>> Signed-off-by: Kinglong Mee <[email protected]>
>> ---
>> fs/nfsd/nfs4state.c | 2 ++
>> 1 file changed, 2 insertions(+)
>>
>> diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c
>> index 105d6fa..1aed9be 100644
>> --- a/fs/nfsd/nfs4state.c
>> +++ b/fs/nfsd/nfs4state.c
>> @@ -999,6 +999,8 @@ static void init_session(struct svc_rqst *rqstp, struct nfsd4_session *new, stru
>> spin_unlock(&nn->client_lock);
>> memcpy(&new->se_fchannel, &cses->fore_channel,
>> sizeof(struct nfsd4_channel_attrs));
>> + memcpy(&new->se_bchannel, &cses->back_channel,
>> + sizeof(struct nfsd4_channel_attrs));
>> if (cses->flags & SESSION4_BACK_CHAN) {
>> struct sockaddr *sa = svc_addr(rqstp);
>> /*
>> --
>> 1.8.1.4
>>
>
On Tue, Dec 03, 2013 at 10:08:59AM +0800, Kinglong Mee wrote:
> On 12/03/2013 12:55 AM, J. Bruce Fields wrote:
> > On Tue, Nov 26, 2013 at 10:25:20PM +0800, Kinglong Mee wrote:
> >> commit 5b6feee9608dce7afd2646f457c93e612526d1d8 forgot
> >> recording the back channel attrs in nfsd4_session.
> >>
> >> nfsd just check the back channel attars by check_backchannel_attrs,
> >> but do not record it in nfsd4_session in the latest kernel.
> >
> > Whoops. So the current server is returning random uninitialized data in
> > those fields, or all zeros?
> >
> > It would be easy enough to write a pynfs test for this, I think.
>
> No, server returns right data for nfsd4_encode_create_session encoding
> back-channel attrs from struct nfsd4_create_session, not from nfsd4_session.
>
> Without this patch, se_bchannel in struct nfsd4_session contains data with all zeros.
Oh, I see.
But actually we never use se_bchannel for anything. Maybe the right
thing is just to remove that field from nfsd4_session.
--b.
>
> thanks,
> Kinglong Mee
>
> >>
> >> Signed-off-by: Kinglong Mee <[email protected]>
> >> ---
> >> fs/nfsd/nfs4state.c | 2 ++
> >> 1 file changed, 2 insertions(+)
> >>
> >> diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c
> >> index 105d6fa..1aed9be 100644
> >> --- a/fs/nfsd/nfs4state.c
> >> +++ b/fs/nfsd/nfs4state.c
> >> @@ -999,6 +999,8 @@ static void init_session(struct svc_rqst *rqstp, struct nfsd4_session *new, stru
> >> spin_unlock(&nn->client_lock);
> >> memcpy(&new->se_fchannel, &cses->fore_channel,
> >> sizeof(struct nfsd4_channel_attrs));
> >> + memcpy(&new->se_bchannel, &cses->back_channel,
> >> + sizeof(struct nfsd4_channel_attrs));
> >> if (cses->flags & SESSION4_BACK_CHAN) {
> >> struct sockaddr *sa = svc_addr(rqstp);
> >> /*
> >> --
> >> 1.8.1.4
> >>
> >
>