[PATCH] thaw FC local port

[PATCH] thaw FC local port

[Hillf Danton]

The state machine of local port is freezed if it could not get ready,
so thawing is necessary.

Signed-off-by: Hillf Danton <dhillf@gmail.com>
---

--- a/drivers/scsi/libfc/fc_lport.c	2010-11-01 19:54:12.000000000 +0800
+++ b/drivers/scsi/libfc/fc_lport.c	2010-12-17 00:08:06.000000000 +0800
@@ -1514,11 +1514,11 @@ void fc_lport_enter_flogi(struct fc_lpor

 	fc_lport_state_enter(lport, LPORT_ST_FLOGI);

-	if (lport->point_to_multipoint) {
-		if (lport->port_id)
+	if (lport->point_to_multipoint)
+		if (lport->port_id) {
 			fc_lport_enter_ready(lport);
-		return;
-	}
+			return;
+		}

 	fp = fc_frame_alloc(lport, sizeof(struct fc_els_flogi));
 	if (!fp)

Re: [PATCH] thaw FC local port

[Robert Love]

On Thu, 2010-12-16 at 08:07 -0800, Hillf Danton wrote:
> The state machine of local port is freezed if it could not get ready,
> so thawing is necessary.
> 
Please start putting a prefix "fcoe: ", "libfc: " or "libfcoe: " on your
patch titles. Also, please provide more meaningful commit messages. I
believe that I've pointed out these two issues with your patches before.

> Signed-off-by: Hillf Danton <dhillf@gmail.com>
> ---
> 
> --- a/drivers/scsi/libfc/fc_lport.c	2010-11-01 19:54:12.000000000 +0800
> +++ b/drivers/scsi/libfc/fc_lport.c	2010-12-17 00:08:06.000000000 +0800
> @@ -1514,11 +1514,11 @@ void fc_lport_enter_flogi(struct fc_lpor
> 
>  	fc_lport_state_enter(lport, LPORT_ST_FLOGI);
> 
> -	if (lport->point_to_multipoint) {
> -		if (lport->port_id)
> +	if (lport->point_to_multipoint)
> +		if (lport->port_id) {
>  			fc_lport_enter_ready(lport);
> -		return;
> -	}
> +			return;
> +		}
> 
>  	fp = fc_frame_alloc(lport, sizeof(struct fc_els_flogi));
>  	if (!fp)

We do not want to proceed with sending a FLOGI here when we're in
point-to-multipoint (VN2VN) mode. VN2VN is a mode where there is no
fabric and we FLOGI directly into the remote port. So for VN2VN the
FLOGI happens in the rport layer.

//Rob

Re: [PATCH] thaw FC local port

[Hillf Danton]

On Fri, Dec 17, 2010 at 7:18 AM, Robert Love <robert.w.love@intel.com> wrote:
> On Thu, 2010-12-16 at 08:07 -0800, Hillf Danton wrote:
>> The state machine of local port is freezed if it could not get ready,
>> so thawing is necessary.
>>
> Please start putting a prefix "fcoe: ", "libfc: " or "libfcoe: " on your
> patch titles. Also, please provide more meaningful commit messages. I
> believe that I've pointed out these two issues with your patches before.
>
>> Signed-off-by: Hillf Danton <dhillf@gmail.com>
>> ---
>>
>> --- a/drivers/scsi/libfc/fc_lport.c   2010-11-01 19:54:12.000000000 +0800
>> +++ b/drivers/scsi/libfc/fc_lport.c   2010-12-17 00:08:06.000000000 +0800
>> @@ -1514,11 +1514,11 @@ void fc_lport_enter_flogi(struct fc_lpor
>>
>>       fc_lport_state_enter(lport, LPORT_ST_FLOGI);
>>
>> -     if (lport->point_to_multipoint) {
>> -             if (lport->port_id)
>> +     if (lport->point_to_multipoint)
>> +             if (lport->port_id) {
>>                       fc_lport_enter_ready(lport);
>> -             return;
>> -     }
>> +                     return;
>> +             }
>>
>>       fp = fc_frame_alloc(lport, sizeof(struct fc_els_flogi));
>>       if (!fp)
>
> We do not want to proceed with sending a FLOGI here when we're in
> point-to-multipoint (VN2VN) mode. VN2VN is a mode where there is no
> fabric and we FLOGI directly into the remote port. So for VN2VN the
> FLOGI happens in the rport layer.

How will lport get out of the LPORT_ST_FLOGI state then?

Thanks

Hillf
>
> //Rob
>
>
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH] thaw FC local port

[Robert Love]

On Fri, 2010-12-17 at 05:46 -0800, Hillf Danton wrote:
> On Fri, Dec 17, 2010 at 7:18 AM, Robert Love <robert.w.love@intel.com> wrote:
> > On Thu, 2010-12-16 at 08:07 -0800, Hillf Danton wrote:
> >> The state machine of local port is freezed if it could not get ready,
> >> so thawing is necessary.
> >>
> > Please start putting a prefix "fcoe: ", "libfc: " or "libfcoe: " on your
> > patch titles. Also, please provide more meaningful commit messages. I
> > believe that I've pointed out these two issues with your patches before.
> >
> >> Signed-off-by: Hillf Danton <dhillf@gmail.com>
> >> ---
> >>
> >> --- a/drivers/scsi/libfc/fc_lport.c   2010-11-01 19:54:12.000000000 +0800
> >> +++ b/drivers/scsi/libfc/fc_lport.c   2010-12-17 00:08:06.000000000 +0800
> >> @@ -1514,11 +1514,11 @@ void fc_lport_enter_flogi(struct fc_lpor
> >>
> >>       fc_lport_state_enter(lport, LPORT_ST_FLOGI);
> >>
> >> -     if (lport->point_to_multipoint) {
> >> -             if (lport->port_id)
> >> +     if (lport->point_to_multipoint)
> >> +             if (lport->port_id) {
> >>                       fc_lport_enter_ready(lport);
> >> -             return;
> >> -     }
> >> +                     return;
> >> +             }
> >>
> >>       fp = fc_frame_alloc(lport, sizeof(struct fc_els_flogi));
> >>       if (!fp)
> >
> > We do not want to proceed with sending a FLOGI here when we're in
> > point-to-multipoint (VN2VN) mode. VN2VN is a mode where there is no
> > fabric and we FLOGI directly into the remote port. So for VN2VN the
> > FLOGI happens in the rport layer.
> 
> How will lport get out of the LPORT_ST_FLOGI state then?
> 

fc_lport_set_local_id

Re: [PATCH] thaw FC local port

[Joe Eykholt]

On 12/17/10 5:46 AM, Hillf Danton wrote:
> On Fri, Dec 17, 2010 at 7:18 AM, Robert Love <robert.w.love@intel.com> wrote:
>> On Thu, 2010-12-16 at 08:07 -0800, Hillf Danton wrote:
>>> The state machine of local port is freezed if it could not get ready,
>>> so thawing is necessary.
>>>
>> Please start putting a prefix "fcoe: ", "libfc: " or "libfcoe: " on your
>> patch titles. Also, please provide more meaningful commit messages. I
>> believe that I've pointed out these two issues with your patches before.
>>
>>> Signed-off-by: Hillf Danton <dhillf@gmail.com>
>>> ---
>>>
>>> --- a/drivers/scsi/libfc/fc_lport.c   2010-11-01 19:54:12.000000000 +0800
>>> +++ b/drivers/scsi/libfc/fc_lport.c   2010-12-17 00:08:06.000000000 +0800
>>> @@ -1514,11 +1514,11 @@ void fc_lport_enter_flogi(struct fc_lpor
>>>
>>>       fc_lport_state_enter(lport, LPORT_ST_FLOGI);
>>>
>>> -     if (lport->point_to_multipoint) {
>>> -             if (lport->port_id)
>>> +     if (lport->point_to_multipoint)
>>> +             if (lport->port_id) {
>>>                       fc_lport_enter_ready(lport);
>>> -             return;
>>> -     }
>>> +                     return;
>>> +             }
>>>
>>>       fp = fc_frame_alloc(lport, sizeof(struct fc_els_flogi));
>>>       if (!fp)
>>
>> We do not want to proceed with sending a FLOGI here when we're in
>> point-to-multipoint (VN2VN) mode. VN2VN is a mode where there is no
>> fabric and we FLOGI directly into the remote port. So for VN2VN the
>> FLOGI happens in the rport layer.
> 
> How will lport get out of the LPORT_ST_FLOGI state then?

By getting a FLOGI accept.  Why else would it need to?
What end-user problem are you trying to solve.

Also, I suggest *not* copying linux-scsi.  This can be handled by
those subscribed to devel@open-fcoe.org.

	Joe

Re: [PATCH] thaw FC local port

[Hillf Danton]

On Sat, Dec 18, 2010 at 3:04 AM, Robert Love <robert.w.love@intel.com> wrote:
> On Fri, 2010-12-17 at 05:46 -0800, Hillf Danton wrote:
>> On Fri, Dec 17, 2010 at 7:18 AM, Robert Love <robert.w.love@intel.com> wrote:
>> > On Thu, 2010-12-16 at 08:07 -0800, Hillf Danton wrote:
>> >> The state machine of local port is freezed if it could not get ready,
>> >> so thawing is necessary.
>> >>
>> > Please start putting a prefix "fcoe: ", "libfc: " or "libfcoe: " on your
>> > patch titles. Also, please provide more meaningful commit messages. I
>> > believe that I've pointed out these two issues with your patches before.
>> >
>> >> Signed-off-by: Hillf Danton <dhillf@gmail.com>
>> >> ---
>> >>
>> >> --- a/drivers/scsi/libfc/fc_lport.c   2010-11-01 19:54:12.000000000 +0800
>> >> +++ b/drivers/scsi/libfc/fc_lport.c   2010-12-17 00:08:06.000000000 +0800
>> >> @@ -1514,11 +1514,11 @@ void fc_lport_enter_flogi(struct fc_lpor
>> >>
>> >>       fc_lport_state_enter(lport, LPORT_ST_FLOGI);
>> >>
>> >> -     if (lport->point_to_multipoint) {
>> >> -             if (lport->port_id)
>> >> +     if (lport->point_to_multipoint)
>> >> +             if (lport->port_id) {
>> >>                       fc_lport_enter_ready(lport);
>> >> -             return;
>> >> -     }
>> >> +                     return;
>> >> +             }
>> >>
>> >>       fp = fc_frame_alloc(lport, sizeof(struct fc_els_flogi));
>> >>       if (!fp)
>> >
>> > We do not want to proceed with sending a FLOGI here when we're in
>> > point-to-multipoint (VN2VN) mode. VN2VN is a mode where there is no
>> > fabric and we FLOGI directly into the remote port. So for VN2VN the
>> > FLOGI happens in the rport layer.
>>
>> How will lport get out of the LPORT_ST_FLOGI state then?
>>
>
> fc_lport_set_local_id
>

You are right, thanks.

Hillf
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html