On Tue, Oct 13, 2009 at 11:56 AM, Sasha Khapyorsky <[email protected]> wrote:
> On 09:15 Tue 13 Oct , Hal Rosenstock wrote:
>> On Mon, Oct 12, 2009 at 2:45 PM, Sasha Khapyorsky <[email protected]>
>> wrote:
>> > On 10:46 Fri 09 Oct , Hal Rosenstock wrote:
>> >>
>> >> Per published MgtWG errata RefID 4576
>> >>
>> >> Signed-off-by: Hal Rosenstock <[email protected]>
>> >> ---
>> >> diff --git a/opensm/include/iba/ib_types.h b/opensm/include/iba/ib_types.h
>> >> index 25ed35f..c820372 100644
>> >> --- a/opensm/include/iba/ib_types.h
>> >> +++ b/opensm/include/iba/ib_types.h
>> >> @@ -4432,7 +4432,7 @@ typedef struct _ib_port_info {
>> >> ib_net16_t p_key_violations;
>> >> ib_net16_t q_key_violations;
>> >> uint8_t guid_cap;
>> >> - uint8_t subnet_timeout; /* cli_rereg(1b), resrv(2b), timeout(5b) */
>> >> + uint8_t subnet_timeout; /* cli_rereg(1b),
>> >> mcast_pkey_trap_suppr(1b), resrv(1b), timeout(5b) */
>> >> uint8_t resp_time_value;
>> >> uint8_t error_threshold; /* local phy errors(4b), overrun
>> >> errors(4b) */
>> >> ib_net16_t max_credit_hint;
>> >> @@ -5282,7 +5282,7 @@ ib_port_info_set_timeout(IN ib_port_info_t * const
>> >> p_pi,
>> >> {
>> >> CL_ASSERT(timeout <= 0x1F);
>> >> p_pi->subnet_timeout =
>> >> - (uint8_t) ((p_pi->subnet_timeout & 0x80) | (timeout & 0x1F));
>> >> + (uint8_t) ((p_pi->subnet_timeout & 0xC0) | (timeout & 0x1F));
>> >
>> > Why not 'p_pi->subnet_timeout & 0xe0' then (IOW why to bother with
>> > reserved bits reset)?
>>
>> Per this erratum (RefID 4576), this bit is no longer reserved. It is
>> now used for MulticastPKeyTrapSuppressionEnabled.
>
> It is bit 6 and I'm about bit 5 which still be reserved - We don't need
> to clear this. IOW I think that here (and in other similar places) should
> be:
>
> p_pi->subnet_timeout = (p_pi->subnet_timeout & 0xe0) | (timeout &
> 0x1f);
Yes, that's better for the future.
-- Hal
>>
>> >
>> >> }
>> >>
>> >> /*
>> >> @@ -5317,7 +5317,7 @@ ib_port_info_set_client_rereg(IN ib_port_info_t *
>> >> const p_pi,
>> >> CL_ASSERT(client_rereg <= 0x1);
>> >> p_pi->subnet_timeout =
>> >> (uint8_t) ((p_pi->
>> >> - subnet_timeout & 0x1F) | ((client_rereg << 7) &
>> >> 0x80));
>> >> + subnet_timeout & 0x5F) | ((client_rereg << 7) &
>> >> 0x80));
>> >
>> > Ditto.
>>
>> See above.
>>
>> >> }
>> >>
>> >> /*
>> >> @@ -5336,6 +5336,43 @@ ib_port_info_set_client_rereg(IN ib_port_info_t *
>> >> const p_pi,
>> >> * SEE ALSO
>> >> *********/
>> >>
>> >> +/****f* IBA Base: Types/ib_port_info_set_mcast_pkey_trap_suppress
>> >> +* NAME
>> >> +* ib_port_info_set_mcast_pkey_trap_suppress
>> >> +*
>> >> +* DESCRIPTION
>> >> +* Sets the encoded multicast pkey trap suppresion enabled bit value
>> >> +* in the PortInfo attribute.
>> >> +*
>> >> +* SYNOPSIS
>> >> +*/
>> >> +static inline void OSM_API
>> >> +ib_port_info_set_mcast_pkey_trap_suppress(IN ib_port_info_t * const p_pi,
>> >> + IN const uint8_t trap_suppress)
>> >> +{
>> >> + CL_ASSERT(trap_suppress <= 0x1);
>> >> + p_pi->subnet_timeout =
>> >> + (uint8_t) ((p_pi->
>> >> + subnet_timeout & 0x9F) | ((trap_suppress << 6) &
>> >> 0x40));
>> >
>> > Ditto.
>>
>> See above.
>>
>> > And whoudn't this:
>> >
>> > p_pi->subnet_timeout = (p_pi->subnet_timeout & 0xbf) |
>> > (trap_suppress << 6)
>> >
>> > just work?
>>
>> Sure, Do you want an updated patch with this change ?
>
> Yes, please.
>
> Sasha
>
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
