[Openib-windows] [IBAL] 128 bit key patch
Alex Estrin
alex.estrin at qlogic.com
Mon Dec 4 06:44:28 PST 2006
Yossi,
Your latest patch seem fixed both problems.
Thanks,
Alex
P.S.
Do you have Silverstorm switch chassis
with slots for IO modules( series 5000, 7000)?
> -----Original Message-----
> From: Yossi Leybovich [mailto:sleybo at mellanox.co.il]
> Sent: Sunday, December 03, 2006 6:58 AM
> To: Alex Estrin
> Cc: openib-windows at openib.org
> Subject: RE: [Openib-windows] [IBAL] 128 bit key patch
>
>
> Hi
>
> I reproduce the problem with the IOC_REMOVE event and you are
> right , I
> missed the ca_guid when I created the remove events
> This patch fix the problem .
>
> About the second problem:
> I guess you mean that the installer does not find any
> matching device ID
> in the INF.
> The patch for the 128 bit key extension does not change the
> generic IOC
> device ID (which is based on io_class/io_subclass/protocol...)
>
> Can you check in the setupapi.log what is the Device ID that Windows
> cant find?
> Can you collects AL logs?
> Can you check if this patch fix the first problem too?
>
> BTW
>
> Here in Mellanox we have only simple target that use one IOC and one
> service entry,
> Is it possible that we get the target you use for our testing and
> regression ?
>
>
> Thanks
> Yossi
>
> Index: kernel/al_ioc_pnp.c
> ===================================================================
> --- kernel/al_ioc_pnp.c (revision 1840)
> +++ kernel/al_ioc_pnp.c (working copy)
> @@ -1701,6 +1701,7 @@
> /* Decrement the IOC PnP manager's query count. */
> if( !cl_atomic_dec( &gp_ioc_pnp->query_cnt ) )
> cl_async_proc_queue( gp_async_pnp_mgr, &gp_ioc_pnp->async_item );
> + AL_EXIT( AL_DBG_PNP );
> return;
> }
>
> @@ -3060,6 +3061,7 @@
> cl_memclr( &rec, sizeof(ib_pnp_iou_rec_t) );
> rec.pnp_rec.pnp_event = IB_PNP_IOU_REMOVE;
> rec.pnp_rec.guid = p_iou->guid;
> + rec.pnp_rec.ca_guid = p_iou->ca_guid;
>
> event.rec_size = sizeof(ib_pnp_iou_rec_t);
> event.p_rec = (ib_pnp_rec_t*)&rec;
> @@ -3154,6 +3156,7 @@
> cl_memclr( &rec, sizeof(ib_pnp_ioc_rec_t) );
> rec.pnp_rec.pnp_event = IB_PNP_IOC_REMOVE;
> rec.pnp_rec.guid = p_ioc->profile.ioc_guid;
> + rec.pnp_rec.ca_guid = p_ioc->p_iou->ca_guid;
>
> event.rec_size = sizeof(ib_pnp_ioc_rec_t);
> event.p_rec = (ib_pnp_rec_t*)&rec;
> @@ -3310,3 +3313,4 @@
>
>
>
> +
>
>
>
> ________________________________
>
> From: Alex Estrin [mailto:alex.estrin at qlogic.com]
> Sent: Saturday, December 02, 2006 8:03 PM
> To: Yossi Leybovich
> Subject: RE: [Openib-windows] [IBAL] 128 bit key patch
>
>
>
>
> ________________________________
>
> From: Yossi Leybovich [mailto:sleybo at mellanox.co.il]
> Sent: Sat 12/2/2006 12:39 PM
> To: Alex Estrin
> Subject: RE: [Openib-windows] [IBAL] 128 bit key patch
>
>
>
>
>
> ________________________________
>
> From: Alex Estrin [mailto:alex.estrin at qlogic.com]
> Sent: Saturday, December 02, 2006 4:11 PM
> To: Yossi Leybovich
> Subject: [Openib-windows] [IBAL] 128 bit key patch
>
>
> Hi Yossi,
>
> It seem extended key patch does not work for me.
> Currently I see following symptoms on a system with one
> HCA/one port in use,
> and module I work with reports 3 IOC x 2 service
> entries:
>
> Difficulty with installation of IOCs.
> Sometimes It could not find matching Device Id( somehow
> mostly happend with a third IOC)
> untill I used clean HW profile.
> [Yossi Leybovich] Does the each IOC report different
> GUID ?
>
> Yes, each IOC reports unique guid.
>
> IOC_REMOVE event does not get through if I pull IB
> cable.
> Manual disable/uninstall devices still works though.
> [Yossi Leybovich] what version of SVN do you use ?, I
> place fix for that on version 552 ?
> 1.4.0. Yes I meant patch you placed in
> Nov. 28th rev 550- 552
>
>
> Thanks,
> Alex
> ________________________________
>
> From: Yossi Leybovich [mailto:sleybo at mellanox.co.il]
> Sent: Wed 11/22/2006 1:49 AM
> To: Alex Estrin
> Subject: RE: [Openib-windows] [IBAL][Patch] reference
> counter for IOC obj in IOC PnP manager
>
>
>
>
>
>
>
More information about the ofw
mailing list