linux-hotplug.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* Missing SAS phy enumeration in newer versions of udev path_id
@ 2012-03-30  7:59 Nils Carlson
  2012-03-30  8:23 ` Hannes Reinecke
  0 siblings, 1 reply; 2+ messages in thread
From: Nils Carlson @ 2012-03-30  7:59 UTC (permalink / raw)
  To: linux-hotplug

Hi,

We recently upgraded our version of udev and I saw that the symlinks 
under /dev/disk/by-path/ have changed since udev started using the 
path_id C code to no longer include phy enumeration for SAS disks. Was 
this change intentional? Would anyone mind if i re-create the behavior 
of the old bash script as we were using it?

Best Regards,
Nils Carlson


^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: Missing SAS phy enumeration in newer versions of udev path_id
  2012-03-30  7:59 Missing SAS phy enumeration in newer versions of udev path_id Nils Carlson
@ 2012-03-30  8:23 ` Hannes Reinecke
  0 siblings, 0 replies; 2+ messages in thread
From: Hannes Reinecke @ 2012-03-30  8:23 UTC (permalink / raw)
  To: linux-hotplug

On 03/30/2012 09:59 AM, Nils Carlson wrote:
> Hi,
> 
> We recently upgraded our version of udev and I saw that the symlinks
> under /dev/disk/by-path/ have changed since udev started using the
> path_id C code to no longer include phy enumeration for SAS disks.
> Was this change intentional? Would anyone mind if i re-create the
> behavior of the old bash script as we were using it?
> 
Not everybody at once ... I've just had several customer calls for
exactly the same issue.

Actually ... the company name seems to be vaguely familiar ...

To quote myself:

> Problem is that SAS has something called 'wide-ports', ie inherent
> port-bonding for SAS. When a system has wide ports, several PHYs
> are merged _physically_ together into a single connector.
> However, on the system side we're only seeing PHYs, so it'
> impossible to tell which of the phy has been used.
>
> So to handle this correctly we would need to generate symlinks for
> each phy in a wide port; however, currently the path_id program is
> only capable of printing out _one_ link per wide port.
>
> And actually, the original path_id script also printed out only
> one symlink, so even there it didn't handle wide ports correctly.
>

I've discussed this issue with Kay Sievers, and we agreed that we
should be fixing this.
It should be relatively easy to update the path_id program to print
out a link for each phy in a wide-port, and then we could do a
static enumeration via udev rules like

$env{ID_SAS_PHY0}="?*", SYMLINK+="/dev/disk/by-path/$ID_SAS_PHY0"
$env{ID_SAS_PHY1}="?*", SYMLINK+="/dev/disk/by-path/$ID_SAS_PHY1"

seeing that there is currently a max of 8 phys per wide port, we
only would need to specify 8 lines here. Which should be too bad.

You're welcome to lend a hand ...

Cheers,

Hannes
-- 
Dr. Hannes Reinecke		      zSeries & Storage
hare@suse.de			      +49 911 74053 688
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2012-03-30  8:23 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2012-03-30  7:59 Missing SAS phy enumeration in newer versions of udev path_id Nils Carlson
2012-03-30  8:23 ` Hannes Reinecke

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).