All of lore.kernel.org
 help / color / mirror / Atom feed
From: Mike Christie <michaelc@cs.wisc.edu>
To: James.Smart@Emulex.Com
Cc: linux-scsi@vger.kernel.org
Subject: Re: [PATCH] [REPOST] SCSI and FC Transport: add netlink support for posting of transport events
Date: Sun, 20 Aug 2006 20:03:02 -0400	[thread overview]
Message-ID: <44E8F836.6010108@cs.wisc.edu> (raw)
In-Reply-To: <44E8605B.1070102@emulex.com>

James Smart wrote:
>> This is because for
>> software iscsi and iser we do not know the pci device that will be used.
>> For software iscsi it could be a bonded device or multiple devices
>> depending on tables getting updated. What would we use for the vendor id
>> in this case?
> 
> Well - in this case, it's not a vendor-specific event being generated.
> It's an event by the iscsi/iser layer. This should fit the "well-known
> to the iscsi/iser transport" event definition and not require a vendor id.
> 
> If we expected to have multiple iscsi/iser stacks present (vs the 1 and
> only
> 1 implementation we have for the other transports), then either that gets
> built into the event message (as a stack identifier) or you could munge the
> same thing into the vendor id (e.g. define a "type" that indicates s/w
> assigned - and the vendor id is the stack identifier (but this method seems
> hoaky)).
> 

Oh yeah, I guess where we missed each other on this is that for iscsi:
software iscsi, software iser, iscsi cards like qlogic which hide most
iscsi details and iscsi cards like broadcom which partially hook into
the software stack all use the same transport class interface, but
sometimes events need to be handled differently in userspace depending
on what card is being used if any. So if you count the iscsi stack
running or partially running in FW/HW in something like a broadcom or
qlogic card and the software stack (iscsi_tcp or iscsi_iser) then we can
and do have multiple iscsi stacks loaded at one time, but have one
transport class that handles common tasks for all stacks.

As a result of doing software iscsi first then software iser, we do not
think in terms of vendors since they do not have one and instead do it
per driver. And we may need to perform a driver specific action in
userspace to handle a common transport class event. We thought this
might be the case for all transports and so we were hoping that common
id could be used, but if not I do not care anymore.

  parent reply	other threads:[~2006-08-21  1:03 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2006-08-18 21:30 [PATCH] [REPOST] SCSI and FC Transport: add netlink support for posting of transport events James Smart
2006-08-19  4:11 ` Mike Christie
2006-08-19 11:56   ` James Smart
2006-08-20  2:26     ` Mike Christie
2006-08-20 13:15       ` James Smart
2006-08-20 13:30         ` James Smart
2006-08-20 22:35         ` Mike Christie
2006-08-20 22:46           ` Mike Christie
2006-08-20 23:01           ` Mike Christie
2006-08-21  0:03         ` Mike Christie [this message]
2006-08-30 17:40 ` James Bottomley
2006-08-30 17:51   ` James Smart

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=44E8F836.6010108@cs.wisc.edu \
    --to=michaelc@cs.wisc.edu \
    --cc=James.Smart@Emulex.Com \
    --cc=linux-scsi@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.