Linux SCSI subsystem development
 help / color / mirror / Atom feed
From: Chandra Seetharaman <sekharan@us.ibm.com>
To: "Moger, Babu" <Babu.Moger@lsi.com>
Cc: device-mapper development <dm-devel@redhat.com>,
	"linux-scsi@vger.kernel.org" <linux-scsi@vger.kernel.org>
Subject: RE: [dm-devel] failover does not work with rdac device handler
Date: Thu, 23 Oct 2008 10:15:04 -0700	[thread overview]
Message-ID: <1224782104.14830.870.camel@chandra-ubuntu> (raw)
In-Reply-To: <E463DF2B2E584B4A82673F53D62C2EF45BCE45E0@cosmail01.lsi.com>

Hi Babu,

There is an another issue w.r.t non-listed devices (like you tested).
The issue is that when the device disappears and reappear, the dh_state
file is not created for them. I wanted to send the patch for that also
together with this one. That is the reason for delay.

chandra
On Thu, 2008-10-23 at 08:12 -0600, Moger, Babu wrote:
> Hello Chandra,
>    I have not seen this patch being submitted to mainstream kernel. When are you planning to submit this patch? Should I open a bugzilla for tracking purposes?
> 
> Thanks
> Babu Moger
> -----Original Message-----
> From: Chandra Seetharaman [mailto:sekharan@us.ibm.com]
> Sent: Tuesday, October 07, 2008 2:19 PM
> To: Moger, Babu
> Cc: device-mapper development; linux-scsi@vger.kernel.org
> Subject: RE: [dm-devel] failover does not work with rdac device handler
> 
> 
> Sorry, I sent an imcomplete patch. Here is the correct one.
> 
> BTW, The panic you saw is caused by the line (one of the lines you
> added):
> 
> dev = container_of(&scsi_dh_data, struct scsi_device, scsi_dh_data);
> 
> in scsi_dh_release(). We cannot use pointer in a structure to get the
> parent structure.
> 
> chandra
> 
> Keep a reference count of attaches, so that same number of detaches are allowed.
> -----------
> 
> Signed-off-by: Chandra Seetharaman <sekharan@us.ibm.com>
> ---
> Index: linux-2.6.27-rc8-git5/drivers/scsi/device_handler/scsi_dh.c
> ===================================================================
> --- linux-2.6.27-rc8-git5.orig/drivers/scsi/device_handler/scsi_dh.c
> +++ linux-2.6.27-rc8-git5/drivers/scsi/device_handler/scsi_dh.c
> @@ -153,12 +153,28 @@ static int scsi_dh_handler_attach(struct
>         if (sdev->scsi_dh_data) {
>                 if (sdev->scsi_dh_data->scsi_dh != scsi_dh)
>                         err = -EBUSY;
> -       } else if (scsi_dh->attach)
> +               else
> +                       kref_get(&sdev->scsi_dh_data->kref);
> +       } else if (scsi_dh->attach) {
>                 err = scsi_dh->attach(sdev);
> +               if (!err) {
> +                       kref_init(&sdev->scsi_dh_data->kref);
> +                       sdev->scsi_dh_data->sdev = sdev;
> +               }
> +       }
> 
>         return err;
>  }
> 
> +static void scsi_dh_release(struct kref *kref)
> +{
> +       struct scsi_dh_data *scsi_dh_data;
> +       scsi_dh_data = container_of(kref, struct scsi_dh_data, kref);
> +
> +       if (scsi_dh_data->scsi_dh && scsi_dh_data->scsi_dh->detach)
> +               scsi_dh_data->scsi_dh->detach(scsi_dh_data->sdev);
> +}
> +
>  /*
>   * scsi_dh_handler_detach - Detach a device handler from a device
>   * @sdev - SCSI device the device handler should be detached from
> @@ -176,11 +192,7 @@ static void scsi_dh_handler_detach(struc
>         if (scsi_dh && scsi_dh != sdev->scsi_dh_data->scsi_dh)
>                 return;
> 
> -       if (!scsi_dh)
> -               scsi_dh = sdev->scsi_dh_data->scsi_dh;
> -
> -       if (scsi_dh && scsi_dh->detach)
> -               scsi_dh->detach(sdev);
> +       kref_put(&sdev->scsi_dh_data->kref, scsi_dh_release);
>  }
> 
>  /*
> Index: linux-2.6.27-rc8-git5/include/scsi/scsi_device.h
> ===================================================================
> --- linux-2.6.27-rc8-git5.orig/include/scsi/scsi_device.h
> +++ linux-2.6.27-rc8-git5/include/scsi/scsi_device.h
> @@ -6,6 +6,7 @@
>  #include <linux/spinlock.h>
>  #include <linux/workqueue.h>
>  #include <linux/blkdev.h>
> +#include <linux/kref.h>
>  #include <scsi/scsi.h>
>  #include <asm/atomic.h>
> 
> @@ -191,6 +192,8 @@ struct scsi_device_handler {
> 
>  struct scsi_dh_data {
>         struct scsi_device_handler *scsi_dh;
> +       struct kref kref;
> +       struct scsi_device *sdev; /* back reference */
>         char buf[0];
>  };
> 
> 
> 


  reply	other threads:[~2008-10-23 17:16 UTC|newest]

Thread overview: 18+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-09-30 17:01 Patch: add Dell product information into rdac device handler Yanqing_Liu
2008-09-30 17:57 ` Chandra Seetharaman
2008-10-01 14:12 ` [dm-devel] " James Bottomley
2008-10-01 21:47 ` failover does not work with " Moger, Babu
2008-10-04  0:12   ` Moger, Babu
2008-10-04  0:50     ` [dm-devel] " Chandra Seetharaman
2008-10-04  3:35       ` Moger, Babu
2008-10-06 19:26         ` Moger, Babu
2008-10-06 19:53           ` [dm-devel] " Chandra Seetharaman
2008-10-06 19:59             ` Moger, Babu
2008-10-06 22:26               ` Moger, Babu
2008-10-07  2:04                 ` [dm-devel] " Chandra Seetharaman
2008-10-07 18:34                   ` Moger, Babu
2008-10-07 19:19                     ` [dm-devel] " Chandra Seetharaman
2008-10-07 20:26                       ` Moger, Babu
2008-10-23 14:12                       ` [dm-devel] " Moger, Babu
2008-10-23 17:15                         ` Chandra Seetharaman [this message]
2008-10-23 17:31                           ` Moger, Babu

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=1224782104.14830.870.camel@chandra-ubuntu \
    --to=sekharan@us.ibm.com \
    --cc=Babu.Moger@lsi.com \
    --cc=dm-devel@redhat.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox