[PATCH] ide/ide-cs: fix order of releasing resources

[PATCH] ide/ide-cs: fix order of releasing resources

[Wolfram Sang]

ide_detach() called first ide_release() and then release_region(). This
produced the following warnings:

  Trying to free nonexistent resource <000000000000c10e-000000000000c10e>
  Trying to free nonexistent resource <000000000000c100-000000000000c107>

This is true, because the callchain inside ide_release() is:

  ide_release -> pcmcia_disable_device -> pcmcia_release_io

So, the whole io-block is already gone for release_region(). To fix
this, just swap the order of releasing (and remove the now obsolete
shadowing).

Signed-off-by: Wolfram Sang <w.sang@pengutronix.de>
Cc: Dominik Brodowski <linux@dominikbrodowski.net>
Cc: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
Cc: "David S. Miller" <davem@davemloft.net>
---
 drivers/ide/ide-cs.c |    8 ++------
 1 files changed, 2 insertions(+), 6 deletions(-)

diff --git a/drivers/ide/ide-cs.c b/drivers/ide/ide-cs.c
index dd63963..43ed330 100644
--- a/drivers/ide/ide-cs.c
+++ b/drivers/ide/ide-cs.c
@@ -122,18 +122,14 @@ static void ide_detach(struct pcmcia_device *link)
 {
     ide_info_t *info = link->priv;
     ide_hwif_t *hwif = info->host->ports[0];
-    unsigned long data_addr, ctl_addr;
 
     dev_dbg(&link->dev, "ide_detach(0x%p)\n", link);
 
-    data_addr = hwif->io_ports.data_addr;
-    ctl_addr  = hwif->io_ports.ctl_addr;
+    release_region(hwif->io_ports.ctl_addr, 1);
+    release_region(hwif->io_ports.data_addr, 8);
 
     ide_release(link);
 
-    release_region(ctl_addr, 1);
-    release_region(data_addr, 8);
-
     kfree(info);
 } /* ide_detach */
 
-- 
1.6.3.3

Re: [PATCH] ide/ide-cs: fix order of releasing resources

[David Miller]

From: Wolfram Sang <w.sang@pengutronix.de>
Date: Tue, 12 Jan 2010 20:34:54 +0100

> ide_detach() called first ide_release() and then release_region(). This
> produced the following warnings:
> 
>   Trying to free nonexistent resource <000000000000c10e-000000000000c10e>
>   Trying to free nonexistent resource <000000000000c100-000000000000c107>
> 
> This is true, because the callchain inside ide_release() is:
> 
>   ide_release -> pcmcia_disable_device -> pcmcia_release_io
> 
> So, the whole io-block is already gone for release_region(). To fix
> this, just swap the order of releasing (and remove the now obsolete
> shadowing).
> 
> Signed-off-by: Wolfram Sang <w.sang@pengutronix.de>

Applied, thanks a lot!

Re: [PATCH] ide/ide-cs: fix order of releasing resources

[Bartlomiej Zolnierkiewicz]

On Thursday 14 January 2010 10:59:38 am David Miller wrote:
> From: Wolfram Sang <w.sang@pengutronix.de>
> Date: Tue, 12 Jan 2010 20:34:54 +0100
> 
> > ide_detach() called first ide_release() and then release_region(). This
> > produced the following warnings:
> > 
> >   Trying to free nonexistent resource <000000000000c10e-000000000000c10e>
> >   Trying to free nonexistent resource <000000000000c100-000000000000c107>
> > 
> > This is true, because the callchain inside ide_release() is:
> > 
> >   ide_release -> pcmcia_disable_device -> pcmcia_release_io
> > 
> > So, the whole io-block is already gone for release_region(). To fix
> > this, just swap the order of releasing (and remove the now obsolete
> > shadowing).
> > 
> > Signed-off-by: Wolfram Sang <w.sang@pengutronix.de>
> 
> Applied, thanks a lot!

Port resources should not be freed before ide_host_remove() call in
ide_release() returns as port and/or devices may still be in use.

FWIW the following modified version of the patch is now in my atang tree:

From: Wolfram Sang <w.sang@pengutronix.de>
Subject: [PATCH] ide/ide-cs: fix order of releasing resources

ide_detach() called first ide_release() and then release_region(). This
produced the following warnings:

  Trying to free nonexistent resource <000000000000c10e-000000000000c10e>
  Trying to free nonexistent resource <000000000000c100-000000000000c107>

This is true, because the callchain inside ide_release() is:

  ide_release -> pcmcia_disable_device -> pcmcia_release_io

So, the whole io-block is already gone for release_region(). To fix
this, just swap the order of releasing (and remove the now obsolete
shadowing).

bzolnier:
- release resources in ide_release() to fix ordering of events
- remove stale FIXME note while at it

Signed-off-by: Wolfram Sang <w.sang@pengutronix.de>
Cc: Dominik Brodowski <linux@dominikbrodowski.net>
Cc: "David S. Miller" <davem@davemloft.net>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
---
 drivers/ide/ide-cs.c |   23 +++++++++++------------
 1 file changed, 11 insertions(+), 12 deletions(-)

Index: b/drivers/ide/ide-cs.c
===================================================================
--- a/drivers/ide/ide-cs.c
+++ b/drivers/ide/ide-cs.c
@@ -131,19 +131,11 @@ static int ide_probe(struct pcmcia_devic
 static void ide_detach(struct pcmcia_device *link)
 {
     ide_info_t *info = link->priv;
-    ide_hwif_t *hwif = info->host->ports[0];
-    unsigned long data_addr, ctl_addr;
 
     DEBUG(0, "ide_detach(0x%p)\n", link);
 
-    data_addr = hwif->io_ports.data_addr;
-    ctl_addr  = hwif->io_ports.ctl_addr;
-
     ide_release(link);
 
-    release_region(ctl_addr, 1);
-    release_region(data_addr, 8);
-
     kfree(info);
 } /* ide_detach */
 
@@ -365,12 +357,19 @@ static void ide_release(struct pcmcia_de
 
     DEBUG(0, "ide_release(0x%p)\n", link);
 
-    if (info->ndev)
-	/* FIXME: if this fails we need to queue the cleanup somehow
-	   -- need to investigate the required PCMCIA magic */
+    if (info->ndev) {
+	ide_hwif_t *hwif = host->ports[0];
+	unsigned long data_addr, ctl_addr;
+
+	data_addr = hwif->io_ports.data_addr;
+	ctl_addr = hwif->io_ports.ctl_addr;
+
 	ide_host_remove(host);
+	info->ndev = 0;
 
-    info->ndev = 0;
+	release_region(ctl_addr, 1);
+	release_region(data_addr, 8);
+    }
 
     pcmcia_disable_device(link);
 } /* ide_release */

Re: [PATCH] ide/ide-cs: fix order of releasing resources

[David Miller]

From: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
Date: Thu, 14 Jan 2010 16:50:17 +0100

> bzolnier:
> - release resources in ide_release() to fix ordering of events
> - remove stale FIXME note while at it

Ok this does look better, Wolfram can you test this?

Re: [PATCH] ide/ide-cs: fix order of releasing resources

[Wolfram Sang]

[-- Attachment #1: Type: text/plain, Size: 423 bytes --]

> > bzolnier:
> > - release resources in ide_release() to fix ordering of events
> > - remove stale FIXME note while at it
> 
> Ok this does look better, Wolfram can you test this?

ACK, looks better. I will test it, might not happen before Sunday, though.

-- 
Pengutronix e.K.                           | Wolfram Sang                |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 197 bytes --]

Re: [PATCH] ide/ide-cs: fix order of releasing resources

[Wolfram Sang]

[-- Attachment #1: Type: text/plain, Size: 367 bytes --]

> > bzolnier:
> > - release resources in ide_release() to fix ordering of events
> > - remove stale FIXME note while at it
> 
> Ok this does look better, Wolfram can you test this?

Works as expected.

-- 
Pengutronix e.K.                           | Wolfram Sang                |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 197 bytes --]

Re: [PATCH] ide/ide-cs: fix order of releasing resources

[David Miller]

From: Wolfram Sang <w.sang@pengutronix.de>
Date: Sat, 16 Jan 2010 19:51:24 +0100

>> > bzolnier:
>> > - release resources in ide_release() to fix ordering of events
>> > - remove stale FIXME note while at it
>> 
>> Ok this does look better, Wolfram can you test this?
> 
> Works as expected.

Great, thanks for testing.

Re: [PATCH] ide/ide-cs: fix order of releasing resources

[Dominik Brodowski]

On Tue, Jan 12, 2010 at 08:34:54PM +0100, Wolfram Sang wrote:
> ide_detach() called first ide_release() and then release_region(). This
> produced the following warnings:
> 
>   Trying to free nonexistent resource <000000000000c10e-000000000000c10e>
>   Trying to free nonexistent resource <000000000000c100-000000000000c107>
> 
> This is true, because the callchain inside ide_release() is:
> 
>   ide_release -> pcmcia_disable_device -> pcmcia_release_io
> 
> So, the whole io-block is already gone for release_region(). To fix
> this, just swap the order of releasing (and remove the now obsolete
> shadowing).
> 
> Signed-off-by: Wolfram Sang <w.sang@pengutronix.de>
> Cc: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
> Cc: "David S. Miller" <davem@davemloft.net>
Acked-by: Dominik Brodowski <linux@dominikbrodowski.net>

Best,
	Dominik