[PATCH] usb/ub: deprecate & schedule for removal the "Low Performance USB Block" driver

[PATCH] usb/ub: deprecate & schedule for removal the "Low Performance USB Block" driver

[Sebastian Andrzej Siewior]

Deprecate this driver. All devices which can be handled by this driver
can also be handled by the usb-storage driver.

Cc: Pete Zaitcev <zaitcev@redhat.com>
Cc: Jens Axboe <jaxboe@fusionio.com>
Signed-off-by: Sebastian Andrzej Siewior <sebastian@breakpoint.cc>
---
 Documentation/feature-removal-schedule.txt |   15 +++++++++++++++
 drivers/block/Kconfig                      |    2 +-
 drivers/block/ub.c                         |    2 ++
 3 files changed, 18 insertions(+), 1 deletions(-)

diff --git a/Documentation/feature-removal-schedule.txt b/Documentation/feature-removal-schedule.txt
index a0ffac0..f534828 100644
--- a/Documentation/feature-removal-schedule.txt
+++ b/Documentation/feature-removal-schedule.txt
@@ -524,3 +524,18 @@ Files:	arch/arm/mach-at91/at91cap9.c
 Why:	The code is not actively maintained and platforms are now hard to find.
 Who:	Nicolas Ferre <nicolas.ferre@atmel.com>
 	Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
+
+----------------------------
+
+What:	Low Performance USB Block driver ("CONFIG_BLK_DEV_UB")
+When:	3.6
+Why:	This driver provides support for USB storage devices like "USB
+	sticks". It translates the requests from the block layer into
+	READ & WRITE requests. It does not use the SCSI layer for device
+	handling and as results sdparm & smartmontool do not work. It
+	does not use libusual which holds various workarounds for
+	certain buggy devices.
+	As of now, it is deactivated in Debian, Fedora and Ubuntu. All
+	current users can switch over to usb-storage (CONFIG_USB_STORAGE)
+	which only drawback is the additional SCSI stack.
+Who:	Sebastian Andrzej Siewior <sebastian@breakpoint.cc>
diff --git a/drivers/block/Kconfig b/drivers/block/Kconfig
index 4e4c8a4..a796407 100644
--- a/drivers/block/Kconfig
+++ b/drivers/block/Kconfig
@@ -354,7 +354,7 @@ config BLK_DEV_SX8
 	  Use devices /dev/sx8/$N and /dev/sx8/$Np$M.
 
 config BLK_DEV_UB
-	tristate "Low Performance USB Block driver"
+	tristate "Low Performance USB Block driver (deprecated)"
 	depends on USB
 	help
 	  This driver supports certain USB attached storage devices
diff --git a/drivers/block/ub.c b/drivers/block/ub.c
index 298ac75..fcec022 100644
--- a/drivers/block/ub.c
+++ b/drivers/block/ub.c
@@ -2440,6 +2440,8 @@ static int __init ub_init(void)
 	int rc;
 	int i;
 
+	pr_info("'Low Performance USB Block' driver is deprecated. "
+			"Please switch to usb-storage\n");
 	for (i = 0; i < UB_QLOCK_NUM; i++)
 		spin_lock_init(&ub_qlockv[i]);
 
-- 
1.7.9.1

Re: [PATCH] usb/ub: deprecate & schedule for removal the "Low Performance USB Block" driver

[Greg Kroah-Hartman]

On Wed, Mar 14, 2012 at 09:52:51PM +0100, Sebastian Andrzej Siewior wrote:
> Deprecate this driver. All devices which can be handled by this driver
> can also be handled by the usb-storage driver.
> 
> Cc: Pete Zaitcev <zaitcev@redhat.com>

I need an ack here from Pete before I can accept this.

Re: [PATCH] usb/ub: deprecate & schedule for removal the "Low Performance USB Block" driver

[Pete Zaitcev]

On Wed, 14 Mar 2012 21:52:51 +0100
Sebastian Andrzej Siewior <sebastian@breakpoint.cc> wrote:

> +What:	Low Performance USB Block driver ("CONFIG_BLK_DEV_UB")
> +When:	3.6

ACK

> +	does not use libusual which holds various workarounds for
> +	certain buggy devices.

Although the above is false, and ub actually started libusual,
I don't mind. It's a transient notice anyway. Once ub is gone,
libusual can be folded back into usb-storage.

-- Pete

Re: [PATCH] usb/ub: deprecate & schedule for removal the "Low Performance USB Block" driver

[Greg Kroah-Hartman]

On Thu, Mar 15, 2012 at 09:33:31AM -0600, Pete Zaitcev wrote:
> On Wed, 14 Mar 2012 21:52:51 +0100
> Sebastian Andrzej Siewior <sebastian@breakpoint.cc> wrote:
> 
> > +What:	Low Performance USB Block driver ("CONFIG_BLK_DEV_UB")
> > +When:	3.6
> 
> ACK
> 
> > +	does not use libusual which holds various workarounds for
> > +	certain buggy devices.
> 
> Although the above is false, and ub actually started libusual,
> I don't mind. It's a transient notice anyway. Once ub is gone,
> libusual can be folded back into usb-storage.

Good point.

Sebastian, care to reword this and resend it?

Also, are you going to be doing the "merge libusual into usb-storage"
after you remove the ub driver?

thanks,

greg k-h

Re: [PATCH] usb/ub: deprecate & schedule for removal the "Low Performance USB Block" driver

[Sebastian Andrzej Siewior]

> > Although the above is false, and ub actually started libusual,
> > I don't mind. It's a transient notice anyway. Once ub is gone,
> > libusual can be folded back into usb-storage.
> 
> Good point.
> 
> Sebastian, care to reword this and resend it?
I want clarify one thing but sure.

> Also, are you going to be doing the "merge libusual into usb-storage"
> after you remove the ub driver?

This sounds like once ub is out, you have to merge those two :) So yes, unless
a use case for it comes up.

> thanks,
> 
> greg k-h

Sebastian

Re: [PATCH] usb/ub: deprecate & schedule for removal the "Low Performance USB Block" driver

[Greg Kroah-Hartman]

On Thu, Mar 15, 2012 at 07:10:45PM +0100, Sebastian Andrzej Siewior wrote:
> > > Although the above is false, and ub actually started libusual,
> > > I don't mind. It's a transient notice anyway. Once ub is gone,
> > > libusual can be folded back into usb-storage.
> > 
> > Good point.
> > 
> > Sebastian, care to reword this and resend it?
> I want clarify one thing but sure.
> 
> > Also, are you going to be doing the "merge libusual into usb-storage"
> > after you remove the ub driver?
> 
> This sounds like once ub is out, you have to merge those two :) So yes, unless
> a use case for it comes up.

I'll take patches from you to do this, as the removal patch of ub will
make libusual not need to be stand-alone.

And you will be the one sending the ub removal patch in 6 months, right?

thanks,

greg k-h

Re: [PATCH] usb/ub: deprecate & schedule for removal the "Low Performance USB Block" driver

[Sebastian Andrzej Siewior]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

On > I'll take patches from you to do this, as the removal patch of ub will
> make libusual not need to be stand-alone.
Okay.

> And you will be the one sending the ub removal patch in 6 months, right?

Linus should release 3.3 this weekend. The removal is scheduled for 3.6 which
means 3 releases so something between 3 * 2 .. 3 * 3 months depending how
long each release will take.

> 
> thanks,
> 
> greg k-h

Sebastian
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.12 (GNU/Linux)
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=uzZf
-----END PGP SIGNATURE-----

[PATCH v2] usb/ub: deprecate & schedule for removal the "Low Performance USB Block" driver

[Sebastian Andrzej Siewior]

Deprecate this driver. All devices which can be handled by this driver
can also be handled by the usb-storage driver.

Cc: Pete Zaitcev <zaitcev@redhat.com>
Cc: Jens Axboe <jaxboe@fusionio.com>
Signed-off-by: Sebastian Andrzej Siewior <sebastian@breakpoint.cc>
---
v1..v2: remove "can't use SCSI commands" and "does not use libusual". I meant
libunusual but it probably does not matter.

 Documentation/feature-removal-schedule.txt |   11 +++++++++++
 drivers/block/Kconfig                      |    2 +-
 drivers/block/ub.c                         |    2 ++
 3 files changed, 14 insertions(+), 1 deletions(-)

diff --git a/Documentation/feature-removal-schedule.txt b/Documentation/feature-removal-schedule.txt
index a0ffac0..d5dc80f 100644
--- a/Documentation/feature-removal-schedule.txt
+++ b/Documentation/feature-removal-schedule.txt
@@ -524,3 +524,14 @@ Files:	arch/arm/mach-at91/at91cap9.c
 Why:	The code is not actively maintained and platforms are now hard to find.
 Who:	Nicolas Ferre <nicolas.ferre@atmel.com>
 	Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
+
+----------------------------
+
+What:	Low Performance USB Block driver ("CONFIG_BLK_DEV_UB")
+When:	3.6
+Why:	This driver provides support for USB storage devices like "USB
+	sticks". As of now, it is deactivated in Debian, Fedora and
+        Ubuntu. All current users can switch over to usb-storage
+        (CONFIG_USB_STORAGE) which only drawback is the additional SCSI
+        stack.
+Who:	Sebastian Andrzej Siewior <sebastian@breakpoint.cc>
diff --git a/drivers/block/Kconfig b/drivers/block/Kconfig
index 4e4c8a4..a796407 100644
--- a/drivers/block/Kconfig
+++ b/drivers/block/Kconfig
@@ -354,7 +354,7 @@ config BLK_DEV_SX8
 	  Use devices /dev/sx8/$N and /dev/sx8/$Np$M.
 
 config BLK_DEV_UB
-	tristate "Low Performance USB Block driver"
+	tristate "Low Performance USB Block driver (deprecated)"
 	depends on USB
 	help
 	  This driver supports certain USB attached storage devices
diff --git a/drivers/block/ub.c b/drivers/block/ub.c
index 298ac75..fcec022 100644
--- a/drivers/block/ub.c
+++ b/drivers/block/ub.c
@@ -2440,6 +2440,8 @@ static int __init ub_init(void)
 	int rc;
 	int i;
 
+	pr_info("'Low Performance USB Block' driver is deprecated. "
+			"Please switch to usb-storage\n");
 	for (i = 0; i < UB_QLOCK_NUM; i++)
 		spin_lock_init(&ub_qlockv[i]);
 
-- 
1.7.9.1

Re: [PATCH v2] usb/ub: deprecate & schedule for removal the "Low Performance USB Block" driver

[Pete Zaitcev]

On Thu, 15 Mar 2012 21:59:36 +0100
Sebastian Andrzej Siewior <sebastian@breakpoint.cc> wrote:

> +What:	Low Performance USB Block driver ("CONFIG_BLK_DEV_UB")
> +When:	3.6
> +Why:	This driver provides support for USB storage devices like "USB
> +	sticks". As of now, it is deactivated in Debian, Fedora and
> +        Ubuntu. All current users can switch over to usb-storage
> +        (CONFIG_USB_STORAGE) which only drawback is the additional SCSI
> +        stack.
> +Who:	Sebastian Andrzej Siewior <sebastian@breakpoint.cc>

Acked-By: Pete Zaitcev <zaitcev@redhat.com>

-- Pete

Re: [PATCH] usb/ub: deprecate & schedule for removal the "Low Performance USB Block" driver

[Sebastian Andrzej Siewior]

On Thu, Mar 15, 2012 at 09:33:31AM -0600, Pete Zaitcev wrote:
> On Wed, 14 Mar 2012 21:52:51 +0100
> Sebastian Andrzej Siewior <sebastian@breakpoint.cc> wrote:
> 
> > +What:	Low Performance USB Block driver ("CONFIG_BLK_DEV_UB")
> > +When:	3.6
> 
> ACK
> 
> > +	does not use libusual which holds various workarounds for
> > +	certain buggy devices.
> 
> Although the above is false, and ub actually started libusual,
> I don't mind. It's a transient notice anyway. Once ub is gone,
> libusual can be folded back into usb-storage.

There might be an miss understanding here. For instance in
drivers/usb/storage/unusual_devs.h we have an entry like:

|UNUSUAL_DEV(  0x03f0, 0x4002, 0x0001, 0x0001,
|                 "HP",
|                 "PhotoSmart R707",
|                 USB_SC_DEVICE, USB_PR_DEVICE, NULL, US_FL_FIX_CAPACITY),
|

That US_FL_FIX_CAPACITY is used later in scsiglue.c to set sdev->fix_capacity
which is then used by sd_read_capacity() to substract one sector from the
final size.
In ub I see ub_sync_read_cap() and I don't see a change. Was my wording wrong
or is this something else?

> -- Pete

Sebastian

Re: [PATCH] usb/ub: deprecate & schedule for removal the "Low Performance USB Block" driver

[Alan Stern]

On Thu, 15 Mar 2012, Sebastian Andrzej Siewior wrote:

> On Thu, Mar 15, 2012 at 09:33:31AM -0600, Pete Zaitcev wrote:
> > On Wed, 14 Mar 2012 21:52:51 +0100
> > Sebastian Andrzej Siewior <sebastian@breakpoint.cc> wrote:
> > 
> > > +What:	Low Performance USB Block driver ("CONFIG_BLK_DEV_UB")
> > > +When:	3.6
> > 
> > ACK
> > 
> > > +	does not use libusual which holds various workarounds for
> > > +	certain buggy devices.
> > 
> > Although the above is false, and ub actually started libusual,
> > I don't mind. It's a transient notice anyway. Once ub is gone,
> > libusual can be folded back into usb-storage.
> 
> There might be an miss understanding here. For instance in
> drivers/usb/storage/unusual_devs.h we have an entry like:
> 
> |UNUSUAL_DEV(  0x03f0, 0x4002, 0x0001, 0x0001,
> |                 "HP",
> |                 "PhotoSmart R707",
> |                 USB_SC_DEVICE, USB_PR_DEVICE, NULL, US_FL_FIX_CAPACITY),
> |
> 
> That US_FL_FIX_CAPACITY is used later in scsiglue.c to set sdev->fix_capacity
> which is then used by sd_read_capacity() to substract one sector from the
> final size.
> In ub I see ub_sync_read_cap() and I don't see a change. Was my wording wrong
> or is this something else?

Pete wasn't talking about the phrase "holds various workarounds...".  
Rather, he was referring to the phrase "does not use libusual".  That
part is wrong; ub _does_ use libusual.

Alan Stern

Re: [PATCH] usb/ub: deprecate & schedule for removal the "Low Performance USB Block" driver

[Sebastian Andrzej Siewior]

On Thu, Mar 15, 2012 at 04:03:02PM -0400, Alan Stern wrote:
> Pete wasn't talking about the phrase "holds various workarounds...".  
> Rather, he was referring to the phrase "does not use libusual".  That
> part is wrong; ub _does_ use libusual.

Okay, so I mixed up those two. "does not use libunsual" would be the correct
thing to day, right?

> Alan Stern

Sebastian

Re: [PATCH] usb/ub: deprecate & schedule for removal the "Low Performance USB Block" driver

[Alan Stern]

On Thu, 15 Mar 2012, Sebastian Andrzej Siewior wrote:

> On Thu, Mar 15, 2012 at 04:03:02PM -0400, Alan Stern wrote:
> > Pete wasn't talking about the phrase "holds various workarounds...".  
> > Rather, he was referring to the phrase "does not use libusual".  That
> > part is wrong; ub _does_ use libusual.
> 
> Okay, so I mixed up those two. "does not use libunsual" would be the correct
> thing to day, right?

Mixed up what two?  What is libunusual?  I have never heard of it.

Alan Stern