Skip to contentSkip to conversation

undeprecate raw driver.

PATCH Dave Jones 10 participants 18 messages May 13, 2007 – May 15, 2007

Subsystems (2)
lore
Share & tools
lore

Dave Jones

author
#0 lore raw

Despite repeated attempts over the last two and half years, this driver seems somewhat persistant. Remove its deprecated status as it has existing users who may not be in a position to migrate their apps to O_DIRECT.

Documentation/feature-removal-schedule.txt +0 -8
index 498ff31..5108619 100644
@@ -41,14 +41,6 @@ Who:	Pavel Machek <[email protected]>
 
 ---------------------------
 
-What:	RAW driver (CONFIG_RAW_DRIVER)
-When:	December 2005
-Why:	declared obsolete since kernel 2.6.3
-	O_DIRECT can be used instead
-Who:	Adrian Bunk <[email protected]>
-
----------------------------
-
 What:	raw1394: requests of type RAW1394_REQ_ISO_SEND, RAW1394_REQ_ISO_LISTEN
 When:	June 2007
 Why:	Deprecated in favour of the more efficient and robust rawiso interface.
drivers/char/Kconfig +4 -5
index abcafac..1e4d688 100644
@@ -993,15 +993,14 @@ config GPIO_VR41XX
 	depends on CPU_VR41XX
 
 config RAW_DRIVER
-	tristate "RAW driver (/dev/raw/rawN) (OBSOLETE)"
+	tristate "RAW driver (/dev/raw/rawN)"
 	depends on BLOCK
 	help
-	  The raw driver permits block devices to be bound to /dev/raw/rawN. 
-	  Once bound, I/O against /dev/raw/rawN uses efficient zero-copy I/O. 
+	  The raw driver permits block devices to be bound to /dev/raw/rawN.
+	  Once bound, I/O against /dev/raw/rawN uses efficient zero-copy I/O.
 	  See the raw(8) manpage for more details.
 
-          The raw driver is deprecated and will be removed soon.
-          Applications should simply open the device (eg /dev/hda1)
+          Applications should preferably open the device (eg /dev/hda1)
           with the O_DIRECT flag.
 
 config MAX_RAW_DEVS
Signature · 1 line
http://www.codemonkey.org.uk
#1 lore raw
Quoted 49 lines · Dave Jones

Despite repeated attempts over the last two and half years, this driver seems somewhat persistant. Remove its deprecated status as it has existing users who may not be in a position to migrate their apps to O_DIRECT.

Documentation/feature-removal-schedule.txt +0 -9
index 498ff31..5108619 100644
@@ -41,14 +41,6 @@ Who:	Pavel Machek <[email protected]>
 
---------------------------
 
-What:	RAW driver (CONFIG_RAW_DRIVER)
-When:	December 2005
-Why:	declared obsolete since kernel 2.6.3
-	O_DIRECT can be used instead
-Who:	Adrian Bunk <[email protected]>
-
----------------------------
-

What: raw1394: requests of type RAW1394_REQ_ISO_SEND, RAW1394_REQ_ISO_LISTEN
When: June 2007
Why: Deprecated in favour of the more efficient and robust rawiso interface.

drivers/char/Kconfig +0 -0
index abcafac..1e4d688 100644
@@ -993,15 +993,14 @@ config GPIO_VR41XX
 	depends on CPU_VR41XX
 

config RAW_DRIVER

drivers/char/Kconfig +1 -1
-	tristate "RAW driver (/dev/raw/rawN) (OBSOLETE)"
+	tristate "RAW driver (/dev/raw/rawN)"

depends on BLOCK
help

drivers/char/Kconfig +2 -2
-	  The raw driver permits block devices to be bound to /dev/raw/rawN. 
-	  Once bound, I/O against /dev/raw/rawN uses efficient zero-copy I/O. 
+	  The raw driver permits block devices to be bound to /dev/raw/rawN.
+	  Once bound, I/O against /dev/raw/rawN uses efficient zero-copy I/O.

See the raw(8) manpage for more details.

drivers/char/Kconfig +1 -2
-          The raw driver is deprecated and will be removed soon.
-          Applications should simply open the device (eg /dev/hda1)
+          Applications should preferably open the device (eg /dev/hda1)
          with the O_DIRECT flag.
Quoted 1 line

config MAX_RAW_DEVS

At least keep the "it's obsolete" Kconfig description. We don't want new users/projects to jump on /dev/raw.

Jan

author
#2 lore raw
Quoted 15 lines · Jan Engelhardt
Quoted 15 lines
drivers/char/Kconfig +1 -1
-	tristate "RAW driver (/dev/raw/rawN) (OBSOLETE)"
+	tristate "RAW driver (/dev/raw/rawN)"

depends on BLOCK
help

drivers/char/Kconfig +2 -2
-	  The raw driver permits block devices to be bound to /dev/raw/rawN. 
-	  Once bound, I/O against /dev/raw/rawN uses efficient zero-copy I/O. 
+	  The raw driver permits block devices to be bound to /dev/raw/rawN.
+	  Once bound, I/O against /dev/raw/rawN uses efficient zero-copy I/O.

See the raw(8) manpage for more details.

drivers/char/Kconfig +1 -2
-          The raw driver is deprecated and will be removed soon.
-          Applications should simply open the device (eg /dev/hda1)
+          Applications should preferably open the device (eg /dev/hda1)
          with the O_DIRECT flag.
Quoted 4 lines
Quoted 1 line

config MAX_RAW_DEVS

At least keep the "it's obsolete" Kconfig description. We don't want new users/projects to jump on /dev/raw.

That'll just lead to some janitor sending more "OMG! OBSOLETE!" cleanup patches getting it re-added to the feature-removal.txt
If we undeprecate it, it won't be going away, so new users of it aren't any more of a problem than they have been the last two and a half years.

Dave

Signature · 1 line
http://www.codemonkey.org.uk
#3 lore raw
In reply to Jan Engelhardt
Quoted 10 lines · Jan Engelhardt
Quoted 4 lines · Dave Jones

Despite repeated attempts over the last two and half years, this driver seems somewhat persistant. Remove its deprecated status as it has existing users who may not be in a position to migrate their apps

At least keep the "it's obsolete" Kconfig description. We don't want new users/projects to jump on /dev/raw.

i just *know* this is a mistake, but i'm going to take one more shot at distinguishing between deprecated and obsolete.

as i understand it, the raw driver is *deprecated*. that is, it's still there, it's still supported, people are still using it but its use is *seriously* discouraged and everyone should be trying to move off of it at their earliest possible convenience.

that is *not* the same as "obsolete" which should mean that that feature is dead, dead, DEAD and *no one* should be using it anymore.

yes, i realize it sounds like splitting hairs, but it's this malleable definition of "deprecated" that's causing all of this trouble in the first place -- the fact that the raw driver is currently listed as "obsolete" when it is, in fact, only deprecated.

in short, do *not* remove its "deprecated" status. rather, remove its "obsolete" status and *make* it deprecated.

rday

Signature · 7 lines
========================================================================
Robert P. J. Day
Linux Consulting, Training and Annoying Kernel Pedantry
Waterloo, Ontario, CANADA

http://fsdev.net/wiki/index.php?title=Main_Page
========================================================================
#4 lore raw
Quoted 2 lines · In article <[email protected]> you wrote:

in short, do *not* remove its "deprecated" status. rather, remove its "obsolete" status and *make* it deprecated.

it is deprecated and obsolete.

Gruss
Bernd

#5 lore raw
Quoted 5 lines · Bernd Eckenfels
Quoted 2 lines · In article <[email protected]> you wrote:

in short, do *not* remove its "deprecated" status. rather, remove its "obsolete" status and *make* it deprecated.

it is deprecated and obsolete.

no, it's not. and, trust me, no one on this list wants me to have to explain all that again. :-)

rday

p.s. before we get into this again where everyone thinks they know what they're talking about, i suggest consulting the official definitions of those two terms as defined at
http://www.w3.org/TR/html4/conform.html:

Deprecated:
----------

"A deprecated element or attribute is one that has been outdated by newer constructs. Deprecated elements are defined in the reference manual in appropriate locations, but are clearly marked as deprecated. Deprecated elements may become obsolete in future versions of HTML.

"User agents should continue to support deprecated elements for reasons of backward compatibility.

"Definitions of elements and attributes clearly indicate which are deprecated."

got that? deprecated means that it's been "outdated," but it's
still defined, it's still supported, it's marked as "deprecated", and it ***may become*** obsolete in future versions. if there's a better description of raw drivers, i don't know where you'd find it.

Obsolete:
--------

"An obsolete element or attribute is one for which there is no guarantee of support by a user agent."

there. see the difference? why is this so difficult to grok? you
can be deprecated, or you can be obsolete. you *can't* be both. so please stop making that argument.

p.p.s. using those definitions, deprecated features are *not* possible candidates for removal, but obsolete features *are*. that's why raw drivers are deprecated, not obsolete.

Signature · 7 lines
========================================================================
Robert P. J. Day
Linux Consulting, Training and Annoying Kernel Pedantry
Waterloo, Ontario, CANADA

http://fsdev.net/wiki/index.php?title=Main_Page
========================================================================
#6 lore raw
Quoted 7 lines · Robert P. J. Day

p.s. before we get into this again where everyone thinks they know what they're talking about, i suggest consulting the official definitions of those two terms as defined at
http://www.w3.org/TR/html4/conform.html:

Deprecated:
----------

[...]

Quoted 5 lines

Obsolete:
--------

"An obsolete element or attribute is one for which there is no guarantee of support by a user agent."

Please quote W3C's entire definition of their notion of obsolete:

"An obsolete element or attribute is one for which there is no guarantee of support by a user agent. Obsolete elements are no longer defined in the specification, but are listed for historical purposes in the changes section of the reference manual."

Quoted 1 line

there. see the difference? why is this so difficult to grok?

[...]

If you apply W3C's term "obsolete" 1:1 to kernel features, then it would read:

"An obsolete feature is one for which there is no guarantee of support by a randomly picked kernel release. Obsolete features are no longer implemented in this release, but are listed for historical purposes in Documentation/ABI/removed/."

Except that the term "obsolete" is already used differently in the context of Linux kernel features; see Documentation/ABI/README.

Also, you say "the official definitions of those terms" were defined at
http://www.w3.org/TR/html4/conform.html. That's not quite true. What you find there are the definitions of those terms as used in the HTML 4 specification. Nothing more.

Signature · 3 lines
Stefan Richter
-=====-=-=== -=-= -==-=
http://arcgraph.de/sr/
#7 lore raw
Quoted 2 lines · Stefan Richter

Except that the term "obsolete" is already used differently in the context of Linux kernel features; see Documentation/ABI/README.

no, it isn't. in fact, it's used in that file *exactly* the way i've been defining it:

"obsolete/"

"This directory documents interfaces that are still remaining in the kernel, but are marked to be removed at some later point in time. The description of the interface will document the reason why it is obsolete and when it can be expected to be removed. The file Documentation/feature-removal-schedule.txt may describe some of these interfaces, giving a schedule for when they will be removed."

note the salient points of what is considered "obsolete":

* yes, it's still in the kernel
* it may already be scheduled to be removed
* it's documented why that interface is ***obsolete*** and is scheduled for removal

look carefully, stefan -- note what's missing from that explanation:
the word "deprecated" which, IMHO, means something quite different.

at this point, i've explained several times already what i think is
the difference between these two terms, and why i think it's an *important* difference and how those terms should be used to guarantee clarity.

you clearly disagree and there's absolutely nothing i can say at
this point that's going to get through to you and i'm as thoroughly tired of trying to explain it to you as i'm sure everyone else is of reading it. so let's just accept that any further discussion on this topic is a waste of time. and we can leave things the way they are, and things will continue to be mis-labelled and people will, on a regular basis, continue to submit patches to remove things that claim to be "obsolete" but which aren't, and this whole freakin' debate will start all over again.

happy now?

rday

Signature · 7 lines
========================================================================
Robert P. J. Day
Linux Consulting, Training and Annoying Kernel Pedantry
Waterloo, Ontario, CANADA

http://fsdev.net/wiki/index.php?title=Main_Page
========================================================================
#8 lore raw
Quoted 7 lines · Robert P. J. Day
Quoted 2 lines · Stefan Richter

Except that the term "obsolete" is already used differently in the context of Linux kernel features; see Documentation/ABI/README.

no, it isn't. in fact, it's used in that file *exactly* the way i've been defining it:

I meant _differently from the term in the HTML spec_.

Signature · 3 lines
Stefan Richter
-=====-=-=== -=-= -==-=
http://arcgraph.de/sr/
#9 lore raw
In reply to Robert P. J. Day
Quoted 9 lines · Robert P. J. Day
Quoted 4 lines · Bernd Eckenfels
Quoted 2 lines · In article <[email protected]> you wrote:

in short, do *not* remove its "deprecated" status. rather, remove its "obsolete" status and *make* it deprecated.

it is deprecated and obsolete.

no, it's not. and, trust me, no one on this list wants me to have to explain all that again. :-)

So why don't we simply say:

- "Users of feature A are urged to migrate to alternative B."
- "Feature C will be removed in February 2038."

Everybody understands the meaning of /that/.

Signature · 3 lines
Stefan Richter
-=====-=-=== -=-= -==-=
http://arcgraph.de/sr/
#10 lore raw

Am 13.05.2007 22:10 schrieb Stefan Richter:

Quoted 4 lines

So why don't we simply say:

- "Users of feature A are urged to migrate to alternative B."
- "Feature C will be removed in February 2038."

Because that's too long to add to a Kconfig tag line. You need a terse label there.

Signature · 4 lines
Tilman Schmidt                          E-Mail: [email protected]
Bonn, Germany
Diese Nachricht besteht zu 100% aus wiederverwerteten Bits.
Ungeöffnet mindestens haltbar bis: (siehe Rückseite)
#11 lore raw
Quoted 8 lines · Tilman Schmidt

Am 13.05.2007 22:10 schrieb Stefan Richter:

Quoted 4 lines

So why don't we simply say:

- "Users of feature A are urged to migrate to alternative B."
- "Feature C will be removed in February 2038."

Because that's too long to add to a Kconfig tag line. You need a terse label there.

Then put "DEPRECATED" there, because you deprecate the enabling of the option.

However, actual users of such a kernel option need the full explanation of what's wrong with the option in order to decide whether to switch it
off. The terse label alone is *insufficient*. It's insufficient even
if you develop a scheme with two or more different labels for different flavors of deprecation. In fact, generically defined finer-grained flavors of "deprecated" will be too fine-grained to be sensibly used in /real/ feature removal processes.

It doesn't lead us anywhere to discuss about deprecation on an abstract level. Discuss concrete features (like the raw driver) with the people who used that feature --- only then you get to know how to proceed with that feature appropriately.

Signature · 3 lines
Stefan Richter
-=====-=-=== -=-= -===-
http://arcgraph.de/sr/
#12 lore raw
In reply to Stefan Richter
Quoted 6 lines · Stefan Richter <stefanr <at> s5r6.in-berlin.de> writes:

So why don't we simply say:

- "Users of feature A are urged to migrate to alternative B."
- "Feature C will be removed in February 2038."

Everybody understands the meaning of /that/.

Why not just use the terms:
* outdated - as a replacement for "deprecated".
* toberemoved - as a replacement for "obsolete".

These terms are easy to understand, short, universally accepted and extremely difficult to get wrong.

#13 lore raw
Quoted 10 lines · On Sun, 13 May 2007 23:01:51 +0000 (UTC) Bob Johnston <[email protected]> wrote:
Quoted 6 lines · Stefan Richter <stefanr <at> s5r6.in-berlin.de> writes:

So why don't we simply say:

- "Users of feature A are urged to migrate to alternative B."
- "Feature C will be removed in February 2038."

Everybody understands the meaning of /that/.

Why not just use the terms:
* outdated - as a replacement for "deprecated".

Because they don't actually mean the same thing ?

Quoted 1 line

* toberemoved - as a replacement for "obsolete".

With spaces, I'd hope.

#14 lore raw
Quoted 4 lines · Alan Cox <alan <at> lxorguk.ukuu.org.uk> writes:
Quoted 2 lines

Why not just use the terms:
* outdated - as a replacement for "deprecated".

Because they don't actually mean the same thing ?

"superseded" would probably be a better word, perhaps lacking the negative connotations of "deprecated"

Quoted 3 lines
Quoted 1 line

* toberemoved - as a replacement for "obsolete".

With spaces, I'd hope.

Sure.

I'm just trying to propose an alternative because it seems that deprecated/obsolete causes some confusion.

#15 lore raw
Quoted 9 lines · Bob Johnston
Quoted 4 lines · Alan Cox <alan <at> lxorguk.ukuu.org.uk> writes:
Quoted 2 lines

Why not just use the terms:
* outdated - as a replacement for "deprecated".

Because they don't actually mean the same thing ?

"superseded" would probably be a better word, perhaps lacking the negative connotations of "deprecated"

except that you *want* the negative connotation associated with the word "deprecated." you don't just want to say something's been superseded. rather, you want to say that it's not only been superseded but that you really want people to *stop using it* and move up to the newer version; otherwise, there is no motivation to upgrade.

i'm just baffled by the resistance to the word "deprecated" since it represents *exactly* the idea you want to get across here. its use in software projects is well-established. why are some people so put off by it?

Quoted 2 lines

I'm just trying to propose an alternative because it seems that deprecated/obsolete causes some confusion.

no, it does *not* cause confusion. by now, everyone who's been following this thread knows *exactly* what those two words are being proposed to represent. what's happening is that they don't *like* that proposed usage.

quite simply, there is no confusion here, there is simply annoying obstinacy. there's a difference.

rday

Signature · 7 lines
========================================================================
Robert P. J. Day
Linux Consulting, Training and Annoying Kernel Pedantry
Waterloo, Ontario, CANADA

http://fsdev.net/wiki/index.php?title=Main_Page
========================================================================
#16 lore raw
Quoted 19 lines · Bob Johnston
Quoted 10 lines
Quoted 4 lines · Alan Cox <alan <at> lxorguk.ukuu.org.uk> writes:
Quoted 2 lines

Why not just use the terms:
* outdated - as a replacement for "deprecated".

Because they don't actually mean the same thing ?

"superseded" would probably be a better word, perhaps lacking the negative connotations of "deprecated"

except that you *want* the negative connotation associated with the word "deprecated." you don't just want to say something's been superseded. rather, you want to say that it's not only been superseded but that you really want people to *stop using it* and move up to the newer version; otherwise, there is no motivation to upgrade.

Quoted 4 lines

i'm just baffled by the resistance to the word "deprecated" since it represents *exactly* the idea you want to get across here. its use in software projects is well-established. why are some people so put off by it?

Let me give this a big AOL-style "me too". Terms such as "deprecated" and "obsolete" have been used by programmers for as long as I can remember, and if you don't know exactly what they mean, you probably should just learn.

A feature is "deprecated" if there is some other feature or mechanism that is considered superior to it in all cases. It may just be more efficient, it may avoid potential pitfalls of the deprecated mechanism.

You should avoid using a deprecated feature in new code and should try to remove use of it from old code. The classic example of a deprecated function is 'gets'. Deprecated features are generally kept for compatability purposes or because standards require them. A function deprecated on one platform may be perfectly fine on another.

You might choose to use a deprecated feature for compatability purposes. It is not urgent to hack use of deprecated features out of existing code. (However, you should evaluate whether the reason for deprecation creates a problem in your code, especially if it's a security issue.)

Deprecated features can reasonably be expected to continue to work as well as they ever have. They should not be newly-broken. If a deprecated feature is broken in a new version, that is a regression.

A feature is "obsolete" if it no longer serves any purpose. It may be so broken that it is no longer even usable. It might still be technically working but not have any actual case of any significance in which you might want to use it. Linux's Sony CDU-535 driver is obsolete (a driver for an old CDROM drive with a specialized ISA controller that transferred data at well less than 1X and supports only a *very* slow PIO mode).

Obsolete features are generally scheduled for removal. They are usually not removed just on the off chance that the removal might harm someone. It may not even be possible to test them and the last time they were tested may have been several major releases ago. One should generally not expect obsolete features to work.

Something obsolete may conflict with never features. For example, a driver that breaks with hotpluggable CPUs. One contemplating using an obsolete feature in a new design should have a damn good reason and might possibly need to have their head examined.

The raw driver *is* deprecated. It is *not* obsolete.

DS

#17 lore raw
In reply to Robert P. J. Day
Quoted 22 lines · Robert P. J. Day
Quoted 8 lines · Jan Engelhardt
Quoted 4 lines · Dave Jones

Despite repeated attempts over the last two and half years, this driver seems somewhat persistant. Remove its deprecated status as it has existing users who may not be in a position to migrate their apps

At least keep the "it's obsolete" Kconfig description. We don't want new users/projects to jump on /dev/raw.

i just *know* this is a mistake, but i'm going to take one more shot at distinguishing between deprecated and obsolete.

as i understand it, the raw driver is *deprecated*. that is, it's still there, it's still supported, people are still using it but its use is *seriously* discouraged and everyone should be trying to move off of it at their earliest possible convenience.

that is *not* the same as "obsolete" which should mean that that feature is dead, dead, DEAD and *no one* should be using it anymore.

Quoted 8 lines

yes, i realize it sounds like splitting hairs, but it's this malleable definition of "deprecated" that's causing all of this trouble in the first place -- the fact that the raw driver is currently listed as "obsolete" when it is, in fact, only deprecated.

in short, do *not* remove its "deprecated" status. rather, remove its "obsolete" status and *make* it deprecated.

Correct. Like the weird lady next door who fancies you, it's old, it's ugly, but it's not likely to go away any time soon.

Signature · 3 lines
Bill Davidsen <[email protected]>
  "We have more to fear from the bungling of the incompetent than from
the machinations of the wicked."  - from Slashdot