Skip to contentSkip to conversation

Delete JFFS (version 1)

PATCH/RFC Jeff Garzik 10 participants 16 messages Dec 12, 2006 – Mar 9, 2007

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

I have created the 'kill-jffs' branch of git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/misc-2.6.git that removes fs/jffs.

I argue that you can count the users (who aren't on 2.4) on one hand, and developers don't seem to have cared for it in ages.

People are already talking about jffs2 replacements, so I propose we zap jffs in 2.6.21.

Jeff

Documentation/feature-removal-schedule.txt +7 -0
index 46f2a55..c008303 100644
@@ -270,3 +270,10 @@ Why:	The new layer 3 independant connection tracking replaces the old
 Who:	Patrick McHardy <[email protected]>
 
 ---------------------------
+
+What:	JFFS (version 1) filesystem
+When:	2.6.21
+Why:	No users or developers
+Who:	Jeff Garzik <[email protected]>
+
+---------------------------
#1 lore raw
Quoted 9 lines · Jeff Garzik

I have created the 'kill-jffs' branch of
git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/misc-2.6.git that removes fs/jffs.

I argue that you can count the users (who aren't on 2.4) on one hand, and developers don't seem to have cared for it in ages.

People are already talking about jffs2 replacements, so I propose we zap jffs in 2.6.21.

I'm usually all for killing broken code, but JFFS isn't really broken is it? Is there some burden it's causing by being in the kernel at the moment?

josh

#10 lore raw
In reply to Jeff Garzik
Quoted 7 lines · Jeff Garzik

I have created the 'kill-jffs' branch of
git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/misc-2.6.git that removes fs/jffs.

I argue that you can count the users (who aren't on 2.4) on one hand, and developers don't seem to have cared for it in ages.

You can count them on one finger, I guess. The last time jffs was proposed for removal, a single person said he still used it and would like to keep it around. And I haven't seen any patches come forward since.

Jörn

Signature · 6 lines
Don't patch bad code, rewrite it.
-- Kernigham and Pike, according to Rusty
-
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#2 lore raw
In reply to Josh Boyer
Quoted 8 lines · Josh Boyer
Quoted 7 lines · Jeff Garzik

I have created the 'kill-jffs' branch of
git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/misc-2.6.git that removes fs/jffs.

I argue that you can count the users (who aren't on 2.4) on one hand, and developers don't seem to have cared for it in ages.

Quoted 6 lines
Quoted 2 lines

People are already talking about jffs2 replacements, so I propose we zap jffs in 2.6.21.

I'm usually all for killing broken code, but JFFS isn't really broken is it? Is there some burden it's causing by being in the kernel at the moment?

It's always been the case that we remove Linux kernel code when the number of users (and more importantly, developers) drops to near-nil.

Every line of code is one more place you have to audit when code changes, one more place to update each time the VFS API is touched.

When it's more likely to get struck by lightning than encounter filesystem X on a random hard drive in the field, filesystem X need not be in the kernel.

IMO, of course :)

Jeff

#3 lore raw
Quoted 16 lines · Jeff Garzik
Quoted 8 lines · Josh Boyer
Quoted 7 lines · Jeff Garzik

I have created the 'kill-jffs' branch of
git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/misc-2.6.git that removes fs/jffs.

I argue that you can count the users (who aren't on 2.4) on one hand, and developers don't seem to have cared for it in ages.

Quoted 6 lines
Quoted 2 lines

People are already talking about jffs2 replacements, so I propose we zap jffs in 2.6.21.

I'm usually all for killing broken code, but JFFS isn't really broken is it? Is there some burden it's causing by being in the kernel at the moment?

Quoted 5 lines

It's always been the case that we remove Linux kernel code when the number of users (and more importantly, developers) drops to near-nil.

Every line of code is one more place you have to audit when code changes, one more place to update each time the VFS API is touched.

Ok, I can buy that.

Quoted 4 lines

When it's more likely to get struck by lightning than encounter filesystem X on a random hard drive in the field, filesystem X need not be in the kernel.

Or flash chip in this case ;)

josh

#4 lore raw
Quoted 23 lines · Josh Boyer
Quoted 16 lines · Jeff Garzik
Quoted 8 lines · Josh Boyer
Quoted 7 lines · Jeff Garzik

I have created the 'kill-jffs' branch of
git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/misc-2.6.git that removes fs/jffs.

I argue that you can count the users (who aren't on 2.4) on one hand, and developers don't seem to have cared for it in ages.

Quoted 6 lines
Quoted 2 lines

People are already talking about jffs2 replacements, so I propose we zap jffs in 2.6.21.

I'm usually all for killing broken code, but JFFS isn't really broken is it? Is there some burden it's causing by being in the kernel at the moment?

Quoted 5 lines

It's always been the case that we remove Linux kernel code when the number of users (and more importantly, developers) drops to near-nil.

Every line of code is one more place you have to audit when code changes, one more place to update each time the VFS API is touched.

Quoted 8 lines

Ok, I can buy that.

Quoted 4 lines

When it's more likely to get struck by lightning than encounter filesystem X on a random hard drive in the field, filesystem X need not be in the kernel.

Or flash chip in this case ;)

More to the point, people have occasionally actually _used_ JFFS instead of JFFS2. I'm all for removing it now.

Signature · 1 line
dwmw2
#5 lore raw
In reply to Jeff Garzik
Quoted 1 line · Jeff Garzik

When it's more likely to get struck by lightning than encounter filesystem X on a random hard drive in the field, filesystem X need not be in the kernel.

As people are already poking me:) I course meant "flash device" not "hard drive".

SATA maintainer's curse, I suppose, to think of all storage devices as hard drives, no matter how incorrect that might be :)

Jeff

#6 lore raw
In reply to Jeff Garzik
Quoted 2 lines · Jeff Garzik

It's always been the case that we remove Linux kernel code when the number of users (and more importantly, developers) drops to near-nil.

So, drivers/net/3c501.c?

Bill

#7 lore raw
Quoted 4 lines · Bill Nottingham
Quoted 1 line · Jeff Garzik

It's always been the case that we remove Linux kernel code when the number of users (and more importantly, developers) drops to near-nil.

So, drivers/net/3c501.c?

Depends on how motivated Alan remains ;-) Historically, if the developer is active, we have occasionally ignored the miniscule userbase.

Jeff

#8 lore raw
In reply to Bill Nottingham
Quoted 5 lines · On Tue, 12 Dec 2006 12:01:25 -0500 Bill Nottingham <[email protected]> wrote:
Quoted 2 lines · Jeff Garzik

It's always been the case that we remove Linux kernel code when the number of users (and more importantly, developers) drops to near-nil.

So, drivers/net/3c501.c?

I think 3c501.c is a case where putting in an "If you use this email .. or it will go away" might be a good idea indeed.

#11 lore raw
In reply to Jeff Garzik
Quoted 9 lines · On Tue, 12 Dec 2006 07:39:26 -0500 Jeff Garzik <[email protected]> wrote:

I have created the 'kill-jffs' branch of
git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/misc-2.6.git that removes fs/jffs.

I argue that you can count the users (who aren't on 2.4) on one hand, and developers don't seem to have cared for it in ages.

People are already talking about jffs2 replacements, so I propose we zap jffs in 2.6.21.

It would be good to provide unignorable notice of this in 2.6.20. Via a loud printk, or perhaps even CONFIG_BROKEN or CONFIG_EMBEDDED.

#12 lore raw
Quoted 7 lines · Andrew Morton
Quoted 4 lines · On Tue, 12 Dec 2006 07:39:26 -0500 Jeff Garzik <[email protected]> wrote:

I have created the 'kill-jffs' branch of
git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/misc-2.6.git that removes fs/jffs.

Quoted 8 lines
Quoted 5 lines

I argue that you can count the users (who aren't on 2.4) on one hand, and developers don't seem to have cared for it in ages.

People are already talking about jffs2 replacements, so I propose we zap jffs in 2.6.21.

It would be good to provide unignorable notice of this in 2.6.20. Via a loud printk, or perhaps even CONFIG_BROKEN or CONFIG_EMBEDDED.

Something like the below?

Make CONFIG_JFFS depend on CONFIG_BROKEN

fs/Kconfig +0 -0
index b3b5aa0..4ac367d 100644
@@ -1204,13 +1204,16 @@ config EFS_FS
 

config JFFS_FS
tristate "Journalling Flash File System (JFFS) support"
- depends on MTD && BLOCK
+ depends on MTD && BLOCK && BROKEN
help
JFFS is the Journalling Flash File System developed by Axis Communications in Sweden, aimed at providing a crash/powerdown-safe file system for disk-less embedded devices. Further information is available at (<http://developer.axis.com/software/jffs/>).

+ NOTE: This filesystem is deprecated and is scheduled for removal in
+ 2.6.21. See Documentation/feature-removal-schedule.txt
+
config JFFS_FS_VERBOSE
int "JFFS debugging verbosity (0 = quiet, 3 = noisy)" depends on JFFS_FS

#13 lore raw
Quoted 26 lines · Josh Boyer
Quoted 2 lines · Andrew Morton

It would be good to provide unignorable notice of this in 2.6.20. Via a loud printk, or perhaps even CONFIG_BROKEN or CONFIG_EMBEDDED.

Something like the below?

Make CONFIG_JFFS depend on CONFIG_BROKEN

fs/Kconfig +0 -0
index b3b5aa0..4ac367d 100644
@@ -1204,13 +1204,16 @@ config EFS_FS
 

config JFFS_FS
tristate "Journalling Flash File System (JFFS) support"

fs/Kconfig +1 -1
-	depends on MTD && BLOCK
+	depends on MTD && BLOCK && BROKEN

help
JFFS is the Journalling Flash File System developed by Axis Communications in Sweden, aimed at providing a crash/powerdown-safe file system for disk-less embedded devices. Further information is available at (<http://developer.axis.com/software/jffs/>).

fs/Kconfig

Quoted 3 lines
fs/Kconfig +3 -0
+	  NOTE: This filesystem is deprecated and is scheduled for removal in
+	  2.6.21.  See Documentation/feature-removal-schedule.txt
+

It would be better to have a printk in the module init, since users with upstream builds won't see the config help.

Cheers, Andreas
--
Andreas Dilger
Principal Software Engineer
Cluster File Systems, Inc.

#14 lore raw
Quoted 3 lines · Andreas Dilger

It would be better to have a printk in the module init, since users with upstream builds won't see the config help.

How many of those are there? This is not ext3, this is a flash filesystem. Users are either those doing the upstream build or they won't ever see a console. Think DSL-Router. ;)

Jörn

Signature · 6 lines
Joern's library part 14:
http://www.sandpile.org/
-
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#9 lore raw
In reply to Alan
Quoted 11 lines · Alan
Quoted 5 lines · On Tue, 12 Dec 2006 12:01:25 -0500 Bill Nottingham <[email protected]> wrote:
Quoted 2 lines · Jeff Garzik

It's always been the case that we remove Linux kernel code when the number of users (and more importantly, developers) drops to near-nil.

So, drivers/net/3c501.c?

I think 3c501.c is a case where putting in an "If you use this email .. or it will go away" might be a good idea indeed.

Maybe even: "the people who are using 3c501's - please chime up and we'll donate better cards to you."

After all, 3c501 is one of the crappier pieces of network cards, I feel sorry for the people using them...

Regards: David

Signature · 3 lines
 /) David Weinehall <[email protected]> /) Northern lights wander      (\
//  Maintainer of the v2.0 kernel   //  Dance across the winter sky //
\)  http://www.acc.umu.se/~tao/    (/   Full colour fire           (/
#15 lore raw
In reply to Jeff Garzik
Quoted 2 lines

I argue that you can count the users (who aren't on 2.4) on one hand, and developers don't seem to have cared for it in ages.

Rather than ask on mailing-lists, it's probably easier to just make the jffs compilation fail (with a #error). This way, if someone uses it, he'll bump into it, no matter how much he wants to ignore messages posted on mailing-lists.

Stefan