Driver removals

Driver removals

[Bill Davidsen]

Just a general thought on removing drivers in general, when a driver is 
removed because there's a better one, it would be good to have either a 
message which shows up at "make oldconfig" time, or a file listing the 
driver(s) which replace it.

Half the resistance to removing drivers is finding what is supposed to 
replace the driver. For instance a list of all the drivers Adrian Bunk 
has suggested to replace sk98lin, so users have something better than 
searching the source code and/or LKML archive to find the next thing to try.

In general, if a driver works and is being used, until it *needs* 
attention I see no reason to replace it. I don't agree that "it forces 
people to try the new driver" is a valid reason, being unmaintained is 
only a problem if it needs maintenance. I am not going to reopen that 
topic, I'm simply noting a general opposition to unfunded mandates, and 
requiring changes to kernel, module and/or rc.local config is just that.

-- 
Bill Davidsen <davidsen@tmr.com>
   "We have more to fear from the bungling of the incompetent than from
the machinations of the wicked."  - from Slashdot

Re: Driver removals

[Adrian Bunk]

On Wed, Feb 13, 2008 at 09:26:26PM -0500, Bill Davidsen wrote:
>...
> In general, if a driver works and is being used, until it *needs*  
> attention I see no reason to replace it. I don't agree that "it forces  
> people to try the new driver" is a valid reason, being unmaintained is  
> only a problem if it needs maintenance. I am not going to reopen that  
> topic, I'm simply noting a general opposition to unfunded mandates, and  
> requiring changes to kernel, module and/or rc.local config is just that.

Keeping a working unmaintained driver in the tree is not a big deal - we 
have hundreds of them.

But you miss the main point that removal of an obsolete driver with a 
new replacement driver forces people to finally report their problems 
with the new driver, thus making the new driver better.

After all, the people who scream loudly that the new driver doesn't work 
for them when the old driver gets removed are the people who should have 
reported their problems with the new driver many years ago...

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

Re: Driver removals

[Bill Davidsen]

Adrian Bunk wrote:
> On Wed, Feb 13, 2008 at 09:26:26PM -0500, Bill Davidsen wrote:
>> ...
>> In general, if a driver works and is being used, until it *needs*  
>> attention I see no reason to replace it. I don't agree that "it forces  
>> people to try the new driver" is a valid reason, being unmaintained is  
>> only a problem if it needs maintenance. I am not going to reopen that  
>> topic, I'm simply noting a general opposition to unfunded mandates, and  
>> requiring changes to kernel, module and/or rc.local config is just that.
> 
> Keeping a working unmaintained driver in the tree is not a big deal - we 
> have hundreds of them.
> 
> But you miss the main point that removal of an obsolete driver with a 
> new replacement driver forces people to finally report their problems 
> with the new driver, thus making the new driver better.
> 
You sure are proud of that new driver! People won't use it because the 
old one is working fine, so you think it's fine to force them to make 
changes in their system to use the new driver. Best case is it works 
after costing the user some time, worst case it doesn't and breaks their 
system, so they stop upgrading the kernel and don't get security fixes.

> After all, the people who scream loudly that the new driver doesn't work 
> for them when the old driver gets removed are the people who should have 
> reported their problems with the new driver many years ago...
> 
Is it not obvious that the problem lies with the "when the old driver 
gets removed" part, there is absolutely no effort needed to keep an old 
driver, and if it's left in until some change requires rewriting every 
module in the kernel, it's likely that either the old hardware or the 
user will die before that ever happens again.

There is no benefit to users, if the old driver didn't work they would 
have switched, there's no saving in support effort because, as you 
pointed out, there are "hundreds of them" now.

This reminds me of Microsoft and XP vs. VISTA. MSFT is stopping sales 
and support of XP to "force people to upgrade" to VISTA. You want to 
"force people to upgrade" to newer drivers. The difference is that MSFT 
at least has money as a reason, as far as I can tell the reason you want 
to force people to use new drivers is because people put all the effort 
into writing the new drivers and now most of us want to use the old one 
if it works. We don't want to change configuration and hope something 
else new works, because we know the old driver works for us and we want 
to use our system instead of helping test the new driver.

I appreciate the effort it took to write new drivers, I believe most 
users able to build their own kernels do. I use new drivers on new 
systems because install picks them and a new system has to go through 
shakedown in any case. I just wish that *you* could appreciate that a 
driver change requires user effort and chance to find bugs in a new 
driver, for each and every system, many of which are at EOL now. I wish 
you valued the user's time as much as users value developer time.

*EOL - end-of-life, if your organization doesn't use the term. the "it's 
paid for, use it but don't spend money on it" phase of ownership.


-- 
Bill Davidsen <davidsen@tmr.com>
   "We have more to fear from the bungling of the incompetent than from
the machinations of the wicked."  - from Slashdot

Re: Driver removals

[Adrian Bunk]

On Fri, Feb 15, 2008 at 02:07:41PM -0500, Bill Davidsen wrote:
> Adrian Bunk wrote:
>> On Wed, Feb 13, 2008 at 09:26:26PM -0500, Bill Davidsen wrote:
>>> ...
>>> In general, if a driver works and is being used, until it *needs*   
>>> attention I see no reason to replace it. I don't agree that "it 
>>> forces  people to try the new driver" is a valid reason, being 
>>> unmaintained is  only a problem if it needs maintenance. I am not 
>>> going to reopen that  topic, I'm simply noting a general opposition 
>>> to unfunded mandates, and  requiring changes to kernel, module and/or 
>>> rc.local config is just that.
>>
>> Keeping a working unmaintained driver in the tree is not a big deal - 
>> we have hundreds of them.
>>
>> But you miss the main point that removal of an obsolete driver with a  
>> new replacement driver forces people to finally report their problems  
>> with the new driver, thus making the new driver better.
>>
> You sure are proud of that new driver! People won't use it because the  
> old one is working fine, so you think it's fine to force them to make  
> changes in their system to use the new driver.

Sometimes what is best in the global picture is not what everyone
subjectively considers to be the best thing for him.

Well, our whole society is based on this principle...

> Best case is it works  
> after costing the user some time, worst case it doesn't and breaks their  
> system, so they stop upgrading the kernel and don't get security fixes.
>...

Instead of sending a bug report?

When removing an obsolete driver adult people suddenly start whining
"the new driver didn't work for me when I tried it one year ago".

And when asking where they reported the bug in the new driver the answer 
is that they didn't report it.

Driver development heavily relies on getting bug reports when something 
doesn't work.

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

Re: Driver removals

[Bill Davidsen]

Adrian Bunk wrote:
> On Fri, Feb 15, 2008 at 02:07:41PM -0500, Bill Davidsen wrote:
>   
>> Adrian Bunk wrote:
>>     
>>> On Wed, Feb 13, 2008 at 09:26:26PM -0500, Bill Davidsen wrote:
>>>       
>>>> ...
>>>> In general, if a driver works and is being used, until it *needs*   
>>>> attention I see no reason to replace it. I don't agree that "it 
>>>> forces  people to try the new driver" is a valid reason, being 
>>>> unmaintained is  only a problem if it needs maintenance. I am not 
>>>> going to reopen that  topic, I'm simply noting a general opposition 
>>>> to unfunded mandates, and  requiring changes to kernel, module and/or 
>>>> rc.local config is just that.
>>>>         
>>> Keeping a working unmaintained driver in the tree is not a big deal - 
>>> we have hundreds of them.
>>>
>>> But you miss the main point that removal of an obsolete driver with a  
>>> new replacement driver forces people to finally report their problems  
>>> with the new driver, thus making the new driver better.
>>>
>>>       
>> You sure are proud of that new driver! People won't use it because the  
>> old one is working fine, so you think it's fine to force them to make  
>> changes in their system to use the new driver.
>>     
>
> Sometimes what is best in the global picture is not what everyone
> subjectively considers to be the best thing for him.
>
> Well, our whole society is based on this principle...
>
>   
>> Best case is it works  
>> after costing the user some time, worst case it doesn't and breaks their  
>> system, so they stop upgrading the kernel and don't get security fixes.
>> ...
>>     
>
> Instead of sending a bug report?
>   

To get the system working.
> When removing an obsolete driver adult people suddenly start whining
> "the new driver didn't work for me when I tried it one year ago".
>
> And when asking where they reported the bug in the new driver the answer 
> is that they didn't report it.
>
> Driver development heavily relies on getting bug reports when something 
> doesn't work.

If you don't see an ethical problem in removing a working driver which 
is not taking support resources, in order to force people to test and 
debug a driver they don't now and never would need, so that it might in 
time offer them the same functionality those users already had... then I 
can never make you see why technological extortion is evil. People have 
always moved to new drivers without pushing because they were *better*, 
guess that model is dead.

-- 
Bill Davidsen <davidsen@tmr.com>
  "Woe unto the statesman who makes war without a reason that will still
  be valid when the war is over..." Otto von Bismark 

Re: Driver removals

[Valdis.Kletnieks]

[-- Attachment #1: Type: text/plain, Size: 980 bytes --]

On Fri, 15 Feb 2008 20:08:13 EST, Bill Davidsen said:

> can never make you see why technological extortion is evil. People have 
> always moved to new drivers without pushing because they were *better*, 
> guess that model is dead.

And the drivers get better because the Code Fairy comes and sprinkles magic
bugfix dust all over them? And the Code Fairy knows where to sprinkle bugfix
dust because it can see where the Bugreport Fairy sprinkled Bugreport Dust?

Yes, people will move without pushing when the drivers are better.  However,
remember that a major cultural motivation for the GPL is the concept of "give
back".  And if a user can't be bothered to even give back enough to say
"wow, this blows, my Frobnozz9800 doesn't work with this driver", that's a
problem with the user.  They're getting it for free, they should at least
give the developers the kindness of a bug report if something is broken...

<insert diatribe about users who just want free-as-in-beer>....

[-- Attachment #2: Type: application/pgp-signature, Size: 226 bytes --]

Re: Driver removals

[Adrian Bunk]

On Fri, Feb 15, 2008 at 08:08:13PM -0500, Bill Davidsen wrote:
>...
> If you don't see an ethical problem in removing a working driver which  
> is not taking support resources, in order to force people to test and  
> debug a driver they don't now and never would need, so that it might in  
> time offer them the same functionality those users already had... then I  
> can never make you see why technological extortion is evil.
>...

You miss one basic principle of free software:

Forks are allowed, so when you don't like the way some software is 
developed you can always release a version of the software that is in 
your eyes better.

And that's nothing evil, after all each distribution kernel is a fork of 
the upstream kernel.

Hey, you can even use the 2.6.16 branch *I* do maintain to avoid what 
you claim was an "ethical problem".

And if you don't like 2.6.16 either for whatever reason there's still no 
ethical problem but only the technical problem of you not getting your 
ass up and doing whatever is "better" in your opinion.

After all, free software is not driven by people whining but by people 
doing stuff.

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

Re: Driver removals

[Willy Tarreau]

On Fri, Feb 15, 2008 at 08:52:27PM -0500, Valdis.Kletnieks@vt.edu wrote:
> On Fri, 15 Feb 2008 20:08:13 EST, Bill Davidsen said:
> 
> > can never make you see why technological extortion is evil. People have 
> > always moved to new drivers without pushing because they were *better*, 
> > guess that model is dead.
> 
> And the drivers get better because the Code Fairy comes and sprinkles magic
> bugfix dust all over them? And the Code Fairy knows where to sprinkle bugfix
> dust because it can see where the Bugreport Fairy sprinkled Bugreport Dust?
> 
> Yes, people will move without pushing when the drivers are better.  However,
> remember that a major cultural motivation for the GPL is the concept of "give
> back".  And if a user can't be bothered to even give back enough to say
> "wow, this blows, my Frobnozz9800 doesn't work with this driver", that's a
> problem with the user.

I don't understand why kernel developers always think that users spend
their whole time testing their new stuff. That is mostly true for a lot
of desktop users, but definitely not for servers. On a server, you may
*ignore* that a new driver exists for years. The basic make oldconfig
does the stuff right.

An old driver must spend some time marked "deprecated", if possible with
the config option changed so that at least *something* informs the admin
that it may soon be removed. It looks like this is something that people
building a kernel every day and never getting more than one week of uptime
do not understand. But there are many people who build once a year and
upgrade that often at most, unless there is a big security issue.

If the old driver simply keeps silently building when marked deprecated,
noone will notice. And as Bill pointed it out, we should also make sure
that when marked deprecated, the old one always refers to the new one so
that the guy noticing this during the build has time to set up a test
machine to try that new driver.

Not everyone has a mouse and a joystick attached to the computers he
builds kernels for...

Willy

Re: Driver removals

[David Newall]

Adrian Bunk wrote:
> Forks are allowed, so when you don't like the way some software is 
> developed you can always release a version of the software that is in 
> your eyes better.
>   
What a silly thought.  Nobody, I should hope, wants multiple Linuxes
competing and diluting the market.  That's what went wrong with UNIX,
and it's what's wrong with BSD (and what gave Linux a foothold.)  Read
your history; and for goodness sake, the less said on this the better.

> After all, free software is not driven by people whining but by people 
> doing stuff.

Howling protest is not whining.  The whining comes from those who want
to kill the old driver, which is reported to be used, works well and is
wanted.  It sounds insecure to want to terminate one's competition with
such extreme prejudice.  Let the old driver die of natural causes, if
that's its destiny.  And don't whine if the new can't make it on its own
merits.

Re: Driver removals

[Bill Davidsen]

Valdis.Kletnieks@vt.edu wrote:
> On Fri, 15 Feb 2008 20:08:13 EST, Bill Davidsen said:
>
>   
>> can never make you see why technological extortion is evil. People have 
>> always moved to new drivers without pushing because they were *better*, 
>> guess that model is dead.
>>     
>
> And the drivers get better because the Code Fairy comes and sprinkles magic
> bugfix dust all over them? And the Code Fairy knows where to sprinkle bugfix
> dust because it can see where the Bugreport Fairy sprinkled Bugreport Dust?
>   

Drivers get better because someone who finds a benefit in them also 
finds a problem. They don't get better by developers looking for 
intermittent, probably load dependent, bugs which effect eight year old 
server hardware which was a low volume item, the developers are unlikely 
to have, and which is in use providing services, not on someone's 
desktop where it can reasonably be rebooted to test patches.
> Yes, people will move without pushing when the drivers are better.  However,
> remember that a major cultural motivation for the GPL is the concept of "give
> back".  And if a user can't be bothered to even give back enough to say
> "wow, this blows, my Frobnozz9800 doesn't work with this driver", that's a
> problem with the user.  They're getting it for free, they should at least
> give the developers the kindness of a bug report if something is broken...
>   

Not when drivers are "better" but when a new driver offers some benefit, 
be it reliability, capability, etc. When the new driver offers not a 
single benefit and the only "feature" is "possible unknown bugs," people 
are not going to change, and I don't think forcing people off working 
drivers is any more ethical than Microsoft killing XP to force people to 
VISTA. Less ethical, actually, because MSFT is looking for profit, and 
they make no pretense of caring about the users in any role but revenue 
stream.
> <insert diatribe about users who just want free-as-in-beer>....
>   

Insert it right next to the diatribe about developers who think that 
because some new feature was a lot of work that Linus *must* put it in 
the kernel, or users show show proper respect and gratitude and disrupt 
their production hardware to test and debug some new code which offers 
zero added functionality on that hardware. If you think downtime is 
"free" then you  have not been working in the right environments, or for 
the right management.

-- 
Bill Davidsen <davidsen@tmr.com>
  "Woe unto the statesman who makes war without a reason that will still
  be valid when the war is over..." Otto von Bismark 

Re: Driver removals

[Valdis.Kletnieks]

[-- Attachment #1: Type: text/plain, Size: 1561 bytes --]

On Sat, 16 Feb 2008 08:39:08 +0100, Willy Tarreau said:

> I don't understand why kernel developers always think that users spend
> their whole time testing their new stuff. That is mostly true for a lot
> of desktop users, but definitely not for servers. On a server, you may
> *ignore* that a new driver exists for years. The basic make oldconfig
> does the stuff right.

Been there, done that.  We got a number of boxes that sit there for ages
between reboots and even longer between upgrades.  Deploying Solaris 10 was
like a 2-year process for some of our boxes (when the boxes are running the
Oracle databases that house the enterprise business systems, and your
organizational budget is pushing the $1B/year mark, everybody gets *really*
cautious to avoid a CLM while upgrading... ;)  And we may actually manage to
finally kill off our last AIX 4.3.3 box this quarter (4.3.3 was released all
the way back in Sep 1999, and EOL'ed back in 2004 - don't ask. :)

Of course, the difference is that we *expect* that there's a good chance that
if we try to subject that sort of box to 3 year's worth of updates, there's a
good chance we'll discover that some driver has been EOL'ed or otherwise
problematic on now-ancient hardware...

(And yes, there's a *lot* of risk-management going on centered around "What if
we have to patch Server23 for a critical kernel security issue?".  Of course,
if it was actually *feasible* to blindly upgrade-and-reboot twice a month, the
job could be done by a much less expensive patch monkey, so it's good job
security ;)





[-- Attachment #2: Type: application/pgp-signature, Size: 226 bytes --]