Multiple functionality breakages in 2.6.12rc3 IDE layer

Multiple functionality breakages in 2.6.12rc3 IDE layer

[Alan Cox]

Ages ago we added an ide_default driver to clean up all the corner cases
like spurious IRQs for a device with no matching driver (eg ide-cd and
no CD driver) as well as ioctls and file access. 

2.6.12rc removes it. Unfortunately it also means that if your only IDE
interface is one you hand configure you can no longer run Linux. It also
changes other aspects of behaviour although they don't look problematic
for most users. You can no longer
	- Control the bus state of an interface
	- Reset an interface
	- Add an interface if none exist
	- Issue raw commands
	- Get an objects bios geometry
	- Read the identify data by ioctl (its still in proc but may be stale)

without having a device specific driver loaded matching the media - and
that only works if its already detected the device correctly.

I don't have the tools at the moment to generate spurious IRQ's for
devices with no driver loaded but it does look like the code may well
then crash. From the way the changes were done it appears the current
IDE maintainers never appreciated that ide_default existed for far more
than just cleaning up ide-proc but also to handle IRQ's, opening of
empty slots, ioctls and power
management ?

The ability to specify the IDE ports on the command line as needed for
some Sony laptop installs have also become "obsolete" over time. They
still appear to work but spew a warning that the user will soon be
screwed.

Alan

Re: Multiple functionality breakages in 2.6.12rc3 IDE layer

[Bill Davidsen]

On Thu, 28 Apr 2005, Alan Cox wrote:

> Ages ago we added an ide_default driver to clean up all the corner cases
> like spurious IRQs for a device with no matching driver (eg ide-cd and
> no CD driver) as well as ioctls and file access. 
> 
> 2.6.12rc removes it. Unfortunately it also means that if your only IDE
> interface is one you hand configure you can no longer run Linux. It also
> changes other aspects of behaviour although they don't look problematic
> for most users. You can no longer
> 	- Control the bus state of an interface
> 	- Reset an interface

But isn't that the stuff we use for swapping drives? Down the drive, down
the interface, swap, and restart? Are these the functions called by hdparm
(-bRU) which are all of a sudden not going to work? Or am I being
alarmist?

> 	- Add an interface if none exist
> 	- Issue raw commands
> 	- Get an objects bios geometry
> 	- Read the identify data by ioctl (its still in proc but may be stale)
> 
> without having a device specific driver loaded matching the media - and
> that only works if its already detected the device correctly.

I missed the discussion of why it was felt that the users would no longer
want to be able to do these things, or the new way to do it.
> 
> I don't have the tools at the moment to generate spurious IRQ's for
> devices with no driver loaded but it does look like the code may well
> then crash. From the way the changes were done it appears the current
> IDE maintainers never appreciated that ide_default existed for far more
> than just cleaning up ide-proc but also to handle IRQ's, opening of
> empty slots, ioctls and power
> management ?
> 
I suspect that's true, but should it not have been discussed first?

> The ability to specify the IDE ports on the command line as needed for
> some Sony laptop installs have also become "obsolete" over time. They
> still appear to work but spew a warning that the user will soon be
> screwed.

I'll have to read the release notes and see who to thank for this
reduction in functionality.

-- 
bill davidsen <davidsen@tmr.com>
  CTO, TMR Associates, Inc
Doing interesting things with little computers since 1979.

Re: Multiple functionality breakages in 2.6.12rc3 IDE layer

[Alan Cox]

On Iau, 2005-04-28 at 19:13, Bill Davidsen wrote:
> On Thu, 28 Apr 2005, Alan Cox wrote:

> But isn't that the stuff we use for swapping drives? Down the drive, down
> the interface, swap, and restart? Are these the functions called by hdparm
> (-bRU) which are all of a sudden not going to work? Or am I being
> alarmist?

True but the only kernels supporting that are 2.4.x-ac. There are
reasons I noticed this and looking at getting 2.6 IDE back to 2.4
standards in this area was one of them.

> I missed the discussion of why it was felt that the users would no longer
> want to be able to do these things, or the new way to do it.

I'm assuming it may be accidental rather than detailed planning. Also
its taken this long to notice so its clearly not that critical to
everyone. Seems to be reasonably sane to fix too.

Alan

Re: Multiple functionality breakages in 2.6.12rc3 IDE layer

[Bill Davidsen]

On Thu, 28 Apr 2005, Alan Cox wrote:

> On Iau, 2005-04-28 at 19:13, Bill Davidsen wrote:
> > On Thu, 28 Apr 2005, Alan Cox wrote:
> 
> > But isn't that the stuff we use for swapping drives? Down the drive, down
> > the interface, swap, and restart? Are these the functions called by hdparm
> > (-bRU) which are all of a sudden not going to work? Or am I being
> > alarmist?
> 
> True but the only kernels supporting that are 2.4.x-ac. There are
> reasons I noticed this and looking at getting 2.6 IDE back to 2.4
> standards in this area was one of them.

The amazing part is that I haven't had a drive fail since I went from 2.4
to 2.6 over a year ago. As you say, it works on your 2.4 kernels, I've "oh
shit" tested it. So if I lose a drive now I'm screwed, not my favorite
operating status.

> 
> > I missed the discussion of why it was felt that the users would no longer
> > want to be able to do these things, or the new way to do it.
> 
> I'm assuming it may be accidental rather than detailed planning. Also
> its taken this long to notice so its clearly not that critical to
> everyone. Seems to be reasonably sane to fix too.

I was being a bit sarcastic about the "missed the discussion" bit, but I'm
pretty sure ripping out the capability was deliberate. Hopefully it's now
going to be evaluated, and then fixed. One thing Linux doesn't seem to do
well is recover failed drives at boot time, it always seems to take a
bunch of fiddling or even a boot from live CD and hand recover.
> 
> Alan
> 

Thanks for jumping into this, with ATAPI storage down to about
$1100(US)/TB it's getting really hard to justify SCSI and real hot swap
hardware.

-- 
bill davidsen <davidsen@tmr.com>
  CTO, TMR Associates, Inc
Doing interesting things with little computers since 1979.

Re: Multiple functionality breakages in 2.6.12rc3 IDE layer

[Bartlomiej Zolnierkiewicz]

On 4/28/05, Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:
> Ages ago we added an ide_default driver to clean up all the corner cases

s/clean up/hide/

> like spurious IRQs for a device with no matching driver (eg ide-cd and
> no CD driver) as well as ioctls and file access.
>  
> 2.6.12rc removes it. Unfortunately it also means that if your only IDE
> interface is one you hand configure you can no longer run Linux. It also
> changes other aspects of behaviour although they don't look problematic
> for most users. You can no longer
>         - Control the bus state of an interface
>         - Reset an interface
>         - Add an interface if none exist
>         - Issue raw commands
>         - Get an objects bios geometry
>         - Read the identify data by ioctl (its still in proc but may be stale)

Details please.

> without having a device specific driver loaded matching the media - and
> that only works if its already detected the device correctly.
> 
> I don't have the tools at the moment to generate spurious IRQ's for
> devices with no driver loaded but it does look like the code may well
> then crash. From the way the changes were done it appears the current
> IDE maintainers never appreciated that ide_default existed for far more
> than just cleaning up ide-proc but also to handle IRQ's, opening of
> empty slots, ioctls and power
> management ?

Maybe you should mail current maintainer before spreading FUD?

No functionality was removed AFAIK, see the patches.  I spend quite 
a bit of time making sure that nothing breaks up (I missed one special
case but somebody already posted patch to LKML fixing it).

These patches were posted at least two times to both linux-ide and
linux-kernel, they were in -mm for ages - were you hiding under the
rock?

> The ability to specify the IDE ports on the command line as needed for
> some Sony laptop installs have also become "obsolete" over time. They
> still appear to work but spew a warning that the user will soon be
> screwed.

This was discussed few times already.

Alan, seriously, what is your problem?

Bartlomiej

Re: Multiple functionality breakages in 2.6.12rc3 IDE layer

[Alan Cox]

On Iau, 2005-04-28 at 21:41, Bartlomiej Zolnierkiewicz wrote:
> On 4/28/05, Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:
> > Ages ago we added an ide_default driver to clean up all the corner cases
> 
> s/clean up/hide/

Matter of opinion.

> > for most users. You can no longer
> >         - Control the bus state of an interface
> >         - Reset an interface
> >         - Add an interface if none exist
> >         - Issue raw commands
> >         - Get an objects bios geometry
> >         - Read the identify data by ioctl (its still in proc but may be stale)
> 
> Details please.

If you need details you shouldn't be maintaining that code. Its obvious
why. You've disabled open() of a device with no bound driver.

> No functionality was removed AFAIK, see the patches.  I spend quite 
> a bit of time making sure that nothing breaks up (I missed one special
> case but somebody already posted patch to LKML fixing it).

Build a kernel without ide-cd. Now try and issue ioctls on it. Doesn't
work any more does it.

> 
> These patches were posted at least two times to both linux-ide and
> linux-kernel, they were in -mm for ages - were you hiding under the
> rock?

No, just doing an MBA thesis, a job, learning a second language and
trying to beat sense into our politicians. Now I come back to look at
the ide layer ready for a 2.6.12 merge and its all a bit messy. The open
code was clean and is now duplicated. Copies of subtly different per
driver gendisk/disk layer open routines have appeared that should be
shared. The default driver handling has been removed and half the
options for obscure systems have been marked obsolete in some Gnome like
purge of functionality that might scare small children.

> > The ability to specify the IDE ports on the command line as needed for
> > some Sony laptop installs have also become "obsolete" over time. They
> > still appear to work but spew a warning that the user will soon be
> > screwed.
> 
> This was discussed few times already.

And the discussion lead to no fixes

> Alan, seriously, what is your problem?

The fact that the IDE layer appears to be getting worse not better,
which given the starting point is a remarkable achievement. 

Alan

Re: Multiple functionality breakages in 2.6.12rc3 IDE layer

[Bartlomiej Zolnierkiewicz]

On 4/29/05, Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:
> On Iau, 2005-04-28 at 21:41, Bartlomiej Zolnierkiewicz wrote:
> > On 4/28/05, Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:
> > > Ages ago we added an ide_default driver to clean up all the corner cases
> >
> > s/clean up/hide/
> 
> Matter of opinion.

No, it hided holes in the locking which is getting fixed finally.

> > > for most users. You can no longer
> > >         - Control the bus state of an interface
> > >         - Reset an interface
> > >         - Add an interface if none exist
> > >         - Issue raw commands
> > >         - Get an objects bios geometry
> > >         - Read the identify data by ioctl (its still in proc but may be stale)
> >
> > Details please.
> 
> If you need details you shouldn't be maintaining that code. Its obvious

Give details or quit whining.

> why. You've disabled open() of a device with no bound driver.

Guess what open() for ide-default was doing in 2.6?

return -ENXIO;

and no it wasn't my change - it was the effect of fixing
locking of the higher layers.

> > No functionality was removed AFAIK, see the patches.  I spend quite
> > a bit of time making sure that nothing breaks up (I missed one special
> > case but somebody already posted patch to LKML fixing it).
> 
> Build a kernel without ide-cd. Now try and issue ioctls on it. Doesn't
> work any more does it.
> 
> >
> > These patches were posted at least two times to both linux-ide and
> > linux-kernel, they were in -mm for ages - were you hiding under the
> > rock?
> 
> No, just doing an MBA thesis, a job, learning a second language and
> trying to beat sense into our politicians. Now I come back to look at

It seems that they influenced you heavily...

> the ide layer ready for a 2.6.12 merge and its all a bit messy. The open
> code was clean and is now duplicated. Copies of subtly different per

You must be joking.

> driver gendisk/disk layer open routines have appeared that should be

Each change was given rationale and detailed changelog, 
maybe you should get familiar with them.  Also look for
patch converting device drivers to sysfs (posted few times).

> shared. The default driver handling has been removed and half the
> options for obscure systems have been marked obsolete in some Gnome like
> purge of functionality that might scare small children.
> 
> > > The ability to specify the IDE ports on the command line as needed for
> > > some Sony laptop installs have also become "obsolete" over time. They
> > > still appear to work but spew a warning that the user will soon be
> > > screwed.
> >
> > This was discussed few times already.
> 
> And the discussion lead to no fixes

So fix the real bugs and leave debugging stuff alone.

> > Alan, seriously, what is your problem?
> 
> The fact that the IDE layer appears to be getting worse not better,
> which given the starting point is a remarkable achievement.

Personal insults are easy, get technical facts.

Bartlomiej

Re: Multiple functionality breakages in 2.6.12rc3 IDE layer

[Alan Cox]

On Gwe, 2005-04-29 at 00:00, Bartlomiej Zolnierkiewicz wrote:
> > why. You've disabled open() of a device with no bound driver.
> 
> Guess what open() for ide-default was doing in 2.6?
> 
> return -ENXIO;
> 
> and no it wasn't my change - it was the effect of fixing
> locking of the higher layers.

Yes so it needed fixing and without all the kref, kmalloc, unique object
structure per ide driver code spew too.

> > The fact that the IDE layer appears to be getting worse not better,
> > which given the starting point is a remarkable achievement.
> 
> Personal insults are easy, get technical facts.

I consider that a technical fact. The last IDE code I maintained fully
in 2.4 had mostly working locking, drive hotplug, open for unbound
drivers, didnt oops on spurious irqs and wasn't losing all sorts of
useful boot options. I had hoped that I wouldnt have to totally fork the
2.6 IDE code in order to get back to where 2.4-ac was and get the
locking working so you can't oops it via /proc

Alan

Re: Multiple functionality breakages in 2.6.12rc3 IDE layer

[Bartlomiej Zolnierkiewicz]

On 4/29/05, Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:
> On Gwe, 2005-04-29 at 00:00, Bartlomiej Zolnierkiewicz wrote:
> > > why. You've disabled open() of a device with no bound driver.
> >
> > Guess what open() for ide-default was doing in 2.6?
> >
> > return -ENXIO;
> >
> > and no it wasn't my change - it was the effect of fixing
> > locking of the higher layers.
> 
> Yes so it needed fixing and without all the kref, kmalloc, unique object
> structure per ide driver code spew too.

IDE is similar to SCSI now in this respect.
Are you claiming that SCSI needs fixing too?

> > > The fact that the IDE layer appears to be getting worse not better,
> > > which given the starting point is a remarkable achievement.
> >
> > Personal insults are easy, get technical facts.
> 
> I consider that a technical fact. The last IDE code I maintained fully
> in 2.4 had mostly working locking, drive hotplug, open for unbound
> drivers, didnt oops on spurious irqs and wasn't losing all sorts of
> useful boot options. I had hoped that I wouldnt have to totally fork the
> 2.6 IDE code in order to get back to where 2.4-ac was and get the
> locking working so you can't oops it via /proc

Somehow you seem to forget that I took maintaince over
2.5.5x (or 2.5.6x) and there was a lot new stuff added when
you were away (i.e. driver-model and IDE code needs to conform
with it to get sane power management and sysfs support)
and that a lot of other things have changed (influencing IDE).

Feel free to fork so you'll be wasting yours time only and not mine.