Re: 2.6.25-rc2: Reported regressions from 2.6.24

["Rafael J. Wysocki" <rjw@sisk.pl>]

On Monday, 18 of February 2008, Jiri Slaby wrote:
> On 02/18/2008 10:49 AM, Ingo Molnar wrote:
> > * Alexey Dobriyan <adobriyan@gmail.com> wrote:
> > 
> >> Why on earth do we have CONFIG_SYSFS_DEPRECATED at all if the only 
> >> sane choice for someone who tests new kernels is to always turn it on?
> > 
> > seconded ...
> > 
> > it's insane how inconvenient certain storage systems are to use, and 
> > people get turned away with NOATBUG. ITISABUG every time user-space 
> > breaks or the user has to do _anything_ to get the box working as it 
> > should.
> 
> No, user needn't to do anything. Just turn that default y option really to y 
> (yes, you are right, the help text of this option is misleading...). It's like 
> you turn off old acpi events option and scream that you acpi daemon doesn't work 
> for instance. There are many instances of this behaviour in the kernel. And yes, 
> many people don't need that option tuyrned on -- I think we need not-y testers 
> too, but it might be my personal feeling.

Well, my definition of a regression from 2.6.24 is that it happens when someone
takes a .config that worked with 2.6.24, configures the kernel with that,
leaving the defaults for the options he is _asked_ _for_, and the resulting
kernel doesn't work as expected.

Thanks,
Rafael