On Fri, 2012-10-26 at 14:16 +0200, Gabor Z. Papp wrote:
> * Greg KH <[email protected]>:
> 
> | > > $ dd if=/dev/zero of=/dev/fd0
> | > >
> | > > ------------[ cut here ]------------
> | > > WARNING: at drivers/block/floppy.c:1041 setup_rw_floppy+0x2f7/0x310 
> [floppy]()
> | > > Hardware name: System Product Name
> | > > floppy_disable_hlt() scheduled for removal in 2012
> | > 
> | > Yes.  I don't understand the point of that warning:
> | > http://bugs.debian.org/667501
> | > 
> | > Ben and Greg, would
> | > 
> | >   f6365201d8a2 x86: Remove the ancient and deprecated disable_hlt()
> | >                and enable_hlt() facility
> | > 
> | > be a candidate for inclusion in the 3.0.y and 3.2.y trees?
> | > 
> | > An alternative would be to revert 3b70b2e5fcf6 ("x86 idle floppy:
> | > deprecate disable_hlt()", 2011-04-01), which in principle seems a
> | > little safer.
> 
> | Gabor, does applying this patch fix this issue?
> 
> Yes, using patch¹ fix this issue with 3.0.48 and 3.2.32.
> 
> ¹http://bugs.debian.org/cgi-bin/bugreport.cgi?msg=10;filename=x86-Remove-the-ancient-and-deprecated-disable_hlt-an.patch;att=1;bug=667501

I find this particular deprecation process deeply flawed.  Since we had
this hack for ages and it wasn't restricted to specific known-broken
CPUs or chipsets, how can we be confident that no later 32-bit PCs
depend on it?  Why was the warning issued to floppy users *before* the
change - with no option to test the new behaviour and quiet the warning
- and not after?  Many distribution users who skip several kernel
versions will never see the warning at all.

(Bonus bug: the warning was not dependent on CONFIG_X86_32.)

Greg, which of these bad options do you think is preferable?

Ben.

-- 
Ben Hutchings
Reality is just a crutch for people who can't handle science fiction.

Attachment: signature.asc
Description: This is a digitally signed message part

Reply via email to