On Fri, May 13, 2011 at 07:03:48PM -0700, Yinghai Lu wrote: > On 05/13/2011 11:59 AM, Greg KH wrote: > > On Fri, May 13, 2011 at 10:04:15AM -0600, Alex Chiang wrote: > >> * Greg KH <greg@xxxxxxxxx>: > >>> On Fri, May 13, 2011 at 03:05:08AM -0600, Alex Chiang wrote: > >>>> * Jesse Barnes <jbarnes@xxxxxxxxxxxxxxxx>: > >>>>> Yinghai Lu <yinghai@xxxxxxxxxx> wrote: > >>>>>> > >>>>>> So remove rescan for all pci devices. less confusing > >>>>> > >>>>> I'm reluctant to break an existing interface like this, even though > >>>>> it's only in the "testing" section of the ABI. > >>>>> > >>>>> Alex, any comments? > >>>> > >>>> I think standard ABI deprecation practices should apply here. > >>> > >>> Ooh, I missed the fact that this series removed a sysfs file, you are > >>> right, we can't do that. > >>> > >>> So no, don't delete this file, please fix any tools that might happen to > >>> be using it, and then, in a year or so, we can revisit this to see if it > >>> makes sense to remove it. > >> > >> I thought we had some "best practices" to do that, like printing > >> out a big scary deprecation message when the file is accessed, > >> and documenting it in Documentation/ABI somewhere. > > > > Printing out such a message would be good, that would allow you to catch > > the tools that are doing it and change them. > > > > Documenting that the file is going away is for the > > Documentation/feature_removal.txt file, not the ABI files. > > > > So that all sounds like a good idea. > > thought if it is with testing, we can remove it if need to. Unfortunatly, no, everything ends up under "testing". And as there are tools that use this, how are you going to handle angry users that have their systems suddenly broken? > > or like this ? > > diff --git a/Documentation/feature-removal-schedule.txt b/Documentation/feature-removal-schedule.txt > index 4a9879b..2811a42 100644 > --- a/Documentation/feature-removal-schedule.txt > +++ b/Documentation/feature-removal-schedule.txt > @@ -559,3 +559,14 @@ Why: These legacy callbacks should no longer be used as i2c-core offers > Who: Jean Delvare <khali@xxxxxxxxxxxx> > > ---------------------------- > + > +What: rescan in /sys/bus/pci/devices/.../ > +When: June 2012 > +Why: That is not right. the device is already there, there is no reason to rescan it. > + We can not get increase resource for them. > + Now We already have rescan for pci_bus. > + So remove rescan for all pci devices. less confusing > + Finally We remove devices, and rescan bus that there were on before. > +Who: Yinghai Lu <yinghai@xxxxxxxxxx> I think the "Why" needs to be cleaned up a bit to make some more sense. thanks, greg k-h -- To unsubscribe from this list: send the line "unsubscribe linux-pci" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
