On Sun, Feb 29, 2004 at 01:53:58AM -0500, Dmitry Torokhov wrote:
> Here is the new set of input patches that I have. You have seen some of
> them, buit this time they are rediffed against 2.6.4-rc1 and in nice order.
I like them very much. Do you have a bitkeeper tree anywhere where I
could pull from, so that I don't have to apply these by hand?
> 01-atkbd-whitespace-fixes.patch
> simple whitespace fixes
>
> 02-atkbd-bad-merge.patch
> clean up bad merge in atkbd module (get rid of MODULE_PARMs,
> atkbd_softrepeat was declared twice)
>
> 03-synaptics-relaxed-proto.patch
> some hardware (PowerBook) require relaxed Synaptics protocol checks,
> but relaxed checks hurt hardware implementing proper protocol when
> device looses sync. With the patch synaptics driver analyzes first
> full data packet and either staus in relaxed mode or switches into
> strict mode.
>
> 04-psmouse-whitespace-fixes.patch
> simple whitespace fixes
>
> 05-psmouse-workaround-noack.patch
> some mice do not ACK "disable streaming mode" command causing psmouse
> driver abort initialization without any indication to the user. This
> is a regression compared to 2.4. Have kernel complain but continue
> with prbing hardware (after all we got valid responce from GET ID
> command).
>
> 06-module-param-array-named.patch
> introduce module_param_array_named() modeled after module_param_named
> that allows mapping array module option to
>
> 07-joystick-module-param.patch
> complete moving input drivers to the new way of handling module
> parameters using module_param()
>
> 08-obsolete-setup.patch
> introduce __obsolete_setup(). This is a drop-in replacement for
> __setup() for truly obsolete options. Kernel will complain when sees
> such an option.
>
> 09-input-obsolete-setup.patch
> document removed or renamed options in input drivers using
> __obsolete_setup() so users will have some clue why old options
> stopped having any effect.
--
Vojtech Pavlik
SuSE Labs, SuSE CR
On Tuesday 02 March 2004 08:02 am, Vojtech Pavlik wrote:
> On Sun, Feb 29, 2004 at 01:53:58AM -0500, Dmitry Torokhov wrote:
>
> > Here is the new set of input patches that I have. You have seen some of
> > them, buit this time they are rediffed against 2.6.4-rc1 and in nice order.
>
> I like them very much. Do you have a bitkeeper tree anywhere where I
> could pull from, so that I don't have to apply these by hand?
>
No, unfortunately I don't have an accessible tree... Hmm, what does it take
to get an account at kernel.bkbits.net?
--
Dmitry
On Tue, Mar 02, 2004 at 12:45:10PM -0500, Dmitry Torokhov wrote:
> On Tuesday 02 March 2004 08:02 am, Vojtech Pavlik wrote:
> > I like them very much. Do you have a bitkeeper tree anywhere where I
> > could pull from, so that I don't have to apply these by hand?
>
> No, unfortunately I don't have an accessible tree... Hmm, what does it take
> to get an account at kernel.bkbits.net?
I already sent private mail to Dmitry, but it occurs to me that other
developers are probably in the same straits.
The easiest thing to do in this case is to host your patches at
bkbits.net (hostme.bkbits.net, which is different than kernel.bkbits.net
-- kernel is a special case). You don't have to change your workflow at
all, after it's been set up; you end up just pushing to the bkbits
repository when you've got changes to submit.
To set up a hosted project, just follow the directions under
http://www.bitkeeper.com/Hosted.html (the table of contents is on the
left). You can clone Linus' tree to populate your tree initially, then
push your csets from your workstation to hostme.
If any kernel developers have difficulty doing this, drop me a line --
I'm trying to make the process easy and error-free, and I appreciate any
suggestions.
If any kernel developers are having BK workflow problems ("how do I
merge my work with other developers'? How do I keep up-to-date without
generating hundreds of merge csets?") the first thing to do is to read
the BK kernel howto, http://lwn.net/2002/0425/a/bk-thing.php3. If your
questions aren't answered there, send them to me or to the
bitkeeper-users list.
-andy