On Sat, Sep 30, 2006 at 07:51:22AM -0700, Mike Emmel wrote: > On 9/30/06, Ville Syrjälä <[EMAIL PROTECTED]> wrote: > > On Fri, Sep 29, 2006 at 09:21:27PM +0200, Attilio Fiandrotti wrote: > > > Hi > > > > > > I just completed a preliminary survey [1] about graphical > > > debian-installer usability on PPC machines and it turned out that there > > > are two distinct kinds of issues that currently block PPC users from > > > using the g-i. > > > > > > *Graphical issues: PPC testers always report crashes when hw > > > acceleration iss enabled, but luckily such crashes can be easily avoided > > > by disabling hardware acceleration and chip-specific modules, or better > > > by removing gfxdrivers them from the filesystem. > > > So, even if not optimal, disabling hw acceleration does the trick and we > > > can live well. > > > > I'm curious about the mach64 problems at least. But I need someone to > > tell me more than 'it crashes'. The driver should work (mmio byte swap > > asm code is included) but I have never tested it because I don't have > > the hardware. > > > > If someone wants to run tests, start with running dfbinfo, and if that > > works run df_dok. > > > > > *Input devices issues: In many cases it turned out that the key to get > > > DFB running on PPCs is disabling the linux_input module. > > > Whenever a success was reported (except for PowerBook6,7 IBook's), it > > > was because the linux_input module was disabled. > > > We also noticed that many i386 laptop touchpads do not to work properly > > > when the linux_input module is enabled, and hence the user has to plug > > > in an USB mouse to use the installer. > > > > linux_input is crap is the touchpad is in absolute mode. Fortunately all > > my laptops have synaptics touchpads which supports relative mode in > > hardware/firmware. I have though about fixing the problem a couple of > > times but since my hardware works I haven't had enough motivation :) I > > was thinking about using tslib for this. > > > I actually dug into it myself I question even supporting linux-input by > default I'd suggest the fix here is to turn it off in the default build until > the underlying drivers become more reasonable. > I was not impressed in the least with the work and it seems that little > or no ongoing work is happening if you visit the sourceforge site. > > If you look at say commit messages > > http://sourceforge.net/mailarchive/forum.php?forum=linuxconsole-commit > > You see a dying project so I think we should simply not build them by default > and wait till either the project restarts and finishes or a new one solves the > problems it was trying to solve. At the moment I consider most of the work > far from finished.
That site was just some initial project which is no longer relevant. The code is in the kernel and all kernel input drivers use the same input framework. -- Ville Syrjälä [EMAIL PROTECTED] http://www.sci.fi/~syrjala/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
