nanogui: Thread: mouse driver

I don't seem to understand the issues with
the mouse driver. If the mouse support can be
compiled into the kernel, (as I've done on my system)
why aren't the mouse events simply picked up from the
kernel?
Jeffrey

Get the facts first, you can distort them later.
Mark Twain

You are confusing the job of the kernel and the job of the userland 
applications.  The kernel exists only as a service, no more no less.  Its job 
is to actually read the PS2 port (or USB, or serial or direct connection to 
the brain stem) and turn it into a recognizable set of data that can be read 
by other programs.  The kernel provides this data in a handy set of device 
files which can be read or ignored as the case may be. 

Each individual user land program then has the responsiblity to read the 
appropriate device file and process the mouse data accordingly.  Because the 
kernel passes the data straight out, userland programs can do a wide range of 
cool stuff.  For example, consider GPM.  GPM can emulate any mouse type it 
wants by simply reading the appropriate raw mouse data, and turning it into a 
common recognizable type.  Microwindows likes this because we can turn input 
from a variety of mice into the same format (PS2), making for a very simple 
set of drivers.  

So in the Microwindows driver, we can read the mouse data at our leisure, and 
handle it however we like.   This way the kernel stays out of the data 
processing business, and we can handle the mouse the way we see fit, and not 
the way the kernel sees it.

Jordan


On Friday 29 June 2001 12:02, Jeffrey Goddard mentioned:
> I don't seem to understand the issues with
> the mouse driver. If the mouse support can be
> compiled into the kernel, (as I've done on my system)
> why aren't the mouse events simply picked up from the
> kernel?
> Jeffrey
>
> Get the facts first, you can distort them later.
> Mark Twain
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: ####@####.####
> For additional commands, e-mail: ####@####.####

OK, I'm not a device driver writer, but the way it seems to me, is that
since
the kernel is taking the connection from (in my case the brain dead
connection)
and putting the data into /dev/mouse(?), then if I am using a PS2 type
device, I shouldn't need GPM to translate, and it just adds a layer of
unnecessary software. Is this true? What am I missing here?
Jeffrey

> You are confusing the job of the kernel and the job of the userland
> applications.  The kernel exists only as a service, no more no less.  Its
job
> is to actually read the PS2 port (or USB, or serial or direct connection
to
> the brain stem) and turn it into a recognizable set of data that can be
read
> by other programs.  The kernel provides this data in a handy set of device
> files which can be read or ignored as the case may be.
> Jordan
>
>

Well, actually, I checked the GPM code, and it appears that it repeats the 
data as msc (which is a Mouse Systems compatible protocol of 5 byte packets), 
so the PS/2 is actually being translated by GPM and we are really reading the 
MSC data (my apologies for confusing that earlier).
 
So the extra level is required in this case.  If that is a hassle for people 
I don't think anyone will have a problem if somebody writes a PS/2 mouse 
driver.  That might actually save some pain and suffering with GPM (like when 
you switch back and forth from console to X11).

Jordan

On Friday 29 June 2001 13:46, MagnaSygg Technical Support mentioned:
> OK, I'm not a device driver writer, but the way it seems to me, is that
> since
> the kernel is taking the connection from (in my case the brain dead
> connection)
> and putting the data into /dev/mouse(?), then if I am using a PS2 type
> device, I shouldn't need GPM to translate, and it just adds a layer of
> unnecessary software. Is this true? What am I missing here?
> Jeffrey


> > You are confusing the job of the kernel and the job of the userland
> > applications.  The kernel exists only as a service, no more no less.  Its
>
> job
>
> > is to actually read the PS2 port (or USB, or serial or direct connection
>
> to
>
> > the brain stem) and turn it into a recognizable set of data that can be
>
> read
>
> > by other programs.  The kernel provides this data in a handy set of
> > device files which can be read or ignored as the case may be.
> > Jordan
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: ####@####.####
> For additional commands, e-mail: ####@####.####

Hi all,
In the device.h the MOUSEDEV has a function 

int     (*Read)(MWCOORD *dx,MWCOORD *dy,MWCOORD *dz,int *bp);

but the actural function is the following the the mou_touscreen

static int PD_Read(MWCOORD *px, MWCOORD *py, MWCOORD *pz, int *pb, int mode)

What's the deal here? I surprise the compiler didn't complain and I've just noticed it. 
Can someone explain this for me please. Thank you.

What am I trying to do?
Well, I'm using the tslib and it reads the screen coordinates as I intended. I'm trying to understand the mouse driver (touchscreen in this case) so I know the client is passing the correct values, as received, to the server.
 

_______________________________________________
Join Excite! - http://www.excite.com
The most personalized portal on the Web!

Anyone care to comment?
A few words is better than nothing.

Thank you.

---------------------------------------------------------------------
Date:   	Fri, 1 Jul 2005 09:52:13 -0400 (EDT)
From:   	ER ####@####.####
[ Add to Address Book | Block Address | Report as Spam ]
To:   	####@####.####
Subject:   	[nanogui] mouse driver


Hi all,
In the device.h the MOUSEDEV has a function

int (*Read)(MWCOORD *dx,MWCOORD *dy,MWCOORD *dz,int *bp);

but the actural function is the following the the mou_touscreen

static int PD_Read(MWCOORD *px, MWCOORD *py, MWCOORD *pz, int *pb, int mode)

What's the deal here? I surprise the compiler didn't complain and I've just noticed it.
Can someone explain this for me please. Thank you.

What am I trying to do?
Well, I'm using the tslib and it reads the screen coordinates as I intended. I'm trying to understand the mouse driver (touchscreen in this case) so I know the client is passing the correct values, as received, to the server.


_______________________________________________
Join Excite! - http://www.excite.com
The most personalized portal on the Web!

I think you can just remove the last parameter from the actual function.
It's not used in the function.  Other driver (mou_gpm.c) only have 4
param. 

> -----Original Message-----
> From: ER ####@####.#### 
> Sent: Friday, July 01, 2005 3:12 PM
> To: ####@####.####
> Subject: RE: [nanogui] mouse driver
> Importance: High
> 
> 
> Anyone care to comment?
> 
> A few words is better than nothing.
> 
> 
> 
> Thank you.
> 
> 
> 
> ---------------------------------------------------------------------
> 
> Date:   	Fri, 1 Jul 2005 09:52:13 -0400 (EDT)
> 
> From:   	ER ####@####.####
> 
> [ Add to Address Book | Block Address | Report as Spam ]
> 
> To:   	####@####.####
> 
> Subject:   	[nanogui] mouse driver
> 
> 
> 
> 
> 
> Hi all,
> 
> In the device.h the MOUSEDEV has a function
> 
> 
> 
> int (*Read)(MWCOORD *dx,MWCOORD *dy,MWCOORD *dz,int *bp);
> 
> 
> 
> but the actural function is the following the the mou_touscreen
> 
> 
> 
> static int PD_Read(MWCOORD *px, MWCOORD *py, MWCOORD *pz, int 
> *pb, int mode)
> 
> 
> 
> What's the deal here? I surprise the compiler didn't complain 
> and I've just noticed it.
> 
> Can someone explain this for me please. Thank you.
> 
> 
> 
> What am I trying to do?
> 
> Well, I'm using the tslib and it reads the screen coordinates 
> as I intended. I'm trying to understand the mouse driver 
> (touchscreen in this case) so I know the client is passing 
> the correct values, as received, to the server.
> 
> 
> 
> _______________________________________________
> Join Excite! - http://www.excite.com
> The most personalized portal on the Web!
> 
> 
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: ####@####.####
> For additional commands, e-mail: ####@####.####
> 

This is actually already fixed in the current CVS of microwin.

* Van Nguyen (vannguye) [Fri,  1 Jul 2005 at 15:27 -0400]
<quote>
> I think you can just remove the last parameter from the actual function.
> It's not used in the function.  Other driver (mou_gpm.c) only have 4
> param. 
> 
> > -----Original Message-----
> > From: ER ####@####.#### 
> > Sent: Friday, July 01, 2005 3:12 PM
> > To: ####@####.####
> > Subject: RE: [nanogui] mouse driver
> > Importance: High
> > 
> > 
> > Anyone care to comment?
> > 
> > A few words is better than nothing.
> > 
> > 
> > 
> > Thank you.
> > 
> > 
> > 
> > ---------------------------------------------------------------------
> > 
> > Date:   	Fri, 1 Jul 2005 09:52:13 -0400 (EDT)
> > 
> > From:   	ER ####@####.####
> > 
> > [ Add to Address Book | Block Address | Report as Spam ]
> > 
> > To:   	####@####.####
> > 
> > Subject:   	[nanogui] mouse driver
> > 
> > 
> > 
> > 
> > 
> > Hi all,
> > 
> > In the device.h the MOUSEDEV has a function
> > 
> > 
> > 
> > int (*Read)(MWCOORD *dx,MWCOORD *dy,MWCOORD *dz,int *bp);
> > 
> > 
> > 
> > but the actural function is the following the the mou_touscreen
> > 
> > 
> > 
> > static int PD_Read(MWCOORD *px, MWCOORD *py, MWCOORD *pz, int 
> > *pb, int mode)
> > 
> > 
> > 
> > What's the deal here? I surprise the compiler didn't complain 
> > and I've just noticed it.
> > 
> > Can someone explain this for me please. Thank you.
> > 
> > 
> > 
> > What am I trying to do?
> > 
> > Well, I'm using the tslib and it reads the screen coordinates 
> > as I intended. I'm trying to understand the mouse driver 
> > (touchscreen in this case) so I know the client is passing 
> > the correct values, as received, to the server.
> > 
> > 
> > 
> > _______________________________________________
> > Join Excite! - http://www.excite.com
> > The most personalized portal on the Web!
> > 
> > 
> > 
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: ####@####.####
> > For additional commands, e-mail: ####@####.####
> > 
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: ####@####.####
> For additional commands, e-mail: ####@####.####
> 
</quote>

	

	
		
___________________________________________________________________________ 
Appel audio GRATUIT partout dans le monde avec le nouveau Yahoo! Messenger 
Téléchargez cette version sur http://fr.messenger.yahoo.com