insmod problem

insmod problem

[Marco Schramel]

Hi,

we use a custom mpc8270 system. Everything works fine.
Now i have to handle some float values in a device driver. If i insert the compiled module my target prints out 

insmod: unresolved symbol __addsf3
insmod: unresolved symbol __floatsisf

Are there any missing libraries ?

Any other ideas?

Best regards

Marco


-- 
---------
Marco Schramel

Re: insmod problem

[Pantelis Antoniou]

Marco Schramel wrote:
> Hi,
> 
> we use a custom mpc8270 system. Everything works fine.
> Now i have to handle some float values in a device driver. If i insert the compiled module my target prints out 
> 
> insmod: unresolved symbol __addsf3
> insmod: unresolved symbol __floatsisf
> 
> Are there any missing libraries ?
> 
> Any other ideas?
> 
> Best regards
> 
> Marco
> 
> 

You're using floating arithmetic in the kernel.

That's not supported.

Regards

Pantelis

Re: insmod problem

[Mark Chambers]

>
>
> You're using floating arithmetic in the kernel.
>
> That's not supported.
>
> Regards
>
> Pantelis
>

Just for my edification:  Is floating point itself not supported (floating
point register
corruption or something like that) or just linking to user space libraries?

Mark Chambers

Re: insmod problem

[Kumar Gala]

On May 13, 2005, at 8:14 AM, Mark Chambers wrote:

> >
>  >
>  > You're using floating arithmetic in the kernel.
>  >
>  > That's not supported.
>  >
>  > Regards
>  >
>  > Pantelis
>  >
>
> Just for my edification:=A0 Is floating point itself not supported=20
> (floating
>  point register
>  corruption or something like that) or just linking to user space=20
> libraries?

Use of floating point registers and instructions.  Its possible to use=20=

FP in rare cases inside the kernel buts its extremely discouraged.

- kumar

MPC885 - USB HCI drivers.

[Guillaume Autran]

Hi,

Is there any driver for the USB host controller interface built-in the 
MPC885 processors.
Does anyone have any experience with it ?

Thanks.
Guillaume.

-- 
=======================================
Guillaume Autran
Senior Software Engineer
MRV Communications, Inc.
Tel: (978) 952-4932 office
E-mail: gautran@mrv.com
======================================= 

Re: MPC885 - USB HCI drivers.

[Kylo Ginsberg]

Guillaume,

There is a driver in the linux-2.4 tree at denx.de, although I have no
experience with it.  Here's info on the cvs repository:=20
http://www.denx.de/e/solution-se.php.  Look in drivers/usb for
m8xxhci.[ch]  This code is not in the 2.4/2.6 trees on kernel.org, not
sure if patches have been submitted.  I don't know the history here.

I will shortly be in need of a 2.6-based driver for the mpc8555e, but
haven't found any patches.  I'll probably look to port this driver, so
if you need to modify it, please post patches.

Cheers,
Kylo


On 5/13/05, Guillaume Autran <gautran@mrv.com> wrote:
> Hi,
>=20
> Is there any driver for the USB host controller interface built-in the
> MPC885 processors.
> Does anyone have any experience with it ?
>=20
> Thanks.
> Guillaume.
>=20
> --
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> Guillaume Autran
> Senior Software Engineer
> MRV Communications, Inc.
> Tel: (978) 952-4932 office
> E-mail: gautran@mrv.com
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
>=20
> _______________________________________________
> Linuxppc-embedded mailing list
> Linuxppc-embedded@ozlabs.org
> https://ozlabs.org/mailman/listinfo/linuxppc-embedded
>

Re: MPC885 - USB HCI drivers.

[Wolfgang Denk]

Dear Kylo,

in message <61cc712d05051614507b688c6e@mail.gmail.com> you wrote:
> 
> There is a driver in the linux-2.4 tree at denx.de, although I have no

No, there is not!

> experience with it.  Here's info on the cvs repository: 
> http://www.denx.de/e/solution-se.php.  Look in drivers/usb for
> m8xxhci.[ch]  This code is not in the 2.4/2.6 trees on kernel.org, not
> sure if patches have been submitted.  I don't know the history here.

This code will NOT work on a MPC885.  It  was  written  for  the  old
MPC823/850  processors,  and it NEVER worked really reliably on these
systems either.

My advice: use a _real_ USB controller.

Best regards,

Wolfgang Denk

-- 
Software Engineering:  Embedded and Realtime Systems,  Embedded Linux
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd@denx.de
For those who like this sort of thing, this is the sort of thing they
like.                                               - Abraham Lincoln

Re: MPC885 - USB HCI drivers.

[Kylo Ginsberg]

On 5/16/05, Wolfgang Denk <wd@denx.de> wrote:
> This code will NOT work on a MPC885.  It  was  written  for  the  old
> MPC823/850  processors,  and it NEVER worked really reliably on these
> systems either.
=20
Ok, thanks.  Good to know!  Do you know of a driver for the 885, or
for the 8555E?

> My advice: use a _real_ USB controller.

Are there known problems with CPM-based USB controllers?  I'm
completely new to the subject, but we are designing a board assuming
we can use the USB controller on an MPC8555E.

Thanks for any comments.  If there are gotchas here, I'd love to find
out about them sooner rather than later.

Cheers,
Kylo

Re: MPC885 - USB HCI drivers.

[Dan Malek]

On May 16, 2005, at 6:39 PM, Kylo Ginsberg wrote:

> Are there known problems with CPM-based USB controllers?

The original CPM USB controller was only supposed to be
a device side controller.  With some software hacks, CPM
microcode downloads, and even some hardware modifications,
you can often get it to work as a host in some static situations.
It doesn't have a built in root hub, you usually concentrate
on making it work with one specific device for a particular
product.  You may find some application notes on the
Freescale web site, but I haven't seem them lately.

Take Wolfgang's advice, it's easier to find a way to
attach a real USB controller than to make this one
work as you probably expect.

Thanks.

	-- Dan

RE: MPC885 - USB HCI drivers.

[Jonathan Masel]

CPM USB controllers have their quirks and erratas, but they do work. Most
erratas are documented but I'm not sure all are - we identified a couple and
reported them back to Freescale, but I don't know if they've updated their
web site in this regard.

I'm not sure what "software hacks" means but they don't require CPM
downloads or hardware modifications to work. It is also not true that they
can only support one device at a time (though yes, you do need an external
hub).

Bottom line: it's not easy getting them to work, but I can assure you that
there are software packages that do so (we have one). They're not free - but
neither are the USB controllers. Your choice!

Jonathan Masel


On May 16, 2005, at 6:39 PM, Kylo Ginsberg wrote:

> Are there known problems with CPM-based USB controllers?

The original CPM USB controller was only supposed to be
a device side controller.  With some software hacks, CPM
microcode downloads, and even some hardware modifications,
you can often get it to work as a host in some static situations.
It doesn't have a built in root hub, you usually concentrate
on making it work with one specific device for a particular
product.  You may find some application notes on the
Freescale web site, but I haven't seem them lately.

Take Wolfgang's advice, it's easier to find a way to
attach a real USB controller than to make this one
work as you probably expect.

Thanks.

	-- Dan

Re: MPC885 - USB HCI drivers.

[Wolfgang Denk]

In message <61cc712d050516153941bc3b26@mail.gmail.com> you wrote:
>
> > My advice: use a _real_ USB controller.
> 
> Are there known problems with CPM-based USB controllers?  I'm

Ummmm... yes, there are. Especially if you want to use it as  a  host
comtroller  (which  is  never  was designed for). Even more, if there
might be other periopherals in your system that might need  a  bit  a
CPM performance. Or ...

No, I better stop here. Please ask your  Freecale  sales  representa-
tive. He may know better. [Or not, maybe.]

> completely new to the subject, but we are designing a board assuming
> we can use the USB controller on an MPC8555E.

You can do that. Anything can be done. But in this case I don't  know
of an efficient way to do it. But my knowledge is limited, obviously.

Best regards,

Wolfgang Denk

-- 
Software Engineering:  Embedded and Realtime Systems,  Embedded Linux
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd@denx.de
Man did not weave the web of life; he  is  merely  a  strand  in  it.
Whatever he does to the web, he does to himself.     - Seattle [1854]

Re: MPC885 - USB HCI drivers.

[Mike Rapoport]

Wolfgang Denk wrote:

>In message <61cc712d050516153941bc3b26@mail.gmail.com> you wrote:
>  
>
>>>My advice: use a _real_ USB controller.
>>>      
>>>
>>Are there known problems with CPM-based USB controllers?  I'm
>>    
>>
>
>Ummmm... yes, there are. Especially if you want to use it as  a  host
>comtroller  (which  is  never  was designed for). Even more, if there
>might be other periopherals in your system that might need  a  bit  a
>CPM performance. Or ...
>
>  
>
Do you know if the same is applicable to CPM2 USB controllers present 
for example on mpc8247?

>No, I better stop here. Please ask your  Freecale  sales  representa-
>tive. He may know better. [Or not, maybe.]
>  
>

The Freescale sales representative we work with actually says that CPM2 
USB host controller is working well.

>Best regards,
>
>Wolfgang Denk
>  
>
-- 
Sincerely yours,
Mike

Re: MPC885 - USB HCI drivers.

[Wolfgang Denk]

Dear Mike,

in message <428995CB.4050301@compulab.co.il> you wrote:
> 
> Do you know if the same is applicable to CPM2 USB controllers present 
> for example on mpc8247?

Sorry, I don't know from personal experience. But from what  I  heard
there might be some "surprises", too.

> The Freescale sales representative we work with actually says that CPM2 
> USB host controller is working well.

Of course they do. They say the same for the 850 and 823 :-(

Best regards,

Wolfgang Denk

-- 
Software Engineering:  Embedded and Realtime Systems,  Embedded Linux
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd@denx.de
I haven't lost my mind -- it's backed up on tape somewhere.

Re: MPC885 - USB HCI drivers.

[Pantelis Antoniou]

Wolfgang Denk wrote:
> Dear Mike,
> 
> in message <428995CB.4050301@compulab.co.il> you wrote:
> 
>>Do you know if the same is applicable to CPM2 USB controllers present 
>>for example on mpc8247?
> 
> 
> Sorry, I don't know from personal experience. But from what  I  heard
> there might be some "surprises", too.
> 
> 
>>The Freescale sales representative we work with actually says that CPM2 
>>USB host controller is working well.
> 
> 
> Of course they do. They say the same for the 850 and 823 :-(
> 
> Best regards,
> 
> Wolfgang Denk
> 

Well I have USB host drivers for both 8xx & 82xx working.

It was not a piece of cake exactly to get them to work
but they seem to work fine even under heavy traffic.

However I use them connected to a specific peripheral
so I don't know how well they fare when used as a PC
style USB host controller.

YMMV

Regards

Pantelis

Re: MPC885 - USB HCI drivers.

[Bryan O'Donoghue]

Pantelis Antoniou wrote:

> Well I have USB host drivers for both 8xx & 82xx working.

Speaking of which... would it not be a good idea, to get these comitted
to the 2.6 tree... at some stage ... at least to stop people
periodically posting to this list... saying "Dear all have spent 2
months, writing code for m8xx_hci drivers", when it's a needless
replication of effort ?

> However I use them connected to a specific peripheral
> so I don't know how well they fare when used as a PC
> style USB host controller.

Also, one easy way to find out, how well or badly said code performs, in
hetrogenous environments, is to... suck it and see.

As I was attempting to intimate above, I'm sure there'd be a legion of
eager people to debug, modify and recode such drivers, if there lived in
an obvious place... in the 2.6 tree.

Just my €0.02.

--

Best,
Bryan

Re: MPC885 - USB HCI drivers.

[Pantelis Antoniou]

Bryan O'Donoghue wrote:
> Pantelis Antoniou wrote:
> 
> 
>>Well I have USB host drivers for both 8xx & 82xx working.
> 
> 
> Speaking of which... would it not be a good idea, to get these comitted
> to the 2.6 tree... at some stage ... at least to stop people
> periodically posting to this list... saying "Dear all have spent 2
> months, writing code for m8xx_hci drivers", when it's a needless
> replication of effort ?
> 

Well, they are too hideous for human eyes :)

> 
>>However I use them connected to a specific peripheral
>>so I don't know how well they fare when used as a PC
>>style USB host controller.
> 
> 
> Also, one easy way to find out, how well or badly said code performs, in
> hetrogenous environments, is to... suck it and see.
> 
> As I was attempting to intimate above, I'm sure there'd be a legion of
> eager people to debug, modify and recode such drivers, if there lived in
> an obvious place... in the 2.6 tree.
> 
> Just my €0.02.
> 

Also they're 2.4 only :)

I'll try to clean them up and post them sometime this week.

But I don't have time left to hack them for 2.6. I'll need
vict^H^H^Holunteers to do it...

Regards

Pantelis


> --
> 
> Best,
> Bryan
> 
> 

Re: MPC885 - USB HCI drivers.

[Mike Rapoport]

Pantelis Antoniou wrote:

> Also they're 2.4 only :)
>
> I'll try to clean them up and post them sometime this week.
>
> But I don't have time left to hack them for 2.6. I'll need
> vict^H^H^Holunteers to do it...
>
> Regards
>
> Pantelis
>
I think you've found one _vict^H^H^Holunteer_  :)
I've started to port the m82xxhci to 2.6. For now it is get compiled 
(mostly). :))

-- 
Sincerely yours,
Mike

Re: MPC885 - USB HCI drivers.

[Guillaume Autran]

Mike Rapoport wrote:

> Pantelis Antoniou wrote:
>
>> Also they're 2.4 only :)
>>
>> I'll try to clean them up and post them sometime this week.
>>
>> But I don't have time left to hack them for 2.6. I'll need
>> vict^H^H^Holunteers to do it...
>>
>> Regards
>>
>> Pantelis
>>
> I think you've found one _vict^H^H^Holunteer_  :)
> I've started to port the m82xxhci to 2.6. For now it is get compiled 
> (mostly). :))
>
Add me to the list !

-- 
=======================================
Guillaume Autran
Senior Software Engineer
MRV Communications, Inc.
Tel: (978) 952-4932 office
E-mail: gautran@mrv.com
======================================= 

linux/drivers/net/fec_8xx/* vs. linux/arch/ppc/8xx_io/fec.c

[Guillaume Autran]

Can anyone tell me what is the difference between:
    linux/arch/ppc/8xx_io/fec.c
and
    linux/drivers/net/fec_8xx/fec_*

They both seem to be mpc8xx fec drivers....

Thanks.
Guillaume.

-- 
=======================================
Guillaume Autran
Senior Software Engineer
MRV Communications, Inc.
Tel: (978) 952-4932 office
E-mail: gautran@mrv.com
======================================= 

Re: linux/drivers/net/fec_8xx/* vs. linux/arch/ppc/8xx_io/fec.c

[Pantelis Antoniou]

Guillaume Autran wrote:
> 
> Can anyone tell me what is the difference between:
>    linux/arch/ppc/8xx_io/fec.c

Old driver (AFAIK doesn't work on 2.6)

> and
>    linux/drivers/net/fec_8xx/fec_*

The new driver for 2.6 (for now).

> 
> They both seem to be mpc8xx fec drivers....
> 
> Thanks.
> Guillaume.
> 

Regards

Pantelis

Re: MPC885 - USB HCI drivers.

[Guillaume Autran]

Hi Pantelis,

Are your drivers compatible with linux kernel 2.6.x ?
I'll be interested to have a look at them if you don't mind.

Also, do you think an USB bus analyzer is a *must have* in order to work 
on the drivers ?

Regards,
Guillaume.


Pantelis Antoniou wrote:

> Wolfgang Denk wrote:
>
>> Dear Mike,
>>
>> in message <428995CB.4050301@compulab.co.il> you wrote:
>>
>>> Do you know if the same is applicable to CPM2 USB controllers 
>>> present for example on mpc8247?
>>
>>
>>
>> Sorry, I don't know from personal experience. But from what  I  heard
>> there might be some "surprises", too.
>>
>>
>>> The Freescale sales representative we work with actually says that 
>>> CPM2 USB host controller is working well.
>>
>>
>>
>> Of course they do. They say the same for the 850 and 823 :-(
>>
>> Best regards,
>>
>> Wolfgang Denk
>>
>
> Well I have USB host drivers for both 8xx & 82xx working.
>
> It was not a piece of cake exactly to get them to work
> but they seem to work fine even under heavy traffic.
>
> However I use them connected to a specific peripheral
> so I don't know how well they fare when used as a PC
> style USB host controller.
>
> YMMV
>
> Regards
>
> Pantelis
>
> _______________________________________________
> Linuxppc-embedded mailing list
> Linuxppc-embedded@ozlabs.org
> https://ozlabs.org/mailman/listinfo/linuxppc-embedded
>

-- 
=======================================
Guillaume Autran
Senior Software Engineer
MRV Communications, Inc.
Tel: (978) 952-4932 office
E-mail: gautran@mrv.com
======================================= 

Re: MPC885 - USB HCI drivers.

[Kylo Ginsberg]

On 5/17/05, Wolfgang Denk <wd@denx.de> wrote:
> Dear Mike,
>=20
> in message <428995CB.4050301@compulab.co.il> you wrote:
> >
> > Do you know if the same is applicable to CPM2 USB controllers present
> > for example on mpc8247?
>=20
> Sorry, I don't know from personal experience. But from what  I  heard
> there might be some "surprises", too.

Thanks for all the feedback.  If we set our sights lower and only aim
to use the CPM USB as a device (specifically an alternative serial
console), similar warnings of "surprises"?  Any source out there for
this purpose?  (I'm inclined to just use a 232-usb converter off a
uart and make it transparent to the target s/w.)

Also, I failed to find app notes or usb-specific errata on the
freescale site.  Perhaps they're tucked away under a specific
processor without a suggestive title?

Kylo

floating point in kernel (Re: insmod problem)

[Roger Larsson]

On Friday 13 May 2005 15.46, Kumar Gala wrote:
> On May 13, 2005, at 8:14 AM, Mark Chambers wrote:
> >  > You're using floating arithmetic in the kernel.
> >  >
> >  > That's not supported.
> >  >
> >  > Regards
> >  >
> >  > Pantelis
> >
> > Just for my edification:=A0 Is floating point itself not supported
> > (floating
> >  point register
> >  corruption or something like that) or just linking to user space
> > libraries?
>
> Use of floating point registers and instructions.  Its possible to use
> FP in rare cases inside the kernel buts its extremely discouraged.
>

The kernel is doing lazy saving of FP registers, only saving for user=20
processes that uses them.  And especially avoiding to save for every
interrupt!

So if you like to use any FP (registers) in kernel you have to:
=2D disable kernel preemption (YES)
=2D save previous value
=2D do your stuff
=2D restore
=2D enable preemption

I remember a preemption bug caused by using MMX/SSE to copy
memory uncached.

/RogerL