i was wrong, i misread the first question, and he told me, GMP is in
userspace. by kernel "handling of the SIMD regs" he meant the switch from
one userspace task to another, i think.
what do you mean by "in a limited context"? Can you use SIMD in kernel
code regs if you avoid preemtion?
joerg
On
Adam Done wrote:
Benjamin Herrenschmidt wrote:
I have no idea what's up at this point... best would be to find somebody
in Canberra with one of these machines that would let me use it to debug
the problem That or somebody willing to hack into the code and try
to find a way to
Benjamin Herrenschmidt wrote:
> On Thu, 2006-02-23 at 17:19 -0800, Adam Done wrote:
>
>>Since last summer/June I have not been able to use current kernels
>>because I have been experiencing regression with the display on my
>>PowerBook TI Radeon Mobility (M6). The last working kernel I can switc
On Thu, 2006-02-23 at 17:19 -0800, Adam Done wrote:
> Since last summer/June I have not been able to use current kernels
> because I have been experiencing regression with the display on my
> PowerBook TI Radeon Mobility (M6). The last working kernel I can switch
> from X (GDM and session) to con
Since last summer/June I have not been able to use current kernels
because I have been experiencing regression with the display on my
PowerBook TI Radeon Mobility (M6). The last working kernel I can switch
from X (GDM and session) to console and back without any problems was
kernel 2.6.11.9. I h
On Thu, 2006-02-23 at 17:17 +0100, Joerg Maier wrote:
> Hi,
>
> you cannot use SIMD registers in kernel as far as i know. The kernel does
> not save these registers when changing userspace/kernel mode (i think for
> performance reason) so these registers have to stay untouched when you are
> in ke
On Thu, 2006-02-23 at 15:42 +0100, Torbjorn Granlund wrote:
> I recently committed several GMP assembly functions for powerpc32
> using Altivec registers and instructions.
>
> I also reinstalled my system (a Pegasos2) recently, from an obsolete
> snapshot I got from somebody on the net, to a much
On Thu, 2006-02-23 at 23:56 +0100, Jan-David Salchow wrote:
> Benjamin Herrenschmidt <[EMAIL PROTECTED]> writes:
>
> > There are several ways.
>
> > The first issue you'll encounter is the DHCP problem. Apple firmware
> > will happily DHCP to obtain an IP address and TFTP down a file if you
> > t
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
>>>Also, have you tried playing with
>>>Option "DisplayPriority" in section "Device" (value can be "AUTO" "BIOS"
>>>or "HIGH").
>
>
> No, I will try that tonight and inform you about the results
>
playing with the above mentioned option didn't hel
Benjamin Herrenschmidt <[EMAIL PROTECTED]> writes:
> There are several ways.
> The first issue you'll encounter is the DHCP problem. Apple firmware
> will happily DHCP to obtain an IP address and TFTP down a file if you
> tell it to do so explicitely (providing the server address and file name
>
Hi,
you cannot use SIMD registers in kernel as far as i know. The kernel does
not save these registers when changing userspace/kernel mode (i think for
performance reason) so these registers have to stay untouched when you are
in kernel mode.
Regards, Joerg.
On Do, 23.02.2006, 15:42, Torbjorn G
I recently committed several GMP assembly functions for powerpc32
using Altivec registers and instructions.
I also reinstalled my system (a Pegasos2) recently, from an obsolete
snapshot I got from somebody on the net, to a much less obsolete
system.
With the new assembly code *and* with the new G
Hi!
* Jason Pope <[EMAIL PROTECTED]> [060222 20:32]:
> I am having the same problem. Did you ever resolve your issue? I know it
> was 3 years ago, but it is worth a try.
When answering to that old mails without setting proper reference or
In-Reply-To-Header to the mail, it might be a good idea to
> before moussemu starts (and after it has been terminated) "ls /dev/input"
> shows event4, event5, mice, mouse1 and uinput
That's odd - why does the event device list start with event4? Which
devices do event4 and event5 correspond to (see /sys/class/input)?
>
> with "mouseemu -right 29 0 -devic
On Wed, Feb 22, 2006 at 06:59:34PM +0100, Andrzej Mendel wrote:
> Dnia 21-02-2006, wto o godzinie 19:04 +0100, Brian Durant napisa??(a):
> > Hi again,
> >
> > I installed Debian (Etch) PPC (I think this is considered to be
> > "testing"). At the end of the install, I got the following message:
>
On Thu, Feb 23, 2006 at 11:04 AM, Marco Giusti wrote:
> is there an alternative for UML? i must test the comunication between
> three computers using ethereal, iptables, ecc. in x86 i lanch three
> instances of uml, but in ppc?
Try qemu.
It seems you can also use MOL to get LOL (linux on linux),
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Benjamin Herrenschmidt wrote:
>
>
> What about the X log ?
sorry, forgot that one, is now here:
http://www.informatik.uni-freiburg.de/~lesch/Xorg.log
> Also, have you tried playing with
> Option "DisplayPriority" in section "Device" (value can b
On Sat, Feb 18, 2006 at 11:54:15PM +0100, Stefano Melchior wrote:
>On Sat, Feb 18, 2006 at 08:58:27PM +0100, Luca Bigliardi wrote:
>Hi Dieter
>> On Sat, Feb 18, 2006 at 07:53 PM, Dieter Schuster wrote:
>>
>> > does anyone of you run user mode linux on PPC?
>>
>> it doesn't work
is there an alte
18 matches
Mail list logo
