endian attribute: setting or register?

endian attribute: setting or register?

[Ben Elliston]

The current base cpu classes include an "endian" attribute -- it's
currently added as a "register" type attribute.  Should it be a
"setting" instead?  That would make more sense -- and by changing it,
we save valuable real estate in tksm's gui for the multitide of
physical cpu registers that may be listed.  Comments?

Ben

Re: endian attribute: setting or register?

[Frank Ch. Eigler]

Hi -

On Thu, May 31, 2001 at 02:49:03PM +1000, Ben Elliston wrote:
: The current base cpu classes include an "endian" attribute -- it's
: currently added as a "register" type attribute.  Should it be a
: "setting" instead?  

Yes and no: some CPUs can switch endianness on the fly.

: [...] by changing it, we save valuable real estate in
: tksm's gui for the multitide of physical cpu registers
: that may be listed.  [...]

tksm should be smarter about the layout of the attribute list windows.


- FChE
-- 
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.5 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE7FiCTVZbdDOm/ZT0RAnAfAJ9qD9Xl6Y2Fid+nAfJEl146kFyNEgCcD01V
VTopfx5jsPaNsAqRfkFI9Kk=
=sQJL
-----END PGP SIGNATURE-----

Re: endian attribute: setting or register?

[Mark Douglas]

wouldn't it make more sense just to read the configuration reg that identifies
the endianess and then behave appropriately?
Mark

"Frank Ch. Eigler" wrote:

> Hi -
>
> On Thu, May 31, 2001 at 02:49:03PM +1000, Ben Elliston wrote:
> : The current base cpu classes include an "endian" attribute -- it's
> : currently added as a "register" type attribute.  Should it be a
> : "setting" instead?
>
> Yes and no: some CPUs can switch endianness on the fly.
>
> : [...] by changing it, we save valuable real estate in
> : tksm's gui for the multitide of physical cpu registers
> : that may be listed.  [...]
>
> tksm should be smarter about the layout of the attribute list windows.
>
> - FChE
>
>   ------------------------------------------------------------------------
>    Part 1.2Type: application/pgp-signature

Re: endian attribute: setting or register?

[Frank Ch. Eigler]

Hi -

On Thu, May 31, 2001 at 08:39:28AM -0700, Mark Douglas wrote:
: wouldn't it make more sense just to read the configuration reg that
: identifies the endianess and then behave appropriately?

Yes, this is how it works for endian-switcheroo architectures.
In most cases, the desired endianness may be configured at
sid startup/reset time.  Ben's comment was only about how
to classify the endianness parameter in the tksm GUI
(regardless how it was acquired or related to hardware
control registers).

- FChE
-- 
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.5 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE7FmWPVZbdDOm/ZT0RAomNAJ9iSg8hPtFpuE0c/P8ZL7QVr4XZSQCeIpww
v+TpwlMRER2CDQYWEGY4hC0=
=QJvT
-----END PGP SIGNATURE-----

Re: endian attribute: setting or register?

[Mark Douglas]

Sorry... just trying to "stick with this" conceptually.
Mark

"Frank Ch. Eigler" wrote:

> Hi -
>
> On Thu, May 31, 2001 at 08:39:28AM -0700, Mark Douglas wrote:
> : wouldn't it make more sense just to read the configuration reg that
> : identifies the endianess and then behave appropriately?
>
> Yes, this is how it works for endian-switcheroo architectures.
> In most cases, the desired endianness may be configured at
> sid startup/reset time.  Ben's comment was only about how
> to classify the endianness parameter in the tksm GUI
> (regardless how it was acquired or related to hardware
> control registers).
>
> - FChE
>
>   ------------------------------------------------------------------------
>    Part 1.2Type: application/pgp-signature