Re: I/O flexpages / new mapping DB issue with overlapping ports

On Thursday 08 June 2006 09:35, Frank Mehnert wrote:
>   [Requesting I/O flexpage port=0x0000 log2_size=0x10 from Sigma0]
>
> io<f0352020>::map (fp=00004027 [00000000,16] base=00000000) to io<f0352040>
> (fp=00004020 [00000000,16])
> Received I/O ports 0000-ffff
>
>   [MR2 was 00004027 after IPC]
>
> --- "enter_kdebug" ---
> /mdb> dumpallio
> 00000000 table [objsize=256  radix=256  count=255] (f0363060)
>   00000100 [1] port: 100, space: f0358000
>    00000200 [1] port: 200, space: f0358000
>    00000300 [1] port: 300, space: f0358000
>    00000400 [1] port: 400, space: f0358000
>      ...
>    0000fd00 [1] port: fd00, space: f0358000
>    0000fe00 [1] port: fe00, space: f0358000
>    0000ff00 [1] port: ff00, space: f0358000
>
>   [PROBLEM1: although we requested the whole I/O space from Sigma0 we did
>    not receive any port below 0x100, right?]

Some more info on this issue:

- root task requests I/O fpage [port=0x0000 log2_size=4] from Sigma0
  => mI shows 
     00000100 table [objsize=8  radix=32  count=1] (f0363060)
       00000100 [1] port: 100, space: f0358000
- root task requests I/O fpage [port=0x0080 log2_size=4] from Sigma0
  => mI shows
     00000100 table [objsize=8  radix=32  count=1] (f0363060)
       00000180 [1] port: 180, space: f0358000
- root task requests I/O fpage [port=0x0100 log2_size=4] from Sigma0
  => mI shows
     00000100 table [objsize=8  radix=32  count=2] (f0363060)
       00000100 [1] port: 100, space: f0358000
         00000108 [1] port: 108, space: f0358000

Strange...

Frank
-- 
## InnoTek Systemberatung GmbH, http://www.innotek.de ##

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.3 (GNU/Linux)

iD8DBQBEiApL6z8pigLf3EcRAs2UAKCUMN9iqRAXsDDio5E2ewaanvfWlQCeOWxW
irrBEoszqXX+VThIwHJqGm4=
=0w7h
-----END PGP SIGNATURE-----