Re: [coreboot] Patch to superiotool - add basic indexing of devices detected at IO ports

[Frantisek Rysanek]

Dear gentlemen,

sorry for a noob question... noone has responded to my suggested 
patch - is there something more I can do?
I can see on the mailing list that the admins are busy doing other 
things, and I understand they are volunteers, so I'd hate to sound 
abusive :-)

I guess the patch may look long, but it's mostly the same one-liners 
over and over in all the vendor-specific files. And it's so simple 
conceptually, that it shouldn't cause any trouble.
I've spent an enjoyable evening coding the patch and I'd like to have 
it merged, for the benefit of others...

Thanks for your time and attention :-)

Frank Rysanek

-----------
On 10 Jan 2011 at 8:49, [e-mail] wrote:

This patch adds a simple table of SuperIO chips already detected in 
the system. This allows superiotool to properly avoid scanning the 
same IO port again by another vendor-specific probe, after some 
device has already been found on that particular port.

The table of chips detected contains entries that, apart from the IO 
port known to be "inhabited", also relate the port to a particular 
"chip entry" (struct superio_registers*).
There's no immediate use for this, but in the future it could 
facilitate further functionality enhancements, such as
- separation of the "dump" function from the "detect" function
- write support (with validity checking of the desired LDN/CR)

As suggested by Mr. Hendricks, the vendor-specific probe functions 
now return a result code, indicating that a particular chip was found
and which one that was, or that none was detected.

The simple "framework" is headquartered in superiotool.c, with some 
necessary mods to superiotool.h. All the existing "vendor-specific" C

files have been adapted to make use of this "framework".

Apart from this addition, the patch scraps some previous attempts at 
the same functionality, and moves some table definitions from 
superiotool.h to superiotool.c, where they seem to belong.

Signed-off-by: Frank Rysanek 
---

Tested on my AEC-6811 with two Winbond SuperIO chips.
Note that the Winbond probe comes last in the "sequence of vendors", 
so my test is not entirely significant :-)

I've tried the recommended method of deriving the patch via 
"svn diff". The resulting patch however fails to apply cleanly (hunk 
failed...).
But if I try the bare "diff -Naur", the resulting patch *does* apply 
just fine. Go figure... 
Attached you'll find the manual diff that works for me.

The following section of this message contains a file attachment
prepared for transmission using the Internet MIME message format.
If you are using Pegasus Mail, or any other MIME-compliant system,
you should be able to save it or view it from within your mailer.
If you cannot, please ask your system administrator for assistance.

---- File information -----------
     File:  device_counting.diff
     Date:  9 Jan 2011, 22:25
     Size:  20401 bytes.
     Type:  Unknown