Patchwork Add board enable for ABIT AV8

adds board enable for ABIT AV8 (first revision).

Signed-off-by: Christoph Grenz <christophg+cb at grenz-bonn.de>

+       {0x1106, 0x0282, 0x147B, 0x1415,  0x1106, 0x3227, 0x147B, 0x1415, 
"^AV8 ",      NULL, NULL,           P3, "abit",        "AV8",                   
0,   OK, board_abit_av8_enable},
        {0x8086, 0x7190,      0,      0,  0x8086, 0x7110,      0,      0, 
"^i440BX-W977 (BM6)$", NULL, NULL,  P3, "abit",        "BM6",                   
0,   OK, intel_piix4_gpo26_lower},
        {0x8086, 0x24d3, 0x147b, 0x1014,  0x8086, 0x2578, 0x147b, 0x1014, 
NULL,         NULL, NULL,           P3, "abit",        "IC7",                   
0,   NT, intel_ich_gpio23_raise},
        {0x8086, 0x2930, 0x147b, 0x1084,  0x11ab, 0x4364, 0x147b, 0x1084, 
NULL,         NULL, NULL,           P3, "abit",        "IP35",                  
0,   OK, intel_ich_gpio16_raise},

On Sat, 1 Oct 2011 08:04:44 +0200
Christoph Grenz <christophg+cb@grenz-bonn.de> wrote:

> adds board enable for ABIT AV8 (first revision).
> 
> Signed-off-by: Christoph Grenz <christophg+cb at grenz-bonn.de>
> 

hello christoph and thanks for the patch!

i guess you own such a board and have tested the board enable yourself?
could you please provide a verbose log of its action?
also we would like to see lspci -nn and lspci -xxnnvvv logs to verify
that the pci ids are well chosen.

i cant really comment the enabling function itself besides that it
maybe it would make sense to add a more general function to set/clear
GPO on that chipset? though that is probably overkill yet.

how did you find out what to do? is there a public datasheet? etc.
please tell us the story :)

btw i think we usually use the designation "rev. 1.0" (or whatever the
vendor used) instead of "first revision". did you name it yourself like
that or was it documented by abit that way?

Hello Stefan,

Am Donnerstag, 6. Oktober 2011, um 23:27:32 schrieb Stefan Tauner:
> On Sat, 1 Oct 2011 08:04:44 +0200
> 
> Christoph Grenz <christophg+cb@grenz-bonn.de> wrote:
> > adds board enable for ABIT AV8 (first revision).
> > 
> > Signed-off-by: Christoph Grenz <christophg+cb at grenz-bonn.de>
> 
> hello christoph and thanks for the patch!
> 
> i guess you own such a board and have tested the board enable yourself?
> could you please provide a verbose log of its action?
> also we would like to see lspci -nn and lspci -xxnnvvv logs to verify
> that the pci ids are well chosen.

Yes, i own it and tested the board enable.
The board's temporarily in use by a family member, but I can provide a verbose 
log in about a week.
Ok, I'll append the lspci outputs for verification when I send the verbose 
log. (btw, I chose the pci ids by the K8 host bride (00:00.0) and the SuperIO 
ISA bridge)
 
> i cant really comment the enabling function itself besides that it
> maybe it would make sense to add a more general function to set/clear
> GPO on that chipset? though that is probably overkill yet.
> 
> how did you find out what to do? is there a public datasheet? etc.
> please tell us the story :)

I disassembled the write enable and the write disable functions from the Award 
BIOS image and reconstructed C code to understand for myself what happens. 
(Only took me half a night, until this weekend I have semester break, so I had 
spare time ;-) )

The ABIT AV8 board uses both the VT8237 chipset write protection, which is 
already handled by the code in chipset_enable.c, and the write protection I 
handled in my patch, which apparently is connected to some GPIO ports. After 
reading the DSDT I found they were called GP22 und GP23 in a region called 
GPOB (0x404C-0x404F). I can't really tell which chip they are connected to and 
found no writes to the same region in another board enable function, so I 
didn't add a general function.

I attached the C code I reconstructed from the bios image to this mail.

> btw i think we usually use the designation "rev. 1.0" (or whatever the
> vendor used) instead of "first revision". did you name it yourself like
> that or was it documented by abit that way?

I named it myself like that because I couldn't find any reference to a 
revision number by abit, but as there is a board AV8-3rd-Eye which slightly 
differs from the original AV8 and as I don't own it I couldn't test if it 
works the same way, so I just wanted to add an indicator which revision I 
tested on. So I could change it to "rev. 1.0" or omit it if that's better.

Kind regards,
Christoph Grenz

Hello,

here the promised flashrom and lspci logs.

As I wrote earlier, the board enable simply sets two bits I could identify as 
GPIOs by their name in the DSDT.

All flashrom actions were done with an earlier erased rom chip.

Kind regards,
Christoph Grenz

Am Donnerstag, 6. Oktober 2011, um 23:27:32 schrieb Stefan Tauner:
> On Sat, 1 Oct 2011 08:04:44 +0200
> 
> Christoph Grenz <christophg+cb@grenz-bonn.de> wrote:
> > adds board enable for ABIT AV8 (first revision).
> > 
> > Signed-off-by: Christoph Grenz <christophg+cb at grenz-bonn.de>
> 
> hello christoph and thanks for the patch!
> 
> i guess you own such a board and have tested the board enable yourself?
> could you please provide a verbose log of its action?
> also we would like to see lspci -nn and lspci -xxnnvvv logs to verify
> that the pci ids are well chosen.
> [...]

On Thu, 13 Oct 2011 03:01:24 +0200
Christoph Grenz <christophg+cb@grenz-bonn.de> wrote:

> Hello,
> 
> here the promised flashrom and lspci logs.
> 
> As I wrote earlier, the board enable simply sets two bits I could identify as 
> GPIOs by their name in the DSDT.
> 
> All flashrom actions were done with an earlier erased rom chip.
> 

thanks, looks good!

can you please...
- drop the mentioning of "first revision" (maybe add "not AV8-3rd-Eye"
  or so to the commit log instead?),
- rename board_abit_av8_enable to board_abit_av8 (yes, there are
  similar wrong names already there, sorry :) 
- remove "board" from /* Enable board GPO ports GP22 & GP23 */ if you
  also think it is (not) appropriate,
- use tabs instead of spaces,
- resend the patch as an attachment (got mangled previously)?