On Thu, Apr 26, 2007 at 06:24:06PM +0200, Mondrian Nuessle wrote:
The attached patch enables flashing on the Iwill DK8-HTX board. Basically, it configures the SuperIO to set the right GPIO pins, so write protection is disabled.
Signed-off-by: Mondrian Nuessle nuessle@uni-mannheim.de
So, this is the second try :-) Only pci vendor id/device id is given, no subsystem id. Therefor, flashing when booted from factory BIOS will only work, if -m iwill:dk8_htx switch is given. I guess this could go in the wiki doc to this board...
And in the flashrom README, yes.
Index: util/flashrom/board_enable.c
--- util/flashrom/board_enable.c (revision 2618) +++ util/flashrom/board_enable.c (working copy) @@ -23,6 +23,56 @@ #include "flash.h" #include "debug.h"
+static int board_iwill_dk8htx(char *name)
I think this can be const char *name as you never modify 'name' (or use it at all).
+{
- /* Extended function index register, either 0x2e or 0x4e */
On what does this depend? Can it be detected at runtime? If so, we should "probe" the value and handle both cases.
+#define EFIR 0x2e
- /* Extended function data register, one plus the index reg. */
+#define EFDR EFIR + 1
- char b;
/* Disable the flash write protect. The flash write protect is* connected to the WinBond w83627hf GPIOs*/- outb(0x87, EFIR); /* sequence to unlock extended functions */
- outb(0x87, EFIR);
- /* activate logical devie */
device
- outb(0x7,EFIR);
- outb(8,EFDR);
Space after comma (and other stuff Luc mentioned in his mail). Please use the Linux kernel coding style and/or run your code through indent, see http://linuxbios.org/Development_Guidelines#Coding_Style.
(yes, I know, lots of code currently in LinuxBIOS does not yet follow those guidelines, but we're slowly moving towards that direction)
- /*lock extended function sagain */
Typo.
Otherwise the patch looks good, but I cannot test it on real hardware, of course.
Uwe.