Re: [LinuxBIOS] [PATCH] Add Intel i810, i82801aa (ICH) and Asus MEW-VM support

[Uwe Hermann]

On Fri, May 25, 2007 at 05:47:51AM -0400, Corey Osgood wrote:
> See attached patch. There's still some work to be done on this, but my
> laptop's out of commission for a week or so, so I figure I'll let anyone
> who's bored do some hacking. As is, it does boot a payload, and probably
> would do a kernel with either acpi or irq tables, but should be
> considered a WIP. Feedback and comments are appreciated!

Nice! This looks very good. I'd like to make some changes before
we commit, though.
> Also, the i82801aa is almost completely copy and pasted from the
> i82801ca, with the device IDs swapped over (bad practice, I know, but if
> ain't broke don't fix it).

This is change number 1 ;-)

The code is indeed fully copy+paste, which we should try to avoid. Can
you change it to modify the i82801ca with #ifdef's to support the
i82801aa, too? In the mainboard code you'd just have to

#define I82801AA 1

or something similar...
> I haven't updated any of the license headers
> to be GPLv2 ones, mainly because I don't want to make a bad assumption.
> Not quite sure what to do with them...

Maybe you can modify the i82801er code? That was originally named ich5_r
and written by Eric and/or Yinghai so we can assume it's GPLv2 (all of
Eric's code is).

If the i82801er is too different, please modify the i82801ca and add the
proper license headers. The i82801ca is based on the i82801er code, and
was written by svn user "magnanisj" (not sure who that is).
> Index: src/northbridge/intel/i810/Config.lb

Poll: do we want i810 or rather i82810? For the southbridges we use the
full names (e.g. i82801aa), why don't we do the same with the
northbridges?

(yes, that would also change i440bx to i82443bx)
> +#include "i810.h"
>
> +/*-----------------------------------------------------------------------------
> +Macros and definitions.
> +-----------------------------------------------------------------------------*/
>
> +/* Uncomment this to enable debugging output. */
> +#define DEBUG_RAM_SETUP 1

Btw, disabling this reduces the number of registers required for romcc,
thus you can get further in the compile...
> for(i = 0; i < DIMM_SOCKETS; i++)

^
> {
> /* First check if a DIMM is actually present */
> if(smbus_read_byte(ctrl->channel0[i], 2) == 4)

^
               space needed

(coding guidelines, http://linuxbios.org/Development_Guidelines#Coding_Style)
> 	dimm_size = smbus_read_byte(ctrl->channel0[i], 31);			

smbus_read_byte -> spd_read_byte

31 -> SPD_BANK_DENSITY
> 	/* The i810 can't handle dimms larger than 128MB per side */
> 	/* Note: the factory BIOS just dies if it spots this :D */

Nice entry for a "Why LinuxBIOS sucks less" list in the wiki ;-)
> 	if(dimm_size < 32)
> 	{
> 		print_err("DIMM row sizes larger than 128MB not 
> 					supported on i810\r\n");
> 		print_err("Treating as 128MB DIMM\r\n");

Will this work? Did you test it?
> 	/* If the dimm is dual-sided, the DRP value is +2 */
> 	/* TODO: Figure out asymetrical configurations */
> 	if((smbus_read_byte(ctrl->channel0[i], 127) | 0xf) == 0xff)

smbus_read_byte -> spd_read_byte

127 -> SPD_INTEL_SPEC_100_MHZ
> /* TODO */
> pci_write_config8(ctrl->d0, PAM, 0x00);

This should probably be set to "Read/Write, all accesses are directed
to system memory DRAM." for all areas (?)
> +static void sdram_enable(int controllers, const struct mem_controller *ctrl)
> +{
> int i;
> /* Todo: this will currently work with either one dual sided or two single
> * sided dimms. Needs to work with 2 dual sided dimms in the long run */
> uint32_t row_offset;
> spd_set_dram_size(ctrl, row_offset);
> /* 1. Apply NOP. */
> PRINT_DEBUG("RAM Enable 1: Apply NOP\r\n");
> do_ram_command(ctrl, RAM_COMMAND_NOP, 0, row_offset);
> udelay(200);

Did you test the delays? It's likely that smaller values will suffice
(not too important, though, RAM init is pretty fast anyway).
> Index: src/northbridge/intel/i810/debug.c
>
> --- src/northbridge/intel/i810/debug.c	(revision 0)
> +++ src/northbridge/intel/i810/debug.c	(revision 0)
> @@ -0,0 +1,35 @@
>
> +static void dump_spd_registers(const struct mem_controller *ctrl)
> +{
> int i;
> print_debug("\r\n");
> for(i = 0; i < 4; i++) {
> unsigned device;
> device = ctrl->channel0[i];
> if (device) {
> 	int j;
> 	print_debug("dimm: "); 
> 	print_debug_hex8(i); 
> 	print_debug(".0: ");
> 	print_debug_hex8(device);
> 	for(j = 0; j < 256; j++) {
> 		int status;
> 		unsigned char byte;
> 		if ((j & 0xf) == 0) {
> 			print_debug("\r\n");
> 			print_debug_hex8(j);
> 			print_debug(": ");
> 		}
> 		status = smbus_read_byte(device, j);
> 		if (status < 0) {
> 			print_debug("bad device\r\n");
> 			break;
> 		}
> 		byte = status & 0xff;
> 		print_debug_hex8(byte);
> 		print_debug_char(' ');
> 	}
> 	print_debug("\r\n");
> }
> }
>
> +}

I'd like to drop this. There are hundreds of debug.c all over the place.
Using src/sdram/generic_dump_spd.c should work just the same, correct?
> +/* Table which returns the ram size in MB when fed the DRP[7:4] or [3:0] value */

ram -> RAM. Please use the correct spelling of all terms, even in
comments.
> +static void pci_domain_set_resources(device_t dev)
> +{
> device_t mc_dev;
>    uint32_t pci_tolm;
>    pci_tolm = find_pci_tolm(&dev->link[0]);
> mc_dev = dev->link[0].children;
> if (mc_dev) {
> /* Figure out which areas are/should be occupied by RAM.
>  * This is all computed in kilobytes and converted to/from
>  * the memory controller right at the edges.
>  * Having different variables in different units is
>  * too confusing to get right.  Kilobytes are good up to
>  * 4 Terabytes of RAM...
>  */
> unsigned long tomk, tolmk;
> int idx;
> int drp_value;
> /* Get the value of the total memory, which we stored in scratch 
>  * in raminit.c. The units are ticks of 8MB, i.e. 1 means 8MB.
>  */
> /*First get the value for DIMM 0 */
> drp_value = pci_read_config8(mc_dev, DRP) & 0xf; 
> /* Translate it to MB and add to tomk */
> tomk = (unsigned long)(translate_i810_to_mb[drp_value]);
> /* Now do the same for DIMM 1 */
> drp_value = pci_read_config8(mc_dev, DRP) >> 4;

Just a minor comment:

drp_value >>= 4

should work, too? No need to read the same value twice from the register...
> /* Convert tomk from MB to KB */
> tomk = tomk << 10;		

tomk *= 1024;
> /* Compute the top of Low memory */
> tolmk = pci_tolm >> 10;

tolmk = pci_tolm / 1024;
> +chip northbridge/intel/i810
> device pci_domain 0 on
> chip southbridge/intel/i82801aa # Southbridge
> 	device pci 1e.0 on end # PCI Bridge
> 	device pci 1f.0 on  # ISA/LPC? Bridge
> 		#chip superio/smsc/lpc47b272 # Super I/O
> 		#	device pnp 2e.0 on end # Floppy
> 		#	device pnp 2e.3 off end # Parallel port
> 		#	device pnp 2e.4 on # COM1
> 		#		io 0x60 = 0x3f8
> 		#		irq 0x70 = 4
> 		#	end
> 		#	device pnp 2e.5 on end # COM2
> 		#	device pnp 2e.4 on end # Power mgmt.
> 		#	#device pnp 2e.5 on end # Mouse
> 		#	device pnp 2e.7 on # Keyboard & Mouse?
> 		#		io 0x60 = 0x60
> 		#		io 0x62 = 0x64
> 		#		irq 0x70 = 1
> 		#		irq 0x72 = 12 # ???
> 		#	end
> 		#end
> 	end
> 	device pci 1f.1 on end # IDE
> 	device pci 1f.2 on end # USB
> 	device pci 1f.3 on end # SMBus
> 	device pci 1f.5 off end # AC'97
> 	device pci 1f.6 off end # AC'97 Modem (MC'97)?
> end
> device pci 0.0 on end # Host bridge

Not sure about this. I thought the order of the devices _does_ matter
here? I.e. lower IDs must come first? Can somebody please clarify this?
> Index: src/mainboard/asus/mew-vm/debug.c
>
> --- src/mainboard/asus/mew-vm/debug.c	(revision 0)
> +++ src/mainboard/asus/mew-vm/debug.c	(revision 0)
> @@ -0,0 +1,66 @@
> +/* Generic debug ops, used by many mainboards */
> +static void print_debug_pci_dev(unsigned dev)
> +{
> print_debug("PCI: ");
> print_debug_hex8((dev >> 16) & 0xff);
> print_debug_char(':');
> print_debug_hex8((dev >> 11) & 0x1f);
> print_debug_char('.');
> print_debug_hex8((dev >> 8) & 7);
>
> +}
>
> +static void print_pci_devices(void)
> +{
> device_t dev;
> for(dev = PCI_DEV(0, 0, 0);
> dev <= PCI_DEV(0, 0x1f, 0x7); 
> dev += PCI_DEV(0,0,1)) {
> uint32_t id;
> id = pci_read_config32(dev, PCI_VENDOR_ID);
> if (((id & 0xffff) == 0x0000) || ((id & 0xffff) == 0xffff) ||
> 	(((id >> 16) & 0xffff) == 0xffff) ||
> 	(((id >> 16) & 0xffff) == 0x0000)) {
> 	continue;
> }
> print_debug_pci_dev(dev);
> print_debug("\r\n");
> }
>
> +}
>
> +static void dump_pci_device(unsigned dev)
> +{
> int i;
> print_debug_pci_dev(dev);
> print_debug("\r\n");
> for(i = 0; i <= 255; i++) {
> unsigned char val;
> if ((i & 0x0f) == 0) {
> 	print_debug_hex8(i);
> 	print_debug_char(':');
> }
> val = pci_read_config8(dev, i);
> print_debug_char(' ');
> print_debug_hex8(val);
> if ((i & 0x0f) == 0x0f) {
> 	print_debug("\r\n");
> }
> }
>
> +}
>
> +static void dump_pci_devices(void)
> +{
> device_t dev;
> for(dev = PCI_DEV(0, 0, 0);
> dev <= PCI_DEV(0, 0x1f, 0x7); 
> dev += PCI_DEV(0,0,1)) {
> uint32_t id;
> id = pci_read_config32(dev, PCI_VENDOR_ID);
> if (((id & 0xffff) == 0x0000) || ((id & 0xffff) == 0xffff) ||
> 	(((id >> 16) & 0xffff) == 0xffff) ||
> 	(((id >> 16) & 0xffff) == 0x0000)) {
> 	continue;
> }
> dump_pci_device(dev);
> }
>
> +}

Dito, please drop this.

I think there's a global debug.c somewhere, and if there isn't we should
create one...
> Index: src/mainboard/asus/mew-vm/auto.c

Oh, I'm not sure this will work or is a good idea.

I'd rather use "mew_vm" (instead of "mew-vm"). Pathnames may appear as
variable names (for instance) in C code, that's also why all directory names
start with a letter (i440bx, not 440bx).

Same for dashes, "mew-vm" is not a valid C variable name.
> +void udelay(int usecs) 
> +{
> int i;
> for(i = 0; i < usecs; i++)
> outb(i&0xff, 0x80);
>
> +}

Is this needed? There's a global version somewhere.
> Index: targets/asus/mew-vm/Config.lb
>
> --- targets/asus/mew-vm/Config.lb	(revision 0)
> +++ targets/asus/mew-vm/Config.lb	(revision 0)

Please add a copyright header, here.
> option ROM_IMAGE_SIZE=0x10000

option ROM_IMAGE_SIZE = 64 * 1024

(this should really be done for all boards, makes the stuff a lot more
readable IMO)

With the above changes I think we can commit the code.

Uwe.