diff mbox

[U-Boot,v6,4/4] mpc85xx: Add board support for the eXMeritus HWW-1U-1A devices

Message ID 1300208664-18339-5-git-send-email-Kyle.D.Moffett@boeing.com
State Superseded
Headers show

Commit Message

Kyle Moffett March 15, 2011, 5:04 p.m. UTC
The eXMeritus HWW-1U-1A unit is a DO-160-certified 13lb 1U chassis
with 3 independent TEMPEST zones.  Two independent P2020 computers may
be found inside each zone.  Complete hardware support is included.

High-level hardware overview:
  * DO-160 certified for passenger aircraft (noncritical)
  * TEMPEST ceritified for RED/BLACK separation
  * 3 zones per chassis, 2 computers per zone (total of 6)
  * Dual-core 1.066GHz P2020 per computer
  * One 2GB DDR2 SO-RDIMM module per computer (upgradable to 4GB)
  * Removable 80GB or 160GB Intel X18-M SSD per computer
  * Front-accessible dual-port E1000E per computer
  * Front-accessible serial console per computer
  * Front-accessible USB port per computer
  * Internal Gigabit crossover within each TEMPEST zone
  * Internal unidirectional fiber links across TEMPEST zones
  * Battery-backed DS1339 I2C RTC on each CPU.

Combined, each 13lb 1U chassis contains 12GB RAM, 12 cores @ 1.066GHz,
12 front-accessible Gigabit Ethernet ports and 960GB of solid-state
storage with a total power consumption of ~200W.

Additional notes:
  * SPD detection is only known to work with the DO-160-certified DIMMs

  * A U-Boot built with 36-bit address-space seems to work, but I don't
    yet have a usable 36-bit kernel or DTB, so it's mostly untested.

  * CPU reset is a little quirky due to hardware misfeature, see the
    extensive comments in the board_reset() function in hww1u1a.c

Signed-off-by: Kyle Moffett <Kyle.D.Moffett@boeing.com>
Cc: Andy Fleming <afleming@gmail.com>
Cc: Kumar Gala <kumar.gala@freescale.com>
---
 MAINTAINERS                       |    4 +
 board/exmeritus/hww1u1a/Makefile  |   54 ++++
 board/exmeritus/hww1u1a/ddr.c     |   34 +++
 board/exmeritus/hww1u1a/gpios.h   |   67 +++++
 board/exmeritus/hww1u1a/hww1u1a.c |  543 +++++++++++++++++++++++++++++++++++++
 board/exmeritus/hww1u1a/law.c     |   34 +++
 board/exmeritus/hww1u1a/tlb.c     |  106 +++++++
 boards.cfg                        |    2 +
 include/configs/HWW1U1A.h         |  474 ++++++++++++++++++++++++++++++++
 9 files changed, 1318 insertions(+), 0 deletions(-)
 create mode 100644 board/exmeritus/hww1u1a/Makefile
 create mode 100644 board/exmeritus/hww1u1a/ddr.c
 create mode 100644 board/exmeritus/hww1u1a/gpios.h
 create mode 100644 board/exmeritus/hww1u1a/hww1u1a.c
 create mode 100644 board/exmeritus/hww1u1a/law.c
 create mode 100644 board/exmeritus/hww1u1a/tlb.c
 create mode 100644 include/configs/HWW1U1A.h

Comments

Wolfgang Denk March 15, 2011, 7:36 p.m. UTC | #1
Dear Kyle Moffett,

In message <1300208664-18339-5-git-send-email-Kyle.D.Moffett@boeing.com> you wrote:
> The eXMeritus HWW-1U-1A unit is a DO-160-certified 13lb 1U chassis
> with 3 independent TEMPEST zones.  Two independent P2020 computers may
> be found inside each zone.  Complete hardware support is included.

Please run checkpatch on your submissions!

...
> +	/* Ok, now go ahead and program all of those in one go */
> +	mpc85xx_gpio_set(	gpio_high|gpio_low|gpio_in,
> +				gpio_high|gpio_low,
> +				gpio_high);

ERROR: space prohibited after that open parenthesis '('
#427: FILE: board/exmeritus/hww1u1a/hww1u1a.c:100:
+       mpc85xx_gpio_set(       gpio_high|gpio_low|gpio_in,

> +	/*
> +	 * If things have been taken out of reset early (for example, by one
> +	 * of the BDI3000 debuggers), then we need to put them back in reset
> +	 * and delay a while before we continue.
> +	 */
> +#define GPIO_RESETS (GPIO_DIMM_RESET|GPIO_USB_RESET|GPIO_GETH0_RESET)
> +	if (mpc85xx_gpio_get(GPIO_RESETS)) {

Please don;t add #defines right in the middle of the code.

> +/*
> + * This little shell function just returns whether or not it's CPU A.
> + * It can be used to select the right device-tree when booting, etc.
> + */
> +int do_hww1u1a_test_cpu_a(cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[])

WARNING: line over 80 characters
#463: FILE: board/exmeritus/hww1u1a/hww1u1a.c:136:
+int do_hww1u1a_test_cpu_a(cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[])

> +U_BOOT_CMD(
> +	hww1u1a_test_cpu_a, 1, 0, do_hww1u1a_test_cpu_a,
> +	"Test if this is CPU A (versus B) on the eXMeritus HWW-1U-1A board",
> +	/*  */" && <command-if-true>\n"
> +	"hww1u1a_test_cpu_a || <command-if-false>\n"

What is this empty comment needed for?

> +	/* Now the serial# part of the hostname */
> +	for (j = 0; serialnr[j]; j++)
> +		if (isalnum(serialnr[j]))
> +			hww1u1a_prompt[i++] = tolower(serialnr[j]);

Braces needed for multiline statements.

...
> +		/* Turn on the "HRESET_REQ" pin (hard-reset request) */
> +		printf("\nRESET: Hardware reset triggered, waiting...\n");
> +		out_be32(&gur->rstcr, 0x2);
> +		while (1)
> +			udelay(10000);
> +	}

Should that not be an infinite wait here?

...
> +/* Enable the U-Boot "memory test" */
> +#define CONFIG_SYS_MEMTEST_START 0x00000000
> +#define CONFIG_SYS_MEMTEST_END   0x7fffffff

I think this has not been tested, right?


Best regards,

Wolfgang Denk
Kyle Moffett March 21, 2011, 4:29 p.m. UTC | #2
Wolfgang,

Thanks for your detailed reviews!

Once I get these last few style issues resolved, what more do I need to do to get this merged?  I don't really want to spam the list with more nearly identical copies of these patches unless I'm sure that all the necessary review items have been taken care of.


On Mar 15, 2011, at 15:36, Wolfgang Denk wrote:
> In message <1300208664-18339-5-git-send-email-Kyle.D.Moffett@boeing.com> you wrote:
>> The eXMeritus HWW-1U-1A unit is a DO-160-certified 13lb 1U chassis
>> with 3 independent TEMPEST zones.  Two independent P2020 computers may
>> be found inside each zone.  Complete hardware support is included.
> 
> Please run checkpatch on your submissions!

I did run it on this patch, although I forgot to run it on the resurrected reset patch (which is also now fixed).  It's a common gripe on the LKML that the tool is overzealous about certain warnings in cases where the "fix" makes the code less readable.

Specifically:

> ...
>> +	/* Ok, now go ahead and program all of those in one go */
>> +	mpc85xx_gpio_set(	gpio_high|gpio_low|gpio_in,
>> +				gpio_high|gpio_low,
>> +				gpio_high);
> 
> ERROR: space prohibited after that open parenthesis '('
> #427: FILE: board/exmeritus/hww1u1a/hww1u1a.c:100:
> +       mpc85xx_gpio_set(       gpio_high|gpio_low|gpio_in,

I could "fix" this code to read:

        mpc85xx_gpio_set(gpio_high|gpio_low|gpio_in,
                                gpio_high|gpio_low, gpio_high);

And it would be much harder to visually compare the three bitmask arguments against each other.


>> +	/*
>> +	 * If things have been taken out of reset early (for example, by one
>> +	 * of the BDI3000 debuggers), then we need to put them back in reset
>> +	 * and delay a while before we continue.
>> +	 */
>> +#define GPIO_RESETS (GPIO_DIMM_RESET|GPIO_USB_RESET|GPIO_GETH0_RESET)
>> +	if (mpc85xx_gpio_get(GPIO_RESETS)) {
> 
> Please don;t add #defines right in the middle of the code.

Fixed, thanks!


>> +/*
>> + * This little shell function just returns whether or not it's CPU A.
>> + * It can be used to select the right device-tree when booting, etc.
>> + */
>> +int do_hww1u1a_test_cpu_a(cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[])
> 
> WARNING: line over 80 characters
> #463: FILE: board/exmeritus/hww1u1a/hww1u1a.c:136:
> +int do_hww1u1a_test_cpu_a(cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[])

This one is only a warning, and it's much more readable to have 1 84-character line than split it across 2 different lines.  Even still, this warning is only issued if you pass "--subjective" to checkpatch, which is documented to "enable more subjective tests".  This thread discusses it further:
  http://lkml.org/lkml/2009/12/15/490

>> +U_BOOT_CMD(
>> +	hww1u1a_test_cpu_a, 1, 0, do_hww1u1a_test_cpu_a,
>> +	"Test if this is CPU A (versus B) on the eXMeritus HWW-1U-1A board",
>> +	/*  */" && <command-if-true>\n"
>> +	"hww1u1a_test_cpu_a || <command-if-false>\n"
> 
> What is this empty comment needed for?

Just a mental placeholder for the fact that the U_BOOT_CMD macro inserts the name of the command in that spot.  Will remove.


>> +	/* Now the serial# part of the hostname */
>> +	for (j = 0; serialnr[j]; j++)
>> +		if (isalnum(serialnr[j]))
>> +			hww1u1a_prompt[i++] = tolower(serialnr[j]);
> 
> Braces needed for multiline statements.

Fixed, thanks!


>> +		/* Turn on the "HRESET_REQ" pin (hard-reset request) */
>> +		printf("\nRESET: Hardware reset triggered, waiting...\n");
>> +		out_be32(&gur->rstcr, 0x2);
>> +		while (1)
>> +			udelay(10000);
>> +	}
> 
> Should that not be an infinite wait here?

At this point if the board does not reset due to hardware failure it's better off hanging than silently falling through.


>> +/* Enable the U-Boot "memory test" */
>> +#define CONFIG_SYS_MEMTEST_START 0x00000000
>> +#define CONFIG_SYS_MEMTEST_END   0x7fffffff
> 
> I think this has not been tested, right?

I'm pretty sure it's been tested, but not very recently; the memory on all the shipped units is ECC anyways.

Cheers,
Kyle Moffett
Wolfgang Denk March 21, 2011, 8:30 p.m. UTC | #3
Dear "Moffett, Kyle D",

In message <5B9D9C87-C278-4AF3-B20C-26ECFF6C09B7@boeing.com> you wrote:
> 
> > WARNING: line over 80 characters
> > #463: FILE: board/exmeritus/hww1u1a/hww1u1a.c:136:
> > +int do_hww1u1a_test_cpu_a(cmd_tbl_t *cmdtp, int flag, int argc, char * c=
> onst argv[])
> 
> This one is only a warning, and it's much more readable to have 1 84-charac=

In U-Boot, it is considered to be an ERROR.

> ter line than split it across 2 different lines.  Even still, this warning =
> is only issued if you pass "--subjective" to checkpatch, which is documente=
> d to "enable more subjective tests".  This thread discusses it further:

No, I get this warning without such flags.  The command I run is just
"checkpatch.pl --no-tree"

BTW - could you please restrict your line length to some 70 characters
or so?  Thanks.

> >> +U_BOOT_CMD(
> >> +	hww1u1a_test_cpu_a, 1, 0, do_hww1u1a_test_cpu_a,
> >> +	"Test if this is CPU A (versus B) on the eXMeritus HWW-1U-1A board",
> >> +	/*  */" && <command-if-true>\n"
> >> +	"hww1u1a_test_cpu_a || <command-if-false>\n"
> >=20
> > What is this empty comment needed for?
> 
> Just a mental placeholder for the fact that the U_BOOT_CMD macro inserts th=
> e name of the command in that spot.  Will remove.

We don't provide usage examples in the help text.  This should be
fixed in the first place.

> >> +		/* Turn on the "HRESET_REQ" pin (hard-reset request) */
> >> +		printf("\nRESET: Hardware reset triggered, waiting...\n");
> >> +		out_be32(&gur->rstcr, 0x2);
> >> +		while (1)
> >> +			udelay(10000);
> >> +	}
> >=20
> > Should that not be an infinite wait here?
> 
> At this point if the board does not reset due to hardware failure it's bett=
> er off hanging than silently falling through.

Why don't you simply call "hang()" then?

> >> +/* Enable the U-Boot "memory test" */
> >> +#define CONFIG_SYS_MEMTEST_START 0x00000000
> >> +#define CONFIG_SYS_MEMTEST_END   0x7fffffff
> >=20
> > I think this has not been tested, right?
> 
> I'm pretty sure it's been tested, but not very recently; the memory on all =
> the shipped units is ECC anyways.

I guess testing it would reveal that it crashes your system because you
are overwriting the exception vectors.

Best regards,

Wolfgang Denk
Kyle Moffett March 21, 2011, 9:15 p.m. UTC | #4
On Mar 21, 2011, at 16:30, Wolfgang Denk wrote:
> In message <5B9D9C87-C278-4AF3-B20C-26ECFF6C09B7@boeing.com> you wrote:
>> 
>>> WARNING: line over 80 characters
>>> #463: FILE: board/exmeritus/hww1u1a/hww1u1a.c:136:
>>> +int do_hww1u1a_test_cpu_a(cmd_tbl_t *cmdtp, int flag, int argc, char *
>>> const argv[])
>> 
>> This one is only a warning, and it's much more readable to have 1 84-charac
> 
> In U-Boot, it is considered to be an ERROR.

Just looking at the last ~200 commits (actually 187, because it ignores merges):

$ git format-patch -o recent-patches -200 origin/master
$ ./checkpatch.pl --no-tree --strict recent-patches/* >checkpatch.log 2>&1
$ grep 'over 80 char' checkpatch.log | wc -l
130

That's 130 lines in the last 200 patches which are over 80 characters?!?!
How are those patches any different from mine?

$ grep '^ERROR:' checkpatch.log | wc -l
113

And that's 113 HARD ERRORS from checkpatch!?!?!

Of those, 32 are "Missing Signed-off-by: line(s)", 20 are "macros with
complex values should be enclosed in parenthesis", 19 are inconsistent
or missing whitespace issues, 4 are "(foo*) instead of "(foo *)", 3 are
invalid UTF-8 errors, 4 are "return is not a function" errors, etc, etc.

Look, I'm really trying to comply with U-Boot coding standards, but I'm really
of pissed off about the inconsistent requirements you are applying to my
patches versus a lot of other things that YOU ARE MERGING on a regular basis.

So why are you picking on my board-specific code so hard here?  This is
extremely frustrating for me and a strong deterrent against us *ever*
contributing to U-Boot again in the future.


>> ter line than split it across 2 different lines.  Even still, this
>> warning is only issued if you pass "--subjective" to checkpatch, which
>> is documented to "enable more subjective tests".
> 
> No, I get this warning without such flags.  The command I run is just
> "checkpatch.pl --no-tree"

What version of checkpatch are you running?  I copied version 0.31 out of
my latest Linux kernel tree, which identical to the latest version from
here:

http://www.kernel.org/pub/linux/kernel/people/apw/checkpatch/

If U-Boot policy is to run checkpatch then you'd better either specify a
particular version and command-line options or be willing to accept the
default output of the latest version.


> BTW - could you please restrict your line length to some 70 characters
> or so?  Thanks.

Sorry about that, sending email through an Exchange server is no fun :-(.
Hopefully I've got it fixed.


>>>> +U_BOOT_CMD(
>>>> +	hww1u1a_test_cpu_a, 1, 0, do_hww1u1a_test_cpu_a,
>>>> +	"Test if this is CPU A (versus B) on the eXMeritus HWW-1U-1A board",
>>>> +	/*  */" && <command-if-true>\n"
>>>> +	"hww1u1a_test_cpu_a || <command-if-false>\n"
>>> =20
>>> What is this empty comment needed for?
>> 
>> Just a mental placeholder for the fact that the U_BOOT_CMD macro inserts
>> the name of the command in that spot.  Will remove.
> 
> We don't provide usage examples in the help text.  This should be
> fixed in the first place.

This *IS* the help text, and not a sample usage.  This is visually and
effectively no different from this text in common/cmd_mp.c:

U_BOOT_CMD(
        cpu, CONFIG_SYS_MAXARGS, 1, cpu_cmd,
        "Multiprocessor CPU boot manipulation and release",
            "<num> reset                 - Reset cpu <num>\n"
        "cpu <num> status                - Status of cpu <num>\n"
        "cpu <num> disable               - Disable cpu <num>\n"
        "cpu <num> release <addr> [args] - Release cpu <num> at <addr> with [args]"
#ifdef CPU_ARCH_HELP
        "\n"
        CPU_ARCH_HELP
#endif
);


>>>> +		/* Turn on the "HRESET_REQ" pin (hard-reset request) */
>>>> +		printf("\nRESET: Hardware reset triggered, waiting...\n");
>>>> +		out_be32(&gur->rstcr, 0x2);
>>>> +		while (1)
>>>> +			udelay(10000);
>>>> +	}
>>> Should that not be an infinite wait here?
>> 
>> At this point if the board does not reset due to hardware failure it's
>> better off hanging than silently falling through.
> 
> Why don't you simply call "hang()" then?

Didn't know it existed at the time the code was written.  Will fix.

Cheers,
Kyle Moffett
Wolfgang Denk March 21, 2011, 9:34 p.m. UTC | #5
Dear "Moffett, Kyle D",

In message <AC0C0781-9E5F-42F7-9DB6-EECF6A5BE840@boeing.com> you wrote:
>
> Just looking at the last ~200 commits (actually 187, because it ignores mer=
> ges):
> 
> $ git format-patch -o recent-patches -200 origin/master
> $ ./checkpatch.pl --no-tree --strict recent-patches/* >checkpatch.log 2>&1
> $ grep 'over 80 char' checkpatch.log | wc -l
> 130
> 
> That's 130 lines in the last 200 patches which are over 80 characters?!?!
> How are those patches any different from mine?

The difference is: They were not detected.

Patches welcome.

> Look, I'm really trying to comply with U-Boot coding standards, but I'm rea=
> lly
> of pissed off about the inconsistent requirements you are applying to my
> patches versus a lot of other things that YOU ARE MERGING on a regular basi=
> s.

The requirements are NOT inconsistent.  It's just that nobody is
perfect, and nobody ever claimed that we manage to get 100% of review
coverage.

> So why are you picking on my board-specific code so hard here?  This is

It's just that the problems got noticed.

> > "checkpatch.pl --no-tree"
> 
> What version of checkpatch are you running?  I copied version 0.31 out of
> my latest Linux kernel tree, which identical to the latest version from
> here:

-> checkpatch.pl --version
Usage: checkpatch.pl [OPTION]... [FILE]...
Version: 0.31
...

> If U-Boot policy is to run checkpatch then you'd better either specify a
> particular version and command-line options or be willing to accept the
> default output of the latest version.

I don't see any version issue here, nor do I use any non-standard
options.

> >>>> +U_BOOT_CMD(
> >>>> +	hww1u1a_test_cpu_a, 1, 0, do_hww1u1a_test_cpu_a,
> >>>> +	"Test if this is CPU A (versus B) on the eXMeritus HWW-1U-1A board",
> >>>> +	/*  */" && <command-if-true>\n"
> >>>> +	"hww1u1a_test_cpu_a || <command-if-false>\n"
> >>> =20
> >>> What is this empty comment needed for?
> >> 
> >> Just a mental placeholder for the fact that the U_BOOT_CMD macro inserts
> >> the name of the command in that spot.  Will remove.
> > 
> > We don't provide usage examples in the help text.  This should be
> > fixed in the first place.
>
> This *IS* the help text, and not a sample usage.  This is visually and
> effectively no different from this text in common/cmd_mp.c:

The synopsis of a command gives the command name and possible options,
and explains what the command does and what the options do.

"name && <command-if-true>" does NOT fall into that pattern.

Look, alternatively I can claim your help message is highly incomplete
as it fails to cover use cases like

	name || <command-if-false>

etc.

Not to mention that the usage message is plain wrong for all boards
that don't have the hush shell enabled.

Best regards,

Wolfgang Denk
Kyle Moffett March 21, 2011, 10:01 p.m. UTC | #6
On Mar 21, 2011, at 17:34, Wolfgang Denk wrote:
> In message <AC0C0781-9E5F-42F7-9DB6-EECF6A5BE840@boeing.com> you wrote:
>> 
>> Just looking at the last ~200 commits (actually 187, because it ignores
>> merges):
>> 
>> $ git format-patch -o recent-patches -200 origin/master
>> $ ./checkpatch.pl --no-tree --strict recent-patches/* >checkpatch.log 2>&1
>> $ grep 'over 80 char' checkpatch.log | wc -l
>> 130
>> 
>> That's 130 lines in the last 200 patches which are over 80 characters?!?!
>> How are those patches any different from mine?
> 
> The difference is: They were not detected.
> 
> Patches welcome.

If those were patches from two years ago and your style policies had
significantly changed since then I would understand.

But those are patches from *LAST MONTH* which you were perfectly happy to
merge from dozens of different developers, but had *I* submitted identical
patches they would have been rejected without even a second glance.

That is inconsistent at best, and in my humble opinion downright rude.


>> Look, I'm really trying to comply with U-Boot coding standards, but I'm
>> really of pissed off about the inconsistent requirements you are applying
>> to my patches versus a lot of other things that YOU ARE MERGING on a
>> regular basis.
> 
> The requirements are NOT inconsistent.  It's just that nobody is
> perfect, and nobody ever claimed that we manage to get 100% of review
> coverage.

I give up.

I apparently cannot rely on the U-Boot *CODE* to understand what the
U-Boot *CODING* style is.

The time investment to get reasonable board support merged into U-Boot is
proving to be *greater* than the time investment to just maintain our
board ports out of tree.

If anyone would like to use our code as a reference or try to get it merged
themselves, I will continue to maintain our GPLed out-of-tree patchset here:
  http://opensource.exmeritus.com/git/

But otherwise I see no valid reason I should waste any more of my time
submitting patches which get torn apart out of hand over issues which are
completely ignored for patches which come in from other maintainers.

In the future we will have to weigh other boot-loader alternatives due to
the unfriendly attitude here.

Good luck.

Cheers,
Kyle Moffett
Wolfgang Denk March 21, 2011, 10:24 p.m. UTC | #7
Dear "Moffett, Kyle D",

In message <DDD31591-33CD-49E4-B303-3477E00933BF@boeing.com> you wrote:
>
> I apparently cannot rely on the U-Boot *CODE* to understand what the
> U-Boot *CODING* style is.

You don't have to rely on the code.  It's clearly documented.

The README says:

	Coding Standards:
	-----------------

	All contributions to U-Boot should conform to the Linux kernel
	coding style; see the file "Documentation/CodingStyle" and the script
	"scripts/Lindent" in your Linux kernel source directory...

http://www.denx.de/wiki/U-Boot/CodingStyle  says the same.

And the referred document says:

			Chapter 2: Breaking long lines and strings
	 
	Coding style is all about readability and maintainability using
	commonly
	available tools.
	 
	The limit on the length of lines is 80 columns and this is a strongly
	preferred limit.
	 
	Statements longer than 80 columns will be broken into sensible chunks.

Now what exactly is unclear here?


And no, you are not the only one who is asked to fix his code beause
of "Line too long" errors. Here just a small collection from the last
4 months or so:

11/11 To:Xiangfu Liu     Re: [U-Boot] [PATCH 1/5] those files are jz4740 base files
      http://article.gmane.org/gmane.comp.boot-loaders.u-boot/88655
12/04 To:Luigi 'Comio'   Re: [U-Boot] [PATCH 5/6 v2] Enable bootstrap support for MIPS architecture.
      http://article.gmane.org/gmane.comp.boot-loaders.u-boot/90325
12/12 To:Macpaul Lin     Re: [U-Boot] [PATCH] ftgmac100: support of gigabit eth ftgmac100
      http://article.gmane.org/gmane.comp.boot-loaders.u-boot/90740
12/12 To:Macpaul Lin     Re: [U-Boot] [PATCH] ftgmac100: support of gigabit eth ftgmac100
      http://article.gmane.org/gmane.comp.boot-loaders.u-boot/90740
03/13 To:Heiko Schocher  Re: [U-Boot] [PATCH 01/20] keymile: rework headerfiles for keymile boards
      http://article.gmane.org/gmane.comp.boot-loaders.u-boot/95728


> But otherwise I see no valid reason I should waste any more of my time
> submitting patches which get torn apart out of hand over issues which are
> completely ignored for patches which come in from other maintainers.

This is simply not true. But I'm not sure if you are still listening
to any rational arguments. I can only ask you to calm down, and
eventually reconsider.  Sorry.

Best regards,

Wolfgang Denk
Kyle Moffett March 21, 2011, 11:07 p.m. UTC | #8
On Mar 21, 2011, at 18:24, Wolfgang Denk wrote:
> In message <DDD31591-33CD-49E4-B303-3477E00933BF@boeing.com> you wrote:
>> 
>> I apparently cannot rely on the U-Boot *CODE* to understand what the
>> U-Boot *CODING* style is.
> 
> You don't have to rely on the code.  It's clearly documented.
> 
> The README says:
> 
> 	Coding Standards:
> 	-----------------
> 
> 	All contributions to U-Boot should conform to the Linux kernel
> 	coding style; see the file "Documentation/CodingStyle" and the script
> 	"scripts/Lindent" in your Linux kernel source directory...
> 
> http://www.denx.de/wiki/U-Boot/CodingStyle  says the same.

So here it is claimed that the U-Boot coding style is the same as the Linux
Kernel coding style.

But previously you said:
>>>> WARNING: line over 80 characters
>>>> #463: FILE: board/exmeritus/hww1u1a/hww1u1a.c:136:
>>>> +int do_hww1u1a_test_cpu_a(cmd_tbl_t *cmdtp, int flag, int argc, char *
>>>> const argv[])
>>> 
>>> This one is only a warning, and it's much more readable to have 1
>>> 84-character line than split it across 2 different lines.  Even still,
>>> this warning is only issued if you pass "--subjective" to checkpatch, which
>>> is documented to "enable more subjective tests".
> 
> In U-Boot, it is considered to be an ERROR.


So which is it?

If U-Boot uses the official Linux Kernel CodingStyle, then a few >80-char
lines are OK if it increases readability, for example by not having to wrap
a simple U_BOOT_CMD function declaration that goes just a whole 4 characters
over the limit.

If U-Boot does NOT use the Linux Kernel CodingStyle and wants to refuse all
patches with over 80 characters then you should copy checkpatch.pl and the
CodingStyle document and change that wording from "strongly preferred" to
"hard requirement", and change the "WARNING" to "ERROR".


> And the referred document says:
> 			Chapter 2: Breaking long lines and strings
> 	 
> 	Coding style is all about readability and maintainability using
> 	commonly
> 	available tools.
> 	 
> 	The limit on the length of lines is 80 columns and this is a strongly
> 	preferred limit.
> 	 
> 	Statements longer than 80 columns will be broken into sensible chunks.
> 
> Now what exactly is unclear here?

You removed the very next paragraph in the Linux CodingStyle file,
which contains:

> The only exception to this is where exceeding 80 columns significantly increases

> readability and does not hide information.


Furthermore, Linus Torvalds himself said in an email:
>> We fixed that to allow checkpatch to skip those warnings, but the fact is, 
>> the fundamnetal problem has always been the "80 character" part.

>> 
>> I don't think any kernel developers use a vt100 any more. And even if they 
>> do, I bet they curse the "24 lines" more than they curse the occasional 
>> 80+ character lines.
>> 
>> I'd be ok with changing the warning to 132 characters, which is another 
>> perfectly fine historical limit. Or we can split the difference, and say 
>> "ok, 106 characters is too much". I don't care. But 80 characters is 
>> causing too many idiotic changes.
>> 
>> There are way worse problems in many patches than long lines. Too complex 
>> expressions. Too deep indentation. Pure crap code. People seem to get way 
>> too hung up on ".. but at least it passes checkpatch". 

The line you were complaining about is a static U_BOOT_CMD function declaration
for goodness sakes!  It's about as common as dirt in the U-Boot code and the
only reason it doesn't fit on one line anymore is because I made the name of the
function *NAME* about 8 characters longer in this version of the patch than it
was in a previous patch.

It's not any more complex than it was before, nor is it any harder to read.

I'm tired and fed up with this whole mess.  If you think it's likely to be
accepted I'll go ahead and submit one more final respin tomorrow, assuming
I feel up to it.  Otherwise, I wish you the best of luck with U-Boot.

Good night.

Cheers,
Kyle Moffett
diff mbox

Patch

diff --git a/MAINTAINERS b/MAINTAINERS
index 4756f14..6644baf 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -320,6 +320,10 @@  Reinhard Meyer <reinhard.meyer@emk-elektronik.de>
 	TOP5200		MPC5200
 	TOP9000		ARM926EJS (AT91SAM9xxx SoC)
 
+Kyle Moffett <Kyle.D.Moffett@boeing.com>
+
+	HWW1U1A		P2020
+
 Tolunay Orkun <torkun@nextio.com>
 
 	csb272		PPC405GP
diff --git a/board/exmeritus/hww1u1a/Makefile b/board/exmeritus/hww1u1a/Makefile
new file mode 100644
index 0000000..b927f59
--- /dev/null
+++ b/board/exmeritus/hww1u1a/Makefile
@@ -0,0 +1,54 @@ 
+#
+# Copyright 2007-2009 Freescale Semiconductor, Inc.
+# (C) Copyright 2001-2006
+# Wolfgang Denk, DENX Software Engineering, wd@denx.de.
+#
+# See file CREDITS for list of people who contributed to this
+# project.
+#
+# This program is free software; you can redistribute it and/or
+# modify it under the terms of the GNU General Public License as
+# published by the Free Software Foundation; either version 2 of
+# the License, or (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, write to the Free Software
+# Foundation, Inc., 59 Temple Place, Suite 330, Boston,
+# MA 02111-1307 USA
+#
+
+include $(TOPDIR)/config.mk
+
+LIB	= $(obj)lib$(BOARD).o
+
+COBJS-y	+= $(BOARD).o
+COBJS-y	+= law.o
+COBJS-y	+= tlb.o
+COBJS-$(CONFIG_DDR_SPD) += ddr.o
+
+SRCS	:= $(SOBJS:.o=.S) $(COBJS-y:.o=.c)
+OBJS	:= $(addprefix $(obj),$(COBJS-y))
+SOBJS	:= $(addprefix $(obj),$(SOBJS))
+
+$(LIB):	$(obj).depend $(OBJS) $(SOBJS)
+	$(call cmd_link_o_target, $(OBJS))
+
+clean:
+	rm -f $(OBJS) $(SOBJS)
+
+distclean:	clean
+	rm -f $(LIB) core *.bak .depend
+
+#########################################################################
+
+# defines $(obj).depend target
+include $(SRCTREE)/rules.mk
+
+sinclude $(obj).depend
+
+#########################################################################
diff --git a/board/exmeritus/hww1u1a/ddr.c b/board/exmeritus/hww1u1a/ddr.c
new file mode 100644
index 0000000..36d02ad
--- /dev/null
+++ b/board/exmeritus/hww1u1a/ddr.c
@@ -0,0 +1,34 @@ 
+/*
+ * Copyright 2009-2010 eXMeritus, A Boeing Company
+ * Copyright 2008-2009 Freescale Semiconductor, Inc.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * Version 2 as published by the Free Software Foundation.
+ */
+
+#include <common.h>
+
+#include <asm/fsl_ddr_sdram.h>
+#include <asm/fsl_ddr_dimm_params.h>
+
+void fsl_ddr_board_options(memctl_options_t *popts,
+				dimm_params_t *pdimm,
+				unsigned int ctrl_num)
+{
+	/*
+	 * We only support one DIMM, so according to the P2020 docs we should
+	 * set the options as follows:
+	 */
+	popts->cs_local_opts[0].odt_rd_cfg = 0;
+	popts->cs_local_opts[0].odt_wr_cfg = 4;
+	popts->cs_local_opts[1].odt_rd_cfg = 0;
+	popts->cs_local_opts[1].odt_wr_cfg = 0;
+	popts->half_strength_driver_enable = 0;
+
+	/* Manually configured for our static clock rate */
+	popts->clk_adjust = 4;
+	popts->cpo_override = 4;
+	popts->write_data_delay = 2;
+	popts->twoT_en = 0;
+}
diff --git a/board/exmeritus/hww1u1a/gpios.h b/board/exmeritus/hww1u1a/gpios.h
new file mode 100644
index 0000000..fa383a6
--- /dev/null
+++ b/board/exmeritus/hww1u1a/gpios.h
@@ -0,0 +1,67 @@ 
+/*
+ * Copyright 2010 eXMeritus, A Boeing Company
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation; either version 2 of
+ * the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.	 See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston,
+ * MA 02111-1307 USA
+ */
+
+#include <asm/mpc85xx_gpio.h>
+
+/* Common CPU A/B GPIOs (GPIO8-GPIO15 and IRQ4-IRQ6) */
+#define GPIO_CPU_ID		(1UL << (31 -  8))
+#define GPIO_BLUE_LED		(1UL << (31 -  9))
+#define GPIO_DIMM_RESET		(1UL << (31 - 10))
+#define GPIO_USB_RESET		(1UL << (31 - 11))
+#define GPIO_UNUSED_12		(1UL << (31 - 12))
+#define GPIO_GETH0_RESET	(1UL << (31 - 13))
+#define GPIO_RS422_RE		(1UL << (31 - 14))
+#define GPIO_RS422_DE		(1UL << (31 - 15))
+#define IRQ_I2CINT		(1UL << (31 - 20))
+#define IRQ_FANINT		(1UL << (31 - 21))
+#define IRQ_DIMM_EVENT		(1UL << (31 - 22))
+
+/* CPU A GPIOS (GPIO0-GPIO7 and IRQ0-IRQ3) */
+#define GPIO_CPUA_UNUSED_0	(1UL << (31 -  0))
+#define GPIO_CPUA_CPU_READY	(1UL << (31 -  1))
+#define GPIO_CPUA_DEBUG_LED2	(1UL << (31 -  2))
+#define GPIO_CPUA_DEBUG_LED1	(1UL << (31 -  3))
+#define GPIO_CPUA_TDIS2B	(1UL << (31 -  4)) /* MAC 2 TX B */
+#define GPIO_CPUA_TDIS2A	(1UL << (31 -  5)) /* MAC 2 TX A */
+#define GPIO_CPUA_TDIS1B	(1UL << (31 -  6)) /* MAC 1 TX B */
+#define GPIO_CPUA_TDIS1A	(1UL << (31 -  7)) /* MAC 1 TX A */
+#define IRQ_CPUA_UNUSED_0	(1UL << (31 - 16))
+#define IRQ_CPUA_UNUSED_1	(1UL << (31 - 17))
+#define IRQ_CPUA_UNUSED_2	(1UL << (31 - 18))
+#define IRQ_CPUA_UNUSED_3	(1UL << (31 - 19))
+
+/* CPU B GPIOS (GPIO0-GPIO7 and IRQ0-IRQ3) */
+#define GPIO_CPUB_RMUX_SEL1B	(1UL << (31 -  0))
+#define GPIO_CPUB_RMUX_SEL0B	(1UL << (31 -  1))
+#define GPIO_CPUB_RMUX_SEL1A	(1UL << (31 -  2))
+#define GPIO_CPUB_RMUX_SEL0A	(1UL << (31 -  3))
+#define GPIO_CPUB_UNUSED_4	(1UL << (31 -  4))
+#define GPIO_CPUB_CPU_READY	(1UL << (31 -  5))
+#define GPIO_CPUB_DEBUG_LED2	(1UL << (31 -  6))
+#define GPIO_CPUB_DEBUG_LED1	(1UL << (31 -  7))
+#define IRQ_CPUB_SD_1A		(1UL << (31 - 16))
+#define IRQ_CPUB_SD_2B		(1UL << (31 - 17))
+#define IRQ_CPUB_SD_2A		(1UL << (31 - 18))
+#define IRQ_CPUB_SD_1B		(1UL << (31 - 19))
+
+/* If it isn't CPU A then it's CPU B */
+static inline unsigned int hww1u1a_is_cpu_a(void)
+{
+	return !mpc85xx_gpio_get(GPIO_CPU_ID);
+}
diff --git a/board/exmeritus/hww1u1a/hww1u1a.c b/board/exmeritus/hww1u1a/hww1u1a.c
new file mode 100644
index 0000000..f0dfd67
--- /dev/null
+++ b/board/exmeritus/hww1u1a/hww1u1a.c
@@ -0,0 +1,543 @@ 
+/*
+ * Copyright 2009-2011 eXMeritus, A Boeing Company
+ * Copyright 2007-2009 Freescale Semiconductor, Inc.
+ *
+ * See file CREDITS for list of people who contributed to this
+ * project.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation; either version 2 of
+ * the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.	 See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston,
+ * MA 02111-1307 USA
+ */
+
+#include <common.h>
+#include <command.h>
+#include <pci.h>
+#include <asm/processor.h>
+#include <asm/mmu.h>
+#include <asm/cache.h>
+#include <asm/immap_85xx.h>
+#include <asm/fsl_pci.h>
+#include <asm/fsl_ddr_sdram.h>
+#include <asm/io.h>
+#include <miiphy.h>
+#include <libfdt.h>
+#include <linux/ctype.h>
+#include <fdt_support.h>
+#include <tsec.h>
+#include <asm/fsl_law.h>
+#include <netdev.h>
+#include <malloc.h>
+#include <i2c.h>
+#include <pca953x.h>
+
+#include "gpios.h"
+
+DECLARE_GLOBAL_DATA_PTR;
+
+int checkboard(void)
+{
+	unsigned int gpio_high = 0;
+	unsigned int gpio_low  = 0;
+	unsigned int gpio_in   = 0;
+	unsigned int i;
+
+	puts("Board: HWW-1U-1A ");
+#ifdef CONFIG_PHYS_64BIT
+	puts("(36-bit addrmap) ");
+#endif
+
+	/*
+	 * First just figure out which CPU we're on, then use that to
+	 * configure the lists of other GPIOs to be programmed.
+	 */
+	mpc85xx_gpio_set_in(GPIO_CPU_ID);
+	if (hww1u1a_is_cpu_a()) {
+		puts("CPU A\n");
+
+		/* We want to turn on some LEDs */
+		gpio_high |= GPIO_CPUA_CPU_READY;
+		gpio_low  |= GPIO_CPUA_DEBUG_LED1;
+		gpio_low  |= GPIO_CPUA_DEBUG_LED2;
+
+		/* Disable the unused transmitters */
+		gpio_low  |= GPIO_CPUA_TDIS1A;
+		gpio_high |= GPIO_CPUA_TDIS1B;
+		gpio_low  |= GPIO_CPUA_TDIS2A;
+		gpio_high |= GPIO_CPUA_TDIS2B;
+	} else {
+		puts("CPU B\n");
+
+		/* We want to turn on some LEDs */
+		gpio_high |= GPIO_CPUB_CPU_READY;
+		gpio_low  |= GPIO_CPUB_DEBUG_LED1;
+		gpio_low  |= GPIO_CPUB_DEBUG_LED2;
+
+		/* Enable the appropriate receivers */
+		gpio_high |= GPIO_CPUB_RMUX_SEL0A;
+		gpio_high |= GPIO_CPUB_RMUX_SEL0B;
+		gpio_low  |= GPIO_CPUB_RMUX_SEL1A;
+		gpio_low  |= GPIO_CPUB_RMUX_SEL1B;
+	}
+
+	/* These GPIOs are common */
+	gpio_in   |= IRQ_I2CINT | IRQ_FANINT | IRQ_DIMM_EVENT;
+	gpio_low  |= GPIO_RS422_RE;
+	gpio_high |= GPIO_RS422_DE;
+
+	/* Ok, now go ahead and program all of those in one go */
+	mpc85xx_gpio_set(	gpio_high|gpio_low|gpio_in,
+				gpio_high|gpio_low,
+				gpio_high);
+
+	/*
+	 * If things have been taken out of reset early (for example, by one
+	 * of the BDI3000 debuggers), then we need to put them back in reset
+	 * and delay a while before we continue.
+	 */
+#define GPIO_RESETS (GPIO_DIMM_RESET|GPIO_USB_RESET|GPIO_GETH0_RESET)
+	if (mpc85xx_gpio_get(GPIO_RESETS)) {
+		ccsr_ddr_t *ddr = (ccsr_ddr_t *)CONFIG_SYS_MPC85xx_DDR_ADDR;
+
+		puts("Debugger detected... extra device reset enabled!\n");
+
+		/* Put stuff into reset and disable the DDR controller */
+		mpc85xx_gpio_set_low(GPIO_RESETS);
+		out_be32(&ddr->sdram_cfg, 0x00000000);
+
+		puts("    Waiting 1 sec for reset...");
+		for (i = 0; i < 10; i++) {
+			udelay(100000);
+			puts(".");
+		}
+		puts("\n");
+	}
+
+	/* Now bring everything back out of reset again */
+	mpc85xx_gpio_set_high(GPIO_RESETS);
+	return 0;
+}
+
+/*
+ * This little shell function just returns whether or not it's CPU A.
+ * It can be used to select the right device-tree when booting, etc.
+ */
+int do_hww1u1a_test_cpu_a(cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[])
+{
+	if (argc > 1)
+		cmd_usage(cmdtp);
+
+	if (hww1u1a_is_cpu_a())
+		return 0;
+	else
+		return 1;
+}
+U_BOOT_CMD(
+	hww1u1a_test_cpu_a, 1, 0, do_hww1u1a_test_cpu_a,
+	"Test if this is CPU A (versus B) on the eXMeritus HWW-1U-1A board",
+	/*  */" && <command-if-true>\n"
+	"hww1u1a_test_cpu_a || <command-if-false>\n"
+);
+
+/* Create a prompt-like string: "uboot@HOSTNAME% " */
+#define PROMPT_PREFIX "uboot@exm"
+#define PROMPT_SUFFIX "% "
+
+/* This function returns a PS1 prompt based on the serial number */
+static char *hww1u1a_prompt;
+const char *hww1u1a_get_ps1(void)
+{
+	unsigned long len, i, j;
+	const char *serialnr;
+
+	/* If our prompt was already set, just use that */
+	if (hww1u1a_prompt)
+		return hww1u1a_prompt;
+
+	/* Use our serial number if present, otherwise "UNPROGRAMMED-[AB]" */
+	serialnr = getenv("serial#");
+	if (!serialnr || !serialnr[0]) {
+		if (hww1u1a_is_cpu_a())
+			serialnr = "999999-XA";
+		else
+			serialnr = "999999-XB";
+	}
+
+	/*
+	 * We will turn the serial number into a hostname by:
+	 *   (A) Delete all non-alphanumerics.
+	 *   (B) Lowercase all letters
+	 *   (C) Prefix "exm"
+	 */
+	for (i = 0, len = 0; serialnr[i]; i++)
+		if (isalnum(serialnr[i]))
+			len++;
+
+	len += sizeof(PROMPT_PREFIX PROMPT_SUFFIX); /* Includes NUL */
+	hww1u1a_prompt = malloc(len);
+	if (!hww1u1a_prompt)
+		return PROMPT_PREFIX "UNKNOWN(ENOMEM)" PROMPT_SUFFIX;
+
+	/* Now actually fill it in */
+	i = 0;
+
+	/* Handle the prefix */
+	for (j = 0; j < sizeof(PROMPT_PREFIX) - 1; j++)
+		hww1u1a_prompt[i++] = PROMPT_PREFIX[j];
+
+	/* Now the serial# part of the hostname */
+	for (j = 0; serialnr[j]; j++)
+		if (isalnum(serialnr[j]))
+			hww1u1a_prompt[i++] = tolower(serialnr[j]);
+
+	/* Finally the suffix */
+	for (j = 0; j < sizeof(PROMPT_SUFFIX); j++)
+		hww1u1a_prompt[i++] = PROMPT_SUFFIX[j];
+
+	/* This should all have added up, but just in case */
+	hww1u1a_prompt[len - 1] = '\0';
+
+	/* Now we're done */
+	return hww1u1a_prompt;
+}
+
+void pci_init_board(void)
+{
+	fsl_pcie_init_board(0);
+}
+
+int board_early_init_r(void)
+{
+	const unsigned int flashbase = CONFIG_SYS_FLASH_BASE;
+	const u8 flash_esel = find_tlb_idx((void *)flashbase, 1);
+
+	/*
+	 * Remap bootflash region to caching-inhibited
+	 * so that flash can be erased properly.
+	 */
+
+	/* Flush d-cache and invalidate i-cache of any FLASH data */
+	flush_dcache();
+	invalidate_icache();
+
+	/* invalidate existing TLB entry for FLASH */
+	disable_tlb(flash_esel);
+
+	set_tlb(1, flashbase, CONFIG_SYS_FLASH_BASE_PHYS,
+			MAS3_SX|MAS3_SW|MAS3_SR, MAS2_I|MAS2_G,
+			0, flash_esel, BOOKE_PAGESZ_256M, 1);
+
+	return 0;
+}
+
+int board_eth_init(bd_t *bis)
+{
+	struct tsec_info_struct tsec_info[4];
+
+	SET_STD_TSEC_INFO(tsec_info[0], 1);
+	SET_STD_TSEC_INFO(tsec_info[1], 2);
+	SET_STD_TSEC_INFO(tsec_info[2], 3);
+
+	if (hww1u1a_is_cpu_a())
+		tsec_info[2].phyaddr = TSEC3_PHY_ADDR_CPUA;
+	else
+		tsec_info[2].phyaddr = TSEC3_PHY_ADDR_CPUB;
+
+	tsec_eth_init(bis, tsec_info, 3);
+	return pci_eth_init(bis);
+}
+
+void ft_board_setup(void *blob, bd_t *bd)
+{
+	phys_addr_t base;
+	phys_size_t size;
+
+	ft_cpu_setup(blob, bd);
+
+	base = getenv_bootm_low();
+	size = getenv_bootm_size();
+
+	fdt_fixup_memory(blob, (u64)base, (u64)size);
+
+	FT_FSL_PCI_SETUP;
+}
+
+/*
+ * Unfortunately we have a bit of a problem with our reset line wiring on the
+ * HWW-1U-1A boards.
+ *
+ * Specifically the "CPU A" reset controls all of the PCI devices on the
+ * IO/backplane card, which means ethernet and SATA controllers for both
+ * CPU A *and* CPU B.
+ *
+ * This is somewhat of a problem if the CPUs are reset independently... CPU B
+ * might fail to come up because his PCI devices are confused, or CPU B might
+ * die because CPU A just reset the SATA controller for his root filesystem.
+ *
+ * Since in practice the CPUs never need to reboot independently we will go
+ * ahead and require GPIO cross-communication between them before either one
+ * will reset.
+ *
+ * The communication protocol uses GPIOs 4-7 on the PCA9554 GPIO expander
+ * found on I2C-2 at address 0x3f.  CPU A uses 4,5 as TX and 6,7 as RX; for
+ * CPU B we use the exact opposite.
+ *
+ * The GPIOs initialize as inputs during poweron, and are pulled high; the
+ * PCA9554 expanders will maintain valid outputs over soft-reset allowing for
+ * reliable operation.
+ *
+ * To avoid duplication, we will assign names to the GPIOs:
+ *            ,---------------------------------------.
+ *            |  GPIO4  |  GPIO5  |  GPIO6  |  GPIO7  |
+ *  ,---------|=========|=========|=========|=========|
+ *  |  CPU A  |   tx1   |   tx2   |   rx1   |   rx2   |
+ *  |---------|---------|---------|---------|---------|
+ *  |  CPU B  |   rx1   |   rx2   |   tx1   |   tx2   |
+ *  `-------------------------------------------------'
+ *
+ * The state machine for each CPU is below.  In simplistic terms, each CPU
+ * waits until the other has cleared its GPIOs, then cause a tx1 high=>low
+ * transition.  Upon seeing the other CPU do the same, it will perform a
+ * tx2 high=>low transition and again wait.
+ *
+ * Once both CPUs have performed and seen both transitions, they will each
+ * set tx1 high (to create an "invalid" state) and then reset.
+ *
+ *       ,--------------.
+ *       | START/ERROR: |<==-.
+ *  ,-==>|   tx2 = 1;   |   || (rx2 == 0)
+ *  ||   |   tx1 = 1;   | ==-'
+ *  ||   `--------------'
+ *  ||          ||
+ *  ||          ||  (rx2 == 1)
+ *  ||          \/
+ *  ||   ,--------------.
+ *  ||   | WAITING:     |<==-.
+ *  |\== |   tx1 = 0;   |   || (rx1 == 1) && (rx2 == 1)
+ *  ||   |              | ==-'
+ *  ||   `--------------'
+ *  ||          ||
+ *  ||          ||  (rx1 == 0) && (rx2 == 1)
+ *  ||          \/
+ *  ||   ,--------------.
+ *  ||   | BOTH_READY:  |<==-.
+ *  `-== |   tx2 = 0;   |   || (rx1 == 0) && (rx2 == 1)
+ *       |              | ==-'
+ *       `--------------'
+ *              ||
+ *              ||  (rx2 == 0)
+ *              \/
+ *       ,--------------.
+ *       | RESETTING:   |
+ *       |   tx1 = 1;   |
+ *       |   reset();   |
+ *       `--------------'
+ *
+ * While in the "BOTH_READY" state a high-to-low rx2 transition must be
+ * followed within 50ms by activation of the HRESET_REQ signal.  That
+ * state MUST NOT be interrupted by any event other than the observed
+ * event (rx1 == high) && (rx2 == high).
+ */
+
+/*
+ * These tiny wrappers hardcode the I2C bus and the chip address to make the
+ * state-machine function easier to read and understand.
+ */
+#define PCA9554 CONFIG_SYS_I2C_PCA953X_ADDR
+static int pca9554_set_low(u8 gpio)
+{
+	i2c_set_bus_num(1);
+	return pca953x_set_val(PCA9554, (1U << gpio), 0x00);
+}
+
+static int pca9554_set_high(u8 gpio)
+{
+	i2c_set_bus_num(1);
+	return pca953x_set_val(PCA9554, (1U << gpio), 0xff);
+}
+
+static int pca9554_set_output(u8 gpio)
+{
+	i2c_set_bus_num(1);
+	return pca953x_set_dir(PCA9554, (1U << gpio), 0x00);
+}
+
+static int pca9554_set_input(u8 gpio)
+{
+	i2c_set_bus_num(1);
+	return pca953x_set_dir(PCA9554, (1U << gpio), 0xff);
+}
+
+int __board_restart(void)
+{
+	ccsr_gur_t *gur = (void *)(CONFIG_SYS_MPC85xx_GUTS_ADDR);
+	u8 tx1, tx2, rx1, rx2, rxmask;
+	int ret;
+
+	/* Our state-machine state */
+	enum {
+		START_ERROR = 0,
+		WAITING,
+		BOTH_READY,
+	} state;
+
+	/* Check what CPU we are and configure accordingly. */
+	if (hww1u1a_is_cpu_a()) {
+		tx1 = 4; tx2 = 5; rx1 = 6; rx2 = 7;
+	} else {
+		rx1 = 4; rx2 = 5; tx1 = 6; tx2 = 7;
+	}
+	rxmask = (1U << rx1) | (1U << rx2);
+
+	printf("RESET: Preparing...");
+
+	/* First clear the "INVERT" setting on all relevant GPIOs */
+	i2c_set_bus_num(1);
+	if (pca953x_set_pol(PCA9554, 0xf0, 0x00) < 0)
+		goto err;
+
+	/* Configure both inputs */
+	if (pca9554_set_input(rx1) < 0)
+		goto err;
+	if (pca9554_set_input(rx2) < 0)
+		goto err;
+
+	/* Make sure the very first thing we do is set the "tx2" GPIO high */
+	if (pca9554_set_high(tx2) < 0)
+		goto err;
+	if (pca9554_set_output(tx2) < 0)
+		goto err;
+
+	/* Now reset our other transmit pin */
+	if (pca9554_set_high(tx1) < 0)
+		goto err;
+	if (pca9554_set_output(tx1) < 0)
+		goto err;
+
+	/* Enter the state machine */
+	state = START_ERROR;
+
+poll_state:
+	i2c_set_bus_num(1);
+	ret = pca953x_get_val(PCA9554);
+	switch (state) {
+	case START_ERROR:
+		if (ctrlc())
+			goto intr;
+
+		if (ret < 0)
+			goto err;
+
+		/* Wait for (rx2 == high) */
+		if (ret & (1U << rx2)) {
+			/* Success! */
+			state = WAITING;
+			goto enter_state;
+		}
+
+		goto poll_state;
+
+	case WAITING:
+		if (ctrlc())
+			goto intr;
+
+		if (ret < 0)
+			goto err;
+
+		/* Wait for (rx1 == low) && (rx2 == high) */
+		ret &= rxmask;
+		if (ret == (1U << rx2)) {
+			/* Success! */
+			state = BOTH_READY;
+			goto enter_state;
+		}
+		if (ret != rxmask) {
+			/* Oops, error */
+			state = START_ERROR;
+			goto enter_state;
+		}
+
+		goto poll_state;
+
+	case BOTH_READY:
+		/*
+		 * Non-interruptable wait for (rx2 == low), however if we
+		 * see (rx1 == high) and (rx2 == high) we should go back
+		 * to START_ERROR.
+		 */
+		if (ret < 0)
+			goto poll_state;
+
+		ret &= rxmask;
+		if (ret == rxmask) {
+			/* Oops, error */
+			state = START_ERROR;
+			goto enter_state;
+		}
+		if (ret & (1U << rx2))
+			goto poll_state;
+
+		/* Success, we have to reset quick now! */
+		(void)pca9554_set_high(tx1);
+
+		/* Turn on the "HRESET_REQ" pin (hard-reset request) */
+		printf("\nRESET: Hardware reset triggered, waiting...\n");
+		out_be32(&gur->rstcr, 0x2);
+		while (1)
+			udelay(10000);
+	}
+
+enter_state:
+	switch (state) {
+	case START_ERROR:
+		printf("\nRESET: Negotiation error!");
+		printf("\nRESET: Starting over...");
+		/* Clear both GPIOs high in the right order */
+		if (pca9554_set_high(tx2) < 0)
+			goto err;
+		if (pca9554_set_high(tx1) < 0)
+			goto err;
+		if (ctrlc())
+			goto intr;
+
+		goto poll_state;
+
+	case WAITING:
+		printf("\nRESET: Waiting for peer...");
+		/* Set tx1 low */
+		if (pca9554_set_low(tx1) < 0)
+			goto err;
+		if (ctrlc())
+			goto intr;
+
+		goto poll_state;
+
+	case BOTH_READY:
+		printf("\nRESET: Both ready, point of no return...");
+		/* Point of no return: set tx2 low */
+		if (pca9554_set_low(tx2) < 0)
+			goto err;
+
+		goto poll_state;
+	}
+
+	/* We only get here through gotos */
+err:
+	printf("\nRESET: I2C communication error!\n");
+intr:
+	printf("\nRESET: Aborted!\n");
+	(void)pca9554_set_high(tx1);
+	(void)pca9554_set_high(tx2);
+	return -1;
+}
diff --git a/board/exmeritus/hww1u1a/law.c b/board/exmeritus/hww1u1a/law.c
new file mode 100644
index 0000000..1281c56
--- /dev/null
+++ b/board/exmeritus/hww1u1a/law.c
@@ -0,0 +1,34 @@ 
+/*
+ * Copyright 2008-2009 Freescale Semiconductor, Inc.
+ *
+ * (C) Copyright 2000
+ * Wolfgang Denk, DENX Software Engineering, wd@denx.de.
+ *
+ * See file CREDITS for list of people who contributed to this
+ * project.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation; either version 2 of
+ * the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston,
+ * MA 02111-1307 USA
+ */
+
+#include <common.h>
+#include <asm/fsl_law.h>
+#include <asm/mmu.h>
+
+struct law_entry law_table[] = {
+	SET_LAW(CONFIG_SYS_FLASH_BASE_PHYS, LAW_SIZE_256M, LAW_TRGT_IF_LBC),
+};
+
+int num_law_entries = ARRAY_SIZE(law_table);
diff --git a/board/exmeritus/hww1u1a/tlb.c b/board/exmeritus/hww1u1a/tlb.c
new file mode 100644
index 0000000..6c65206
--- /dev/null
+++ b/board/exmeritus/hww1u1a/tlb.c
@@ -0,0 +1,106 @@ 
+/*
+ * Copyright 2009-2010 eXMeritus, A Boeing Company
+ * Copyright 2008-2009 Freescale Semiconductor, Inc.
+ *
+ * (C) Copyright 2000
+ * Wolfgang Denk, DENX Software Engineering, wd@denx.de.
+ *
+ * See file CREDITS for list of people who contributed to this
+ * project.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation; either version 2 of
+ * the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston,
+ * MA 02111-1307 USA
+ */
+
+#include <common.h>
+#include <asm/mmu.h>
+
+struct fsl_e_tlb_entry tlb_table[] = {
+	/* TLB 0 - for temp stack in cache */
+	SET_TLB_ENTRY(0,	CONFIG_SYS_INIT_RAM_ADDR      +  0 * 1024,
+				CONFIG_SYS_INIT_RAM_ADDR_PHYS +  0 * 1024,
+				MAS3_SX|MAS3_SW|MAS3_SR, 0,
+				0, 0, BOOKE_PAGESZ_4K, 0),
+
+	SET_TLB_ENTRY(0,	CONFIG_SYS_INIT_RAM_ADDR      +  4 * 1024,
+				CONFIG_SYS_INIT_RAM_ADDR_PHYS +  4 * 1024,
+				MAS3_SX|MAS3_SW|MAS3_SR, 0,
+				0, 0, BOOKE_PAGESZ_4K, 0),
+
+	SET_TLB_ENTRY(0,	CONFIG_SYS_INIT_RAM_ADDR      +  8 * 1024,
+				CONFIG_SYS_INIT_RAM_ADDR_PHYS +  8 * 1024,
+				MAS3_SX|MAS3_SW|MAS3_SR, 0,
+				0, 0, BOOKE_PAGESZ_4K, 0),
+
+	SET_TLB_ENTRY(0,	CONFIG_SYS_INIT_RAM_ADDR      + 12 * 1024,
+				CONFIG_SYS_INIT_RAM_ADDR_PHYS + 12 * 1024,
+				MAS3_SX|MAS3_SW|MAS3_SR, 0,
+				0, 0, BOOKE_PAGESZ_4K, 0),
+
+	/* TLB 1 */
+	/* *I*** - Boot page */
+	SET_TLB_ENTRY(1,	CONFIG_BPTR_VIRT_ADDR,
+				CONFIG_BPTR_VIRT_ADDR,
+				MAS3_SX|MAS3_SW|MAS3_SR, MAS2_I|MAS2_G,
+				0, 0, BOOKE_PAGESZ_4K, 1),
+
+	/* *I*G* - CCSRBAR */
+	SET_TLB_ENTRY(1,	CONFIG_SYS_CCSRBAR,
+				CONFIG_SYS_CCSRBAR_PHYS,
+				MAS3_SX|MAS3_SW|MAS3_SR, MAS2_I|MAS2_G,
+				0, 1, BOOKE_PAGESZ_1M, 1),
+
+	/*
+	 * W**G* - FLASH (Will be *I*G* after relocation to RAM)
+	 *
+	 * This maps both SPI FLASH chips (128MByte per chip)
+	 */
+	SET_TLB_ENTRY(1,	CONFIG_SYS_FLASH_BASE,
+				CONFIG_SYS_FLASH_BASE_PHYS,
+				MAS3_SX|MAS3_SR, MAS2_W|MAS2_G,
+				0, 2, BOOKE_PAGESZ_256M, 1),
+
+	/*
+	 * *I*G* - PCI memory
+	 *
+	 * We have 1.5GB total PCI-E memory space to map and we want to use
+	 * the minimum possible number of TLB entries.  Since Book-E TLB
+	 * entries are sized in powers of 4, we use 1GB + 256MB + 256MB.
+	 */
+	SET_TLB_ENTRY(1,	CONFIG_SYS_PCIE3_MEM_VIRT,
+				CONFIG_SYS_PCIE3_MEM_PHYS,
+				MAS3_SX|MAS3_SW|MAS3_SR, MAS2_I|MAS2_G,
+				0, 3, BOOKE_PAGESZ_1G, 1),
+	SET_TLB_ENTRY(1,	CONFIG_SYS_PCIE3_MEM_VIRT + 0x40000000,
+				CONFIG_SYS_PCIE3_MEM_PHYS + 0x40000000,
+				MAS3_SX|MAS3_SW|MAS3_SR, MAS2_I|MAS2_G,
+				0, 4, BOOKE_PAGESZ_256M, 1),
+	SET_TLB_ENTRY(1,	CONFIG_SYS_PCIE3_MEM_VIRT + 0x50000000,
+				CONFIG_SYS_PCIE3_MEM_PHYS + 0x50000000,
+				MAS3_SX|MAS3_SW|MAS3_SR, MAS2_I|MAS2_G,
+				0, 5, BOOKE_PAGESZ_256M, 1),
+
+	/*
+	 * *I*G* - PCI I/O
+	 *
+	 * This one entry covers all 3 64k PCI-E I/O windows
+	 */
+	SET_TLB_ENTRY(1,	CONFIG_SYS_PCIE3_IO_VIRT,
+				CONFIG_SYS_PCIE3_IO_PHYS,
+				MAS3_SX|MAS3_SW|MAS3_SR, MAS2_I|MAS2_G,
+				0, 6, BOOKE_PAGESZ_256K, 1),
+};
+
+int num_tlb_entries = ARRAY_SIZE(tlb_table);
diff --git a/boards.cfg b/boards.cfg
index 45c3102..743e016 100644
--- a/boards.cfg
+++ b/boards.cfg
@@ -487,6 +487,8 @@  sbc8560                      powerpc     mpc85xx     sbc8560             -
 sbc8560_33                   powerpc     mpc85xx     sbc8560             -              -           sbc8560
 sbc8560_66                   powerpc     mpc85xx     sbc8560             -              -           sbc8560
 socrates                     powerpc     mpc85xx     socrates
+HWW1U1A                      powerpc     mpc85xx     hww1u1a             exmeritus
+HWW1U1A_36BIT                powerpc     mpc85xx     hww1u1a             exmeritus      -           HWW1U1A:36BIT
 MPC8536DS                    powerpc     mpc85xx     mpc8536ds           freescale      -           MPC8536DS
 MPC8536DS_36BIT              powerpc     mpc85xx     mpc8536ds           freescale      -           MPC8536DS:36BIT
 MPC8536DS_NAND               powerpc     mpc85xx     mpc8536ds           freescale      -           MPC8536DS:NAND
diff --git a/include/configs/HWW1U1A.h b/include/configs/HWW1U1A.h
new file mode 100644
index 0000000..8472802
--- /dev/null
+++ b/include/configs/HWW1U1A.h
@@ -0,0 +1,474 @@ 
+/*
+ * Copyright 2009-2010 eXMeritus, A Boeing Company
+ *
+ * See file CREDITS for list of people who contributed to this
+ * project.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation; either version 2 of
+ * the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston,
+ * MA 02111-1307 USA
+ */
+
+/*
+ * HardwareWall HWW-1U-1A airborne unit configuration file
+ */
+#ifndef __CONFIG_H
+#define __CONFIG_H
+
+/* High-level system configuration options */
+#define CONFIG_BOOKE		/* Power/PowerPC Book-E			*/
+#define CONFIG_E500		/* e500 (Power ISA v2.03 with SPE)	*/
+#define CONFIG_MPC85xx		/* MPC8540/60/55/41/48 family		*/
+#define CONFIG_FSL_ELBC		/* FreeScale Enhanced LocalBus Cntlr	*/
+#define CONFIG_FSL_LAW		/* FreeScale Local Access Window	*/
+#define CONFIG_P2020		/* FreeScale P2020			*/
+#define CONFIG_HWW1U1A		/* eXMeritus HardwareWall HWW-1U-1A	*/
+#define CONFIG_MP		/* Multiprocessing support		*/
+#define CONFIG_HWCONFIG		/* Use hwconfig from environment	*/
+
+#define CONFIG_L2_CACHE			/* L2 cache enabled		*/
+#define CONFIG_BTB			/* Branch predition enabled	*/
+
+#define CONFIG_PANIC_HANG		/* No board reset on panic	*/
+#define CONFIG_BOARD_EARLY_INIT_R	/* Call board_early_init_r()	*/
+#define CONFIG_CMD_REGINFO		/* Dump various CPU regs	*/
+
+/*
+ * Allow the use of 36-bit physical addresses in "HWW_1U_1A_36BIT" mode.
+ * NOTE: This is only very minimally tested and has known compatibility
+ * issues with the existing kernel images.
+ */
+#define CONFIG_ENABLE_36BIT_PHYS
+#ifdef CONFIG_36BIT
+# define CONFIG_PHYS_64BIT
+# define CONFIG_ADDR_MAP
+# define CONFIG_SYS_NUM_ADDR_MAP 16 /* Number of entries in TLB1 */
+#endif
+
+/* Reserve plenty of RAM for malloc (we have 2GB+) */
+#define CONFIG_SYS_MALLOC_LEN (1024 * 1024)
+
+/* How much L2 cache do we map so we can use it as RAM */
+#define CONFIG_SYS_INIT_RAM_LOCK
+#define CONFIG_SYS_INIT_RAM_SIZE 0x00004000
+
+/* This is our temporary global data area just above the stack */
+#define CONFIG_SYS_GBL_DATA_OFFSET \
+	(CONFIG_SYS_INIT_RAM_SIZE - GENERATED_GBL_DATA_SIZE)
+
+/* The stack grows down from the global data area */
+#define CONFIG_SYS_INIT_SP_OFFSET CONFIG_SYS_GBL_DATA_OFFSET
+
+/* Enable IRQs and watchdog with a 1000Hz system decrementer */
+#define CONFIG_CMD_IRQ
+#define CONFIG_SYS_HZ 1000
+
+
+/* -------------------------------------------------------------------- */
+
+/*
+ * Clock crystal configuration:
+ *  (1) SYS: 66.666MHz +/- 50ppm (drives CPU/PCI/DDR)
+ *  (2) CCB: Multiplier from SYS_CLK
+ *  (3) RTC: 25.000MHz +/- 50ppm (sampled against CCB clock)
+ */
+#define CONFIG_SYS_CLK_FREQ 66666000/*Hz*/
+#define CONFIG_DDR_CLK_FREQ 66666000/*Hz*/
+
+
+/* -------------------------------------------------------------------- */
+
+/*
+ * Memory map
+ *
+ * 0x0000_0000  0x7fff_ffff    2G  DDR2 ECC SDRAM
+ * 0x8000_0000  0x9fff_ffff  512M  PCI-E Bus 1
+ * 0xa000_0000  0xbfff_ffff  512M  PCI-E Bus 2 (unused)
+ * 0xc000_0000  0xdfff_ffff  512M  PCI-E Bus 3
+ * 0xe000_0000  0xe7ff_ffff  128M  Spansion FLASH
+ * 0xe800_0000  0xefff_ffff  128M  Spansion FLASH
+ * 0xffd0_0000  0xffd0_3fff   16K  L1 boot stack (TLB0)
+ * 0xffe0_0000  0xffef_ffff    1M  CCSR
+ * 0xffe0_5000  0xffe0_5fff    4K    Enhanced LocalBus Controller
+ */
+
+/* Virtual Memory Map */
+#define CONFIG_SYS_DDR_SDRAM_BASE	0x00000000
+#define CONFIG_SYS_SDRAM_BASE		0x00000000
+#define CONFIG_SYS_PCIE3_MEM_VIRT	0x80000000
+#define CONFIG_SYS_PCIE2_MEM_VIRT	0xa0000000
+#define CONFIG_SYS_PCIE1_MEM_VIRT	0xc0000000
+#define CONFIG_SYS_FLASH_BASE		0xe0000000
+#define CONFIG_SYS_CCSRBAR_DEFAULT	0xff700000 /* CCSRBAR @ reset */
+#define CONFIG_SYS_PCIE3_IO_VIRT	0xffc00000
+#define CONFIG_SYS_PCIE2_IO_VIRT	0xffc10000
+#define CONFIG_SYS_PCIE1_IO_VIRT	0xffc20000
+#define CONFIG_SYS_INIT_RAM_ADDR	0xffd00000
+#define CONFIG_SYS_CCSRBAR		0xffe00000 /* CCSRBAR @ runtime */
+
+#define CONFIG_SYS_PCIE3_MEM_SIZE	0x20000000 /* 512M */
+#define CONFIG_SYS_PCIE2_MEM_SIZE	0x20000000 /* 512M */
+#define CONFIG_SYS_PCIE1_MEM_SIZE	0x20000000 /* 512M */
+#define CONFIG_SYS_PCIE3_IO_SIZE	0x00010000 /* 64k */
+#define CONFIG_SYS_PCIE2_IO_SIZE	0x00010000 /* 64k */
+#define CONFIG_SYS_PCIE1_IO_SIZE	0x00010000 /* 64k */
+
+/* Physical Memory Map */
+#ifdef CONFIG_PHYS_64BIT
+#define CONFIG_SYS_PCIE3_MEM_PHYS     0xc00000000ull
+#define CONFIG_SYS_PCIE2_MEM_PHYS     0xc20000000ull
+#define CONFIG_SYS_PCIE1_MEM_PHYS     0xc40000000ull
+#define CONFIG_SYS_FLASH_BASE_PHYS    0xfe0000000ull
+#define CONFIG_SYS_PCIE3_IO_PHYS      0xfffc00000ull
+#define CONFIG_SYS_PCIE2_IO_PHYS      0xfffc10000ull
+#define CONFIG_SYS_PCIE1_IO_PHYS      0xfffc20000ull
+#define CONFIG_SYS_INIT_RAM_ADDR_PHYS 0xfffd00000ull
+#define CONFIG_SYS_INIT_RAM_ADDR_PHYS_HIGH 0xf		/* for ASM code */
+#define CONFIG_SYS_INIT_RAM_ADDR_PHYS_LOW  0xffd00000	/* for ASM code */
+#define CONFIG_SYS_CCSRBAR_PHYS       0xfffe00000ull
+#else
+#define CONFIG_SYS_PCIE3_MEM_PHYS      0x80000000ul
+#define CONFIG_SYS_PCIE2_MEM_PHYS      0xa0000000ul
+#define CONFIG_SYS_PCIE1_MEM_PHYS      0xc0000000ul
+#define CONFIG_SYS_FLASH_BASE_PHYS     0xe0000000ul
+#define CONFIG_SYS_PCIE3_IO_PHYS       0xffc00000ul
+#define CONFIG_SYS_PCIE2_IO_PHYS       0xffc10000ul
+#define CONFIG_SYS_PCIE1_IO_PHYS       0xffc20000ul
+#define CONFIG_SYS_INIT_RAM_ADDR_PHYS  0xffd00000ul
+#define CONFIG_SYS_INIT_RAM_ADDR_PHYS_HIGH 0x0		/* for ASM code */
+#define CONFIG_SYS_INIT_RAM_ADDR_PHYS_LOW  0xffd00000	/* for ASM code */
+#define CONFIG_SYS_CCSRBAR_PHYS        0xffe00000ul
+#endif
+
+/* Legacy code needs this */
+#define CONFIG_SYS_IMMR CONFIG_SYS_CCSRBAR
+
+
+/* -------------------------------------------------------------------- */
+
+/* U-Boot image (MONITOR_BASE == TEXT_BASE) */
+#define CONFIG_RESET_VECTOR_ADDRESS	0xeffffffc /* Top address in flash */
+#define CONFIG_SYS_TEXT_BASE		0xeff80000 /* Start of U-Boot image */
+#define CONFIG_SYS_MONITOR_BASE		CONFIG_SYS_TEXT_BASE
+#define CONFIG_SYS_MONITOR_LEN		0x80000 /* 512kB (4 flash sectors) */
+
+/*
+ * U-Boot Environment Image:  The two sectors immediately below U-Boot
+ * form the U-Boot environment (regular and redundant).
+ */
+#define CONFIG_ENV_IS_IN_FLASH	/* The environment image is stored in FLASH */
+#define CONFIG_ENV_OVERWRITE	/* Allow "protected" variables to be erased */
+#define CONFIG_ENV_SECT_SIZE	0x20000 /* 128kB (1 flash sector) */
+#define CONFIG_ENV_ADDR        (CONFIG_SYS_MONITOR_BASE - CONFIG_ENV_SECT_SIZE)
+#define CONFIG_ENV_ADDR_REDUND (CONFIG_ENV_ADDR         - CONFIG_ENV_SECT_SIZE)
+
+/* Only use 8kB of each environment sector for data */
+#define CONFIG_ENV_SIZE		0x2000 /* 8kB */
+#define CONFIG_ENV_SIZE_REDUND	0x2000 /* 8kB */
+
+
+/* -------------------------------------------------------------------- */
+
+/* Serial Console Configuration */
+#define CONFIG_CONS_INDEX 1
+#define CONFIG_SYS_NS16550
+#define CONFIG_SYS_NS16550_SERIAL
+#define CONFIG_SYS_NS16550_REG_SIZE 1
+#define CONFIG_SYS_NS16550_CLK get_bus_freq(0)
+
+#define CONFIG_BAUDRATE 115200
+#define CONFIG_SYS_BAUDRATE_TABLE \
+	{300, 600, 1200, 2400, 4800, 9600, 19200, 38400, 115200}
+
+#define CONFIG_SYS_NS16550_COM1	(CONFIG_SYS_CCSRBAR+0x4500)
+#define CONFIG_SYS_NS16550_COM2	(CONFIG_SYS_CCSRBAR+0x4600)
+
+/* Echo back characters received during a serial download */
+#define CONFIG_LOADS_ECHO
+
+/* Allow a serial-download to temporarily change baud */
+#define CONFIG_SYS_LOADS_BAUD_CHANGE
+
+
+/* -------------------------------------------------------------------- */
+
+/* PCI and PCI-Express Support */
+#define CONFIG_PCI		/* Enable PCI/PCIE			*/
+#define CONFIG_PCI_PNP		/* Scan PCI busses			*/
+#define CONFIG_CMD_PCI		/* Enable the "pci" command		*/
+#define CONFIG_FSL_PCI_INIT	/* Common FreeScale PCI initialization	*/
+#define CONFIG_FSL_PCIE_RESET	/* We have PCI-E reset errata		*/
+#define CONFIG_SYS_PCI_64BIT	/* PCI resources are 64-bit		*/
+#define CONFIG_PCI_SCAN_SHOW	/* Display PCI scan during boot		*/
+
+/* Enable 2 of the 3 PCI-E controllers */
+#define CONFIG_PCIE3
+#undef  CONFIG_PCIE2
+#define CONFIG_PCIE1
+
+/* Display human-readable names when initializing */
+#define CONFIG_SYS_PCIE3_NAME "Intel 82571EB"
+#define CONFIG_SYS_PCIE2_NAME "Unused"
+#define CONFIG_SYS_PCIE1_NAME "Silicon Image SIL3531"
+
+/*
+ * PCI bus addresses
+ * Memory space is mapped 1-1, but I/O space must start from 0.
+ */
+#define CONFIG_SYS_PCIE3_IO_BUS  0x00000000
+#define CONFIG_SYS_PCIE2_IO_BUS  0x00000000
+#define CONFIG_SYS_PCIE1_IO_BUS  0x00000000
+#ifdef CONFIG_PHYS_64BIT
+#define CONFIG_SYS_PCIE3_MEM_BUS 0xe0000000
+#define CONFIG_SYS_PCIE2_MEM_BUS 0xe0000000
+#define CONFIG_SYS_PCIE1_MEM_BUS 0xe0000000
+#else
+#define CONFIG_SYS_PCIE3_MEM_BUS 0x80000000
+#define CONFIG_SYS_PCIE2_MEM_BUS 0xa0000000
+#define CONFIG_SYS_PCIE1_MEM_BUS 0xc0000000
+#endif
+
+
+/* -------------------------------------------------------------------- */
+
+/* Generic FreeScale hardware I2C support */
+#define CONFIG_HARD_I2C
+#define CONFIG_FSL_I2C
+#define CONFIG_CMD_I2C
+#define CONFIG_I2C_MULTI_BUS
+#define CONFIG_SYS_I2C_OFFSET  0x3000
+#define CONFIG_SYS_I2C2_OFFSET 0x3100
+
+/* I2C bus configuration */
+#define CONFIG_SYS_I2C_SPEED 400000
+#define CONFIG_SYS_I2C_SLAVE 0x7F
+
+/* DDR2 SO-RDIMM SPD EEPROM is at I2C0-0x51 */
+#define CONFIG_SYS_SPD_BUS_NUM 0
+#define SPD_EEPROM_ADDRESS 0x51
+
+/* DS1339 RTC is at I2C0-0x68 (I know it says "DS1337", it's a DS1339) */
+#define CONFIG_CMD_DATE
+#define CONFIG_RTC_DS1337
+#define CONFIG_SYS_RTC_BUS_NUM 0
+#define CONFIG_SYS_I2C_RTC_ADDR	0x68
+/* Turn off RTC square-wave output to save battery */
+#define CONFIG_SYS_RTC_DS1337_NOOSC
+
+/* PCA9554 is at I2C1-0x3f (I know it says "PCA953X", it's a PCA9554) */
+#define CONFIG_PCA953X
+#define CONFIG_CMD_PCA953X
+#define CONFIG_CMD_PCA953X_INFO
+#define CONFIG_SYS_I2C_PCA953X_ADDR 0x3f
+
+
+/* -------------------------------------------------------------------- */
+
+/* FreeScale DDR2/3 SDRAM Controller */
+#define CONFIG_FSL_DDR2		/* Our SDRAM slot is DDR2		*/
+#define CONFIG_DDR_ECC		/* Enable ECC by default		*/
+#define CONFIG_DDR_SPD		/* Detect DDR config from SPD EEPROM	*/
+#define CONFIG_SPD_EEPROM	/* ...why 2 config variables for this?	*/
+#define CONFIG_VERY_BIG_RAM	/* Allow 2GB+ of RAM			*/
+#define CONFIG_CMD_SDRAM
+
+/* Standard P2020 DDR controller parameters */
+#define CONFIG_NUM_DDR_CONTROLLERS 1
+#define CONFIG_DIMM_SLOTS_PER_CTLR 1
+#define CONFIG_CHIP_SELECTS_PER_CTRL 2
+
+/* Make sure to tell the DDR controller to preinitialze all of RAM */
+#define CONFIG_MEM_INIT_VALUE 0xDEADBEEF
+#define CONFIG_ECC_INIT_VIA_DDRCONTROLLER
+
+/* Enable the U-Boot "memory test" */
+#define CONFIG_SYS_MEMTEST_START 0x00000000
+#define CONFIG_SYS_MEMTEST_END   0x7fffffff
+
+
+/* -------------------------------------------------------------------- */
+
+/* FLASH Memory Configuration (2x 128MB SPANSION FLASH) */
+#define CONFIG_FLASH_CFI_DRIVER
+#define CONFIG_SYS_FLASH_CFI
+#define CONFIG_SYS_FLASH_EMPTY_INFO
+#define CONFIG_SYS_FLASH_AMD_CHECK_DQ7
+
+/* Flash banks (2x 128MB) */
+#define FLASH0_PHYS (CONFIG_SYS_FLASH_BASE_PHYS + 0x8000000ull)
+#define FLASH1_PHYS (CONFIG_SYS_FLASH_BASE_PHYS + 0x0000000ull)
+#define CONFIG_SYS_MAX_FLASH_BANKS 2
+#define CONFIG_SYS_MAX_FLASH_SECT 1024
+#define CONFIG_SYS_FLASH_BANKS_LIST { FLASH0_PHYS, FLASH1_PHYS }
+
+/*
+ * Flash access modes and timings (values are the defaults after a RESET).
+ *
+ * NOTE: These could probably be optimized but are more than sufficient for
+ * this particular system for the moment.
+ */
+#define FLASH_BRx (BR_PS_16 | BR_MS_GPCM | BR_V)
+#define FLASH_ORx (OR_GPCM_CSNT | OR_GPCM_ACS_DIV2 | OR_GPCM_XACS \
+		| OR_GPCM_SCY_15 | OR_GPCM_TRLX | OR_GPCM_EHTR | OR_GPCM_EAD)
+
+/* Configure both flash banks */
+#define CONFIG_SYS_BR0_PRELIM (FLASH_BRx | BR_PHYS_ADDR(FLASH0_PHYS))
+#define CONFIG_SYS_BR1_PRELIM (FLASH_BRx | BR_PHYS_ADDR(FLASH1_PHYS))
+#define CONFIG_SYS_OR0_PRELIM (FLASH_ORx | OR_AM_128MB)
+#define CONFIG_SYS_OR1_PRELIM (FLASH_ORx | OR_AM_128MB)
+
+/* Flash timeouts (in ms) */
+#define CONFIG_SYS_FLASH_ERASE_TOUT 60000UL /* Erase (60s)	*/
+#define CONFIG_SYS_FLASH_WRITE_TOUT   500UL /* Write (0.5s)	*/
+
+/* Quiet flash testing */
+#define CONFIG_SYS_FLASH_QUIET_TEST
+
+/* Make program/erase count down from 45/5 (9....8....7....) */
+#define CONFIG_FLASH_SHOW_PROGRESS 45
+
+
+/* -------------------------------------------------------------------- */
+
+/* Ethernet Device Support */
+#define CONFIG_NET_MULTI		/* We have multiple devices	*/
+#define CONFIG_MII			/* Enable MII PHY code		*/
+#define CONFIG_MII_DEFAULT_TSEC		/* ??? Copied from P2020DS	*/
+#define CONFIG_PHY_GIGE			/* Support Gigabit PHYs		*/
+#define CONFIG_ETHPRIME "e1000#0"	/* Default to external ports	*/
+
+/* Turn on various helpful networking commands */
+#define CONFIG_CMD_DHCP
+#define CONFIG_CMD_MII
+#define CONFIG_CMD_NET
+#define CONFIG_CMD_PING
+
+/* On-chip FreeScale P2020 "tsec" Ethernet (oneway fibers and peer) */
+#define CONFIG_TSEC_ENET
+#define CONFIG_TSEC1
+#define CONFIG_TSEC2
+#define CONFIG_TSEC3
+#define CONFIG_TSEC1_NAME "owt0"
+#define CONFIG_TSEC2_NAME "owt1"
+#define CONFIG_TSEC3_NAME "peer"
+#define TSEC1_FLAGS (TSEC_GIGABIT | TSEC_REDUCED)
+#define TSEC2_FLAGS (TSEC_GIGABIT | TSEC_REDUCED)
+#define TSEC3_FLAGS (TSEC_GIGABIT | TSEC_REDUCED)
+#define TSEC1_PHYIDX 0
+#define TSEC2_PHYIDX 0
+#define TSEC3_PHYIDX 0
+#define TSEC1_PHY_ADDR      2
+#define TSEC2_PHY_ADDR      3
+#define TSEC3_PHY_ADDR      4
+#define TSEC3_PHY_ADDR_CPUA 4
+#define TSEC3_PHY_ADDR_CPUB 5
+
+/* PCI-E dual-port E1000 (external ethernet ports) */
+#define CONFIG_E1000
+#define CONFIG_E1000_SPI
+#define CONFIG_CMD_E1000
+
+/* We need the SPI infrastructure to poke the E1000's EEPROM */
+#define CONFIG_SPI
+#define CONFIG_SPI_X
+#define CONFIG_CMD_SPI
+#define MAX_SPI_BYTES 32
+
+
+/* -------------------------------------------------------------------- */
+
+/* USB Thumbdrive Device Support */
+#define CONFIG_USB_EHCI
+#define CONFIG_USB_EHCI_FSL
+#define CONFIG_USB_STORAGE
+#define CONFIG_EHCI_HCD_INIT_AFTER_RESET
+#define CONFIG_CMD_USB
+
+/* Partition and Filesystem support */
+#define CONFIG_DOS_PARTITION
+#define CONFIG_CMD_EXT2
+#define CONFIG_CMD_FAT
+
+
+/* -------------------------------------------------------------------- */
+
+/* Command line configuration. */
+#define CONFIG_CMDLINE_EDITING		/* Enable command editing	*/
+#define CONFIG_COMMAND_HISTORY		/* Enable command history	*/
+#define CONFIG_AUTO_COMPLETE		/* Enable command completion	*/
+#define CONFIG_SYS_LONGHELP		/* Enable detailed command help	*/
+#define CONFIG_SYS_MAXARGS 128		/* Up to 128 command-line args	*/
+#define CONFIG_SYS_PBSIZE 8192		/* Allow up to 8k printed lines	*/
+#define CONFIG_SYS_CBSIZE 4096		/* Allow up to 4k command lines	*/
+#define CONFIG_SYS_BARGSIZE 4096	/* Allow up to 4k boot args	*/
+#define CONFIG_SYS_HUSH_PARSER		/* Enable a fancier shell	*/
+#define CONFIG_SYS_PROMPT_HUSH_PS2 "> "	/* Command-line continuation	*/
+
+/* A little extra magic here for the prompt */
+#define CONFIG_SYS_PROMPT hww1u1a_get_ps1()
+#ifndef __ASSEMBLY__
+const char *hww1u1a_get_ps1(void);
+#endif
+
+/* Include a bunch of default commands we probably want */
+#include <config_cmd_default.h>
+
+/* Other helpful shell-like commands */
+#define CONFIG_MD5
+#define CONFIG_SHA1
+#define CONFIG_CMD_MD5SUM
+#define CONFIG_CMD_SHA1
+#define CONFIG_CMD_ASKENV
+#define CONFIG_CMD_SETEXPR
+
+
+/* -------------------------------------------------------------------- */
+
+/* Image manipulation and booting */
+
+/* We use the OpenFirmware-esque "Flattened Device Tree" */
+#define CONFIG_OF_LIBFDT
+#define CONFIG_OF_BOARD_SETUP
+#define CONFIG_OF_STDOUT_VIA_ALIAS
+
+/*
+ * For booting Linux, the board info and command line data
+ * have to be in the first 16 MB of memory, since this is
+ * the maximum mapped by the Linux kernel during initialization.
+ */
+#define CONFIG_CMD_ELF
+#define CONFIG_SYS_BOOTMAPSZ 0x1000000	/* Maximum kernel memory map	*/
+#define CONFIG_SYS_BOOTM_LEN 0x1000000	/* Maximum kernel size of 16MB	*/
+
+/* This is the default address for commands with an optional address arg */
+#define CONFIG_LOADADDR		  100000
+#define CONFIG_SYS_LOAD_ADDR	0x100000
+
+#define CONFIG_BOOTDELAY 20
+#define CONFIG_BOOTCOMMAND "echo Not yet flashed"
+#define CONFIG_BOOTARGS ""
+#define CONFIG_BOOTARGS_DYNAMIC "console=ttyS0,${baudrate}n1"
+
+/* Extra environment parameters */
+#define CONFIG_EXTRA_ENV_SETTINGS					\
+	"preboot=setenv bootargs \"${bootargs} "CONFIG_BOOTARGS_DYNAMIC"\"\0" \
+	"perf_mode=stable\0"						\
+	"memctl_intlv_ctl=2\0"						\
+	"flkernel=0xe8020000\0"						\
+	"flinitramfs=0xe8800000\0"					\
+	"fldevicetree=0xeff20000\0"					\
+	"flbootm=bootm ${flkernel} ${flinitramfs} ${fldevicetree}\0"	\
+	"flboot=run preboot; run flbootm\0"
+
+#endif	/* __CONFIG_H */