[PATCH] decstation: Document more MB ASIC register bits

[PATCH] decstation: Document more MB ASIC register bits

[Maciej W. Rozycki]

 Document a few more register bits provided by the MB ASIC used on R4000SC
(KN04) and R4400SC (KN05) CPU daughtercards with the DECstation.  

 Reverse-engineered and not documented anywhere else to the best of my
knowledge.  Bit names appended to the last underscore the same as reported
by the firmware in register dumps.

Signed-off-by: Maciej W. Rozycki <macro@linux-mips.org>
---
patch-mips-2.6.23-rc5-20070904-dec-kn4k-mb-int-0
diff -up --recursive --new-file linux-mips-2.6.23-rc5-20070904.macro/include/asm-mips/dec/kn05.h linux-mips-2.6.23-rc5-20070904/include/asm-mips/dec/kn05.h
--- linux-mips-2.6.23-rc5-20070904.macro/include/asm-mips/dec/kn05.h	2005-07-20 05:00:28.000000000 +0000
+++ linux-mips-2.6.23-rc5-20070904/include/asm-mips/dec/kn05.h	2007-10-11 00:11:52.000000000 +0000
@@ -6,7 +6,7 @@
  *	KN04-CA) and DECsystem 5900/260 (KN05) R4k CPU card MB ASIC
  *	definitions.
  *
- *	Copyright (C) 2002, 2003, 2005  Maciej W. Rozycki
+ *	Copyright (C) 2002, 2003, 2005, 2008  Maciej W. Rozycki
  *
  *	This program is free software; you can redistribute it and/or
  *	modify it under the terms of the GNU General Public License
@@ -54,11 +54,11 @@
  */
 #define KN4K_MB_INT_TC		(1<<0)		/* TURBOchannel? */
 #define KN4K_MB_INT_RTC		(1<<1)		/* RTC? */
-#define KN4K_MB_INT_MT		(1<<3)		/* ??? */
+#define KN4K_MB_INT_MT		(1<<3)		/* I/O ASIC cascade */
 
 /*
  * Bits for the MB control & status register.
- * Set to 0x00bf8001 on my system by the ROM.
+ * Set to 0x00bf8001 for KN05 and to 0x003f8000 for KN04 by the firmware.
  */
 #define KN4K_MB_CSR_PF		(1<<0)		/* PreFetching enable? */
 #define KN4K_MB_CSR_F		(1<<1)		/* ??? */
@@ -69,7 +69,8 @@
 #define KN4K_MB_CSR_IM		(1<<13)		/* ??? */
 #define KN4K_MB_CSR_NC		(1<<14)		/* ??? */
 #define KN4K_MB_CSR_EE		(1<<15)		/* (bus) Exception Enable? */
-#define KN4K_MB_CSR_MSK		(0x1f<<16)	/* ??? */
+#define KN4K_MB_CSR_MSK		(0x1f<<16)	/* CPU Int[4:0] mask */
 #define KN4K_MB_CSR_FW		(1<<21)		/* ??? */
+#define KN4K_MB_CSR_W		(1<<31)		/* ??? */
 
 #endif /* __ASM_MIPS_DEC_KN05_H */

Re: [PATCH] decstation: Document more MB ASIC register bits

[Ralf Baechle]

On Fri, Jun 13, 2008 at 12:25:36AM +0100, Maciej W. Rozycki wrote:

>  Document a few more register bits provided by the MB ASIC used on R4000SC
> (KN04) and R4400SC (KN05) CPU daughtercards with the DECstation.  
> 
>  Reverse-engineered and not documented anywhere else to the best of my
> knowledge.  Bit names appended to the last underscore the same as reported
> by the firmware in register dumps.

You realize that documentation the DECstation code endangers the safety of
your job as DECstation maintainer, do you ;-)

Queued for 2.6.27, thanks!

  Ralf

Re: [PATCH] decstation: Document more MB ASIC register bits

[Maciej W. Rozycki]

On Fri, 13 Jun 2008, Ralf Baechle wrote:

> You realize that documentation the DECstation code endangers the safety of
> your job as DECstation maintainer, do you ;-)

 I do, but this is a carefully considered compromise.  I am not bold
enough to assert I shall live indefinitely -- actually the opposite is
quite likely -- so I do my best to provide means for long-term support
regardless of the future of the atoms comprising my body at the moment.

  Maciej

Re: [PATCH] decstation: Document more MB ASIC register bits

[Ralf Baechle]

On Fri, Jun 13, 2008 at 05:43:16PM +0100, Maciej W. Rozycki wrote:

> > You realize that documentation the DECstation code endangers the safety of
> > your job as DECstation maintainer, do you ;-)
> 
>  I do, but this is a carefully considered compromise.  I am not bold
> enough to assert I shall live indefinitely -- actually the opposite is
> quite likely -- so I do my best to provide means for long-term support
> regardless of the future of the atoms comprising my body at the moment.

Don't worry about the protons, they're supposed to have at least a half-
life of 10^35 years.

  Ralf

Re: [PATCH] decstation: Document more MB ASIC register bits

[Maciej W. Rozycki]

On Fri, 13 Jun 2008, Ralf Baechle wrote:

> >  I do, but this is a carefully considered compromise.  I am not bold
> > enough to assert I shall live indefinitely -- actually the opposite is
> > quite likely -- so I do my best to provide means for long-term support
> > regardless of the future of the atoms comprising my body at the moment.
> 
> Don't worry about the protons, they're supposed to have at least a half-
> life of 10^35 years.

 That's reassuring, thanks!

  Maciej