r913 - clients/lib clients/lib/x86 cpu/x86/pc cpu/x86/pc/neptune

Author: wmb
Date: 2008-09-02 07:42:53 +0200 (Tue, 02 Sep 2008)
New Revision: 913

Modified:
   clients/lib/callofw.c
   clients/lib/x86/makefile
   cpu/x86/pc/neptune/fw.bth
   cpu/x86/pc/reenter.fth
Log:
Neptune - allow rerunning ispVME without rebooting.

Modified: clients/lib/callofw.c
===================================================================
--- clients/lib/callofw.c	2008-09-02 02:45:16 UTC (rev 912)
+++ clients/lib/callofw.c	2008-09-02 05:42:53 UTC (rev 913)
@@ -6,14 +6,10 @@
  */

#include <stdarg.h>
-#include <asm/callofw.h>
-#include <linux/types.h>
-#include <linux/spinlock.h>
+typedef unsigned long u32;

-u32 (*call_firmware)(u32 *);
+extern u32 call_firmware(u32 *);

-static DEFINE_SPINLOCK(prom_lock);
-
 #define MAXARGS 20
 int callofw(char *name, int numargs, int numres, ...)
 {
@@ -24,9 +20,6 @@
    int *intp;
    unsigned long flags;

-	if (call_firmware == NULL)
-		return -1;
-
    argarray[0] = (u32)name;
    argarray[1] = numargs;
    argarray[2] = numres;
@@ -38,9 +31,7 @@
    while (numargs--)
    	argarray[argnum++] = va_arg(ap, int);

-	spin_lock_irqsave(&prom_lock, flags);
    retval = call_firmware(argarray);
-	spin_unlock_irqrestore(&prom_lock, flags);

if (retval == 0) {
    	while (numres--) {

Modified: clients/lib/x86/makefile
===================================================================
--- clients/lib/x86/makefile	2008-09-02 02:45:16 UTC (rev 912)
+++ clients/lib/x86/makefile	2008-09-02 05:42:53 UTC (rev 913)
@@ -3,51 +3,23 @@
 # -fno-builtin has the effect of suppressing some warnings about
 # functions that conflict with gcc builtins
 CC=gcc
-CFLAGS=-g -m32 -fno-builtin -fno-stack-protector
+CFLAGS=-g -m32 -fno-builtin -fno-stack-limit -fno-stack-protector

all: libobp.a hello hello.elf start.o

# Create a library file containing all the library routines

-libobp.a: lib.o printf.o wrappers.o malloc.o strings.o printf.o debug.o main.o intprop.o regprop.o strprop.o mem.o
-	ar rcv libobp.a lib.o malloc.o wrappers.o strings.o printf.o debug.o main.o intprop.o regprop.o strprop.o mem.o
+OBJS=lib.o printf.o wrappers.o malloc.o strings.o printf.o debug.o main.o intprop.o regprop.o strprop.o mem.o callofw.o
+
+libobp.a: ${OBJS}
+	ar rcv libobp.a ${OBJS}
    ranlib libobp.a

# Build machine-independent library routines

-main.o: ../main.c
-	${CC} ${CFLAGS} -c ../main.c
+%.o: ../%.c
+	${CC} ${CFLAGS} -c $<

-lib.o: ../lib.c
-	${CC} ${CFLAGS} -c ../lib.c
-
-printf.o: ../printf.c
-	${CC} ${CFLAGS} -c ../printf.c
-
-debug.o: ../debug.c
-	${CC} ${CFLAGS} -c ../debug.c
-
-strings.o: ../strings.c
-	${CC} ${CFLAGS} -c ../strings.c
-
-mem.o: ../mem.c
-	${CC} ${CFLAGS} -c ../mem.c
-
-intprop.o: ../intprop.c
-	${CC} ${CFLAGS} -c ../intprop.c
-
-regprop.o: ../regprop.c
-	${CC} ${CFLAGS} -c ../regprop.c
-
-strprop.o: ../strprop.c
-	${CC} ${CFLAGS} -c ../strprop.c
-
-wrappers.o: ../wrappers.c
-	${CC} ${CFLAGS} -c ../wrappers.c
-
-malloc.o: ../malloc.c
-	${CC} ${CFLAGS} -c ../malloc.c
-
 # Build processor-specific startup code and call gateway

start.o: start.s

Modified: cpu/x86/pc/neptune/fw.bth
===================================================================
--- cpu/x86/pc/neptune/fw.bth	2008-09-02 02:45:16 UTC (rev 912)
+++ cpu/x86/pc/neptune/fw.bth	2008-09-02 05:42:53 UTC (rev 913)
@@ -165,10 +165,11 @@
 fload ${BP}/ofw/elf/elf.fth
 fload ${BP}/ofw/elf/elfdebug.fth
 [ifdef] virtual-mode
-: (elf-map-in) ( va size -- )  swap 1 -rot  allocate-aligned  drop  ;
-\ Depends on the assumption that physical memory is mapped 1:1 already
-\ : (elf-map-in) ( va size -- )  0 mem-claim  drop  ;
-' (elf-map-in) is elf-map-in
+\ This version dynamically allocates physical memory
+\ : (elf-map-in) ( va size -- )  swap 1 -rot  allocate-aligned  drop  ;
+
+\ This version depends on the assumption that physical memory is mapped 1:1 already
+: (elf-map-in)  ( va size -- )  2drop  ;
 [else]
 : (elf-map-in)  ( va size -- )  2drop  ;
 [then]
@@ -178,6 +179,7 @@
 \ Reboot and re-entry code
 fload ${BP}/cpu/x86/pc/reboot.fth	\ Restart the client program
 fload ${BP}/cpu/x86/pc/reenter.fth	\ Various entries into Forth
+true to client-rerun?

headerless
 [ifdef] virtual-mode
@@ -198,8 +200,9 @@

\ The LPC FLASH chips decode several addresses below the FLASH "memory" array,
 \ for "registers" used for programming and access to general purpose inputs.
-\ The FPGA programming program wants to access the GPI register at virtual
-\ address = physical address, so we pre-claim that range of virtual addresses.
+\ We pre-claim that range of virtual addresses just in case someone needs to
+\ access the LPC device with virtual-address = physical address.  (OFW and
+\ ispVME both use dynamically-mapped virtual addresses for that region.)

: claim-lpc  ( -- )
    (initial-claim)
@@ -404,13 +407,8 @@
 lpc-gpi0-pa value lpc-gpi-pa

: vme ( -- )
-   ." Running Lattice programmer "
-
-   lpc-gpi-pa lpc-gpi-pa h# 100 -1 mmu-map
    \ verbose-cif
    " rom:\isp u:\neptune.vme" $boot
-	
-   lpc-gpi-pa h# 100 mmu-unmap
 ;

fload ${BP}/cpu/x86/pc/neptune/versions.fth

Modified: cpu/x86/pc/reenter.fth
===================================================================
--- cpu/x86/pc/reenter.fth	2008-09-02 02:45:16 UTC (rev 912)
+++ cpu/x86/pc/reenter.fth	2008-09-02 05:42:53 UTC (rev 913)
@@ -28,6 +28,8 @@

\ Temporary hacks until we implement these functions correctly

+0 value client-rerun?
+
 defer user-interface ' quit to user-interface
 also client-services definitions
 \ " reenter.fth: Implement 'exit' client service correctly" ?reminder
@@ -38,6 +40,7 @@
    " restore"  stdout @  ['] $call-method  catch  if  3drop  then
    " restore"  stdin  @  ['] $call-method  catch  if  3drop  then

+   client-rerun? 0= to already-go?
    user-interface
 ;
 : enter  ( -- )  interact  ;