powerpc 476, Little-endian, pte fault

public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed

* powerpc 476, Little-endian, pte fault
@ 2011-10-31  5:30 Santosh Kumar
  2011-10-31  5:51 ` Michael Neuling
  2011-10-31  7:48 ` Shan Hai
  0 siblings, 2 replies; 11+ messages in thread
From: Santosh Kumar @ 2011-10-31  5:30 UTC (permalink / raw)
  To: linux-kernel

KERNEL: linux 2.6.39.4
POWERPC: 476, little endian.

I am trying to get linux 2.6.39.4 up on PPC 476 i have done done
Big-endian to little endian Changes in:
1) bitops header file.
2) while reading the device tree.
3) the PTE read/computed in head_32.S
4) added E bit in the TLB entries.

with all the above changes the kernel_init is done but and inited is mounted.

But while spawning init process the kernel continuously hits pte
faults at address 0x100000fc and never comes out. Please let me know
where i should be looking into.

-
Thanks
Santosh Kumar .A

Vision without Action is a daydream... Action without Vision is a nightmare...

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: powerpc 476, Little-endian, pte fault
  2011-10-31  5:30 powerpc 476, Little-endian, pte fault Santosh Kumar
@ 2011-10-31  5:51 ` Michael Neuling
       [not found]   ` <CAJw04pAh5QDQMkUHbj9=rRPiUCZejimHiydCJFMAoZeSd_ydhA@mail.gmail.com>
  2011-10-31  7:48 ` Shan Hai
  1 sibling, 1 reply; 11+ messages in thread
From: Michael Neuling @ 2011-10-31  5:51 UTC (permalink / raw)
  To: Santosh Kumar; +Cc: linux-kernel, linuxppc-dev, Tony Breeds

Adding linuxppc-dev list to the CC

> KERNEL: linux 2.6.39.4
> POWERPC: 476, little endian.
> 
> I am trying to get linux 2.6.39.4 up on PPC 476 i have done done
> Big-endian to little endian Changes in:

Can you explain what you are trying to do in more detail?  What does "i
have done Big-endian to little endian Changes" mean?

> 1) bitops header file.
> 2) while reading the device tree.
> 3) the PTE read/computed in head_32.S
> 4) added E bit in the TLB entries.
> 
> with all the above changes the kernel_init is done but and inited is
> mounted.

Can you post your patch?

Are you trying to boot the kernel in little endian or just run userspace
in little endian?  

> But while spawning init process the kernel continuously hits pte
> faults at address 0x100000fc and never comes out. Please let me know
> where i should be looking into.

Is this the very first userspace instruction?

Mikey


^ permalink raw reply	[flat|nested] 11+ messages in thread

[parent not found: <CAJw04pAh5QDQMkUHbj9=rRPiUCZejimHiydCJFMAoZeSd_ydhA@mail.gmail.com>]

* Re: powerpc 476, Little-endian, pte fault
       [not found]   ` <CAJw04pAh5QDQMkUHbj9=rRPiUCZejimHiydCJFMAoZeSd_ydhA@mail.gmail.com>
@ 2011-10-31  7:38     ` Santosh Kumar
  2011-10-31  9:49     ` Michael Neuling
  1 sibling, 0 replies; 11+ messages in thread
From: Santosh Kumar @ 2011-10-31  7:38 UTC (permalink / raw)
  To: Michael Neuling; +Cc: linux-kernel, linuxppc-dev, Tony Breeds

This is not the first user-space instruction. While executing this
process kernel has added two TLB entries with TID(process id) 1. while
trying to map 0x100000fc it is raising pte faults .

Santosh Kumar .A

Vision without Action is a daydream... Action without Vision is a nightmare...



On 31 October 2011 12:05, Santosh Kumar <santoshkumar.a@gmail.com> wrote:
> I have built a cross compiler for ppc440 in little endian mode and
> using it to build the kernel.
>
> Yes i am running Linux in Little-Endian. This is the first user space
> process. I wrote the below program and running it as init from
> /sbin/init. I have also set the permissions with chmod +s.
>
> main()
> {
>
> while(1){
> printf("hello world");
> sleep(1);
>  }
> }
>
> I have attached the patch.
>
> -
> Santosh Kumar .A
>
> Vision without Action is a daydream... Action without Vision is a nightmare...
>
>
>
> On 31 October 2011 11:21, Michael Neuling <mikey@neuling.org> wrote:
>> Adding linuxppc-dev list to the CC
>>
>>> KERNEL: linux 2.6.39.4
>>> POWERPC: 476, little endian.
>>>
>>> I am trying to get linux 2.6.39.4 up on PPC 476 i have done done
>>> Big-endian to little endian Changes in:
>>
>> Can you explain what you are trying to do in more detail?  What does "i
>> have done Big-endian to little endian Changes" mean?
>>
>>> 1) bitops header file.
>>> 2) while reading the device tree.
>>> 3) the PTE read/computed in head_32.S
>>> 4) added E bit in the TLB entries.
>>>
>>> with all the above changes the kernel_init is done but and inited is
>>> mounted.
>>
>> Can you post your patch?
>>
>> Are you trying to boot the kernel in little endian or just run userspace
>> in little endian?
>>
>>> But while spawning init process the kernel continuously hits pte
>>> faults at address 0x100000fc and never comes out. Please let me know
>>> where i should be looking into.
>>
>> Is this the very first userspace instruction?
>>
>> Mikey
>>
>>
>

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: powerpc 476, Little-endian, pte fault
       [not found]   ` <CAJw04pAh5QDQMkUHbj9=rRPiUCZejimHiydCJFMAoZeSd_ydhA@mail.gmail.com>
  2011-10-31  7:38     ` Santosh Kumar
@ 2011-10-31  9:49     ` Michael Neuling
  2011-10-31 11:23       ` Benjamin Herrenschmidt
  1 sibling, 1 reply; 11+ messages in thread
From: Michael Neuling @ 2011-10-31  9:49 UTC (permalink / raw)
  To: Santosh Kumar; +Cc: linux-kernel, linuxppc-dev, Tony Breeds, Ian Munsie

> I have built a cross compiler for ppc440 in little endian mode and
> using it to build the kernel.
> 
> Yes i am running Linux in Little-Endian. This is the first user space
> process. I wrote the below program and running it as init from
> /sbin/init. I have also set the permissions with chmod +s.
> 
> main()
> {
> 
> while(1){
> printf("hello world");
> sleep(1);
>  }
> }

Does libc even support little endian on PPC?

> I have attached the patch.

This is a pretty huge patch:

 115 files changed, 44479 insertions(+), 7398 deletions(-)

It seems to include a new platform as well as a bunch of unrelated junk.

I suggest you need to break this down into something more digestible.
Like remove all the junk in the patch.  Then add the support for the new
platform (invader? platform).  Then start looking at little endian.
Unless you do this, it's unlikely anyone here is going to be able to
help.

When you get to the little endian work, you might want to take a look at
this patch series from Ian Munsie:

http://lists.ozlabs.org/pipermail/linuxppc-dev/2010-October/086165.html

Mikey

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: powerpc 476, Little-endian, pte fault
  2011-10-31  9:49     ` Michael Neuling
@ 2011-10-31 11:23       ` Benjamin Herrenschmidt
  2011-11-01  3:02         ` Santosh Kumar
  0 siblings, 1 reply; 11+ messages in thread
From: Benjamin Herrenschmidt @ 2011-10-31 11:23 UTC (permalink / raw)
  To: Michael Neuling; +Cc: Santosh Kumar, linuxppc-dev, Ian Munsie, linux-kernel

On Mon, 2011-10-31 at 20:49 +1100, Michael Neuling wrote:
> > I have built a cross compiler for ppc440 in little endian mode and
> > using it to build the kernel.
> > 
> > Yes i am running Linux in Little-Endian. This is the first user space
> > process. I wrote the below program and running it as init from
> > /sbin/init. I have also set the permissions with chmod +s.
> > 
> > main()
> > {
> > 
> > while(1){
> > printf("hello world");
> > sleep(1);
> >  }
> > }
> 
> Does libc even support little endian on PPC?

Ian did a port a while back for uClibc, is that at least partially based
on it ?

> > I have attached the patch.
> 
> This is a pretty huge patch:
> 
>  115 files changed, 44479 insertions(+), 7398 deletions(-)
> 
> It seems to include a new platform as well as a bunch of unrelated junk.
>
> I suggest you need to break this down into something more digestible.
> Like remove all the junk in the patch.  Then add the support for the new
> platform (invader? platform).  Then start looking at little endian.
> Unless you do this, it's unlikely anyone here is going to be able to
> help.
>
> When you get to the little endian work, you might want to take a look at
> this patch series from Ian Munsie:
> 
> http://lists.ozlabs.org/pipermail/linuxppc-dev/2010-October/086165.html

Right, the new patch should be if possible based on Ian's series or at
least a cleaned / rebased variant of it. Then split in bits so we can
review it properly.

Cheers,
Ben.

> Mikey
> _______________________________________________
> Linuxppc-dev mailing list
> Linuxppc-dev@lists.ozlabs.org
> https://lists.ozlabs.org/listinfo/linuxppc-dev



^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: powerpc 476, Little-endian, pte fault
  2011-10-31 11:23       ` Benjamin Herrenschmidt
@ 2011-11-01  3:02         ` Santosh Kumar
  2011-11-01 13:44           ` Peter Bergner
  0 siblings, 1 reply; 11+ messages in thread
From: Santosh Kumar @ 2011-11-01  3:02 UTC (permalink / raw)
  To: Benjamin Herrenschmidt
  Cc: Michael Neuling, linuxppc-dev, Ian Munsie, linux-kernel

[-- Attachment #1: Type: text/plain, Size: 2302 bytes --]

I have attached the patch with only the relevant changes.

The patch is not based on uLIBC patch.

Glibc doesnt support little endian for PPC, but after minor changes to
makefile i got it working. With the compiler i am using i could get
2.6.31 on ppc440 working. I am using the same compiler as 476 & 440
instruction is almost the same.

Thanks
Santosh Kumar .A

Vision without Action is a daydream... Action without Vision is a nightmare...



On 31 October 2011 16:53, Benjamin Herrenschmidt
<benh@kernel.crashing.org> wrote:
> On Mon, 2011-10-31 at 20:49 +1100, Michael Neuling wrote:
>> > I have built a cross compiler for ppc440 in little endian mode and
>> > using it to build the kernel.
>> >
>> > Yes i am running Linux in Little-Endian. This is the first user space
>> > process. I wrote the below program and running it as init from
>> > /sbin/init. I have also set the permissions with chmod +s.
>> >
>> > main()
>> > {
>> >
>> > while(1){
>> > printf("hello world");
>> > sleep(1);
>> >  }
>> > }
>>
>> Does libc even support little endian on PPC?
>
> Ian did a port a while back for uClibc, is that at least partially based
> on it ?
>
>> > I have attached the patch.
>>
>> This is a pretty huge patch:
>>
>>  115 files changed, 44479 insertions(+), 7398 deletions(-)
>>
>> It seems to include a new platform as well as a bunch of unrelated junk.
>>
>> I suggest you need to break this down into something more digestible.
>> Like remove all the junk in the patch.  Then add the support for the new
>> platform (invader? platform).  Then start looking at little endian.
>> Unless you do this, it's unlikely anyone here is going to be able to
>> help.
>>
>> When you get to the little endian work, you might want to take a look at
>> this patch series from Ian Munsie:
>>
>> http://lists.ozlabs.org/pipermail/linuxppc-dev/2010-October/086165.html
>
> Right, the new patch should be if possible based on Ian's series or at
> least a cleaned / rebased variant of it. Then split in bits so we can
> review it properly.
>
> Cheers,
> Ben.
>
>> Mikey
>> _______________________________________________
>> Linuxppc-dev mailing list
>> Linuxppc-dev@lists.ozlabs.org
>> https://lists.ozlabs.org/listinfo/linuxppc-dev
>
>
>

[-- Attachment #2: linux_ppc_476.patch --]
[-- Type: application/octet-stream, Size: 42386 bytes --]

diff -uNr ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/boot/div64.S ../linux-2.6.39.4/arch/powerpc/boot/div64.S
--- ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/boot/div64.S	2011-08-04 01:13:28.000000000 +0530
+++ ../linux-2.6.39.4/arch/powerpc/boot/div64.S	2011-10-22 14:16:46.616117143 +0530
@@ -17,8 +17,13 @@
 
 	.globl __div64_32
 __div64_32:
+#ifdef CONFIG_INVADER
+	lwz	r5,4(r3)	# get the dividend into r5/r6
+	lwz	r6,0(r3)
+#else
 	lwz	r5,0(r3)	# get the dividend into r5/r6
 	lwz	r6,4(r3)
+#endif
 	cmplw	r5,r4
 	li	r7,0
 	li	r8,0
@@ -53,7 +58,12 @@
 	mullw	r10,r0,r4	# and get the remainder
 	add	r8,r8,r0
 	subf	r6,r10,r6
+#ifdef CONFIG_INVADER
 4:	stw	r7,0(r3)	# return the quotient in *r3
 	stw	r8,4(r3)
+#else
+4:	stw	r7,0(r3)	# return the quotient in *r3
+	stw	r8,4(r3)
+#endif
 	mr	r3,r6		# return the remainder in r3
 	blr
diff -uNr ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/boot/Makefile ../linux-2.6.39.4/arch/powerpc/boot/Makefile
--- ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/boot/Makefile	2011-08-04 01:13:28.000000000 +0530
+++ ../linux-2.6.39.4/arch/powerpc/boot/Makefile	2011-10-22 06:47:05.354137527 +0530
@@ -37,7 +37,7 @@
 
 DTC_FLAGS	?= -p 1024
 
-$(obj)/4xx.o: BOOTCFLAGS += -mcpu=405
+$(obj)/4xx.o: BOOTCFLAGS += -mcpu=440
 $(obj)/ebony.o: BOOTCFLAGS += -mcpu=405
 $(obj)/cuboot-hotfoot.o: BOOTCFLAGS += -mcpu=405
 $(obj)/cuboot-taishan.o: BOOTCFLAGS += -mcpu=405
diff -uNr ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/boot/wrapper ../linux-2.6.39.4/arch/powerpc/boot/wrapper
--- ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/boot/wrapper	2011-08-04 01:13:28.000000000 +0530
+++ ../linux-2.6.39.4/arch/powerpc/boot/wrapper	2011-10-22 06:38:26.286134445 +0530
@@ -38,6 +38,7 @@
 dts=
 cacheit=
 binary=
+gzip=
 gzip=.gz
 
 # cross-compilation prefix
@@ -50,7 +51,6 @@
 object=arch/powerpc/boot
 objbin=$object
 dtc=scripts/dtc/dtc
-
 # directory for working files
 tmpdir=.
 
@@ -260,7 +260,6 @@
 	vmz="$vmz.$$"
     fi
 fi
-
 vmz="$vmz$gzip"
 
 # Extract kernel version information, some platforms want to include
@@ -277,12 +276,13 @@
 case "$platform" in
 uboot)
     rm -f "$ofile"
-    ${MKIMAGE} -A ppc -O linux -T kernel -C gzip -a $membase -e $membase \
+    ${MKIMAGE} -A ppc -O linux -T kernel -a $membase -e $membase \
 	$uboot_version -d "$vmz" "$ofile"
     if [ -z "$cacheit" ]; then
 	rm -f "$vmz"
     fi
     exit 0
+    echo "not exiting"
     ;;
 esac
 
diff -uNr ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/include/asm/bitops.h ../linux-2.6.39.4/arch/powerpc/include/asm/bitops.h
--- ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/include/asm/bitops.h	2011-08-04 01:13:28.000000000 +0530
+++ ../linux-2.6.39.4/arch/powerpc/include/asm/bitops.h	2011-10-22 06:38:25.203141637 +0530
@@ -318,13 +318,34 @@
 	return __test_and_clear_bit(nr ^ BITOP_LE_SWIZZLE, addr);
 }
 
+static inline unsigned long find_next_zero_bit_le(const void *addr,
+                unsigned long size, unsigned long offset)
+{
+        return find_next_zero_bit(addr, size, offset);
+}
+
+static inline unsigned long find_next_bit_le(const void *addr,
+                unsigned long size, unsigned long offset)
+{
+        return find_next_bit(addr, size, offset);
+}
+
+static inline unsigned long find_first_zero_bit_le(const void *addr,
+                unsigned long size)
+{
+        return find_first_zero_bit(addr, size);
+}
+
 #define find_first_zero_bit_le(addr, size) \
 	find_next_zero_bit_le((addr), (size), 0)
+/*
 unsigned long find_next_zero_bit_le(const void *addr,
 				    unsigned long size, unsigned long offset);
 
 unsigned long find_next_bit_le(const void *addr,
 				    unsigned long size, unsigned long offset);
+*/
+
 /* Bitmap functions for the ext2 filesystem */
 
 #define ext2_set_bit_atomic(lock, nr, addr) \
diff -uNr ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/include/asm/byteorder.h ../linux-2.6.39.4/arch/powerpc/include/asm/byteorder.h
--- ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/include/asm/byteorder.h	2011-08-04 01:13:28.000000000 +0530
+++ ../linux-2.6.39.4/arch/powerpc/include/asm/byteorder.h	2011-10-22 06:38:25.254182204 +0530
@@ -7,6 +7,12 @@
  * as published by the Free Software Foundation; either version
  * 2 of the License, or (at your option) any later version.
  */
+
+#ifdef CONFIG_INVADER
+#include <linux/byteorder/little_endian.h>
+#else
 #include <linux/byteorder/big_endian.h>
+#endif
+
 
 #endif /* _ASM_POWERPC_BYTEORDER_H */
diff -uNr ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/include/asm/elf.h ../linux-2.6.39.4/arch/powerpc/include/asm/elf.h
--- ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/include/asm/elf.h	2011-08-04 01:13:28.000000000 +0530
+++ ../linux-2.6.39.4/arch/powerpc/include/asm/elf.h	2011-10-22 06:38:25.140139532 +0530
@@ -118,13 +118,21 @@
 # define ELF_GREG_TYPE	elf_greg_t32
 # define ELF_ARCH	EM_PPC
 # define ELF_CLASS	ELFCLASS32
+#ifdef CONFIG_INVADER
+# define ELF_DATA	ELFDATA2LSB
+#else
 # define ELF_DATA	ELFDATA2MSB
+#endif
 #endif /* __powerpc64__ */
 
 #ifndef ELF_ARCH
 # define ELF_ARCH	EM_PPC64
 # define ELF_CLASS	ELFCLASS64
+#ifdef CONFIG_INVADER
+# define ELF_DATA	ELFDATA2LSB
+#else
 # define ELF_DATA	ELFDATA2MSB
+#endif
   typedef elf_greg_t64 elf_greg_t;
   typedef elf_gregset_t64 elf_gregset_t;
 #else

 iff -uNr ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/include/asm/pgtable-ppc32.h ../linux-2.6.39.4/arch/powerpc/include/asm/pgtable-ppc32.h
--- ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/include/asm/pgtable-ppc32.h	2011-08-04 01:13:28.000000000 +0530
+++ ../linux-2.6.39.4/arch/powerpc/include/asm/pgtable-ppc32.h	2011-10-28 13:54:33.350779238 +0530
@@ -63,8 +63,12 @@
 #ifdef CONFIG_HIGHMEM
 #define KVIRT_TOP	PKMAP_BASE
 #else
+#ifdef CONFIG_INVADER
+#define KVIRT_TOP	(0xbe000000UL)	
+#else
 #define KVIRT_TOP	(0xfe000000UL)	/* for now, could be FIXMAP_BASE ? */
 #endif
+#endif
 
 /*
  * ioremap_bot starts at that address. Early ioremaps move down from there,
diff -uNr ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/include/asm/prom.h ../linux-2.6.39.4/arch/powerpc/include/asm/prom.h
--- ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/include/asm/prom.h	2011-08-04 01:13:28.000000000 +0530
+++ ../linux-2.6.39.4/arch/powerpc/include/asm/prom.h	2011-10-22 06:38:25.211141792 +0530
@@ -20,6 +20,10 @@
 #include <asm/irq.h>
 #include <asm/atomic.h>
 
+#ifdef CONFIG_INVADER
+#include <asm/io.h>
+#endif
+
 #define HAVE_ARCH_DEVTREE_FIXUPS
 
 #ifdef CONFIG_PPC32
diff -uNr ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/include/asm/reg.h ../linux-2.6.39.4/arch/powerpc/include/asm/reg.h
--- ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/include/asm/reg.h	2011-08-04 01:13:28.000000000 +0530
+++ ../linux-2.6.39.4/arch/powerpc/include/asm/reg.h	2011-10-22 06:38:25.222139709 +0530
@@ -839,7 +839,11 @@
 #define PVR_403GC	0x00200200
 #define PVR_403GCX	0x00201400
 #define PVR_405GP	0x40110000
+#ifdef CONFIG_INVADER
+#define PVR_476		0x7ff52080
+#else
 #define PVR_476		0x11a52000
+#endif
 #define PVR_STB03XXX	0x40310000
 #define PVR_NP405H	0x41410000
 #define PVR_NP405L	0x41610000
 
diff -uNr ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/kernel/cacheinfo.c ../linux-2.6.39.4/arch/powerpc/kernel/cacheinfo.c
--- ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/kernel/cacheinfo.c	2011-08-04 01:13:28.000000000 +0530
+++ ../linux-2.6.39.4/arch/powerpc/kernel/cacheinfo.c	2011-10-24 17:02:11.354148243 +0530
@@ -203,7 +203,12 @@
 	if (!cache_size)
 		return -ENODEV;
 
+	printk("cache size: 0x%u\n", be32_to_cpup(cache_size));
+#ifdef CONFIG_INVADER
+	*ret = be32_to_cpup(cache_size);
+#else
 	*ret = *cache_size;
+#endif
 	return 0;
 }
 
@@ -238,7 +243,12 @@
 	if (!line_size)
 		return -ENODEV;
 
+	printk("lne size: 0x%u\n", be32_to_cpup(line_size));
+#ifdef CONFIG_INVADER
+	*ret = be32_to_cpup(line_size);
+#else
 	*ret = *line_size;
+#endif
 	return 0;
 }
 
diff -uNr ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/kernel/cputable.c ../linux-2.6.39.4/arch/powerpc/kernel/cputable.c
--- ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/kernel/cputable.c	2011-08-04 01:13:28.000000000 +0530
+++ ../linux-2.6.39.4/arch/powerpc/kernel/cputable.c	2011-10-22 06:38:27.763108473 +0530
@@ -1853,6 +1853,22 @@
 		.machine_check		= machine_check_47x,
 		.platform		= "ppc470",
 	},
+#ifdef CONFIG_INVADER
+        { /* 476 others */
+                .pvr_mask               = 0xffffffff,
+                .pvr_value              = 0x7ff52080,
+                .cpu_name               = "476fp",
+                .cpu_features           = CPU_FTRS_47X,
+                .cpu_user_features      = COMMON_USER_BOOKE |
+                        PPC_FEATURE_HAS_FPU,
+                .mmu_features           = MMU_FTR_TYPE_47x |
+                        MMU_FTR_USE_TLBIVAX_BCAST | MMU_FTR_LOCK_BCAST_INVAL,
+                .icache_bsize           = 32,
+                .dcache_bsize           = 32,
+                .machine_check          = machine_check_47x,
+                .platform               = "ppc470",
+        },
+#endif
 	{	/* default match */
 		.pvr_mask		= 0x00000000,
 		.pvr_value		= 0x00000000,
diff -uNr ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/kernel/head_44x.S ../linux-2.6.39.4/arch/powerpc/kernel/head_44x.S
--- ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/kernel/head_44x.S	2011-08-04 01:13:28.000000000 +0530
+++ ../linux-2.6.39.4/arch/powerpc/kernel/head_44x.S	2011-10-30 16:00:57.409103255 +0530
@@ -259,8 +259,14 @@
 
 	/* Compute pte address */
 	rlwimi  r12, r10, PPC44x_PTE_ADD_SHIFT, PPC44x_PTE_ADD_MASK_BIT, 28
+
+#ifdef CONFIG_INVADER
+	lwz	r11, 4(r12)		/* Get high word of pte entry */
+	lwz	r12, 0(r12)		/* Get low word of pte entry */
+#else
 	lwz	r11, 0(r12)		/* Get high word of pte entry */
 	lwz	r12, 4(r12)		/* Get low word of pte entry */
+#endif
 
 	lis	r10,tlb_44x_index@ha
 
@@ -355,8 +361,13 @@
 
 	/* Compute pte address */
 	rlwimi	r12, r10, PPC44x_PTE_ADD_SHIFT, PPC44x_PTE_ADD_MASK_BIT, 28
+#ifdef CONFIG_INVADER
+	lwz	r11, 4(r12)		/* Get high word of pte entry */
+	lwz	r12, 0(r12)		/* Get low word of pte entry */
+#else
 	lwz	r11, 0(r12)		/* Get high word of pte entry */
 	lwz	r12, 4(r12)		/* Get low word of pte entry */
+#endif
 
 	lis	r10,tlb_44x_index@ha
 
@@ -508,8 +519,12 @@
 	/* Compute pte address */
 	rlwimi  r12,r10,PPC44x_PTE_ADD_SHIFT,PPC44x_PTE_ADD_MASK_BIT,28
 	beq	2f			/* Bail if no table */
-	lwz	r11,0(r12)		/* Get high word of pte entry */
 
+#ifdef CONFIG_INVADER
+	lwz	r11,4(r12)		/* Get high word of pte entry */
+#else
+	lwz	r11,0(r12)		/* Get high word of pte entry */
+#endif
 	/* XXX can we do better ? maybe insert a known 0 bit from r11 into the
 	 * bottom of r12 to create a data dependency... We can also use r10
 	 * as destination nowadays
@@ -517,8 +532,12 @@
 #ifdef CONFIG_SMP
 	lwsync
 #endif
-	lwz	r12,4(r12)		/* Get low word of pte entry */
 
+#ifdef CONFIG_INVADER
+	lwz	r12,0(r12)		/* Get low word of pte entry */
+#else
+	lwz	r12,4(r12)		/* Get low word of pte entry */
+#endif
 	andc.	r13,r13,r12		/* Check permission */
 
 	 /* Jump to common tlb load */
@@ -592,15 +611,23 @@
 	rlwimi  r12,r10,PPC44x_PTE_ADD_SHIFT,PPC44x_PTE_ADD_MASK_BIT,28
 	beq	2f			/* Bail if no table */
 
+#ifdef CONFIG_INVADER
+	lwz	r11,4(r12)		/* Get high word of pte entry */
+#else
 	lwz	r11,0(r12)		/* Get high word of pte entry */
+#endif
 	/* XXX can we do better ? maybe insert a known 0 bit from r11 into the
 	 * bottom of r12 to create a data dependency... We can also use r10
 	 * as destination nowadays
 	 */
-#ifdef CONFIG_SMP
+#ifndef CONFIG_SMP
 	lwsync
 #endif
+#ifdef CONFIG_INVADER
+	lwz	r11,0(r12)		/* Get high word of pte entry */
+#else
 	lwz	r12,4(r12)		/* Get low word of pte entry */
+#endif
 
 	andc.	r13,r13,r12		/* Check permission */
 
@@ -638,8 +665,12 @@
 	rlwimi	r10,r12,29,30,30		/* DIRTY -> SW position */
 	and	r11,r12,r10			/* Mask PTE bits to keep */
 	andi.	r10,r12,_PAGE_USER		/* User page ? */
+#ifdef CONFIG_INVADER
+	ori	r11,r11,PPC47x_TLB2_E
+#endif
 	beq	1f				/* nope, leave U bits empty */
 	rlwimi	r11,r11,3,26,28			/* yes, copy S bits to U */
+
 1:	tlbwe	r11,r13,2
 
 	/* Done...restore registers and get out of here.
@@ -935,6 +966,13 @@
  */
 
 head_start_47x:
+//	.align
+	nop
+	nop
+	nop
+	
+	li r3, 0x3
+	rlwinm r23, r3, 1, 31, 31
 	/* Load our current PID->MMUCR TID and MSR IS->MMUCR STS */
 	mfspr	r3,SPRN_PID			/* Get PID */
 	mfmsr	r4				/* Get MSR */
@@ -964,8 +1002,7 @@
 clear_all_utlb_entries:
 
 	#; Set initial values.
-
-	addis		r3,0,0x8000
+	addis		r3,0,0xa000	/* Invader change: quick fix?? */
 	addi		r4,0,0
 	addi		r5,0,0
 	b		clear_utlb_entry
@@ -976,6 +1013,7 @@
 
 clear_utlb_entry:
 
+//	isync	//not needed ??
 	tlbwe		r4,r3,0
 	tlbwe		r5,r3,1
 	tlbwe		r5,r3,2
@@ -987,12 +1025,15 @@
 	cmpwi		r4,0
 	bne		clear_utlb_entry
 
-	#; Restore original entry.
 
+	#; Restore original entry.
+#ifndef CONFIG_INVDER
+// Invader change 
 	oris	r23,r23,0x8000  /* specify the way */
 	tlbwe		r24,r23,0
 	tlbwe		r25,r23,1
 	tlbwe		r26,r23,2
+#endif
 
 /*
  * Configure and load pinned entry into TLB for the kernel core
@@ -1018,7 +1059,12 @@
 					/* ERPN is 0 for first 4GB page */
 	/* Word 2 */
 	li	r5,0
+#ifdef CONFIG_INVADER
+	ori	r5,r5,(PPC47x_TLB2_S_RWX|PPC47x_TLB2_E)
+#else
 	ori	r5,r5,PPC47x_TLB2_S_RWX
+#endif
+
 #ifdef CONFIG_SMP
 	ori	r5,r5,PPC47x_TLB2_M
 #endif
@@ -1062,14 +1108,19 @@
 
 	/* Word 0 */
 	lis	r3,PPC44x_EARLY_DEBUG_VIRTADDR@h
-	ori	r3,r3,PPC47x_TLB0_VALID | PPC47x_TLB0_TS | PPC47x_TLB0_1M
+	ori	r3,r3, (PPC47x_TLB0_VALID | PPC47x_TLB0_TS | PPC47x_TLB0_1M)
 
 	/* Word 1 */
 	lis	r4,CONFIG_PPC_EARLY_DEBUG_44x_PHYSLOW@h
 	ori	r4,r4,CONFIG_PPC_EARLY_DEBUG_44x_PHYSHIGH
 
-	/* Word 2 */
+#ifdef CONFIG_INVADER
+	/* Removing M flag as there is no SMP support for now*/
+	li	r5,(PPC47x_TLB2_S_RW | PPC47x_TLB2_I | PPC47x_TLB2_G | PPC47x_TLB2_E)
+
+#else
 	li	r5,(PPC47x_TLB2_S_RW | PPC47x_TLB2_IMG)
+#endif
 
 	/* Bolted in way 0, bolt slot 5, we -hope- we don't hit the same
 	 * congruence class as the kernel, we need to make sure of it at
diff -uNr ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/kernel/irq.c ../linux-2.6.39.4/arch/powerpc/kernel/irq.c
--- ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/kernel/irq.c	2011-08-04 01:13:28.000000000 +0530
+++ ../linux-2.6.39.4/arch/powerpc/kernel/irq.c	2011-10-24 16:36:08.295113827 +0530
@@ -757,7 +757,11 @@
 
 	/* If host has no translation, then we assume interrupt line */
 	if (host->ops->xlate == NULL)
+#ifdef CONFIG_INVADER
+		hwirq = be32_to_cpu(intspec[0]);
+#else
 		hwirq = intspec[0];
+#endif
 	else {
 		if (host->ops->xlate(host, controller, intspec, intsize,
 				     &hwirq, &type))
diff -uNr ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/kernel/misc_32.S ../linux-2.6.39.4/arch/powerpc/kernel/misc_32.S
--- ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/kernel/misc_32.S	2011-08-04 01:13:28.000000000 +0530
+++ ../linux-2.6.39.4/arch/powerpc/kernel/misc_32.S	2011-10-26 15:51:22.562068744 +0530
@@ -60,6 +60,32 @@
 /*
  * This returns the high 64 bits of the product of two 64-bit numbers.
  */
+
+#ifdef CONFIG_INVADER
+_GLOBAL(mulhdu)
+       cmpwi   r5,0
+       cmpwi   cr1,r4,0
+       mr      r10,r3
+       mulhwu  r3,r3,r6
+       beq     1f
+       mulhwu  r0,r10,r5
+       mullw   r7,r10,r6
+       addc    r7,r0,r7
+       addze   r3,r3
+1:     beqlr   cr1             /* all done if high part of A is 0 */
+       mr      r10,r4
+       mullw   r9,r4,r6
+       mulhwu  r4,r4,r6
+       beq     2f
+       mullw   r0,r10,r5
+       mulhwu  r8,r10,r5
+       addc    r7,r0,r7
+       adde    r3,r3,r8
+       addze   r4,r4
+2:     addc    r3,r3,r9
+       addze   r4,r4
+       blr
+#else
 _GLOBAL(mulhdu)
 	cmpwi	r6,0
 	cmpwi	cr1,r3,0
@@ -83,6 +109,7 @@
 2:	addc	r4,r4,r9
 	addze	r3,r3
 	blr
+#endif
 
 /*
  * sub_reloc_offset(x) returns x - reloc_offset().
@@ -605,6 +632,42 @@
  *  lshrdi3: logical right shift
  *  ashldi3: left shift
  */
+#ifdef CONFIG_INVADER
+_GLOBAL(__ashrdi3)
+       subfic  r6,r5,32
+       srw     r3,r3,r5        # LSW = count > 31 ? 0 : LSW >> count
+       addi    r7,r5,32        # could be xori, or addi with -32
+       slw     r6,r4,r6        # t1 = count > 31 ? 0 : MSW << (32-count)
+       rlwinm  r8,r7,0,32      # t3 = (count < 32) ? 32 : 0
+       sraw    r7,r4,r7        # t2 = MSW >> (count-32)
+       or      r3,r3,r6        # LSW |= t1
+       slw     r7,r7,r8        # t2 = (count < 32) ? 0 : t2
+       sraw    r4,r4,r5        # MSW = MSW >> count
+       or      r3,r3,r7        # LSW |= t2
+       blr
+
+_GLOBAL(__ashldi3)
+       subfic  r6,r5,32
+       slw     r4,r4,r5        # MSW = count > 31 ? 0 : MSW << count
+       addi    r7,r5,32        # could be xori, or addi with -32
+       srw     r6,r3,r6        # t1 = count > 31 ? 0 : LSW >> (32-count)
+       slw     r7,r3,r7        # t2 = count < 32 ? 0 : LSW << (count-32)
+       or      r4,r4,r6        # MSW |= t1
+       slw     r3,r3,r5        # LSW = LSW << count
+       or      r4,r4,r7        # MSW |= t2
+       blr
+
+_GLOBAL(__lshrdi3)
+       subfic  r6,r5,32
+       srw     r3,r3,r5        # LSW = count > 31 ? 0 : LSW >> count
+       addi    r7,r5,32        # could be xori, or addi with -32
+       slw     r6,r4,r6        # t1 = count > 31 ? 0 : MSW << (32-count)
+       srw     r7,r4,r7        # t2 = count < 32 ? 0 : MSW >> (count-32)
+       or      r3,r3,r6        # LSW |= t1
+       srw     r4,r4,r5        # MSW = MSW >> count
+       or      r3,r3,r7        # LSW |= t2
+       blr
+#else
 _GLOBAL(__ashrdi3)
 	subfic	r6,r5,32
 	srw	r4,r4,r5	# LSW = count > 31 ? 0 : LSW >> count
@@ -639,6 +702,7 @@
 	srw	r3,r3,r5	# MSW = MSW >> count
 	or	r4,r4,r7	# LSW |= t2
 	blr
+#endif 
 
 /*
  * 64-bit comparison: __ucmpdi2(u64 a, u64 b)
diff -uNr ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/kernel/module_32.c ../linux-2.6.39.4/arch/powerpc/kernel/module_32.c
--- ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/kernel/module_32.c	2011-08-04 01:13:28.000000000 +0530
+++ ../linux-2.6.39.4/arch/powerpc/kernel/module_32.c	2011-10-22 06:38:27.155103324 +0530
@@ -46,9 +46,15 @@
 	r_addend = 0;
 	for (i = 0; i < num; i++)
 		/* Only count 24-bit relocs, others don't need stubs */
+#ifdef CONFIG_INVADER
+		if ( ((ELF32_R_TYPE(rela[i].r_info) == R_PPC_REL24) && (ELF32_R_TYPE(rela[i].r_info) == R_PPC_REL14) &&
+		    (r_info != ELF32_R_SYM(rela[i].r_info) ||
+		     r_addend != rela[i].r_addend)) ) {
+#else
 		if (ELF32_R_TYPE(rela[i].r_info) == R_PPC_REL24 &&
 		    (r_info != ELF32_R_SYM(rela[i].r_info) ||
 		     r_addend != rela[i].r_addend)) {
+#endif
 			_count_relocs++;
 			r_info = ELF32_R_SYM(rela[i].r_info);
 			r_addend = rela[i].r_addend;
diff -uNr ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/kernel/prom.c ../linux-2.6.39.4/arch/powerpc/kernel/prom.c
--- ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/kernel/prom.c	2011-08-04 01:13:28.000000000 +0530
+++ ../linux-2.6.39.4/arch/powerpc/kernel/prom.c	2011-10-22 12:32:28.777113831 +0530
@@ -103,6 +103,8 @@
 		memcpy(p, initial_boot_params, size);
 		initial_boot_params = (struct boot_param_header *)p;
 		DBG("Moved device tree to 0x%p\n", p);
+// invader debug
+		printk("Moved device tree to 0x%p\n", p);
 	}
 
 	DBG("<- move_device_tree\n");
@@ -300,7 +302,7 @@
 		 */
 		if (initial_boot_params && initial_boot_params->version >= 2) {
 			if (intserv[i] ==
-					initial_boot_params->boot_cpuid_phys) {
+					be32_to_cpu(initial_boot_params->boot_cpuid_phys)) {
 				found = 1;
 				break;
 			}
@@ -344,8 +346,14 @@
 		 * it uses 0x0f000001.
 		 */
 		prop = of_get_flat_dt_prop(node, "cpu-version", NULL);
+		printk(KERN_DEBUG "INVADER PVR: 0x%x\n", be32_to_cpup(prop));
+#ifdef CONFIG_INVADER
+		if (prop && (be32_to_cpup(prop) & 0xff000000) == 0x0f000000)
+			identify_cpu(0, be32_to_cpup(prop));
+#else
 		if (prop && (*prop & 0xff000000) == 0x0f000000)
 			identify_cpu(0, *prop);
+#endif
 
 		identical_pvr_fixup(node);
 	}
@@ -478,6 +486,7 @@
 				if ((base + size) > 0x80000000ul)
 					size = 0x80000000ul - base;
 			}
+			printk(KERN_DEBUG "memblk Add1 base: 0x%llx, size: 0x%llx", base, size);
 			memblock_add(base, size);
 		} while (--rngs);
 	}
@@ -516,6 +525,7 @@
 	memstart_addr = min((u64)memstart_addr, base);
 
 	/* Add the chunk to the MEMBLOCK list */
+	printk(KERN_DEBUG "memblk Add2 base: 0x%llx, size: 0x%llx", base, size);
 	memblock_add(base, size);
 }
 
@@ -541,12 +551,18 @@
 	unsigned long self_base;
 	unsigned long self_size;
 
+	// Invader Change
 	reserve_map = (u64 *)(((unsigned long)initial_boot_params) +
+#ifdef CONFIG_INVADER
+					be32_to_cpu(initial_boot_params->off_mem_rsvmap));
+#else
 					initial_boot_params->off_mem_rsvmap);
+#endif
 
 	/* before we do anything, lets reserve the dt blob */
 	self_base = __pa((unsigned long)initial_boot_params);
-	self_size = initial_boot_params->totalsize;
+	self_size = be32_to_cpu(initial_boot_params->totalsize);
+
 	memblock_reserve(self_base, self_size);
 
 #ifdef CONFIG_BLK_DEV_INITRD
@@ -560,13 +576,18 @@
 	 * Handle the case where we might be booting from an old kexec
 	 * image that setup the mem_rsvmap as pairs of 32-bit values
 	 */
-	if (*reserve_map > 0xffffffffull) {
+	if (be64_to_cpu(*reserve_map) > 0xffffffffull) {
 		u32 base_32, size_32;
 		u32 *reserve_map_32 = (u32 *)reserve_map;
 
 		while (1) {
+#ifdef CONFIG_INVADER
+			base_32 = be32_to_cpu(*(reserve_map_32++));
+			size_32 = be32_to_cpu(*(reserve_map_32++));
+#else
 			base_32 = *(reserve_map_32++);
 			size_32 = *(reserve_map_32++);
+#endif
 			if (size_32 == 0)
 				break;
 			/* skip if the reservation is for the blob */
@@ -579,11 +600,17 @@
 	}
 #endif
 	while (1) {
+// Invader Changes
+#ifdef CONFIG_INVADER
+		base = be64_to_cpu(*(reserve_map++));
+		size = be64_to_cpu(*(reserve_map++));
+#else
 		base = *(reserve_map++);
 		size = *(reserve_map++);
+#endif
 		if (size == 0)
 			break;
-		DBG("reserving: %llx -> %llx\n", base, size);
+		DBG("reserving2: %llx -> %llx\n", base, size);
 		memblock_reserve(base, size);
 	}
 }
@@ -912,8 +939,8 @@
 {
 	struct dentry *d;
 
-	flat_dt_blob.data = initial_boot_params;
-	flat_dt_blob.size = initial_boot_params->totalsize;
+	flat_dt_blob.data = (initial_boot_params);
+	flat_dt_blob.size = be32_to_cpu(initial_boot_params->totalsize);
 
 	d = debugfs_create_blob("flat-device-tree", S_IFREG | S_IRUSR,
 				powerpc_debugfs_root, &flat_dt_blob);
diff -uNr ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/kernel/prom_init.c ../linux-2.6.39.4/arch/powerpc/kernel/prom_init.c
--- ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/kernel/prom_init.c	2011-08-04 01:13:28.000000000 +0530
+++ ../linux-2.6.39.4/arch/powerpc/kernel/prom_init.c	2011-10-22 06:38:27.610103651 +0530
@@ -789,7 +789,11 @@
 } fake_elf = {
 	.elfhdr = {
 		.e_ident = { 0x7f, 'E', 'L', 'F',
+#ifdef CONFIG_INVADER
+			     ELFCLASS32, ELFDATA2LSB, EV_CURRENT },
+#else
 			     ELFCLASS32, ELFDATA2MSB, EV_CURRENT },
+#endif
 		.e_type = ET_EXEC,	/* yeah right */
 		.e_machine = EM_PPC,
 		.e_version = EV_CURRENT,

diff -uNr ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/kernel/vmlinux.lds.S ../linux-2.6.39.4/arch/powerpc/kernel/vmlinux.lds.S
--- ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/kernel/vmlinux.lds.S	2011-08-04 01:13:28.000000000 +0530
+++ ../linux-2.6.39.4/arch/powerpc/kernel/vmlinux.lds.S	2011-10-22 06:38:27.713103244 +0530
@@ -34,11 +34,15 @@
 jiffies = jiffies_64;
 #else
 OUTPUT_ARCH(powerpc:common)
+#ifdef CONFIG_INVADER
 jiffies = jiffies_64 + 4;
+#else
+jiffies = jiffies_64 + 4;
+#endif
 #endif
 SECTIONS
 {
-	. = 0;
+	. = 0x00;
 	reloc_start = .;
 
 	. = KERNELBASE;
diff -uNr ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/lib/div64.S ../linux-2.6.39.4/arch/powerpc/lib/div64.S
--- ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/lib/div64.S	2011-08-04 01:13:28.000000000 +0530
+++ ../linux-2.6.39.4/arch/powerpc/lib/div64.S	2011-10-22 06:38:28.053112501 +0530
@@ -17,8 +17,13 @@
 #include <asm/processor.h>
 
 _GLOBAL(__div64_32)
+#ifdef CONFIG_INVADER
+	lwz	r5,4(r3)	# get the dividend into r5/r6
+	lwz	r6,0(r3)
+#else
 	lwz	r5,0(r3)	# get the dividend into r5/r6
 	lwz	r6,4(r3)
+#endif
 	cmplw	r5,r4
 	li	r7,0
 	li	r8,0
@@ -53,7 +58,12 @@
 	mullw	r10,r0,r4	# and get the remainder
 	add	r8,r8,r0
 	subf	r6,r10,r6
+#ifdef CONFIG_INVADER
+4:	stw	r7,4(r3)	# return the quotient in *r3
+	stw	r8,0(r3)
+#else
 4:	stw	r7,0(r3)	# return the quotient in *r3
 	stw	r8,4(r3)
+#endif
 	mr	r3,r6		# return the remainder in r3
 	blr
diff -uNr ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/Makefile ../linux-2.6.39.4/arch/powerpc/Makefile
--- ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/Makefile	2011-08-04 01:13:28.000000000 +0530
+++ ../linux-2.6.39.4/arch/powerpc/Makefile	2011-10-22 06:42:29.464137538 +0530
@@ -17,8 +17,8 @@
 # Set default 32 bits cross compilers for vdso and boot wrapper
 CROSS32_COMPILE ?=
 
-CROSS32CC		:= $(CROSS32_COMPILE)gcc
-CROSS32AR		:= $(CROSS32_COMPILE)ar
+CROSS32CC		:= $(CROSS32_COMPILE)gcc -mlittle-endian -mno-mfcrf
+CROSS32AR		:= $(CROSS32_COMPILE)ar -EL
 
 ifeq ($(HAS_BIARCH),y)
 ifeq ($(CROSS32_COMPILE),)
@@ -57,9 +57,9 @@
 UTS_MACHINE := $(OLDARCH)
 
 ifeq ($(HAS_BIARCH),y)
-override AS	+= -a$(CONFIG_WORD_SIZE)
-override LD	+= -m elf$(CONFIG_WORD_SIZE)ppc
-override CC	+= -m$(CONFIG_WORD_SIZE)
+override AS	+= -a$(CONFIG_WORD_SIZE) -mlittle-endian
+override LD	+= -m elf$(CONFIG_WORD_SIZE)ppc -EL
+override CC	+= -m$(CONFIG_WORD_SIZE) -mlittle-endian -mno-mfcrf
 override AR	:= GNUTARGET=elf$(CONFIG_WORD_SIZE)-powerpc $(AR)
 endif
 
@@ -68,10 +68,10 @@
 LDFLAGS_vmlinux	:= $(LDFLAGS_vmlinux-yy)
 
 CFLAGS-$(CONFIG_PPC64)	:= -mminimal-toc -mtraceback=none  -mcall-aixdesc
-CFLAGS-$(CONFIG_PPC32)	:= -ffixed-r2 -mmultiple
+CFLAGS-$(CONFIG_PPC32)	:= -ffixed-r2
 KBUILD_CPPFLAGS	+= -Iarch/$(ARCH)
 KBUILD_AFLAGS	+= -Iarch/$(ARCH)
-KBUILD_CFLAGS	+= -msoft-float -pipe -Iarch/$(ARCH) $(CFLAGS-y)
+KBUILD_CFLAGS	+= -msoft-float -pipe -Iarch/$(ARCH) $(CFLAGS-y) 
 CPP		= $(CC) -E $(KBUILD_CFLAGS)
 
 CHECKFLAGS	+= -m$(CONFIG_WORD_SIZE) -D__powerpc__ -D__powerpc$(CONFIG_WORD_SIZE)__
@@ -119,7 +119,7 @@
 
 # Never use string load/store instructions as they are
 # often slow when they are implemented at all
-KBUILD_CFLAGS		+= -mno-string
+#KBUILD_CFLAGS		+= -mno-string
 
 ifeq ($(CONFIG_6xx),y)
 KBUILD_CFLAGS		+= -mcpu=powerpc
@@ -130,7 +130,7 @@
 KBUILD_CFLAGS		+= -mno-sched-epilog
 endif
 
-cpu-as-$(CONFIG_4xx)		+= -Wa,-m405
+cpu-as-$(CONFIG_4xx)		+= -Wa,-m440
 cpu-as-$(CONFIG_6xx)		+= -Wa,-maltivec
 cpu-as-$(CONFIG_POWER4)		+= -Wa,-maltivec
 cpu-as-$(CONFIG_E500)		+= -Wa,-me500
diff -uNr ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/mm/44x_mmu.c ../linux-2.6.39.4/arch/powerpc/mm/44x_mmu.c
--- ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/mm/44x_mmu.c	2011-08-04 01:13:28.000000000 +0530
+++ ../linux-2.6.39.4/arch/powerpc/mm/44x_mmu.c	2011-10-24 17:26:29.012226884 +0530
@@ -165,7 +165,12 @@
 		"tlbwe	%0,%3,2\n"
 		:
 		: "r" (PPC47x_TLB2_SW | PPC47x_TLB2_SR |
+#ifdef CONFIG_INVADER
+		       PPC47x_TLB2_SX | PPC47x_TLB2_E
+#else
 		       PPC47x_TLB2_SX
+#endif
+
 #ifdef CONFIG_SMP
 		       | PPC47x_TLB2_M
 #endif
@@ -223,6 +228,7 @@
 	 */
 	BUG_ON(first_memblock_base != 0);
 
+	printk(KERN_DEBUG "setup_initial_memory_limit: 0x%llx, 0x%llx\n", first_memblock_base, first_memblock_size);
 	/* 44x has a 256M TLB entry pinned at boot */
 	memblock_set_current_limit(min_t(u64, first_memblock_size, PPC_PIN_SIZE));
 }
diff -uNr ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/platforms/invader/invader_setup.c ../linux-2.6.39.4/arch/powerpc/platforms/invader/invader_setup.c
--- ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/platforms/invader/invader_setup.c	1970-01-01 05:30:00.000000000 +0530
+++ ../linux-2.6.39.4/arch/powerpc/platforms/invader/invader_setup.c	2011-10-24 15:53:49.024532082 +0530
@@ -0,0 +1,87 @@
+/*
+	LSI HEADER ???
+*/
+
+#include <linux/init.h>
+#include <linux/of_platform.h>
+
+#include <asm/machdep.h>
+#include <asm/prom.h>
+#include <asm/udbg.h>
+#include <asm/time.h>
+#include <asm/mpic.h>
+#include <asm/ppc4xx.h>
+#include <asm/irq.h>
+#include <asm/pci-bridge.h>
+
+#define DEBUG 1
+
+static struct of_device_id invader_of_bus_ids[] = {
+	{ .compatible = "ibm,plb6", },
+        { .compatible = "simple-bus", },
+	{ .compatible = "chrp,open-pic", },
+	{},
+};
+
+static int __init invader_device_probe(void)
+{
+        of_platform_bus_probe(NULL, invader_of_bus_ids, NULL);
+
+        return 0;
+}
+machine_device_initcall(invader, invader_device_probe);
+
+static int __init invader_probe(void)
+{
+	unsigned long root = of_get_flat_dt_root();
+
+	if (!of_flat_dt_is_compatible(root, "LSI,Invader"))
+		return 0;
+
+	ppc_pci_set_flags(PPC_PCI_REASSIGN_ALL_RSRC);
+
+	return 1;
+}
+
+static void __init invader_init_irq(void)
+{
+        struct mpic *mpic;
+        struct resource r;
+        struct device_node *np = NULL;
+
+        /* Find top level interrupt controller */
+        for_each_node_with_property(np, "interrupt-controller") {
+                if (of_get_property(np, "interrupts", NULL) == NULL)
+                        break;
+        }
+        if (np == NULL)
+                panic("Invader: Can't find top level interrupt controller");
+
+        if (of_device_is_compatible(np, "chrp,open-pic")) {
+                /* The MPIC driver will get everything it needs from the
+                 * device-tree, just pass 0 to all arguments
+                 */
+                mpic = mpic_alloc(np, 0, MPIC_PRIMARY|MPIC_WANTS_RESET, 0, 0, " MPIC ");
+                BUG_ON(mpic == NULL);
+                mpic_init(mpic);
+                ppc_md.get_irq = mpic_get_irq;
+        } else
+                panic("Invader: Unrecognized top level interrupt controller");
+
+	printk("\n\n\t\t\tMPIC initialized\n\n\n");
+}
+
+static void __init invader_progress(char *s, unsigned short hex)
+{
+        printk("*** %04x : %s\n", hex, s ? s : "");
+}
+
+define_machine(invader) {
+	.name			= "Invader",
+	.probe			= invader_probe,
+        .progress               = invader_progress,
+        .init_IRQ       	= invader_init_irq,
+        .get_irq        	= mpic_get_irq,
+        .calibrate_decr         = generic_calibrate_decr,
+        .restart                = ppc4xx_reset_system,
+};
diff -uNr ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/platforms/invader/Kconfig ../linux-2.6.39.4/arch/powerpc/platforms/invader/Kconfig
--- ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/platforms/invader/Kconfig	1970-01-01 05:30:00.000000000 +0530
+++ ../linux-2.6.39.4/arch/powerpc/platforms/invader/Kconfig	2011-10-22 06:38:25.833134467 +0530
@@ -0,0 +1,8 @@
+config INVADER
+	bool "Invader"
+	depends on PPC_47x
+	help 
+	  This option enabled support for LSI PPC476 "Invader"
+		evaluation board.
+
+
diff -uNr ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/platforms/invader/Makefile ../linux-2.6.39.4/arch/powerpc/platforms/invader/Makefile
--- ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/platforms/invader/Makefile	1970-01-01 05:30:00.000000000 +0530
+++ ../linux-2.6.39.4/arch/powerpc/platforms/invader/Makefile	2011-10-22 06:38:25.828374811 +0530
@@ -0,0 +1 @@
+obj-$(CONFIG_INVADER)       += invader_setup.o
diff -uNr ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/platforms/Kconfig ../linux-2.6.39.4/arch/powerpc/platforms/Kconfig
--- ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/platforms/Kconfig	2011-08-04 01:13:28.000000000 +0530
+++ ../linux-2.6.39.4/arch/powerpc/platforms/Kconfig	2011-10-22 06:38:26.083147732 +0530
@@ -20,6 +20,7 @@
 source "arch/powerpc/platforms/44x/Kconfig"
 source "arch/powerpc/platforms/40x/Kconfig"
 source "arch/powerpc/platforms/amigaone/Kconfig"
+source "arch/powerpc/platforms/invader/Kconfig"
 
 config KVM_GUEST
 	bool "KVM Guest support"
diff -uNr ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/platforms/Makefile ../linux-2.6.39.4/arch/powerpc/platforms/Makefile
--- ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/platforms/Makefile	2011-08-04 01:13:28.000000000 +0530
+++ ../linux-2.6.39.4/arch/powerpc/platforms/Makefile	2011-10-22 06:38:25.910151396 +0530
@@ -22,3 +22,4 @@
 obj-$(CONFIG_PPC_PS3)		+= ps3/
 obj-$(CONFIG_EMBEDDED6xx)	+= embedded6xx/
 obj-$(CONFIG_AMIGAONE)		+= amigaone/
+obj-$(CONFIG_INVADER)		+= invader/
diff -uNr ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/sysdev/dcr.c ../linux-2.6.39.4/arch/powerpc/sysdev/dcr.c
--- ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/sysdev/dcr.c	2011-08-04 01:13:28.000000000 +0530
+++ ../linux-2.6.39.4/arch/powerpc/sysdev/dcr.c	2011-10-24 16:41:24.331106125 +0530
@@ -133,7 +133,12 @@
 	if (dr == NULL || ds & 1 || index >= (ds / 8))
 		return 0;
 
+	printk(KERN_DEBUG " dcr-reg: 0x%x", be32_to_cpu(dr[index * 2]));
+#ifdef CONFIG_INVADER
+	return be32_to_cpu(dr[index * 2]);
+#else
 	return dr[index * 2];
+#endif
 }
 EXPORT_SYMBOL_GPL(dcr_resource_start);
 
@@ -145,7 +150,12 @@
 	if (dr == NULL || ds & 1 || index >= (ds / 8))
 		return 0;
 
+	printk(KERN_DEBUG " dcr-reg2: 0x%x", be32_to_cpu(dr[index * 2]+1));
+#ifdef CONFIG_INVADER
+	return be32_to_cpu(dr[index * 2 + 1]);
+#else
 	return dr[index * 2 + 1];
+#endif
 }
 EXPORT_SYMBOL_GPL(dcr_resource_len);
 
@@ -177,6 +187,8 @@
 
 	/* Maybe could do some better range checking here */
 	ret = of_translate_address(dp, p);
+	
+	printk(KERN_DEBUG " translate-dcr-reg: 0x%llx", ret);
 	if (ret != OF_BAD_ADDR)
 		ret += (u64)(stride) * (u64)dcr_n;
 	if (out_stride)
diff -uNr ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/sysdev/mpic.c ../linux-2.6.39.4/arch/powerpc/sysdev/mpic.c
--- ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/sysdev/mpic.c	2011-08-04 01:13:28.000000000 +0530
+++ ../linux-2.6.39.4/arch/powerpc/sysdev/mpic.c	2011-10-25 18:07:53.637770104 +0530
@@ -296,7 +296,7 @@
 
 	dbasep = of_get_property(node, "dcr-reg", NULL);
 
-	rb->dhost = dcr_map(node, *dbasep + offset, size);
+	rb->dhost = dcr_map(node, be32_to_cpup(dbasep) + offset, size);
 	BUG_ON(!DCR_MAP_OK(rb->dhost));
 }
 
@@ -1179,8 +1179,9 @@
 	 */
 	if (paddr == 0 && !(mpic->flags & MPIC_USES_DCR)) {
 		const u32 *reg = of_get_property(node, "reg", NULL);
+		printk(KERN_DEBUG "mpic: reg: 0x%x\n", *reg);
 		BUG_ON(reg == NULL);
-		paddr = of_translate_address(node, reg);
+		paddr = be32_to_cpu(of_translate_address(node, reg));
 		BUG_ON(paddr == OF_BAD_ADDR);
 	}
 
diff -uNr ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/sysdev/ppc4xx_soc.c ../linux-2.6.39.4/arch/powerpc/sysdev/ppc4xx_soc.c
--- ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/arch/powerpc/sysdev/ppc4xx_soc.c	2011-08-04 01:13:28.000000000 +0530
+++ ../linux-2.6.39.4/arch/powerpc/sysdev/ppc4xx_soc.c	2011-10-25 18:04:40.509769915 +0530
@@ -93,7 +93,11 @@
 		of_node_put(np);
 		return -ENODEV;
 	}
+#ifdef CONFIG_INVADER
+	l2_size = be32_to_cpu(prop[0]);
+#else
 	l2_size = prop[0];
+#endif
 
 	/* Map DCRs */
 	dcrreg = of_get_property(np, "dcr-reg", &len);
@@ -103,8 +107,13 @@
 		of_node_put(np);
 		return -ENODEV;
 	}
+#ifdef CONFIG_INVADER
+	dcrbase_isram = be32_to_cpu(dcrreg[0]);
+	dcrbase_l2c = be32_to_cpu(dcrreg[2]);
+#else
 	dcrbase_isram = dcrreg[0];
 	dcrbase_l2c = dcrreg[2];
+#endif
 
 	/* Get and map irq number from device tree */
 	irq = irq_of_parse_and_map(np, 0);
@@ -199,6 +208,7 @@
 	struct device_node *np;
 	u32 reset_type = DBCR0_RST_SYSTEM;
 	const u32 *prop;
+	u32	propVal=0;
 
 	np = of_find_node_by_type(NULL, "cpu");
 	if (np) {
@@ -210,8 +220,17 @@
 		 * 2 - PPC4xx chip reset
 		 * 3 - PPC4xx system reset (default)
 		 */
+#ifdef CONFIG_INVADER
+		if (prop)
+		{
+			propVal = be32_to_cpup(prop);
+			if (((propVal>= 1) && (propVal <= 3)))
+				reset_type = propVal << 28;
+		}
+#else
 		if ((prop) && ((prop[0] >= 1) && (prop[0] <= 3)))
 			reset_type = prop[0] << 28;
+#endif
 	}
 
 	mtspr(SPRN_DBCR0, mfspr(SPRN_DBCR0) | reset_type);
diff -uNr ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/fs/ext2/balloc.c ../linux-2.6.39.4/fs/ext2/balloc.c
--- ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/fs/ext2/balloc.c	2011-08-04 01:13:28.000000000 +0530
+++ ../linux-2.6.39.4/fs/ext2/balloc.c	2011-10-22 06:38:04.625107806 +0530
@@ -18,6 +18,9 @@
 #include <linux/buffer_head.h>
 #include <linux/capability.h>
 
+#ifdef CONFIG_INVADER
+//#include <asm-generic/bitops/le.h>
+#endif
 /*
  * balloc.c contains the blocks allocation and deallocation routines
  */
diff -uNr ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/include/asm-generic/bitops/non-atomic.h ../linux-2.6.39.4/include/asm-generic/bitops/non-atomic.h
--- ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/include/asm-generic/bitops/non-atomic.h	2011-08-04 01:13:28.000000000 +0530
+++ ../linux-2.6.39.4/include/asm-generic/bitops/non-atomic.h	2011-10-22 12:40:27.954406170 +0530
@@ -105,4 +105,25 @@
 	return 1UL & (addr[BIT_WORD(nr)] >> (nr & (BITS_PER_LONG-1)));
 }
 
+static inline int test_bit_dbg(int nr, const volatile unsigned long *addr, volatile unsigned long *ret, volatile unsigned long *ret1, volatile unsigned long *ret2)
+{
+	if(ret)
+		*ret = (addr[BIT_WORD(nr)] >> (nr & (BITS_PER_LONG-1)));
+
+	if (ret1)
+		*ret1 = 1UL&(*ret);
+
+	if (ret2)
+		*ret2 = 1UL & (addr[BIT_WORD(nr)] >> (nr & (BITS_PER_LONG-1)));
+
+#ifdef CONFIG_INVADER
+        return (addr[BIT_WORD(nr)] >> (nr & (BITS_PER_LONG-1)));
+#else
+        return 1UL & (addr[BIT_WORD(nr)] >> (nr & (BITS_PER_LONG-1)));
+#endif
+}
+
+
+
+
diff -uNr ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/init/main.c ../linux-2.6.39.4/init/main.c
--- ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/init/main.c	2011-08-04 01:13:28.000000000 +0530
+++ ../linux-2.6.39.4/init/main.c	2011-10-30 14:04:41.250103375 +0530
@@ -664,7 +664,8 @@
 	int count = preempt_count();
 	int ret;
 
-	if (initcall_debug)
+//	if (initcall_debug)
+	if (1)
 		ret = do_one_initcall_debug(fn);
 	else
 		ret = fn();
@@ -729,6 +730,7 @@
 static void run_init_process(const char *init_filename)
 {
 	argv_init[0] = init_filename;
+	printk(KERN_DEBUG "RUNNING run_init_process %s\n", init_filename);
 	kernel_execve(init_filename, argv_init, envp_init);
 }
 
@@ -747,6 +749,7 @@
 
 	current->signal->flags |= SIGNAL_UNKILLABLE;
 
+	printk(KERN_DEBUG "ramdisk_execute_command: %s, execute_command: %s\n",ramdisk_execute_command ,execute_command);
 	if (ramdisk_execute_command) {
 		run_init_process(ramdisk_execute_command);
 		printk(KERN_WARNING "Failed to execute %s\n",
@@ -765,10 +768,10 @@
 					"defaults...\n", execute_command);
 	}
 	run_init_process("/sbin/init");
-	run_init_process("/etc/init");
+/*	run_init_process("/etc/init");
 	run_init_process("/bin/init");
 	run_init_process("/bin/sh");
-
+*/
 	panic("No init found.  Try passing init= option to kernel. "
 	      "See Linux Documentation/init.txt for guidance.");
 }
diff -uNr ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/kernel/printk.c ../linux-2.6.39.4/kernel/printk.c
--- ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/kernel/printk.c	2011-08-04 01:13:28.000000000 +0530
+++ ../linux-2.6.39.4/kernel/printk.c	2011-10-24 11:52:01.814164824 +0530
@@ -1485,7 +1485,9 @@
 	}
 	console_unlock();
 	console_sysfs_notify();
-
+#ifdef CONFIG_INVADER
+	keep_bootcon = 1;
+#endif
 	/*
 	 * By unregistering the bootconsoles after we enable the real console
 	 * we get the "console xxx enabled" message on all the consoles -
diff -uNr ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/Makefile ../linux-2.6.39.4/Makefile
--- ../../../crosscompiler_with_gdb_c_c++_objc/work/linux-2.6.39.4/Makefile	2011-08-04 01:13:28.000000000 +0530
+++ ../linux-2.6.39.4/Makefile	2011-10-22 06:39:21.397105230 +0530
@@ -237,6 +237,7 @@
 
 HOSTCC       = gcc
 HOSTCXX      = g++
+#Invader Changes
 HOSTCFLAGS   = -Wall -Wmissing-prototypes -Wstrict-prototypes -O2 -fomit-frame-pointer
 HOSTCXXFLAGS = -O2
 
@@ -320,9 +321,9 @@
 
 # Make variables (CC, etc...)
 
-AS		= $(CROSS_COMPILE)as
-LD		= $(CROSS_COMPILE)ld
-CC		= $(CROSS_COMPILE)gcc
+AS		= $(CROSS_COMPILE)as -mlittle-endian
+LD		= $(CROSS_COMPILE)ld -EL
+CC		= $(CROSS_COMPILE)gcc -mlittle-endian -mno-mfcrf
 CPP		= $(CC) -E
 AR		= $(CROSS_COMPILE)ar
 NM		= $(CROSS_COMPILE)nm
@@ -359,10 +360,12 @@
 		   -fno-strict-aliasing -fno-common \
 		   -Werror-implicit-function-declaration \
 		   -Wno-format-security \
-		   -fno-delete-null-pointer-checks
+		   -fno-delete-null-pointer-checks \
+		   -mlittle-endian \
+		   -fsigned-char
 KBUILD_AFLAGS_KERNEL :=
 KBUILD_CFLAGS_KERNEL :=
-KBUILD_AFLAGS   := -D__ASSEMBLY__
+KBUILD_AFLAGS   := -D__ASSEMBLY__ -mlittle-endian
 KBUILD_AFLAGS_MODULE  := -DMODULE
 KBUILD_CFLAGS_MODULE  := -DMODULE
 KBUILD_LDFLAGS_MODULE := -T $(srctree)/scripts/module-common.lds
@@ -545,6 +548,7 @@
 ifdef CONFIG_CC_OPTIMIZE_FOR_SIZE
 KBUILD_CFLAGS	+= -Os
 else
+#Invader Changes
 KBUILD_CFLAGS	+= -O2
 endif
 


^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: powerpc 476, Little-endian, pte fault
  2011-11-01  3:02         ` Santosh Kumar
@ 2011-11-01 13:44           ` Peter Bergner
  2011-11-01 17:14             ` Santosh Kumar
  0 siblings, 1 reply; 11+ messages in thread
From: Peter Bergner @ 2011-11-01 13:44 UTC (permalink / raw)
  To: Santosh Kumar
  Cc: Benjamin Herrenschmidt, linuxppc-dev, Michael Neuling, Ian Munsie,
	linux-kernel

On Tue, 2011-11-01 at 08:32 +0530, Santosh Kumar wrote:
> I am using the same compiler as 476 & 440 instruction is almost the same.

Well the 476 implements ISA 2.05, which I think has added a fair amount
over the 440.  Not to mention the 476 core that has been released has
a FP unit.  I'll note that GCC has support for a 476 with and without
a FP unit, even though AFAIK, we only ship one with a FP unit.
The relevant options are -mcpu=476 (no FP unit) and -mcpu=476fp
(with FP unit).  Basically, -mcpu=476 is equivalent to
-mcpu=476fp -msoft-float.



> @@ -53,7 +58,12 @@
>         mullw   r10,r0,r4       # and get the remainder
>         add     r8,r8,r0
>         subf    r6,r10,r6
> +#ifdef CONFIG_INVADER
>  4:     stw     r7,0(r3)        # return the quotient in *r3
>         stw     r8,4(r3)
> +#else
> +4:     stw     r7,0(r3)        # return the quotient in *r3
> +       stw     r8,4(r3)
> +#endif
>         mr      r3,r6           # return the remainder in r3
>         blr

This looks like a typo, since you didn't actually swap the offsets
on the stw's like you did in all of the other patch hunks.


Peter




^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: powerpc 476, Little-endian, pte fault
  2011-11-01 13:44           ` Peter Bergner
@ 2011-11-01 17:14             ` Santosh Kumar
  2011-11-01 18:44               ` Peter Bergner
  0 siblings, 1 reply; 11+ messages in thread
From: Santosh Kumar @ 2011-11-01 17:14 UTC (permalink / raw)
  To: Peter Bergner
  Cc: Benjamin Herrenschmidt, linuxppc-dev, Michael Neuling, Ian Munsie,
	linux-kernel

> The relevant options are -mcpu=476 (no FP unit) and -mcpu=476fp> (with FP unit).  Basically, -mcpu=476 is equivalent to> -mcpu=476fp -msoft-float.
Yes what you have mentioned is right.

I had a problem configuring the GCC for 476 in little endian mode,
therefore I am using 440 compiler. As this compiler doesn't accept
-mcpu=476 i am using -mcpu=440.

So is this PTE fault related to the compiler options?

Thanks,
Santosh Kumar .A

On 1 November 2011 19:14, Peter Bergner <bergner@vnet.ibm.com> wrote:
> On Tue, 2011-11-01 at 08:32 +0530, Santosh Kumar wrote:
>> I am using the same compiler as 476 & 440 instruction is almost the same.
>
> Well the 476 implements ISA 2.05, which I think has added a fair amount
> over the 440.  Not to mention the 476 core that has been released has
> a FP unit.  I'll note that GCC has support for a 476 with and without
> a FP unit, even though AFAIK, we only ship one with a FP unit.
> The relevant options are -mcpu=476 (no FP unit) and -mcpu=476fp
> (with FP unit).  Basically, -mcpu=476 is equivalent to
> -mcpu=476fp -msoft-float.
>
>
>
>> @@ -53,7 +58,12 @@
>>         mullw   r10,r0,r4       # and get the remainder
>>         add     r8,r8,r0
>>         subf    r6,r10,r6
>> +#ifdef CONFIG_INVADER
>>  4:     stw     r7,0(r3)        # return the quotient in *r3
>>         stw     r8,4(r3)
>> +#else
>> +4:     stw     r7,0(r3)        # return the quotient in *r3
>> +       stw     r8,4(r3)
>> +#endif
>>         mr      r3,r6           # return the remainder in r3
>>         blr
>
> This looks like a typo, since you didn't actually swap the offsets
> on the stw's like you did in all of the other patch hunks.
>
>
> Peter
>
>
>
>

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: powerpc 476, Little-endian, pte fault
  2011-11-01 17:14             ` Santosh Kumar
@ 2011-11-01 18:44               ` Peter Bergner
  0 siblings, 0 replies; 11+ messages in thread
From: Peter Bergner @ 2011-11-01 18:44 UTC (permalink / raw)
  To: Santosh Kumar
  Cc: Benjamin Herrenschmidt, linuxppc-dev, Michael Neuling, Ian Munsie,
	linux-kernel

On Tue, 2011-11-01 at 22:44 +0530, Santosh Kumar wrote:
> I had a problem configuring the GCC for 476 in little endian mode,

What type of problem?  I assume the binutils you are building
against has 476 support too, correct?  You'll need that.

> So is this PTE fault related to the compiler options?

This could be caused by anything, but I highly doubt this is
caused by a compiler option issue.  My $$ are that you either
have a bug in your kernel patch or there are more places in
the kernel source that need patching.

Peter

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: powerpc 476, Little-endian, pte fault
  2011-10-31  5:30 powerpc 476, Little-endian, pte fault Santosh Kumar
  2011-10-31  5:51 ` Michael Neuling
@ 2011-10-31  7:48 ` Shan Hai
  2011-10-31  9:17   ` Santosh Kumar
  1 sibling, 1 reply; 11+ messages in thread
From: Shan Hai @ 2011-10-31  7:48 UTC (permalink / raw)
  To: Santosh Kumar; +Cc: linux-kernel

On Mon, Oct 31, 2011 at 11:00:14AM +0530, Santosh Kumar wrote:
> KERNEL: linux 2.6.39.4
> POWERPC: 476, little endian.
> 
> I am trying to get linux 2.6.39.4 up on PPC 476 i have done done
> Big-endian to little endian Changes in:
> 1) bitops header file.
> 2) while reading the device tree.
> 3) the PTE read/computed in head_32.S
> 4) added E bit in the TLB entries.
> 
> with all the above changes the kernel_init is done but and inited is mounted.
> 
> But while spawning init process the kernel continuously hits pte
> faults at address 0x100000fc and never comes out. Please let me know
> where i should be looking into.
> 

Please check whether the following commit is there in your tree.
2efaca927f5cd7ecd0f1554b8f9b6a9a2c329c03

Cheers
Shan Hai

> -
> Thanks
> Santosh Kumar .A
> 
> Vision without Action is a daydream... Action without Vision is a nightmare...
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: powerpc 476, Little-endian, pte fault
  2011-10-31  7:48 ` Shan Hai
@ 2011-10-31  9:17   ` Santosh Kumar
  0 siblings, 0 replies; 11+ messages in thread
From: Santosh Kumar @ 2011-10-31  9:17 UTC (permalink / raw)
  To: Shan Hai; +Cc: linux-kernel

i tried this but the fixup_user_fault is called as the path followed
is different. Also i see that the flags passed to handle_mm_fault is 0
i.e the fault is not on a write access. Please let me know if i have
to use this fix some where else too.

thanks
Santosh Kumar .A

Vision without Action is a daydream... Action without Vision is a nightmare...



On 31 October 2011 13:18, Shan Hai <haishan.bai@gmail.com> wrote:
> On Mon, Oct 31, 2011 at 11:00:14AM +0530, Santosh Kumar wrote:
>> KERNEL: linux 2.6.39.4
>> POWERPC: 476, little endian.
>>
>> I am trying to get linux 2.6.39.4 up on PPC 476 i have done done
>> Big-endian to little endian Changes in:
>> 1) bitops header file.
>> 2) while reading the device tree.
>> 3) the PTE read/computed in head_32.S
>> 4) added E bit in the TLB entries.
>>
>> with all the above changes the kernel_init is done but and inited is mounted.
>>
>> But while spawning init process the kernel continuously hits pte
>> faults at address 0x100000fc and never comes out. Please let me know
>> where i should be looking into.
>>
>
> Please check whether the following commit is there in your tree.
> 2efaca927f5cd7ecd0f1554b8f9b6a9a2c329c03
>
> Cheers
> Shan Hai
>
>> -
>> Thanks
>> Santosh Kumar .A
>>
>> Vision without Action is a daydream... Action without Vision is a nightmare...
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>> Please read the FAQ at  http://www.tux.org/lkml/
>

^ permalink raw reply	[flat|nested] 11+ messages in thread

end of thread, other threads:[~2011-11-01 18:55 UTC | newest]

Thread overview: 11+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2011-10-31  5:30 powerpc 476, Little-endian, pte fault Santosh Kumar
2011-10-31  5:51 ` Michael Neuling
     [not found]   ` <CAJw04pAh5QDQMkUHbj9=rRPiUCZejimHiydCJFMAoZeSd_ydhA@mail.gmail.com>
2011-10-31  7:38     ` Santosh Kumar
2011-10-31  9:49     ` Michael Neuling
2011-10-31 11:23       ` Benjamin Herrenschmidt
2011-11-01  3:02         ` Santosh Kumar
2011-11-01 13:44           ` Peter Bergner
2011-11-01 17:14             ` Santosh Kumar
2011-11-01 18:44               ` Peter Bergner
2011-10-31  7:48 ` Shan Hai
2011-10-31  9:17   ` Santosh Kumar

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox