[PATCH 2.6.26-rc5 0/2] cris: compile fixes for 2.6.26-rc5

[PATCH 2.6.26-rc5 0/2] cris: compile fixes for 2.6.26-rc5

[Hinko Kočevar]

Hi,

Current status of cris architecture is that current GIT tree at 2.6.26-rc5 does
not even compile.

To remedy this I've created two patches that are sent as a reply to this post:
1. add dummy ops for serial debug port
2. make compile succeed when building with O= (srctree != objtree)

Best regards,
Hinko

-- 
ČETRTA POT, d.o.o., Kranj
Planina 3
4000 Kranj
Slovenia, Europe
Tel. +386 (0) 4 280 66 03
E-mail: hinko.kocevar@cetrtapot.si
Http: www.cetrtapot.si

Re: [PATCH 2.6.26-rc5 1/2] cris: compile fixes for 2.6.26-rc5

[Hinko Kočevar]

[-- Attachment #1: Type: text/plain, Size: 102 bytes --]

Add dummy ops for serial debug port.

Signed-off-by: Hinko Kocevar <hinko.kocevar@cetrtapot.si>

---


[-- Attachment #2: 001-cris-debugport-add-dummy-ops.patch --]
[-- Type: text/x-patch, Size: 969 bytes --]

diff --git a/arch/cris/arch-v10/kernel/debugport.c b/arch/cris/arch-v10/kernel/debugport.c
index 04d5eee..429a281 100644
--- a/arch/cris/arch-v10/kernel/debugport.c
+++ b/arch/cris/arch-v10/kernel/debugport.c
@@ -432,6 +432,13 @@ dummy_write_room(struct tty_struct *tty)
 	return 8192;
 }
 
+static struct tty_operations dummy_ops = {
+	.open = &dummy_open,
+	.close = &dummy_close,
+	.write = &dummy_write,
+	.write_room = &dummy_write_room,
+};
+
 void __init
 init_dummy_console(void)
 {
@@ -448,10 +455,7 @@ init_dummy_console(void)
 		B115200 | CS8 | CREAD | HUPCL | CLOCAL; /* is normally B9600 default... */
 	dummy_driver.flags = TTY_DRIVER_REAL_RAW | TTY_DRIVER_DYNAMIC_DEV;
 
-	dummy_driver.open = dummy_open;
-	dummy_driver.close = dummy_close;
-	dummy_driver.write = dummy_write;
-	dummy_driver.write_room = dummy_write_room;
+	dummy_driver.ops = &dummy_ops;
 	if (tty_register_driver(&dummy_driver))
 		panic("Couldn't register dummy serial driver\n");
 }

Re: [PATCH 2.6.26-rc5 1/2] cris: compile fixes for 2.6.26-rc5

[Alan Cox]

> +static struct tty_operations dummy_ops = {
> +	.open = &dummy_open,
> +	.close = &dummy_close,
> +	.write = &dummy_write,
> +	.write_room = &dummy_write_room,
> +};

Looks ok. From 2.6.26 onward you should be able to omit write/write_room
if you want writes to error, but you will need the dummies if you want
them to vanish.

> @@ -448,10 +455,7 @@ init_dummy_console(void)
>  		B115200 | CS8 | CREAD | HUPCL | CLOCAL; /* is normally B9600 default... */

It should also be setting c_ispeed/c_ospeed

>  	dummy_driver.flags = TTY_DRIVER_REAL_RAW | TTY_DRIVER_DYNAMIC_DEV;

Alan

Re: [PATCH 2.6.26-rc5 1/2] cris: compile fixes for 2.6.26-rc5

[Hinko Koc(evar]

[-- Attachment #1: Type: text/plain, Size: 170 bytes --]

Stab 2

Add dummy ops for serial debug port.
Add setting of c_ispeed/c_ospeed as suggested by Alan Cox.

Signed-off-by: Hinko Kocevar <hinko.kocevar@cetrtapot.si>

---



[-- Attachment #2: 001-cris-debugport-add-dummy-ops.patch --]
[-- Type: text/x-patch, Size: 1102 bytes --]

diff --git a/arch/cris/arch-v10/kernel/debugport.c b/arch/cris/arch-v10/kernel/debugport.c
index 04d5eee..36eee3d 100644
--- a/arch/cris/arch-v10/kernel/debugport.c
+++ b/arch/cris/arch-v10/kernel/debugport.c
@@ -432,6 +432,13 @@ dummy_write_room(struct tty_struct *tty)
 	return 8192;
 }
 
+static struct tty_operations dummy_ops = {
+	.open = &dummy_open,
+	.close = &dummy_close,
+	.write = &dummy_write,
+	.write_room = &dummy_write_room,
+};
+
 void __init
 init_dummy_console(void)
 {
@@ -447,11 +454,10 @@ init_dummy_console(void)
 	dummy_driver.init_termios.c_cflag =
 		B115200 | CS8 | CREAD | HUPCL | CLOCAL; /* is normally B9600 default... */
 	dummy_driver.flags = TTY_DRIVER_REAL_RAW | TTY_DRIVER_DYNAMIC_DEV;
+	dummy_driver.init_termios.c_ispeed = 115200;
+	dummy_driver.init_termios.c_ospeed = 115200;
 
-	dummy_driver.open = dummy_open;
-	dummy_driver.close = dummy_close;
-	dummy_driver.write = dummy_write;
-	dummy_driver.write_room = dummy_write_room;
+	dummy_driver.ops = &dummy_ops;
 	if (tty_register_driver(&dummy_driver))
 		panic("Couldn't register dummy serial driver\n");
 }

Re: [PATCH 2.6.26-rc5 1/2] cris: compile fixes for 2.6.26-rc5

[Alan Cox]

On Fri, 06 Jun 2008 14:12:26 +0200
Hinko Koc(evar <hinko.kocevar@cetrtapot.si> wrote:

> Stab 2
> 
> Add dummy ops for serial debug port.
> Add setting of c_ispeed/c_ospeed as suggested by Alan Cox.
> 
> Signed-off-by: Hinko Kocevar <hinko.kocevar@cetrtapot.si>

Acked-by: Alan Cox <alan@redhat.com>

Re: [PATCH 2.6.26-rc5 2/2] cris: compile fixes for 2.6.26-rc5

[Hinko Kočevar]

[-- Attachment #1: Type: text/plain, Size: 130 bytes --]

Make compile succeed when building with O= (srctree != objtree).

Signed-off-by: Hinko Kocevar <hinko.kocevar@cetrtapot.si>

---


[-- Attachment #2: 002-cris-compile-with-src-ne-obj-dir.patch --]
[-- Type: text/x-patch, Size: 1059 bytes --]

diff --git a/arch/cris/arch-v10/boot/compressed/Makefile b/arch/cris/arch-v10/boot/compressed/Makefile
index 4a031cb..89ad821 100644
--- a/arch/cris/arch-v10/boot/compressed/Makefile
+++ b/arch/cris/arch-v10/boot/compressed/Makefile
@@ -5,7 +5,7 @@
 CC = gcc-cris -melf $(LINUXINCLUDE)
 ccflags-y += -O2
 LD = ld-cris
-ldflags-y += -T $(obj)/decompress.ld
+ldflags-y += -T $(srctree)/$(src)/decompress.ld
 OBJECTS = $(obj)/head.o $(obj)/misc.o
 OBJCOPY = objcopy-cris
 OBJCOPYFLAGS = -O binary --remove-section=.bss
diff --git a/arch/cris/arch-v10/boot/rescue/Makefile b/arch/cris/arch-v10/boot/rescue/Makefile
index 2e5045b..32b9a91 100644
--- a/arch/cris/arch-v10/boot/rescue/Makefile
+++ b/arch/cris/arch-v10/boot/rescue/Makefile
@@ -6,7 +6,7 @@ CC = gcc-cris -mlinux $(LINUXINCLUDE)
 ccflags-y += -O2
 asflags-y += -traditional
 LD = gcc-cris -mlinux -nostdlib
-ldflags-y += -T $(obj)/rescue.ld
+ldflags-y += -T $(srctree)/$(src)/rescue.ld
 OBJCOPY = objcopy-cris
 OBJCOPYFLAGS = -O binary --remove-section=.bss
 obj-$(CONFIG_ETRAX_AXISFLASHMAP) = head.o

Re: [PATCH 2.6.26-rc5 0/2] cris: compile fixes for 2.6.26-rc5

[Adrian Bunk]

On Fri, Jun 06, 2008 at 01:36:26PM +0200, Hinko Kočevar wrote:
> Hi,
> 
> Current status of cris architecture is that current GIT tree at 2.6.26-rc5 does
> not even compile.
>...

@Mikael, Jesper:

BTW:
What happened with the stuff you have?
It seems you missed to merge for 2.6.26?

Can you try to not miss the merge window for 2.6.27, and also publish 
your queued patches in a way that they might get into the -next tree
(Stephen Cc'ed for possibly discussing details)?

> Best regards,
> Hinko

TIA
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

Re: [PATCH 2.6.26-rc5 0/2] cris: compile fixes for 2.6.26-rc5

[Jesper Nilsson]

On Fri, Jun 06, 2008 at 08:32:22PM +0200, Adrian Bunk wrote:
> On Fri, Jun 06, 2008 at 01:36:26PM +0200, Hinko Ko??evar wrote:
> > Hi,
> >
> > Current status of cris architecture is that current GIT tree at 2.6.26-rc5 does
> > not even compile.
> >...
> 
> @Mikael, Jesper:
> 
> BTW:
> What happened with the stuff you have?
> It seems you missed to merge for 2.6.26?

Yeah, unfortunately I got swamped just as the 2.6.26 merge window opened up.
I'll push them out for 2.6.27.

> Can you try to not miss the merge window for 2.6.27, and also publish
> your queued patches in a way that they might get into the -next tree
> (Stephen Cc'ed for possibly discussing details)?

Have I understood correctly that a tree exported for -next can be rebased
without causing any problems? I've hesitated a bit to create a -next tree
to avoid problems with later merges.

> > Best regards,
> > Hinko
> 
> TIA
> Adrian

/^JN - Jesper Nilsson
-- 
               Jesper Nilsson -- jesper.nilsson@axis.com

Re: [PATCH 2.6.26-rc5 0/2] cris: compile fixes for 2.6.26-rc5

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 761 bytes --]

Hi Jasper,

On Tue, 10 Jun 2008 11:49:37 +0200 Jesper Nilsson <Jesper.Nilsson@axis.com> wrote:
>
> Have I understood correctly that a tree exported for -next can be rebased
> without causing any problems? I've hesitated a bit to create a -next tree
> to avoid problems with later merges.

Yes, rebases are OK (but not encouraged :-)).  The rules are that stuff in
the tree for -next must have been posted somewhere appropriate, reviewed,
and unit tested as much as possible and be reasonably expected to be in
the next merge window.

I also merge trees containing fixes for Linus' current tree.

Also, I accept git trees or quilt series.
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 197 bytes --]

Re: [PATCH linux-next 1/1] cris: look for linker scripts in srctree not in objtree

[Hinko Kocevar]

[-- Attachment #1: Type: text/plain, Size: 182 bytes --]

When building kernel with O= linker looks for linker scripts in output tree, while they are located in source tree.

Signed-off-by: Hinko Kocevar <hinko.kocevar@cetrtapot.si>

---



[-- Attachment #2: cris-look-for-linker-scripts-in-srctree.patch --]
[-- Type: text/x-patch, Size: 1025 bytes --]

diff --git a/arch/cris/arch-v10/boot/compressed/Makefile b/arch/cris/arch-v10/boot/compressed/Makefile
index 9ec5f87..cc6527d 100644
--- a/arch/cris/arch-v10/boot/compressed/Makefile
+++ b/arch/cris/arch-v10/boot/compressed/Makefile
@@ -4,7 +4,7 @@
 
 asflags-y += $(LINUXINCLUDE)
 ccflags-y += -O2 $(LINUXINCLUDE)
-ldflags-y += -T $(obj)/decompress.ld
+ldflags-y += -T $(srctree)/arch/$(ARCH)/boot/compressed/decompress.ld
 OBJECTS = $(obj)/head.o $(obj)/misc.o
 OBJCOPYFLAGS = -O binary --remove-section=.bss
 
diff --git a/arch/cris/arch-v10/boot/rescue/Makefile b/arch/cris/arch-v10/boot/rescue/Makefile
index bea8b9c..476007d 100644
--- a/arch/cris/arch-v10/boot/rescue/Makefile
+++ b/arch/cris/arch-v10/boot/rescue/Makefile
@@ -4,7 +4,7 @@
 
 ccflags-y += -O2 $(LINUXINCLUDE)
 asflags-y += $(LINUXINCLUDE)
-ldflags-y += -T $(obj)/rescue.ld
+ldflags-y += -T $(srctree)/arch/$(ARCH)/boot/rescue/rescue.ld
 OBJCOPYFLAGS = -O binary --remove-section=.bss
 obj-$(CONFIG_ETRAX_AXISFLASHMAP) = head.o
 OBJECT := $(obj)/head.o

Re: [PATCH linux-next 1/1] cris: look for linker scripts in srctree not in objtree

[Jesper Nilsson]

Hi!

On Tue, Jul 01, 2008 at 04:08:44PM +0200, Hinko Kocevar wrote:
> When building kernel with O= linker looks for linker scripts in output tree, while they are located in source tree.
>
> Signed-off-by: Hinko Kocevar <hinko.kocevar@cetrtapot.si>

I think this would be cleaner:

diff --git a/arch/cris/arch-v10/boot/compressed/Makefile b/arch/cris/arch-v10/boot/compressed/Makefile
--- a/arch/cris/arch-v10/boot/compressed/Makefile
+++ b/arch/cris/arch-v10/boot/compressed/Makefile
@@ -4,7 +4,7 @@
 
 asflags-y += $(LINUXINCLUDE)
 ccflags-y += -O2 $(LINUXINCLUDE)
-ldflags-y += -T $(obj)/decompress.ld
+ldflags-y += -T $(srctree)/$(src)/decompress.ld
 OBJECTS = $(obj)/head.o $(obj)/misc.o
 OBJCOPYFLAGS = -O binary --remove-section=.bss
 
diff --git a/arch/cris/arch-v10/boot/rescue/Makefile b/arch/cris/arch-v10/boot/rescue/Makefile
--- a/arch/cris/arch-v10/boot/rescue/Makefile
+++ b/arch/cris/arch-v10/boot/rescue/Makefile
@@ -4,7 +4,7 @@
 
 ccflags-y += -O2 $(LINUXINCLUDE)
 asflags-y += $(LINUXINCLUDE)
-ldflags-y += -T $(obj)/rescue.ld
+ldflags-y += -T $(srctree)/$(src)/rescue.ld
 OBJCOPYFLAGS = -O binary --remove-section=.bss
 obj-$(CONFIG_ETRAX_AXISFLASHMAP) = head.o
 OBJECT := $(obj)/head.o


> ---
> diff --git a/arch/cris/arch-v10/boot/compressed/Makefile b/arch/cris/arch-v10/boot/compressed/Makefile
> index 9ec5f87..cc6527d 100644
> --- a/arch/cris/arch-v10/boot/compressed/Makefile
> +++ b/arch/cris/arch-v10/boot/compressed/Makefile
> @@ -4,7 +4,7 @@
>  
>  asflags-y += $(LINUXINCLUDE)
>  ccflags-y += -O2 $(LINUXINCLUDE)
> -ldflags-y += -T $(obj)/decompress.ld
> +ldflags-y += -T $(srctree)/arch/$(ARCH)/boot/compressed/decompress.ld

>  OBJECTS = $(obj)/head.o $(obj)/misc.o
>  OBJCOPYFLAGS = -O binary --remove-section=.bss
>  
> diff --git a/arch/cris/arch-v10/boot/rescue/Makefile b/arch/cris/arch-v10/boot/rescue/Makefile
> index bea8b9c..476007d 100644
> --- a/arch/cris/arch-v10/boot/rescue/Makefile
> +++ b/arch/cris/arch-v10/boot/rescue/Makefile
> @@ -4,7 +4,7 @@
>  
>  ccflags-y += -O2 $(LINUXINCLUDE)
>  asflags-y += $(LINUXINCLUDE)
> -ldflags-y += -T $(obj)/rescue.ld
> +ldflags-y += -T $(srctree)/arch/$(ARCH)/boot/rescue/rescue.ld
>  OBJCOPYFLAGS = -O binary --remove-section=.bss
>  obj-$(CONFIG_ETRAX_AXISFLASHMAP) = head.o
>  OBJECT := $(obj)/head.o

Thanks!

/^JN - Jesper Nilsson
-- 
               Jesper Nilsson -- jesper.nilsson@axis.com

Re: [PATCH linux-next 1/1] cris: look for linker scripts in srctree not in objtree

[Sam Ravnborg]

On Tue, Jul 01, 2008 at 06:02:41PM +0200, Jesper Nilsson wrote:
> Hi!
> 
> On Tue, Jul 01, 2008 at 04:08:44PM +0200, Hinko Kocevar wrote:
> > When building kernel with O= linker looks for linker scripts in output tree, while they are located in source tree.
> >
> > Signed-off-by: Hinko Kocevar <hinko.kocevar@cetrtapot.si>
> 
> I think this would be cleaner:
> 
> diff --git a/arch/cris/arch-v10/boot/compressed/Makefile b/arch/cris/arch-v10/boot/compressed/Makefile
> --- a/arch/cris/arch-v10/boot/compressed/Makefile
> +++ b/arch/cris/arch-v10/boot/compressed/Makefile
> @@ -4,7 +4,7 @@
>  
>  asflags-y += $(LINUXINCLUDE)
>  ccflags-y += -O2 $(LINUXINCLUDE)
> -ldflags-y += -T $(obj)/decompress.ld
> +ldflags-y += -T $(srctree)/$(src)/decompress.ld
>  OBJECTS = $(obj)/head.o $(obj)/misc.o
>  OBJCOPYFLAGS = -O binary --remove-section=.bss
>  

linker scripts are by definition named *.lds in the kernel.
And if you need to preprocess them then name them *.lds.S as
we do in arch/$ARCH/kernel/*

As this is not a generated linker script then the $(srctree)/$(src) prefix
is correct and the $(obj) prefix was plain wrong.

So:
Acked-by: Sam Ravnborg <sam@ravnborg.org>

	Sam

Re: [PATCH linux-next 1/1] cris: look for linker scripts in srctree not in objtree

[Hinko Kocevar]

Jesper Nilsson wrote:
> Hi!
> 
> On Tue, Jul 01, 2008 at 04:08:44PM +0200, Hinko Kocevar wrote:
>> When building kernel with O= linker looks for linker scripts in output tree, while they are located in source tree.
>>
>> Signed-off-by: Hinko Kocevar <hinko.kocevar@cetrtapot.si>
> 
> I think this would be cleaner:
> 
> diff --git a/arch/cris/arch-v10/boot/compressed/Makefile b/arch/cris/arch-v10/boot/compressed/Makefile
> --- a/arch/cris/arch-v10/boot/compressed/Makefile
> +++ b/arch/cris/arch-v10/boot/compressed/Makefile
> @@ -4,7 +4,7 @@
>  
>  asflags-y += $(LINUXINCLUDE)
>  ccflags-y += -O2 $(LINUXINCLUDE)
> -ldflags-y += -T $(obj)/decompress.ld
> +ldflags-y += -T $(srctree)/$(src)/decompress.ld
>  OBJECTS = $(obj)/head.o $(obj)/misc.o
>  OBJCOPYFLAGS = -O binary --remove-section=.bss
>  
> diff --git a/arch/cris/arch-v10/boot/rescue/Makefile b/arch/cris/arch-v10/boot/rescue/Makefile
> --- a/arch/cris/arch-v10/boot/rescue/Makefile
> +++ b/arch/cris/arch-v10/boot/rescue/Makefile
> @@ -4,7 +4,7 @@
>  
>  ccflags-y += -O2 $(LINUXINCLUDE)
>  asflags-y += $(LINUXINCLUDE)
> -ldflags-y += -T $(obj)/rescue.ld
> +ldflags-y += -T $(srctree)/$(src)/rescue.ld
>  OBJCOPYFLAGS = -O binary --remove-section=.bss
>  obj-$(CONFIG_ETRAX_AXISFLASHMAP) = head.o
>  OBJECT := $(obj)/head.o
> 
> 

It certainly is.

-- 
ČETRTA POT, d.o.o., Kranj
Planina 3
4000 Kranj
Slovenia, Europe
Tel. +386 (0) 4 280 66 03
E-mail: hinko.kocevar@cetrtapot.si
Http: www.cetrtapot.si