Skip to contentSkip to conversation

POST: support for km_arm and mem_regions test definition

RFC PATCH Superseded Valentin Longchamp 4 participants 26 messages Aug 3, 2011 – Sep 2, 2011

Subsystems (4)

v1 of 4 Superseded Latest is v4 →

lore
Share & tools
lore
#0 lore raw

cover POST: support for km_arm and mem_regions test definition

This series adds support for POST on the km_arm boards. These boards use a jumper to run some self-tests at the board power-up. There are some adaptations for POST to run on the ARM architecture.

This series defines a new mem_regions POST test. This test also takes place before relocation, but it only tests some regions of the RAM so that it is quicker.

Valentin Longchamp (8):
POST/arm: adaptations needed for POST on ARM to work
POST: add post_log_res field for post results in global data
POST: make env test flags fetching optional
POST: drivers/memory.c coding style cleanup
POST: add new memory regions test
POST/km_arm: add POST memory tests infrastructure
km_arm: change CONFIG_SYS_TEXT_BASE to end of RAM
km_arm: enable POST for these boards

 arch/arm/include/asm/global_data.h      |    5 +
 arch/arm/lib/board.c                    |    2 +
 arch/blackfin/include/asm/global_data.h |    1 +
 arch/nios2/include/asm/global_data.h    |    1 +
 arch/powerpc/include/asm/global_data.h  |    1 +
 arch/sparc/include/asm/global_data.h    |    1 +
 board/keymile/km_arm/km_arm.c           |   11 ++
 include/configs/km/km_arm.h             |   10 ++-
 include/post.h                          |    7 ++
 post/board/km_arm/Makefile              |   29 +++++
 post/board/km_arm/memory.c              |   37 +++++++
 post/drivers/memory.c                   |  171 +++++++++++++++++++------------
 post/post.c                             |   29 ++++--
 post/tests.c                            |   14 +++
 14 files changed, 244 insertions(+), 75 deletions(-)
 create mode 100644 post/board/km_arm/Makefile
 create mode 100644 post/board/km_arm/memory.c
#1 lore raw

PATCH 1/8 POST/arm: adaptations needed for POST on ARM to work

For post to run on ARM, 3 things are needed:
- post_log_word to be defined in gd
- a post.h include in arch/arm/lib/board.c
- most ARM boards will set POST_WORD in RAM, so we need gd->ram, and thus DECLARE_GLOBAL_DATA_PTR is needed in post.h

---
 arch/arm/include/asm/global_data.h |    4 ++++
 arch/arm/lib/board.c               |    2 ++
 include/post.h                     |    6 ++++++
 3 files changed, 12 insertions(+), 0 deletions(-)
arch/arm/include/asm/global_data.h +4 -0
index 4fc51fd..b2c336a 100644
@@ -75,6 +75,10 @@ typedef	struct	global_data {
 #endif
 	void		**jt;		/* jump table */
 	char		env_buf[32];	/* buffer for getenv() before reloc. */
+#if defined(CONFIG_POST) || defined(CONFIG_LOGBUFFER)
+	unsigned long	post_log_word; /* Record POST activities */
+	unsigned long	post_init_f_time; /* When post_init_f started */
+#endif
 } gd_t;
 
 /*
arch/arm/lib/board.c +2 -0
index 90709d0..db8abac 100644
@@ -48,6 +48,8 @@
 #include <nand.h>
 #include <onenand_uboot.h>
 #include <mmc.h>
+#include <post.h>
+#include <logbuff.h>
 
 #ifdef CONFIG_BITBANGMII
 #include <miiphy.h>
include/post.h +6 -0
index 3f259b7..0801956 100644
@@ -76,6 +76,12 @@
 #endif
 #endif /* CONFIG_SYS_POST_WORD_ADDR */
 
+/*
+ * some ARM implementations have to use gd->ram_size, since POST_WORD is
+ * defined in RAM
+ */
+DECLARE_GLOBAL_DATA_PTR;
+
 static inline ulong post_word_load (void)
 {
 	return in_le32((volatile void *)(_POST_WORD_ADDR));
Signature · 1 line
1.7.1
#2 lore raw

PATCH 1/8 POST/arm: adaptations needed for POST on ARM to work

include/post.h line 83

Quoted 8 lines · On Wednesday, August 03, 2011 08:37:00 Valentin Longchamp wrote:
include/post.h +5 -0
 
+/*
+ * some ARM implementations have to use gd->ram_size, since POST_WORD is
+ * defined in RAM
+ */
+DECLARE_GLOBAL_DATA_PTR;

i'm not sure about this. no other header has been allowed to do this in the past, and i dont think we should start now.
-mike
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 836 bytes
Desc: This is a digitally signed message part.
Url : http://lists.denx.de/pipermail/u-boot/attachments/20110814/42854d03/attachment.pgp

#3 lore raw

PATCH 1/8 POST/arm: adaptations needed for POST on ARM to work

Quoted 4 lines · Mike Frysinger
Quoted 3 lines · On Wednesday, August 03, 2011 08:37:00 Valentin Longchamp wrote:
include/post.h +0 -0
Quoted 9 lines

include/post.h line 83

Quoted 5 lines
include/post.h +5 -0
+/*
+ * some ARM implementations have to use gd->ram_size, since POST_WORD is
+ * defined in RAM
+ */
+DECLARE_GLOBAL_DATA_PTR;

i'm not sure about this. no other header has been allowed to do this in the past, and i dont think we should start now.
-mike

OK. Then we should move the post_word_load and post_word_store function to post/post.c. Would this be accepted ?

Signature · 13 lines
Valentin Longchamp
Embedded Software Engineer
Hardware and Chip Integration
______________________________________
KEYMILE AG
Schwarzenburgstr. 73
CH-3097 Liebefeld
Phone +41 31 377 1318
Fax   +41 31 377 1212
valentin.longchamp at keymile.com
www.keymile.com
______________________________________
KEYMILE: A Specialist as a Partner
#4 lore raw

PATCH 1/8 POST/arm: adaptations needed for POST on ARM to work

Quoted 14 lines · On Monday, August 15, 2011 05:09:42 Valentin Longchamp wrote:
Quoted 4 lines · Mike Frysinger
Quoted 3 lines · On Wednesday, August 03, 2011 08:37:00 Valentin Longchamp wrote:
include/post.h +0 -0
Quoted 8 lines
Quoted 5 lines

+/*
+ * some ARM implementations have to use gd->ram_size, since POST_WORD is + * defined in RAM
+ */
+DECLARE_GLOBAL_DATA_PTR;

i'm not sure about this. no other header has been allowed to do this in the past, and i dont think we should start now.

Quoted 2 lines

OK. Then we should move the post_word_load and post_word_store function to post/post.c. Would this be accepted ?

that would add overhead that most people dont need. i guess the only other option would be to add a CONFIG_POST_EXTERNAL_WORD_FUNCS and then post.h would just define the two funcs as externs. it'd be up to the board porters to define them however they want.
-mike
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 836 bytes
Desc: This is a digitally signed message part.
Url : http://lists.denx.de/pipermail/u-boot/attachments/20110816/73142da2/attachment.pgp

#5 lore raw

PATCH 1/8 POST/arm: adaptations needed for POST on ARM to work

Quoted 15 lines · Mike Frysinger
Quoted 14 lines · On Monday, August 15, 2011 05:09:42 Valentin Longchamp wrote:
Quoted 12 lines · Mike Frysinger
Quoted 8 lines · On Wednesday, August 03, 2011 08:37:00 Valentin Longchamp wrote:
include/post.h +2 -0
 
+/*
+ * some ARM implementations have to use gd->ram_size, since POST_WORD

is + * defined in RAM
+ */
+DECLARE_GLOBAL_DATA_PTR;

i'm not sure about this. no other header has been allowed to do this in the past, and i dont think we should start now.

Quoted 8 lines
Quoted 2 lines

OK. Then we should move the post_word_load and post_word_store function to post/post.c. Would this be accepted ?

that would add overhead that most people dont need. i guess the only other option would be to add a CONFIG_POST_EXTERNAL_WORD_FUNCS and then post.h would just define the two funcs as externs. it'd be up to the board porters to define them however they want.
-mike

We don't want externs. Why would moving it into post.c introduce any overhead ?

#6 lore raw

PATCH 1/8 POST/arm: adaptations needed for POST on ARM to work

Quoted 19 lines · On Thursday, August 18, 2011 06:07:06 Marek Vasut wrote:
Quoted 18 lines · Mike Frysinger
Quoted 16 lines · On Monday, August 15, 2011 05:09:42 Valentin Longchamp wrote:
Quoted 4 lines · Mike Frysinger
Quoted 3 lines · On Wednesday, August 03, 2011 08:37:00 Valentin Longchamp wrote:
include/post.h +0 -0
Quoted 8 lines
Quoted 5 lines

+/*
+ * some ARM implementations have to use gd->ram_size, since POST_WORD is + * defined in RAM
+ */
+DECLARE_GLOBAL_DATA_PTR;

i'm not sure about this. no other header has been allowed to do this in the past, and i dont think we should start now.

OK. Then we should move the post_word_load and post_word_store function to post/post.c. Would this be accepted ?

Quoted 7 lines
Quoted 4 lines

that would add overhead that most people dont need. i guess the only other option would be to add a CONFIG_POST_EXTERNAL_WORD_FUNCS and then post.h would just define the two funcs as externs. it'd be up to the board porters to define them however they want.

We don't want externs. Why would moving it into post.c introduce any overhead ?

because the current code expands into a single memory read/write for many arches. moving it into post.c already means making it into an extern and now people have to call an external function instead of inlining the memory access.
-mike
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 836 bytes
Desc: This is a digitally signed message part.
Url : http://lists.denx.de/pipermail/u-boot/attachments/20110818/694883a9/attachment.pgp

#7 lore raw

PATCH 1/8 POST/arm: adaptations needed for POST on ARM to work

Quoted 25 lines · Mike Frysinger
Quoted 24 lines · On Thursday, August 18, 2011 06:07:06 Marek Vasut wrote:
Quoted 22 lines · Mike Frysinger
Quoted 14 lines · On Monday, August 15, 2011 05:09:42 Valentin Longchamp wrote:
Quoted 4 lines · Mike Frysinger
Quoted 3 lines · On Wednesday, August 03, 2011 08:37:00 Valentin Longchamp wrote:
include/post.h +0 -0
Quoted 8 lines
Quoted 5 lines

+/*
+ * some ARM implementations have to use gd->ram_size, since POST_WORD is + * defined in RAM
+ */
+DECLARE_GLOBAL_DATA_PTR;

i'm not sure about this. no other header has been allowed to do this in the past, and i dont think we should start now.

Quoted 2 lines

OK. Then we should move the post_word_load and post_word_store function to post/post.c. Would this be accepted ?

that would add overhead that most people dont need. i guess the only other option would be to add a CONFIG_POST_EXTERNAL_WORD_FUNCS and then post.h would just define the two funcs as externs. it'd be up to the board porters to define them however they want.

Quoted 7 lines
Quoted 2 lines

We don't want externs. Why would moving it into post.c introduce any overhead ?

because the current code expands into a single memory read/write for many arches. moving it into post.c already means making it into an extern and now people have to call an external function instead of inlining the memory access.

I don't think I follow you here ... why won't you be able to inline that stuff if it's in post.c ?

Quoted 1 line

-mike

#8 lore raw

PATCH 1/8 POST/arm: adaptations needed for POST on ARM to work

Quoted 29 lines · On Thursday, August 18, 2011 13:01:56 Marek Vasut wrote:
Quoted 28 lines · Mike Frysinger
Quoted 26 lines · On Thursday, August 18, 2011 06:07:06 Marek Vasut wrote:
Quoted 15 lines · Mike Frysinger
Quoted 14 lines · On Monday, August 15, 2011 05:09:42 Valentin Longchamp wrote:
Quoted 12 lines · Mike Frysinger
Quoted 8 lines · On Wednesday, August 03, 2011 08:37:00 Valentin Longchamp wrote:
include/post.h +2 -0
 
+/*
+ * some ARM implementations have to use gd->ram_size, since

POST_WORD is + * defined in RAM
+ */
+DECLARE_GLOBAL_DATA_PTR;

i'm not sure about this. no other header has been allowed to do this in the past, and i dont think we should start now.

Quoted 7 lines
Quoted 2 lines

OK. Then we should move the post_word_load and post_word_store function to post/post.c. Would this be accepted ?

that would add overhead that most people dont need. i guess the only other option would be to add a CONFIG_POST_EXTERNAL_WORD_FUNCS and then post.h would just define the two funcs as externs. it'd be up to the board porters to define them however they want.

We don't want externs. Why would moving it into post.c introduce any overhead ?

Quoted 7 lines
Quoted 4 lines

because the current code expands into a single memory read/write for many arches. moving it into post.c already means making it into an extern and now people have to call an external function instead of inlining the memory access.

I don't think I follow you here ... why won't you be able to inline that stuff if it's in post.c ?

grep the tree. post.c isnt the only consumer of these funcs.
-mike
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 836 bytes
Desc: This is a digitally signed message part.
Url : http://lists.denx.de/pipermail/u-boot/attachments/20110818/0f474f10/attachment.pgp

#9 lore raw

PATCH 2/8 POST: add post_log_res field for post results in global data

In reply to Valentin Longchamp

The current post_log_word in global data is currently split into 2x 16 bits: half for the test start, half for the test success. Since we alredy have more than 16 POST tests defined and more could be defined, this may result in an overflow and the post_output_backlog would not work for the tests defined further of these 16 positions.

An additional field is added to global data so that we can now support up to 32 (depending of architecture) tests. The post_log_word is only used to record the start of the test and the new field post_log_res for the test success (or failure). The post_output_backlog is for this change also adapted.

---
 arch/arm/include/asm/global_data.h      |    1 +
 arch/blackfin/include/asm/global_data.h |    1 +
 arch/nios2/include/asm/global_data.h    |    1 +
 arch/powerpc/include/asm/global_data.h  |    1 +
 arch/sparc/include/asm/global_data.h    |    1 +
 post/post.c                             |    9 +++++----
 6 files changed, 10 insertions(+), 4 deletions(-)
arch/arm/include/asm/global_data.h +1 -0
index b2c336a..14b9255 100644
@@ -77,6 +77,7 @@ typedef	struct	global_data {
 	char		env_buf[32];	/* buffer for getenv() before reloc. */
 #if defined(CONFIG_POST) || defined(CONFIG_LOGBUFFER)
 	unsigned long	post_log_word; /* Record POST activities */
+	unsigned long	post_log_res; /* success of POST test */
 	unsigned long	post_init_f_time; /* When post_init_f started */
 #endif
 } gd_t;
arch/blackfin/include/asm/global_data.h +1 -0
index eba5e93..62cd631 100644
@@ -50,6 +50,7 @@ typedef struct global_data {
 	unsigned long env_valid;	/* Checksum of Environment valid? */
 #if defined(CONFIG_POST) || defined(CONFIG_LOGBUFFER)
 	unsigned long post_log_word;	/* Record POST activities */
+	unsigned long post_log_res; 	/* success of POST test */
 	unsigned long post_init_f_time;	/* When post_init_f started */
 #endif
 
arch/nios2/include/asm/global_data.h +1 -0
index 2c4a719..1f0bbe8 100644
@@ -34,6 +34,7 @@ typedef	struct	global_data {
 	unsigned long	env_valid;	/* Checksum of Environment valid */
 #if defined(CONFIG_POST) || defined(CONFIG_LOGBUFFER)
 	unsigned long	post_log_word;	/* Record POST activities */
+	unsigned long	post_log_res; /* success of POST test */
 	unsigned long	post_init_f_time; /* When post_init_f started */
 #endif
 	void		**jt;		/* Standalone app jump table */
arch/powerpc/include/asm/global_data.h +1 -0
index a33ca2f..fbfe8c0 100644
@@ -160,6 +160,7 @@ typedef	struct	global_data {
 #endif
 #if defined(CONFIG_POST) || defined(CONFIG_LOGBUFFER)
 	unsigned long	post_log_word;  /* Record POST activities */
+	unsigned long	post_log_res; /* success of POST test */
 	unsigned long	post_init_f_time;  /* When post_init_f started */
 #endif
 #ifdef CONFIG_BOARD_TYPES
arch/sparc/include/asm/global_data.h +1 -0
index 9b14674..4b62250 100644
@@ -58,6 +58,7 @@ typedef struct global_data {
 #endif
 #if defined(CONFIG_POST) || defined(CONFIG_LOGBUFFER)
 	unsigned long post_log_word;	/* Record POST activities */
+	unsigned long post_log_res;	/* success of POST test */
 	unsigned long post_init_f_time;	/* When post_init_f started */
 #endif
 #ifdef CONFIG_BOARD_TYPES
post/post.c +5 -4
index 852d6a5..61acf8d 100644
@@ -121,6 +121,7 @@ void post_bootmode_init (void)
 
 	/* Reset activity record */
 	gd->post_log_word = 0;
+	gd->post_log_res = 0;
 }
 
 int post_bootmode_get (unsigned int *last_test)
@@ -144,12 +145,12 @@ int post_bootmode_get (unsigned int *last_test)
 /* POST tests run before relocation only mark status bits .... */
 static void post_log_mark_start ( unsigned long testid )
 {
-	gd->post_log_word |= (testid)<<16;
+	gd->post_log_word |= testid;
 }
 
 static void post_log_mark_succ ( unsigned long testid )
 {
-	gd->post_log_word |= testid;
+	gd->post_log_res |= testid;
 }
 
 /* ... and the messages are output once we are relocated */
@@ -158,9 +159,9 @@ void post_output_backlog ( void )
 	int j;
 
 	for (j = 0; j < post_list_size; j++) {
-		if (gd->post_log_word & (post_list[j].testid<<16)) {
+		if (gd->post_log_word & (post_list[j].testid)) {
 			post_log ("POST %s ", post_list[j].cmd);
-			if (gd->post_log_word & post_list[j].testid)
+			if (gd->post_log_res & post_list[j].testid)
 				post_log ("PASSED\n");
 			else {
 				post_log ("FAILED\n");
Signature · 1 line
1.7.1
#10 lore raw

PATCH 2/8 POST: add post_log_res field for post results in global data

Quoted 13 lines · Valentin Longchamp

The current post_log_word in global data is currently split into 2x 16 bits: half for the test start, half for the test success. Since we alredy have more than 16 POST tests defined and more could be defined, this may result in an overflow and the post_output_backlog would not work for the tests defined further of these 16 positions.

An additional field is added to global data so that we can now support up to 32 (depending of architecture) tests. The post_log_word is only used to record the start of the test and the new field post_log_res for the test success (or failure). The post_output_backlog is for this change also adapted.

[...]

post/post.c int post_bootmode_get (unsigned int *last_test) · line 148

Quoted 6 lines
post/post.c +1 -1
@@ -144,12 +145,12 @@ int post_bootmode_get (unsigned int *last_test)
 /* POST tests run before relocation only mark status bits .... */
 static void post_log_mark_start ( unsigned long testid )
 {
-	gd->post_log_word |= (testid)<<16;
+	gd->post_log_word |= testid;

I think you can just assign it if you changed the meaning. Still, what if we will have more than 32 tests?

Quoted 7 lines
post/post.c +1 -1
 }
 
 static void post_log_mark_succ ( unsigned long testid )
 {
-	gd->post_log_word |= testid;
+	gd->post_log_res |= testid;
 }
#11 lore raw

PATCH 2/8 POST: add post_log_res field for post results in global data

Quoted 17 lines · Marek Vasut
Quoted 13 lines · Valentin Longchamp

The current post_log_word in global data is currently split into 2x 16 bits: half for the test start, half for the test success. Since we alredy have more than 16 POST tests defined and more could be defined, this may result in an overflow and the post_output_backlog would not work for the tests defined further of these 16 positions.

An additional field is added to global data so that we can now support up to 32 (depending of architecture) tests. The post_log_word is only used to record the start of the test and the new field post_log_res for the test success (or failure). The post_output_backlog is for this change also adapted.

[...]

Quoted 10 lines

post/post.c int post_bootmode_get (unsigned int *last_test) · line 148

Quoted 6 lines
post/post.c +1 -1
@@ -144,12 +145,12 @@ int post_bootmode_get (unsigned int *last_test)
 /* POST tests run before relocation only mark status bits .... */
 static void post_log_mark_start ( unsigned long testid )
 {
-	gd->post_log_word |= (testid)<<16;
+	gd->post_log_word |= testid;

I think you can just assign it if you changed the meaning. Still, what if we will have more than 32 tests?

Yes, this will not work anymore when we have more than 32 tests. But we already have more than 16 tests so there was already a flaw. This at least corrects this flaw.

The problem is that this has to stay in the global_data that has to remain small, so a u32 for bit setting has the great advantage of being small.

Signature · 13 lines
Valentin Longchamp
Embedded Software Engineer
Hardware and Chip Integration
______________________________________
KEYMILE AG
Schwarzenburgstr. 73
CH-3097 Liebefeld
Phone +41 31 377 1318
Fax   +41 31 377 1212
valentin.longchamp at keymile.com
www.keymile.com
______________________________________
KEYMILE: A Specialist as a Partner
#12 lore raw

PATCH 2/8 POST: add post_log_res field for post results in global data

Quoted 30 lines · Valentin Longchamp
Quoted 23 lines · Marek Vasut
Quoted 13 lines · Valentin Longchamp

The current post_log_word in global data is currently split into 2x 16 bits: half for the test start, half for the test success. Since we alredy have more than 16 POST tests defined and more could be defined, this may result in an overflow and the post_output_backlog would not work for the tests defined further of these 16 positions.

An additional field is added to global data so that we can now support up to 32 (depending of architecture) tests. The post_log_word is only used to record the start of the test and the new field post_log_res for the test success (or failure). The post_output_backlog is for this change also adapted.

[...]

Quoted 6 lines
post/post.c +0 -0
@@ -144,12 +145,12 @@ int post_bootmode_get (unsigned int *last_test)
 
 /* POST tests run before relocation only mark status bits .... */
 static void post_log_mark_start ( unsigned long testid )
 {
 
Quoted 5 lines

post/post.c int post_bootmode_get (unsigned int *last_test) · line 148

Quoted 2 lines
post/post.c +1 -1
-	gd->post_log_word |= (testid)<<16;
+	gd->post_log_word |= testid;

I think you can just assign it if you changed the meaning. Still, what if we will have more than 32 tests?

Quoted 6 lines

Yes, this will not work anymore when we have more than 32 tests. But we already have more than 16 tests so there was already a flaw. This at least corrects this flaw.

The problem is that this has to stay in the global_data that has to remain small, so a u32 for bit setting has the great advantage of being small.

I'd love to see a more competent test framework which would run in later stages of boot. That'll remove the huge load of tests that's on the POST framework.

Detlev, any suggestions ? ;-)

#13 lore raw

PATCH 3/8 POST: make env test flags fetching optional

In reply to Valentin Longchamp

Some boards have the environment variables defined in a slow EEPROM. post_run accesses these environment variables to define which tests have to be run (in post_get_flags). This is very slow before the code relocation on some boards with a slow I2C EEPROM for environement variables.

This patch adds a config option to skip the fetching of the test flags in the environment variables. The test flags assigned to the tests then only are the ones statically defined for the test in post/tests.c.

---
 post/post.c |   20 +++++++++++++++-----
 1 files changed, 15 insertions(+), 5 deletions(-)
post/post.c +15 -5
index 61acf8d..1ee0a29 100644
@@ -191,7 +191,8 @@ static void post_bootmode_test_off (void)
 	post_word_store (word);
 }
 
-static void post_get_flags (int *test_flags)
+#ifndef CONFIG_POST_SKIP_ENV_FLAGS
+static void post_get_env_flags(int *test_flags)
 {
 	int  flag[] = {  POST_POWERON,   POST_NORMAL,   POST_SLOWTEST,
 			 POST_CRITICAL };
@@ -204,10 +205,6 @@ static void post_get_flags (int *test_flags)
 	int last;
 	int i, j;
 
-	for (j = 0; j < post_list_size; j++) {
-		test_flags[j] = post_list[j].flags;
-	}
-
 	for (i = 0; i < varnum; i++) {
 		if (getenv_f(var[i], list, sizeof (list)) <= 0)
 			continue;
@@ -245,6 +242,19 @@ static void post_get_flags (int *test_flags)
 			name = s + 1;
 		}
 	}
+}
+#endif
+
+static void post_get_flags(int *test_flags)
+{
+	int j;
+
+	for (j = 0; j < post_list_size; j++)
+		test_flags[j] = post_list[j].flags;
+
+#ifndef CONFIG_POST_SKIP_ENV_FLAGS
+	post_get_env_flags(test_flags);
+#endif
 
 	for (j = 0; j < post_list_size; j++) {
 		if (test_flags[j] & POST_POWERON) {
Signature · 1 line
1.7.1
#14 lore raw

PATCH 4/8 POST: drivers/memory.c coding style cleanup

In reply to Valentin Longchamp

This is needed for a further patch adding a new memory test.

---
 post/drivers/memory.c |  113 +++++++++++++++++++++++++------------------------
 1 files changed, 57 insertions(+), 56 deletions(-)
post/drivers/memory.c +57 -56
index c2b711e..69c5dbe 100644
@@ -250,7 +250,7 @@ static int memory_post_dataline(unsigned long long * pmem)
 			hi = (temp64>>32) & 0xffffffff;
 			lo = temp64 & 0xffffffff;
 
-			post_log ("Memory (date line) error at %08x, "
+			post_log("Memory (date line) error at %08x, "
 				  "wrote %08x%08x, read %08x%08x !\n",
 					  pmem, pathi, patlo, hi, lo);
 			ret = -1;
@@ -281,7 +281,7 @@ static int memory_post_addrline(ulong *testaddr, ulong *base, ulong size)
 			}
 #endif
 			if(readback == *testaddr) {
-				post_log ("Memory (address line) error at %08x<->%08x, "
+				post_log("Memory (address line) error at %08x<->%08x, "
 					"XOR value %08x !\n",
 					testaddr, target, xor);
 				ret = -1;
@@ -291,7 +291,7 @@ static int memory_post_addrline(ulong *testaddr, ulong *base, ulong size)
 	return ret;
 }
 
-static int memory_post_test1 (unsigned long start,
+static int memory_post_test1(unsigned long start,
 			      unsigned long size,
 			      unsigned long val)
 {
@@ -303,13 +303,13 @@ static int memory_post_test1 (unsigned long start,
 	for (i = 0; i < size / sizeof (ulong); i++) {
 		mem[i] = val;
 		if (i % 1024 == 0)
-			WATCHDOG_RESET ();
+			WATCHDOG_RESET();
 	}
 
-	for (i = 0; i < size / sizeof (ulong) && ret == 0; i++) {
+	for (i = 0; i < size / sizeof (ulong) && !ret; i++) {
 		readback = mem[i];
 		if (readback != val) {
-			post_log ("Memory error at %08x, "
+			post_log("Memory error at %08x, "
 				  "wrote %08x, read %08x !\n",
 					  mem + i, val, readback);
 
@@ -317,13 +317,13 @@ static int memory_post_test1 (unsigned long start,
 			break;
 		}
 		if (i % 1024 == 0)
-			WATCHDOG_RESET ();
+			WATCHDOG_RESET();
 	}
 
 	return ret;
 }
 
-static int memory_post_test2 (unsigned long start, unsigned long size)
+static int memory_post_test2(unsigned long start, unsigned long size)
 {
 	unsigned long i;
 	ulong *mem = (ulong *) start;
@@ -333,13 +333,13 @@ static int memory_post_test2 (unsigned long start, unsigned long size)
 	for (i = 0; i < size / sizeof (ulong); i++) {
 		mem[i] = 1 << (i % 32);
 		if (i % 1024 == 0)
-			WATCHDOG_RESET ();
+			WATCHDOG_RESET();
 	}
 
-	for (i = 0; i < size / sizeof (ulong) && ret == 0; i++) {
+	for (i = 0; i < size / sizeof (ulong) && !ret; i++) {
 		readback = mem[i];
 		if (readback != (1 << (i % 32))) {
-			post_log ("Memory error at %08x, "
+			post_log("Memory error at %08x, "
 				  "wrote %08x, read %08x !\n",
 					  mem + i, 1 << (i % 32), readback);
 
@@ -347,13 +347,13 @@ static int memory_post_test2 (unsigned long start, unsigned long size)
 			break;
 		}
 		if (i % 1024 == 0)
-			WATCHDOG_RESET ();
+			WATCHDOG_RESET();
 	}
 
 	return ret;
 }
 
-static int memory_post_test3 (unsigned long start, unsigned long size)
+static int memory_post_test3(unsigned long start, unsigned long size)
 {
 	unsigned long i;
 	ulong *mem = (ulong *) start;
@@ -363,13 +363,13 @@ static int memory_post_test3 (unsigned long start, unsigned long size)
 	for (i = 0; i < size / sizeof (ulong); i++) {
 		mem[i] = i;
 		if (i % 1024 == 0)
-			WATCHDOG_RESET ();
+			WATCHDOG_RESET();
 	}
 
-	for (i = 0; i < size / sizeof (ulong) && ret == 0; i++) {
+	for (i = 0; i < size / sizeof (ulong) && !ret; i++) {
 		readback = mem[i];
 		if (readback != i) {
-			post_log ("Memory error at %08x, "
+			post_log("Memory error at %08x, "
 				  "wrote %08x, read %08x !\n",
 					  mem + i, i, readback);
 
@@ -377,13 +377,13 @@ static int memory_post_test3 (unsigned long start, unsigned long size)
 			break;
 		}
 		if (i % 1024 == 0)
-			WATCHDOG_RESET ();
+			WATCHDOG_RESET();
 	}
 
 	return ret;
 }
 
-static int memory_post_test4 (unsigned long start, unsigned long size)
+static int memory_post_test4(unsigned long start, unsigned long size)
 {
 	unsigned long i;
 	ulong *mem = (ulong *) start;
@@ -393,13 +393,13 @@ static int memory_post_test4 (unsigned long start, unsigned long size)
 	for (i = 0; i < size / sizeof (ulong); i++) {
 		mem[i] = ~i;
 		if (i % 1024 == 0)
-			WATCHDOG_RESET ();
+			WATCHDOG_RESET();
 	}
 
-	for (i = 0; i < size / sizeof (ulong) && ret == 0; i++) {
+	for (i = 0; i < size / sizeof (ulong) && !ret; i++) {
 		readback = mem[i];
 		if (readback != ~i) {
-			post_log ("Memory error at %08x, "
+			post_log("Memory error@%08x, "
 				  "wrote %08x, read %08x !\n",
 					  mem + i, ~i, readback);
 
@@ -407,47 +407,48 @@ static int memory_post_test4 (unsigned long start, unsigned long size)
 			break;
 		}
 		if (i % 1024 == 0)
-			WATCHDOG_RESET ();
+			WATCHDOG_RESET();
 	}
 
 	return ret;
 }
 
-static int memory_post_tests (unsigned long start, unsigned long size)
+static int memory_post_tests(unsigned long start, unsigned long size)
 {
 	int ret = 0;
 
-	if (ret == 0)
-		ret = memory_post_dataline ((unsigned long long *)start);
-	WATCHDOG_RESET ();
-	if (ret == 0)
-		ret = memory_post_addrline ((ulong *)start, (ulong *)start, size);
-	WATCHDOG_RESET ();
-	if (ret == 0)
-		ret = memory_post_addrline ((ulong *)(start + size - 8),
+	if (!ret)
+		ret = memory_post_dataline((unsigned long long *)start);
+	WATCHDOG_RESET();
+	if (!ret)
+		ret = memory_post_addrline((ulong *)start, (ulong *)start,
+						size);
+	WATCHDOG_RESET();
+	if (!ret)
+		ret = memory_post_addrline((ulong *)(start + size - 8),
 					    (ulong *)start, size);
-	WATCHDOG_RESET ();
-	if (ret == 0)
-		ret = memory_post_test1 (start, size, 0x00000000);
-	WATCHDOG_RESET ();
-	if (ret == 0)
-		ret = memory_post_test1 (start, size, 0xffffffff);
-	WATCHDOG_RESET ();
-	if (ret == 0)
-		ret = memory_post_test1 (start, size, 0x55555555);
-	WATCHDOG_RESET ();
-	if (ret == 0)
-		ret = memory_post_test1 (start, size, 0xaaaaaaaa);
-	WATCHDOG_RESET ();
-	if (ret == 0)
-		ret = memory_post_test2 (start, size);
-	WATCHDOG_RESET ();
-	if (ret == 0)
-		ret = memory_post_test3 (start, size);
-	WATCHDOG_RESET ();
-	if (ret == 0)
-		ret = memory_post_test4 (start, size);
-	WATCHDOG_RESET ();
+	WATCHDOG_RESET();
+	if (!ret)
+		ret = memory_post_test1(start, size, 0x00000000);
+	WATCHDOG_RESET();
+	if (!ret)
+		ret = memory_post_test1(start, size, 0xffffffff);
+	WATCHDOG_RESET();
+	if (!ret)
+		ret = memory_post_test1(start, size, 0x55555555);
+	WATCHDOG_RESET();
+	if (!ret)
+		ret = memory_post_test1(start, size, 0xaaaaaaaa);
+	WATCHDOG_RESET();
+	if (!ret)
+		ret = memory_post_test2(start, size);
+	WATCHDOG_RESET();
+	if (!ret)
+		ret = memory_post_test3(start, size);
+	WATCHDOG_RESET();
+	if (!ret)
+		ret = memory_post_test4(start, size);
+	WATCHDOG_RESET();
 
 	return ret;
 }
@@ -502,11 +503,11 @@ int memory_post_test(int flags)
 			ret = memory_post_tests(vstart, memsize);
 		} else {			/* POST_NORMAL */
 			unsigned long i;
-			for (i = 0; i < (memsize >> 20) && ret == 0; i++) {
-				if (ret == 0)
+			for (i = 0; i < (memsize >> 20) && !ret; i++) {
+				if (!ret)
 					ret = memory_post_tests(vstart +
 						(i << 20), 0x800);
-				if (ret == 0)
+				if (!ret)
 					ret = memory_post_tests(vstart +
 						(i << 20) + 0xff800, 0x800);
 			}
Signature · 1 line
1.7.1
#16 lore raw

PATCH 5/8 POST: add new memory regions test

In reply to Valentin Longchamp

This test is similar to the actual POST memory test but quicker and far less complete. It checks the address and data lines and then only tests some regularly placed sub regions of the RAM.
This can be useful when we want to test the RAM but we do not have enough time to run the full memory test.

The POST memory test code was rearranged in order to avoid code duplication between the two tests but the memory test functionnality remains the same.

---
 include/post.h        |    1 +
 post/drivers/memory.c |   78 +++++++++++++++++++++++++++++++++++++------------
 post/tests.c          |   14 +++++++++
 3 files changed, 74 insertions(+), 19 deletions(-)
include/post.h +1 -0
index 0801956..e13d320 100644
@@ -193,6 +193,7 @@ extern int post_hotkeys_pressed(void);
 #define CONFIG_SYS_POST_CODEC		0x00200000
 #define CONFIG_SYS_POST_COPROC		0x00400000
 #define CONFIG_SYS_POST_FLASH		0x00800000
+#define CONFIG_SYS_POST_MEM_REGIONS	0x01000000
 
 #endif /* CONFIG_POST */
 
post/drivers/memory.c +59 -19

120 lines not on this page · raw message

post/tests.c +14 -0
index bfb9cb5..c3f4d9c 100644
@@ -55,6 +55,7 @@ extern int fpga_post_test (int flags);
 extern int lwmon5_watchdog_post_test(int flags);
 extern int sysmon1_post_test(int flags);
 extern int coprocessor_post_test(int flags);
+extern int memory_regions_post_test(int flags);
 
 extern int sysmon_init_f (void);
 
@@ -316,6 +317,19 @@ struct post_test post_list[] =
 	CONFIG_SYS_POST_FLASH
     },
 #endif
+#if CONFIG_POST & CONFIG_SYS_POST_MEM_REGIONS
+    {
+	"Memory regions test",
+	"mem_regions",
+	"This test checks regularly placed regions of the RAM.",
+	POST_ROM | POST_SLOWTEST | POST_PREREL,
+	&memory_regions_post_test,
+	NULL,
+	NULL,
+	CONFIG_SYS_POST_MEM_REGIONS
+    },
+#endif
+
 };
 
 unsigned int post_list_size = ARRAY_SIZE(post_list);
Signature · 1 line
1.7.1
#17 lore raw

PATCH 6/8 POST/km_arm: add POST memory tests infrastructure

In reply to Valentin Longchamp

This patch adds a board support for km_arm in the POST framework. It consists of a memory test configuration with the redefinition of the arch_memory_test_prepare function.

---
 post/board/km_arm/Makefile |   29 +++++++++++++++++++++++++++++
 post/board/km_arm/memory.c |   37 +++++++++++++++++++++++++++++++++++++
 2 files changed, 66 insertions(+), 0 deletions(-)
 create mode 100644 post/board/km_arm/Makefile
 create mode 100644 post/board/km_arm/memory.c
post/board/km_arm/Makefile new file +29 -0
new file mode 100644
index 0000000..6e8a494
@@ -0,0 +1,29 @@
+#
+# (C) Copyright 2002-2006
+# Wolfgang Denk, DENX Software Engineering, wd at 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 $(OBJTREE)/include/autoconf.mk
+
+LIB	= libpostkm_arm.o
+
+COBJS-$(CONFIG_HAS_POST)	+= memory.o
+
+include $(TOPDIR)/post/rules.mk
post/board/km_arm/memory.c new file +37 -0
new file mode 100644
index 0000000..dc466a6
@@ -0,0 +1,37 @@
+/*
+ * (C) Copyright 2011
+ * Valentin Longchamp, Keymile AG Bern, <[email protected]>
+ *
+ * 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., 51 Franklin Street, Fifth Floor, Boston,
+ * MA 02110-1301 USA
+ */
+
+#include <common.h>
+#include <post.h>
+
+int arch_memory_test_prepare(u32 *vstart, u32 *size, phys_addr_t *phys_offset)
+{
+	bd_t *bd = gd->bd;
+
+	*vstart = CONFIG_SYS_SDRAM_BASE;
+
+	/* we go up to relocation plus a 1 MB margin */
+	*size = CONFIG_SYS_TEXT_BASE - (1<<20);
+
+	return 0;
+}
Signature · 1 line
1.7.1
#18 lore raw

PATCH 6/8 POST/km_arm: add POST memory tests infrastructure

Quoted 9 lines · On Wednesday, August 03, 2011 08:37:05 Valentin Longchamp wrote:

This patch adds a board support for km_arm in the POST framework. It consists of a memory test configuration with the redefinition of the arch_memory_test_prepare function.

---
 post/board/km_arm/Makefile |   29 +++++++++++++++++++++++++++++
 post/board/km_arm/memory.c |   37 +++++++++++++++++++++++++++++++++++++

i'm not sure post/board/<board>/ is something we want to continue. putting the post logic into your normal board/<board>/ dir works fine. might i suggest board/km_arm/post.c ...
-mike
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 836 bytes
Desc: This is a digitally signed message part.
Url : http://lists.denx.de/pipermail/u-boot/attachments/20110814/54e1ead2/attachment.pgp

#19 lore raw

PATCH 6/8 POST/km_arm: add POST memory tests infrastructure

Quoted 11 lines · Mike Frysinger
Quoted 9 lines · On Wednesday, August 03, 2011 08:37:05 Valentin Longchamp wrote:

This patch adds a board support for km_arm in the POST framework. It consists of a memory test configuration with the redefinition of the arch_memory_test_prepare function.

---
 post/board/km_arm/Makefile |   29 +++++++++++++++++++++++++++++
 post/board/km_arm/memory.c |   37 +++++++++++++++++++++++++++++++++++++
Quoted 4 lines

i'm not sure post/board/<board>/ is something we want to continue. putting the post logic into your normal board/<board>/ dir works fine. might i suggest board/km_arm/post.c ...
-mike

Fine for me, I will change it for the next version of the series.

Signature · 13 lines
Valentin Longchamp
Embedded Software Engineer
Hardware and Chip Integration
______________________________________
KEYMILE AG
Schwarzenburgstr. 73
CH-3097 Liebefeld
Phone +41 31 377 1318
Fax   +41 31 377 1212
valentin.longchamp at keymile.com
www.keymile.com
______________________________________
KEYMILE: A Specialist as a Partner
#20 lore raw

PATCH 6/8 POST/km_arm: add POST memory tests infrastructure

In reply to Mike Frysinger
Quoted 9 lines · Mike Frysinger
Quoted 8 lines · On Wednesday, August 03, 2011 08:37:05 Valentin Longchamp wrote:

This patch adds a board support for km_arm in the POST framework. It consists of a memory test configuration with the redefinition of the arch_memory_test_prepare function.

---

Quoted 7 lines
Quoted 2 lines
 post/board/km_arm/Makefile |   29 +++++++++++++++++++++++++++++
 post/board/km_arm/memory.c |   37 +++++++++++++++++++++++++++++++++++++

i'm not sure post/board/<board>/ is something we want to continue. putting the post logic into your normal board/<board>/ dir works fine. might i suggest board/km_arm/post.c ...
-mike

post/board/... seems fine to me actually ... there might be more post tests for the board so poluting usual board directory would be stupid.

#21 lore raw

PATCH 6/8 POST/km_arm: add POST memory tests infrastructure

Quoted 14 lines · On Thursday, August 18, 2011 06:44:34 Marek Vasut wrote:
Quoted 13 lines · Mike Frysinger
Quoted 11 lines · On Wednesday, August 03, 2011 08:37:05 Valentin Longchamp wrote:

This patch adds a board support for km_arm in the POST framework. It consists of a memory test configuration with the redefinition of the arch_memory_test_prepare function.

---

 post/board/km_arm/Makefile |   29 +++++++++++++++++++++++++++++
 post/board/km_arm/memory.c |   37

+++++++++++++++++++++++++++++++++++++

Quoted 6 lines
Quoted 3 lines

i'm not sure post/board/<board>/ is something we want to continue. putting the post logic into your normal board/<board>/ dir works fine. might i suggest board/km_arm/post.c ...

post/board/... seems fine to me actually ... there might be more post tests for the board so poluting usual board directory would be stupid.

no it wouldnt. post tests that are board specific belong in the board specific location: board/<board>/. creating a mirror of that existing hierarchy is stupid, and most board dirs have very few files, so there really isn't all that much to clutter up in the first place.
-mike
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 836 bytes
Desc: This is a digitally signed message part.
Url : http://lists.denx.de/pipermail/u-boot/attachments/20110902/50dd8add/attachment.pgp

#22 lore raw

PATCH 7/8 km_arm: change CONFIG_SYS_TEXT_BASE to end of RAM

In reply to Valentin Longchamp

This allows to test a larger part of the RAM in the memory tests.

---
 include/configs/km/km_arm.h |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
include/configs/km/km_arm.h +1 -1
index b1243ad..54c13ce 100644
@@ -54,7 +54,7 @@
 
 #include "asm/arch/config.h"
 
-#define CONFIG_SYS_TEXT_BASE	0x04000000	/* code address after reloc */
+#define CONFIG_SYS_TEXT_BASE	0x07d00000	/* code address after reloc */
 #define CONFIG_SYS_MEMTEST_START 0x00400000	/* 4M */
 #define CONFIG_SYS_MEMTEST_END	0x007fffff	/*(_8M -1) */
 #define CONFIG_SYS_LOAD_ADDR	0x00800000	/* default load adr- 8M */
Signature · 1 line
1.7.1
#23 lore raw

PATCH 8/8 km_arm: enable POST for these boards

In reply to Valentin Longchamp

The current km_arm boards have a Power-On test jumper. When this jumper is set, this triggers some Power-On tests on the board.

This patch enables the support of this jumper for starting the memory_regions test when the jumper is set.

---
 board/keymile/km_arm/km_arm.c |   11 +++++++++++
 include/configs/km/km_arm.h   |    8 ++++++++
 2 files changed, 19 insertions(+), 0 deletions(-)
board/keymile/km_arm/km_arm.c +11 -0
index a8f2b23..f12c730 100644
@@ -448,6 +448,17 @@ int get_scl(void)
 }
 #endif
 
+#if defined(CONFIG_POST)
+
+#define KM_POST_EN_L   44
+
+int post_hotkeys_pressed(void)
+{
+	int val = kw_gpio_get_value(KM_POST_EN_L);
+	return !val;
+}
+#endif
+
 #if defined(CONFIG_SYS_EEPROM_WREN)
 int eeprom_write_enable(unsigned dev_addr, int state)
 {
include/configs/km/km_arm.h +8 -0
index 54c13ce..64ab140 100644
@@ -271,4 +271,12 @@ int get_scl(void);
 /* address for the bootcount (taken from end of RAM) */
 #define BOOTCOUNT_ADDR          (CONFIG_KM_RESERVED_PRAM)
 
+/* enable POST tests with log */
+#define CONFIG_POST	(CONFIG_SYS_POST_MEM_REGIONS)
+#define CONFIG_POST_SKIP_ENV_FLAGS
+#define CONFIG_CMD_DIAG
+
+/* we take the address after the bootcounter */
+#define CONFIG_SYS_POST_WORD_ADDR ((gd->ram_size) - BOOTCOUNT_ADDR + 4)
+
 #endif /* _CONFIG_KM_ARM_H */
Signature · 1 line
1.7.1
#24 lore raw

PATCH 8/8 km_arm: enable POST for these boards

Hello.

Quoted 19 lines · Valentin Longchamp

The current km_arm boards have a Power-On test jumper. When this jumper is set, this triggers some Power-On tests on the board.

This patch enables the support of this jumper for starting the memory_regions test when the jumper is set.

---
  board/keymile/km_arm/km_arm.c |   11 +++++++++++
  include/configs/km/km_arm.h   |    8 ++++++++
  2 files changed, 19 insertions(+), 0 deletions(-)
board/keymile/km_arm/km_arm.c +0 -0
index a8f2b23..f12c730 100644
@@ -448,6 +448,17 @@ int get_scl(void)
  }

board/keymile/km_arm/km_arm.c int get_scl(void) · line 458

Quoted 10 lines
board/keymile/km_arm/km_arm.c +8 -0
  #endif
 
+#if defined(CONFIG_POST)
+
+#define KM_POST_EN_L   44
+
+int post_hotkeys_pressed(void)
+{
+	int val = kw_gpio_get_value(KM_POST_EN_L);
+	return !val;

WHy not just:

return !kw_gpio_get_value(KM_POST_EN_L);

WBR, Sergei

#25 lore raw

PATCH 8/8 km_arm: enable POST for these boards

Quoted 14 lines · Sergei Shtylyov

board/keymile/km_arm/km_arm.c int get_scl(void) · line 458

Quoted 9 lines · Valentin Longchamp
board/keymile/km_arm/km_arm.c +8 -0
+#if defined(CONFIG_POST)
+
+#define KM_POST_EN_L   44
+
+int post_hotkeys_pressed(void)
+{
+    int val = kw_gpio_get_value(KM_POST_EN_L);
+    return !val;

WHy not just:

Quoted 3 lines
    return !kw_gpio_get_value(KM_POST_EN_L);

WBR, Sergei

Yeah, you are right. I will change it for the next version of the series.

Signature · 13 lines
Valentin Longchamp
Embedded Software Engineer
Hardware and Chip Integration
______________________________________
KEYMILE AG
Schwarzenburgstr. 73
CH-3097 Liebefeld
Phone +41 31 377 1318
Fax   +41 31 377 1212
valentin.longchamp at keymile.com
www.keymile.com
______________________________________
KEYMILE: A Specialist as a Partner