[U-Boot] [PATCH 1/5] arm/km: adapt bootcounter evaluation

[U-Boot] [PATCH 1/5] arm/km: adapt bootcounter evaluation

[Holger Brunck]

The bootcounter (stored in the RAM) is not enough protected with the 4 Bytes
BOOTCOUNT_MAGIC against bit errors due to short power loss or holding a system
in RESET. It has been seen, that the bootcounter value has been changed due to
a bit flip on a system holding in RESET, but the BOOTCOUNT_MAGIC was still valid.

A bit pattern with 4000 bytes (after BOOTCOUNT_MAGIC) has been implemented,
which should be enough to detect a bit error.

Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
Signed-off-by: Valentin Longchamp <valentin.longchamp@keymile.com>
cc: Prafulla Wadaskar <prafulla@marvell.com>
---
 board/keymile/km_arm/km_arm.c |   32 +++++++++++++++++++++++++++-----
 1 files changed, 27 insertions(+), 5 deletions(-)

diff --git a/board/keymile/km_arm/km_arm.c b/board/keymile/km_arm/km_arm.c
index fb8f79c..cfd0de7 100644
--- a/board/keymile/km_arm/km_arm.c
+++ b/board/keymile/km_arm/km_arm.c
@@ -408,6 +408,15 @@ int hush_init_var(void)
 #endif
 
 #if defined(CONFIG_BOOTCOUNT_LIMIT)
+const ulong patterns[]      = {	0x00000000,
+				0xFFFFFFFF,
+				0xFF00FF00,
+				0x0F0F0F0F,
+				0xF0F0F0F0};
+const ulong NBR_OF_PATTERNS = sizeof(patterns)/sizeof(*patterns);
+const ulong OFFS_PATTERN    = 3;
+const ulong REPEAT_PATTERN  = 1000;
+
 void bootcount_store(ulong a)
 {
 	volatile ulong *save_addr;
@@ -419,21 +428,34 @@ void bootcount_store(ulong a)
 	save_addr = (ulong*)(size - BOOTCOUNT_ADDR);
 	writel(a, save_addr);
 	writel(BOOTCOUNT_MAGIC, &save_addr[1]);
+
+	for (i = 0; i < REPEAT_PATTERN; i++)
+		writel(patterns[i % NBR_OF_PATTERNS],
+			&save_addr[i+OFFS_PATTERN]);
+
 }
 
 ulong bootcount_load(void)
 {
 	volatile ulong *save_addr;
 	volatile ulong size = 0;
-	int i;
+	ulong counter = 0;
+	int i, tmp;
+
 	for (i = 0; i < CONFIG_NR_DRAM_BANKS; i++) {
 		size += gd->bd->bi_dram[i].size;
 	}
 	save_addr = (ulong*)(size - BOOTCOUNT_ADDR);
-	if (readl(&save_addr[1]) != BOOTCOUNT_MAGIC)
-		return 0;
-	else
-		return readl(save_addr);
+
+	counter = readl(&save_addr[0]);
+
+	/* Is the counter reliable, check in the big pattern for bit errors */
+	for (i = 0; (i < REPEAT_PATTERN) && (counter != 0); i++) {
+		tmp = readl(&save_addr[i+OFFS_PATTERN]);
+		if (tmp != patterns[i % NBR_OF_PATTERNS])
+			counter = 0;
+	}
+	return counter;
 }
 #endif
 
-- 
1.7.1

[U-Boot] [PATCH 2/5] arm/km: portl2 environment address update to P1B

[Holger Brunck]

From: Thomas Herzmann <thomas.herzmann@keymile.com>

The environment eeprom is now at a different MUX address.

Signed-off-by: Thomas Herzmann <thomas.herzmann@keymile.com>
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
cc: Prafulla Wadaskar <prafulla@marvell.com>
---
 include/configs/portl2.h |    6 +++++-
 1 files changed, 5 insertions(+), 1 deletions(-)

diff --git a/include/configs/portl2.h b/include/configs/portl2.h
index a8543a5..e436cfe 100644
--- a/include/configs/portl2.h
+++ b/include/configs/portl2.h
@@ -46,7 +46,11 @@
 #define CONFIG_PORTL2
 
 #define KM_IVM_BUS	"pca9544a:70:9" /* I2C2 (Mux-Port 1)*/
-#define KM_ENV_BUS	"pca9544a:70:a" /* I2C2 (Mux-Port 2)*/
+/*
+ * Note: This is only valid for HW > P1A if you got an outdated P1A
+ *       use KM_ENV_BUS  "pca9544a:70:a"
+ */
+#define KM_ENV_BUS	"pca9544a:70:d"	/* I2C2 (Mux-Port 5)*/
 
 /*
  * portl2 has a fixed link to the XMPP backplane
-- 
1.7.1

[U-Boot] [PATCH 3/5] arm/km: add boardid and hwkey to kernel command line

[Holger Brunck]

We need in some cases a possibility for the kernel to distinguish
on which board he is running. On powerpc we did this with different
dts files. On arm currently we can't do this, so add boardid and
hwkey to the kernel command line and use it later on in the kernel
code.

Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
---
 include/configs/km/km_arm.h |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)

diff --git a/include/configs/km/km_arm.h b/include/configs/km/km_arm.h
index 8476eb7..00410c9 100644
--- a/include/configs/km/km_arm.h
+++ b/include/configs/km/km_arm.h
@@ -69,7 +69,8 @@
 
 /* architecture specific default bootargs */
 #define CONFIG_KM_DEF_BOOT_ARGS_CPU					\
-		"bootcountaddr=${bootcountaddr} ${mtdparts}"
+		"bootcountaddr=${bootcountaddr} ${mtdparts}"		\
+		" boardid=0x${IVM_BoardId} hwkey=0x${IVM_HWKey}"
 
 #define CONFIG_KM_DEF_ENV_CPU						\
 	"boot=bootm ${load_addr_r} - -\0"				\
-- 
1.7.1

[U-Boot] [PATCH 4/5] arm/km: trigger reconfiguration for the Xilinx FPGA

[Holger Brunck]

The Xilinx FPGA must be reconfigured each time the unit
reboots. The FPGA is connected to the GPIO pin 39 from kirkwood.
This patch triggers this pin for km_kirkwood_pci targets.

Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
Signed-off-by: Valentin Longchamp <valentin.longchamp@keymile.com>
cc: Prafulla Wadaskar <prafulla@marvell.com>
---
 board/keymile/km_arm/km_arm.c |    7 ++++++-
 boards.cfg                    |    2 +-
 include/configs/km_kirkwood.h |    3 +++
 3 files changed, 10 insertions(+), 2 deletions(-)

diff --git a/board/keymile/km_arm/km_arm.c b/board/keymile/km_arm/km_arm.c
index cfd0de7..7e4facc 100644
--- a/board/keymile/km_arm/km_arm.c
+++ b/board/keymile/km_arm/km_arm.c
@@ -257,7 +257,12 @@ int board_early_init_f(void)
 	kw_gpio_set_valid(KM_KIRKWOOD_ENV_WP, 38);
 	kw_gpio_direction_output(KM_KIRKWOOD_ENV_WP, 1);
 #endif
-
+#if defined(CONFIG_KM_RECONFIG_XLX)
+	/* trigger the reconfiguration of the xilinx fpga */
+	kw_gpio_set_valid(KM_XLX_PROGRAM_B_PIN, 1);
+	kw_gpio_direction_output(KM_XLX_PROGRAM_B_PIN, 0);
+	kw_gpio_direction_input(KM_XLX_PROGRAM_B_PIN);
+#endif
 	return 0;
 }
 
diff --git a/boards.cfg b/boards.cfg
index 8a5bfc1..afd73ef 100644
--- a/boards.cfg
+++ b/boards.cfg
@@ -129,7 +129,7 @@ davinci_schmoogie            arm         arm926ejs   schmoogie           davinci
 davinci_sffsdr               arm         arm926ejs   sffsdr              davinci        davinci
 davinci_sonata               arm         arm926ejs   sonata              davinci        davinci
 km_kirkwood                  arm         arm926ejs   km_arm              keymile        kirkwood	km_kirkwood:KM_DISABLE_PCI
-km_kirkwood_pci              arm         arm926ejs   km_arm              keymile        kirkwood	km_kirkwood
+km_kirkwood_pci              arm         arm926ejs   km_arm              keymile        kirkwood	km_kirkwood:KM_RECONFIG_XLX
 mgcoge3un                    arm         arm926ejs   km_arm              keymile        kirkwood
 portl2                       arm         arm926ejs   km_arm              keymile        kirkwood
 inetspace_v2                 arm         arm926ejs   netspace_v2         LaCie          kirkwood    netspace_v2:INETSPACE_V2
diff --git a/include/configs/km_kirkwood.h b/include/configs/km_kirkwood.h
index e51b270..ed36124 100644
--- a/include/configs/km_kirkwood.h
+++ b/include/configs/km_kirkwood.h
@@ -54,4 +54,7 @@
 #define KM_IVM_BUS	"pca9544a:70:9"	/* I2C2 (Mux-Port 1)*/
 #define KM_ENV_BUS	"pca9544a:70:d"	/* I2C2 (Mux-Port 5)*/
 
+/* GPIO Pin from kirkwood connected to PROGRAM_B pin of the xilinx FPGA */
+#define KM_XLX_PROGRAM_B_PIN    39
+
 #endif /* _CONFIG_KM_KIRKWOOD */
-- 
1.7.1

[U-Boot] [PATCH 5/5] arm/km: enable jffs2 cmds

[Holger Brunck]

Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
Signed-off-by: Valentin Longchamp <valentin.longchamp@keymile.com>
cc: Prafulla Wadaskar <prafulla@marvell.com>
---
 include/configs/km/km_arm.h |    1 -
 1 files changed, 0 insertions(+), 1 deletions(-)

diff --git a/include/configs/km/km_arm.h b/include/configs/km/km_arm.h
index 00410c9..0271e7e 100644
--- a/include/configs/km/km_arm.h
+++ b/include/configs/km/km_arm.h
@@ -256,7 +256,6 @@ int get_scl(void);
 #if defined(CONFIG_SYS_NO_FLASH)
 #define CONFIG_KM_UBI_PARTITION_NAME   "ubi0"
 #undef	CONFIG_FLASH_CFI_MTD
-#undef	CONFIG_CMD_JFFS2
 #undef	CONFIG_JFFS2_CMDLINE
 #endif
 
-- 
1.7.1

[U-Boot] [PATCH 1/5] arm/km: adapt bootcounter evaluation

[Holger Brunck]

Hi Prafulla,

On 09/14/2011 10:41 AM, Holger Brunck wrote:
> The bootcounter (stored in the RAM) is not enough protected with the 4 Bytes
> BOOTCOUNT_MAGIC against bit errors due to short power loss or holding a system
> in RESET. It has been seen, that the bootcounter value has been changed due to
> a bit flip on a system holding in RESET, but the BOOTCOUNT_MAGIC was still valid.
> 
> A bit pattern with 4000 bytes (after BOOTCOUNT_MAGIC) has been implemented,
> which should be enough to detect a bit error.
> 
> Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
> Signed-off-by: Valentin Longchamp <valentin.longchamp@keymile.com>
> cc: Prafulla Wadaskar <prafulla@marvell.com>
> ---
>  board/keymile/km_arm/km_arm.c |   32 +++++++++++++++++++++++++++-----
>  1 files changed, 27 insertions(+), 5 deletions(-)
> 

...ping...

Or should this serie not go through your tree? If so then please let me know.

Same question for these two patches:
http://lists.denx.de/pipermail/u-boot/2011-September/102534.html
http://lists.denx.de/pipermail/u-boot/2011-September/102535.html

Best regards
Holger

[U-Boot] [PATCH 1/5] arm/km: adapt bootcounter evaluation

[Prafulla Wadaskar]

> -----Original Message-----
> From: Holger Brunck [mailto:holger.brunck at keymile.com]
> Sent: Friday, October 28, 2011 6:43 PM
> To: Prafulla Wadaskar
> Cc: u-boot at lists.denx.de; Valentin Longchamp
> Subject: Re: [PATCH 1/5] arm/km: adapt bootcounter evaluation
> 
> Hi Prafulla,
> 
> On 09/14/2011 10:41 AM, Holger Brunck wrote:
> > The bootcounter (stored in the RAM) is not enough protected
> with the 4 Bytes
> > BOOTCOUNT_MAGIC against bit errors due to short power loss or
> holding a system
> > in RESET. It has been seen, that the bootcounter value has
> been changed due to
> > a bit flip on a system holding in RESET, but the
> BOOTCOUNT_MAGIC was still valid.
> >
> > A bit pattern with 4000 bytes (after BOOTCOUNT_MAGIC) has
> been implemented,
> > which should be enough to detect a bit error.
> >
> > Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
> > Signed-off-by: Valentin Longchamp
> <valentin.longchamp@keymile.com>
> > cc: Prafulla Wadaskar <prafulla@marvell.com>
> > ---
> >  board/keymile/km_arm/km_arm.c |   32
> +++++++++++++++++++++++++++-----
> >  1 files changed, 27 insertions(+), 5 deletions(-)
> >
> 
> ...ping...
> 
> Or should this serie not go through your tree? If so then
> please let me know.

Generally not, this was done earlier by Wolfgang or Albert.
So I was assuming someone will pull.

Hi Wolfgang,
Please flag if I need to pull them.
 
Regards..
Prafulla . . .

> 
> Same question for these two patches:
> http://lists.denx.de/pipermail/u-boot/2011-
> September/102534.html
> http://lists.denx.de/pipermail/u-boot/2011-
> September/102535.html
> 
> Best regards
> Holger

[U-Boot] [PATCH 1/5] arm/km: adapt bootcounter evaluation

[Holger Brunck]

On 10/31/2011 08:29 AM, Prafulla Wadaskar wrote:
> 
> 
>> -----Original Message-----
>> From: Holger Brunck [mailto:holger.brunck at keymile.com]
>> Sent: Friday, October 28, 2011 6:43 PM
>> To: Prafulla Wadaskar
>> Cc: u-boot at lists.denx.de; Valentin Longchamp
>> Subject: Re: [PATCH 1/5] arm/km: adapt bootcounter evaluation
>>
>> Hi Prafulla,
>>
>> On 09/14/2011 10:41 AM, Holger Brunck wrote:
>>> The bootcounter (stored in the RAM) is not enough protected
>> with the 4 Bytes
>>> BOOTCOUNT_MAGIC against bit errors due to short power loss or
>> holding a system
>>> in RESET. It has been seen, that the bootcounter value has
>> been changed due to
>>> a bit flip on a system holding in RESET, but the
>> BOOTCOUNT_MAGIC was still valid.
>>>
>>> A bit pattern with 4000 bytes (after BOOTCOUNT_MAGIC) has
>> been implemented,
>>> which should be enough to detect a bit error.
>>>
>>> Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
>>> Signed-off-by: Valentin Longchamp
>> <valentin.longchamp@keymile.com>
>>> cc: Prafulla Wadaskar <prafulla@marvell.com>
>>> ---
>>>  board/keymile/km_arm/km_arm.c |   32
>> +++++++++++++++++++++++++++-----
>>>  1 files changed, 27 insertions(+), 5 deletions(-)
>>>
>>
>> ...ping...
>>
>> Or should this serie not go through your tree? If so then
>> please let me know.
> 
> Generally not, this was done earlier by Wolfgang or Albert.
> So I was assuming someone will pull.
> 

Hm, in the past the patches for keymile ARM boards go through your tree Prafulla.

Anyway Wolfgang or Albert, who should pick up the patches for our Keymile
Kirkwood boards?

This two series have only board specific patches:
http://lists.denx.de/pipermail/u-boot/2011-September/101328.html
http://lists.denx.de/pipermail/u-boot/2011-September/102534.html

Best regards
Holger

[U-Boot] [PATCH 1/5] arm/km: adapt bootcounter evaluation

[Prafulla Wadaskar]

> -----Original Message-----
> From: Holger Brunck [mailto:holger.brunck at keymile.com]
> Sent: Monday, October 31, 2011 1:44 PM
> To: Prafulla Wadaskar
> Cc: u-boot at lists.denx.de; Valentin Longchamp;
> albert.u.boot at aribaud.net; Wolfgang Denk
> Subject: Re: [PATCH 1/5] arm/km: adapt bootcounter evaluation
> 
> On 10/31/2011 08:29 AM, Prafulla Wadaskar wrote:
> >
> >
> >> -----Original Message-----
> >> From: Holger Brunck [mailto:holger.brunck at keymile.com]
> >> Sent: Friday, October 28, 2011 6:43 PM
> >> To: Prafulla Wadaskar
> >> Cc: u-boot at lists.denx.de; Valentin Longchamp
> >> Subject: Re: [PATCH 1/5] arm/km: adapt bootcounter
> evaluation
> >>
> >> Hi Prafulla,
> >>
> >> On 09/14/2011 10:41 AM, Holger Brunck wrote:
> >>> The bootcounter (stored in the RAM) is not enough protected
> >> with the 4 Bytes
> >>> BOOTCOUNT_MAGIC against bit errors due to short power loss
> or
> >> holding a system
> >>> in RESET. It has been seen, that the bootcounter value has
> >> been changed due to
> >>> a bit flip on a system holding in RESET, but the
> >> BOOTCOUNT_MAGIC was still valid.
> >>>
> >>> A bit pattern with 4000 bytes (after BOOTCOUNT_MAGIC) has
> >> been implemented,
> >>> which should be enough to detect a bit error.
> >>>
> >>> Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
> >>> Signed-off-by: Valentin Longchamp
> >> <valentin.longchamp@keymile.com>
> >>> cc: Prafulla Wadaskar <prafulla@marvell.com>
> >>> ---
> >>>  board/keymile/km_arm/km_arm.c |   32
> >> +++++++++++++++++++++++++++-----
> >>>  1 files changed, 27 insertions(+), 5 deletions(-)
> >>>
> >>
> >> ...ping...
> >>
> >> Or should this serie not go through your tree? If so then
> >> please let me know.
> >
> > Generally not, this was done earlier by Wolfgang or Albert.
> > So I was assuming someone will pull.
> >
> 
> Hm, in the past the patches for keymile ARM boards go through
> your tree Prafulla.

Yes, I remember, there are few patches which are pulled by them.

> 
> Anyway Wolfgang or Albert, who should pick up the patches for
> our Keymile
> Kirkwood boards?

Since those are Kirkwood board specific, I must pull them.
I am pulling them, will provide pull request soon
Once again thanks for notifying.

Regards..
Prafulla . . .

> 
> This two series have only board specific patches:
> http://lists.denx.de/pipermail/u-boot/2011-
> September/101328.html
> http://lists.denx.de/pipermail/u-boot/2011-
> September/102534.html
> 
> Best regards
> Holger

[U-Boot] [PATCH 1/5] arm/km: adapt bootcounter evaluation

[Prafulla Wadaskar]

> -----Original Message-----
> From: Holger Brunck [mailto:holger.brunck at keymile.com]
> Sent: Monday, October 31, 2011 1:44 PM
> To: Prafulla Wadaskar
> Cc: u-boot at lists.denx.de; Valentin Longchamp;
> albert.u.boot at aribaud.net; Wolfgang Denk
> Subject: Re: [PATCH 1/5] arm/km: adapt bootcounter evaluation
> 
> On 10/31/2011 08:29 AM, Prafulla Wadaskar wrote:
...snip...

> 
> Anyway Wolfgang or Albert, who should pick up the patches for
> our Keymile
> Kirkwood boards?
> 
> This two series have only board specific patches:
> http://lists.denx.de/pipermail/u-boot/2011-
> September/101328.html
> http://lists.denx.de/pipermail/u-boot/2011-
> September/102534.html

Applied both these patches to u-boot-marvell.git master branch

Regards..
Prafulla . .

[U-Boot] [PATCH 1/5] arm/km: adapt bootcounter evaluation

[Holger Brunck]

Hi Prafulla,

On 10/31/2011 09:48 AM, Prafulla Wadaskar wrote:
>> Anyway Wolfgang or Albert, who should pick up the patches for
>> our Keymile
>> Kirkwood boards?
>>
>> This two series have only board specific patches:
>> http://lists.denx.de/pipermail/u-boot/2011-
>> September/101328.html
>> http://lists.denx.de/pipermail/u-boot/2011-
>> September/102534.html
> 
> Applied both these patches to u-boot-marvell.git master branch
> 

please pull the two patch series, not only these two patches which are the first
patches of the two series. All in all it consists of seven patches.

Thanks.

Regards
Holger