arm/imx: serial number / UID support

PATCH In review Baruch Siach 3 participants 8 messages Aug 2, 2010 – Aug 4, 2010

Subsystems (3)

Arm (32-bit) Arm sub-architectures Arm i.MX and Layerscape

lore

Share & tools

lore

Baruch Siach

author

2010-08-02 11:59 UTC #0 lore raw

cover arm/imx: serial number / UID support

This patch series adds support for the serial ATAG, and uses this functionality to pass the i.MX25 chip UID as a serial number to the Linux kernel.

Baruch Siach (3):
arm: add support for the serial ATAG
imx25: add support for UID read from eFuse
mx25 3ds: set CPU UID

 arch/arm/boards/freescale-mx25-3-stack/3stack.c |    2 ++
 arch/arm/include/asm/armlinux.h                 |    4 ++++
 arch/arm/lib/armlinux.c                         |   19 +++++++++++++++++++
 arch/arm/mach-imx/imx25.c                       |   12 ++++++++++++
 arch/arm/mach-imx/include/mach/generic.h        |    1 +
 arch/arm/mach-imx/include/mach/imx25-regs.h     |    3 +++
 6 files changed, 41 insertions(+), 0 deletions(-)

_______________________________________________
barebox mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/barebox

Baruch Siach

author

2010-08-02 11:59 UTC #1 lore raw

PATCH 1/3 arm: add support for the serial ATAG

Signed-off-by Baruch Siach <[email protected]>

---
 arch/arm/include/asm/armlinux.h |    4 ++++
 arch/arm/lib/armlinux.c         |   19 +++++++++++++++++++
 2 files changed, 23 insertions(+), 0 deletions(-)

arch/arm/include/asm/armlinux.h +4 -0

index f4104fb..a92d4c1 100644
@@ -7,6 +7,7 @@ void armlinux_set_bootparams(void *params);
 void armlinux_set_architecture(int architecture);
 void armlinux_add_dram(struct device_d *dev);
 void armlinux_set_revision(unsigned int);
+void armlinux_set_serial(u64);
 #else
 static inline void armlinux_set_bootparams(void *params)
 {

@@ -24,6 +25,9 @@ static inline void armlinux_set_revision(unsigned int)
 {
 }
 
+void armlinux_set_serial(u64)
+{
+}
 #endif
 
 #endif /* __ARCH_ARMLINUX_H */

arch/arm/lib/armlinux.c +19 -0

index 17ae057..040fd34 100644
@@ -46,6 +46,7 @@ static int armlinux_architecture = 0;
 static void *armlinux_bootparams = NULL;
 
 static unsigned int system_rev;
+static u64 system_serial;
 
 static void setup_start_tag(void)
 {

@@ -121,6 +122,19 @@ static void setup_revision_tag(void)
 	}
 }
 
+static void setup_serial_tag(void)
+{
+	if (system_serial) {
+		params->hdr.tag = ATAG_SERIAL;
+		params->hdr.size = tag_size(tag_serialnr);
+
+		params->u.serialnr.low  = system_serial & 0xffffffff;
+		params->u.serialnr.high = system_serial >> 32;
+
+		params = tag_next(params);
+	}
+}
+
 #if 0
 static void setup_initrd_tag(ulong initrd_start, ulong initrd_end)
 {

@@ -155,6 +169,7 @@ static void setup_tags(void)
 		setup_initrd_tag (initrd_start, initrd_end);
 #endif
 	setup_revision_tag();
+	setup_serial_tag();
 	setup_end_tag();
 
 	printf("commandline: %s\n"

@@ -186,6 +201,10 @@ void armlinux_set_revision(unsigned int rev)
 	system_rev = rev;
 }
 
+void armlinux_set_serial(u64 serial)
+{
+	system_serial = serial;
+}
 
 #ifdef CONFIG_CMD_BOOTM
 int do_bootm_linux(struct image_data *data)

Signature · 7 lines

1.7.1


_______________________________________________
barebox mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/barebox

Uwe Kleine-König

2010-08-02 12:13 UTC #2 lore raw

Re: [PATCH 1/3] arm: add support for the serial ATAG

Hi Baruch,

Quoted 6 lines

Signed-off-by Baruch Siach <[email protected]>

---
 arch/arm/include/asm/armlinux.h |    4 ++++
 arch/arm/lib/armlinux.c         |   19 +++++++++++++++++++
 2 files changed, 23 insertions(+), 0 deletions(-)

I like it. Last time I was tempted to implement this myself I would have added a command to set the serial from the shell as in my case the serial number was saved in env and passed to Linux via a kernel parameter.

So you can count that as an Ack.

Thanks
Uwe

Signature · 7 lines

Pengutronix e.K.                           | Uwe Kleine-König            |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |

_______________________________________________
barebox mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/barebox

Baruch Siach

author

2010-08-02 11:59 UTC #3 lore raw

PATCH 2/3 imx25: add support for UID read from eFuse

In reply to Baruch Siach

Signed-off-by Baruch Siach <[email protected]>

---
 arch/arm/mach-imx/imx25.c                   |   12 ++++++++++++
 arch/arm/mach-imx/include/mach/generic.h    |    1 +
 arch/arm/mach-imx/include/mach/imx25-regs.h |    3 +++
 3 files changed, 16 insertions(+), 0 deletions(-)

arch/arm/mach-imx/imx25.c +12 -0

index 00a1e4e..de70247 100644
@@ -16,6 +16,8 @@
  */
 
 #include <common.h>
+#include <mach/imx-regs.h>
+#include <asm/io.h>
 
 #include "gpio.h"

@@ -28,3 +30,13 @@ void *imx_gpio_base[] = {
 
 int imx_gpio_count = ARRAY_SIZE(imx_gpio_base) * 32;
 
+u64 imx_uid(void)
+{
+	u64 uid = 0;
+	int i;
+
+	for (i = 0; i < 8; i++)
+		uid = (uid << 8) | readb(IMX_IIM_BASE + IIM_UID + i*4);
+
+	return uid;
+}

arch/arm/mach-imx/include/mach/generic.h +1 -0

index 48ed336..4b89838 100644
@@ -6,6 +6,7 @@ int imx_silicon_revision(void);
 #define IMX35_CHIP_REVISION_1_0   0x10
 #define IMX35_CHIP_REVISION_2_0   0x20
 
+u64 imx_uid(void);
 
 
 #ifdef CONFIG_ARCH_IMX1

arch/arm/mach-imx/include/mach/imx25-regs.h +3 -0

index e91e7b6..7c2b5f9 100644
@@ -140,5 +140,8 @@
 /* important definition of some bits of WCR */
 #define WCR_WDE 0x04
 
+/* IIM fuse definitions */
+#define IIM_UID		0x820
+
 #endif /* __ASM_ARCH_MX25_REGS_H */

Signature · 7 lines

1.7.1


_______________________________________________
barebox mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/barebox

Baruch Siach

author

2010-08-02 11:59 UTC #4 lore raw

PATCH 3/3 mx25 3ds: set CPU UID

In reply to Baruch Siach

Signed-off-by Baruch Siach <[email protected]>

---
 arch/arm/boards/freescale-mx25-3-stack/3stack.c |    2 ++
 1 files changed, 2 insertions(+), 0 deletions(-)

arch/arm/boards/freescale-mx25-3-stack/3stack.c +2 -0

index a77a02d..cd228cc 100644
@@ -36,6 +36,7 @@
 #include <nand.h>
 #include <mach/imx-flash-header.h>
 #include <mach/iomux-mx25.h>
+#include <mach/generic.h>
 #include <linux/err.h>
 #include <i2c/i2c.h>
 #include <i2c/mc34704.h>

@@ -275,6 +276,7 @@ static int imx25_devices_init(void)
 	armlinux_add_dram(&sdram0_dev);
 	armlinux_set_bootparams((void *)0x80000100);
 	armlinux_set_architecture(MACH_TYPE_MX25_3DS);
+	armlinux_set_serial(imx_uid());
 
 	return 0;
 }

Signature · 7 lines

1.7.1


_______________________________________________
barebox mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/barebox

Jean-Christophe PLAGNIOL-VILLARD

2010-08-02 12:06 UTC #5 lore raw

In reply to Baruch Siach

Hi,

serial number is more for board ref than CPU ref as you can easly read it from the register.

So why pass it via the ATAG?

Best Regards,
J.

Quoted 17 lines · Baruch Siach

This patch series adds support for the serial ATAG, and uses this functionality to pass the i.MX25 chip UID as a serial number to the Linux kernel.
Baruch Siach (3):
arm: add support for the serial ATAG
imx25: add support for UID read from eFuse
mx25 3ds: set CPU UID
 arch/arm/boards/freescale-mx25-3-stack/3stack.c |    2 ++
 arch/arm/include/asm/armlinux.h                 |    4 ++++
 arch/arm/lib/armlinux.c                         |   19 +++++++++++++++++++
 arch/arm/mach-imx/imx25.c                       |   12 ++++++++++++
 arch/arm/mach-imx/include/mach/generic.h        |    1 +
 arch/arm/mach-imx/include/mach/imx25-regs.h     |    3 +++
 6 files changed, 41 insertions(+), 0 deletions(-)

Quoted 4 lines

_______________________________________________
barebox mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/barebox

_______________________________________________
barebox mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/barebox

Baruch Siach

author

2010-08-04 06:46 UTC #6 lore raw

Hi Jean-Christophe,

Quoted 4 lines · Jean-Christophe PLAGNIOL-VILLARD

serial number is more for board ref than CPU ref as you can easly read it from the register.
So why pass it via the ATAG?

The point of the serial number ATAG is to uniquely identify the system. Getting this unique ID from the CPU is as good as any other IMO. I'm not aware of any other UID source on the mx25 3DS system. Besides, having the serial number nicely listed in /proc/cpuinfo is far easier than reading registers.

Anyway, patches 1 and 2 in this series are useful on their own, I think. I intend to use this functionality for our own board. I've added the mx25 3ds two lines patch to justify the bloat :).

baruch

Quoted 18 lines

Best Regards,
J.
Quoted 15 lines · Baruch Siach
This patch series adds support for the serial ATAG, and uses this functionality to pass the i.MX25 chip UID as a serial number to the Linux kernel.
Baruch Siach (3):
arm: add support for the serial ATAG
imx25: add support for UID read from eFuse
mx25 3ds: set CPU UID
 arch/arm/boards/freescale-mx25-3-stack/3stack.c |    2 ++
 arch/arm/include/asm/armlinux.h                 |    4 ++++
 arch/arm/lib/armlinux.c                         |   19 +++++++++++++++++++
 arch/arm/mach-imx/imx25.c                       |   12 ++++++++++++
 arch/arm/mach-imx/include/mach/generic.h        |    1 +
 arch/arm/mach-imx/include/mach/imx25-regs.h     |    3 +++
 6 files changed, 41 insertions(+), 0 deletions(-)

Signature · 8 lines

                                                     ~. .~   Tk Open Systems
=}------------------------------------------------ooO--U--Ooo------------{=
   - [email protected] - tel: +972.2.679.5364, http://www.tkos.co.il -

_______________________________________________
barebox mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/barebox

Jean-Christophe PLAGNIOL-VILLARD

2010-08-04 10:30 UTC #7 lore raw

Quoted 8 lines · Baruch Siach

Hi Jean-Christophe,
Quoted 4 lines · Jean-Christophe PLAGNIOL-VILLARD
serial number is more for board ref than CPU ref as you can easly read it from the register.
So why pass it via the ATAG?

Quoted 8 lines

The point of the serial number ATAG is to uniquely identify the system. Getting this unique ID from the CPU is as good as any other IMO. I'm not aware of any other UID source on the mx25 3DS system. Besides, having the serial number nicely listed in /proc/cpuinfo is far easier than reading registers.
Anyway, patches 1 and 2 in this series are useful on their own, I think. I intend to use this functionality for our own board. I've added the mx25 3ds two lines patch to justify the bloat :).

I agree about the other patch I just do not think we need to pass the CPUID to the kernel as he can easly get it by himself IMHO

Best Regards,
J.

_______________________________________________
barebox mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/barebox