Skip to contentSkip to conversation

83xx, kmeter1: added bootcount feature

PATCH New Heiko Schocher 3 participants 5 messages Jan 28, 2009 – Jan 29, 2009

lore
Share & tools
lore
#0 lore raw

PATCH 31/31 83xx, kmeter1: added bootcount feature

add CONFIG_BOUTCOUNT_LIMIT feature for 83xx CPU.
activate this feature for the kmeter1 board.

---
 cpu/mpc83xx/cpu.c                |   27 +++++++++++++++++++++++++++
 include/configs/keymile-common.h |    2 --
 2 files changed, 27 insertions(+), 2 deletions(-)
cpu/mpc83xx/cpu.c +27 -0
index 587fca3..7c12672 100644
@@ -34,6 +34,9 @@
 #include <libfdt.h>
 #include <tsec.h>
 #include <netdev.h>
+#ifdef CONFIG_BOOTCOUNT_LIMIT
+#include <asm/io.h>
+#endif
 
 DECLARE_GLOBAL_DATA_PTR;
 
@@ -385,3 +388,27 @@ int cpu_eth_init(bd_t *bis)
 #endif
 	return 0;
 }
+
+#ifdef CONFIG_BOOTCOUNT_LIMIT
+
+#if !defined(CONFIG_BOOTCOUNT_ADDR)
+#define CONFIG_BOOTCOUNT_ADDR	0x11bff8
+#endif
+
+void bootcount_store (ulong a)
+{
+	void *reg = (unsigned __iomem *)(CONFIG_SYS_IMMR + CONFIG_BOOTCOUNT_ADDR);
+	out_be32 (reg, a);
+	out_be32 (reg + 4, BOOTCOUNT_MAGIC);
+}
+
+ulong bootcount_load (void)
+{
+	void *reg = (unsigned __iomem *)(CONFIG_SYS_IMMR + CONFIG_BOOTCOUNT_ADDR);
+
+	if (in_be32 (reg + 4) != BOOTCOUNT_MAGIC)
+		return 0;
+	else
+		return in_be32 (reg);
+}
+#endif /* CONFIG_BOOTCOUNT_LIMIT */
include/configs/keymile-common.h +0 -2
index 3736cd8..6a25262 100644
@@ -27,9 +27,7 @@
 /* Do boardspecific init for all boards */
 #define CONFIG_BOARD_EARLY_INIT_R       1
 
-#if defined(CONFIG_MGCOGE) || defined(CONFIG_MGSUVD) || defined(CONFIG_KMSUPX4)
 #define CONFIG_BOOTCOUNT_LIMIT
-#endif
 
 /*
  * Command line configuration.
Signature · 5 lines
1.6.0.6

-- 
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
#1 lore raw

PATCH 31/31 83xx, kmeter1: added bootcount feature

cpu/mpc83xx/cpu.c int cpu_eth_init(bd_t *bis) · line 395

Quoted 6 lines · On Wed, 28 Jan 2009 10:41:02 +0100 Heiko Schocher <[email protected]> wrote:
cpu/mpc83xx/cpu.c +4 -0
 
+#ifdef CONFIG_BOOTCOUNT_LIMIT
+
+#if !defined(CONFIG_BOOTCOUNT_ADDR)
+#define CONFIG_BOOTCOUNT_ADDR	0x11bff8

it's CONFIG_SYS and magic number, but now that I've found it, sorry, I can't accept this - not all 83xx have a QE and thus this MURAM. Even if they did this would require changing the device tree muram node property, to indicate its size has shrunk 8 bytes, right? In fact, not all QE's have 48Kbytes of MURAM either - the 8323 only has 16Kbytes. Problem is, I don't know of a better place to put the bootcount. Ideas?

Kim

#2 lore raw

PATCH 31/31 83xx, kmeter1: added bootcount feature

Hello Kim,

Quoted 5 lines · Kim Phillips
Quoted 1 line · On Wed, 28 Jan 2009 10:41:02 +0100 Heiko Schocher <[email protected]> wrote:
cpu/mpc83xx/cpu.c +0 -0
Quoted 7 lines

cpu/mpc83xx/cpu.c int cpu_eth_init(bd_t *bis) · line 395

Quoted 4 lines
cpu/mpc83xx/cpu.c +4 -0
+#ifdef CONFIG_BOOTCOUNT_LIMIT
+
+#if !defined(CONFIG_BOOTCOUNT_ADDR)
+#define CONFIG_BOOTCOUNT_ADDR	0x11bff8

it's CONFIG_SYS and magic number, but now that I've found it, sorry, I can't accept this - not all 83xx have a QE and thus this MURAM. Even

Hmm.. maybe we make this dependent on CONFIG_QE ?

Quoted 2 lines

if they did this would require changing the device tree muram node property, to indicate its size has shrunk 8 bytes, right? In fact, not

Yes, we should do this.

Quoted 1 line

all QE's have 48Kbytes of MURAM either - the 8323 only has 16Kbytes.

Thats why you could define it with CONFIG_BOOTCOUNT_ADDR, where exactly this 8 bytes are.

Quoted 2 lines

Problem is, I don't know of a better place to put the bootcount. Ideas?

I thought to make it as on 82xx (using parameter Ram of SCC1), because UCC5 should be compatible to the SCC1 see 8360ERM.pdf Table 19-11 on page 19-20. But on the 8360, after reset, the complete parameter RAM is initialized with 0 ... so that didn't work. Other places I couldn't found for this feature :-(

bye
Heiko

Signature · 2 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
#3 lore raw

PATCH 31/31 83xx, kmeter1: added bootcount feature

Quoted 16 lines · On Thu, 29 Jan 2009 10:22:06 +0100 Heiko Schocher <[email protected]> wrote:

Hello Kim,

Quoted 5 lines · Kim Phillips
Quoted 1 line · On Wed, 28 Jan 2009 10:41:02 +0100 Heiko Schocher <[email protected]> wrote:
cpu/mpc83xx/cpu.c +0 -0
Quoted 7 lines

cpu/mpc83xx/cpu.c int cpu_eth_init(bd_t *bis) · line 395

Quoted 4 lines
cpu/mpc83xx/cpu.c +4 -0
+#ifdef CONFIG_BOOTCOUNT_LIMIT
+
+#if !defined(CONFIG_BOOTCOUNT_ADDR)
+#define CONFIG_BOOTCOUNT_ADDR	0x11bff8

it's CONFIG_SYS and magic number, but now that I've found it, sorry, I can't accept this - not all 83xx have a QE and thus this MURAM. Even

Quoted 1 line

Hmm.. maybe we make this dependent on CONFIG_QE ?

still, the 8323 and 8360 would need different address values...so not only CONFIG_QE.

Quoted 4 lines
Quoted 2 lines

if they did this would require changing the device tree muram node property, to indicate its size has shrunk 8 bytes, right? In fact, not

Yes, we should do this.

I hope the QE drivers don't significantly drop their performance given a muram size is not a multiple of a power of 2.

Quoted 4 lines
Quoted 1 line

all QE's have 48Kbytes of MURAM either - the 8323 only has 16Kbytes.

Thats why you could define it with CONFIG_BOOTCOUNT_ADDR, where exactly this 8 bytes are.

it would be nice if all 83xx had the same address though.

Quoted 8 lines
Quoted 2 lines

Problem is, I don't know of a better place to put the bootcount. Ideas?

I thought to make it as on 82xx (using parameter Ram of SCC1), because UCC5 should be compatible to the SCC1 see 8360ERM.pdf Table 19-11 on page 19-20. But on the 8360, after reset, the complete parameter RAM is initialized with 0 ... so that didn't work. Other places I couldn't found for this feature :-(

you thinking is right, I just can't think of another place to put it - and I suspect this isn't something that can be put in memory that lies outside the IMMR.

if you really want this, make it only for the QE based chips, and include the device tree modification. I still don't know if this should be enabled by default in the configs, however - debugging QE performance down to this change would be relatively hard to find.

Kim

#4 lore raw

PATCH 31/31 83xx, kmeter1: added bootcount feature

Dear Kim Phillips,

Quoted 4 lines · In message <[email protected]> you wrote:

you thinking is right, I just can't think of another place to put it - and I suspect this isn't something that can be put in memory that lies outside the IMMR.

We could use any registers that are guaranteed to maintaion their current values through a core reset. Documentation usually gives little or no guarantees, though.

Best regards,

Wolfgang Denk

Signature · 5 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
Microsoft Multimedia:
You have nice graphics, sound and animations when the system crashes.