Skip to contentSkip to conversation

Kirkwood: print_cpuinfo fixed for valid devid revid

PATCH Superseded Prafulla Wadaskar 3 participants 3 messages Sep 16, 2010 Arm (32-bit)

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 Kirkwood: print_cpuinfo fixed for valid devid revid

Earlier Device Identification register was used to detect the type for SoC, considering 88F6282 support to be added, It is not possible to detect the same using current algorithm.

With this patch, device ID is being read using PCIE devid register, also valid chip revision ID will also be read and displayed

---
 arch/arm/cpu/arm926ejs/kirkwood/cpu.c    |   35 ++++++++++++++++++++---------
 arch/arm/include/asm/arch-kirkwood/cpu.h |    2 +
 2 files changed, 26 insertions(+), 11 deletions(-)
arch/arm/cpu/arm926ejs/kirkwood/cpu.c +24 -11
index 82c978b..5feb0b5 100644
@@ -54,10 +54,11 @@ unsigned char get_random_hex(void)
 	u8 outbuf[BUFLEN];
 
 	/*
-	 * in case of 88F6281/88F6192 A0,
+	 * in case of 88F6281/88F6282/88F6192 A0,
 	 * Bit7 need to reset to generate random values in KW_REG_UNDOC_0x1470
-	 * Soc reg offsets KW_REG_UNDOC_0x1470 and KW_REG_UNDOC_0x1478 are reserved regs and
-	 * Does not have names at this moment (no errata available)
+	 * Soc reg offsets KW_REG_UNDOC_0x1470 and KW_REG_UNDOC_0x1478 are
+	 * reserved regs and does not have names at this moment
+	 * (no errata available)
 	 */
 	writel(readl(KW_REG_UNDOC_0x1478) & ~(1 << 7), KW_REG_UNDOC_0x1478);
 	for (i = 0; i < BUFLEN; i++) {
@@ -271,20 +272,32 @@ static void kw_sysrst_check(void)
 #if defined(CONFIG_DISPLAY_CPUINFO)
 int print_cpuinfo(void)
 {
-	char *name = "Unknown";
+	char *rev, name[6];
+	u16 devid = (readl(KW_REG_PCIE_DEVID) >> 16) & 0xffff;
+	u8 revid = readl(KW_REG_PCIE_REVID) & 0xff;
 
-	switch (readl(KW_REG_DEVICE_ID) & 0x03) {
-	case 1:
-		name = "88F6192_A0";
+	if ((readl(KW_REG_DEVICE_ID) & 0x03) > 2) {
+		debug("Error.. %s:Unsupported Kirkwood SoC\n", __FUNCTION__);
+		return -1;
+	}
+
+	switch (revid) {
+	case 0:
+		rev = "Z0";
 		break;
 	case 2:
-		name = "88F6281_A0";
+		rev = "A0";
+		break;
+	case 3:
+		rev = "A1";
 		break;
 	default:
-		printf("SoC:   Unsupported Kirkwood\n");
-		return -1;
+		rev = "??";
+		break;
 	}
-	printf("SoC:   Kirkwood %s\n", name);
+
+	sprintf(name, "%04x", devid);
+	printf("SoC:   Kirkwood 88F%s_%s\n", name, rev);
 	return 0;
 }
 #endif /* CONFIG_DISPLAY_CPUINFO */
arch/arm/include/asm/arch-kirkwood/cpu.h +2 -0
index b3022a3..d28c51a 100644
@@ -35,6 +35,8 @@
 #define KWGBE_PORT_SERIAL_CONTROL1_REG(_x)	\
 		((_x ? KW_EGIGA0_BASE : KW_EGIGA1_BASE) + 0x44c)
 
+#define KW_REG_PCIE_DEVID		(KW_REG_PCIE_BASE + 0x00)
+#define KW_REG_PCIE_REVID		(KW_REG_PCIE_BASE + 0x08)
 #define KW_REG_DEVICE_ID		(KW_MPP_BASE + 0x34)
 #define KW_REG_SYSRST_CNT		(KW_MPP_BASE + 0x50)
 #define SYSRST_CNT_1SEC_VAL		(25*1000000)
Signature · 1 line
1.5.3.3
#1 lore raw

PATCH 1/1 Kirkwood: print_cpuinfo fixed for valid devid revid

Hi Prafulla,

Le 16/09/2010 16:52, Prafulla Wadaskar a ?crit :

Review

Quoted 1 line
(quoted diff) +0 -0

b/arch/arm/cpu/arm926ejs/kirkwood/cpu.c

Quoted 8 lines

index 82c978b..5feb0b5 100644

arch/arm/cpu/arm926ejs/kirkwood/cpu.c +0 -0
@@ -271,20 +272,32 @@ static void kw_sysrst_check(void)
  #if defined(CONFIG_DISPLAY_CPUINFO)
  int print_cpuinfo(void)
  {

arch/arm/cpu/arm926ejs/kirkwood/cpu.c static void kw_sysrst_check(void) · line 280

Quoted 10 lines
arch/arm/cpu/arm926ejs/kirkwood/cpu.c +5 -4
-	char *name = "Unknown";
+	char *rev, name[6];
+	u16 devid = (readl(KW_REG_PCIE_DEVID)>>  16)&  0xffff;
+	u8 revid = readl(KW_REG_PCIE_REVID)&  0xff;
 
-	switch (readl(KW_REG_DEVICE_ID)&  0x03) {
-	case 1:
-		name = "88F6192_A0";
+	if ((readl(KW_REG_DEVICE_ID)&  0x03)>  2) {
+		debug("Error.. %s:Unsupported Kirkwood SoC\n", __FUNCTION__);

Maybe you could add the actual content of the device ID register so that the operator can read it on the console?

Quoted 8 lines
arch/arm/cpu/arm926ejs/kirkwood/cpu.c +6 -0
+		return -1;
+	}
+
+	switch (revid) {
+	case 0:
+		rev = "Z0";
  		break;
  	case 2:

arch/arm/cpu/arm926ejs/kirkwood/cpu.c static void kw_sysrst_check(void) · line 295

Quoted 10 lines
arch/arm/cpu/arm926ejs/kirkwood/cpu.c +5 -3
-		name = "88F6281_A0";
+		rev = "A0";
+		break;
+	case 3:
+		rev = "A1";
  		break;
  	default:
-		printf("SoC:   Unsupported Kirkwood\n");
-		return -1;
+		rev = "??";

Same here with this "??": replacing it with the actual revision ID register value would allow the operator to read it directly on the console.

Amicalement,

Signature · 1 line
Albert.
#2 lore raw

PATCH 1/1 Kirkwood: print_cpuinfo fixed for valid devid revid

In reply to Prafulla Wadaskar

Dear Prafulla Wadaskar,

Quoted 10 lines · In message <[email protected]> you wrote:

Earlier Device Identification register was used to detect the type for SoC, considering 88F6282 support to be added, It is not possible to detect the same using current algorithm.

With this patch, device ID is being read using PCIE devid register, also valid chip revision ID will also be read and displayed

...

arch/arm/cpu/arm926ejs/kirkwood/cpu.c static void kw_sysrst_check(void) · line 281

Quoted 3 lines
arch/arm/cpu/arm926ejs/kirkwood/cpu.c +3 -0
+	if ((readl(KW_REG_DEVICE_ID) & 0x03) > 2) {
+		debug("Error.. %s:Unsupported Kirkwood SoC\n", __FUNCTION__);
+		return -1;

Please use a printf() instead of debug() here - error messages should always be printed.

arch/arm/cpu/arm926ejs/kirkwood/cpu.c static void kw_sysrst_check(void) · line 300

Quoted 2 lines
arch/arm/cpu/arm926ejs/kirkwood/cpu.c +2 -0
+	sprintf(name, "%04x", devid);
+	printf("SoC:   Kirkwood 88F%s_%s\n", name, rev);

Why do you need a separate sprintf() here? Why not simply:

	printf("SoC:   Kirkwood 88F%04x_%s\n", devid, rev);

?

Best regards,

Wolfgang Denk

Signature · 6 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
"Am besten betrachten Sie Fehlermeldungen als eine  Art  Psycho-Test,
mit  dem  herausgefunden  werden soll, wie belastbar Sie sind."
 - Dr. R. Wonneberger, Kompaktf?hrer LaTeX, Kap. 1.6: Fehlermeldungen