Skip to contentSkip to conversation

Random small sunxi fixes

PATCH New megous at megous.com 4 participants 11 messages Feb 9, 2019 – Feb 19, 2019

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

cover Random small sunxi fixes

From: Ondrej Jirman <[email protected]>

Patch 1 and 2 are fixes for compilation failures with disabled network support on sunxi platform.

Patch 3 is a fix for error path bug in decode_timing_property(), I've found while debugging a missing LCD panel output.

Patch 4 is typo fix.

regards,
Ondřej Jirman

Ondrej Jirman (4):
sunxi: Fix build when CONFIG_CMD_NET is disabled
sunxi: Fix build when CONFIG_CMD_PXE or CONFIG_CMD_DHCP are disabled
lib: decode_timing_property should return -NOENT when proprty is not
found
power: axp818: Fix typo in axp_set_dldo

 board/sunxi/board.c            |  7 ++++++-
 drivers/power/axp818.c         |  2 +-
 include/configs/sunxi-common.h | 16 ++++++++++++++--
 lib/fdtdec.c                   |  2 +-
 4 files changed, 22 insertions(+), 5 deletions(-)
Signature · 1 line
2.20.1
#1 lore raw

PATCH 1/4 sunxi: Fix build when CONFIG_CMD_NET is disabled

From: Ondrej Jirman <[email protected]>

---
 board/sunxi/board.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)
board/sunxi/board.c +6 -1
index 98bc3cd0c1..e918bdf36b 100644
@@ -774,9 +774,12 @@ static void setup_environment(const void *fdt)
 {
 	char serial_string[17] = { 0 };
 	unsigned int sid[4];
+#if defined CONFIG_CMD_NET
 	uint8_t mac_addr[6];
 	char ethaddr[16];
-	int i, ret;
+	int i;
+#endif
+	int ret;
 
 	ret = sunxi_get_sid(sid);
 	if (ret == 0 && sid[0] != 0) {
@@ -801,6 +804,7 @@ static void setup_environment(const void *fdt)
 		if ((sid[3] & 0xffffff) == 0)
 			sid[3] |= 0x800000;
 
+#if defined CONFIG_CMD_NET
 		for (i = 0; i < 4; i++) {
 			sprintf(ethaddr, "ethernet%d", i);
 			if (!fdt_get_alias(fdt, ethaddr))
@@ -824,6 +828,7 @@ static void setup_environment(const void *fdt)
 
 			eth_env_set_enetaddr(ethaddr, mac_addr);
 		}
+#endif
 
 		if (!env_get("serial#")) {
 			snprintf(serial_string, sizeof(serial_string),
Signature · 1 line
2.20.1
#2 lore raw

PATCH 1/4 sunxi: Fix build when CONFIG_CMD_NET is disabled

Hi Ondřej,

Quoted 34 lines · megous at megous.com

From: Ondrej Jirman <[email protected]>

---
 board/sunxi/board.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)
board/sunxi/board.c +5 -1
index 98bc3cd0c1..e918bdf36b 100644
@@ -774,9 +774,12 @@ static void setup_environment(const void *fdt)
 {
 	char serial_string[17] = { 0 };
 	unsigned int sid[4];
+#if defined CONFIG_CMD_NET
 	uint8_t mac_addr[6];
 	char ethaddr[16];
-	int i, ret;
+	int i;
+#endif
+	int ret;
 
 	ret = sunxi_get_sid(sid);
 	if (ret == 0 && sid[0] != 0) {
@@ -801,6 +804,7 @@ static void setup_environment(const void *fdt)
 		if ((sid[3] & 0xffffff) == 0)
 			sid[3] |= 0x800000;
 
+#if defined CONFIG_CMD_NET
 		for (i = 0; i < 4; i++) {
 			sprintf(ethaddr, "ethernet%d", i);
 			if (!fdt_get_alias(fdt, ethaddr))

board/sunxi/board.c static void setup_environment(const void *fdt) · line 834

Quoted 7 lines
board/sunxi/board.c +1 -0
@@ -824,6 +828,7 @@ static void setup_environment(const void *fdt)
 
 			eth_env_set_enetaddr(ethaddr, mac_addr);
 		}
+#endif
 
 		if (!env_get("serial#")) {
 			snprintf(serial_string, sizeof(serial_string),

This patch is not needed anymore. We had a similar one also in 2018.03, but after 2019.01 upgrade it wasn't needed. It was fixed in Apr-2018:
9925f1dbc38c:
net: Move enetaddr env access code to env config instead of net config

Tomas

#3 lore raw

PATCH 1/4 sunxi: Fix build when CONFIG_CMD_NET is disabled

Hi Tomas,

Quoted 43 lines · Tomas Novotny

Hi Ondřej,

Quoted 39 lines · megous at megous.com

From: Ondrej Jirman <[email protected]>

---
 board/sunxi/board.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)
board/sunxi/board.c +6 -1
index 98bc3cd0c1..e918bdf36b 100644
@@ -774,9 +774,12 @@ static void setup_environment(const void *fdt)
 {
 	char serial_string[17] = { 0 };
 	unsigned int sid[4];
+#if defined CONFIG_CMD_NET
 	uint8_t mac_addr[6];
 	char ethaddr[16];
-	int i, ret;
+	int i;
+#endif
+	int ret;
 
 	ret = sunxi_get_sid(sid);
 	if (ret == 0 && sid[0] != 0) {
@@ -801,6 +804,7 @@ static void setup_environment(const void *fdt)
 		if ((sid[3] & 0xffffff) == 0)
 			sid[3] |= 0x800000;
 
+#if defined CONFIG_CMD_NET
 		for (i = 0; i < 4; i++) {
 			sprintf(ethaddr, "ethernet%d", i);
 			if (!fdt_get_alias(fdt, ethaddr))
@@ -824,6 +828,7 @@ static void setup_environment(const void *fdt)
 
 			eth_env_set_enetaddr(ethaddr, mac_addr);
 		}
+#endif
 
Quoted 7 lines

board/sunxi/board.c static void setup_environment(const void *fdt) · line 834

Quoted 2 lines
board/sunxi/board.c +0 -0
 		if (!env_get("serial#")) {
 			snprintf(serial_string, sizeof(serial_string),

This patch is not needed anymore. We had a similar one also in 2018.03, but after 2019.01 upgrade it wasn't needed. It was fixed in Apr-2018:
9925f1dbc38c:
net: Move enetaddr env access code to env config instead of net config

Indeed. I just tested with all my boards, and it's not necessary anymore.

The other one still is though.

thanks,
o.

Quoted 1 line

Tomas

#4 lore raw

PATCH 2/4 sunxi: Fix build when CONFIG_CMD_PXE or CONFIG_CMD_DHCP are disabled

In reply to megous at megous.com

From: Ondrej Jirman <[email protected]>

---
 include/configs/sunxi-common.h | 16 ++++++++++++++--
 1 file changed, 14 insertions(+), 2 deletions(-)
include/configs/sunxi-common.h +14 -2
index b01d1c3c84..d421b93509 100644
@@ -399,6 +399,18 @@ extern int soft_i2c_gpio_scl;
 #define BOOT_TARGET_DEVICES_USB(func)
 #endif
 
+#ifdef CONFIG_CMD_PXE
+#define BOOT_TARGET_DEVICES_PXE(func) func(PXE, pxe, na)
+#else
+#define BOOT_TARGET_DEVICES_PXE(func)
+#endif
+
+#ifdef CONFIG_CMD_DHCP
+#define BOOT_TARGET_DEVICES_DHCP(func) func(DHCP, dhcp, na)
+#else
+#define BOOT_TARGET_DEVICES_DHCP(func)
+#endif
+
 /* FEL boot support, auto-execute boot.scr if a script address was provided */
 #define BOOTENV_DEV_FEL(devtypeu, devtypel, instance) \
 	"bootcmd_fel=" \
@@ -414,8 +426,8 @@ extern int soft_i2c_gpio_scl;
 	BOOT_TARGET_DEVICES_MMC(func) \
 	BOOT_TARGET_DEVICES_SCSI(func) \
 	BOOT_TARGET_DEVICES_USB(func) \
-	func(PXE, pxe, na) \
-	func(DHCP, dhcp, na)
+	BOOT_TARGET_DEVICES_PXE(func) \
+	BOOT_TARGET_DEVICES_DHCP(func)
 
 #ifdef CONFIG_OLD_SUNXI_KERNEL_COMPAT
 #define BOOTCMD_SUNXI_COMPAT \
Signature · 1 line
2.20.1
#6 lore raw

PATCH 3/4 lib: decode_timing_property should return -NOENT when proprty is not found

In reply to megous at megous.com

From: Ondrej Jirman <[email protected]>

Length may be unitialized when fdt_getprop fails.

---
 lib/fdtdec.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
lib/fdtdec.c +1 -1
index 18663ce6bd..3510de78e1 100644
@@ -980,7 +980,7 @@ static int decode_timing_property(const void *blob, int node, const char *name,
 	if (!prop) {
 		debug("%s: could not find property %s\n",
 		      fdt_get_name(blob, node, NULL), name);
-		return length;
+		return -ENOENT;
 	}
 
 	if (length == sizeof(u32)) {
Signature · 1 line
2.20.1
#7 lore raw

PATCH 3/4 lib: decode_timing_property should return -NOENT when proprty is not found

Hi,

Quoted 4 lines · On Sat, 9 Feb 2019 at 06:44, <[email protected]> wrote:

From: Ondrej Jirman <[email protected]>

Length may be unitialized when fdt_getprop fails.

Not according to the docs:

* if lenp is non-NULL, *lenp contains the length of the property
* value (>=0)
* NULL, on error
* if lenp is non-NULL, *lenp contains an error code (<0):
* -FDT_ERR_NOTFOUND, node does not have named property
* -FDT_ERR_BADOFFSET, nodeoffset did not point to FDT_BEGIN_NODE
* tag
* -FDT_ERR_BADMAGIC,
* -FDT_ERR_BADVERSION,
* -FDT_ERR_BADSTATE,
* -FDT_ERR_BADSTRUCTURE,
* -FDT_ERR_TRUNCATED, standard meanings

I suspect there might be a bug. Can you provide a few more details?

Quoted 20 lines
---
 lib/fdtdec.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
lib/fdtdec.c +1 -2
index 18663ce6bd..3510de78e1 100644
@@ -980,7 +980,7 @@ static int decode_timing_property(const void *blob, int node, const char *name,
        if (!prop) {
                debug("%s: could not find property %s\n",
                      fdt_get_name(blob, node, NULL), name);
-               return length;
+               return -ENOENT;
        }
 
        if (length == sizeof(u32)) {
--
Quoted 2 lines

2.20.1

Regards,
Simon

#8 lore raw

PATCH 3/4 lib: decode_timing_property should return -NOENT when proprty is not found

Hi,

Quoted 10 lines · Simon Glass

Hi,

Quoted 4 lines · On Sat, 9 Feb 2019 at 06:44, <[email protected]> wrote:

From: Ondrej Jirman <[email protected]>

Length may be unitialized when fdt_getprop fails.

Not according to the docs:

Quoted 12 lines

* if lenp is non-NULL, *lenp contains the length of the property
* value (>=0)
* NULL, on error
* if lenp is non-NULL, *lenp contains an error code (<0):
* -FDT_ERR_NOTFOUND, node does not have named property
* -FDT_ERR_BADOFFSET, nodeoffset did not point to FDT_BEGIN_NODE
* tag
* -FDT_ERR_BADMAGIC,
* -FDT_ERR_BADVERSION,
* -FDT_ERR_BADSTATE,
* -FDT_ERR_BADSTRUCTURE,
* -FDT_ERR_TRUNCATED, standard meanings

Ah, you're right.

Quoted 1 line

I suspect there might be a bug. Can you provide a few more details?

The return value of multiple decode_timing_property() calls is ORed together in such a way in fdtdec_decode_display_timing() that it's pretty meaningless what it is for each call, other than that it's < 0.

So you can ignore this patch.

regards,
o.

Quoted 25 lines
Quoted 20 lines
---
 lib/fdtdec.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
lib/fdtdec.c +1 -2
index 18663ce6bd..3510de78e1 100644
@@ -980,7 +980,7 @@ static int decode_timing_property(const void *blob, int node, const char *name,
        if (!prop) {
                debug("%s: could not find property %s\n",
                      fdt_get_name(blob, node, NULL), name);
-               return length;
+               return -ENOENT;
        }
 
        if (length == sizeof(u32)) {
--
Quoted 2 lines

2.20.1

Regards,
Simon

#9 lore raw

PATCH 4/4 power: axp818: Fix typo in axp_set_dldo

In reply to megous at megous.com

From: Ondrej Jirman <[email protected]>

---
 drivers/power/axp818.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/power/axp818.c +1 -1
index c737da1180..834919ddd4 100644
@@ -161,7 +161,7 @@ int axp_set_dldo(int dldo_num, unsigned int mvolt)
 	cfg = axp818_mvolt_to_cfg(mvolt, 700, 3300, 100);
 	if (dldo_num == 2 && mvolt > 3300)
 		cfg += 1 + axp818_mvolt_to_cfg(mvolt, 3400, 4200, 200);
-	ret = pmic_bus_write(AXP818_ELDO1_CTRL + (dldo_num - 1), cfg);
+	ret = pmic_bus_write(AXP818_DLDO1_CTRL + (dldo_num - 1), cfg);
 	if (ret)
 		return ret;
 
Signature · 1 line
2.20.1
#10 lore raw

PATCH 4/4 power: axp818: Fix typo in axp_set_dldo

Quoted 1 line · On Sat, Feb 9, 2019 at 7:59 PM <[email protected]> wrote:

Add commit message, it look some fix of using DLDO1 instead of ELDO1