pinctrl: imx1-core: Fix debug output pin array index

PATCH New Markus Pargmann 3 participants 7 messages Jul 8, 2015 – Jul 10, 2015

Subsystems (4)

Stable branch Arm i.MX and Layerscape Pin control Pin controller - Freescale

lore

Share & tools

lore

Markus Pargmann

author

2015-07-08 14:16 UTC #0 lore raw

PATCH 1/1 pinctrl: imx1-core: Fix debug output pin array index

The pins do not have a 1:1 mapping from index to pin_id. Unfortunately the debug output assumes exactly that.

The first driver using imx1-core was imx27, which had exactly this 1:1 mapping. It was accidently removed when removing all unused pads which were listed:
607af165c047 (pinctrl: i.MX27: Remove nonexistent pad definitions)

The patch fixes this issue by printing the pin_id directly and not the pin name.

Cc <[email protected]>

Reported-by Chris Ruehl <[email protected]>

Signed-off-by Markus Pargmann <[email protected]>

---
 drivers/pinctrl/freescale/pinctrl-imx1-core.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

drivers/pinctrl/freescale/pinctrl-imx1-core.c +1 -3

index 5fd4437cee15..f08858769167 100644
@@ -403,14 +403,12 @@ static int imx1_pinconf_set(struct pinctrl_dev *pctldev,
 			     unsigned num_configs)
 {
 	struct imx1_pinctrl *ipctl = pinctrl_dev_get_drvdata(pctldev);
-	const struct imx1_pinctrl_soc_info *info = ipctl->info;
 	int i;
 
 	for (i = 0; i != num_configs; ++i) {
 		imx1_write_bit(ipctl, pin_id, configs[i] & 0x01, MX1_PUEN);
 
-		dev_dbg(ipctl->dev, "pinconf set pullup pin %s\n",
-			info->pins[pin_id].name);
+		dev_dbg(ipctl->dev, "pinconf set pullup pin %d\n", pin_id);
 	}
 
 	return 0;

Signature · 1 line

2.1.4

Uwe Kleine-König

2015-07-09 07:12 UTC #1 lore raw

PATCH 1/1 pinctrl: imx1-core: Fix debug output pin array index

Hello Markus,

Cc += Chris Ruehl

Quoted 10 lines · Markus Pargmann

The pins do not have a 1:1 mapping from index to pin_id. Unfortunately the debug output assumes exactly that.
The first driver using imx1-core was imx27, which had exactly this 1:1 mapping. It was accidently removed when removing all unused pads which were listed:
607af165c047 (pinctrl: i.MX27: Remove nonexistent pad definitions)
The patch fixes this issue by printing the pin_id directly and not the pin name.

Knowing a bit about the imx pinctrl drivers I failed to understand what you wrote here. Probably because I first though that "1:1 mapping" is a hardware property. What about:

Subject: pinctrl: imx1-core: Fix debug output in .pin_config_set callback

imx1_pinconf_set assumes that the array of pins in struct imx1_pinctrl_soc_info can be indexed by pin id to get the pinctrl_pin_desc for a pin. This used to be correct up to commit 607af165c047 ("pinctrl: i.MX27: Remove nonexistent pad definitions") which removed some entries from the array and so made it wrong to access the array by pin id.

Implement the easiest fix by not resolving the pin id to a name but printing the id instead.

Fixes 607af165c047 ("pinctrl: i.MX27: Remove nonexistent pad definitions")

Cc stable at vger.kernel.org

Reported-by Chris Ruehl <[email protected]>

Signed-off-by Markus Pargmann <[email protected]>

Having the pad name in the output is nice, is it worth to search for the right pinctrl_pin_desc in the array? The array is still sorted, so a binary search would do, maybe a function for this already exists?

How did Chris notice the error? Just a bogus output, or did it crash the kernel? That would be worth to note in the commit log, too.

Otherwise the change looks fine.

Best regards
Uwe

Signature · 2 lines

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

Chris Ruehl

2015-07-09 08:28 UTC #2 lore raw

PATCH 1/1 pinctrl: imx1-core: Fix debug output pin array index

Uwe,

Quoted 40 lines · Uwe Kleine-K?nig

Hello Markus,
Cc += Chris Ruehl
Quoted 10 lines · Markus Pargmann
The pins do not have a 1:1 mapping from index to pin_id. Unfortunately the debug output assumes exactly that.
The first driver using imx1-core was imx27, which had exactly this 1:1 mapping. It was accidently removed when removing all unused pads which were listed:
607af165c047 (pinctrl: i.MX27: Remove nonexistent pad definitions)
The patch fixes this issue by printing the pin_id directly and not the pin name.
Knowing a bit about the imx pinctrl drivers I failed to understand what you wrote here. Probably because I first though that "1:1 mapping" is a hardware property. What about:
Subject: pinctrl: imx1-core: Fix debug output in .pin_config_set callback
imx1_pinconf_set assumes that the array of pins in struct imx1_pinctrl_soc_info can be indexed by pin id to get the pinctrl_pin_desc for a pin. This used to be correct up to commit 607af165c047 ("pinctrl: i.MX27: Remove nonexistent pad definitions") which removed some entries from the array and so made it wrong to access the array by pin id.
Implement the easiest fix by not resolving the pin id to a name but printing the id instead.
Fixes 607af165c047 ("pinctrl: i.MX27: Remove nonexistent pad definitions")
Cc stable at vger.kernel.org
Reported-by Chris Ruehl <[email protected]>
Signed-off-by Markus Pargmann <[email protected]>
Having the pad name in the output is nice, is it worth to search for the right pinctrl_pin_desc in the array? The array is still sorted, so a binary search would do, maybe a function for this already exists?

Quoted 7 lines

How did Chris notice the error? Just a bogus output, or did it crash the kernel? That would be worth to note in the commit log, too.
Otherwise the change looks fine.
Best regards
Uwe

I had a crash on a array overrun.

I'd applied a patch to the linux-usb (which was spit out by the maintainer perl script)
chris at wheezyvm:~/kernel.d/linux-next$ cat 0001-Pinctrl-imx1-fix-wrong-pin-name-resolving.patch
From 50d56e5f626b2ea86211818cf487514c96f60487 Mon Sep 17 00:00:00 2001
From: Chris Ruehl <[email protected]>
Date: Sat, 23 May 2015 15:02:44 +0800
Subject: [PATCH] Pinctrl: imx1: fix wrong pin-name resolving

Bug in function imx1_pinconf_set() cause crash when princtrl debug is enabled and the pin_id becomes larger then the info->pins[] contains.

imx27-pinctrl 10015000.iomuxc: request pin 134 (MX27_PAD_UART2_TXD) for 1000b000.serial
imx27-pinctrl 10015000.iomuxc: request pin 135 (MX27_PAD_UART2_RXD) for 1000b000.serial
imx27-pinctrl 10015000.iomuxc: request pin 131 (MX27_PAD_UART2_CTS) for 1000b000.serial
imx27-pinctrl 10015000.iomuxc: request pin 132 (MX27_PAD_UART2_RTS) for 1000b000.serial
imx27-pinctrl 10015000.iomuxc: enable function uart group uart2-1 imx27-pinctrl 10015000.iomuxc: imx1_pmx_set, pin 0x86, function 0, gpio 0, direction 1, oconf 0, iconfa 0, iconfb 0
imx27-pinctrl 10015000.iomuxc: imx1_pmx_set, pin 0x87, function 0, gpio 0, direction 0, oconf 0, iconfa 0, iconfb 0
imx27-pinctrl 10015000.iomuxc: imx1_pmx_set, pin 0x83, function 0, gpio 0, direction 1, oconf 0, iconfa 0, iconfb 0
imx27-pinctrl 10015000.iomuxc: imx1_pmx_set, pin 0x84, function 0, gpio 0, direction 0, oconf 0, iconfa 0, iconfb 0
imx27-pinctrl 10015000.iomuxc: num_configs=1 PinID=134 imx27-pinctrl 10015000.iomuxc: pinconf set pullup pin MX27_PAD_UART1_RTS imx27-pinctrl 10015000.iomuxc: num_configs=1 PinID=135 imx27-pinctrl 10015000.iomuxc: pinconf set pullup pin MX27_PAD_RTCK imx27-pinctrl 10015000.iomuxc: num_configs=1 PinID=131 imx27-pinctrl 10015000.iomuxc: pinconf set pullup pin MX27_PAD_UART1_TXD imx27-pinctrl 10015000.iomuxc: num_configs=1 PinID=132 imx27-pinctrl 10015000.iomuxc: pinconf set pullup pin MX27_PAD_UART1_RXD
...
imx27-pinctrl 10015000.iomuxc: write: register 0xf4415508 offset 4 value 0x3 imx27-pinctrl 10015000.iomuxc: write: register 0xf4415510 offset 4 value 0x0 imx27-pinctrl 10015000.iomuxc: write: register 0xf4415518 offset 4 value 0x0 imx27-pinctrl 10015000.iomuxc: imx1_pmx_set, pin 0xb5, function 0, gpio 0, direction 1, oconf 0, iconfa 0, iconfb 0
imx27-pinctrl 10015000.iomuxc: num_configs=1 PinID=171

Unable to handle kernel paging request at virtual address 6c61765f
pgd = c0004000
6c61765f] *pgd=00000000
Internal error: Oops: 5 [#1] ARM
Modules linked in:
CPU: 0 PID: 1 Comm: swapper Not tainted 4.1.0-rc4-next-20150522-dirty #8
Hardware name: GTSYS i.MX27GTSIR (Device Tree Support)
task: ce832000 ti: ce848000 task.ti: ce848000
PC is at strnlen+0x28/0x3c
LR is at string.isra.4+0x34/0xcc
pc : [<c01ae188>]    lr : [<c01af9a4>]    psr: 20000093

sp : ce849a88 ip : ce849a98 fp : ce849a94
r10: c05d7e3a r9 : c05d81e4 r8 : 00000000
r7 : 6c61765f r6 : c05d81e4 r5 : ffffffff r4 : c05d7e3a r3 : 6c61765f r2 : 6c61765f r1 : 6c61765e r0 : 6c61765f
Flags: nzCv IRQs off FIQs on Mode SVC_32 ISA ARM Segment kernel

Signed-off-by Chris Ruehl <[email protected]>

---
  drivers/pinctrl/freescale/pinctrl-imx1-core.c |    6 ++++--
  1 file changed, 4 insertions(+), 2 deletions(-)

(quoted diff) +0 -0

b/drivers/pinctrl/freescale/pinctrl-imx1-core.c
index 5ac59fb..8408bd8 100644

drivers/pinctrl/freescale/pinctrl-imx1-core.c +4 -2

@@ -403,14 +403,16 @@ static int imx1_pinconf_set(struct pinctrl_dev *pctldev,
  			     unsigned num_configs)
  {
  	struct imx1_pinctrl *ipctl = pinctrl_dev_get_drvdata(pctldev);
-	const struct imx1_pinctrl_soc_info *info = ipctl->info;
+	struct pin_desc *desc;
  	int i;
 
+	desc = pin_desc_get(pctldev, pin_id);
+
  	for (i = 0; i != num_configs; ++i) {
  		imx1_write_bit(ipctl, pin_id, configs[i] & 0x01, MX1_PUEN);
 
  		dev_dbg(ipctl->dev, "pinconf set pullup pin %s\n",
-			info->pins[pin_id].name);
+			desc->name);
  	}
 
  	return 0;

Signature · 12 lines

1.7.10.4




-- 
GTSYS Limited RFID Technology
9/F, Unit E, R07, Kwai Shing Industrial Building Phase 2,
42-46 Tai Lin Pai Road, Kwai Chung, N.T., Hong Kong
Tel (852) 9079 9521

Disclaimer: http://www.gtsys.com.hk/email/classified.html

Markus Pargmann

author

2015-07-09 13:30 UTC #3 lore raw

PATCH 1/1 pinctrl: imx1-core: Fix debug output pin array index

Hi,

Quoted 48 lines · Chris Ruehl

Uwe,
Quoted 45 lines · Uwe Kleine-K?nig
Hello Markus,
Cc += Chris Ruehl
Quoted 10 lines · Markus Pargmann
The pins do not have a 1:1 mapping from index to pin_id. Unfortunately the debug output assumes exactly that.
The first driver using imx1-core was imx27, which had exactly this 1:1 mapping. It was accidently removed when removing all unused pads which were listed:
607af165c047 (pinctrl: i.MX27: Remove nonexistent pad definitions)
The patch fixes this issue by printing the pin_id directly and not the pin name.
Knowing a bit about the imx pinctrl drivers I failed to understand what you wrote here. Probably because I first though that "1:1 mapping" is a hardware property. What about:
Subject: pinctrl: imx1-core: Fix debug output in .pin_config_set callback
imx1_pinconf_set assumes that the array of pins in struct imx1_pinctrl_soc_info can be indexed by pin id to get the pinctrl_pin_desc for a pin. This used to be correct up to commit 607af165c047 ("pinctrl: i.MX27: Remove nonexistent pad definitions") which removed some entries from the array and so made it wrong to access the array by pin id.
Implement the easiest fix by not resolving the pin id to a name but printing the id instead.
Fixes 607af165c047 ("pinctrl: i.MX27: Remove nonexistent pad definitions")
Cc stable at vger.kernel.org
Reported-by Chris Ruehl <[email protected]>
Signed-off-by Markus Pargmann <[email protected]>
Having the pad name in the output is nice, is it worth to search for the right pinctrl_pin_desc in the array? The array is still sorted, so a binary search would do, maybe a function for this already exists?
How did Chris notice the error? Just a bogus output, or did it crash the kernel? That would be worth to note in the commit log, too.
Otherwise the change looks fine.

Quoted 7 lines

Quoted 2 lines
Best regards
Uwe
I had a crash on a array overrun.
I'd applied a patch to the linux-usb (which was spit out by the maintainer perl script)

Yes, as there is already a patch from Chris my patch can be dropped. We should continue the discussion on your patch instead.

Quoted 72 lines

chris at wheezyvm:~/kernel.d/linux-next$ cat 0001-Pinctrl-imx1-fix-wrong-pin-name-resolving.patch
From 50d56e5f626b2ea86211818cf487514c96f60487 Mon Sep 17 00:00:00 2001
From: Chris Ruehl <[email protected]>
Date: Sat, 23 May 2015 15:02:44 +0800
Subject: [PATCH] Pinctrl: imx1: fix wrong pin-name resolving
Bug in function imx1_pinconf_set() cause crash when princtrl debug is enabled and the pin_id becomes larger then the info->pins[] contains.
imx27-pinctrl 10015000.iomuxc: request pin 134 (MX27_PAD_UART2_TXD) for 1000b000.serial
imx27-pinctrl 10015000.iomuxc: request pin 135 (MX27_PAD_UART2_RXD) for 1000b000.serial
imx27-pinctrl 10015000.iomuxc: request pin 131 (MX27_PAD_UART2_CTS) for 1000b000.serial
imx27-pinctrl 10015000.iomuxc: request pin 132 (MX27_PAD_UART2_RTS) for 1000b000.serial
imx27-pinctrl 10015000.iomuxc: enable function uart group uart2-1 imx27-pinctrl 10015000.iomuxc: imx1_pmx_set, pin 0x86, function 0, gpio 0, direction 1, oconf 0, iconfa 0, iconfb 0
imx27-pinctrl 10015000.iomuxc: imx1_pmx_set, pin 0x87, function 0, gpio 0, direction 0, oconf 0, iconfa 0, iconfb 0
imx27-pinctrl 10015000.iomuxc: imx1_pmx_set, pin 0x83, function 0, gpio 0, direction 1, oconf 0, iconfa 0, iconfb 0
imx27-pinctrl 10015000.iomuxc: imx1_pmx_set, pin 0x84, function 0, gpio 0, direction 0, oconf 0, iconfa 0, iconfb 0
imx27-pinctrl 10015000.iomuxc: num_configs=1 PinID=134 imx27-pinctrl 10015000.iomuxc: pinconf set pullup pin MX27_PAD_UART1_RTS imx27-pinctrl 10015000.iomuxc: num_configs=1 PinID=135 imx27-pinctrl 10015000.iomuxc: pinconf set pullup pin MX27_PAD_RTCK imx27-pinctrl 10015000.iomuxc: num_configs=1 PinID=131 imx27-pinctrl 10015000.iomuxc: pinconf set pullup pin MX27_PAD_UART1_TXD imx27-pinctrl 10015000.iomuxc: num_configs=1 PinID=132 imx27-pinctrl 10015000.iomuxc: pinconf set pullup pin MX27_PAD_UART1_RXD
...
imx27-pinctrl 10015000.iomuxc: write: register 0xf4415508 offset 4 value 0x3 imx27-pinctrl 10015000.iomuxc: write: register 0xf4415510 offset 4 value 0x0 imx27-pinctrl 10015000.iomuxc: write: register 0xf4415518 offset 4 value 0x0 imx27-pinctrl 10015000.iomuxc: imx1_pmx_set, pin 0xb5, function 0, gpio 0, direction 1, oconf 0, iconfa 0, iconfb 0
imx27-pinctrl 10015000.iomuxc: num_configs=1 PinID=171
Unable to handle kernel paging request at virtual address 6c61765f
pgd = c0004000
6c61765f] *pgd=00000000
Internal error: Oops: 5 [#1] ARM
Modules linked in:
CPU: 0 PID: 1 Comm: swapper Not tainted 4.1.0-rc4-next-20150522-dirty #8
Hardware name: GTSYS i.MX27GTSIR (Device Tree Support)
task: ce832000 ti: ce848000 task.ti: ce848000
PC is at strnlen+0x28/0x3c
LR is at string.isra.4+0x34/0xcc
pc : [<c01ae188>]    lr : [<c01af9a4>]    psr: 20000093
sp : ce849a88 ip : ce849a98 fp : ce849a94
r10: c05d7e3a r9 : c05d81e4 r8 : 00000000
r7 : 6c61765f r6 : c05d81e4 r5 : ffffffff r4 : c05d7e3a r3 : 6c61765f r2 : 6c61765f r1 : 6c61765e r0 : 6c61765f
Flags: nzCv IRQs off FIQs on Mode SVC_32 ISA ARM Segment kernel
Signed-off-by Chris Ruehl <[email protected]>
---
 drivers/pinctrl/freescale/pinctrl-imx1-core.c |    6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/pinctrl/freescale/pinctrl-imx1-core.c +0 -0
index 5ac59fb..8408bd8 100644

drivers/pinctrl/freescale/pinctrl-imx1-core.c static int imx1_pinconf_set(struct pinctrl_dev *pctldev, · line 409

Quoted 8 lines

drivers/pinctrl/freescale/pinctrl-imx1-core.c +2 -1

@@ -403,14 +403,16 @@ static int imx1_pinconf_set(struct pinctrl_dev *pctldev,
 			     unsigned num_configs)
 {
 	struct imx1_pinctrl *ipctl = pinctrl_dev_get_drvdata(pctldev);
-	const struct imx1_pinctrl_soc_info *info = ipctl->info;
+	struct pin_desc *desc;
 	int i;
 
+	desc = pin_desc_get(pctldev, pin_id);

For the rest of this function pin_desc_get() is not necessary. This is not a simple function call but a radix tree lookup as far as I can see. I am not sure how important this debugging information is and how expensive this function call is?!

Best Regards,

Markus

Signature · 11 lines

Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20150709/509ef35f/attachment.sig>

Uwe Kleine-König

2015-07-09 18:06 UTC #4 lore raw

PATCH 1/1 pinctrl: imx1-core: Fix debug output pin array index

Hi,

Quoted 12 lines · Markus Pargmann

Quoted 12 lines

b/drivers/pinctrl/freescale/pinctrl-imx1-core.c
index 5ac59fb..8408bd8 100644

drivers/pinctrl/freescale/pinctrl-imx1-core.c +1 -1

@@ -403,14 +403,16 @@ static int imx1_pinconf_set(struct pinctrl_dev *pctldev,
 			     unsigned num_configs)
 {
 	struct imx1_pinctrl *ipctl = pinctrl_dev_get_drvdata(pctldev);
-	const struct imx1_pinctrl_soc_info *info = ipctl->info;
+	struct pin_desc *desc;
 	int i;

Quoted 6 lines

drivers/pinctrl/freescale/pinctrl-imx1-core.c static int imx1_pinconf_set(struct pinctrl_dev *pctldev, · line 409
Quoted 1 line
drivers/pinctrl/freescale/pinctrl-imx1-core.c +1 -0
+	desc = pin_desc_get(pctldev, pin_id);
For the rest of this function pin_desc_get() is not necessary. This is not a simple function call but a radix tree lookup as far as I can see. I am not sure how important this debugging information is and how expensive this function call is?!

If it's moved into the call to pr_debug it should be optimized away when debugging is off. That together with a merge of Chris' and mine commit log and I'm happy.

Best regards
Uwe

Signature · 2 lines

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

Markus Pargmann

author

2015-07-10 10:02 UTC #5 lore raw

PATCH 1/1 pinctrl: imx1-core: Fix debug output pin array index

Hi,

Quoted 21 lines · Uwe Kleine-K?nig

Hi,
Quoted 12 lines · Markus Pargmann
Quoted 12 lines
b/drivers/pinctrl/freescale/pinctrl-imx1-core.c
index 5ac59fb..8408bd8 100644
drivers/pinctrl/freescale/pinctrl-imx1-core.c +1 -1
@@ -403,14 +403,16 @@ static int imx1_pinconf_set(struct pinctrl_dev *pctldev,
 			     unsigned num_configs)
 {
 	struct imx1_pinctrl *ipctl = pinctrl_dev_get_drvdata(pctldev);
-	const struct imx1_pinctrl_soc_info *info = ipctl->info;
+	struct pin_desc *desc;
 	int i;
 
Quoted 6 lines
drivers/pinctrl/freescale/pinctrl-imx1-core.c static int imx1_pinconf_set(struct pinctrl_dev *pctldev, · line 409
Quoted 1 line
drivers/pinctrl/freescale/pinctrl-imx1-core.c +1 -0
+	desc = pin_desc_get(pctldev, pin_id);
For the rest of this function pin_desc_get() is not necessary. This is not a simple function call but a radix tree lookup as far as I can see. I am not sure how important this debugging information is and how expensive this function call is?!

Quoted 3 lines

If it's moved into the call to pr_debug it should be optimized away when debugging is off. That together with a merge of Chris' and mine commit log and I'm happy.

Ok. If you change that I am happy as well.

Best Regards,

Markus

Signature · 11 lines

Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20150710/9ef8930b/attachment.sig>

Chris Ruehl

2015-07-10 01:24 UTC #6 lore raw

PATCH 1/1 pinctrl: imx1-core: Fix debug output pin array index

In reply to Markus Pargmann

Quoted 148 lines · Markus Pargmann

Hi,
Quoted 48 lines · Chris Ruehl
Uwe,
Quoted 45 lines · Uwe Kleine-K?nig
Hello Markus,
Cc += Chris Ruehl
Quoted 10 lines · Markus Pargmann
The pins do not have a 1:1 mapping from index to pin_id. Unfortunately the debug output assumes exactly that.
The first driver using imx1-core was imx27, which had exactly this 1:1 mapping. It was accidently removed when removing all unused pads which were listed:
607af165c047 (pinctrl: i.MX27: Remove nonexistent pad definitions)
The patch fixes this issue by printing the pin_id directly and not the pin name.
Knowing a bit about the imx pinctrl drivers I failed to understand what you wrote here. Probably because I first though that "1:1 mapping" is a hardware property. What about:
Subject: pinctrl: imx1-core: Fix debug output in .pin_config_set callback
imx1_pinconf_set assumes that the array of pins in struct imx1_pinctrl_soc_info can be indexed by pin id to get the pinctrl_pin_desc for a pin. This used to be correct up to commit 607af165c047 ("pinctrl: i.MX27: Remove nonexistent pad definitions") which removed some entries from the array and so made it wrong to access the array by pin id.
Implement the easiest fix by not resolving the pin id to a name but printing the id instead.
Fixes 607af165c047 ("pinctrl: i.MX27: Remove nonexistent pad definitions")
Cc stable at vger.kernel.org
Reported-by Chris Ruehl <[email protected]>
Signed-off-by Markus Pargmann <[email protected]>
Having the pad name in the output is nice, is it worth to search for the right pinctrl_pin_desc in the array? The array is still sorted, so a binary search would do, maybe a function for this already exists?
How did Chris notice the error? Just a bogus output, or did it crash the kernel? That would be worth to note in the commit log, too.
Otherwise the change looks fine.
Quoted 7 lines
Quoted 2 lines
Best regards
Uwe
I had a crash on a array overrun.
I'd applied a patch to the linux-usb (which was spit out by the maintainer perl script)
Yes, as there is already a patch from Chris my patch can be dropped. We should continue the discussion on your patch instead.
Quoted 72 lines
chris at wheezyvm:~/kernel.d/linux-next$ cat 0001-Pinctrl-imx1-fix-wrong-pin-name-resolving.patch
From 50d56e5f626b2ea86211818cf487514c96f60487 Mon Sep 17 00:00:00 2001
From: Chris Ruehl <[email protected]>
Date: Sat, 23 May 2015 15:02:44 +0800
Subject: [PATCH] Pinctrl: imx1: fix wrong pin-name resolving
Bug in function imx1_pinconf_set() cause crash when princtrl debug is enabled and the pin_id becomes larger then the info->pins[] contains.
imx27-pinctrl 10015000.iomuxc: request pin 134 (MX27_PAD_UART2_TXD) for 1000b000.serial
imx27-pinctrl 10015000.iomuxc: request pin 135 (MX27_PAD_UART2_RXD) for 1000b000.serial
imx27-pinctrl 10015000.iomuxc: request pin 131 (MX27_PAD_UART2_CTS) for 1000b000.serial
imx27-pinctrl 10015000.iomuxc: request pin 132 (MX27_PAD_UART2_RTS) for 1000b000.serial
imx27-pinctrl 10015000.iomuxc: enable function uart group uart2-1 imx27-pinctrl 10015000.iomuxc: imx1_pmx_set, pin 0x86, function 0, gpio 0, direction 1, oconf 0, iconfa 0, iconfb 0
imx27-pinctrl 10015000.iomuxc: imx1_pmx_set, pin 0x87, function 0, gpio 0, direction 0, oconf 0, iconfa 0, iconfb 0
imx27-pinctrl 10015000.iomuxc: imx1_pmx_set, pin 0x83, function 0, gpio 0, direction 1, oconf 0, iconfa 0, iconfb 0
imx27-pinctrl 10015000.iomuxc: imx1_pmx_set, pin 0x84, function 0, gpio 0, direction 0, oconf 0, iconfa 0, iconfb 0
imx27-pinctrl 10015000.iomuxc: num_configs=1 PinID=134 imx27-pinctrl 10015000.iomuxc: pinconf set pullup pin MX27_PAD_UART1_RTS imx27-pinctrl 10015000.iomuxc: num_configs=1 PinID=135 imx27-pinctrl 10015000.iomuxc: pinconf set pullup pin MX27_PAD_RTCK imx27-pinctrl 10015000.iomuxc: num_configs=1 PinID=131 imx27-pinctrl 10015000.iomuxc: pinconf set pullup pin MX27_PAD_UART1_TXD imx27-pinctrl 10015000.iomuxc: num_configs=1 PinID=132 imx27-pinctrl 10015000.iomuxc: pinconf set pullup pin MX27_PAD_UART1_RXD
...
imx27-pinctrl 10015000.iomuxc: write: register 0xf4415508 offset 4 value 0x3 imx27-pinctrl 10015000.iomuxc: write: register 0xf4415510 offset 4 value 0x0 imx27-pinctrl 10015000.iomuxc: write: register 0xf4415518 offset 4 value 0x0 imx27-pinctrl 10015000.iomuxc: imx1_pmx_set, pin 0xb5, function 0, gpio 0, direction 1, oconf 0, iconfa 0, iconfb 0
imx27-pinctrl 10015000.iomuxc: num_configs=1 PinID=171
Unable to handle kernel paging request at virtual address 6c61765f
pgd = c0004000
6c61765f] *pgd=00000000
Internal error: Oops: 5 [#1] ARM
Modules linked in:
CPU: 0 PID: 1 Comm: swapper Not tainted 4.1.0-rc4-next-20150522-dirty #8
Hardware name: GTSYS i.MX27GTSIR (Device Tree Support)
task: ce832000 ti: ce848000 task.ti: ce848000
PC is at strnlen+0x28/0x3c
LR is at string.isra.4+0x34/0xcc
pc : [<c01ae188>]    lr : [<c01af9a4>]    psr: 20000093
sp : ce849a88 ip : ce849a98 fp : ce849a94
r10: c05d7e3a r9 : c05d81e4 r8 : 00000000
r7 : 6c61765f r6 : c05d81e4 r5 : ffffffff r4 : c05d7e3a r3 : 6c61765f r2 : 6c61765f r1 : 6c61765e r0 : 6c61765f
Flags: nzCv IRQs off FIQs on Mode SVC_32 ISA ARM Segment kernel
Signed-off-by Chris Ruehl <[email protected]>
---
  drivers/pinctrl/freescale/pinctrl-imx1-core.c |    6 ++++--
  1 file changed, 4 insertions(+), 2 deletions(-)
drivers/pinctrl/freescale/pinctrl-imx1-core.c +0 -0
index 5ac59fb..8408bd8 100644
drivers/pinctrl/freescale/pinctrl-imx1-core.c static int imx1_pinconf_set(struct pinctrl_dev *pctldev, · line 409
Quoted 8 lines
drivers/pinctrl/freescale/pinctrl-imx1-core.c +2 -1
@@ -403,14 +403,16 @@ static int imx1_pinconf_set(struct pinctrl_dev *pctldev,
  			     unsigned num_configs)
  {
  	struct imx1_pinctrl *ipctl = pinctrl_dev_get_drvdata(pctldev);
-	const struct imx1_pinctrl_soc_info *info = ipctl->info;
+	struct pin_desc *desc;
  	int i;
 
+	desc = pin_desc_get(pctldev, pin_id);
For the rest of this function pin_desc_get() is not necessary. This is not a simple function call but a radix tree lookup as far as I can see. I am not sure how important this debugging information is and how expensive this function call is?!

Quoted 4 lines

Best Regards,
Markus

Hi Markus,

I found it helpful to debug my boards configuration, and have the pin debug on, then I see the output only while the kernel initiation phase.

If you look around, the pin_desc_get() function is used quite often in other parts of the pinctrl section.
drivers/pinctrl/pinctrl-falcon.c
drivers/pinctrl/pinconf.c
drivers/pinctrl/pinmux.c
drivers/pinctrl/core.c
(me) drivers/pinctrl/freescale/pinctrl-imx1-core.c drivers/pinctrl/pinctrl-single.c

Regards
Chris

Signature · 6 lines

GTSYS Limited RFID Technology
9/F, Unit E, R07, Kwai Shing Industrial Building Phase 2,
42-46 Tai Lin Pai Road, Kwai Chung, N.T., Hong Kong
Tel (852) 9079 9521

Disclaimer: http://www.gtsys.com.hk/email/classified.html