Re: [PATCH 07/14 v3] bq27x00: Cache battery registers

[Grazvydas Ignotas <notasas@gmail.com>]

[-- Attachment #1: Type: text/plain, Size: 2104 bytes --]

On Tue, Feb 15, 2011 at 1:48 PM, Grazvydas Ignotas <notasas@gmail.com> wrote:
> On Tue, Feb 15, 2011 at 12:14 AM, Lars-Peter Clausen <lars@metafoo.de> wrote:
>> On 02/14/2011 10:58 PM, Grazvydas Ignotas wrote:
>>> This is wrong, as read function returns negative values for bq27500
>>> when discharging. That's why read function used to pass value through
>>> argument before your series (return value was for error code).
>>
>> I don't think so. The register is 16bit wide and it is read as a unsigned. So
>> in the non error case bq27x00_read will always return >= 0.
>> The value is later reinterpreted as a signed 16bit.(See the other lines you
>> quoted underneath).
>
> Hmh, right..
>
>> Did you experience any actual problem with current being wrong?
>
> Yes, the returned current values were randomly jumping between -500000
> and 600000 while the device was discharging, so I thought
> uninitialized values were being returned (this never happened before
> the series; no errors in dmesg). I'll need to debug a bit more I
> guess..

ok this series seem to trigger unrelated bug, probably due to lots of
reads when cache is being updated, the attached patch seems to help.
Feel free to integrate it or I'll send it separately after your
patches are merged.

However there is bigger problem, compiling as module and doing
insmod/rmmod/insmod causes kernel OOPS:

[  882.575714] BUG: sleeping function called from invalid context at
arch/arm/mm/fault.c:295
[  882.584350] in_atomic(): 0, irqs_disabled(): 128, pid: 1154, name: insmod
...
[  882.959930] Unable to handle kernel NULL pointer dereference at
virtual address 00000000
[  882.968444] pgd = c14c8000
[  882.977905] Internal error: Oops: 817 [#1]
...
[  883.304412] [<c007eed8>] (__queue_work+0x140/0x260) from
[<c007f044>] (queue_work_on+0x2c/0x34)
[  883.313568] [<c007f044>] (queue_work_on+0x2c/0x34) from
[<bf008390>] (bq27x00_update+0x1f8/0x220 [bq27x00_battery])
[  883.324584] [<bf008390>] (bq27x00_update+0x1f8/0x220
[bq27x00_battery]) from [<bf0084c8>] (bq27x00_battery_poll+0x14/0x48
[bq27x00_battery])

full backtrace attached.

[-- Attachment #2: backtrace --]
[-- Type: application/octet-stream, Size: 18766 bytes --]

[-- Attachment #3: 0001-bq27x00-Use-single-i2c_transfer-call-for-property-re.patch --]
[-- Type: text/x-diff, Size: 1651 bytes --]

From 8fc5bdaf033d834d1728cd70c240d31bc457336c Mon Sep 17 00:00:00 2001
From: Grazvydas Ignotas <notasas@gmail.com>
Date: Tue, 15 Feb 2011 23:27:35 +0200
Subject: [PATCH] bq27x00: Use single i2c_transfer call for property read

Doing this by using 2 calls sometimes results in unexpected
values being returned on OMAP3 i2c controller.

Signed-off-by: Grazvydas Ignotas <notasas@gmail.com>
---
 drivers/power/bq27x00_battery.c |   27 +++++++++++----------------
 1 files changed, 11 insertions(+), 16 deletions(-)

diff --git a/drivers/power/bq27x00_battery.c b/drivers/power/bq27x00_battery.c
index 740a0ac..59e68db 100644
--- a/drivers/power/bq27x00_battery.c
+++ b/drivers/power/bq27x00_battery.c
@@ -565,31 +565,26 @@ static DEFINE_MUTEX(battery_mutex);
 static int bq27x00_read_i2c(struct bq27x00_device_info *di, u8 reg, bool single)
 {
 	struct i2c_client *client = to_i2c_client(di->dev);
-	struct i2c_msg msg;
+	struct i2c_msg msg[2];
 	unsigned char data[2];
 	int ret;
 
 	if (!client->adapter)
 		return -ENODEV;
 
-	msg.addr = client->addr;
-	msg.flags = 0;
-	msg.len = 1;
-	msg.buf = data;
-
-	data[0] = reg;
-	ret = i2c_transfer(client->adapter, &msg, 1);
-
-	if (ret < 0)
-		return ret;
-
+	msg[0].addr = client->addr;
+	msg[0].flags = 0;
+	msg[0].buf = &reg;
+	msg[0].len = sizeof(reg);
+	msg[1].addr = client->addr;
+	msg[1].flags = I2C_M_RD;
+	msg[1].buf = data;
 	if (single)
-		msg.len = 1;
+		msg[1].len = 1;
 	else
-		msg.len = 2;
+		msg[1].len = 2;
 
-	msg.flags = I2C_M_RD;
-	ret = i2c_transfer(client->adapter, &msg, 1);
+	ret = i2c_transfer(client->adapter, msg, ARRAY_SIZE(msg));
 	if (ret < 0)
 		return ret;
 
-- 
1.6.3.3