Re: [i2c] [crash, bisected] I2C, kobject (ffff81003e8c4160): tried to init an initialized object
* Ingo Molnar [EMAIL PROTECTED] wrote:
Jean,
-tip testing found another (similar) i2c crash with latest -git, on a
Core2Duo laptop:
calling nforce2_init+0x0/0x1b
initcall nforce2_init+0x0/0x1b returned 0 after 0 msecs
calling nforce2_s4985_init+0x0/0x275
BUG: unable to handle kernel NULL pointer dereference at 000c
IP: [c04af2cf] i2c_smbus_xfer+0x27/0x3f0
*pdpt = 009cb001 *pde =
Oops: [#1] SMP
turning off I2C_NFORCE2_S4985 makes the system boot up fine. Workaround
patch below.
Ingo
---
commit 15e524060d13e8fac3fe2fb6c6519be6de9ab8a8
Author: Ingo Molnar [EMAIL PROTECTED]
Date: Thu Jul 17 11:29:44 2008 +0200
i2c workaround: turn off I2C_NFORCE2_S4985 for now
I2C_NFORCE2_S4985 causes this crash:
calling nforce2_init+0x0/0x1b
initcall nforce2_init+0x0/0x1b returned 0 after 0 msecs
calling nforce2_s4985_init+0x0/0x275
BUG: unable to handle kernel NULL pointer dereference at 000c
IP: [c04af2cf] i2c_smbus_xfer+0x27/0x3f0
*pdpt = 009cb001 *pde =
Oops: [#1] SMP
Pid: 1, comm: swapper Not tainted (2.6.26-tip #3033)
EIP: 0060:[c04af2cf] EFLAGS: 00010246 CPU: 0
EIP is at i2c_smbus_xfer+0x27/0x3f0
EAX: 0003 EBX: ECX: EDX: 0018
ESI: 0003 EDI: 0018 EBP: f7c6ff24 ESP: f7c6fea8
DS: 007b ES: 007b FS: 00d8 GS: SS: 0068
Process swapper (pid: 1, ti=f7c6e000 task=f7c7 task.ti=f7c6e000)
Stack: 0035 0003 c0104237 0046 0046
f7c6ff50 c09debf0 0001007b fffe007b 00d8 ffef
c01213cb
0060 0286 f7c6ff1c 0004 0286 0246 f7c6ff24
c014673a
Call Trace:
[c0104237] ? common_interrupt+0x23/0x28
[c01213cb] ? vprintk+0x290/0x29b
[c014673a] ? ftrace_record_ip+0x19f/0x1e6
[c01a1385] ? sysfs_add_file_mode+0x57/0x76
[c099632f] ? nforce2_s4985_init+0x2d/0x275
[c013319c] ? ktime_get_ts+0x51/0x55
[c01331b8] ? ktime_get+0x18/0x34
[c0971887] ? kernel_init+0x11f/0x24a
[c0996302] ? nforce2_s4985_init+0x0/0x275
[c011cbe8] ? schedule_tail+0x1e/0x4d
[c01037aa] ? ret_from_fork+0x6/0x1c
[c0971768] ? kernel_init+0x0/0x24a
[c0971768] ? kernel_init+0x0/0x24a
[c0104477] ? kernel_thread_helper+0x7/0x10
===
Code: 5e 5f 5d c3 55 89 e5 57 56 53 83 ec 70 e8 ca 51 c5 ff 89 45 8c 8b 75
0c 83 e1 14 8a 45 08 89 d7 8b 5d 8c 88 45 8b 89 f0 88 45 8a 8b 43 0c 83 78 04
00 0f b7 c1 89 45 84 74 3a 83 c3 20 89 d8 e8
EIP: [c04af2cf] i2c_smbus_xfer+0x27/0x3f0 SS:ESP 0068:f7c6fea8
Kernel panic - not syncing: Fatal exception
Pid: 1, comm: swapper Tainted: G D 2.6.26-tip #3033
[c01208ee] panic+0x44/0xeb
[c0104c8d] die+0x123/0x13a
[c0110f38] do_page_fault+0x632/0x6f2
[c0110906] ? do_page_fault+0x0/0x6f2
[c06585ca] error_code+0x72/0x78
[c04af2cf] ? i2c_smbus_xfer+0x27/0x3f0
[c0104237] ? common_interrupt+0x23/0x28
[c01213cb] ? vprintk+0x290/0x29b
[c014673a] ? ftrace_record_ip+0x19f/0x1e6
[c01a1385] ? sysfs_add_file_mode+0x57/0x76
[c099632f] nforce2_s4985_init+0x2d/0x275
[c013319c] ? ktime_get_ts+0x51/0x55
[c01331b8] ? ktime_get+0x18/0x34
[c0971887] kernel_init+0x11f/0x24a
[c0996302] ? nforce2_s4985_init+0x0/0x275
[c011cbe8] ? schedule_tail+0x1e/0x4d
[c01037aa] ? ret_from_fork+0x6/0x1c
[c0971768] ? kernel_init+0x0/0x24a
[c0971768] ? kernel_init+0x0/0x24a
[c0104477] kernel_thread_helper+0x7/0x10
===
with this config:
http://redhat.com/~mingo/misc/config-Thu_Jul_17_10_47_42_CEST_2008.bad
Signed-off-by: Ingo Molnar [EMAIL PROTECTED]
---
drivers/i2c/busses/Kconfig |3 +++
1 files changed, 3 insertions(+), 0 deletions(-)
diff --git a/drivers/i2c/busses/Kconfig b/drivers/i2c/busses/Kconfig
index 6ee997b..2213fab 100644
--- a/drivers/i2c/busses/Kconfig
+++ b/drivers/i2c/busses/Kconfig
@@ -149,6 +149,9 @@ config I2C_NFORCE2
config I2C_NFORCE2_S4985
tristate SMBus multiplexing on the Tyan S4985
depends on I2C_NFORCE2 EXPERIMENTAL
+ # broke a T60 Core2Duo with:
+ # http://redhat.com/~mingo/misc/config-Thu_Jul_17_10_47_42_CEST_2008.bad
+ depends on 0
help
Enabling this option will add specific SMBus support for the Tyan
S4985 motherboard. On this 4-CPU board, the SMBus is multiplexed
___
i2c mailing list
i2c@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/i2c
Re: [i2c] [crash, bisected] I2C, kobject (ffff81003e8c4160): tried to init an initialized object
* Ingo Molnar [EMAIL PROTECTED] wrote: -tip testing found another (similar) i2c crash with latest -git, on a Core2Duo laptop: found another bootup crash on an Athlon 64 X2 testsystem: [ 19.61] bus: 'pci': add driver amd756_smbus [ 19.615204] initcall amd756_init+0x0/0x20 returned 0 after 9 msecs [ 19.62] calling amd756_s4882_init+0x0/0x2d7 [ 19.624242] [ cut here ] [ 19.628415] WARNING: at kernel/mutex.c:134 mutex_lock_nested+0xc0/0x267() [ 19.63] Pid: 1, comm: swapper Not tainted 2.6.26-tip #20066 [ 19.63] [ 19.63] Call Trace: [ 19.63] [802358f5] warn_on_slowpath+0x58/0x7f [ 19.63] [80263ac6] ? ftrace_record_ip+0x20c/0x263 [ 19.63] [80514faa] ? _raw_spin_unlock+0x8e/0x93 [ 19.63] [80507c44] ? debug_locks_off+0x4/0x3c [ 19.63] [80263ac6] ? ftrace_record_ip+0x20c/0x263 [ 19.63] [8020b3a4] ? mcount_call+0x5/0x31 [ 19.63] [80861c19] mutex_lock_nested+0xc0/0x267 [ 19.63] [806e8ffd] ? i2c_smbus_xfer+0x59/0x49c [ 19.63] [80cc8575] ? amd756_s4882_init+0x0/0x2d7 [ 19.63] [8020b3a4] ? mcount_call+0x5/0x31 [ 19.63] [806e8ffd] i2c_smbus_xfer+0x59/0x49c [ 19.63] [80263ac6] ? ftrace_record_ip+0x20c/0x263 [ 19.63] [80514faa] ? _raw_spin_unlock+0x8e/0x93 [ 19.63] [80cc8584] ? amd756_s4882_init+0xf/0x2d7 [ 19.63] [80263ac6] ? ftrace_record_ip+0x20c/0x263 [ 19.63] [80cc8575] ? amd756_s4882_init+0x0/0x2d7 [ 19.63] [8020b3a4] ? mcount_call+0x5/0x31 [ 19.63] [80cc8575] ? amd756_s4882_init+0x0/0x2d7 [ 19.63] [80cc85b6] amd756_s4882_init+0x41/0x2d7 [ 19.63] [8024e325] ? getnstimeofday+0x3f/0xa1 [ 19.63] [8024bf25] ? ktime_get_ts+0x4e/0x53 [ 19.63] [80cc8575] ? amd756_s4882_init+0x0/0x2d7 [ 19.63] [80ca49a0] kernel_init+0x143/0x2a4 [ 19.63] [80863721] ? _spin_unlock_irq+0x2b/0x37 [ 19.63] [8020c9f9] child_rip+0xa/0x11 [ 19.63] [80ca485d] ? kernel_init+0x0/0x2a4 [ 19.63] [8020c9ef] ? child_rip+0x0/0x11 [ 19.63] [ 19.63] ---[ end trace 0e79313125261797 ]--- [ 19.63] BUG: unable to handle kernel NULL pointer dereference at [ 19.63] IP: [80861c66] mutex_lock_nested+0x10d/0x267 with this config: http://redhat.com/~mingo/misc/config-Thu_Jul_17_11_34_08_CEST_2008.bad i already had I2C_NFORCE2_S4985 off here, so it's a different one. Ingo ___ i2c mailing list i2c@lm-sensors.org http://lists.lm-sensors.org/mailman/listinfo/i2c
Re: [i2c] [crash, bisected] I2C, kobject (ffff81003e8c4160): tried to init an initialized object
* Alexey Dobriyan [EMAIL PROTECTED] wrote: On Thu, Jul 17, 2008 at 11:33:30AM +0200, Ingo Molnar wrote: * Ingo Molnar [EMAIL PROTECTED] wrote: Jean, -tip testing found another (similar) i2c crash with latest -git, on a Core2Duo laptop: calling nforce2_init+0x0/0x1b initcall nforce2_init+0x0/0x1b returned 0 after 0 msecs calling nforce2_s4985_init+0x0/0x275 BUG: unable to handle kernel NULL pointer dereference at 000c IP: [c04af2cf] i2c_smbus_xfer+0x27/0x3f0 *pdpt = 009cb001 *pde = Oops: [#1] SMP turning off I2C_NFORCE2_S4985 makes the system boot up fine. Workaround patch below. Well, it's pretty easy oops unless I misread something. It happened on not nforce2 motherboard, right? yeah. nforce2_init will just register PCI driver, no matching PCI ids = nforce2_set_reference() isn't called = nforce2_smbus stays NULL. Second module loads and tries to i2c_smbus_xfer(NULL, ) in module_init() hook which oopses. can try patches, but got other 5 fresh upstream oopses/hangs/crashes to investigate now. (so i went for the minimal-effort turn-off-the-failing-part approach) Ingo ___ i2c mailing list i2c@lm-sensors.org http://lists.lm-sensors.org/mailman/listinfo/i2c
Re: [i2c] Access 16bits address register in a device through I2C
On Fri, Jul 18, 2008 at 12:13:30AM +0800, Wen Wang wrote: Hi all, I have a problem writing an I2C device driver and would greatly appreciate if someone could provide some enlightenment. I'm working on a project needs to access registers with 16bits address through I2C. But unfortunately I can not find a way to achieve this, cause it seems that current kernel I2C framework only supports 8bits address register. The easiest way is to create a function that uses a pair of i2c messages, the first message sends the 16bits of address pointer to the device, the second one either reads or writes data. So I'm wondering do I miss something or what can I do? Any ideas? Thanks Wen ___ i2c mailing list i2c@lm-sensors.org http://lists.lm-sensors.org/mailman/listinfo/i2c -- Ben ([EMAIL PROTECTED], http://www.fluff.org/) 'a smiley only costs 4 bytes' ___ i2c mailing list i2c@lm-sensors.org http://lists.lm-sensors.org/mailman/listinfo/i2c
Re: [i2c] Access 16bits address register in a device through I2C
Hi Wen, On Fri, Jul 18, 2008 at 12:13:30AM +0800, Wen Wang wrote: So I'm wondering do I miss something or what can I do? Any ideas? For an example what Ben suggested (2 messages), have a look at the at24-driver, which just went through the merge-window. http://article.gmane.org/gmane.linux.drivers.i2c/1661 Check functions 'at24_eeprom_[read|write]' for details. All the best, Wolfram -- Dipl.-Ing. Wolfram Sang | http://www.pengutronix.de Pengutronix - Linux Solutions for Science and Industry signature.asc Description: Digital signature ___ i2c mailing list i2c@lm-sensors.org http://lists.lm-sensors.org/mailman/listinfo/i2c
Re: [i2c] [crash, bisected] I2C, kobject (ffff81003e8c4160): tried to init an initialized object (was: Re: [GIT PULL] i2c updates for 2.6.27, round 1)
On Wed, Jul 16, 2008 at 11:07:02AM +0200, Jean Delvare wrote: btw., Greg, the kobject debugging check and output is really nice, informative and useful! It likely pinpoints the real bug here. Is this something new in 2.6.26-gitX? I want to know if the bug is something new or if it could be already present in 2.6.26 and earlier kernels and went unnoticed so far. It is in 2.6.26 as I haven't merged any patches with Linus yet, portions of it might be in older kernels, but not in the same type of format. thanks, greg k-h ___ i2c mailing list i2c@lm-sensors.org http://lists.lm-sensors.org/mailman/listinfo/i2c
Re: [i2c] [PATCH] hwmon: (adm9240) Convert to a new-style i2c driver
On Wed, 16 Jul 2008 16:32:57 +0200, Jean Delvare [EMAIL PROTECTED] wrote:
The new-style adm9240 driver implements the optional detect() callback
to cover the use cases of the legacy driver.
Signed-off-by: Jean Delvare [EMAIL PROTECTED]
Cc: Grant Coady [EMAIL PROTECTED]
Tested both as compiled in, and as modules, with 2.6.26-git6, no problems
found, userspace:
~$ sensors --version
sensors version 3.0.2 with libsensors version 3.0.2
sensors-detect, sensors and 'sensors -s' behaved as expected.
Acked-by: Grant Coady [EMAIL PROTECTED]
---
drivers/hwmon/adm9240.c | 93 +++
1 file changed, 46 insertions(+), 47 deletions(-)
--- linux-2.6.26-rc9.orig/drivers/hwmon/adm9240.c 2008-07-12
09:20:31.0 +0200
+++ linux-2.6.26-rc9/drivers/hwmon/adm9240.c 2008-07-12 10:27:31.0
+0200
@@ -130,25 +130,37 @@ static inline unsigned int AOUT_FROM_REG
return SCALE(reg, 1250, 255);
}
-static int adm9240_attach_adapter(struct i2c_adapter *adapter);
-static int adm9240_detect(struct i2c_adapter *adapter, int address, int kind);
+static int adm9240_probe(struct i2c_client *client,
+ const struct i2c_device_id *id);
+static int adm9240_detect(struct i2c_client *client, int kind,
+struct i2c_board_info *info);
static void adm9240_init_client(struct i2c_client *client);
-static int adm9240_detach_client(struct i2c_client *client);
+static int adm9240_remove(struct i2c_client *client);
static struct adm9240_data *adm9240_update_device(struct device *dev);
/* driver data */
+static const struct i2c_device_id adm9240_id[] = {
+ { adm9240, adm9240 },
+ { ds1780, ds1780 },
+ { lm81, lm81 },
+ { }
+};
+MODULE_DEVICE_TABLE(i2c, adm9240_id);
+
static struct i2c_driver adm9240_driver = {
+ .class = I2C_CLASS_HWMON,
.driver = {
.name = adm9240,
},
- .attach_adapter = adm9240_attach_adapter,
- .detach_client = adm9240_detach_client,
+ .probe = adm9240_probe,
+ .remove = adm9240_remove,
+ .id_table = adm9240_id,
+ .detect = adm9240_detect,
+ .address_data = addr_data,
};
/* per client data */
struct adm9240_data {
- enum chips type;
- struct i2c_client client;
struct device *hwmon_dev;
struct mutex update_lock;
char valid;
@@ -532,28 +544,17 @@ static const struct attribute_group adm9
/*** sensor chip detect and driver install ***/
-static int adm9240_detect(struct i2c_adapter *adapter, int address, int kind)
+/* Return 0 if detection is successful, -ENODEV otherwise */
+static int adm9240_detect(struct i2c_client *new_client, int kind,
+struct i2c_board_info *info)
{
- struct i2c_client *new_client;
- struct adm9240_data *data;
- int err = 0;
+ struct i2c_adapter *adapter = new_client-adapter;
const char *name = ;
+ int address = new_client-addr;
u8 man_id, die_rev;
if (!i2c_check_functionality(adapter, I2C_FUNC_SMBUS_BYTE_DATA))
- goto exit;
-
- if (!(data = kzalloc(sizeof(*data), GFP_KERNEL))) {
- err = -ENOMEM;
- goto exit;
- }
-
- new_client = data-client;
- i2c_set_clientdata(new_client, data);
- new_client-addr = address;
- new_client-adapter = adapter;
- new_client-driver = adm9240_driver;
- new_client-flags = 0;
+ return -ENODEV;
if (kind == 0) {
kind = adm9240;
@@ -566,7 +567,7 @@ static int adm9240_detect(struct i2c_ada
!= address) {
dev_err(adapter-dev, detect fail: address match,
0x%02x\n, address);
- goto exit_free;
+ return -ENODEV;
}
/* check known chip manufacturer */
@@ -581,7 +582,7 @@ static int adm9240_detect(struct i2c_ada
} else {
dev_err(adapter-dev, detect fail: unknown manuf,
0x%02x\n, man_id);
- goto exit_free;
+ return -ENODEV;
}
/* successful detect, print chip info */
@@ -600,20 +601,31 @@ static int adm9240_detect(struct i2c_ada
} else if (kind == lm81) {
name = lm81;
}
+ strlcpy(info-type, name, I2C_NAME_SIZE);
- /* fill in the remaining client fields and attach */
- strlcpy(new_client-name, name, I2C_NAME_SIZE);
- data-type = kind;
- mutex_init(data-update_lock);
+ return 0;
+}
- if ((err = i2c_attach_client(new_client)))
- goto exit_free;
+static int adm9240_probe(struct i2c_client *new_client,
+ const struct i2c_device_id *id)
+{
+ struct adm9240_data *data;
+ int err;
+
+ data =
Re: [i2c] Access 16bits address register in a device through I2C
Hi Guys, Thank you for your kindly heads up, that really helps me a lot :) BRs Wen On Fri, Jul 18, 2008 at 1:06 AM, Wolfram Sang [EMAIL PROTECTED] wrote: Hi Wen, On Fri, Jul 18, 2008 at 12:13:30AM +0800, Wen Wang wrote: So I'm wondering do I miss something or what can I do? Any ideas? For an example what Ben suggested (2 messages), have a look at the at24-driver, which just went through the merge-window. http://article.gmane.org/gmane.linux.drivers.i2c/1661 Check functions 'at24_eeprom_[read|write]' for details. All the best, Wolfram -- Dipl.-Ing. Wolfram Sang | http://www.pengutronix.de Pengutronix - Linux Solutions for Science and Industry -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.6 (GNU/Linux) iD8DBQFIf3wpD27XaX1/VRsRAjXBAKCx28nzPbkyMx3aN6KNN03jNFIRLwCgiyAL +zHcrnvXfuzWngEpflY5KZg= =sUwl -END PGP SIGNATURE- ___ i2c mailing list i2c@lm-sensors.org http://lists.lm-sensors.org/mailman/listinfo/i2c
