Hi Steve,
the thread is getting longer and longer.
Am 02.05.2010 um 02:10 schrieb Steve Arnold:
On Sat, 1 May 2010 21:18:38 +0200
"Dr. H. Nikolaus Schaller" <h...@computer.org> wrote:
By looking into source file: this means
rtc_class_open(CONFIG_RTC_HCTOSYS_DEVICE) returns NULL.
CONFIG_RTC_HCTOSYS_DEVICE is "rtc0".
The "normal" or standard devices you should get with udev are:
lrwxrwxrwx 1 root root 4 2010-05-01 07:29 /dev/rtc -> rtc0
crw-rw---- 1 root root 254, 0 2010-05-01 07:29 /dev/rtc0
What I see is
crw-rw---- 1 root audio 10, 135 Jan 1 1970 /dev/rtc
and no link. This means that the "old" driver is active.
But this is all consistent with /sys/class/misc/rtc/dev or .../uevent
assuming you have the actual device driver(s) loaded.
So I think there are two possibilities for the rtc0 device not
showing up: 1) the i2c dependency is not being met, or 2) maybe
udev is having trouble creating the device.
$ fgrep RTC .config
# CONFIG_RTC is not set
CONFIG_RTC_PCF8563=y
# CONFIG_RTC_JZ is not set
CONFIG_RTC_LIB=y
CONFIG_RTC_CLASS=y
CONFIG_RTC_HCTOSYS=y
CONFIG_RTC_HCTOSYS_DEVICE="rtc0"
# CONFIG_RTC_DEBUG is not set
# RTC interfaces
CONFIG_RTC_INTF_SYSFS=y
# CONFIG_RTC_INTF_PROC is not set
CONFIG_RTC_INTF_DEV=y
# CONFIG_RTC_INTF_DEV_UIE_EMUL is not set
# CONFIG_RTC_DRV_TEST is not set
# I2C RTC drivers
# CONFIG_RTC_DRV_DS1307 is not set
# CONFIG_RTC_DRV_DS1374 is not set
# CONFIG_RTC_DRV_DS1672 is not set
# CONFIG_RTC_DRV_MAX6900 is not set
# CONFIG_RTC_DRV_RS5C372 is not set
# CONFIG_RTC_DRV_ISL1208 is not set
# CONFIG_RTC_DRV_X1205 is not set
# CONFIG_RTC_DRV_PCF8563 is not set
CONFIG_RTC_DRV_PCF8583=y
# CONFIG_RTC_DRV_M41T80 is not set
# SPI RTC drivers
# Platform RTC drivers
# CONFIG_RTC_DRV_CMOS is not set
# CONFIG_RTC_DRV_DS1553 is not set
# CONFIG_RTC_DRV_STK17TA8 is not set
# CONFIG_RTC_DRV_DS1742 is not set
# CONFIG_RTC_DRV_M48T86 is not set
# CONFIG_RTC_DRV_M48T59 is not set
# CONFIG_RTC_DRV_V3020 is not set
# on-CPU RTC drivers
Hm. Finally I had an idea:
$ find /sys -name '*rtc*'
/sys/class/misc/rtc
/sys/class/rtc
/sys/module/rtc_pcf8583
So it may be as simple as changing
CONFIG_RTC_HCTOSYS_DEVICE="rtc0"
to
CONFIG_RTC_HCTOSYS_DEVICE="rtc"
This was too optimistic... Does not change anything (except that "rtc"
appears in the error message instead of "rtc0").
It looks like you have two Phillips drivers enabled; I would
disable one of them. Have you tried the RTC_JZ driver yet? Or is
it just a stub?
In general, I think the rtc0 device stuff in dmesg really depends on
what gets created under /dev rather than what you see under /sys.
If the right device appears in /sys then you should get something
under /dev (with an appropriate udev rule).
I have compared the udev rules with an i386 lenny - it is the same
/etc/udev/rules.d/50-udev.rules (here it says SUBSYSTEM=="rtc",
DRIVERS=="rtc_cmos", SYMLINK+="rtc")
/etc/udev/rules.d/90-permissions.rules (this is where the root:audio
comes from)
And by old and new rtc drivers I meant the (old) ones under:
Device Drivers --->
Character devices --->
Enhanced Real Time Clock Support (legacy PC RTC driver)
this is disabled
Generic /dev/rtc emulation
...
HPET Timer <= this one works with both
...
does not show up.
But there is an entry
[*] Philips PCF8563 Real Time CLock (I2C Bus)
I will try to disable this as well.
vs. the ones under:
Device Drivers --->
Real Time Clock --->
/sys/class/rtc/rtcN (sysfs)
[*]
/proc/driver/rtc (procfs for rtc0)
[ ]
/dev/rtcN (character devices)
[*]
...
and then, I find: <*> Philips PCF8583
Hm. I think I did read at
http://projects.kwaak.net/twiki/bin/view/Epc700/HardwareLayout
that we have an PCF8563 and not a PCF8583 inside?
The first two options under the (old) Char device section only
appear if the (newer) Real Time Clock section is disabled. There
are lots of device-specific drivers in the latter section
(including the Phillips drivers) but not the JZ driver (at least in
the vanilla kernel). Question: does RTC_JZ appear under the bottom
section "on-CPU RTC drivers"?
No.
It looks to me like your "CONFIG_RTC_PCF8563=y" is an older driver,
while the newer one is dsiabled (along with the RTC_JZ driver).
What happens if you disable the first one and enable the second one?
What happens if you enable "CONFIG_RTC_DRV_PCF8563" instead?
My $.02 would be to make sure the (old) ones are disabled and only
enable stuff under the new Real Time Clock section. Not sure about
HPET in this context; I guess you can try it both ways and see...
Here's what I get on an embedded powerpc board with an older 2.6.15
kernel and udev-079-r1 install:
lrwxrwxrwx 1 root root 8 Apr 18 13:15 /dev/rtc -> misc/rtc
so I guess the version of the kernel and/or udev is old enough that
it doesn't even create the rtc0 device (or a symlink). Do you see
anything under /sys/devices like this:
/sys/devices/pnp0/00:05/rtc
/sys/devices/pnp0/00:05/rtc/rtc0
No. Just usb, serial, cpu, lcd and other stuff but not rtc.
On the above embedded board I only see this:
/sys/class/misc/rtc
After changing the config, I find /sys/class/rtc and /sys/modules/
rtc_pcf8563 but they are empty.
Anyway, hope this helps somehow...
Definitively!
But changing the config did not make it work.
I have now added some printk() to the hcttosys functions and
struct rtc_device *rtc_class_open(char *name)
{
struct device *dev;
struct rtc_device *rtc = NULL;
printk(KERN_EMERG "rtc_class_open(name=%s)\n", name);
down(&rtc_class->sem);
list_for_each_entry(dev, &rtc_class->devices, node) {
printk(KERN_EMERG "dev->bus_id=%s name=%s\n", dev->bus_id,
name);
if (strncmp(dev->bus_id, name, BUS_ID_SIZE) == 0) {
dev = get_device(dev);
if (dev)
rtc = to_rtc_device(dev);
break;
}
}
if (rtc) {
if (!try_module_get(rtc->owner)) {
put_device(dev);
rtc = NULL;
}
}
up(&rtc_class->sem);
return rtc;
}
shows that list_for_each_entry(dev, &rtc_class->devices, node) { }
does not loop. I.e. the rtc_class->devices appears to be empty.
Additional printk() have shown that rtc_init is called, the
class_create() is successful and both rtc_sysfs_init() and
rtc_dev_init() are called. Both are included in the .config. Disabling
the CONFIG_RTC_INTF_DEV had no visible effect (except removing my
printk message).
So, something is wrong with the initialization of these kernel
structures. So the question is where rtc_class->devices should be
initialized/populated.
Finally I have studied http://www.mjmwired.net/kernel/Documentation/rtc.txt
and accodting to that document /sys/class/rtc should not be an empty
directory but show rtc0...
BR,
Nikolaus
_______________________________________________
Mipsbook-devel mailing list
Mipsbook-devel@linuxtogo.org
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/mipsbook-devel
