[U-Boot-Users] rtc/ds1374.c turns clock OFF when it means to turn it ON
rtc/ds1374.c mistakes the "don't enable the oscillator" bit for a "do enable the oscillator bit" and sets it to 1, stopping the clock. My guess is that nobody has actually used rtc_set or rtc_reset with this driver.
We've been using this chip for a few years with a proprietary OS and are now switching to u-boot and Linux. I revised the u-boot driver based on our stable driver. Besides fixing the oscillator bit, I set it up to read all the registers with a single IIC transaction for efficiency and to see any error code from the IIC subsystem.
I would attach the new driver here but I can't figure out how. I'm typing through a broken "hello new user, we don't yet trust you" web page. If somebody could send instructions to andrew@cesa.opbu.xerox.com, I'd like to submit this revision.
Andrew Klossner
Andrew,
On Friday 23 February 2007 22:57, Andrew Klossner wrote:
rtc/ds1374.c mistakes the "don't enable the oscillator" bit for a "do enable the oscillator bit" and sets it to 1, stopping the clock. My guess is that nobody has actually used rtc_set or rtc_reset with this driver.
A quick look shows that this driver is used by the Freescale MPC8349EMDS & MPC8360EMDS board ports. Some of the Freescale guys (Kim?) should jump in here.
We've been using this chip for a few years with a proprietary OS and are now switching to u-boot and Linux. I revised the u-boot driver based on our stable driver. Besides fixing the oscillator bit, I set it up to read all the registers with a single IIC transaction for efficiency and to see any error code from the IIC subsystem.
I would attach the new driver here but I can't figure out how. I'm typing through a broken "hello new user, we don't yet trust you" web page. If somebody could send instructions to andrew@cesa.opbu.xerox.com, I'd like to submit this revision.
For submitting patches please take a look at the section "Submitting Patches:" at the end of the README file. Nowadays patches created using git are even better (like in the Linux kernel development).
Best regards, Stefan
===================================================================== DENX Software Engineering GmbH, HRB 165235 Munich, CEO: Wolfgang Denk Office: Kirchenstr. 5, D-82194 Groebenzell, Germany =====================================================================
On Fri, Feb 23, 2007 at 09:57:12PM +0000, Andrew Klossner wrote:
rtc/ds1374.c mistakes the "don't enable the oscillator" bit for a "do enable the oscillator bit" and sets it to 1, stopping the clock. My guess is that nobody has actually used rtc_set or rtc_reset with this driver.
It doesn't set it, it clears it -- rtc_write is called with "set" set to FALSE.
It's certainly not the clearest way to do it; there should probably be separate functions rtc_set_bits and rtc_clear_bits (the "set" parameter is always passed as a constant, so the runtime check isn't actually being used).
-Scott
participants (3)
-
Andrew Klossner -
Scott Wood -
Stefan Roese