Skip to contentSkip to conversation

X86: Impossible select Enhanced Real Time Clock Support (legacy PC RTC driver)

DISCUSSION Stefani Seibold 5 participants 7 messages Apr 13, 2014 – May 6, 2014

lore
Share & tools
lore
#0 lore raw

Since some kernel version it is impossible to select the Enhanced Real Time Clock Support (legacy PC RTC driver) because RTC_LIB is set by default in arch/x86/Kconfig, but the rule for selecting CONFIG_RTC is RTC_LIB=n. So the code of driver/char/rtc is still useless.

This breaks the API since there is no more misc device /dev/rtc available without a udev rule or a link to /dev/rtc0.

#1 lore raw

[adding more patch committers]

Quoted 7 lines · Stefani Seibold

Since some kernel version it is impossible to select the Enhanced Real Time Clock Support (legacy PC RTC driver) because RTC_LIB is set by default in arch/x86/Kconfig, but the rule for selecting CONFIG_RTC is RTC_LIB=n. So the code of driver/char/rtc is still useless.

This breaks the API since there is no more misc device /dev/rtc available without a udev rule or a link to /dev/rtc0.

Signature · 1 line
~Randy
#2 lore raw
Quoted 10 lines · Randy Dunlap

[adding more patch committers]

Quoted 7 lines · Stefani Seibold

Since some kernel version it is impossible to select the Enhanced Real Time Clock Support (legacy PC RTC driver) because RTC_LIB is set by default in arch/x86/Kconfig, but the rule for selecting CONFIG_RTC is RTC_LIB=n. So the code of driver/char/rtc is still useless.

This breaks the API since there is no more misc device /dev/rtc available without a udev rule or a link to /dev/rtc0.

So yea.. I feel like that /dev/rtcN renaming (and API) break (which was a huge and annoying pain) was back in the 2.6.18-ish era?

But it sounds like the driver/char/rtc bit is dead code, and needs a cleanup? Or is there some use of that code that you need that the generic RTC layer doesn't have?

thanks
-john

#3 lore raw
Quoted 5 lines · On Mon, 28 Apr 2014 11:16:37 -0700 John Stultz <[email protected]> wrote:
Quoted 2 lines
Quoted 2 lines

This breaks the API since there is no more misc device /dev/rtc available without a udev rule or a link to /dev/rtc0.

So yea.. I feel like that /dev/rtcN renaming (and API) break (which was a huge and annoying pain) was back in the 2.6.18-ish era?

It's very old and linking/renaming is around since then.

Quoted 3 lines

But it sounds like the driver/char/rtc bit is dead code, and needs a cleanup? Or is there some use of that code that you need that the generic RTC layer doesn't have?

as far as I know, no recent distribution is using the old code
anymore and the new (well, not so new anymore) framework provides the same functionality.

Signature · 6 lines
 Best regards,

 Alessandro Zummo,
  Tower Technologies - Torino, Italy

  http://www.towertech.it
#4 lore raw
Quoted 15 lines · Alessandro Zummo
Quoted 5 lines · On Mon, 28 Apr 2014 11:16:37 -0700 John Stultz <[email protected]> wrote:
Quoted 2 lines
Quoted 2 lines

This breaks the API since there is no more misc device /dev/rtc available without a udev rule or a link to /dev/rtc0.

So yea.. I feel like that /dev/rtcN renaming (and API) break (which was a huge and annoying pain) was back in the 2.6.18-ish era?

It's very old and linking/renaming is around since then.

Quoted 3 lines

But it sounds like the driver/char/rtc bit is dead code, and needs a cleanup? Or is there some use of that code that you need that the generic RTC layer doesn't have?

The driver/char/rtc is not working code and should be a wrapper to the first rtc device.

Quoted 3 lines

as far as I know, no recent distribution is using the old code
anymore and the new (well, not so new anymore) framework provides the same functionality.

Embedded distribution use it and depending on the /dev/rtc device. So what is missing is the /dev/rtc misc device 10,135.

- Stefani

#5 lore raw
In reply to Alessandro Zummo
Quoted 7 lines · Alessandro Zummo
Quoted 3 lines

But it sounds like the driver/char/rtc bit is dead code, and needs a cleanup? Or is there some use of that code that you need that the generic RTC layer doesn't have?

as far as I know, no recent distribution is using the old code anymore and the new (well, not so new anymore) framework provides the same functionality.

Well, the handling of update-ended interrupts is still missing, as is
some non-x86 stuff I posted a fix for recently (still pending review, can you please look into it sometime?).

Maciej

#6 lore raw
Quoted 3 lines · On Tue, 6 May 2014 00:34:41 +0100 (BST) "Maciej W. Rozycki"…

Well, the handling of update-ended interrupts is still missing, as is
some non-x86 stuff I posted a fix for recently (still pending review, can you please look into it sometime?).

Done, sorry for the delay.

Signature · 6 lines
 Best regards,

 Alessandro Zummo,
  Tower Technologies - Torino, Italy

  http://www.towertech.it