FS#7135 - CONFIG_TIMER_STATS and CONFIG_USB_SUSPEND in kernel

Attached to Project: Arch Linux
Opened by Xilon (Xilon) - Saturday, 12 May 2007, 07:56 GMT
Last edited by Tobias Powalowski (tpowa) - Tuesday, 20 November 2007, 08:24 GMT
Task Type Feature Request
Category Kernel
Status Closed
Assigned To Tobias Powalowski (tpowa)
Architecture All
Severity Low
Priority Normal
Reported Version 0.8 Voodoo
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 2
Private No

Details

I just wanted to check out powertop, but it requires CONFIG_TIMER_STATS to work properly, I believe this is only available in 2.6.21. It also suggests to enable the CONFIG_USB_SUSPEND to allow USB suspending.

I'd appreciate if these options would be enabled, unless there's a reason for them not to be.
This task depends upon

Closed by  Tobias Powalowski (tpowa)
Tuesday, 20 November 2007, 08:24 GMT
Reason for closing:  Fixed
Comment by Tobias Powalowski (tpowa) - Saturday, 12 May 2007, 08:34 GMT
USB_SUSPEND breaks sane scanners, that was the reason for removing it
Comment by Attila (attila) - Sunday, 13 May 2007, 08:15 GMT
Cite from Documentation/hrtimer/timer_stats.txt: "It is not intended for production usage as it adds significant overhead to the (hr)timer code and the (hr)timer data structures."

I think it is better to make an own kernel package if you want something which is not needed by a lot of people or has the disadvantage with a "significant overhead". Since a longer time it is very easy to handle this with a copy of the files from /var/abs or from the homepage because of the excellent work of the devs. I have own kernel packages for both pc's here too because i have rare things which i want but could not be a good idea to use it in a kernel for a distribution. Give this a try, than you have the possibility to use powertop earlier because you don't have to wait for the decision about it and you can say to the kernel devs that it works good without breaking something else.
Comment by Xilon (Xilon) - Sunday, 13 May 2007, 08:32 GMT
Fair enough. CONFIG_TIMER_STATS isn't really that useful as far as I know, so i don't really care about that, too bad USP_SUSPEND breaks other things though.
Comment by Tobias Powalowski (tpowa) - Sunday, 13 May 2007, 08:38 GMT
closing this one now, it's not possible to implement it now
Comment by Scott H (stonecrest) - Monday, 12 November 2007, 08:38 GMT
  • Field changed: Percent Complete (100% → 0%)
Can we reconsider this?

According to the latest documentation at http://www.mjmwired.net/kernel/Documentation/hrtimer/timer_stats.txt: If enabled in the config but not used it has almost zero runtime overhead, and a relatively small data structure overhead. Even if collection is enabled runtime all the locking is per-CPU and lookup is hashed.
Comment by Scott H (stonecrest) - Sunday, 18 November 2007, 05:47 GMT
Looks like it's enabled in the testing kernel...

Loading...