Messages in this thread | | | Date | Fri, 6 Aug 2010 23:26:48 +0200 | From | Wim Van Sebroeck <> | Subject | Re: [PATCH 00/15] hpwdt: make NMI code a config option (+ other cleanup) |
| |
Hi Dan,
> So overall: patches look good, but let's re-order the patches a bit (so that we first clean-up the driver and then add the NMI related changes (If we ever need to revert some things then we at least have a clean-driver before we start bisecting the NMI changes)). > And secondly: let's try to have the ifdef's out of the init and exit procedures. > > If you can look at how we can get rid of the ifdef's in init and exit, then I will reorder the patches and change the 3 timer patches and put that allready in a git tree.
I reorganised the sequence of patches. I'll sent them to you for verification (and will put a copy in linux-watchdog mailing list).
the patches now look like: watchdog: hpwdt (1/12): clean-up include-files. watchdog: hpwdt (2/12): Group options that affect watchdog behavior together watchdog: hpwdt (3/12): Group NMI sourcing specific items together watchdog: hpwdt (4/12): Despecificate driver from iLO2 watchdog: hpwdt (5/12): Make x86 assembly ifdef guard more strict watchdog: hpwdt (6/12): Introduce SECS_TO_TICKS() macro watchdog: hpwdt (7/12): allow full range of timer values supported by hardware watchdog: hpwdt (8/12): implement WDIOC_GETTIMELEFT watchdog: hpwdt (9/12): hpwdt_pretimeout reorganization watchdog: hpwdt (10/12): Construct status message w/ kasprintf and emit it with dev_info watchdog: hpwdt (11/12): Use "decoding" instead of "sourcing" watchdog: hpwdt (12/12): Make NMI decoding a compile-time option
(patch 12 can still be improved).
Kind regards, Wim.
| |