[lkml]   [2011]   [May]   [5]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
    SubjectRe: [rtc-linux] Re: [PATCH 0/3] repair RTC subsys (for i.MX)
    On Wed, May 04, 2011 at 03:30:59PM -0700, john stultz wrote:
    > On Wed, 2011-05-04 at 17:31 +0200, Wolfram Sang wrote:
    > > The recent updates to the RTC subsystem (removing UIE interrupts and use
    > > alarms instead) introduced two problems for i.MX (and the subsys in
    > > general, I'd think):
    > >
    > > a) because registering the rtc now calls get_alarm(), the requirement
    > > has been added for a lot of drivers that drvdata is properly set up
    > > _before_ registering. rtc-mxc did not do that (probably bad; but as said
    > > like a lot of other rtc-drivers currently) and oopsed.
    > >
    > > b) the callbacks to the rtc-core for update_irqs have been removed, but
    > > irq-handlers are still there, now being unused cruft. In case of
    > > rtc-mxc, this is only one if-block, but for rtc-mc13xxx.c this is a
    > > seperate handler.
    > >
    > > From a glimpse, most platform drivers seem to have at least one of these
    > > problems now :( John, am I correct or am I missing something?
    > I've tried to go through and clean up most of the b) issues, although
    > some have apparently slipped by. Please let me know of any others you
    > ran across.

    I discovered it when working on the rtc-stmp3xxx driver. For this, I
    will send a patch series in a few minutes, because there are more things
    to be done there. Grepping for 'RTC_UF' should point to other areas of

    > And indeed we've hit a few of issue a) already, so I should probably run
    > through and do a full audit.

    I hacked a q'n'd coccinelle-script pointing out potential candidates. It
    found two more for which I will send patches, too. But I think a
    seperate audit would be a good idea, just to be sure.

    > Your patches look fine to me. Do you intend to push them or should I
    > queue them up?

    I'd say they should go via your tree.



    Pengutronix e.K. | Wolfram Sang |
    Industrial Linux Solutions | |
    [unhandled content-type:application/pgp-signature]
     \ /
      Last update: 2011-05-05 11:33    [W:0.022 / U:8.856 seconds]
    ©2003-2017 Jasper Spaans. hosted at Digital OceanAdvertise on this site