On Sun, Sep 04, 2005 Nish Aravamudan wrote: > On 9/4/05, Johannes Stezenbach <[EMAIL PROTECTED]> wrote: > > On Sun, Sep 04, 2005 Nish Aravamudan wrote: > > > On 9/4/05, Johannes Stezenbach <[EMAIL PROTECTED]> wrote: > > > > > > > > -#define UP_TIMEOUT (HZ/4) > > > > +#define UP_TIMEOUT (HZ*7/25) > > > > > > #define UP_TIMEOUT msecs_to_jiffies(280) > > > #define UP_TIMEOUT (7*msecs_to_jiffies(40) > > > > I agree it's nicer to read, but AFAIK not required for correctness? > > If so, then we'll fix those up in linuxtv.org CVS and submit > > cleanup patches later. > > Yeah, it's correct with the three current values of HZ (100, 250 and > 1000), but if you try a not-so-clean value (like Con did with 864, or > something), you might run into rounding issues. msecs_to_jiffies() > should take care of them (or will be a single point to do so > eventually).
Well, if msecs_to_jiffies() is the new way of specifying timeouts we'd have a lot more to fix up in our tree. But something like a remote control key-up timeout doesn't need much precision. Generally I see nothing wrong with HZ/4, but something like HZ*20/1000 could be problematic with small or odd HZ values. Agreed? Or is it desired that people generally use msecs_to_jiffies()? Johannes - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/