On Mon, Mar 19, 2018 at 04:43:28PM -0400, Joe Lawrence wrote: > On 03/19/2018 04:17 PM, Peter Zijlstra wrote:
> I don't have the commit offhand that changed the printk behavior, but > from observation: 4bcc595ccd80 ("printk: reinstate KERN_CONT for printing continuation lines") and thereabout I think. > > Does pr_cont("\n") DTRT? it seems like something weird. > > > > Yeah, pr_cont() is kind of a hack. It will terminate a line if that's > the first newline, but as demonstrated above, I don't think it likes > extra newline chars. Fair enough I suppose, just a bit weird. > A better fix would be to marshal the text into temp buffer then dump it > out. Dunno if you prefer that kind of churn for these stats. Yeah, not worth it, simple is better here.