POC: Alternative solution: Re: [PATCH 0/4] printk: reimplement LOG_CONT handling

Joe Perches joe at perches.com
Fri Aug 14 23:08:02 EDT 2020


On Fri, 2020-08-14 at 19:33 -0700, Linus Torvalds wrote:
> On Fri, Aug 14, 2020 at 4:52 PM Joe Perches <joe at perches.com> wrote:
> > On Fri, 2020-08-14 at 15:46 -0700, Linus Torvalds wrote:
> > > This is why I think any discussion that says "people should buffer
> > > their lines themselves and we should get rid if pr_cont()" is
> > > fundamentally broken.
> > > 
> > > Don't go down that hole. I won't take it. It's wrong.
> > 
> > I don't think it's wrong per se.
> 
> It's *absolutely* and 100% wrong.
> 
> Yes, any random *user* of pr_cont() can decide to buffer on it's own.

Which I believe is the point of the discussion,
not the complete removal of KERN_CONT.

> But when the discussion is about printk() - the implementation, not
> the users - then it's complete and utter BS to talk about trying to
> get rid of pr_cont().
> 
> See the difference?

Sure, but I fail to see where anyone said get rid of pr_cont
in this thread.  It seems all that was discussed was just
various schemes to improve coalescing output.





More information about the kexec mailing list