Messages in this thread | | | Date | Mon, 21 Jan 2008 22:04:54 +0900 | From | Tejun Heo <> | Subject | Re: [PATCHSET] printk: implement printk_header() and merging printk, take #2 |
| |
Matthew Wilcox wrote: > On Mon, Jan 21, 2008 at 02:13:52PM +0900, Tejun Heo wrote: >> In a nutshell, printk_header() lets you do the following atomically >> (against other messages). >> >> code: >> printk(KERN_INFO "ata1.00: ", "line0\nline1\nline2\n"); >> >> output: >> <6>ata1.00: line0 >> <6> line1 >> <6> line2 > > I think this is a really bad idea. It's much better to have: > > <6>ata1.00: line0 > <6>ata1.00: line1 > <6>ata1.00: line2 > > That way you can grep for ata1.00 and get all messages relevant to that > device.
It has some pros and cons. Having indentation makes things easier on human eyes but more difficult on scripts, but not too difficult - indented lines can easily be matched with a bit of scripting. Changing the behavior is easy but I'm still inclined toward indentation because printing header every line loses information about message boundaries. Maybe there's a way to satisfy both like omitting the separator from the second line on.
ata1.00: line0 ata1.00 line1 ata1.00 line2
Hmm... Any better ideas?
-- tejun
| |