On Sat, 03 Jan 2004 09:53:13 EST, James Morris said: > > What's that first printk doing? It seems to just kick out a blank line: > This will ensure that the log record starts fresh on a new line, and is > not potentially mixed with a previously unterminated printk. Ahh. Makes sense, even if not much else in the kernel cares about it (I think the only way to trigger the problem it's protecting against is if one CPU is in the middle of a "build a message with multiple printk's w/o \n" section, and either another CPU/pre-empt does a printk, or if there's an LSM call in the middle of the series of printk's. I'm going to work around the issue in userspace by installing Modular Syslog (http://msyslog.sf.net) and using its regexp support to lose the blank lines, so no biggie. (Still need to go back and feed the fixes that msyslog needed back to the developers, it had a few small issues).
This archive was generated by hypermail 2b30 : Sat Jan 03 2004 - 15:22:30 PST