[lkml]   [1997]   [May]   [18]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectCurses writes
I know I've asked this before, but I've never really gotten an answer.

I've noticed for quite a while now that writes to the console sometimes
suck an inordinate amount of CPU and/or bus until they're done writing.

For instance, one can fire up a program that makes heavy use of both the
CPU and the system bus, such as Timidity, and then fire up a curses
program such as ncftp. Then run an 'ls' on the v2.1 kernel directory.
Timidity will skip, indicating it either didn't get its bus bandwidth, or
it didn't get its CPU. This occurs even if Timidity is chucked in at a -20

I've had this behavior confirmed by at least one other person. I first
noticed it in late 2.0.x, early 2.1.x, but that doesn't mean that it
didn't happen before then. I think someone pointed out a problem with
this and distributed a patch that used larger block writes or something of
that nature, but I don't know whether it was incorporated into the kernel
source tree or not.

Is this something to do with a bug in the scrup() function in console.c?


Good government. Good government. Sit. Stay.

 \ /
  Last update: 2005-03-22 13:39    [W:0.048 / U:34.140 seconds]
©2003-2018 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site