Hello Clemens, Thanks once again for this patch. See my notes below. On Tue, Oct 11, 2011 at 1:40 PM, Clemens Ladisch <clemens@xxxxxxxxxx> wrote: > Signed-off-by: Clemens Ladisch <clemens@xxxxxxxxxx> > > --- a/man2/sched_rr_get_interval.2 > +++ b/man2/sched_rr_get_interval.2 > @@ -98,18 +98,24 @@ > .\" for SCHED_RR processes. For RR processes the timeslice is always > .\" DEF_TIMESLICE, or 0.1 secs. > .\" Was the following really true for older kernels? > -.\" Must build a 2.6.8 kernel and retest, and rewrite the following text > -.\" appropriately. > +.\" It is again true for at least Linux 2.6.12, but stopped in 2.6.24. (At this point, I think we can simply drop these source code comments altogether. Your patch supplied the needed details.) > POSIX does not specify any mechanism for controlling the size of the > round-robin time quantum. > However, Linux provides a (nonportable) method of doing this. > -The quantum can be controlled by adjusting the process's nice value (see > +The quantum of both > +.B SCHED_RR > +and > +.B SCHED_OTHER > +processes can be controlled by adjusting the process's nice value (see I don't really want to document the SCHED_OTHER detail here. This page is about SCHED_RR processes. The point about SCHED_OTHER is an implementation detail that isn't relevant to SCHED_RR. > .BR setpriority (2)). > Assigning a negative (i.e., high) nice value results in a longer quantum; > assigning a positive (i.e., low) nice value results in a shorter quantum. > The default quantum is 0.1 seconds; > the degree to which changing the nice value affects the > quantum has varied somewhat across kernel versions. > +Since Linux 2.6.24, the > +.B SCHED_RR > +quantum cannot be modified. Here's the goodies! > .\" .SH BUGS > .\" As of Linux 1.3.81 > .\" .BR sched_rr_get_interval () So, I tweaked your patch somewhat, and applied the patch below for man-pages-3.36. Cheers, Michael --- a/man2/sched_rr_get_interval.2 +++ b/man2/sched_rr_get_interval.2 @@ -102,16 +102,9 @@ is available define in .IR <unistd.h> . .SS Linux Notes -.\" FIXME -.\" From at least Linux 2.6.9, the following is no longer true -.\" for SCHED_RR processes. For RR processes the timeslice is always -.\" DEF_TIMESLICE, or 0.1 secs. -.\" Was the following really true for older kernels? -.\" Must build a 2.6.8 kernel and retest, and rewrite the following text -.\" appropriately. POSIX does not specify any mechanism for controlling the size of the round-robin time quantum. -However, Linux provides a (nonportable) method of doing this. +Older Linux kernels provide a (nonportable) method of doing this. The quantum can be controlled by adjusting the process's nice value (see .BR setpriority (2)). Assigning a negative (i.e., high) nice value results in a longer quantum; @@ -119,6 +112,11 @@ assigning a positive (i.e., low) nice value results in a shorter quantum. The default quantum is 0.1 seconds; the degree to which changing the nice value affects the quantum has varied somewhat across kernel versions. +Since Linux 2.6.24, +.\" commit a4ec24b48ddef1e93f7578be53270f0b95ad666c +the +.BR SCHED_RR +quantum cannot be modified. .\" .SH BUGS .\" As of Linux 1.3.81 .\" .BR sched_rr_get_interval () -- Michael Kerrisk Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/ Author of "The Linux Programming Interface"; http://man7.org/tlpi/ -- To unsubscribe from this list: send the line "unsubscribe linux-man" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
