Re: Posix compliant CLOCK_PROCESS/THREAD_CPUTIME_ID V4
From: George Anzinger
Date: Wed Sep 29 2004 - 14:29:47 EST
Christoph Lameter wrote:
On Wed, 29 Sep 2004, George Anzinger wrote:
Christoph Lameter wrote:
George asked for a test program so I wrote one and debugged the patch.
The test program uses syscall to bypass glibc processing. I have been
working on a patch for glibc but that gets a bit complicated
because backwards compatibility has to be kept. Maybe tomorrow.
Found also that glibc allows the setting of these clocks so I also
implemented that and used it in the test program. Setting these
clocks modifies stime and utime directly, which may not be such a good
idea. Do we really need to be able to set these clocks?
Another way of doing this is to save these values in the task structure. If
null, use the direct value of stime, utime, if not, adjust by the saved value
(i.e. saved value would represent time zero).
But this would require two additional int field in task_t just for that
rarely used functionality.
Exactly. What does the community want? An alternative is to allocate a small
block of memory for this and managed it from the posix-timers.c code. It would
only be referenced by get_clock and set_clock. And be released by the
exit_itimers() code.
Please, when sending patches, attach them. This avoids problems with mailers,
on both ends, messing with white space. They still appear in line, at least in
some mailers (mozilla in my case).
The custom on lkml, for Linus and Andrew is to send them inline. I also
prefer them inline. Will try to remember sending attachments when sending a
patch to you.
I think they WILL be inline as well as attached if you attach them. The
difference is that in both presentations neither mailer will mess with white
space. This means that long lines will not be wrapped and tabs vs space will
not be changed.
Try sending yourself one and see it this is not true for your mailer.
As to the test program, what happens when you attempt to set up a timer on these
clocks? (No, I don't think it should work, but we DO want to properly error
out. And the test should verify that this happens.) By the way, if you use the
support package from sourceforge, you will find a lot of test harness stuff.
That is an interesting issue. If that would work correctly one could
trigger an signal if more than a certain amount of cputime is used.
It looks though that it will create an interrupt based on real time.
SuS says:
Each implementation defines a set of clocks that can be used as timing
bases for per-process timers. All implementations support a clock_id of
CLOCK_REALTIME.
So restrict timer_create to CLOCK_REALTIME and CLOCK_MONOTONIC? Is it
necessary to be able to derive a timer from a timer derives from those
two?
something like the following (just inlined for the discussion ...)?
NO. This is handled through the dispatch table (as set up when you register the
clock). You just supply a timer_create() function that returns the right error.
Likewise, attempts to use clock_nanosleep(). The issue with clock_nanosleep,
however, is that it, at this time, is not sent through the dispatch table. This
should be changed to, again call the same error function.
--- linux-2.6.9-rc2.orig/kernel/posix-timers.c 2004-09-28 20:29:28.000000000 -0700
+++ linux-2.6.9-rc2/kernel/posix-timers.c 2004-09-29 11:12:37.814713085 -0700
@@ -585,8 +585,8 @@
sigevent_t event;
int it_id_set = IT_ID_NOT_SET;
- if ((unsigned) which_clock >= MAX_CLOCKS ||
- !posix_clocks[which_clock].res)
+ if ((unsigned) which_clock != CLOCK_REALTIME &&
+ (unsigned) which_clock != CLOCK_MONOTONIC)
return -EINVAL;
new_timer = alloc_posix_timer();
--
George Anzinger george@xxxxxxxxxx
High-res-timers: http://sourceforge.net/projects/high-res-timers/
Preemption patch: http://www.kernel.org/pub/linux/kernel/people/rml
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/