| Age | Commit message (Collapse) | Author | |
|---|---|---|---|
| 2001-11-11 | Let ltsleep take a const wmesg. | Artur Grabowski | |
| 2001-11-06 | Replace inclusion of <vm/foo.h> with the correct <uvm/bar.h> when necessary. | Miod Vallat | |
| (Look ma, I might have broken the tree) | |||
| 2001-09-13 | Remove a comment that just doesn't make any sense. | Artur Grabowski | |
| 2001-08-07 | Change tsleep into ltsleep. | Artur Grabowski | |
| ltsleep takes an additional argument - a simplelock and unlocks it when it's safe to do so. tsleep now becomes a wrapper around ltsleep. From NetBSD | |||
| 2001-06-27 | remove old vm | Artur Grabowski | |
| 2001-06-24 | cold is in systm.h now | Michael Shalayeff | |
| 2001-05-26 | indentation. | Artur Grabowski | |
| 2001-03-25 | Reintroduce wakeup call | Constantine Sapuntzakis | |
| 2001-03-15 | Print a '*' in front of curproc in ps in ddb. | Artur Grabowski | |
| 2001-02-27 | Add wakeup_n and wakeup_one. wakeup_n will wakeup up to n sleeping processes | Constantine Sapuntzakis | |
| 2001-02-19 | When doing an assertion for phz, just do it once when we set phz, | Artur Grabowski | |
| not once per process. | |||
| 2000-11-10 | Change the ktrace interface functions from taking the trace vnode to taking the | Artur Grabowski | |
| traced proc. The vnode is in the proc and all functions need the proc. | |||
| 2000-08-03 | s/principal/priciple/; from netbsd | Michael Shalayeff | |
| 2000-07-06 | Typo in comment and some cleanup of roundrobin. | Artur Grabowski | |
| 2000-06-27 | Slight optimization of wakeup. | Artur Grabowski | |
| 2000-06-05 | Changes to exit handling. | Artur Grabowski | |
| cpu_exit no longer frees the vmspace and u-area. This is now handled by a separate kernel thread "reaper". This is to avoid sleeping locks in the critical path of cpu_exit where we're not allowed to sleep. From NetBSD | |||
| 2000-04-19 | Remove the roundrobin_attempts hack and replace it with per-process scheduling | Artur Grabowski | |
| flags (much nicer for future smp work). Add two generic functions yield() and preempt(). Use preepmt() in uio when we are told to yield. Based on my idea, code written by Jason Thorpe from NetBSD. | |||
| 2000-03-23 | Don't reinitialize the tsleep and ITIMER_REAL timers all the time. | Artur Grabowski | |
| The function and the argument never change. | |||
| 2000-03-23 | use the new timeout interface for tsleep. | Artur Grabowski | |
| 2000-03-23 | Adapt roundrobin and schedcpu to the new timeout API. | Artur Grabowski | |
| 2000-03-03 | Keep track of the number of times we trigger a reschedule before the | Artur Grabowski | |
| context switch actually happens. | |||
| 2000-03-03 | Use the LIST_FIRST macro to get the head of zombproc list. | Artur Grabowski | |
| 2000-03-03 | Use LIST_ macros instead of internal field names to walk the allproc list. | Artur Grabowski | |
| 1999-09-05 | Use stathz to calculate CPU time when available; fixes CPU calculation | Thorsten Lockert | |
| problems when stathz runs at different speed than hz/profhz. | |||
| 1999-08-15 | Adopt NetBSD fix for scheduler problems (nice was broken). From the NetBSD | Paul Janzen | |
| commit messages: Scheduler bug fixes and reorganization * fix the ancient nice(1) bug, where nice +20 processes incorrectly steal 10 - 20% of the CPU, (or even more depending on load average) * provide a new schedclock() mechanism at a new clock at schedhz, so high platform hz values don't cause nice +0 processes to look like they are niced * change the algorithm slightly, and reorganize the code a lot * fix percent-CPU calculation bugs, and eliminate some no-op code === nice bug === Correctly divide the scheduler queues between niced and compute-bound processes. The current nice weight of two (sort of, see `algorithm change' below) neatly divides the USRPRI queues in half; this should have been used to clip p_estcpu, instead of UCHAR_MAX. Besides being the wrong amount, clipping an unsigned char to UCHAR_MAX is a no-op, and it was done after decay_cpu() which can only _reduce_ the value. It has to be kept <= NICE_WEIGHT * PRIO_MAX - PPQ or processes can scheduler-penalize themselves onto the same queue as nice +20 processes. (Or even a higher one.) === New schedclock() mechanism === Some platforms should be cutting down stathz before hitting the scheduler, since the scheduler algorithm only works right in the vicinity of 64 Hz. Rather than prescale hz, then scale back and forth by 4 every time p_estcpu is touched (each occurance an abstraction violation), use p_estcpu without scaling and require schedhz to be generated directly at the right frequency. Use a default stathz (well, actually, profhz) / 4, so nothing changes unless a platform defines schedhz and a new clock. [ To do: Define these for alpha, where hz==1024, and nice was totally broke.] === Algorithm change === The nice value used to be added to the exponentially-decayed scheduler history value p_estcpu, in _addition_ to be incorporated directly (with greater weight) into the priority calculation. At first glance, it appears to be a pointless increase of 1/8 the nice effect (pri = p_estcpu/4 + nice*2), but it's actually at least 3x that because it will ramp up linearly but be decayed only exponentially, thus converging to an additional .75 nice for a loadaverage of one. I killed this: it makes the behavior hard to control, almost impossible to analyze, and the effect (~~nothing at for the first second, then somewhat increased niceness after three seconds or more, depending on load average) pointless. === Other bugs === hz -> profhz in the p_pctcpu = f(p_cpticks) calcuation. Collect scheduler functionality. Try to put each abstraction in just one place. | |||
| 1999-04-21 | Improved ps formatting. | Alex Feldman | |
| 1999-02-26 | uvm allocation and name changes | Artur Grabowski | |
| 1998-11-15 | GC unnecessary declaration | Artur Grabowski | |
| 1998-02-03 | bad types; wileyc@sekiya.twics.co.jp | Theo de Raadt | |
| 1997-11-06 | Updates for VFS Lite 2 + soft update. | Constantine Sapuntzakis | |
| 1997-11-04 | fix printf formatting of "ps" (aka "show all proc") so that lines never | chuck | |
| overflow (always hated that). replaced "/m" flag with: /a == show process address info /n == show normal process info [currently the default] /w == show process wait/emul info | |||
| 1997-10-06 | back out vfs lite2 till after 2.2 | Theo de Raadt | |
| 1997-10-06 | VFS Lite2 Changes | Constantine Sapuntzakis | |
| 1997-07-28 | two unneeded variables; enami@ba2.so-net.or.jp | Theo de Raadt | |
| 1997-01-19 | asm -> __asm | briggs | |
| 1996-11-23 | remrq -> remrunqueue | kstailey | |
| 1996-05-02 | sync syscalls, no sys/cpu.h | Theo de Raadt | |
| 1996-04-21 | partial sync with netbsd 960418, more to come | Theo de Raadt | |
| 1996-03-03 | From NetBSD: 960217 merge | Niklas Hallqvist | |
| 1995-10-18 | initial import of NetBSD tree | Theo de Raadt | |
 |
index : src | |
| OpenBSD base system |
| summaryrefslogtreecommitdiff |