summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
1999-08-15Please give us dmesg from a GENERIC kernel only, if possible.Ian Darwin
1999-08-15Adopt NetBSD fix for scheduler problems (nice was broken). From the NetBSDPaul 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.
1999-08-15pciide & atapiscsikstailey
1999-08-14Missing -I/usr/include/ssl, evanc@concer.toAngelos D. Keromytis
1999-08-14note support for ASIX 88140A/88141-based PCI Ethernet adaptersAaron Campbell
1999-08-14regenAaron Campbell
1999-08-14add ASIX88141Aaron Campbell
1999-08-14Driver for ASIX88140A/88141 Ethernet; from FreeBSDAaron Campbell
1999-08-14Add -c flag that uses DIOCRLDINFO to clear (and update) the in-core label.Todd C. Miller
1999-08-14Typo fix.Hakan Olsson
1999-08-14syncFederico G. Schwindt
1999-08-14Support for 3CXEM556B-INT.Federico G. Schwindt
1999-08-14dynamic chunks, snake styleTodd T. Fries
1999-08-14debugging off by default. ok mickey@Todd T. Fries
1999-08-14updateMichael Shalayeff
1999-08-14updateMichael Shalayeff
1999-08-14fix debug printfs, ipl nameMichael Shalayeff
1999-08-14this work, smirk, rock, sock (rhymes by fries@ ;)Michael Shalayeff
1999-08-14cpu has interruptMichael Shalayeff
1999-08-14next try to print cpu model number.Michael Shalayeff
attach clock interrupt.
1999-08-14clock_intr()Michael Shalayeff
1999-08-14s/STACK_GROWS_UP/MACHINE_STACK_GROWS_UP/Michael Shalayeff
1999-08-14proper ipl values.Michael Shalayeff
spl's do not mask Ibit anymore, prevent int lossage.
1999-08-14littel perfomance rearngement, also cr23 and cr28 for ddbMichael Shalayeff
1999-08-14proper breakpoint instructionMichael Shalayeff
1999-08-14better CLKF_BASEPRIMichael Shalayeff
1999-08-14clock_intr() protoMichael Shalayeff
1999-08-14bigger stack frame sizeMichael Shalayeff
1999-08-14proper debugging ifdefsMichael Shalayeff
1999-08-14more registers to print and playMichael Shalayeff
1999-08-14enable interrupts after device configuration, provide cold_hook() for things ↵Michael Shalayeff
like lasi to enable power management
1999-08-14export cr23Michael Shalayeff
1999-08-14rework interruptions framewrok a little to use that space in the IV table.Michael Shalayeff
cpu_switch, setrunqueu, remrunqueu now tested and seem working. trapall needs some more work for user mode. optimizations here and there.
1999-08-14do the clock interrupt the narrow way.Michael Shalayeff
fix soft interrupts. some more debugs and other fixes.
1999-08-13Work with newest version of OpenSSL.Angelos D. Keromytis
1999-08-13Correct ssl include file(s) path.Angelos D. Keromytis
1999-08-13Oops, missed updating distfile version.Angelos D. Keromytis
1999-08-13Update the include paths.Angelos D. Keromytis
1999-08-13syncTheo de Raadt
1999-08-13AmbiCom AMB8002TTheo de Raadt
1999-08-13well of course it is Ethernet...Theo de Raadt
1999-08-13Use macros defined in pcmciavar.h.Federico G. Schwindt
1999-08-13Xircom NetWave support. Donations welcomed.Federico G. Schwindt
1999-08-13shutdown hook exists for the wrong reasonsTheo de Raadt
1999-08-13some repairsAaron Campbell
1999-08-13Missing commaTodd C. Miller
1999-08-13shorten dmesg output.Federico G. Schwindt
1999-08-13Display the correct name. Thanks niklas@.Federico G. Schwindt
1999-08-13Remove unncessary printfs as well as NetBSD ifdefs.Federico G. Schwindt
Make dmesg output shorter.
1999-08-13Better fix for fdesc/procfs panic.Constantine Sapuntzakis
generated by cgit v1.2.3 (git 2.25.1) at 2025-01-16 23:37:56 +0000