- more req[uest]->xfer changes.
- get the corresponding NetBSD Id's right
ohci.c
- move untimeout above print statement
- remove usb_delay that panics the system (tsleep in intr context) when
ohcidebug > 5.
ugen.c
- create the devices for endpoints with make_dev.
uhub.c
- change from using usbdebug to uhubdebug
- add more debugging statements
files (opt_*.h) automatically (if they are in ${SRCS}).
Clean vnode_if.[ch] automatically (if one of them is in ${SRCS}, not just
if VFS_KLD is defined).
There are some complications to avoid using the "@" symlink before it
is built.
(kern.randompid), which is currently defaulted off. Use ARC4 (RC4) for our
random number generation, which will not get me executed for violating
crypto laws; a Good Thing(tm).
Reviewed and Approved by: bde, imp
commit to kern_synch.c:
----------------------------
revision 1.55
date: 1999/02/23 02:56:03; author: ross; state: Exp; lines: +39 -10
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 schedclk() 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 schedclk() mechansism === 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. 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 wieght) 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.
----------------------------
The details are a little different in FreeBSD:
=== nice bug === Fixing this is the main point of this commit. We use
essentially the same clipping rule as NetBSD (our limit on p_estcpu
differs by a scale factor). However, clipping at all is fundamentally
bad. It gives free CPU the hoggiest hogs once they reach the limit, and
reaching the limit is normal for long-running hogs. This will be fixed
later.
=== New schedclk() mechanism === We don't use the NetBSD schedclk()
(now schedclock()) mechanism. We require (real)stathz to be about 128
and scale by an extra factor of 2 compared with NetBSD's statclock().
We scale p_estcpu instead of scaling the clock. This is more accurate
and flexible.
=== Algorithm change === Same change.
=== Other bugs === The p_pctcpu bug was fixed long ago. We don't try as
hard to abstract functionality yet.
Related changes: the new limit on p_estcpu must be exported to kern_exit.c
for clipping in wait1().
Agreed with by: dufault
commit to kern_synch.c:
----------------------------
revision 1.55
date: 1999/02/23 02:56:03; author: ross; state: Exp; lines: +39 -10
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 schedclk() 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 schedclk() mechansism === 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. 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 wieght) 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.
----------------------------
The details are a little different in FreeBSD:
=== nice bug === Fixing this is the main point of this commit. We use
essentially the same clipping rule as NetBSD (our limit on p_estcpu
differs by a scale factor). However, clipping at all is fundamentally
bad. It gives free CPU the hoggiest hogs once they reach the limit, and
reaching the limit is normal for long-running hogs. This will be fixed
later.
=== New schedclk() mechanism === We don't use the NetBSD schedclk()
(now schedclock()) mechanism. We require (real)stathz to be about 128
and scale by an extra factor of 2 compared with NetBSD's statclock().
We scale p_estcpu instead of scaling the clock. This is more accurate
and flexible.
=== Algorithm change === Same change.
=== Other bugs === The p_pctcpu bug was fixed long ago. We don't try as
hard to abstract functionality yet.
Related changes: the new limit on p_estcpu must be exported to kern_exit.c
for clipping in wait1().
Agreed with by: dufault
X server, is not responding to the VT switching protocol. (This part
of the code has been somewhat wrong in -CURRENT, but -STABLE has the
correct code...)
PCI SCSI controllers. This driver also supports the following Symbios/LSI
PCI SCSI chips: 53C810A, 53C825A, 53C860, 53C875, 53C876, 53C885, 53C895.
However, it does NOT support earlier chips as the following ones: 53C810,
53C815, 53C825.
See README.sym for more details.
Submitted-by: Gerard Roudier <groudier@club-internet.fr>
and extend. The new function containing the code is named schedclock()
as in NetBSD, but it has slightly different semantics (it already handles
incrementation of p->p_cpticks, and it should handle any calling frequency).
Agreed with in principle by: dufault
Add MD_ROOT and MD_ROOT_SIZE options to the md driver.
Make the md driver handle MFS_ROOT and MFS_ROOT_SIZE options for compatibility.
Add md driver to GENERIC, PCCARD and LINT.
This is a cleanup which removes the need for some of the worse hacks in
MFS: We really want to have a rootvnode but MFS on a preloaded image
doesn't really have one. md is a true device, so it is less trouble.
This has been tested with make release, and if people remember to add
the "md" pseudo-device to their kernels, PicoBSD should be just fine
as well. If people have no other use for MFS, it can be removed from
the kernel.
- Convert to new bus attachment scheme. Thanks to Blaz Zupan for doing
the initial work here. One thing I changed was to have the attach
and detach routines work like the PCI drivers, which means that in
theory you should be able to load and unload the driver like the PCI
NIC drivers, however the pccard support for this hasn't settled down
yet so it doesn't quite work. Once the pccard work is done, I'll have
to revisit this.
- Add device wi0 to PCCARD. If we're lucky, people should be able to
install via their WaveLAN cards now.
- Add support for signal strength caching. The wicontrol utility has
also been updated to allow zeroing and displaying the signal strength
cache.
- Add a /sys/modules/wi directory and fix a Makefile to builf if_wi.ko.
Currently this module is only built for the i386 platform, though once
the pccard stuff is done it should be able to work on the alpha too.
(Theoretically you should be able to plug one of the WaveLAN/IEEE ISA
cards into an alpha with an ISA slot, but we'll see how that turns out.
- Update LINT to use only device wi0. There is no true ISA version of
the WaveLAN/IEEE so we'll never use an ISA attachment.
- Update files.i386 so that if_wi is dependent on card.
from vm_map_pageable(). At the point they called, vm_map_pageable()
holds a read (or shared) lock on the map. The purpose
of vm_map_{clear,set}_recursive() is to disable/enable repeated
write (or exclusive) lock requests by the same process.
- Use 'or' operation to change b_flags.
- SCSI HDD device is 'da', not 'sd'.
Submitted by: kura@tim.hi-ho.ne.jp (Tomohiko Kurahashi) and
chi@bd.mbn.or.jp (Chiharu Shibata)
with NetBSD and the Single Unix Specification v2.
This updates some structures with other, almost equivalent types and
effort is under way to get the whole more consistent.
Also removes a double definition of INET6 and some other clean-ups.
Reviewed by: green, bde, phk
Some part obtained from: NetBSD, SUSv2 specification
by 2 with people just adding numbers on the end of the ethernet subtypes.
We now have an additional 14 subtypes available in ethernet.
Use one of them immediatly for homePNA.
Reviewed by: Garrett Wollman <wollman@khavrinen.lcs.mit.edu>
parameter a char ** instead of a const char **. This make these
kernel routines consistent with the corresponding libc userland
routines.
Which is actually 'correct' is debatable, but consistency and
following the spec was deemed more important in this case.
Reviewed by (in concept): phk, bde
is neither IFF_LOOPBACK or IFF_POINTOPOINT. It's quite common
(and probably more correct) to route local IP numbers via lo0
and it makes configuration easier to assign the hostname address
to local POINTOPOINT links too.
This message usually remains hidden because the loopback interface
gets the highest interface number at boot time, but when the
ethernet interface is added later, the message can get pretty
annoying.
Also, fix a typo.
Not objected to by: freebsd-net
we instead use 0(%esp), which is per-cpu, already pretty much
guarenteed to be locked into the cache, and does not stress the cache's
set associativity. invlpg might also be a good choice (suggested by
Ingo).
Obtained from: Linus Torvalds <torvalds@transmeta.com>
MP lock for the last time. The use of a locked instruction to
cpu-private memory is 3x faster then CPUID and 3x faster then the
use of a locked instruction to shared memory (the lock itself).
Instruction serialization is required to ensure that any pending
memory ops are properly flushed prior to the release of the lock,
due to out-of-order instruction execution by the cpu.
Files sysdep.[ch] are now in ${MACHINE_ARCH} subdirectory. Internal
#if's used to identify the platform where removed.
Make rule for target testmain was greatly simplified, because it was
easier simplifying it than changing it to support the new location of
sysdep.[ch].
(a repo-copy was done on sysdep.[ch], of course)
vm_map always failed because vm_map_lookup() looked at
"vm_map_entry->wired_count" instead of "(vm_map_entry->eflags &
MAP_ENTRY_USER_WIRED)". The effect was that many page
wiring operations by sysctl were (silently) failing.
Since an osigcontext is smaller, if you check for a valid (much larger sized)
ucontext_t and it fails, we bogusly would reject the osigcontext as per
rev 1.378. Instead, check for osigcontext range validity first, and
ucontext_t later. This unbreaks Netscape.
Pointed to the right commit by: peter
