the NTP kernel PLL is disabled, and acquire_timer0() is enabled, thus
opening the door for microtime() (and hence gettimeofday()) to return
bogus timestamps. This option is necessary for the `pca' driver to
work, but is implemented to underscore the fact that accurate timekeeping
and the `pca' driver are incompatible at present. If someone writes a version
of microtime() that works when the `pca' driver is being used, this can get
junked.
1) check va before clearing the page clean flag. Not doing so was
causing the vnode pager error 5 messages when paging from
NFS. (pmap.c)
2) put back interrupt protection in idle_loop. Bruce didn't think
it was necessary, John insists that it is (and I agree). (swtch.s)
3) various improvements to the clustering code (vm_machdep.c). It's
now enabled/used by default.
4) bad disk blocks are now handled properly when doing clustered IOs.
(wd.c, vm_machdep.c)
5) bogus bad block handling fixed in wd.c.
6) algorithm improvements to the pageout/pagescan daemons. It's amazing
how well 4MB machines work now.
1) Removed all instances of disable_intr()/enable_intr() and changed
them back to splimp/splx. The previous method was done to improve
the performance, but Bruces recent changes to inline spl* have
made this unnecessary.
2) Cleaned up vm_machdep.c considerably. Probably fixed a few bugs, too.
3) Added a new mechanism for collecting page statistics - now done by
a new system process "pagescan". Previously this was done by the
pageout daemon, but this proved to be impractical.
4) Improved the page usage statistics gathering mechanism - performance is
much improved in small memory machines.
5) Modified mbuf.h to enable the support for an external free routine when
using mbuf clusters. Added appropriate glue in various places to
allow this to work.
6) Adapted a suggested change to the NFS code from Yuval Yurom to take
advantage of #5.
7) Added fault/swap statistics support.
finally have the f**king documentation!):
1) Changed all the numeric register offsets to symbolic ones (it should
have been this way originally).
2) If 16 bit, disable the shared memory when not using it. Apparantly
switching between 8/16bit mode makes the Ultra unhappy unless
this is done (i.e. it trashes the bus).
- ansi prototypes in lpt.c
- a bit of tidying in lpt.c
- ioctl in lpt.c for switching between polling and using interrupts
- added lpt.h - needed for ioctl to allow switching between polling
and interrupt-driven modes.
1) fixed some bugs related to the bounce buffer code
2) vnode pager now supports clustered pageouts
3) experimental code for clustering all I/O via a new "cldisksort"
4) added >16MB check to Bustek driver
5) made some experimental algorithmic changes to the pageout daemon
6) fixed bugs in truncating mapped files (esp when mapped via NFS)
7) reorganized vnode pager I/O code
class of locale data, but could be extended to handle other locale
classes, as well as message catalogues and other non-locale i18n
support.
I have left the old _ctype_ array in place, and moved the ctype.h
header to octype.h, so that existing shared binaries will still be
able to find and use it as they require.
See /usr/src/share/locale for information on how to create new locale
data files (eventually this procedure will be improved). I'd like to
have a family of locale files for various countries, languages, and
character sets, so please contribute some.
This code was originally written by Paul Borman and contributed to
4.4; I did the integration, and have somewhat tested it. crt0.c
probably ought to call setlocale() if it doesn't already, but I'd like
for people to create some locale files and try things manually first
before I make every program do this.
---
This list of changes is in approximately chronological order (oldest first).
o Many cosmetic changes - renamed comintr1 -> siointr1, moved
things around and fixed whitespace.
o Reduced SLIP latency (FRAME_END hack) from 20-30 ms to 16 ms
at 115200 bps (you won't notice the average 10 ms improvement
on slow lines). ppp seems to use only counted transfers so
there's no similar hack available. It's too hard for the
driver to know the count.
o Temporary #ifdefs for new and old interrupt handling
(OLD_INTERRUPT_HANDLING decided by setsofttty() not being
externally defined.
o Don't test for the IIR_NOPEND bit being set - test for the
non-fifo part of the iir equalling it like the docs say to.
States with other IIR_NOPEND set in combination with the
other iir bits are undefined. The docs may be stupid - the
old test would not have broken when the fifo bits were
introduced.
o Noted more problems with DTR wait.
o Rewrote console stuff. Still some initialization and state
preservation problems. Same for kgdb stuff. The driver
doesn't do anything about the console close bug. It needs
to be fixed entirely in i386/cons.c. I like chmr's version
where the the console driver revectors the device open and
close routines.
o Temporary (?) #ifdefs for references to tty buffers.
o Noted further things to do in (2 comments about 3 places) for
phk's change to not touch RTS unless it is being used for flow
control.
o Temporary #ifdefs for timestamp handling. It needs fixing.
The microtime() call breaks the first rule of writing fast
interrupt handlers: NO calls to functions that might do slow
and bad things. microtime() enables interrupts. This turns
out to be only moderately harmful. Also, I want the timestamp
copy outside of the normal interrupt handler.
o Don't init com->tp early for the !DONT_MALLOC_TTYS case -
both sides are NULL.
o Worry about com->tp == NULL in siopoll. I don't see how you
survived the (incc <= 0 || !(tp->state & TS_ISOPEN)) test.
Perhaps early sttys or comcontrols set up the tp's for _all_
the ports before this code is reached.
list of changes, I've made the following additional changes:
1) i386/include/ipl.h renamed to spl.h as the name conflicts with the
file of the same name in i386/isa/ipl.h.
2) changed all use of *mask (i.e. netmask, biomask, ttymask, etc) to
*_imask (net_imask, etc).
3) changed vestige of splnet use in if_is to splimp.
4) got rid of "impmask" completely (Bruce had gotten rid of netmask),
and are now using net_imask instead.
5) dozens of minor cruft to glue in Bruce's changes.
These require changes I made to config(8) as well, and thus it must
be rebuilt.
-DG
from Bruce Evans:
sio:
o No diff is supplied. Remove the define of setsofttty(). I hope
that is enough.
*.s:
o i386/isa/debug.h no longer exists. The event counters became too
much trouble to maintain. All function call entry and exception
entry counters can be recovered by using profiling kernel (the new
profiling supports all entry points; however, it is too slow to
leave enabled all the time; it also). Only BDBTRAP() from debug.h
is now used. That is moved to exception.s. It might be worth
preserving SHOW_BITS() and calling it from _mcount() (if enabled).
o T_ASTFLT is now only set just before calling trap().
o All exception handlers set SWI_AST_MASK in cpl as soon as possible
after entry and arrange for _doreti to restore it atomically with
exiting. It is not possible to set it atomically with entering
the kernel, so it must be checked against the user mode bits in
the trap frame before committing to using it. There is no place
to store the old value of cpl for syscalls or traps, so there are
some complications restoring it.
Profiling stuff (mostly in *.s):
o Changes to kern/subr_mcount.c, gcc and gprof are not supplied yet.
o All interesting labels `foo' are renamed `_foo' and all
uninteresting labels `_bar' are renamed `bar'. A small change
to gprof allows ignoring labels not starting with underscores.
o MCOUNT_LABEL() is to provide names for counters for times spent
in exception handlers.
o FAKE_MCOUNT() is a version of MCOUNT() suitable for exception
handlers. Its arg is the pc where the exception occurred. The
new mcount() pretends that this was a call from that pc to a
suitable MCOUNT_LABEL().
o MEXITCOUNT is to turn off any timer started by MCOUNT().
/usr/src/sys/i386/i386/exception.s:
o The non-BDB BPTTRAP() macros were doing a sti even when interrupts
were disabled when the trap occurred. The sti (fixed) sti is
actually a no-op unless you have my changes to machdep.c that make
the debugger trap gates interrupt gates, but fixing that would
make the ifdefs messier. ddb seems to be unharmed by both
interrupts always disabled and always enabled (I had the branch in
the fix back to front for some time :-().
o There is no known pushal bug.
o tf_err can be left as garbage for syscalls.
/usr/src/sys/i386/i386/locore.s:
o Fix and update BDE_DEBUGGER support.
o ENTRY(btext) before initialization was dangerous.
o Warm boot shot was longer than intended.
/usr/src/sys/i386/i386/machdep.c:
o DON'T APPLY ALL OF THIS DIFF. It's what I'm using, but may require
other changes.
Use the following:
o Remove aston() and setsoftclock().
Maybe use the following:
o No netisr.h.
o Spelling fix.
o Delay to read the Rebooting message.
o Fix for vm system unmapping a reduced area of memory
after bounds_check_with_label() reduces the size of
a physical i/o for a partition boundary. A similar
fix is required in kern_physio.c.
o Correct use of __CONCAT. It never worked here for non-
ANSI cpp's. Is it time to drop support for non-ANSI?
o gdt_segs init. 0xffffffffUL is bogus because ssd_limit
is not 32 bits. The replacement may have the same
value :-), but is more natural.
o physmem was one page too low. Confusing variable names.
Don't use the following:
o Better numbers of buffers. Each 8K page requires up to
16 buffer headers. On my system, this results in 5576
buffers containing [up to] 2854912 bytes of memory.
The usual allocation of about 384 buffers only holds
192K of disk if you use it on an fs with a block size
of 512.
o gdt changes for bdb.
o *TGT -> *IDT changes for bdb.
o #ifdefed changes for bdb.
/usr/src/sys/i386/i386/microtime.s:
o Use the correct asm macros. I think asm.h was copied from Mach
just for microtime and isn't used now. It certainly doesn't
belong in <sys>. Various macros are also duplicated in
sys/i386/boot.h and libc/i386/*.h.
o Don't switch to and from the IRR; it is guaranteed to be selected
(default after ICU init and explicitly selected in isa.c too, and
never changed until the old microtime clobbered it).
/usr/src/sys/i386/i386/support.s:
o Non-essential changes (none related to spls or profiling).
o Removed slow loads of %gs again. The LDT support may require
not relying on %gs, but loading it is not the way to fix it!
Some places (copyin ...) forgot to load it. Loading it clobbers
the user %gs. trap() still loads it after certain types of
faults so that fuword() etc can rely on it without loading it
explicitly. Exception handlers don't restore it. If we want
to preserve the user %gs, then the fastest method is to not
touch it except for context switches. Comparing with
VM_MAXUSER_ADDRESS and branching takes only 2 or 4 cycles on
a 486, while loading %gs takes 9 cycles and using it takes
another.
o Fixed a signed branch to unsigned.
/usr/src/sys/i386/i386/swtch.s:
o Move spl0() outside of idle loop.
o Remove cli/sti from idle loop. sw1 does a cli, and in the
unlikely event of an interrupt occurring and whichqs becoming
zero, sw1 will just jump back to _idle.
o There's no spl0() function in asm any more, so use splz().
o swtch() doesn't need to be superaligned, at least with the
new mcounting.
o Fixed a signed branch to unsigned.
o Removed astoff().
/usr/src/sys/i386/i386/trap.c:
o The decentralized extern decls were inconsistent, of course.
o Fixed typo MATH_EMULTATE in comments. */
o Removed unused variables.
o Old netmask is now impmask; print it instead. Perhaps we
should print some of the new masks.
o BTW, trap() should not print anything for normal debugger
traps.
/usr/src/sys/i386/include/asmacros.h:
o DON'T APPLY ALL OF THIS DIFF. Just use some of the null macros
as necessary.
/usr/src/sys/i386/include/cpu.h:
o CLKF_BASEPRI() changes since cpl == SWI_AST_MASK is now normal
while the kernel is running.
o Don't use var++ to set boolean variables. It fails after a mere
4G times :-) and is slower than storing a constant on [3-4]86s.
/usr/src/sys/i386/include/cpufunc.h:
o DON'T APPLY ALL OF THIS DIFF. You need mainly the include of
<machine/ipl.h>. Unfortunately, <machine/ipl.h> is needed by
almost everything for the inlines.
/usr/src/sys/i386/include/ipl.h:
o New file. Defines spl inlines and SWI macros and declares most
variables related to hard and soft interrupt masks.
/usr/src/sys/i386/isa/icu.h:
o Moved definitions to <machine/ipl.h>
/usr/src/sys/i386/isa/icu.s:
o Software interrupts (SWIs) and delayed hardware interrupts (HWIs)
are now handled uniformally, and dispatching them from splx() is
more like dispatching them from _doreti. The dispatcher is
essentially *(handler[ffs(ipending & ~cpl)]().
o More care (not quite enough) is taken to avoid unbounded nesting
of interrupts.
o The interface to softclock() is changed so that a trap frame is
not required.
o Fast interrupt handlers are now handled more uniformally.
Configuration is still too early (new handlers would require
bits in <machine/ipl.h> and functions to vector.s).
o splnnn() and splx() are no longer here; they are inline functions
(could be macros for other compilers). splz() is the nontrivial
part of the old splx().
/usr/src/sys/i386/isa/ipl.h
o New file. Supposed to have only bus-dependent stuff. Perhaps
the h/w masks should be declared here.
/usr/src/sys/i386/isa/isa.c:
o DON'T APPLY ALL OF THIS DIFF. You need only things involving
*mask and *MASK and comments about them. netmask is now a pure
software mask. It works like the softclock mask.
/usr/src/sys/i386/isa/vector.s:
o Reorganize AUTO_EOI* macros.
o Option FAST_INTR_HANDLER_USERS_ES for people who don't trust
fastintr handlers.
o fastintr handlers need to metamorphose into ordinary interrupt
handlers if their SWI bit has become set. Previously, sio had
unintended latency for handling output completions and input
of SLIP framing characters because this was not done.
/usr/src/sys/net/netisr.h:
o The machine-dependent stuff is now imported from <machine/ipl.h>.
/usr/src/sys/sys/systm.h
o DON'T APPLY ALL OF THIS DIFF. You need mainly the different
splx() prototype. The spl*() prototypes are duplicated as
inlines in <machine/ipl.h> but they need to be duplicated here
in case there are no inlines. I sent systm.h and cpufunc.h
to Garrett. We agree that spl0 should be replaced by splnone
and not the other way around like I've done.
/usr/src/sys/kern/kern_clock.c
o splsoftclock() now lowers cpl so the direct call to softclock()
works as intended.
o softclock() interface changed to avoid passing the whole frame
(some machines may need another change for profile_tick()).
o profiling renamed _profiling to avoid ANSI namespace pollution.
(I had to improve the mcount() interface and may as well fix it.)
The GUPROF variant doesn't actually reference profiling here,
but the 'U' in GUPROF should mean to select the microtimer
mcount() and not change the interface.
/dev/mcd0a instead of /dev/mcd0d. This is more conforming to the /dev/cd0a
for the SCSI cdrom drives. It breaks the convention d the whole drive.
But the question is, do we really need partitions on cdrom drives ?
put vers.o at the end of the loader line. We are simply jumping in the
moment into the first location of the text segment in 386bsd. So the
linking order is very important :-). With the addition of the const
types in newvers.sh we jumped into them. I have experimented with an
entry point specification, but was unsuccessfull. Someone else should
look at this.
devices.i386:
files.i386:
Added entries for a Sony cdrom driver.
1) tty.c: gather all the info about the processes before calling ttyprintf
(which may block).
2) syscons.c: handle asynchronous output properly (data structures may
be corrupted otherwise).
Example:
Application use port cua01
Getty open ttyd1 (allocates rawq,outq,etc) and waits while application
done
Application quits, sioclose issued, ttyfree issued (getty calls revoke)
Getty awakes and goes to panic into initrb (NULL rawq)
a zero cylinder and/or head count, set them and the sector count
to some fictious values. The same already happens if the request
sense is unsuccessfull.
1) A new mechanism has been added to prevent pages from being paged
out called "vm_page_hold". Similar to vm_page_wire, but
much lower overhead.
2) Scheduling algorithm has been changed to improve interactive
performance.
3) Paging algorithm improved.
4) Some vnode and swap pager bugs fixed.
now HUPCL set only in bidir case for callin lines
(this prevents set HUPCL on mouse)
comhardclose:
in addition to HUPCL case now DTR dropped for bidir case
if line was active in and no carrier present now.
(this prevents DTR sleep on mouse)
Subject: Re: Bugs with floppy drives
Date: Tue, 8 Mar 94 9:11:54 CST
The transfer speed was only set in the retry after error, not when
switching drives.
Eliminates vm_fault overhead on process startup and
mmap referenced data for in-memory pages.
(process startup time using in-memory segments *much* faster)
2) Even more efficient pmap code. Code partially cleaned up.
More comments yet to follow.
(generally more efficient pte management)
3) Pageout clustering ( in addition to the FreeBSD V1.1 pagein
clustering.)
(much faster paging performance on non-write behind disk
subsystems, slightly faster performance on other systems.)
4) Slightly changed vm_pageout code for more efficiency and
better statistics. Also, resist swapout a little more.
(less likely to pageout a recently used page)
5) Slight improvement to the page table page trap efficiency.
(generally faster system VM fault performance)
6) Defer creation of unnamed anonymous regions pager until needed.
(speeds up shared memory bss creation)
7) Remove possible deadlock from swap_pager initialization.
8) Enhanced procfs to provide "vminfo" about vm objects and user
pmaps.
9) Increased MCLSHIFT/MCLBYTES from 2K to 4K to improve net &
socket performance and to prepare for things to come.
John Dyson
dyson@implode.root.com
David Greenman
davidg@root.com
it still looks a little suspicious that so many of the status codes are missing
so I'm not going to adopt all of the existing ones yet. Try to be more
descriptive in the use of hex constants.
Changes _only_ take effect if `options LAPTOP' is set.
Note: This one is distinctly dodgy. When my IDE drive spins back up from sleep
mode, it generates this `extra interrupt' condition by spinning back up and
generating an intr, though without any particular action required. This
message coming out every time is rather annoying, and thus disabled.
However, what I'm not at all sure of is whether or not all IDE drives will
behave in the same way, or if perhaps it needs to be done in a more complicated
fashion by detecting this more involved "I've spun up and am just saying hi"
condition. This is a simple change and easy to back out/ammend if anybody has
any better ideas.
hack in the moment for testing purposes and to get the drive going
again.
0x20 means empty drive.
0x30 means closed drive with CDROM inserted.
0x80 means drive pulled out, but door closed.
0xa0 means drive pulled out and door open.
Luckily none of these values are the same as that reported for Ethernet
cards ( 0 for WD8003E, 0x40 for WD8013EPC, 0x60 for NE2000).
The bad part is, the probe code gets the WD8003E so hosed, that it is
no longer usable after it. No problem with the WD8013EPC.
my previous fix too (using wdp_heads controller value) and check
0 heads case too.
Other fixes from Bruce:
2) Fix dk_timeout from 2 to about 4 seconds.
3) wdcontrol not retried on internal error.
4) wdwait return check changed ( "!=0" to "<0" into wdsetctlr,
"<0" to "!=0" into wdgetctlr).
This inetrface should be used from now on.
pseudo device pty xx still keeps its meaning: a maximum of
xx ptys is allowed.
A ringbuffer is now 2040 bytes long, per Garrett Wollman's request.
The changes are inspired by the way NetBSD did it (thanks for that!),
though I made it slihghtly different, including the interface so
at least 75% of the allocated space is deallocated when the tty is
closed.
Note further that it is easy to modify the ringbuffer length runtime.
This will have to wait untill some later date...
-Guido
This inetrface should be used from now on.
pseudo device pty xx still keeps its meaning: a maximum of
xx ptys is allowed.
A ringbuffer is now 2040 bytes long, per Garrett Wollman's request.
The changes are inspired by the way NetBSD did it (thanks for that!),
though I made it slihghtly different, including the interface so
at least 75% of the allocated space is deallocated when the tty is
closed.
Note further that it is easy to modify the ringbuffer length runtime.
This will have to wait untill some later date...
-Guido
>From: bde@kralizec.zeta.org.au (Bruce Evans)
Subject: cache botch in bootstrap
Date: Sun, 20 Feb 94 18:14:35 +1100
The cache in the bootstrap loader didn't allow for the device changing.
This caused surprisingly little trouble - the cache is only for a single
track (or part of a track), and the first access to a new device is
always for an early sector, so there is no problem unless the last access
to the old device was for an early sector. I saw the problem on a system
with BSD on wd1 and no label on wd0. Everything worked if the the device
name wd(1,a) was specified before the default of wd(0,a) was tried, but
when the default was tried, it failed on the first sector because there
was no label, and then the first sector on wd0 was always used instead
of the first sector on wd1, so there was apparently no label on wd1
either.
Bruce
Subject: Bugfix for SB16 with DSP version 4 and above
No description sent, but it appears to fix a major number problem
with certain models of the SB16.
Small hack, if heads > 16, output a warning and then set the head
count to 16. This stops the infinite loop on this error and allows
people to later fix the DOS parameter table later with fdisk.
run Exabyte 8505 on 1742's. This may not be the final solution, but it
makes it work. It may be better to change the DELAY(10) to DELAY(30)
inside the loop instead of increasing the loop count from 100 to 300.
Subject: Bug Fix for ft.c - please commit
Date: Mon, 14 Feb 94 10:22:54 CST
This fixes a bug where the system can crash if the tape is used
after the floppy has been used.
idea considering the old version (but I haven't time for attic surgery - this
is just a caution). This is part of (and the major reason for) the new ld
changes.
a pv chain was being traversed while interrupts were
fully enabled in pmap_remove_all ... this is bogus, and
has been fixed in pmap.c. (sorry for adding the splimp)
Description:
The integer overflow instruction (into) and the interrupt instruction with
value 4 (int #4) both give rise to SIGBUS signals rather than SIGFPE. The
problem is that overflow is a trap not a fault (unlike the BOUND instruction).
>From: Adam David <adam@veda.is>
Subject: SCSI timeout in sdopen()
Date: Sun, 6 Feb 1994 21:40:48 GMT
Index: sys FreeBSD-current
Reference: FreeBSD-current/sys/9
Description:
sdopen() times out on slow SCSI devices. This code was moved into
a different file, which might explain why it got missed again.
Repeat-By:
using slow SCSI disks
When the keyboard is probed, the LED's blink quickly and
"Keyboard reset failed" is printed on the console. The
init routine keeps trying endlessly with the same behavior
as above.
I got the latest -current sup sources (06-Feb-94 12:00 GMT) to work using the
old syscons.c. The following patch makes the new syscons work:
2) Make SNAKE_SAVER like default, if no saver specified in "options"
3) Remove #ifdef STAR_SAVER before line /* make screensaver happy */
this code needed in any case.
Subject: syscons-1.3
Date: Sat, 29 Jan 94 23:33:50 MET
But here is the (hopefully) final syscons-1.3....
....
I've changed sgetc so it works as the pccons parallel
(it now uses a scgetc internally).
[
There were a couple changes that Bruce Evans sent me that were applied
to this version along with some changes that S'ren didn't incorporate
into the final version. There will be only minor changes if anything
from this version to his final release.
]
o merged and fixed timeout code from sos.
o merged DOS partition support from sos.
o fixed "extra" interrupts.
o check if malloc retuns NULL :-(.
o print drive size as in NetBSD
o after an error, return the residual count, not 0.
o give up early for > 16 heads.
o cleaned up the old-drive detection in wdgetctlr().
o rewrote wddump(). I'd been putting this off because
I'd had the dump disabled since 0.0 and had no idea
if even the standard version worked. It didn't:
(0) syscons' sgetc() interface broke some keyboard
stuff.
(1) CADDR1 is mapped using pmap() and pmap got a NULL
pointer trap (at least when I got to wddump() using
call diediedie from ddb) because the pte for CADDR1
is only supposed to be hacked on directly (e.g. in
physcopyseg()).
(2) bad144 handling was not done.
(3) it was slooow (3600 sectors/minute) because my
controller doesn't cache writes.
o miscellaneous other cleanups, e.g., removed scattered
patchkit/terry dates.
o lots of reformatting.
To be done:
o Merge/fix TIH/NetBSD bad144 code (doesn't belong in any
particular driver. Why aren't we using i386/dkbad.c?).
o Merge/fix Dyson/NetBSD clustering code (large parts
should be shared).
o Fix/extend the partition in use bits. Support extended
partitions. This should be shared by all disk drivers. Swap
to a DOS partition so that the swap space can be shared with
linux.
o Don't use polling except for initialization. Need to
reorganise the state machine. Then "extra" interrupts
shouldn't happen (except maybe one for initialization).
o Fix disklabel, boot and driver inconsistencies with
bad144 in standard versions.
o Look at latest linux clustering methods. Our disksort()
gets in the way of clustering.
o cleaned up the ATA changes (needs more work. I think
the ATA specs say that the only thing really wrong with
the original version of wx is that there was no test
for drive-ready before some commands. 400 nsec delay
suffices for almost everything and 400 nsec delay is
usually "free").
