Fix a bug in the initialization of the busreset_args that left the B channel
args unitialized and the A channel ones initialized to B's vales. Oops.
If we get a NO_IDENT sequencer interrupt (the reconnecting target didn't
issue an identify or botched it), reset the bus instead of panicing. We
should be able to recover from this error.
In the AWAITING_MSG handler, order messages by severity. Since the message
we send is based on a flag on the SCB, it is possible, during error recovery,
to get more than one flag set. This is fine since any time a new flag is
set, it is meant to take us to a more draconian level of recovery. This
also ensures that we don't lose any "history" of what the command has gone
through.
When we reset the bus, reset the "send ordered tag" bitmask.
Clear some additional interrupt status when we perform a bus reset.
a race condition in how SDTR and WDTR negotiation are handled, fixes for multi-lun
non-tagged device recovery, and ensuring that the timedout scbs in the waiting queue
are cleaned up.
Fix a problem with SCB paging that caused bogus residuals to be reported.
Cleanup of the disconnected list was broken in the SCB paging case
(confusion of NULLand SCB_LIST_NULL)
Implement a clean mechanism for determining that we have exited the timeout
state and test for this in ahc_done instead of all over the place.
Bring back the use of AAP (Auto Access Pause) I don't think it was the
true cause of the bus hangs people were reporting.
We want to reset the bus if we've been through an Abort action, not if
we are a recovery SCB (one implies the other, but not vice-versa).
will increase the overhead of queueing a command, but some recent bug reports
make me believe that AAP isn't really working and that we are losing some
SCBs from the input queue. Hopefully this will cure that problem.
Fix some bugs in the error recovery code. Mainly these could cause us to
inadvertantly forget to untimeout an SCB that was recovered causing later
confusion.
at runtime.
etc/make.conf:
Nuked HAVE_FPU option.
lib/msun/Makefile:
Always build the i387 objects. Copy the i387 source files at build
time so that the i387 objects have different names. This is simpler
than renaming the files in the cvs repository or repeating half of
bsd.lib.mk to add explicit rules.
lib/msun/src/*.c:
Renamed all functions that have an i387-specific version by adding
`__generic_' to their names.
lib/msun/src/get_hw_float.c:
New file for getting machdep.hw_float from the kernel.
sys/i386/include/asmacros.h:
Abuse the ENTRY() macro to generate jump vectors and associated code.
This works much like PIC PLT dynamic initialization. The PIC case is
messy. The old i387 entry points are renamed. Renaming is easier
here because the names are given by macro expansions.
"boot.config" (relative to the root directory on the 'a' partition
on the first BSD slice) if it exists. If it doesn't exist, then
the only visible changes should be that the kernel name isn't reset
to "/kernel" after looking it up fails and that the default name
is now "kernel".
The new function readfile() can be used for other things:
- reading help messages.
- reading splash screens.
- reading userconfig info.
Changed it from 4 to 16 for i386's. It can be anything for i386's,
but compiler options limit it to a power of 2, and assembler and
linker deficiencies limit it to a small power of 2 (<= 16).
We use 16 in the kernel to get smaller tables (see Makefile.i386 and
<machine/asmacros.h>). We still use the default of 4 in user mode.
Use HISTCOUNTER instead of (*kcount) in the definition of KCOUNT()
for consistency with other macros.
available in buffer when buffer was completely empty.
It now correctly reports the total buffer space available.
Reviewed by: jkh, davidg
Obtained from: Linux 1.3.20's sound driver code
filed in the hardware SCB not changing during the course of a transaction.
Since the sequencer now DMAs the hardware SCB back up to the host when it
detects a residual, this is no longer the case. I added a field to the
"software" scb to mirror this information and it is now used for doing the
residual calculation.
changes, so don't expect to be able to run the kernel as-is (very well)
without the appropriate Lite/2 userland changes.
The system boots and can mount UFS filesystems.
Untested: ext2fs, msdosfs, NFS
Known problems: Incorrect Berkeley ID strings in some files.
Mount_std mounts will not work until the getfsent
library routine is changed.
Reviewed by: various people
Submitted by: Jeffery Hsu <hsu@freebsd.org>
ULTRAENB->FAST20
Add a missing ahc_run_done_queue if a BRKADDRINT occurs. This should never
happen (haven't heard of one happening), but it was still a bug.
Brought the ordered tag sending code up into the tag code to be clearer.
If we decide we should send an ordered tag, only do so for the target that
timed out instead of all targets.
Initialize the STAILQ in ahc_serach__qinfifo. This was causing a panic
during some recovery operations.
Remove the unused varable maxtarget.
complained so it cannot be entirely bad :-)
I include the email that probably explains it for people who already know:
> >Compiling with -O3 inlines functions. However the function that is being
> >inlined in makeinfo.c (add_word_args()) is a vararg function and must not be
> >inlined.
> >
> >The code in question is K&R style, and AFIK, there is no way for the compiler
> >to determine that the function uses vararg. Either change the code to use
> >prototypes, or use stdarg, or add a directive to prevent inlining.
>
> Not declaring a varargs function as varargs before it is used gives
> undefined behaviour.
>
> However, in practice the bug is probably in FreeBSD's <varargs.h>, which
> doesn't use gcc's __builtin_next_arg(). gcc should notice that it is
> used and not inline functions that have it. <stdarg.h.> uses it, but I
> think there's another gcc builtin that it should be using.
Patch attached. The ellipsis causes gcc to flag this as a varargs function,
and the name "__builtin_va_alist" is special cased in gcc to hide the last
argument in the arglist.
Reviewed by: bde & phk
Submitted by: jlemon@americantv.com (Jonathan Lemon)
The PS/2 mouse device responds to a reset command with a sequence of
ACK(fa), RESULT(aa) and ID(00). Most PS/2 mice immediately returns
ACK, but spend sometime before sending RESULT. The Armada takes time
before ACK; extra delay is necessary before the call to read ACK.
The problem was reported in comp.unix.bsd.freebsd.misc and the patch
was tested by the reporter. No PR was filed, by the way.
to search the QINFIFO to remove any possible command that is waiting
otherwise our abort request may not be held up still waiting for the
first command to complete.
Fix a few panics during error recovery:
1) Stupid mistake in the "no SCB match handler" where I was using the wrong
variable (busy_scbid instead of scb_index).
2) Unbusy the target of an abort request if the command we are trying to
abort is an untagged transaction. If we don't, we get a fatal NO_MATCH_BUSY
condition which "should never happen".
3) When an abort completes, turn off ahc->in_timeout or else the next timeout
will hit the protective "scb timesout again" panic.
4) Fix a typo that caused the requeued "abort" SCB to have its TAG_ENB and
disconnect bits to be cleared (missing ~) so that devices would complain
about overlapped commands.
Be sure to turn off the unexpected busfree interrupt after we do a bus
reset since we are expecting the bus to go free in that case.
Return XS_TIMEOUT instead of XS_DRIVERSTUFFUP in certain scenarios. XS_TIMEOUT
allows for retries, XS_DRIVERSTUFFUP does not.
Allow commands with SDTR and WDTR negotiation to be tagged. The SCSI II spec
says that you probably should not do this for fear of hitting bogus devices.
The driver did this in the past for almost two years without any problem,
and not doing it causes problems during error recovery to a tag capable device
as the number of openings is higher than two and we'll start sending it
tagged commands causing "overlapped commands attempted" type errors. The
real fix needs to happen in the generic SCSI layer which can limit the
number and type of transactions to a device during error recovery efficiently.
Give ourselves at least 100ms to perform a request sense instead of relying
on the original timeout to be long enough to complete this new command as
well as the one that generated the condition.
Removed some redundant code.
with <= 100 usec between each character arrival time. This didn't happen
until rev.1.75 of clock.c because DELAY(100) used to delay for closer to
80 usec than 100 usec, and the minimum time between character arrivals is
87.8 usec at the maximum supported speed of 115200 bps 8N1.
Clear DCD timestamp flag on close (the input timestamp flag is already
cleared).
key "print scrn".
It used to stop at the first non-open vty, now it skips the non-open
ones and thereby enable one to cycle around all open vty by pressing
"print scrn".
I have code to calibrate the overhead fairly accurately, but there
is little point in using it since it is most accurate on machines
where an estimate of 0 works well. On slow machines, the accuracy
of DELAY() has a large variance since it is limited by the resolution
of getit() even if the initial delay is calibrated perfectly.
Use fixed point and long longs to speed up scaling in DELAY().
The old method slowed down a lot when the frequency became variable.
Assume the default frequency for short delays so that the fixed
point calculation can be exact.
Fast scaling is only important for small delays. Scaling is done
after looking at the counter and outside the loop, so it doesn't
decrease accuracy or resolution provided it completes before the
delay is up. The comment in the code is still confused about this.
- don't uselessly initialize the fifo "DMA" bit at attach time.
- initialize the fifo "DMA" bit at open time. Without this, the device
interrupts for every character received, reducing input performance
to that of an 8250.
- don't uselessly initialize the fifo trigger level to 8 (scaled to
256) at attach time.
- don't scale the fifo trigger level to 512 bytes. The driver's pseudo-
dma buffer has size 256, so it can't handle bursts of size 512 or 256.
It should be able to handle the second lowest ftl (2 scaled to 64).
- don't reset the fifos in siostop(). Reset triggers a hardware bug
involving wedging of the output interrupt bit This workaround
unfortunately requires ESP support to be configured.
If we can, use timeouts instead of DELAYs when dealing with a bus reset.
This prevents us from holding up the whole machine for a noticible amount
of time (especially for a real time app).
Make a pass over the timeout/error handling code. Aborts are more
reliable. We actually handle parity errors correctly now instead of
locking up the bus. Added code to properly clean up disconnected SCBs
down on the card during error handling. Improved robustness in several
areas.
If we are using defaults, but are an Ultra card, negotiate at 20MHz instead
of 10.
Don't attempt to handle any commands for 100ms after a reset has occured.
This is the minimum time before a target will respond to selection. Also
disable the busfree interrupt before doing a bus reset. This prevents the
driver from getting confused by an "unexpected busfree".
SEM* and SHM*. These are already supported in the options files. I
mostly used the default value plus 1. This ensures that the LINT kernel
depends on the options headers.
Style nit. Backslashes in macro weren't aligned.
aic7xxx.c:
Preserve the value of STPWEN in SXFRCTL1 during initialization. STPWEN
controls low byte termination and is setup by the PCI probe front end.
Disabling npx0 works right now.
Don't reference `npxdriver' if npx0 is not configured. Not configuring
npx0 doesn't quite work yet.
Don't clear potential non-npx pcb flags in setregs().
should be nearly impossible to overflow the QOUTFIFO (worst case 9 command
have to complete with at least 6 of them requiring paging on an aic7850),
so don't take the additional PIO hit to guard against this condition. If we
don't see our interrupt in time, the system has bigger problems elsewhere.
If this ever does happen, the timeout handler will notice and retry the
command.
Remove the ABORT_TAG sequencer interrupt handler. This condition can't happen
with the new SCB paging scheme.
Fix a few bugs noticed by Dan Eischen <deischen@iworks.InterWorks.org>
that could prevent ULTRA from being negotiated to drives above ID 7 and
also could allow an SCB to be passed to ahc_done twice during error recovery.
Fix a bug notice by Rory Bolt <rory@atBackup.com>. It turns out that a
sequencer reset will actually start the sequencer running regardless of the
state of the pause bit. This could lead to strange problems with loading
the sequencer.
not lazy-fault page table pages. Update the copyout support to take
that into account. This should fix some segfault problems on such
machines.
After a short test period, we'll move this into 2.2.
Submitted by: Stephen McKay <syssgm@devetir.qld.gov.au>
an X seesion. Really stupid error of me, and I've been looking at
this code SO many times. Thanks to Kazutaka YOKOTA for seeing this..
Submitted by: Kazutaka YOKOTA
one in draw_mouse causes spontanious hangs on my p5-100 when I
move the mouse excessively. Forgot that on the last commit, so
using the mouse or destructive cursor would produce large amounts
of flicker..
to -current.
Thanks goes to Ulrike Nitzsche <ulrike@ifw-dresden.de> for giving me
a chance to test this. Only the PCI driver is tested though.
One final patch will follow in a separate commit. This is so that
everything up to here can be dragged into 2.2, if we decide so.
Reviewed by: joerg
Submitted by: Matt Thomas <matt@3am-software.com>
importing it onto a vendor branch first, in the hope that this will
make future maintenance easier.
The conflicts are (hopefully) unimportant. More commits that actually
bring this into the source tree will follow.
Submitted by: Matt Thomas (thomas@lkg.dec.com)
cur_console is NULL when copy_font() is first called from scinit(). This
is apparently harmless when scinit() is called early from sccninit() -
page 0 is apparently mapped r/w then, and 0->status contains suitable
garbage. However, when there is a serial console, scinit() is first
called from scattach() when the page tables are completely initialized,
so the NULL pointer causes a panic.
Submitted by: bruce
called early for console i/o. The timer is usually in BIOS mode
if it isn't explicitly initialized. Then it counts twice as fast
and has a max count of 65535 instead of 11932. The larger count
tended to cause infinite loops for delays of > 20 us. Such delays
are rare. For syscons and kbdio, DELAY() is only called early
enough to matter for ddb input after booting with -d, and the delay
is too small to matter (and too small to be correct) except in the
PC98 case. For pcvt, DELAY() is not used for small delays (pcvt
uses its own broken routine instead of the standard broken one),
but some versions call DELAY() with a large arg when they unnecessarily
initialize the keyboard for doing console output. The problem is
more serious for pcvt because there is always some early console
output.
Guard against the i8254 timer being partially or incorrectly
initialized. This would have prevented the endless loop.
Should be in 2.2.
variable `kern.maxvnodes' which gives much better control over vnode
allocation than EXTRAVNODES (except in -current between 1995/10/28 and
1996/11/12, kern.maxvnodes was read-only and thus useless).
I have no idea if this works since I don't have one of the cards to test.
I also don't know what the LINT and GENERIC entries should look like,
so I just made up some values for now and left them commented out.
Someone who knows the factory settings for a Pro/10, please contact me!
Submitted-By: Javier Martín Rueda <jmrueda@diatel.upm.es>
when allocating memory for network buffers at interrupt time. This is due
to inadequate checking for the new mcl_map. Fixed by merging mb_map and
mcl_map into a single mb_map.
Reviewed by: wollman
This will make a number of things easier in the future, as well as (finally!)
avoiding the Id-smashing problem which has plagued developers for so long.
Boy, I'm glad we're not using sup anymore. This update would have been
insane otherwise.
previous hackery involving struct in_ifaddr and arpcom. Get rid of the
abominable multi_kludge. Update all network interfaces to use the
new machanism. Distressingly few Ethernet drivers program the multicast
filter properly (assuming the hardware has one, which it usually does).
compile again. The code to protect users from combining the dedicated
PCCARD drivers and the generic code is a warning if the above option
is included in the config file.
Demanded by: bde
(helptext from Philippe Regnauld)
- Make introfunc() work with serial terminals.
(submitted by Jean-Marc Zucconi)
- Eliminate excessive statusline redraw during screen updates.
(requested by Peter Wemm)
- Some trivial output formatting dinks.
> <sys/param.h> doesn't include <sys/time.h>
>
> I removed the NAPM check since it's wasteful to check twice. apmprobe()
> checks the unit number, and that's the right check.
Submitted by: bde
swapgeneric.c hasn't had anything to do with swapping for some time.
It just makes the -a boot option actually work, and breaks the static
configuration of rootdev and dumpdev. It should be reorganized to
only support -a.
etc..), plus add a better display suspend function.
- Changed the Copyright's to reflect the new 'jp.FreeBSD.org' email
address.
Submitted by: nate & HOSOKAWA, Tatsumi <hosokawa@jp.FreeBSD.org>
- the operands for bt, bts, arpl and `enter' were reversed.
- btr was reported as bts (with the correct operand order).
- cmpxchg was misplaced. It was misplaced differently in the
comments. It is misplaced differently again in the i486 manual.
I put it where the i586 manual and gas say it is.
- fucompp was misplaced.
- the rr table for(s) some versions of fstp, fcom and fcomp was non-null.
This caused some invalid opcodes to be reported as "" instead of as
"<bad instruction>".
- the word and long versions of the fi* instructions were reversed.
- aaa and daa were reversed.
Fixed bugs involving unusual operand sizes:
- 32-bit registers weren't always forced for bswap or for moves to and
from special registers.
- the operand sizes weren't reported for [l]call or [l]jmp.
- displacements weren't truncated mod 2^16 when the operand size was
16-bit.
- too-large displacements and offsets were fetched, and too-large
offsets were reported, when the operand size was 16-bit.
- sign extended immediate bytes were extended too far when the operand
size was 16-bit.
Fixed bugs involving usual operand sizes:
- 8-bit source registers weren't forced for mov[sz]b[wl].
- 16-bit source registers weren't forced for mov[sz]w[wl].
- immediate bytes were sometimes reported as sign extended even for
byte operations. Same for immediate words in word operations.
- the immediate byte was not reported as sign extended for `push'.
Finished Pentium support:
- cpuid, cmpxchg8b and rsm were missing.
Finished i287 support:
- fneni, fndisi and fsetpm were missing. These are harmless nops on
later FPUs.
Improvements:
- report invalid opcodes 0xd6 and 0xf1 using .byte. They are special
in not causing invalid operand exceptions when executed.
- report the immediate byte for unusual aam and aad instuctions.
Immediate bytes other than 0x0a always worked and are documented to
work on Pentiums.
wdparams from short into u_short. If wdp_cylinders is short, it
overflows and cause serious sign extension bug when large IDE HDD is
used. These members are only used for initialization of u_long
variables in both 3.0-current and RELENG_2_2 branch.
I believe this should be in 2.2.
Reviewed by: Bruce Evans <bde@zeta.org.au>
This code was sent to me by Bruce Evans, and seems to fix some
possible kernel panic in case of an execution error. It did not
cause any problems on my system, but I did never observe the
problem this patch is supposed to fix, anyway.
This patch is a NOP, unless the kernel is built with "options
USER_LDT", and doesn't affect the GENERIC kernel for this reason.
I want to have it in 2.2: it fixes a bug ...
Submitted by: bde
taken from the voxware-3.5 distribution. Also some changes to the SB
and MPU IRQs to reflect more common/default settings.
Submitted-By: Brian Campbell <brianc@netrover.com>
at device attach time, instead of allocating and freeing buffers as
necessary. But he or she forgot to remove the line that invalidated
the buffer when the device is closed. Therefore, after using the
device for the first time, the buffer was incorrectly invalidated and
that caused a page fault on the second, and subsequent uses.
Closes PR # kern/2319: Using Genius GS-4500 scanner...
Submitted by: jmrueda@diatel.upm.es (Javier Martmn Rueda)
also implies VM_PROT_EXEC. We support it that way for now,
since the break system call by default gives VM_PROT_ALL. Now
we have a better chance of coalesing map entries when mixing
mmap/break type operations. This was contributing to excessive
numbers of map entries on the modula-3 runtime system. The
problem is still not "solved", but the situation makes more
sense.
Eventually, when we work on architectures where VM_PROT_READ
is orthogonal to VM_PROT_EXEC, we will have to visit this
issue carefully (esp. regarding security issues.)
is still broken - it doesn't restore the floating point state.
2.2-BETA users should disable it using npx0 flags 0x04 the same as
2.2-ALPHA users should have.
Closes PR # kern/1065.
While i was at it, also reject IO requests that are not an integer
multiple of the device blocksize.
Submitted by: vak@crox.net.kiae.su (Serge V.Vakulenko)
Confirmed by: Georg-W. Koltermann (gwk@cray.com)
workings of #error in particular. He also broke the 2.2 build with this
change, leading me to wonder whether or not the changes were ever even
tested. Folks, I'm happy to see people work directly on 2.2 like
this and will continue to encourage Nate to make direct commits, but
please TEST before committing! I think that's a more than reasonable
prerequisite, and this code could never have worked at all, leading me to
believe that Nate skipped this most basic of steps.
Don't allow people to use the 'dedicated' drivers at the same time as
the generic support code, as it can cause all sorts of problems
including kernel crashes.
[ definite 2.2 material ]
Broke locking on named pipes in the same way as locking on non-vnodes
(wrong errno). This will be fixed later.
The fix involves negative logic. Named pipes are now distinguished from
other types of files with vnodes, and there is additional code to handle
vnodes and named pipes in the same way only where that makes sense (not
for lseek, locking or TIOCSCTTY).
This makes unexpected faults (in an interrupt handler) more likely
to crash properly. It could be done even better (more robustly and
more efficiently) using lazy fault handling.
They were garbage that happened to be 0 in many cases. (real_to_prot()
happens to leave the value of the real-mode %cr0 in %eax and the
memory-size BIOS calls usually don't touch the upper bits of %eax.
The upper 16 bits of %cr0 are usually 0 at boot time, at least on
486's.
Should be in 2.2.
configure() where it always belonged. It was originally slightly
misplaced after configure(). Rev.138 left it completely misplaced
before the DEVFS, DRIVERS and CONFIGURE sysinits by not moving it
together with configure().
Restored the printing of bootinfo.bi_n_bios_used now that it can
be nonzero.
EISA slots to probe. This is mainly intended to allow installing the
system on an HP Netserver with an on-board AIC7xxx EISA SCSI
controller, that is sitting on EISA slot # 11.
Documentation updates explaining this hack will follow shortly.
Note that this can go away again as soon as the EISA device probing
is more intelligent about the address space clash with the PCI address
space.
2.2 candidate.
Not objected by: freebsd-core :)
to TAILQs. Fix places which referenced these for no good reason
that I can see (the references remain, but were fixed to compile
again; they are still questionable).
the profiling level in config and decide what to do in makefiles.
Makefile.i386:
Align functions to 16-byte boundaries if profiling is enabled. This
will allow a fourfold reduction in the size of the profiling buffers.
I've added an installation from optical disk drive facility.
This enables FreeBSD to be installed from an optical disk, which
may be formatted in "super floppy" style or sliced into MSDOS-FS
and UFS partitions.
Note: ncr.c should be reviewed by Stefan Esser <se@freebsd.org>
and cd.c by Joerg Wunsch <joerg@freebsd.org> before bringing this
into 2.2.
Submitted-By: Shunsuke Akiyama <akiyama@kme.mei.co.jp>
makerom checksum check calculation was a no-op
nb8390.c had a bug which caused packet_len to be incorrect
for packets which wrapped in the buffer.
Submitted by: Linux developers (I lost the email with their names)
partition is 0, which is outside the slice unless the disk is
dangerously dedicated. There is little error checking later, and
following garbage metadata outside the slice usually caused crashes,
Empty 'a' partitions are fairly easy to create by cloning the label
for the whole disk.
Fixed a missing newline in a related error message.
type to be int so that errors can be returned.
2) Use the new SIOCSIFMTU ether_ioctl support in the few drivers that are
using ether_ioctl().
3) In if_fxp.c: treat if_bpf as a token, not as a pointer. Don't bother
testing for FXP_NTXSEG being reached in fxp_start()...just check for
non-NULL 'm'. Change fxp_ioctl() to use ether_ioctl().
ACTUALLY_LKM_NOT_KERNEL until the real fix comes through. Whatever
flag that is used to indicate building LKMs will have be be substituted
in the future. This allows proper selection of the usage of the
efficient, single instruction in the kernel, but alas, doesn't allow
for use in userland or LKMs :-(. Them's the breaks!!!
with one of the following (I486_CPU,I586_CPU,I686_CPU) so it can take
advantage of the very quick bswap insn. This keeps LKMs from being
built to take advantage of the insn, but also makes sure that the LKMs
can be run on all CPUs. (The LKMs don't pick up the CPU options :-( ).
function ed_attach_NE2000_pci() in if_ed.c passes
an uninitialized block of memory (got with malloc())
to ed_attach. This prevents a proper initialization
of the device descriptor and in my case causes a panic
during the probe, while printing out device info.
Reviewed by: phk
Submitted by: Luigi Rizzo <luigi@iet.unipi.it>
list of IP setsockopts the Linux emulator recognizes.
Explicitly disallow IP_HDRINCL since Linux's handling of
raw output is different than BSD's.
Closes PR#kern/2111.
Submitted by: y-nakaga@ccs.mt.nec.co.jp (Yoshihisa NAKAGAWA)
1. All the suggestions earlier made by Bruce: renaming some symbols,
stricter error checking, removing redundant code, etc.
2. The `psm' driver preserves the default counter resolution and
report rate, whatever they are after reset. (Based on reports and
suggestion from Nate and Rob Bolin).
3. The `psm' driver now does not check the so-called sync. bit in the
first byte of the data packet by default, so that the tapping feature
of ALPUS GlidePoint works (based on reports from Louis Mamakos). I
tested the code with ALPUS Desktop GlidePoint (M/N GP101) and found
no problem; tapping worked. It appears ALPUS produces several models
of GlidePoint. I hope the other models are OK too.
The check code can still be activated by defining the PSM_CHECKSYNC
option in the config file. (The bit checking slightly reduces, if not
completely eliminates, weird mouse behavior cased by unsynchronized
mouse data packets. It also helps us to detect if the mouse interrupt
can ever be lost. But, well, if there are devices which cannot be
supported this way...)
4. The `psm' driver does not include the protocol emulation code by
default. The code can still be compiled in if the PSM_EMULATION option
is specified in the config file. Louis Mamakos suggests the emulation
code is putting too much in the kernel, and `moused' works well.
I will think about this later and decide if the entire emulation
code should be removed.
5. And, of course, the fix in `scprobe()' from Bruce to cure the
UserConfig problem. My code in `kbdio.c' is slightly different from
his patch, but has the same effect. There still is a possibility that
`scprobe()' gets confused, if, for whatever reasons, the user holds
down a key for very long time during the boot process. But we cannot
cope with everything, can we?
Submitted by: Kazutaka YOKOTA (yokota@zodiac.mech.utsunomiya-u.ac.jp)
decrease the size of buffer_map to approx 2/3 of what it used to be
(buffer_map can be smaller now.) The original commit of these changes
increased the size of buffer_map to the point where the system would
not boot on large systems -- now large systems with large caches will
have even less problems than before.
are always together with Framing Errors and they were incorrectly
treated as FE's and discarded.
Reorganized the BREAK/FE/PE tests.
Found by: NIST-PCTS
with sio devices (not perfectly, since there is no way to flush the tx
holding register on 8250-16450's. I'm not sure if resetting the fifos
flushes the tx shift register).
Reminded by: NIST-PCTS
is completely empty. There is no interrupt for output completion, so
poll for it every 10 ms after output is nearly complete. Now ttywait()
works right.
Reminded by: NIST-PCTS
registers.) Also clean up some namespace pollution, and remove
gcc-1 support (nothing really works with it anymore anyway.)
Submitted by: Bruce Evans <bde@freebsd.org> and me.
consistent stack frame in fastmove() so that only one new fault handler
is necessary.
Should be in 2.2. Harmless until the i586 versions are reenabled.
Per Wayne Scott of Intel, the old sequence took 20cycles!!! on a P6.
Another nice side-benefit is that the kernel is about 3K smaller!!!
Submitted by: Wayne Scott <wscott@ichips.intel.com>
the keyboard IO port definitions from <machine/console.h>.
Note: I do not have a machine available to test this yet, I sent this
for review to several people who have not replied.
This is also totally busted in 2.2, which has half of the syscons/psm/kbdio
changes in it. This really desperately needs to be tested and fixed
in 2.2 as well.
SCBs in response to a busfree.
When re-queueing an SCB that returned with QUEUE FULL status, reset its timeout.
Ensure that aborted SCBs have an error code set in there xs before it gets
passed back up with scsi_done.
Fix a few KNF nits.
uses /usr/include/sys/*, which may point to a different build tree. I'm
not sure that this is necessary, but there was a question mark over what
/usr/include/sys points to when building the "user mode" binaries in the
kernel code, especially when building the smp tree.
I suspect that the "right" line here is to use ${INCLUDES}, but that
causes warnings about unused static inline functions in stdio.h and ctype.h
The 'getchar' function in syscons (sccngetc) is used by UserConfig to
get keyboard input from the user. When it was modified to use the
shared keyboard port routines it used the port passed in during the
probe routine. Since the probe routine was not yet called, the port was
set to 0, which is obviously not going to work.
Pre-initialize sc_port to IO_KBD which is really a kludge, but it's how
the previous driver did it's job.
Found by: remote GDB
substantially increasing buffer space. Specifically, we double
the number of buffers, but allocate only half the amount of memory
per buffer. Note that VDIR files aren't cached unless instantiated
in a buffer. This will significantly improve caching.
do reset it from the QOUTCNT register inside a pause/unpause. This now happens
once per command complete interrupt in the paging case (one interrupt can be
for multiple completed commands). I may introduce a counter and do a lazy
update in the future, similar to what is done with the QINCNT.
Enhance the QUEUE FULL condition handling so that the number of openings will
be reduced. This has become more important now that the driver is faster.
This code really belongs in the gerneric SCSI layer, as will be the case once
3.0 gets the code from the 'SCSI' branch.
Add some #if 0'd out trace code I've been using to help debug sequencer
problems.
Fix the SCB paging problem that I was seeing. This was only on my 7850
controller and stems from the fact that its QINFIFO can only handle 3bit
SCB identifiers. This means that you can only have 8 transactions open at
a time with the current paging scheme to these controllers. The code added
to enforce this is generic in that it tests for the number of relevent bits
that the QINFIFO can store and adjusts the max accordingly. It may be possible
to come up with a scheme that allows for more than 8 commands at a time, but
I don't know that it is worth the effort simply to fix a low end card. The
aic7880 still can do 255.
This problem may be related to what Andrey was seeing since I don't have n
aic7770 rev E chip here to test on, but as soon as someone probes one of these
cards with this new code, the dmesg output will tell the whole story.
aic7xxx cards failing on certain motherboards, reverse the logic used to
control this feature. AHC_FORCE_PIO is replaced with AHC_ALLOW_MEMIO.
GENERIC no longer needs to specify the AHC_FORCE_PIO option since this is
the default.
user supplies a bad address, because they push a lot of stuff that the
fault handler doesn't know about onto the stack. This has been broken
for more than half a year despite being tested for almost half a year
in -current.
synthesizer. The utilities for this will appear as port submissions soon
afterwards, according to the submitter.
Submitted-By: Randall Hopper <rhh@ct.picker.com>
Written-By: Takashi Iwai <iwai@dragon.mm.t.u-tokyo.ac.jp>
with the copyright stuff fixed so soon (this should be merged into 2.2 when
you have a chance, Poul).
This is the new AWE32 driver, with support for the AWE32's fancy MIDI
synthesizer. The utilities for this will appear as port submissions soon
afterwards, according to the submitter.
Submitted-By: Randall Hopper <rhh@ct.picker.com>
Written-By: Takashi Iwai <iwai@dragon.mm.t.u-tokyo.ac.jp>
routines. An older version of this was tested successfully on all of my
systems with PS/2 mice. This was brought in without testing because it
is necessary due to the previously committed syscons changes.
Submitted by: Kazutaka Yokota <yokota@zodiac.mech.utsunomiya-u.ac.jp>
syscons and psm, curtesy Kazutaka Yokota with minor changes by
me. This contains an update of the psm driver as well.
This also fixes the breakage that I introduced to the psm driver by
making syscons poll for keyboard events in the atempt to fix the
hanging keyboard problem.
It works perfectly for me, and I'd like to hear from all that
have had keyboard/ps/2 mouse problems if this is the cure...
Submitted by: Kazutaka YOKOTA (yokota@zodiac.mech.utsunomiya-u.ac.jp)
wdreg.h: Delete wd_ctlr macro. PC98 version of wd.c treats it as a
variable.
GENERIC98: Delete ep0 entry. Current ep driver write I/O port 0x100.
This clobbers ICW of i8259, because upper 8bits of address line is not
masked on mother board.
if_fe.c: Merge from revision 1.18 of sys/i386/isa/if_fe.c.
pc98.c: Globalize dmapageport, because SCSI driver use this
variable.
wd82371.c: Yet another merge.
These are 2.2 candidates.
Submitted by: The FreeBSD(98) Development Team
almost complete control over RTS (control of its initial value is still
missing).
This fixes PR 1644 for sio.
The author of PR 1644 wants it in 2.1.6 and 2.2. This may be safe since
the complications are only in rarely used cases that I hope I've covered.
sys/pc98/i386/machdep.c: sync with i386/i386/machdep.c
sys/pc98/conf/options.pc98: sync with i386/conf/options.i386
sys/i386/isa/sound: DMA auto initialize mode support for PC98.
contributed by: Akio Morita <amorita@bird.scphys.kyoto-u.ac.jp>
Definite 2.2 material, I believe.
Submitted by: The FreeBSD (98) Development Team
one too small for (hex) 12345678 and 4 too small for -1234567890. Large
values can be created by config and userconfig although not (previously)
by visual userconfig.
Fixed a sign extension bug for backspacing on "negative" hex values in
editval().
Increased field width and range for `flags' so that all possible values
can be displayed and edited.
actually harmless.
2. Fixed code to match comment in scintr().
3. Don't allow even root to take control of the machine when securelevel > 0.
I've secured the accesses to PSL_IOPL in all drivers and asked pst to
review it, but he seems to be busy. Write access to /dev/kmem and
other critival devices currently leaks across raisings of securelevel
via open fd's, so there may as well be a similar leak for PSL_IOPL.
4. (Most important.) Don't corrupt memory beyond the screen buffers if
the cursor happens to be off the 80x25 screen when syscons starts.
5. Fix console cursor update (not perfect yet).
Submitted by: bruce
~
getting the same behavior using the flags, which can be done inside of
UserConfig. (Also document other syscons flags which were previously
undocumented).
Requested by: bde
negative-logic flags (flags 0x01 and 0x02 for npx0, defaulting to unset = on).
This changes the default from off to on. The options have been in current
for several months with no problems reported.
Added a boot-time negative-logic flag for the old I5886_FAST_BCOPY option
which went away too soon (flag 0x04 for npx0, defaulting to unset = on).
Added a boot-time way to set the memory size (iosiz in config, iosize in
userconfig for npx0).
LINT:
Removed old options. Documented npx0's flags and iosiz.
options.i386:
Removed old options.
identcpu.c:
Don't set the function pointers here. Setting them has to be delayed
until after userconfig has had a chance to disable them and until after
a good npx0 has been detected.
machdep.c:
Use npx0's iosize instead of MAXMEM if it is nonzero.
support.s:
Added vectors and glue code for copyin() and copyout().
Fixed ifdefs for i586_bzero().
Added ifdefs for i586_bcopy().
npx.c:
Set the function pointers here.
Clear hw_float when an npx exists but is too broken to use.
Restored style from a year or three ago in npxattach().
for get_diskinfo(), whose BIOS call sets %es:%edi in some cases, although
most documentation says that it doesn't change %edi in the cases that
happened to matter (for hard disks).
This shall be in 2.1.6 and 2.2.
Submitted by: Tor.Egge@idt.ntnu.no
(except I kept the unnecessary preservation of %edx and %ecx)
are adding new drivers...) to match, as best I can tell, majors.i386.
Improve behaviour when attempting to save changes for devices that should
not be changeable. Now correclty avoids non-device items, PCI devices and
devices with no isa_device structure.
Submitted by: (observations from) joerg, bde
uses one or the other. This required some changes to the ahc_reset()
function, and how early the probes had to allocate their softc.
Turn the AHC_IN/OUT* macros into inline functions and lowercase their names
to indicate this change. Geting AHC_OUTSB to work as a macro doing
conditional memory mapped I/O would have been too gross.
Stop setting STPWEN in the main driver and let the PCI front end do it
instead. It knows better.
Add the clearing of the QOUTQCNT variable during command complete processing
in the SCB paging case.
Go back to doing unconditional retries for the QUEUE FULL status condition.
This is really a kludge, but the code to handle it properly is on the SCSI
branch and will not make it into 2.2.
Kazutaka YOKOTA
KATO Takenori
Bruce Evans
Justin T. Gibbs
Alexander Langer
Peter Wemm
Mike Smith
David Greenman
Søren Schmidt
Mike Pritchard
Steven Wallace
Julian Elischer
Darren Reed
John Dyson
Poul-Henning Kamp
Joerg Wunsch
David E. O'Brien
Garrett Wollman
Jordan K. Hubbard
Nate Williams
John Fieber
Stefan Eßer
Andrey A. Chernov
Martin Renters
Satoshi Asami
Bill Fenner