c393eae7c5
event_var.h changed location.
1996-05-30 01:40:07 +00:00
9a9cf8c3f5
Put back a `#if defined(SUN4M)' for now.
1996-05-30 00:57:35 +00:00
e86275a0b6
Some slight optimizations.
1996-05-30 00:02:09 +00:00
b3666a87b4
Swiatch to `sys/dev/sun/event*'.
1996-05-29 21:57:51 +00:00
3d1e64415a
Switch to `sys/dev/sun/event.c'.
1996-05-29 21:56:38 +00:00
f9f5f23ee3
We really need a ring buffer for every unit.
...
Eliminate extraneous ttyalloc() calls.
1996-05-29 21:45:28 +00:00
4899447ffc
Add missing splx(), per Jason.
1996-05-29 21:08:28 +00:00
57e61589e7
Fix two cases of handling stale page table information:
...
1. when reading the referenced/modified bits the TLB entry must be flushed
before reading the in-core version.
2. when wrapping up an entire segment in pmap_page_protect(), flush the PTPs
from the TLB to prevent a table-walking operation to pick up stale - or
possibly bogus - PTEs.
(hopefully I'll get a few of my hairs back now..)
1996-05-29 20:58:38 +00:00
5583238255
impliment ttylist stats based on disk stats.
1996-05-29 02:02:42 +00:00
65d186d8e8
find a couple minor nits from previous change.
1996-05-29 01:58:09 +00:00
6ce8e31376
change zsinfo to a zs_softc, and impliment intr counts. mostly from OpenBSD.
1996-05-28 14:06:28 +00:00
08bd090b26
count audio interupts (from OpenBSD).
1996-05-28 13:36:09 +00:00
9913321bff
CMP_PTE_USER_READ4M: check against PPROT_RWX_RWX; this is the canonical
...
stack protection (saves many excursions into trap.c on window underflows
as was the sole intention of the macro..).
1996-05-27 20:46:39 +00:00
d94f5b92f9
prototype mmu_pagein4m() #if DEBUG is on.
1996-05-27 20:29:03 +00:00
75eb006fa2
No longer need to call mmu_pagein() through a pointer.
1996-05-27 01:13:02 +00:00
522c1d367a
mmu_pagein() is only useful on sun4/sun4c. For now, keep a `mmu_pagein4m()'
...
within `#ifdef DEBUG' for monitoring.
Push user windows to stack in pmap_extract() if we need to switch contexts.
1996-05-27 01:12:34 +00:00
4fa564e10d
Fix case where sequence state wasn't updated properly, resulting in
...
`disk_unbusy' panic.
1996-05-27 00:10:40 +00:00
43b3743977
Implement CMP_PTE_USER_WRITE4M more accurately.
...
CMP_PTE_USER_READ4M needs some more work..
1996-05-26 22:33:31 +00:00
2382529ecd
Some more #defines.
1996-05-26 22:30:20 +00:00
f699a85094
sync with reality.
1996-05-24 23:18:23 +00:00
3c13903472
Enable just enough interrupts to get "L1-A" through. Patch from Jason to
...
avoid spurious floppy interrupts on Jason's machine...
1996-05-23 22:35:11 +00:00
bc77c57c13
espattach() requires DMA to be configured. For now, panic if it is not.
1996-05-21 19:09:43 +00:00
68691c178c
Allow all esp dma channels to be called `dma'.
1996-05-21 19:07:30 +00:00
dc7e7814d9
Remove the `espdma' special case; all `esp' related dma channels are now
...
called `dma' on all architectures. Note: `ledma' remains in place.
Also delete a couple of unneeded `needs-flag's in files.sparc.
1996-05-21 19:06:26 +00:00
71943d50df
Protect body of iommuattach() to be more resilient against configuration
...
file oddities.
1996-05-21 07:25:07 +00:00
6da494dde9
Put a `return' back that went missing in rev 1.56
1996-05-20 10:49:20 +00:00
722d0da592
RCS Id police.
1996-05-20 00:56:39 +00:00
1bd6f1f147
We don't generally want to profile `genassym' (easy part of PR#2434).
1996-05-19 20:58:51 +00:00
0e4623e2f7
set the console struct winsize to the values given in the prom (sun4c/sun4m)
...
or the eeprom (sun4) when opening the console. (note: sun4 code untested
but jason claims it will `Just Work').
1996-05-19 13:00:43 +00:00
fae3ba5c32
disable prom sync in romboot() and romhalt(), rather than before calling them.
1996-05-19 04:12:53 +00:00
d292da40b2
Remove obsolete HWTOSW/SWTOHW macros.
...
VA2PA: flush TLB before proceeding with L2 probe (per the manual) [4m].
More KNF.
1996-05-19 00:32:15 +00:00
98a9570a8e
Make sure DELAY(0) returns within an hour..
1996-05-19 00:25:16 +00:00
9ce523cca7
disable the prom `sync' command before calling romboot() or
...
romhalt() (idea from OpenBSD).
1996-05-18 12:36:49 +00:00
4cba75e24a
put promdev definition into <machine/bsd_openprom.h>.
...
also, minor KNF.
1996-05-18 12:35:18 +00:00
b6988904e0
put promdev definition into <machine/bsd_openprom.h>.
1996-05-18 12:27:40 +00:00
d7bc39fad1
don't set tmp twice.
1996-05-18 12:22:49 +00:00
68b2ff1bdb
use __P
1996-05-18 12:19:14 +00:00
da5c96baa1
Correct an oversight and a silly past-o in last change.
1996-05-17 22:54:40 +00:00
05379d078e
Add `_kernel_text' symbol for kvm_mkdb(8).
1996-05-17 20:07:53 +00:00
d9277a9036
Add Sun4m power0 power status/control register to GENERIC configuration
1996-05-17 18:02:56 +00:00
2d2098c11b
Check for `configuration parent' differently: compare against "sbus"
...
rather than "ledma".
1996-05-16 22:57:32 +00:00
b9f70e90d8
Add `flags' to the esp entries: disable disconnect & synch negotiation
...
by default until further notice.
1996-05-16 22:47:47 +00:00
e509f9aaeb
Change attributes of `dma', `espdma' and `ledma' (the attributes on the latter
...
two didn't seem to make much sense anyway..) to allow `esp' devices to be
attached to one of `sbus', `dma' and `espdma'.
Remove the wildcarded `espdma?' and `ledma?' attachments of `esp' and `le'
respectively, in favour of `dma?' and `lebuffer?' (but the latter is not
yet implemented), which seems to better match reality: additional SBus
SCSI/Lance boards call themselves `dma' and `lebuffer'.
1996-05-16 22:38:04 +00:00
3f1283a92f
Re-arrange code that looks for the esp driver to allow for the fact that
...
any "dma" device (i.e. not just "espdma") can have an OBP child. This is
the case with (at least some) SBus SCSI boards.
1996-05-16 21:45:35 +00:00
50e92fd71d
Use configuration file flags to disable reselection and/or sync negotiation,
...
on a per target basis (until the driver can sort things out on its own).
Test against "sbus" in stead of "espdma" to find out where in the
configuration tree we are: an esp can be the child of a "dma" on SBus
add-on boards.
1996-05-16 20:31:28 +00:00
0b59ea6fbd
Replace a couple of (*foo_p)()'s with direct calls. All instances occurred
...
within in functions that were already mmu-arch specific.
Some formatting nits.
1996-05-16 19:19:33 +00:00
4bbf23d04c
Copyright police (s/Harvard University/Harvard College/).
1996-05-16 15:56:54 +00:00
fe454d89f3
Implement a hack to give pmap a better chance of working on SS10's with
...
no external L2 cache.
XXX this is ugly and will go away when cpu_softc gets done for the 1.3 release
Also, copyright police (s/Harvard University/Harvard College/).
1996-05-16 14:30:54 +00:00
81ba0732e0
Document the "flags" for the si and sw controllers. While I'm here,
...
do some re-arrangement (cleanup), and document which devices exist on
which machines, so that people can be better informed when they trim
down their kernels. There are a LOT of comments in this file now!
1996-05-16 03:13:27 +00:00
81b0485bfe
These haven't been kept up to date.
1996-05-15 23:57:50 +00:00
9ad87ec76e
remove some RCS id's we don't need.
1996-05-15 02:13:39 +00:00
16b514adc6
KNF some debug statements.
1996-05-14 13:57:29 +00:00
ebafcd57c2
Remove some comments in eeprom_uio() now that they're no longer relevant.
...
Add some whitespace to eeprom_take().
1996-05-13 21:39:53 +00:00
1a79065f9f
Enable DMA by default on the 4/100 "sw" controller; the DMA bugs appear
...
to have been fixed.
1996-05-13 01:53:45 +00:00
540f9550c0
Make `make depend' work for libraries.
1996-05-11 16:11:30 +00:00
ac9df157aa
Changed struct ifnet to have a pointer to the softc of the underlying
...
device and a printable "external name" (name + unit number), thus eliminating
if_name and if_unit. Updated interface to (*if_watchdog)() and (*if_start)()
to take a struct ifnet *, rather than a unit number.
1996-05-07 01:28:28 +00:00
ae87feee24
Make the MI LANCE driver standalone, and use cfattach to resolve
...
naming conflicts between bus attachments on ports that can have
multiple instances of the LANCE.
Changed struct ifnet to have a pointer to the softc of the underlying
device and a printable "external name" (name + unit number), thus eliminating
if_name and if_unit. Updated interface to (*if_watchdog)() and (*if_start)()
to take a struct ifnet *, rather than a unit number.
1996-05-07 01:27:26 +00:00
0e0d4f1ec2
Make the MI LANCE driver standalone, and use cfattach to resolve
...
naming conflicts between bus attachments on ports that can have
multiple instances of the LANCE.
1996-05-07 01:24:48 +00:00
0ba8565c86
mountroot: return error code to caller, not -1.
1996-05-04 19:36:01 +00:00
4ac7468812
Various simplifications and corrections from Chuck Cranor (makes `oclock'
...
calibration work.. thanks!).
1996-05-02 18:17:33 +00:00
504582ba83
Simpler and more accurate delay() function from Chuck Cranor.
1996-05-02 18:15:08 +00:00
3fff2a8f83
Fix apparent paste-o ... spurious "}"
1996-04-30 00:56:45 +00:00
fa75ea39e1
Move delay() to locore, so we control code generation and can avoid
...
touching memory locations whose caching disposition can perturb timings.
1996-04-29 21:08:37 +00:00
e147a74bbb
Chip needs much more time to think things over (up to 2 seconds to complete
...
a `recalibrate' command if no flop present).
1996-04-29 12:07:32 +00:00
dc5ff52131
dk_establish() returns void.
1996-04-28 20:25:59 +00:00
646acc51d0
Map the pixels explicitly on the VME32 bus.
1996-04-23 19:34:05 +00:00
ee8f3dc1b8
Turn delay() into a calibrated loop. It's more accurate and compiles into
...
a `leaf' routine, avoiding unexpected large delays by overflowing the
register windows.
1996-04-23 19:25:25 +00:00
6b768c677a
Now that we have lehwinit() for the Sun4m, make sure that we wait 20ms
...
after changing the cable type, as specified in the chip documentation.
Also, sanity-check that sc_dma is valid in case a Sun4m ever exists without
a ledma.
1996-04-22 03:57:29 +00:00
90bc31d478
remove include of <sys/cpu.h>
1996-04-22 02:50:20 +00:00
a897ec996a
Remove dependency on <sys/cpu.h>
1996-04-22 02:42:00 +00:00
bebf990150
Oops the name of the UTP/AUI bit changed under me...
1996-04-22 02:39:49 +00:00
5253b892e9
Add a hardware dependent initialization function lehwinit()
...
to support selecting UTP media on sun4m
1996-04-22 02:37:10 +00:00
709298a10a
- Changed Sun4m DMA chip support to detect and use maximum available SBus
...
burst size when transferring data.
- Changed ledma attach code to pay attention to the PROM's notion of what
cable type is being used. Note that this patch does not fix the problem
recently discussed on port-sparc; in most cases the PROM doesn't know
what cable type is being used. The default is now TP rather than AUI,
though. A complete fix is forthcoming.
1996-04-22 02:34:53 +00:00
d7d306aeb7
change LANCE copy & zero functions' names to start with amd7990_, and
...
remove their 'integrate' (usually defined to be 'static') keywords.
when lance drivers are split up by attachment, more than one file will
reference the copy/zero functions (i.e. not just the file that pulls in
am7990.c... and eventually inclusion of am7990.c should go away entirely).
1996-04-18 00:25:11 +00:00
a0b7376b16
Add prototype and fix return type of blink() so "options BLINK" works
...
with -Wall -Werror
1996-04-13 17:40:03 +00:00
b3897170bd
Haven't used "options RCONSOLE" for a long time. Update to the new
...
"options RASTERCONSOLE". Noticed by cgd.
1996-04-11 21:54:13 +00:00
e17cddf913
If the power device is configured, don't forget to compile the driver.
...
Also, add some whitespace for readability.
1996-04-11 21:05:42 +00:00
41a08fd254
Include fdc.h, not fd.h, due to change in files.sparc.
1996-04-11 19:20:03 +00:00
c0919485d4
Move some of the SCSI target juggling to autoconf's fake_bootpath() routine.
...
Fix incorrect interpretation of bootpath's val[2] as a SCSI unit.
1996-04-10 20:48:35 +00:00
e4394459a1
`bootdv' is gone. Instead, set the device pointer in the `struct bootpath'
...
that is passed in the attach arguments.
1996-04-10 20:46:05 +00:00
c28fcc3539
Use extra value in `struct bootpath' to describe the PROM's bootpath better.
...
Use the bootpath[] array in setroot() to determine partition info in case
of boot devices of type DV_DISK. Also, precook more SCSI device info in
fake_bootpath() to simplify dk_establish().
1996-04-10 20:40:14 +00:00
5f2d466997
Make room for another `value' in struct bootpath. Drop the global `bootdv'
...
in favour of storing the device pointer in the bootpath[] element corresponding
to the boot device. This also allows the opportunity to get to "intermediate"
devices (e.g. busses) should the need for this ever arise.
1996-04-10 20:33:38 +00:00
4c26e95dde
Initialize relevant elements of the `romaux' structure in romprop().
1996-04-09 15:24:00 +00:00
1a1a8ba8c9
Undo unintended change in previous commit: _mcount is static.
1996-04-08 20:55:36 +00:00
940ccbc5df
Fix typo that caused sun4 machines to be misidentified as sun4c-class.
...
Fixes reported problem of binstall not stripping the a.out header from
the boot block on sun4 machines.
1996-04-07 20:00:12 +00:00
77366064b8
Fix oversight in sun4m changes: the CPU attach code moved, so the CPU
...
was never getting attached on a sun4 machine. Fix this so that the
cache and FPU get enabled on sun4 machines. (Believe me, a 4/260 is
slot enough, *with* the cache...)
1996-04-07 06:02:20 +00:00
56fed0c102
Fix slight oversight in previous change.
1996-04-07 05:50:59 +00:00
d68e57ee64
Make this compile again on a SUN4-only kernel. (Avoid "unused foo"
...
warnings from gcc -Wall. While the optimizer is smart enough to
eliminate dead code, it looks like the rest of gcc doesn't Get It.)
1996-04-07 04:48:46 +00:00
e50fc72424
make netbsd/sparc boot on sun4 once again. the sun4m changes introduced
...
config lines for obio0 devices without addresses (thus cf->cf_loc[0] took
the default value of -1). we now ignore these entries on a sun4 since
they could never be valid (eliminates a 'panic:alignment' fault at bootup).
1996-04-05 21:50:05 +00:00
a7c600676f
make netbsd/sparc boot on sun4 once again. the sun4m changes introduced
...
config lines for obio0 devices without addresses (thus cf->cf_loc[0] took
the default value of -1). this caused autoconf.c to map memreg0 to
obio physical address -1 (invalid). as soon as we took a memory fault
the code in memfault_sun4 [locore] would get a Watchdog timeout when trying to
read memreg0 (a.k.a. par_err_reg).
1996-04-05 21:44:25 +00:00
8eeab3102b
Probe clock registers on sun4s to make sure we attach the correct device
...
(sun4m has clocks at obio too).
1996-04-04 23:55:47 +00:00
8e427a346f
Check for frame pointers we can handle (i.e. in kernel space). This prevents
...
data faults when tracing proc0's context.
1996-04-04 23:25:35 +00:00
1372d856d7
cpu.c: correctly identify Sun4m associative caches
...
kgdb_stub.c: make it compile again with Sun4m-only kernel
1996-04-04 23:06:32 +00:00
1068aa5751
Add back support for cached DVMA on SuperSPARC (sun4m) machines that was lost
...
during merge
1996-04-04 23:05:22 +00:00
b0fd2dabcb
Include fd.c if (fdc | fd), so we're not obliged to define a device
...
at fdc.
1996-04-04 22:57:55 +00:00
83f9bdc359
update for the fact that config_found() and config_rootfound() now
...
return pointers. (Check vs. NULL, rather than just boolean tests.)
1996-04-04 06:25:00 +00:00
7bc2baeb01
update for the fact that 'alldevs' is now a TAILQ.
1996-04-04 00:27:24 +00:00
7311f0da23
Late-breaking patch from Aaron.
1996-04-01 21:09:39 +00:00
10ffa5da40
Fix warnings when SUN4M is not defined.
1996-04-01 19:16:41 +00:00
3dbbc8d3e9
Fix another printf format argument.
1996-04-01 17:37:08 +00:00
pk