touched any user-space address recently. This is efficient
for things that stay in the kernel for a while, waking up
to handle some I/O then going back to sleep (i.e. nfsd).
If and when such a process returns to user-mode, it will
fault and be given a real context at that time.
This also makes context switch faster, because all we need
to do there for the MMU is slam the context register.
since today, they will have the same size as the on-the-wire-packet on each
architecture.
Problem was reported by George Harvey for the m68k architecture.
* Make the ring buffer size and water marks patchable, and allocate the buffer
separately.
* Do the ttymalloc() at attach time.
* Reorganize the receive buffer so the status and data pair are next to each
other. This is slightly faster.
* Make sure we actually do turn off interrupts in comclose() if we have DDB
configured and it's not the console. (D'oh!!!!)
* When we exhaust the current transmit run, turn off transmit interrupts in
comintr(), so we're fairly sure we don't get another one.
* Nuke the silly lsrmap[] idea; it's slower in the normal case.
* Cache the l_rint pointer in the soft interrupt routine.
* Carrier detect (TS_CARR_ON) is based on the actual DCD bit, even if it's
being ignored.
* Set TS_WOPEN early on in zsopen().
* Don't disable interrupts on the console during close if we have DDB.
Inert changes:
* Don't handle ZS_HWFLAG_NO_DCD here; the frontend does it.
* Deprecate `register'.
* Use SET(), CLR(), and ISSET().
More performance changes:
* Rototill receive handling; use a backpressure mechanism to prevent livelock.
* Output silo/ibuf overflow warnings at most once per minute, from a callout.
* When we exhaust the current transmit run, turn off transmit interrupts in
zstty_txint(), so we're fairly sure we don't get another one.
* Make the ring buffer size and water marks patchable, and allocate the buffer
separately.
* Do the ttymalloc() at attach time.
* Reorganize the receive buffer so the status and data pair are next to each
other. This is slightly faster.
* Make sure we actually do turn off interrupts in comclose() if we have DDB
configured and it's not the console. (D'oh!!!!)
* When we exhaust the current transmit run, turn off transmit interrupts in
comintr(), so we're fairly sure we don't get another one.
* Nuke the silly lsrmap[] idea; it's slower in the normal case.
* Cache the l_rint pointer in the soft interrupt routine.
--entries / remove kernel entry from locore / reorganize vector entry
--code. Enables access to stack frames transitively w.r.t. palcode
--vectors, e.g. upward traceback works, inverting (possibly several)
--kernel vectors. Until now, all trackbacks ended at the first-reached
--instance of trap(), which was totally useless as there is no mystery
--to trap->panic->cpu_reboot
destination register bit pattern with 1.0), which automatically provides
corner case handling.
Missing ftwotox emulation originally reported by Norman Mackenzie in PR 4237,
but he proposed a different implementation.
- Map in all physical memory first in system space. This reduces
pmap_copy_page() and pmap_zero_page to 3 resp. 2 instructions.
- Have fized position user page tables. Makes the pv_table smaller
and all reverse references faster (and simpler).
- Remove the wiring code. Nobody doesn't even know what a DR32 is anymore.
- Simulate page reference bit by setting page invalid, as suggested by
Rich Draves in a paper for 1991 Mach Usenix Symposium.
This reduced the time spent in the pmap module to between 70-75% of
the previous; and made process startup _much_ faster.
- Map in all physical memory first in system space. This reduces
pmap_copy_page() and pmap_zero_page to 3 resp. 2 instructions.
- Have fized position user page tables. Makes the pv_table smaller
and all reverse references faster (and simpler).
- Remove the wiring code. Nobody doesn't even know what a DR32 is anymore.
- Simulate page reference bit by setting page invalid, as suggested by
Rich Draves in a paper for 1991 Mach Usenix Symposium.
This reduced the time spent in the pmap module to between 70-75% of
the previous; and made process startup _much_ faster.
- Map in all physical memory first in system space. This reduces
pmap_copy_page() and pmap_zero_page to 3 resp. 2 instructions.
- Have fized position user page tables. Makes the pv_table smaller
and all reverse references faster (and simpler).
- Remove the wiring code. Nobody doesn't even know what a DR32 is anymore.
- Simulate page reference bit by setting page invalid, as suggested by
Rich Draves in a paper for 1991 Mach Usenix Symposium.
This reduced the time spent in the pmap module to between 70-75% of
the previous; and made process startup _much_ faster.
Supports changing media with ifconfig's "media" directive.
This splits the 3c503 functionality out of the ISA "ed" driver, much the
same way NE2000 support was split out.
control any more, and the speed of changing DTR isn't really an issue. Also,
the old code created a bug where zsparam() might have failed to set some
registers.
Change the interface to zs_hwiflow(); just have it use zst_rx_blocked.
Make zs_modem() a tiny bit faster.
Do RTS updates immediately; do not allow them to be held.
always located at the end of onboard RAM. This allows locore.s to zero
any offboard RAM to initialise the parity bit which most VMEbus RAM cards
have. Without this, many cards buserr on the first read access.
Thanks to Herb Peyerl for the idea.
Stop output as fast as possible when DCD is deasserted.
Do *not* automatically drop DTR when DCD is deasserted.
Only check for rr0 bits that we care about (DTR and DCD).
Make sure we turn on DTR and/or RTS as appropriate during open.
Rearrange close sequence so the tty is flushed before turning off interrupts.
Deal with `softcar' and the console device by silently asserting CLOCAL and
|HUPCL, as in com driver.
Do *not* fiddle with DTR when changing the line speed.
Make sure we update the tty's carrier status when CLOCAL or MDMBUF is changed.
Only change rr1 when we actually need to turn on or off TIE.
From /sys/news3400/news3400/locore.s, with id
@(#)locore.s 8.3 (Berkeley) 9/23/93
Kazumasa Utashiro notes that the pmax cacheflush routines don't work:
#ifndef NOTDEF /* I don't know why Ralph's code doesn't work. KU:XXX */
It's because pmax hardware wries the COP0 bit to external branch
logic. news3400s don't, and so the bc0f loop fails. It will also
fail on some other models of pmax, but we dont' support them.
Surround the relevant framgents in locore_r200.S with "#ifdef pmax".
Longer-term, the cacheflush entry in the locore callback may have
to be a CPU baseboard-specific entry, not just CPU-version specific.
intr_alloc_mask) into one place, comment them, and defopt them.
- Rename pcic_intr_alloc_mask to pcic_isa_intr_alloc_mask, since it's
an ISA-specific thing.
- When allocating/establishing the PCIC's interrupt (for card events),
do error checking, and pay attention to the intr_alloc_mask.
for the built-in SCSI on NEC Versa docking stations, and if a card
allocates that IRQ, it will never get interrupts. This caused the
default kernel to not work on these laptops, as IRQ 10 was often the
first free IRQ.
things happen if we are the console.
Restore to the prevoius value (not to hardwired 8N1) because this
could be set by the serial console initialization.
Closed PR kern/4373 (Dave Huang)
in the ISR (from the ISA "ed" driver's generic 8390 probe routine).
Also, add a big comment describing exactly what we're checking when
probing for the dp8390 (also lifted from the ISA "ed" driver).
The RST check should help weed-out non-matches before they get to the
fairly invasive memory check.
should creation/initialization of a ccb fail. Also, don't panic in
these situations. Instead, simply return an error condition and allow
the caller to deal with it. It may be the case that we were able to
create one ccb what we can grab, and continue to hobble along.
Also, fix the scb freelist fencepost, like the aha/bha/uha drivers.
should creation/initialization of a ecb fail. Also, don't panic in
these situations. Instead, simply return an error condition and allow
the caller to deal with it. It may be the case that we were able to
create one ecb what we can grab, and continue to hobble along.
should creation/initialization of a mscp fail. Also, don't panic in
these situations. Instead, simply return an error condition and allow
the caller to deal with it. It may be the case that we were able to
create one mscp what we can grab, and continue to hobble along.
Also, fix the mscp freelist fencepost error, similar to aha.c and bha.c
should creation/initialization of a ccb fail. Also, don't panic in
these situations. Instead, simply return an error condition and allow
the caller to deal with it. It may be the case that we were able to
create one ccb what we can grab, and continue to hobble along.
shoud creation/initialization of a ccb fail. Also, don't panic in
these situations. Instead, simply return an error condition and allow
the caller to deal with it. It may be the case that we were able to
create one ccb what we can grab, and continue to hobble along.
- add a type field to the isapnp_pin struct and set to IST_EDGE or
IST_LEVEL depending on the card configuration data.
- use the irq type field in all the drivers instead of hardcoding it.
Adjust the number of ccbs allocated during initialization.
Every ccb could potentially lock 64k of memory for dma buffers if bounce
buffers are used. Instead of BHA_CCB_MAX ccbs using 2MByte, only
sc_link.openings ccbs per device are allocated. Thus we now use only
256KByte per device present.
(hannken)
Fix a couple of errors in bha_create_ccbs():
- Don't ever let the number of CCBs be creater than BHA_CCB_MAX.
- Fix a fencepost that caused the last CCB allocated to never be put
on the freelist.
(thorpej)
change KEYSYM_HOLE to KEYSYM_NOP most places in those tables.
Map the `R' function keys with KEYSYM_FUNC_R instead of the
arrow key strings so that numlock indirection works.
- at end of attach, explicitely select an existing drive. This fixes hangs
some users reported (such as the one reported in port-i386/4247).
- Some atapi cdrom drives (e.g. Nec 24x) don't enables their registers before
a controller reset is issued. The controller probe routine is changed as
follow: issue a controller reset. If fail, test atapi signature on slave.
If fail, wait 5s and retry a reset. If the second reset fail, return(0).
If the first reset succeed, test presence of a master drive:
atapi signature, and if this fail RO/RW registers test. If no master, test
atapi signature on slave. If no slave, return 0.
pread64(), pwrite(), pwrite64():
* The `offset' argument specifies an absolute offset within the file.
* After performing the read resp. write operation, restore the original
offset within the file.
Also, change the argument structure used by svr4_sys_prwite64() to
svr4_sys_pwrite64_args, instead of the 32-bit offset version.
"A sequence step of 0 after a select with ATN can be a selection
timeout, or it can also indicate the target did not respond with
a message out phase. The latter will occur on very old SCSI
devices which do not respond to the ATN signal and go directly to
the command phase".
* Check for a disklabel matching the known values in an install diskimage.
If found, update incore disklabel's RAW_PART with the size reported
by the disk, clobbering the size used by vnd(4).*
* If geometry info is bogus or /missing, supply a fake geometry
(as in sd.c). Saves readdisklabel() and sysinst from divide-by-zero errors.
* lint: RAWPART -> RAW_PART.
overflow (always hated that).
replaced "/m" flag with:
/a == show process address info
/n == show normal process info [currently the default]
/w == show process wait/emul info
floating point stuff removed].
the new kprintf replaces the 3 different (and buggy) versions of
printf that were in the kernel before (kprintf, sprintf, and db_printf),
thus reducing duplicated code by 2/3's. this fixes (or adds) several
printf formats. examples:
%#x - previously only supported by db_printf [not printf/sprintf]
%8.8s - printf would print "000chuck" for "chuck" before
%5p - printf would print "0x 1" for value 1 before
XXX: new kprintf still supports several non-standard '%' formats that
are supposed to eventually be removed:
%: - passes an additional format string and argument list recursively
%b - used to decode error registers
%r - int, but print in radix "db_radix" [DDB only]
%z - 'signed hex' [DDB only]
%n - unsigned int, but print in radix "db_radix" [DDB only]
note that DDB's "%n" conflicts with standard "%n" which takes the
number of characters written so far and stores it into the integer
indicated by the "int *" pointer arg. yuck!
while here, add comments for each function explaining what it is
supposed to do.
In ext2fs, an inode is deleted either when mode == 0 or dtime != 0. If
dtime != 0, reset others fields before using the inode, or we could end
up with the wrong v_op in ext2fs_vinit.
While I'm there, kill a unused variable in ext2fs_readwrite
* Make sure all uses of the data pointer are type-cast appropriately.
* Eliminate the super-user check in the KIOCSKEY case (back as it was).
* Return ENOTTY if the ioctl command code is unrecognized.
to the stat(2) family and msync(2). This uses a primitive function
versioning scheme.
This reverts the libc shared library major version from 13 to 12, and
adds a few new interfaces to bring us to libc version 12.20.
From Frank van der Linden <fvdl@NetBSD.ORG>.
Instead, print a diagnostic and return. (Some drivers do this already.)
Also, normalize the diagnostic message, and fix some places where the
printfs were getting ugly.
Instead, print a diagnostic and return. (Some drivers do this already.)
Also, normalize the diagnostic message, and fix some places where the
printfs were getting ugly.
(_BYTE_ORDER, _BIG_ENDIAN, _LITTLE_ENDIAN).
Define old names from the ANSI ones if not _POSIX_SOURCE.
* Define _QUAD_HIGHWORD and _QUAD_LOWWORD properly when
_BYTE_ORDER == _BIG_ENDIAN.
- Work around a bug in the 82557 that causes the receiver to lock up
in certain conditions by kicking the multicast address filter if we
haven't heard anything come down the wire for some period of time.
- Fix a bug that could cause TxCB descriptor chains to cross page boundaries
on the Alpha.
- Remove some unneeded register masking.
- Fix a bug where too much data was copied from the config template, causing
memory corruption.
- Fix handing of if_timer (it was be cleared too early in some cases).
- Attempt to reduce the chances of receiver overrun by doubling the
number of receive DMA segments, and processing receive interrupts
before transmit interrupts.
- Remove a gratuitous assignment.
- Fix a bug where incoming packets were counted twice.
further. Devices are grouped by types and within the types segregated
by bus. Lots of comments fixed up, lots of nits cleaned. Extensive
comments from Jason and Charles incorporated.
We also really need an automatic way to generate INSTALL from GENERIC,
but that's a project for another day...
the CyberVision64 driver. Modified by Tobias Abt with some bugfixes
from Bernd Ernesti.
ZorroII is at the moment not supported and there is a small problem
with the Console driver where you just get a black screen, but the
system boots and you can use X11 without a problem.
pk