needs some testing, but it seems to produce sound. The driver was written
by me, but since I don't have the hardware the debugging and testing was
done by Andreas Gustafsson <gson@araneus.fi>, Chuck Cranor
<chuck@maria.wustl.edu>, and Phil Nelson <phil@cs.wwu.edu>. Thanks.
a change marked XXX, where gcc -Wall pointed out a bit of the original
code that looked wrong. The jury's still out on whether or not it was
really wrong, but it looked _very_ suspicious.)
Reference Design NetBSD source code, obtained from the pages under
http://www.research.digital.com/SRC/iag . Some of this code (badly)
needs to be cleaned up, and as-is it doesn't compile. However, getting
it in the tree is a start.
or not rather than orring in a USER bit into the fault code variable.
This simplifies the fault switch statement and some of the addition
checks that are made of the fault code.
Move the printing of data abort information out into a separate function
rather then duplicating a lot of similar printf() statements.
Removed the fault_code variables from the prefetch abort handler as it
is not actually used for anything.
flow, by setting the "can fast forward" flag in the packet header, and
giving a chance for filters to clear the flag. If the flag is still
set after the filters have given it a chance, the packet will be used
to create a fast-forward flow entry.
later incremented correctedly, causing the wrong data to be paged out, which
then caused general lossage later when the data was paged in and the process
tried to use it. found by pk.
Page tables no longer occupy space in the processes vm_map. This
means that page tables can now be allocated from pmap_enter() when the
first entry is made in the page table rather than relying on taking
a section fault and faulting in a page table. This also lowers
the overheads in creating and deleting page tables and simplfies
cpu_fork().
A result of this change is that problems with mapping zero page and
an associated pagetable when execing from a shared vm_map disappear
fixing PR5178.
the burst size allowed, but rather a fixed number of packets, as
described in the Internet Draft. Default allowed burst is 4 packets,
per the Draft.
Make the use of CWM and the allowed burst size tunable via sysctl.
the "header already complete" flag. This allows BPF writers to spoof
layer 2 source addresses (providing the layer 2 in use supports it) in
applications where this is necessary. From Greg Smith <greg@nas.nasa.gov>.
use the Ethernet source address speficied in the sockaddr rather than
the interface's Ethernet address, and then fall through to the AF_UNSPEC
case. From Greg Smith <greg@nas.nasa.gov>.
if_fddisubr.c to fastpath IP forwarding. If ip_forward successfully
forwards a packet, it will create a cache (ipflow) entry. ether_input
and fddi_input will first call ipflow_fastforward with the received
packet and if the packet passes enough tests, it will be forwarded (the
ttl is decremented and the cksum is adjusted incrementally).
conditional (tcp_compat_42). The kernel config option TCP_COMPAT_42
will still enable this by default, or disable this by default if the
option is not included (i.e. current behavior). This will be made a
sysctl soon.
Cute buglet: you can end up with zero CCBs if there were no targets
seen by the adapter. Always leave a minimum so the adapter can
finish attaching- it may be there w/o targets for a reason.
rule was to update the timestamp if the sequence numbers are in range. New
rule adds a check that the timestamp is advancing, thus preventing our notion
of the most recent timestamp from incorrectly moving backwards.
function is meant to glue the new pckbd driver to the old console code
on ports which use the old PC-style console code. This will eventually
be purged when all of the old legacy consoles have gone away.
all of the fragments. Use the mtu of route in preference of the MTU of the
interface when doing fragmentation decisions. (ie. Fragment to the path
mtu if it is available).
function is intended to process any deferred pmap operations, and is not
meant as a general means of flushing the TLB. Since this pmap module does
not defer any operations, make pmap_update() do nothing.
caller to pass an optional 3rd argument, which is the previous level
PTE corresponding the virtual address. If this argument is non-NULL,
the table walk otherwise necessary will be bypassed.
the capacity based on the c/h/s numbers. In fact, don't use the c/h/s
numbers for much of anything.
For ATA-4 drives or later, always use LBA mode, since it's now required.
Collectively, this allows >8GB disks (like the 12GB Bigfoot) to work.
handlers. (Only slot_ignore() and slot_noint() need this, and we already
have a place to put this information.) Adjust add_nubus_intr() so that if
the client_data arg is specified as NULL, pass the slot number as
client_data to the interrupt handler.
MACHINE_NEW_NONCONTIG interface implementation, which is now enabled
on all systems. Thanks to Jason Thorpe for his work on the hp300 port,
from which most of the code used here was derived.
XXX In spite of the fact that it works on the hp300, UVM does not (yet)
work well under load on mac68k.
adjoining regions (except for the segment starting at the address
we loaded the kernel). This simplifies cpu_init_kcore_hdr()
considerably, but the real win will be with MACHINE_NEW_NONCONTIG.
Compute the disk block addr at command queing time rather than exec time.
This fix a bug which could lead to data corruption on disk: when a command
was reexecuted after an error condition (from wdcunwedge), the partition
offset was re-added to the block addr, leading to a transfert at the wrong
disk block.
This should fix the problem reported by some laptop's users, where the
first disk read after a suspend/resume returned garbage.
prototype declearation for method to override wbflush() callback
vector with model-specific code. Used on DEC r2000a machines with
writebuffers which indicate writebuffer drain via cp0 usability bit.
