movs.
XXX As our CPP seems to hate the 060SP, I couldn't use assym.h for the
PCB_ONFAULT definition, but had to hardwire 64 in the code. This needs
to be fixed ASAP, and will be done in the upcoming reorganization of
the 060sp Makefiles.
of the reserved KVAs (vpage[]) used by these routines. This avoids the
context switching in there.
* In pmap_redzone(), remove the first kernel address (at KERNBASE) from
the kernel maps instead of just turning off the PTE `valid' bits.
The latter doesn't create a "red zone" at all, but causes the VM to
allocate a new page if the red zone is touched.
uncached (if needed) chunks of memory of the two sizes (256 and 1024
bytes) that the SRMMU ever needs.
* Change pmap_changeprot4m() so that is does not rely setpte4m() because
setpte4m() cannot be used for anything other than VAs that are in the
pmap associated with the current context.
In the same vain, make sure that kvm_uncache() switches to context 0.
Define a constant for the SuperMac Spectrum/24 series III display card.
Thanks go to Luca Falzoni <falzoni@jetai.unipv.it> for trying out the
code for the Dayna ethernet support.
a function) which specifies how many lines will be skipped when scrolling
up when the bottom of the screen is reached. Dumb framebuffers skip 10
lines (as before) because the copies are s ... l ... o ... w, but it's
silly to skip 10 lines on VGA, since the copies are much faster, so we
only skip one in that case.
memory-mapped or i/o-mapped access to the device registers, and always
choose memory-mapped if it is enabled. In particular, do _not_ explictly
disable the space we decide to not use, as this confuses some versions
of Alpha console software (which are arguably buggy because of this
problem).
Also, fix a logic error pointed out by Ross Harvey <ross@teraflop.com>
that would cause memory-mapped access to never be enabled.
unmounting all of the file systems. If we encounter a condition where
all of the dirty buffers could not flush, then don't unmount file systems,
since it might be likely to wedge.
- MIT syntax to Motorola syntax moveml register lists (for better readability)
- comment alignment
- remove commented out debug or alternate code from early DraCo support
development.
(I wanted to do this stuff for a long time, but always had "more important"
things to do.)
with large amounts of memory and not spew. This doesn't really replace
fixing this problem better later, but it works for now. Basically, if memory
is greater than 128MB, start upping the sizes of some maps.
don't buck the trend and move the call to initmsgbuf() to cpu_startup()
in machdep.c (like most other ports do) where we can setup things right
in one go. (This means that the CSRG copyright will no longer be in
the kernel logs).
- Don't overload t_maxseg. Previous behavior was to set it to the min
of the peer's advertised MSS, our advertised MSS, and tcp_mssdflt
(for non-local networks). This breaks PMTU discovery running on
either host. Instead, remember the MSS we advertise, and use it
as appropriate (in silly window avoidance).
- Per last bullet, split tcp_mss() into several functions for handling
MSS (ours and peer's), and performing various tasks when a connection
becomes ESTABLISHED.
- Introduce a new function, tcp_segsize(), which computes the max size
for every segment transmitted in tcp_output(). This will eventually
be used to hook in PMTU discovery.
by bus type / address (for diskless boot)
-add some (experimental, commented out) code which could be used for
identification of boot disks - should help for BIOS remappings
The former "cyloffset" argument is used as an (indirect) pointer to the
information structures. "cyloffset" has always been set to zero in the
past, so this should be save to do.
-add a function ("lookup_bootinfo") to extract data
-use passed console information, if present, for console attachment
-arrange definitions for console and KGDB device a bit nicer
msgbuf. Note that old 'dmesg' and 'syslogd' binaries will continue running,
though old 'dmesg' binaries will output a few bytes of junk at the start of
the buffer, and will miss a few bytes at the end of the buffer.
msgbuf. Note that old 'dmesg' and 'syslogd' binaries will continue running,
though old 'dmesg' binaries will output a few bytes of junk at the start of
the buffer, and will miss a few bytes at the end of the buffer.
NOTE: due to hardware requirement of the AGA chipset, the framebuffer width
has to be a multiple of 64 bits. An 800x600 display will use a memory width
of 832 bits, which the current X server can't handle. To get a usable
display, the width needs to be 832 or 768.
set on viking MXCC machines while using transparent memory access mode. This
becomes apparent when booting on multi-processor viking machines.
While I'm here, arrange to tempt fate a little less by using `AC' mode
for the shortest possible periods of time.
Special thanks to Matt Ragan for running my test kernels.
instead of bp->b_bcount. The I/O operation may have changed bp->b_bcount,
and not all of the kernel virtual memory will be freed. This will result
in the VM system putting pages back on the free list which are still
mapped. This may cause a deadlock later trying to enter pages into
the kernel pmap. [See PR#4120 for more details.]
-optionally, use Martin's direct hardware access to serial lines
-optionally, new console autoselection method: select serial console
if a character is seen within 1 second
-move awaitkey() here, it can be shared by different bootloaders and
it is used by console autoselection
-collect information about console selection in a "bootinfo" defined
structure, for passing to the kernel
separated "nfswrapper.c"/"nfs.c".
This nfs.c differs from the stand/lib/libsa one in that it does not
have a separate mount() function - the mount is done in open().
on EV56 and later processors that have the "amask BWX" bit clear. These
instructions will be used to implement non-swizzle bus access functions
on newer systems, such as the new AlphaStation 500s with EV56 and 21172
PCI chipsets.
See "Alpha Architecture Handbook, Version 3", DEC order number EC-QD2KB-TE.
* support chip clocks != COM_FREQ, by introducing sc_frequency (for the
mainline code) and adding a frequency parameter right after the rate
parameter to comcnattach() and com_kgdb_attach().
- Make com_isa and com_multi initialize sc_frequency to COM_FREQ.
- Make i386/machdep.c and alpha/dec_xxx.c call com*attach() with the freq.
parameter.
* supio_attach_args get two more fields: a sc_ipl and a sc_arg, both ints.
- com_supio uses the first for interupt establishment (all childs will, as
soon as they exist) and the 2nd for sc_frequency.
- drsupio passes sc_ipl alway as 5, and for the "com"s, sc_arg as 16*115200
- hyper will pass sc_ipl as 6, and sc_arg as 16 * 460800
- Fix a few bugs in the software single-stepping support code, where
VMS PALcode ops were being checked, rather than OSF/1 PALcode ops,
causing erroneous results in the "exception return" and "unconditional
branch" predicates.
- Add the BWX instructions ("ldbu", "ldwu", "stb", "stw") to the
"load" and "store" predicates.
easier to add instructions that the disassembler doesn't know about
(the opcode subfunction number is now printed).
- Add the "amask" and "implver" operate subfunctions.
- Add the "ldbu", "ldwu", "stb", and "stw" major opcodes (BWX instructions).
An IMB intruction must be executed after software or I/O devices
write into the instruction stream or modify the instruction
stream virtual address mapping, and before the new value is
fetched as an instruction.
We were missing calls to IMB after mappings were changed, which caused
systems with large I-caches (e.g. my AlphaStation 500) to fail miserably
when mapping in new pages of program text, or when context switching
(I couldn't even get the shell from init!).
reads. This is necessary because of newer AlphaStation firmware doing
the Wrong Thing with target aborts behind PCI-PCI bridges, much like they
do the Wrong Thing with master aborts. Reported by Matthias Drochner.
use the `a_entry' field from the kernel's header as a base.
Pass on a new DDB magic number to the kernel, so kernels linked for a
different KERNBASE value can apply a correction when booted from old
bootblocks.
prevents booting DDB kernels that are built with a different value of KERNBASE.
Apply a correction based on the magic number passed in by the bootblocks.
New bootblocks without the hard-coded KERNBASE reference will be assigned
a new magic number.
reset sequence to complete. Just sending the reset command and rushing
off -- like it used to do -- might cause the keyboard's response to get
lost in case the remaining kernel initialization decides to interact
with the console (RB_ASK!) by calling cngetc(), which goes off polling
within the PROM.
is