Age | Commit message (Collapse) | Author |
|
|
|
"sure" deraadt@
|
|
that .text, .rodata and .data are aligned on BATC boundaries) and merge
all phdrs into the first BCS section.
|
|
dbatc. Add batc mapping support for 88110 systems as well.
|
|
DBATC may span over parts of .rodata.
|
|
data area and the kernel image, whenever possible.
On 88100/88200 systems, use BATC mappings to map the kernel text (and the
kernel data for non-MULTIPROCESSOR kernels). 88110 to follow soon.
|
|
|
|
not get lost if the chip is reset. Also try to mask the `TX start' interrupt,
but this doesn't seem to work.
|
|
probably due to incorrect programming of the system board ``Diagnostic Control
Register'', for which I do not have documentation yet; commited so as not to
lose this work, since the machine I was testing on has apparently commited
suicide and will no longer POST.
|
|
ok miod
|
|
readdoslabel(). Ditto all the MD variables whose addresses are passed
to readdoslabel() via partoffp.
Fix some 512-byte block vs disk sector confusion in hppa and sgi.
All the DL_GETxxxxx() defines return disk sector values. All
DL_SETxxxx() take disk sector values. These changes should be no-ops
until a drive using non-512-byte-sectors is encountered.
ok deraadt@
|
|
|
|
from the network, with symbols. This will be useful to load bsd.rd on systems
with the 3MB bootfile limit in the stupid PROM. Crank version to 0.4.
|
|
the .h file is missing... so remove the #if 0 chunk
|
|
|
|
ok miod
|
|
when it's not welcome.
|
|
the end of the commandline. Clean up a few stale comments. Crank boot blocks
version to 0.3.
|
|
- switch to perl for better data structures and (thus) speed
- fix a couple glitches in the interpretation of the stabs output
- compress the strings by putting them in one big array and overlaying
suffixes
- all sizes and offsets are <64k, so use u_short for them
This results in ~60% reduction in the resulting text size and it now
takes less than a second to create on fast platforms.
ok miod@
|
|
|
|
|
|
400 and 4300. Limited to PIO mode only for now, until the DMA controller is
tamed. Heavily based upon the MI aic(4) driver.
|
|
for older PROM will fail to load BCS binaries with only a .text section,
and will also fail to load an empty .data section.
Tested on AV4300.
|
|
address of a controller only know by its SCM name and number, and the boot
device detection code will now attempt to match the controller first, and
then grandchildren of it (if the boot device is a disk or a tape).
|
|
to the kernel will now match the device the kernel is loaded from, not the
device the boot blocks are loaded from (this allows one to netboot the
boot blocks to boot a kernel from a local disk).
|
|
Needs more work in order to cover all cases, but enough to get internal
disks on the first controller on models 4600 and 530 to get matched.
This makes the aviion port finally self-hosting on aforementioned models,
finally! Next step: bsd.rd...
|
|
|
|
cranking boot blocks version because it's unlikely anyone but me is running
them at the moment.
|
|
|
|
|
|
default kernel name if only a boot device is specified.
This makes boot paths like "sd()" auto-boot without asking for a kernel name.
|
|
SCSI tape support.
Drop sync negotiation from oosiop code, we can run without it.
|
|
Currently limited to oosiop(4) controllers, and thus models 4600 and 530.
There are some rough edges to polish to avoid manual interaction (kernel
currently can't figure its boot device when booted from disk), but the hard
(and hair-pulling) work is over.
And we can have kernel symbols again!
|
|
|
|
set up the delay constant.
|
|
mvme88k, to be able to unmap the page at address zero in the kernel, and have
the vbr page read-only after being initialized.
|
|
`extended interrupt' masks. This is how the hardware works, and this makes the
code simpler than attempting to aggregate them.
No functional change.
|
|
|
|
found, do not attempt to check for any other label scheme: the VDM signature
implies the disk doesn't use MBR or whole-disk partitionining.
This makes sure that, once a boot area has been set up on a disk without any
form a label yet, the bounding constraint it creates will not be cleared by
readdoslabel().
It is now possible to correctly set up and label a vnd image with a boot area.
|
|
a bit worthwhile (yet crude), and give it the ability to initialize the boot
area, instead of having installboot do it - we will need the boot area set up
before attempting to invoke disklabel on the root disk.
|
|
more consistent;
feedback/ok miod
|
|
|
|
partitioning, and an optional boot area) and real VDIT.
If a VDM label is found, but no VDIT is found, search for a native label in
sector 1 (where VDIT would start).
In either case, make sure the OpenBSD area of the disk does not enclose the
boot area.
This allows PROM bootable disks, not shared with DG/UX, to be set up with
minimal hassle. To be done by the upcoming installboot(8) soon.
|
|
100/200/300 would lack it. They are single-processor systems anyway.
|
|
|
|
|
|
to the PROM at halt/reboot time; this lets systems reboot again.
|
|
disk layout, and if a vdmpart instance named "OpenBSD" is found (but not a
vdmaggr!!!), assume this is the OpenBSD part of the disk and the native label
is found within it.
Native DG/UX partitions (vdmpart or vdmaggr) will NOT be reported in the
spoofed label of a shared disk.
|
|
with ddb entered early with boot -d or from UKC).
|
|
|