| Age | Commit message (Collapse) | Author |
|---|
|
from NetBSD.
|
|
512) bytes; ffs is inextricably tied to using b_blkno and disklabel
always uses sectorsize units.
Thus use DEV_BSIZE units for all fields describing ffs filesystems
and convert to/from sectors where required. This enables the creation
and use of ffs filesystems on non-512 byte sectorsize devices.
This diff allows i386 and sgi (the two test platforms) to find
disklabels that are not on a sectorsize boundary. Same change to
further archs coming.
This is a no-op on 512-byte sectorsize devices.
This work triggered by jsing@'s need to create ffs filesystems on
sgi cdroms so we can create cdrom install media for sgi.
sgi testing by jsing@
ok jsing@ pedro@ "looks sane" beck@ weingart@
|
|
|
|
declaring another.
ok miod@
|
|
|
|
problem where the clock would stop ticking on some CPUs because of lost
ticks.
|
|
- Let the VLAN driver know that the hw can send and receive full size
VLAN tagged packets.
Tested by maja@
|
|
on amd64.
Evaluating the _PDC seems to lead to instability (PR5878 and a report by
jmc@) on lower end machines its too close to release and I don't have
enough nickels.
ok deraadt@
|
|
|
|
(using a recursion limit). now OpenBSD can be installed in a
subsequent DOS/Linux-compatible extended partition.
ok krw@ deraadt@
|
|
This will let them get some more testing. Re-order things a little while
i'm here.
prompted by and ok deraadt@.
|
|
looking for the boot device; PR #5895
|
|
ok deraadt@
|
|
prints it. These should be the last offenders.
|
|
|
|
endianness conversion on pci bridges.
ok deraadt@ jsing@
|
|
developer documentation some time ago.
|
|
ok deraadt@
|
|
|
|
|
|
on an attachment dependent flag; the Tadople SPARCLE needs it.
ok kettenis@ deraadt@
|
|
in the MI code.
|
|
|
|
|
|
process_write_regs().
|
|
|
|
userland is allowed to change in psr.
|
|
so don't use any in the 88110-specific parts of locore.
|
|
a kernel 10-15% faster on the t1k. Don't do this for SMALL_KERNEL though,
otherwise the 4MB locked mapping for the data segment will be too small to
include a ramdisk.
|
|
|
|
the motherboard eisa id, instead of using a ``one size fits all'' value which
is too large on more than half the eisa-capable alpha designs.
The id -> slot # logic is based on the alpha ECU configuration files, so we
should not perform worse than ECU itself (and see all slots ECU sees too).
|
|
above the stack. Fix by mapping PIE into the first quarter of the address
space before stack. "will do for now" miod@
|
|
several minutes by making pmap_resident_count() return 0 for the kernel pmap.
discussed with miod@
|
|
we depended on the pageability of the kernel_map to map them which worked
most of the time, until we hit the most improbable case - when cpu_switchto
wanted to fault in the missing gdt page, which at this moment is Verboten.
kettenis@ ok
|
|
provided by the PROM gets flushed out of the TLB before we install our own
trap table. Makes RAMDISK kernels work on T2 systems.
|
|
|
|
problem and lots of register dumps to help the skilled user to fix the hardware.
Not compiled if option SMALL_KERNEL.
From NetBSD.
|
|
sense on m68k. Give USRSTACK its real (HP-UX compatible) value instead of
computing it from HIGHPAGES.
No functional change.
|
|
dino_intr_map(); rather than adding the missing cast, make the intent of
the code clearer by explicitenly testing for PCI_INTERRUPT_LINE being ff.
While there, enable the out-of-extent-range checks in dino_memmap() and
dino_memalloc() even if no option DEBUG, but return failure instead of
panicing.
discussed with and ok kettenis@ marco@
|
|
after it was published. In particular, they changed the maximum cache
aliasing boundary from 1MB to 16MB.
It turns that on the PA-8700 the aliasing boundary is actually 4MB
(reported as such by the firmware at least). There are some comments
in the Linux code that suggest that HP never actually built PA-RISC
CPUs with an 8MB or 16MB aliasing boundary.
So raise the aliasing boundary to 4MB. This fixes the weird ps(1) problem
where it didn't print its own arguments correctly.
|
|
them going negative - this consists of identifying a number of cases of
IO not going through the buffer cache and marking those buffers with
B_RAW - as well as fixing nfs_bio to show pending writes and reads through
the buffer cache via NFS
still has a problem with mishandling the counters I believe in the
async/sync fallback case where counters stay positive which will be
addressed seperately.
ok tedu@ deraadt@
|
|
|
|
bother trying to probe more.
|
|
configures when it can. ok kettenis@
|
|
per-platform implementation, instead of selected members of it; this allows
us to get rid of some globals, and paves the way for better bridge support
on some models.
|
|
uvm_map.c from requiring <machine/exec.h>.
|
|
fp; from NetBSD
ok millert@
|
|
- math.h shouldn't define FLT_EVAL_METHOD, but float.h should (per
C99). remove from math.h, and add proper definitions in float.h
ok millert@
|
|
SUN4V to let it suspend strands (why does everyone invent own words for
hyperthreads?). This gives a huge performance boost when most of the
cpus are idle.
kettenis@ ok
|
|
per-processor soft interrupt register is used for hardware clock-tick
counter interrupts. So make smp_signotify() use IPL_SOFTINT instead
such that we don't unintentionally trigger a spurious clock interrupt.
This seems to fix the timekeeping anomalies on the t1k.
tested by art@
|
