| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
|
|
(i.e. IP27 and IP30 sgi kernels).
|
|
be sure to rerun config(8) on your kernel configuration files.
|
|
|
|
|
|
on LOMLite2.
|
|
the Yeelong Lemote mips-based netbook. Kernel bits only for now, needs
polishing; most of this work done during h2k9 last month, although the
porting effort started earlier this year.
|
|
the Yeelong Lemote mips-based netbook. Kernel bits only for now, needs
polishing; most of this work done during h2k9 last month, although the
porting effort started earlier this year.
|
|
the Yeelong Lemote mips-based netbook. Kernel bits only for now, needs
polishing; most of this work done during h2k9 last month, although the
porting effort started earlier this year.
|
|
them; allows userland with stack beyond 2GB to run.
|
|
|
|
|
|
a vDisk server. This makes OpenBSD running in a guest domain survive
a reboot of a control domain or service domain.
|
|
|
|
|
|
the Yeelong Lemote mips-based netbook. Kernel bits only for now, needs
polishing; most of this work done during h2k9 last month, although the
porting effort started earlier this year.
|
|
the Yeelong Lemote mips-based netbook. Kernel bits only for now, needs
polishing; most of this work done during h2k9 last month, although the
porting effort started earlier this year.
|
|
spotted by ckuethe and must be in immediately before my build gets there..
|
|
disk server.
|
|
Fixes most laptops out there on resume. Okay deraadt@.
|
|
code):
pmap_flush_cache(vaddr_t, vsize_t) and pmap_flush_page(paddr_t) to flush
the cache for virtual addresses and physical pages respectively using
the clflush instruction. These apis will shortly be used by the agp
bus_dma functions to avoid doing a wbinvd on each dmamap_sync.
ok kettenis@, some comments from miod@
|
|
instruction.
ok kettenis@ as part of a larger diff.
|
|
the cpu.
ok kettenis@ as part of a larger diff.
|
|
context.S fixes allows these settings to work for kernels linked in CKSEG0.
|
|
wired TLB entries).
|
|
the current 8KB.
|
|
unmaps quite a bit more than intended, which causes weird things like
watchdog resets.
|
|
used by the virtual disk server.
|
|
boundaries, entirely contained within a kernel page; issue raised by drahn@
|
|
|
|
rather than <mips64/param.h>.
For now, kernels are kept at 4KB to give people some time to build 16KB
compatible binaries; this will change before the end of this release cycle.
Use of 16KB page size kernels yields a 18% speedup (which, offset by the
1.6% slowdown caused by the pmap changes, yields a 16.6% overall speedup).
|
|
the u area after each context switch, simply convert it to an XKPHYS address,
so that access to it never faults. Idea from NetBSD.
|
|
directly allocating pages from uvm; this will allow us to eventually use
a different kernel page size without having to alter the pmap structures
layout.
No functional change; measured slowdown of 1.6% for 4KB page kernels.
|
|
page size, be sure to pick the strictest interval so as not to incorrectly
claim ARCBios reserved data not (kernel) page size aligned as free memory.
No functional change since the kernel uses the same page size as ARCBios
(at the moment).
|
|
to consider the probe as failed; tlb pair #0 does not deserve to be magic
in any way.
|
|
honoured, and we may end up loaded at an uncached address (e.g. CKSEG1 on
O2). This would cause the kernel to run with instruction cache disabled.
Work around this by simply jumping to the address we really want to run at,
at the beginning of locore.
|
|
disable interrupts and unconditionnaly use proc0 u area as the stack, so that
once ARCBios loses the upper 32 bits of the stack pointer, accessing the
stack does not fault (proc0 u area is allocated in CKSEG0 for this reason).
|
|
request to other bricks connected to our node.
|
|
|
|
resulting in kernels starting up the other cpu's incorrectly on processors
with very long pipelines, perhaps some bug instruction prefetch dependency
in the mptramp.
|
|
control bit to enabled use of the extra 16, in order to be able to be used
as drop-in R5000 replacement without modifying RM7000-unaware software.
Because of this, when a 64 TLB RM7000 processor is detected, check that
the upper 16 really will be used before deciding how many TLB the kernel
needs to invalidate whe ASID wraps.
|
|
|
|
ok miod@
|
|
ok miod@
|
|
lot of room for improvement in the driver, but atleast it works now.
OK kettenis@, ratchov@
|
|
the the msr paths are sane.
ok deraadt@
|
|
so don't try to do EST on machines we don't know the
bus_clock on for now.
ok deraadt@
|
|
The firmware is loaded from /etc/firmware/afb, which will be provided soon.
|
