cached value will do, or if the very latest total must be fetched. It can
be called thousands of times a second and fetching the totals impacts not
only the calling LWP but other CPUs doing unrelated activity in the VM
system.
Background:
- All m68k ports have fixed PAGE_SIZE value in their kernels,
but each port uses different PAGE_SIZE value (4096 or 8192)
due to historical reasons.
- Currently module(7) binaries are built per each port so
all m68k kernel sources don't support run-time variable PAGE_SIZE.
- MI <uvm/uvm_param.h> assumes that the port supports a variable
PAGE_SIZE on module(7) builds if both MAX_PAGE_SIZE and MIN_PAGE_SIZE
are defined and they have different values.
- On the other hand, jemalloc(3) checks MAX_PAGE_SHIFT in
src/external/bsd/jemalloc/include/jemalloc/internal/jemalloc_internal_defs.h
for internal optimization.
- m68k ports share userland binaries (especially pkgsrc binaries)
among all ports, so we need to define MAX_PAGE_SHIFT as 13 to
support m68k ports where PAGE_SIZE==8192.
(though this would affect only if static binaries built on
4k page hosts are executed on 8k page hosts)
To solve these inconsistency on PAGE_SIZE definitions,
we should have an independent PAGE_SIZE related definitions
for userland, but it requires major reorganization.
For now (especially for netbsd-9) we define MAX/MIN PAGE_SIZE and
PAGE_SHIFT values in <m68k/vmparam.h> only in !defined(_KERNEL) case.
Discussed on source-changes-d@ and tech-kern@ with christos@ and thorpej@:
https://mail-index.netbsd.org/source-changes-d/2020/01/thread1.html#012035https://mail-index.netbsd.org/tech-kern/2020/01/thread1.html#025954
Should be pulled up to netbsd-9.
- Remove all code that should be MI, leaving the bare minimum under arch/.
- Make the required actions very explicit.
- Pass in LWP pointer for convenience.
- When a trap is required on another CPU, have the IPI set it locally.
- Expunge cpu_did_resched().
sys/net/if_ethersubr.c if we can.
- Add ec_ifmedia into struct ethercom.
- ec_mii in struct ethercom is kept and used as it is. It might be used in
future. Note that some Ethernet drivers which _DOESN'T_ use mii(4) use
ec_mii for keeping the if_media. Those should be changed in future.
On ports without a GENERIC kernel config enable in individul files, e.g evbmips.
Omit on:
atari, dreamcast, emips, epoc32, evbppc/VIRTEX*, ia64, luna68x, mvme68k,
mvmeppc, playstation2, riscv, sun2, sun3, x68k, zaurus due to resource
constraints or port infancy.
userspace. The old fetch(9) and store(9) APIs (fubyte(), fuword(),
subyte(), suword(), etc.) are retired and replaced with new ufetch(9)
and ustore(9) APIs that can return proper error codes, etc. and are
implemented consistently across all platforms. The interrupt-safe
variants are no longer supported (and several of the existing attempts
at fuswintr(), etc. were buggy and not actually interrupt-safe).
Also augmement the ucas(9) API, making it consistently available on
all plaforms, supporting uniprocessor and multiprocessor systems, even
those that do not have CAS or LL/SC primitives.
Welcome to NetBSD 8.99.37.
shared copy: m68040_writeback(). It is essentially a copy of the Atari
version, with some minor cosmetic tweaks and one small performance optimization
from the mvme68k port.
Tested by rjs@ on a Quadra 950. (Thanks!)
- Every driver that was removed and whose number hasn't already been
reused is now listed with a commented-out "obsolete" line.
- The format of these has been systematized. Future format changes can
probably be safely done with a script.
- This does not include a few cases of assignments that only lasted a
couple days, or stuff from before major reorgs. Some of these may
be included nonetheless, because there was a lot of ground to cover
and therefore not a lot of time to dig into history in detail.
Note that the obsolete listings do not mean the major numbers can
never be reused; that's up to portmasters and/or core. It does mean
that they won't be reused by accident, however, which in some cases
(depending on the driver, how widely used it was, its family of device
nodes, their default permissions, etc.) can be quite dangerous.
Note that some of the things now explicitly listed as obsolete are
really ancient history. My scan went back as far as when the majors
files were added. (But not before that.)
set or not, in the same way as libcompat.
- Specify OPT_MODULAR in the port Makefile instead of KERN_AS.
Now, KERN_AS=library is used for kernels without module(7) for all ports.
OK christos
This change:
* Removes "options PERFCTRS", the associated includes, and the associated
ifdefs. In doing so, it removes several XXXSMPs in the MI code, which is
good.
* Removes the PMC code of ARM XSCALE.
* Removes all the pmc.h files. They were all empty, except for ARM XSCALE.
* Reorders the x86 PMC code not to rely on the legacy pmc.h file. The
definitions are put in sysarch.h.
* Removes the kern/sys_pmc.c file, and along with it, the sys_pmc_control
and sys_pmc_get_info syscalls. They are marked as OBSOL in kern,
netbsd32 and rump.
* Removes the pmc_evid_t and pmc_ctr_t types.
* Removes all the associated man pages. The sets are marked as obsolete.
compatibility with BIOC[GS]SEESENT ioctl. The userland interface is the same
as FreeBSD.
This change also fixes a bug that the direction is misunderstand on some
environment by passing the direction to bpf_mtap*() instead of checking
m->m_pkthdr.rcvif.
- bad indentation. next68k en.c and mvme68k le_poll.c fixes real issues
in error handling, the rest are NFCI.
- pass 68030 flags as appropriate for mvme68k.
- next68k nextrom.c has -Warray-bounds ignored for an odd expression
that appears to run before relocation, and needs manual offsets
added which trips bounds array checking.
with this all m68k ports build with GCC 6.
We now have 2 variables automatically set in elf_machdep.h:
ARCH_ELFSIZE: the size for userland binaries
KERN_ELFSIZE: the size for the kernel binaries
DB_ELFSIZE has been deleted and KERN_ELFSIZE should have always the
same values DB_ELFSIZE used to have.
In sys/exec_elf.h, if ELFSIZE is not set, it is set to KERN_ELFSIZE
for the kernel and ARCH_ELFSIZE for userland. These defaults should
eliminate the need for most manual ELFSIZE setting.
- new series of files that are useful for saying "i want everything
since netbsd 1.4", etc.
- use the fact COMPAT_* options have future dependancies to remove
many redundant options.
removes about 3000 lines total across kernel configuration files.
tested about 30 random kernels in the changed list.
