on the amount of physical memory and limited by NMBCLUSTERS if present.
Architectures without direct mapping also limit it based on the kmem_map
size, which is used as backing store. On i386 and ARM, the maximum KVA
used for mbuf clusters is limited to 64MB by default.
The old default limits and limits based on GATEWAY have been removed.
key_registered_sb_max is hard-wired to a value derived from 2048
clusters.
#if NBPFILTER is no longer required in the client. This change
doesn't yet add support for loading bpf as a module, since drivers
can register before bpf is attached. However, callers of bpf can
now be modularized.
Dynamically loadable bpf could probably be done fairly easily with
coordination from the stub driver and the real driver by registering
attachments in the stub before the real driver is loaded and doing
a handoff. ... and I'm not going to ponder the depths of unload
here.
Tested with i386/MONOLITHIC, modified MONOLITHIC without bpf and rump.
- use SYSMAP_VA, SEGSHIFT for Sysmap address and table entries
- use TIA_SIZE, TIB_SIZE for sizes of table entries
Now page size dependent numbers are almost replaced with proper macro.
Tested on atari, hp300, and news68k.
- move macro SEGSHIFT, NBSEG, and SEGOFSET from <machine/param.h>
to <m68k/pmap_motorola.h> where they should belong
- also simplify NBSEG macro for both 020/030 and 040/060 cases
- also move m68k_trunc_seg(), m68k_round_seg(), and m68k_seg_offset()
(that use SEGOFSET) from <m68k/param.h> to <m68k/pmap_motorola.h>
- add comments about each segment size value used on pmap_motorola
implementation
- add TIA_SIZE, TIA_SHIFT, TIB_SIZE, and TIB_SHIFT macro which can be
used for ste/pte index sizes
- use proper vaddr_t cast instead of unsigned for SEGOFSET/PGOFSET macro
Tested on atari, hp300, and news68k.
They were incorrect since 1997 on amiga and atari, and since 2002
on other ports, but fortunately they don't look so fatal.
Anyway, these values will be moved into <m68k/pmap_motorola.h> soon
since they are quite pmap implementation dependent.
from MD pmap bootstrap sources to common pmap_bootstrap_finalize().
Tested on atari, hp300, mac68k, and news68k.
XXX: Why is protection_codes[] array initialized at run time?
- Replace most remaining uses of l_addr with uvm_lwp_getuarea() or lwp_getpcb().
- Amend assembly in ports where it accesses PCB via struct user.
- Rename L_ADDR to L_PCB in few places. Reduce sys/user.h inclusions.
- Addresses the issue described in PR/38828.
- Some simplification in threading and sleepq subsystems.
- Eliminates pmap_collect() and, as a side note, allows pmap optimisations.
- Eliminates XS_CTL_DATA_ONSTACK in scsipi code.
- Avoids few scans on LWP list and thus potentially long holds of proc_lock.
- Cuts ~1.5k lines of code. Reduces amd64 kernel size by ~4k.
- Removes __SWAP_BROKEN cases.
Tested on x86, mips, acorn32 (thanks <mpumford>) and partly tested on
acorn26 (thanks to <bjh21>).
Discussed on <tech-kern>, reviewed by <ad>.
The head of the list is now a pv_header, which contains the first pv_entry
as well as a 16-bit attributes field (replaces the pmap_attributes array
plus the pv_entry::pv_flags field) as a 16-bit count of caller-specified
cache-inhibited mappings.
Tested on hp300 (shared pmap_motorola.c), changes to atari and amiga copies
are purely mechanical.
using implementation of old src/usr.sbin/mdsetimage sources
which supports misc executable formats without LGPL'ed libbfd.
No particular comments on port-atari@.
XXX1: amiga also has the similar utility in amiga/stand/binpatch
but it has slightly different options.
XXX2: Is it worth to put this utility into MI src/usr.sbin to patch
rtc_offset etc. in GENERIC kernel binaries in distribution?
There are still about 1600 left, but they have ',' or /* ... */
in the actual variable definitions - which my awk script doesn't handle.
There are also many that need () -> (void).
(The script does handle misordered arguments.)
-DM68040, and -DM68060 into sys/arch/m68k/Makefile.cmachflags and
include from both atari and amiga. Other m68k ports which build
kernels that support multiple CPU types can also switch across to
this.
a sane kernel config which defines both of them
- Cleanup all the #if defined(FOO) entries to be consistent and have
a matching /* FOO */ on the #endif line
- No change to generated ATARITT/FALCON/HADES/MILAN-ISAIDE/MILAN-PCIIDE
- Rename BOOT to SMALL030 and trim a little more to save space.
- Drop BOOTX
virtual_end, and protection_codes[] to common <m68k/pmap_motorola.h>,
and also make protection_codes[] unsigned.
XXX: avail_start and avail_end should also be moved, but it causes
XXX: -Wshadow warnings in uvm/uvm_page.c:uvm_page_physload() and
XXX: I don't have a good idea of alternative names for now.
there are two variants of boot.atari per machine types (std or milan)
and they should be selected and copied from /usr/mdec/{std,milan} dirs
to /targetroot during installation.
Problem on default installation with sysinst was reported by David Ross.
Add asm sources which provide a fixed entry point for raw bootxx and bootxxx
because modern aggressive gcc4 may reorder functions in the same source
and the first function in a C source won't always appear at the beginning
in its object.
Also overhaul various files in the stand directory:
- rename stand/Makefile.inc to stand/Makefile.booters and
explicitly include it from each Makefile because the stand
directory contains non standalone programs like installboot and
Makefile.inc will be included implicitly from all Makefiles in SUBDIRs
- put more common options into Makefile.booters so that
all boot programs use proper options (-Os etc.)
- make standalone boot programs compile without installed ${DESTDIR}:
- create machine and m68k symlinks in ${.OBJDIR} in all boot programs
- set appropriate make environments to suppress errors and warnings
- <string.h> isn't there in _STANDALONE case
- put #ifdef TOSTOOLS (looks equivarent with !_STANDALONE)
to some files in stand/tostools
- use ${LD} ${LINKFLAGS} rather than ${CC} ${LDFLAGS}
- create raw binaries on ${PROG} target rather than on beforeinstall
and stop weird renaming on beforeinstall and afterinstall
- print proper error message if bootxx size is larger than limit
- create ${DESTDIR}/usr/mdec/{milan,std} directories on beforeinstall
in all boot programs
(XXX: how can we put MD /usr/mdec/foo directories into src/etc/mtree?)
- prototype warnsfy
- cleanup Makefiles for readablity and remove unnecessary rules
- no need to use daddr_t in installboot because the AHDI label doesn't
support 64 bit block numbers
Tested by David Ross (PR submitter) on port-atari, and
finally NetBSD/atari on TT030 is now fully functional.
Should be pulled up to netbsd-4 (where gcc4 was initially imported)
and netbsd-5.
Rewrite a weird calculation marked "XXX fix calculations XXX"
for an index value of lev2 segment table for the kernel Sysptmap
with more meaningful expressions and PAGE_SIZE independent macro.
Also use "~0" rather than "-1" for an unsigned bitmap value.
Should produce the same results and no functional change.
Tested on Falcon with 68060 by Tuomo Makinen.
via (atari specific) add_sicallback(), as a workaround hack.
Many drivers which use the MD sicallback depend on BASEPRI() macro
defined in <machine/cpu.h> to check nested interrupts,
but functions invoked from MI softint(9) won't run at IPL_SOFT
any longer and the BASEPRI() macro doesn't return expected value
as the past.
Fixes lost interrupt problem on Falcon wdc(4) reported by
Tuomo Makinen on port-atari, and also confirmed by him.
Should be pulled up to netbsd-5.
arch/amiga/amiga/amiga_init.c 1.95-1.97, 1.100-1.102
arch/amiga/amiga/pmap.c 1.123, 1.126-1.131, 1.142, 1.145
arch/amiga/amiga/pmap_bootstrap.c 1.1, 1.2, 1.3, 1.4
arch/amiga/conf/files.amiga 1.137
arch/amiga/include/pcb.h 1.16
arch/amiga/include/pmap.h 1.44
arch/amiga/include/pte.h 1.19
arch/amiga/include/vmparam.h 1.35
- change kernel address space layout (i.e. move Sysmap to the end of KVA)
to match other m68k ports, which should be done on yamt-km merge
- rework pmap_zero_page(), pmap_copy_page() and some more other functions
to make them compatible with m68k/pmap_motorola.c
- move pmap_bootstrap() into a separate file to allow using either
the atari pmap.c or the m68k common pmap_motorola.c
- use common m68k include files, pcb.h, pmap_motorola.h, and pte_motorola.h
- misc style and cosmetic changes
Most merge work is done by Tuomo Makinen, and
miscellaneous fixes and cleanups are done by me.
Tested on Falcon with both 68030 and 68060 by Tuomo, and
also tested on TT030 by David Ross on port-atari.
Should be pulled up to netbsd-4 (it requires at least Sysmap changes)
and netbsd-5.
rather than calling brelse(9) after read.
Tested by Tuomo Makinen on port-atari.
Should fix pool panics during installation reported by David Ross
on port-atari:
http://mail-index.NetBSD.org/port-atari/2008/11/13/msg000113.html
and actually close PR port-atari/39850.
which is now the "API" provided by the pmap module. pmap_kernel()
remains as the syntactic sugar.
Bonus cosmetics round: move all the pmap_t pointer typedefs into
uvm_pmap.h.
Thanks to Greg Oster for providing cpu muscle for doing test builds.
phases, so move the initialization of the ksyms mutex back into main via
a function called ksyms_init. Rename the existing (but quite different)
ksyms_init* variations into ksyms_addsyms_elf() and ksyms_addsyms_explicit()
and adapt machdep code accordingly.
doshutdownhooks(9): shutdown hooks registered by shutdownhook_establish(9)
expect to be called with interrupts disabled, but shutdown hooks
registered with pmf_device_register1(9) expect to be called with
interrupts enabled. So I have made two changes:
1 Do not call pmf_system_shutdown() from doshutdownhooks(). Instead,
change every call to doshutdownhooks() to a call to doshutdownhooks()
followed by a call to pmf_system_shutdown(). No functional change
is intended by this change.
2 Make i386 re-enable interrupts briefly while it calls
pmf_system_shutdown(). I leave it to others either to fix the
other ports, or to factor out some MI shutdown code, as joerg@
suggests, and fix that. Note that a functional change *is* intended
by this change.
I hope that this patch will stop us from flip-flopping between
calling doshutdownhooks() and pmf_system_shutdown() sometimes with
and sometimes without interrupts enabled.
Please commit changes to 'GENERIC.in' and other source files *first*, then
run 'makeconf' and commit the generated files to ensure the 'Created from:'
tags are correct.
Then follow it. Noted by tsutsui@
discussed with tsutsui@ on port-m68k. OK'ed by jdc@.
For atari:
- export ipl2psl_table[] and make it uint16_t
- make makeiplcookie(9) inline
- put PSL_S bit into ipl2psl_table[] rather than adding it in makeiplcookie(9)
- vme/if_le_vme.c: fix a wrong usage of IPL_NET
we no longer need to guard against access from hardware interrupt handlers.
Additionally, if cloning a process with CLONE_SIGHAND, arrange to have the
child process share the parent's lock so that signal state may be kept in
sync. Partially addresses PR kern/37437.
except the ppbus stuff (which doesn't compile) and ulpt(4) which is
unrelated and can be dealt with separately.
As usual, it comes with related cosmetic changes.
- Add a lot of missing selinit() and seldestroy() calls.
- Merge selwakeup() and selnotify() calls into a single selnotify().
- Add an additional 'events' argument to selnotify() call. It will
indicate which event (POLL_IN, POLL_OUT, etc) happen. If unknown,
zero may be used.
Note: please pass appropriate value of 'events' where possible.
Proposed on: <tech-kern>
them in the mi "files" file, and remove include statements from md files.
These shouldn't pull in additional kernel code when not in use, so it
shouldn't do any harm except a risk of namespace collisions which
should be easy to fix.
- Reduce available SPL levels for hardware devices to none, vm, sched, high.
- Acquire kernel_lock only for interrupts at IPL_VM.
- Implement threaded soft interrupts.
SEMMNI, SEMMNS, SEMUME and SHMMAXPGS.
They can be tweaked via sysctl now. Ports that were setting values on
them weren't touched, I only removed the ones that were commented out.
This branch was a major cleanup and rototill of many of the various OEA
cpu based PPC ports that focused on sharing as much code as possible
between the various ports to eliminate near-identical copies of files in
every tree. Additionally there is a new PIC system that unifies the
interface to interrupt code for all different OEA ppc arches. The work
for this branch was done by a variety of people, too long to list here.
TODO:
bebox still needs work to complete the transition to -renovation.
ofppc still needs a bunch of work, which I will be looking at.
ev64260 still needs to be renovated
amigappc was not attempted.
NOTES:
pmppc was removed as an arch, and moved to a evbppc target.
transfer as a write to ensure the memory is writable before starting any
transfer. The fault status information does not reflect this in the 'read'
status bit (i.e. it shows up as a read access), so faults with a RMW access
to non-writable memory was not getting the correct protection. The page would
be read-only and the instruction would fault over and over.
A specific example is when a process forks, and the child process attempts
to execute a RMW access to a data page, which is read-only because it's CoP
Copy-On-Write.
When checking if the page needs to be writablek, also check the locked transfer
and treat any locked transfer as a write.
68060 already handled this correctly, since it has separate read and write
fault bits, and both are set on a RMW access and the trap code was checking
the write status bit.
Fixes PR#36848.
need to understand the locking around that field. Instead of setting
B_ERROR, set b_error instead. b_error is 'owned' by whoever completes
the I/O request.
was optimizing away modifications to the frame contents (it's not nice to
trick gcc). Pass the pointer as the first argument to reduce the number
of places that would be changed otherwise. Fixes the getcwd regression
test on most m68k ports.
from proc_trampoline to match the other ports).
A DIAGNOSTIC kernel will now boot and run. LOCKDEBUG still doesn't work yet.
Also, my amiga no longer loses time.
curcpu()->ci_want_resched is now required, so get rid of the
want_resched global. Also, cpu_need_resched() now takes two arguments.
The l_next and l_back fields in struct lwp no longer exists, so delete
references to it from genassym.cf.
Also, m68k/m68k/proc_subr.s is no longer used, says mhitch, so remove its use.
This brings the atari BOOT kernel back to a buildable state.
an SR value or an IPL_* constant).
- Take advange of the smaller ipl_cookie_t to shrink kmutex_t from
16 bytes to 8 bytes by overlapping storage where possible.
- Implement a RAS-based _lock_cas() for mc68010 systems (Sun2). See
sun68k/sun68k/isr.c.
Tested on various m68k platforms, but NOT Sun2. In any case, at least
Sun2 compiles now.
and use temporary variables on register accesses
- also remove volatile from *memkva member in struct grfabs_et_priv
because it isn't used to refer memory
XXX: I'm not sure if bm->plane should be volatile or not
but I don't think previous code treats it as volatile anyway.
(though I'm not sure how compiler handled "volatile caddr_t")
int _bus_dmatag_subregion(bus_dma_tag_t tag,
bus_addr_t min_addr,
bus_addr_t max_addr,
bus_dma_tag_t *newtag,
int flags)
void _bus_dmatag_destroy(bus_dma_tag_t tag)
that allow a (normally broken/limited) device to restrict the bus address
range it can talk to. this is used by bce(4) to limit DMA addresses to
1GB range, the maximum the chip can address.
all this is from Yorick Hardy <yhardy@uj.ac.za> with input from several
people on tech-kern.
XXX: bus_dma(9) needs an update still.
debugging and its main use is in device drivers. Its used there to signal
that the flagged buffer has a special meaning or should be handled
differently.
OK'd by Bill Sudenmund on tech-kern.
by Slava Semushin <slava.semushin@gmail.com>.
To verify that no nasty side effects of duplicate includes (or their
removal) have an effect here, I've compiled an i386/ALL kernel with
and without the patch, and the only difference in the resulting .o
files was in shifted line numbers in some assert() calls.
The comparison of the .o files was based on the output of "objdump -D".
Thanks to martin@ for the input on testing.
- finish implementing splraiseipl (and makeiplcookie).
http://mail-index.NetBSD.org/tech-kern/2006/07/01/0000.html
- complete workqueue(9) and fix its ipl problem, which is reported
to cause audio skipping.
- fix netbt (at least compilation problems) for some ports.
- fix PR/33218.
version.
Add disk_blocksize(9) so that disk drivers can record the physical
block size of a disk if it is different to DEV_BSIZE. Right now this
simply initialises dk_blkshift and dk_byteshift according to the
supplied block size. This information is used in the MI version of
bounds_check_with_label().
dumpdev. this occurs when we try to set the dumpdev to a device
with no driver loaded. this fixes PR#34872.
in sys_swapctl, if bdevsw_lookup() fails, set dumpdev = NODEV
before calling cpu_dumpconf(). (this also fixes PR#34872.)
XXX: cpu_dumpconf() should probably be changed to take a dumpdev
XXX: and return an error in such cases, but that is a much more
XXX: intrusive change.
XXX2: this is only run-tested on sparc64 and compile tested on a
XXX2: couple of platforms.
Uvm changes over 17 months ago resulted in the 68040/060 segment table
page being entered with pmap_kenter(), which does not record the mapping
in the pv table. Attempting to inhibit caching of that page as required
by the 68060 hardware no longer changes the PTE and caused varying degrees
of multiple faulting, sometimes resulting in an unusable system. Apparently
very few people attempted to run a 68060 based system since that change.
Fix to to change the caching bits directly rather than using pmap_changebit().
roy