- fix the loaned case in uvm_pagefree().
- redo uvmexp.swpgonly accounting to work with page loaning.
add an assertion before each place we adjust uvmexp.swpgonly.
- fix uvm_km_pgremove() to always free any swap space associated with
the range being removed.
- get rid of UVM_LOAN_WIRED flag. instead, we just make sure that
pages loaned to the kernel are never on the page queues.
this allows us to assert that pages are not loaned and wired
at the same time.
- add yet more assertions.
pages loaned to the kernel. this implies that we also need to
call pmap_kremove() before uvm_km_free().
other general cleanup: remove argument names from prototypes,
rename some variables, etc.
problems when executing programs where text is copied to a page without
a dma sync (like NFS data bcopy'd to a buffer cache page).
From discussion with Jason Thorpe and Eduardo Horvath.
the irq to hardware interrupt mask map instead of using two different
interrelated arrays and a bit mask that I've got wrong more than once
when updating by hand (including when I recently added the keyboard/mouse
interrupt mappings). Now only the galaxy_intr_map[] array needs to
be updated when adding extra interrupts.
o Implement SPDIF selection/monitoring function.
Now, say, playback to both analog output and SPDIF is possible.
o Implement support of AUX input, MIC preamp and MIC recording gain.
o L-R selection of record source doesn't seem to work on cmpci.
Therefore, the *.swap mixer functions are and deleted.
o Following SB mixer registers don't exist on cmpci, and they are removed.
input gain, output gain, AGC, equalization (bass, treble)
o Other mixer changes, including
inputs.XXX.mute -> (deleted)
outputs.XXX.mute -> inputs.XXX.mute
inputs.pc_speaker -> inputs.speaker
spdif.* -> reorganized to spdif.input.*, spdif.output.*
o Current status:
I have tested these and confirmed to work fine.
- Output and recording from Line-in, AUX, CD and MIC analog inputs,
- Output and recording from FM synthesizer,
- Output from PC speaker input,
- Output wave playback.
- SPDIF (44.1kHz) input selection (#1, #2 (6ch version only),
wave to spdin), phase selection, monitoring and recording,
- SPDIF (44.1kHz) playback, through (SPDIF in to SPDIF out)
and monitoring.
I haven't tested these but may work.
- SPDIF 48kHz input and output,
- Full-duplex operation,
- Recording wave output.
I don't think these are working.
- Legacy (wave + FM synthesizer) to SPDIF output (and the monitoring),
- Exchanging front and rear outputs,
- Surround.
These are not implemented.
- 4ch / 6ch support,
- Joystick port support.
- Make the 32-bit SPARC profile support work with the GCC 2.95.3
SPARC ELF compiler, which uses a different name for mcount.
- Make the 64-bit SPARC profile support header look more like the 32-bit
SPARC header (no functional change -- 64-bit SPARC already used the
correct mcount name).
which have the Tekram TRM-S1040 ASIC.
This driver is written by Rui-Xiang Guo <rxg@ms25.url.com.tw>,
and a number of cosmetic changes by me.
Tested on i386 by the author, and on macppc and sparc64 by me.
XXX On arc, kernel got panic in ltsleep() called from scsipi_execute_xs(),
XXX but I'm not sure what is wrong...
file name data and the entry starts with \0\0. Apparently this happens
occasionally, don't know if it's mtools (probably), MS Windows or
NetBSD msdosfs fault. When this happens, NetBSD msdosfs was not
able to open the file, where neither mtools nor MS Windows had any
problem with it. So, it's appropriate to add this fix in any case.
we need later in the code. This fixes a fatal kernel fault in
pmap_modified_emulation if a user application tries to access a kernel
address that is section-mapped.
Add a diagnostic that detects attempts to call pmap_kenter_pa with a
va that is section-mapped.
- Set MARK_START to where we expect to be loading the kernel (0xf0100000).
- The ARM OpenFirmare bindings document describes how the client
program is loaded: OFW allocates and maps 6MB of memory at load-base
(0xf0000000), loads the client program, and then unmaps the memory from
the end of the client program to the end of the allocated region. Then
transfers control to the client program. We must emulate this behavior
to load the kernel: allocate 5MB at 0xf0100000 (where we expect to load
the kernel), load the kernel, then unmap the area after the kernel.
We can now load DHCP and load the kernel via NFS before getting the
dreaded Data Abort.
caching on for a page just because we are clearing the writable bit in
the PTE: this is incompatible with the way pmap_vac_me_harder works,
and the code in the modified emulation handler doesn't know about
recalculating the cachable attributes (nor should it, IMO).
Also, if we are invalidating a page, flush its TLB entry; for some
reason we were only doing this when clearing the Write or modified
bits.
These patches together seem to solve the random seg-faults that were
still occuring occasionally under heavy paging.
a non-_STANDALONE environment (e.g. installboot(8)): internalize and
externalize the exec, program, and section headers as necessary.
Reviewed and OK'd by Christos.
unset on unload), similarily to what i386/machdep.c does #ifdef
COMPAT_SVR4. This makes the svr4 LKM actually work on i386.
XXX kind of ugly, but doing this more generic way would be overkill
restore %fs/%gs appropriately.
Fixes kern/14275 - compat svr4 works on i386 again :)
Thanks to MOCHIDA Shuji for initial investigation on the issue, that helped
to find the bug a lot.
and non-standard inttype-like types, pull in <sys/types.h> if
_KERNEL or _STANDALONE and <inttypes.h> otherwise, and use standard
inttype types.
Discussed with and OK'd by Christos.
the system's byte-order:
- host to {big,little}-endian {16,32}
- {big,little}-endian {16,32} to host
These are not intended to be used in libsa directly, but are rather
intended to be used by host tools which may use libsa routines (such
as loadfile()) which need to use explicit byte-ordering.
(either the current protection or the max protection) that reference
vnodes associated with a file system mounted with the NOEXEC option.
uvm_mmap(): Don't allow PROT_EXEC mappings to be established of vnodes
which are associated with a file system mounted with the NOEXEC option.
executable mappings. Stop overloading VTEXT for this purpose (VTEXT
also has another meaning).
- Rename vn_marktext() to vn_markexec(), and use it when executable
mappings of a vnode are established.
- In places where we want to set VTEXT, set it in v_flag directly, rather
than making a function call to do this (it no longer makes sense to
use a function call, since we no longer overload VTEXT with VEXECMAP's
meaning).
VEXECMAP suggested by Chuq Silvers.
as discussed in tech-net several weeks ago. It turned out that
KAME had already added this functionality to the IPv6 stack, so
I followed their example in adding the sysctl variables
net.inet.icmp.rediraccept and net.inet.icmp.redirtimeout.
device drivers:
- Various native device entries in cdevsw/bdevsw.
- Rework the interrupt infrastructure to provide more flexibility to
the platform-dependent back-end. Rewrite the "ofwgen" simulated
interrupt routines to reflect the changes.
- Clear out the BAT registers and set the fixed battable entries before
calling the platform init routine. The platform init routine is allowed
to set entries in the battable.
- Don't call the platform cons_init routine until after translation is
enabled -- we might need translation to work in order to access bus
space.
- calculate the offset and length of the postbl before byteswapping.
problem noted by der Mouse.
- use offsetof() to determine # of fields to calculate in initial
loop, rather than hard-coding in `52 fields'
- improve comments.
THAT accurate and microtime(9) is painlessly slow on i386 currently.
This speeds up small transfers much. The gain for large transfers
is less significant, but notable too.
Bottleneck was found by Andreas Persson (Re: kern/14246).
Performance improvement with PIII on 661 Mhz according to hbench (with
PIPE_MINDIRECT=8192):
buffersize before after
512 17 49
1024 33 110
2048 52 143
4096 77 163
8192 142 190
64K 577 662
128K 372 392
on PAGE_SIZE. The overhead of setting up Page Loan is pretty much constant
irregardless of page size, so it makes more sense to use fixed constant.
According to hbench, the overhead of Page Loan setup is still significantly
bigger than the performance gain for 4096 byte buffers on i386
(PIII/600Mhz). The difference is smaller on 386DX, but Page Loan is
still not faster for this case.
Also, there is some other code out there which expects 4KB writes
to not block even for 'blocking' write, since it works this
way on some other operating systems.
Partially addresses kern/14246 by Andreas Persson.
and it introduced problems (EBUSY error when opening the driver for
writing in securelevel >= 1, plus manipulating some unitialized data at
the end of chrtoblktbl[])
Make sure the CPUCLASS_686 entry has really 17 (i.e. 16 + default)
name entries as it's supposed to, so that code won't crash when
run on Intel CPUCLASS_686 processor which doesn't have name entry
in the table.
Reported and fix provided by Naoto Morishima in kern/14380.
from sppp_attach.
When destroying the interface, call sppp_detach for proper cleanup.
This avoids a crash from the slow timeout handler for no longer existing
interfaces (spotted by Rémi Zara).
Make sure that an unlabeled device gets at least RAW_PART
in the default in-core disklabel, instead of only a single
partition, so that applying a new disklabel can work normally.
is higher than SPL_HIGH (maybe we should be fixing SPL_HIGH).
If IPL_STATCLOCK is defined, initialize spl_masks[_SPL_STATCLOCK] from
it; otherwise initialize use IPL_CLOCK.
%fs/%gs as appropriate.
XXX Note that the new sigcontext uses the new sigset_t instead of old
int sc_mask. The new FreeBSD sigcontext doesn't contain any backward
compatibility sc_mask, so basically old FreeBSD application making
use of sigcontext are hosed. This can't be fixed in NetBSD compat code.
vnode, we should not attempt to remove the namecache entry. this is because
vget() can sleep (eg. if VXLOCK is set because the vnode is being reclaimed),
and so multiple threads can end up in this context at the same time.
if this happens, each thread ends up removing the cache entry, but
the code to remove the entry assumes that the entry is still valid.
so we should just leave the (now stale) entry in the cache.
if another thread finds the entry again before it is reused,
that thread will notice that the entry is stale and remove it safely.
fixes PR 14042.
the etc Makefile override that by putting USETOOLS into $.MAKEOVERRIDES
This way the default for kernel compiles is still to use the installed
toolchain instead of depending on $TOOLDIR. $TOOLDIR can be used by
simply adding USETOOLS=yes to the command line as usual.
Adjust each ports template to set the default no setting and also pull in
bsd.own.mk if they weren't already to ensure they'll build correctly
with the new toolchain setup.
