on for a while:
- fine grained TLB shootdown for SMP on i386
- ranged TLB shootdowns.. eg: specify a range of pages to shoot down with
a single IPI, since the IPI is very expensive. Adjust some callers
that used to trigger this inside tight loops to do a ranged shootdown
at the end instead.
- PG_G support for SMP on i386 (options ENABLE_PG_G)
- defer PG_G activation till after we decide what we are going to do with
PSE and the 4MB pages at the start of the kernel. This should solve
some rumored strangeness about stale PG_G entries getting stuck
underneath the 4MB pages.
- add some instrumentation for the fine TLB shootdown
- convert some asm instruction wrappers from functions to inlines. gcc
seems to do a fair bit better with this.
- [temporarily!] pessimize the tlb shootdown IPI handlers. I will fix
this again shortly.
This has been working fairly well for me for a while, but I have tweaked
it again prior to commit since my last major testing round. The only
outstanding problem that I know of is PG_G related, which is why there
is an option for it (not on by default for SMP). I have seen a world
speedups by a few percent (as much as 4 or 5% in one case) but I have
*not* accurately measured this - I am a bit sceptical of these numbers.
I don't believe anyone is quite using the sparc64 kernel sources in CVS
yet -- things aren't just quite ready (but almost). So this commit should
be OK to make.
- fix the warnings, they are there for a reason!
- add -DNO_ERROR to your make(1) command.
- add 'makeoptions NO_WERROR=true' to your kernel config.
- add 'nowerror' to conf/files* that have warnings that should be fixed
due to tracking 3rd party vendor code.
- add 'nowerror' to conf/files* where the warning is false due to a
compiler bug and fixing it with brute force would be too expensive.
There are some very sloppy warnings in our kernel build, come on folks!
'make release' uses -DNO_WERROR intentionally.
window to the user stack while in a nested kernel trap. We do this for
entry to the kernel from user mode, but if we get an interrupt in kernel
mode while there are still user windows in the cpu, and we attempt to spill
to the user stack, we may take too many nested traps and overflow the trap
stack, causing a red state exception. This is needed by upcoming changes
to allow the user tsb to not be locked in the tlb.
Reviewed by: tmm
the inter-value histogram for 2000 samples. If the width is 3 or less
for 10 consequtive samples, we trust the counter to be good, otherwise
we use the *_safe() method.
This method may be too strict, but the worst which can happen is that
we take the performance hit of the *_safe() method when we should not.
Make the *_safe() method more discriminating by mandating that the three
samples do not span more than 15 ticks on the counter.
Disable the PCI-ident based probing as a means to recognize good
counters.
Inspiration from: dillon and msmith
were destined for a broadcast IP address. All TCP packets with a
broadcast destination must be ignored. The system only ignored packets
that were _link-layer_ broadcasts or multicast. We need to check the
IP address too since it is quite possible for a broadcast IP address
to come in with a unicast link-layer address.
Note that the check existed prior to CSRG revision 7.35, but was
removed. This commit effectively backs out that nine-year-old change.
PR: misc/35022
virtual page number in a much more convenient way; all in one piece. This
greatly simplifies the comparison for a matching tte, and allows the fault
handlers to be much simpler due to not having to load wierd masks.
Rewrite the tlb fault handlers to account for the new format. These are also
written to allow faults on the user tsb inside of the fault handlers; the
kernel fault handler must be aware of this and not clobber the other's
registers. The faults do not yet occur due to other support that is needed
(and still under my desk).
Bug fixes from: tmm
deep in <stand.h> to eventually include <time.h> to declare the user
version.
This is not quite the right place to declare it, but <stand.h> would
be worse because time() is very MD so it isn't in libstand.
Many places in the boot sources still get the user version using only
1 layer of pollution (#include <sys/time.h>. Some pollute themselves
directly (#include <time.h>). But the boot Makefiles are too broken
to enable warnings for redeclarations.
but never accept'ed, so they must be destroyed. Originally, unp_drop()
detected this situation by checking if so->so_head is non-NULL.
However, since revision 1.54 of uipc_socket.c (Feb 1999), so->so_head
is set to NULL before calling soabort(), so any unix-domain sockets
waiting to be accept'ed are leaked if the server socket is closed.
Resolve this by moving the socket destruction code into uipc_abort()
itself, and making it unconditional (the other caller of unp_drop()
never needs the socket to be destroyed). Use unp_detach() to avoid
the original code duplication when destroying the socket.
PR: kern/17895
Reviewed by: dwmalone (an earlier version of the patch)
MFC after: 1 week
latch the acpi timer, resulting in weird deltas. The problem is severe
enough to adversely effect the timecounter code.
Default to the 'safe' version of the get-timecount function. The probe
will override it if a known-good chipset is found. This is temporary
until a more complete solution is found.
Reviewed by: phk
our feet when we look inside timecounter structures.
Make the "sync_other" code more robust by never overwriting the
tc_next field.
Add counters for the bin[up]time functions.
Call tc_windup() in tc_init() and switch_timecounter() to make sure
we all the fields set right.
Most of the contents are commented out as they are as-yet untested.
However, I wanted the contents to match our other arches, so that when
people make changes to {i386,alpha,ia64}, they will also make the same
changes here.
AIOGCAP ioctl reports software-emulated formats. It defaults to on. People
who use performance-sensitive audio software and do not want it to pick a
software-emulated audio format instead of one supported by their hardware
should turn it off.
This unbreaks isdnphone(1) on systems with PCM-only sound cards.
Approved by: cg
pmap_qenter and pmap_qremove in preference to pmap_kenter/pmap_kremove.
The former maps in multiple pages at a time, and so can do a ranged
flush. Don't assume that pmap_kenter and pmap_kremove will flush the tlb,
even though they still do. It will not once the MI code is updated to use
pmap_qenter and pmap_qremove.
will be used to reduce the number of tlb shootdown ipis in an smp system
by sending one ipi for a whole range of pages, instead of one per page.
Munge the context demap operations slightly to support demapping a non-primary
context.
If we don't do this here there's a 1 instruction race where an interrupt
could come in and crash the user process due to having no stack.
2. Pass %fsr to the user trap handler in %l4. Since %fsr can only be loaded
from or stored to memory, we need to do some contortions and temporarily
save it to the alternate global stack.
3. Reload the pcb and pcpu registers for traps in kernel mode, for sanity.
Submitted by: tmm (1, 2)
New locks are:
- pgrpsess_lock which locks the whole pgrps and sessions,
- pg_mtx which protects the pgrp members, and
- s_mtx which protects the session members.
Please refer to sys/proc.h for the coverage of these locks.
Changes on the pgrp/session interface:
- pgfind() needs the pgrpsess_lock held.
- The caller of enterpgrp() is responsible to allocate a new pgrp and
session.
- Call enterthispgrp() in order to enter an existing pgrp.
- pgsignal() requires a pgrp lock held.
Reviewed by: jhb, alfred
Tested on: cvsup.jp.FreeBSD.org
(which is a quad-CPU machine running -current)
Use ACPI_SUCCESS/ACPI_FAILURE consistently.
The AcpiGetInto* interfaces are obsoleted by ACPI_ALLOCATE_BUFFER.
Convert to using a kthread rather than timeout() to avoid problems
with the interpreter sleeping.
Use ACPI_SUCCESS/ACPI_FAILURE consistently.
The AcpiGetInto* interfaces are obsoleted by ACPI_ALLOCATE_BUFFER.
Use _ADR as well as _BBN to get our bus number.
Use ACPI_SUCCESS/ACPI_FAILURE consistently.
The ACPI global lock acquire takes a timeout value. I'm not sure what
we should do about timeouts on it; a deadlock against this lock is
catastrophic.
Use ACPI_SUCCESS/ACPI_FAILURE consistently.
The AcpiGetInto* interfaces are obsoleted by ACPI_ALLOCATE_BUFFER.
Kill off the timeouts that used to read _BIF and _BST. These are
invoked when the battery is actually read. timeout() is dangerous
in combination with ACPI, as the interpreter can block.
This driver still needs more work.
Use ACPI_SUCCESS/ACPI_FAILURE consistently.
The AcpiGetInto* interfaces are obsoleted by ACPI_ALLOCATE_BUFFER.
Add AcpiBatteryIsPresent helper to determine whether a battery device
is inserted.
Add ACPI_ALL_DRIVERS to the list of debug layers, now that we own the
namespace for this.
Pr:
watchpoint support for debugging (under LOADER_DEBUG). Claim the
physical and virtual addresses used to map the kernel from the prom;
we map it ourselves behind the scenes though. Add a reboot command.
Submitted by: tmm
While in userland, keep the thread's ucred reference in a shadow
field so that the usual place to store it is NULL.
If DIAGNOSTIC is not set, the thread ucred is kept valid until the next
kernel entry, at which time it is checked against the process cred
and possibly corrected. Produces a BIG speedup in
kernels with INVARIANTS set. (A previous commit corrected it
for the non INVARIANTS case already)
Reviewed by: dillon@freebsd.org
Remove bowrite(), it is now unused.
This is the first step in getting entirely rid of BIO_ORDERED which is
a generally accepted evil thing.
Approved by: mckusick
so that this is safe even if VARIABLE is longer than kern.argmax.
There is another instance of CFILES which might need the same treatment,
and might be noticed when doing a "make links".
The same has to be done in RELENG_4 (on some different file).
Noticed-by: picobsd cross-compiling LINT
Suggested-by: Alfred (bright@mu.org), des@freebsd.org
MFC-after: 3 days
counter type, as threatened in rev.1.8 (the density doesn't need to
be recorded since it can be derived from other fields). This doesn't
affect binary compatibility, but new utilities won't be able to depend
on the contents of this field because libc/gmon/gmon.c was broken --
it wrote garbage to the spare fields.
Added a history counter type field to struct gmonparam. This breaks
binary compatibility a little, since kgmon wanted to read the whole
struct. Fixed kgmon to only depend on reading the critical earlier
parts of the struct. This should also fix 6+ year old breakage of
binary compatibility when the profrate field was added.
Only initialize the new field in struct gmon for now, so that the
compatibility code for this (in kgmon) gets tested. The compatibility
code has to guesstimate the value. The new field in struct gmonparam
is for the kernel to initialize so that kgmon doesn't have to guess.
- P_INMEM checks in all the functions. P_INMEM must be checked because
PHOLD() is broken. The old bits had bogus locking (using sched_lock)
to lock P_INMEM. After removing the P_INMEM checks, we were left with
just the bogus locking.
- large comments. They were too large, but better than nothing.
Remove obfuscations that were gained in transition in rev.1.76:
- PROC_REG_ACTION() is even more of an obfuscation than PROC_ACTION().
The change copies procfs_machdep.c rev.1.22 of i386/procfs_machdep.c
verbatim except for "fixing" the old-style function headers and adjusting
function names and comments. It doesn't remove the bogus locking.
Approved by: des
OUT status. We are, apparently, required to force the f/w to log back in
if we want to try and talk to that disk again. This means either issuing
a LOGIN LOCAL LOOP PORT mailbox command, or by issuing a LIP. I've elected
to issue a LIP because this has a better chance of waking up the disk which
clearly just crashed and burned.
These should not occur at all. If they do, they should be darned rare.
MFC after: 1 week
It doesn't actually do it yet though. This adds a flag to config so
that we can exclude certain vendor files from this even when the rest
of the kernel has it on. make -DNO_WERROR would also bypass all of it.
so that after the first time we can follow the pointer instead
of having to scan the list.
This was the intended behaviour from day one.
PR: 34639
MFC-after: 3 days
the next commit actually doing the:
return val; -> return (val);
changes. This commit was done in preparation for getting ``struct
modules'' locked down.
Reviewed by: bde
Approved by: dfr
call VOP_CLOSE() with vp unlocked; clean up the return path a little,
in as much as our namei/vnode operation return paths can be cleared
up. For a return case that was apparently never taken, this sure
is ugly.
Reviewed by: jeffr
count that would otherwise be on one of the free queues. This eliminates a
panic when broken programs unmap memory that still has pending IO from raw
devices.
Reviewed by: dillon, alc
- Allow the OOM killer to target processes currently locked in
memory. These very often are the ones doing the memory hogging.
- Drop the wakeup priority of processes currently sleeping while
waiting for their page fault to complete. In order for the OOM
killer to work well, the killed process and other system processes
waiting on memory must be allowed to wakeup first.
Reviewed by: dillon
MFC after: 1 week
- Leave 10 processes for root-only use, the previous
value of 1 was insufficient to run ps ax | more.
- Remove the printing of "proc: table full". When the table
really is full, this would flood the screen/logs, making
the problem tougher to deal with.
- Force any process trying to fork beyond its user's maximum
number of processes to sleep for .5 seconds before returning
failure. This turns 2000 rampaging fork monsters into 2000
harmlessly snoozing fork monsters.
Reviewed by: dillon, peter
MFC after: 1 week
If the credential on an incoming thread is correct, don't bother
reaquiring it. In the same vein, don't bother dropping the thread cred
when going to userland. We are guaranteed to need it when we come back,
(which we are guaranteed to do).
step and the others are reservations for coming code.
All will be stubbed in this kernel in the next commit.
This will allow people to easily make KSE binaries for userland testing
(the syscalls will be in libc) but they will still need a real KSE kernel
to test it. (libc looks in /sys to decide what it should add stubs for).
to notify other nodes about the address change. Otherwise, they
might try and keep using the old address until their arp table
entry times out and the address is refreshed.
Maybe this ought to be done for INET6 addresses as well but i have
no idea how to do it. It should be pretty straightforward though.
MFC-after: 10 days
deprecated in favor of the POSIX-defined lowercase variants.
o Change all occurrences of NTOHL() and associated marcros in the
source tree to use the lowercase function variants.
o Add missing license bits to sparc64's <machine/endian.h>.
Approved by: jake
o Clean up <machine/endian.h> files.
o Remove unused __uint16_swap_uint32() from i386's <machine/endian.h>.
o Remove prototypes for non-existent bswapXX() functions.
o Include <machine/endian.h> in <arpa/inet.h> to define the
POSIX-required ntohl() family of functions.
o Do similar things to expose the ntohl() family in libstand, <netinet/in.h>,
and <sys/param.h>.
o Prepend underscores to the ntohl() family to help deal with
complexities associated with having MD (asm and inline) versions, and
having to prevent exposure of these functions in other headers that
happen to make use of endian-specific defines.
o Create weak aliases to the canonical function name to help deal with
third-party software forgetting to include an appropriate header.
o Remove some now unneeded pollution from <sys/types.h>.
o Add missing <arpa/inet.h> includes in userland.
Tested on: alpha, i386
Reviewed by: bde, jake, tmm
to be able to use 48bit addressing mode, but says the 48bit
size of the disk is 0, which according to spec means it can
address zero sectors in 48bit mode, why then say it supports
48bit mode at all..
buffer length, determine if the pointer is to a valid string. Currently,
the only check is whether a '\0' appears in the buffer. This is useful
when pulling in a structure from userland that may contain one or more
strings, and validity testing must be performed on elements of the
structure. When copying normal string arguments, copyinstr() is
expected to be used.
VOP_CLOSE() on the vnode, so that VOP_OPEN() and VOP_CLOSE() calls
are symmetric in all failure cases. This prevents an 'open' reference
from being leaked in that unlikely failure scenario.
should require a shared lock, rather than an exclusive lock, which can
improve performance. No actual code change here, since a number of
VFS locking fixes are in the works.
The use of the zone allocator may or may not be overkill.
There is an XXX: over in ufs/ufs/ufs_vnops.c that jlemon may need
to revisit.
This shaves about 60 bytes of struct vnode which on my laptop means
600k less RAM used for vnodes.
ucontext_t. Forward declare struct __ucontext in <sys/signal.h> and
remove reliance on <sys/ucontext.h> being included.
While I'm here, also hide osigcontext types from userland; suggested
by bde.
Namespace pollution noticed by: Kevin Day <toasty@shell.dragondata.com>
If you want QLogic to look at a potential f/w problem for FC cards, you really
have to provide them info in the format they expect. This involves dumping
a lot of hardware registers (> 300 16 bit registers) and a lot of SRAM
(> 128KB minimum). Thus all of this code is #ifdef protected which will
become an option so that the memory allocation of where to dump the crash
image is pretty expensive. It's worth it if you have a reproducible problem
because they have some tools that can tell them, given the f/w version,
the precise state of everything.
MFC after: 1 week
Appologies for making this one bulk commit, but I have tested all these
changes together and don't want to break anything by trying to disentangle
it.
o Make debugging a sysctl/tunable
o Remove flags word from yenta chip info, it is unused
o Make 16-bit card I/O range and 32-bit card I/O range tunables
o Start the rename of pccbb to cbb to match NetBSD by misc renames.
o Kill the now bogus list of softcs to create kthread. Instead, just
create the kthread in the attach routine.
o Remove sc_ from some structure names. It isn't needed.
o Refine chipset lookup code.
o Match generic PCI <-> CardBus bridges. We specifically don't generically
match PCI PCMCIA bridges because they are not, with one exception, yenta
devices.
o Add some comments about the why we need to have a function table ala
OLDCARD
o The PCI interrupt routing by using the ExCA registers is needed for
for all bridges, per the spec, not just TI ones.
o Collapse TOPIC95 and TOPIC95B.
o Using the ToPIC 97 and 100 datasheets, try to support these bridges better,
but more work is needed.
o Generally clarify some XXX comments and add them in a few places where
things didn't look right to me.
o Move interrupt generating register access until after we establish an ISR.
o Add support for YV and XV cards. X and Y are numbers to be determined
later (but maybe never).
o factor powerup code for 16-bit and 32-bit cards.
o When a card supports more than one voltage, prefer the lowest supported
volage. Windows does this, and MS's design guides imply this is the
right thing to do.
o Document race between kthread_exit(0) and kldunload's unmapping of pages
that John Baldwin and I discovered.
o Debounce the CSC interrupt a little better.
o When a 16-bit card is inserted when we don't have a pccard child,
warn about it better. Ditto for 32-bit card.
o Ack ALL the interrupt bits that we get, not just 0x1.
o maybe a couple minor style nits corrected.
reaquiring it. In the same vein, don't bother dropping the thread cred
when goinf ot userland. We are guaranteed to nned it when we come back,
(which we are guaranteed to do).
Reviewed by: jhb@freebsd.org, bde@freebsd.org (slightly different version)
the structure definitions come from NetBSD to make it easier to share card
definitions. The driver only acts as a shim between the pci bus and the
sio driver. Later pci parallel ports could also be supported through this
driver. Support for most single and multiport pci serial cards should be
as simple as adding its definition to pucdata.c
Tested with the following pci cards:
Moxa Industio CP-114, 4 port RS-232,RS-422/485
Syba Tech Ltd. PCI-4S2P-550-ECP, 4 port RS-232 + 2 parallel ports
Netmos NM9835 PCI-2S-550, 2 port RS-232
- clobbering of jsp's $Id$ by FreeBSD's old $Id$.
- long lines in recent KSE changes (procfs_ctl.c).
- other style bugs in KSE changes (most related to an shadowed variable
in procfs_status.c -- the td in the outer scope is obfuscated by
PFS_FILL_ARGS).
Approved by: des
- clobbering of jsp's $Id$ by FreeBSD's old $Id$.
- lost Berkeley id in procfs_dbregs.c
- long lines in recent KSE changes.
- various gratuitous differences between procfs_*regs.c.
patch from a year ago: give file flags their own type. This does not
(yet) change the type used by system calls or library functions.
The underlying type was chosen to match what is returned by stat().
- Remove change for my local configuration that slipped in with
the last commit; I am having problems booting when multiple SCSI
disks are attached, so I will change this part as soon as I find
a solution, anyway.
- Remove two constants that were needed in conjuction with the
NetBSD disklabel header. Use the FreeBSD equivalents.
To boot from NetBSD/sparc64 partitions, define LABELOFFSET to
be 128.
- Do not use the complete open firmware path to filter out cdrom drives.
No path containing "cdrom" is detected as a disk now.
- Simplify some code.
be compiled. Old tty ioctls are still used (possibly ifdef'ed) in at
least the following programs in the src tree:
atc des ee fontedit gdb gdbserver lock ntp perl5 tcsh telnet top vttest
rp.c:
Unremoved used variables so that the support for old ioctls actually
compiles.
Not tested at runtime by: bde
ACPI_NO_SEMAPHORES, ASR_MEASURE_PERFORMANCE, AST_DEBUG, ATAPI_DEBUG,
ATA_DEBUG, BKTR_ALLOC_PAGES, BROOKTREE_ALLOC_PAGES, CAPABILITIES,
COMPAT_SUNOS, CV_DEBUG, MAXFILES, METEOR_TEST_VIDEO, NDEVFSINO,
NDEVFSOVERFLOW, NETGRAPH_BRIDGE, NETSMB, NETSMBCRYPTO, PFIL_HOOKS,
SIMOS, SMBFS, VESA_DEBUG, VGA_DEBUG.
Start using #! to comment out negative options and ## to comment out
broken options.
atapi-all.c:
Fixed rotted bits that were hiding under ATAPI_DEBUG.
atapi-cd.c:
#include "opt_ata.h" so that ACD_DEBUG is actually visible.
ata/atapi-tape.c
#include "opt_ata.h" so that AST_DEBUG is actually visible.
In order of importance:
+ each cluster now uses private data structures (filtering and
local address tables) so you can treat them as fully independent
switches. This part of the work was supported by:
Cisco Systems, Inc. - NSITE lab, RTP, NC.
+ cleaned up the handling of configuration, so the system will behave
much better when real or pseudo devices are dynamically attached
or detached. It should also not panic anymore on systems with large
number of devices, closing a few existings PRs on the topic.
+ while at it, add support for VLAN. This means that a FreeBSD box
can now work as a real VLAN switch, with trunk interfaces etc.
As an example:
ifconfig vlan0 vlan 3 vlandev dc0
ifconfig vlan1 vlan 4 vlandev dc0
net.link.ether.bridge_cfg="vlan0:3,dc1:3,vlan1:4,dc1:4"
uses dc0 as a trunk interface, and dc1 and dc3 as ports on vlans 3 and 4
You get the idea...
NOTA BENE: by default bridge_cfg is initialised to "" so even if
you enable bridging, no packets will be bridged until you set the
list of interfaces on which you want this to happen.
+ large restructuring of the code, moving private vars and types from
bridge.h to bridge.c.
+ added a lot of comments to the code to explain how to use it.
This escaped because DEVICE_POLLING is disabled in LINT being
not compatible with SMP. In fact, it is only a runtime problem,
so if we could recognize that we are building a LINT kernel
we could as well disable the check for SMP being defined.
Reported-by: Joe Clarke
boot sequence.
The new pmap.c is based on NetBSD's newer pmap.c (for the mpc6xx processors)
which is 70% faster than the older code that the original pmap.c was based
on. It has also been based on the framework established by jake's initial
sparc64 pmap.c.
There is no change to how far the kernel gets (it makes it to the mountroot
prompt in psim) but the new pmap code is a lot cleaner.
Obtained from: NetBSD (pmap code)
break scheduling because negative priorities were most fixed up by
converting kg_pri_user back to the correct type.
Fixed some style bugs in previous commit (non-terminated sentence fragments
and regressions in comments).
SMP we'd like as much feedback as possible from users about possible
locking problems as early as possible.
To negate most of the performance impact I've also enabled
WITNESS_SKIPSPIN. I've done this as we've been running WITNESS
over the spinlock code for a while without incident and it goes a
long way to making the performance problems of WITNESS much more
bearable.
Users who should be running current should know about turning WITNESS
off for performance reasons.
That said and done, WITNESS could/should be made into a tuneable,
but we'll leave that as an excersize to those that want to disable
it without a kernel recompile.
descriptors. This simplifies code for jumbo frames.
- Cleaned up coding conventions to make code more unix-like.
- Cleaned up code in if_em_fxhw.c and if_em_phy.c.
Added relevant comments.
MFC after: 1 week
support for managing both streaming caches on psycho pairs).
Use explicit bus space accesses instead of mapping the device memory into
kva.
Move DVMA allocation to the map creation/dma memory allocation functions.
disable interrupts completely, and stxa_sync(), which performs a store
immediately followed by a membar #Sync with interrupts disabled (this
is needed for writes to diagnostic registers).
