First and most importantly, I threw out the thread priority-twiddling
implementation of KeRaiseIrql()/KeLowerIrq()/KeGetCurrentIrql() in
favor of a new scheme that uses sleep mutexes. The old scheme was
really very naughty and sought to provide the same behavior as
Windows spinlocks (i.e. blocking pre-emption) but in a way that
wouldn't raise the ire of WITNESS. The new scheme represents
'DISPATCH_LEVEL' as the acquisition of a per-cpu sleep mutex. If
a thread on cpu0 acquires the 'dispatcher mutex,' it will block
any other thread on the same processor that tries to acquire it,
in effect only allowing one thread on the processor to be at
'DISPATCH_LEVEL' at any given time. It can then do the 'atomic sit
and spin' routine on the spinlock variable itself. If a thread on
cpu1 wants to acquire the same spinlock, it acquires the 'dispatcher
mutex' for cpu1 and then it too does an atomic sit and spin to try
acquiring the spinlock.
Unlike real spinlocks, this does not disable pre-emption of all
threads on the CPU, but it does put any threads involved with
the NDISulator to sleep, which is just as good for our purposes.
This means I can now play nice with WITNESS, and I can safely do
things like call malloc() when I'm at 'DISPATCH_LEVEL,' which
you're allowed to do in Windows.
Next, I completely re-wrote most of the event/timer/mutex handling
and wait code. KeWaitForSingleObject() and KeWaitForMultipleObjects()
have been re-written to use condition variables instead of msleep().
This allows us to use the Windows convention whereby thread A can
tell thread B "wake up with a boosted priority." (With msleep(), you
instead have thread B saying "when I get woken up, I'll use this
priority here," and thread A can't tell it to do otherwise.) The
new KeWaitForMultipleObjects() has been better tested and better
duplicates the semantics of its Windows counterpart.
I also overhauled the IoQueueWorkItem() API and underlying code.
Like KeInsertQueueDpc(), IoQueueWorkItem() must insure that the
same work item isn't put on the queue twice. ExQueueWorkItem(),
which in my implementation is built on top of IoQueueWorkItem(),
was also modified to perform a similar test.
I renamed the doubly-linked list macros to give them the same names
as their Windows counterparts and fixed RemoveListTail() and
RemoveListHead() so they properly return the removed item.
I also corrected the list handling code in ntoskrnl_dpc_thread()
and ntoskrnl_workitem_thread(). I realized that the original logic
did not correctly handle the case where a DPC callout tries to
queue up another DPC. It works correctly now.
I implemented IoConnectInterrupt() and IoDisconnectInterrupt() and
modified NdisMRegisterInterrupt() and NdisMDisconnectInterrupt() to
use them. I also tried to duplicate the interrupt handling scheme
used in Windows. The interrupt handling is now internal to ndis.ko,
and the ndis_intr() function has been removed from if_ndis.c. (In
the USB case, interrupt handling isn't needed in if_ndis.c anyway.)
NdisMSleep() has been rewritten to use a KeWaitForSingleObject()
and a KeTimer, which is how it works in Windows. (This is mainly
to insure that the NDISulator uses the KeTimer API so I can spot
any problems with it that may arise.)
KeCancelTimer() has been changed so that it only cancels timers, and
does not attempt to cancel a DPC if the timer managed to fire and
queue one up before KeCancelTimer() was called. The Windows DDK
documentation seems to imply that KeCantelTimer() will also call
KeRemoveQueueDpc() if necessary, but it really doesn't.
The KeTimer implementation has been rewritten to use the callout API
directly instead of timeout()/untimeout(). I still cheat a little in
that I have to manage my own small callout timer wheel, but the timer
code works more smoothly now. I discovered a race condition using
timeout()/untimeout() with periodic timers where untimeout() fails
to actually cancel a timer. I don't quite understand where the race
is, using callout_init()/callout_reset()/callout_stop() directly
seems to fix it.
I also discovered and fixed a bug in winx32_wrap.S related to
translating _stdcall calls. There are a couple of routines
(i.e. the 64-bit arithmetic intrinsics in subr_ntoskrnl) that
return 64-bit quantities. On the x86 arch, 64-bit values are
returned in the %eax and %edx registers. However, it happens
that the ctxsw_utow() routine uses %edx as a scratch register,
and x86_stdcall_wrap() and x86_stdcall_call() were only preserving
%eax before branching to ctxsw_utow(). This means %edx was getting
clobbered in some cases. Curiously, the most noticeable effect of this
bug is that the driver for the TI AXC110 chipset would constantly drop
and reacquire its link for no apparent reason. Both %eax and %edx
are preserved on the stack now. The _fastcall and _regparm
wrappers already handled everything correctly.
I changed if_ndis to use IoAllocateWorkItem() and IoQueueWorkItem()
instead of the NdisScheduleWorkItem() API. This is to avoid possible
deadlocks with any drivers that use NdisScheduleWorkItem() themselves.
The unicode/ansi conversion handling code has been cleaned up. The
internal routines have been moved to subr_ntoskrnl and the
RtlXXX routines have been exported so that subr_ndis can call them.
This removes the incestuous relationship between the two modules
regarding this code and fixes the implementation so that it honors
the 'maxlen' fields correctly. (Previously it was possible for
NdisUnicodeStringToAnsiString() to possibly clobber memory it didn't
own, which was causing many mysterious crashes in the Marvell 8335
driver.)
The registry handling code (NdisOpen/Close/ReadConfiguration()) has
been fixed to allocate memory for all the parameters it hands out to
callers and delete whem when NdisCloseConfiguration() is called.
(Previously, it would secretly use a single static buffer.)
I also substantially updated if_ndis so that the source can now be
built on FreeBSD 7, 6 and 5 without any changes. On FreeBSD 5, only
WEP support is enabled. On FreeBSD 6 and 7, WPA-PSK support is enabled.
The original WPA code has been updated to fit in more cleanly with
the net80211 API, and to eleminate the use of magic numbers. The
ndis_80211_setstate() routine now sets a default authmode of OPEN
and initializes the RTS threshold and fragmentation threshold.
The WPA routines were changed so that the authentication mode is
always set first, followed by the cipher. Some drivers depend on
the operations being performed in this order.
I also added passthrough ioctls that allow application code to
directly call the MiniportSetInformation()/MiniportQueryInformation()
methods via ndis_set_info() and ndis_get_info(). The ndis_linksts()
routine also caches the last 4 events signalled by the driver via
NdisMIndicateStatus(), and they can be queried by an application via
a separate ioctl. This is done to allow wpa_supplicant to directly
program the various crypto and key management options in the driver,
allowing things like WPA2 support to work.
Whew.
where applicable. The main reason for this change is that
the location of make.conf is not constant and can be
modified via __MAKE_CONF. This change also improves
hyper-text linkage in our manpages.
MFC after: 2 weeks
specified. The result is that the package will be listed with a '?'
as if it was not found within the INDEX file, so fix this behaviour.
* Remove trailing spaces.
PR: bin/87136
Submitted by: Sean Farley <sean-freebsd at farley dot org>
MFC after: 3 days
routine, create all the child bio objects before starting the
requests, rather than starting them as created. This closes a race
whereby some number of child operations could complete before the
rest were ever created, and prematurely freeing the parent bio.
This fixes the panics installing in VMWare and qemu
and add a comment about related magic (many lines)).
__kernel_cos[f]() needs a trick to reduce the error to below 1 ulp
when |x| >= 0.3 for the range-reduced x. Modulo other bugs, naive
code that doesn't use the trick would have an error of >= 1 ulp
in about 0.00006% of cases when |x| >= 0.3 for the unreduced x,
with a maximum relative error of about 1.03 ulps. Mistransation
of the trick from the double precision case resulted in errors in
about 0.2% of cases, with a maximum relative error of about 1.3 ulps.
The mistranslation involved not doing implicit masking of the 32-bit
float word corresponding to to implicit masking of the lower 32-bit
double word by clearing it.
sinf() uses __kernel_cosf() for half of all cases so its errors from
this bug are similar. tanf() is not affected.
The error bounds in the above and in my other recent commit messages
are for amd64. Extra precision for floats on i386's accidentally masks
this bug, but only if k_cosf.c is compiled with -O. Although the extra
precision helps here, this is accidental and depends on longstanding
gcc precision bugs (not clipping extra precision on assignment...),
and the gcc bugs are mostly avoided by compiling without -O. I now
develop libm mainly on amd64 systems to simplify error detection and
debugging.
updated by a process holding the snapshot lock. Another process updating a
different inode in the same inodeblock will do copy on write checks and lock in
the opposite direction.
The snapshot code force a copy on write of these blocks manually (cf. start of
expunge_ufs[12]) and these inode blocks are later put on snapblklist.
This partial fix is to 'drain' the relevant ffs_copyonwrite() operation after
installing new snapblklist. This is not a 100% solution since a failed block
allocation can cause implicit fsync() which might deadlock before the new
snapblklist has been installed.
file is flushed by a process not holding snaplk (e.g. bufdaemon). Another
process might hold snaplk and try to access the block due to ffs_copyonwrite
processing.
the cg map buffer being held when writing indirect blocks. The process ends up
in ffs_copyonwrite(), attempting to get snaplk while holding the cg map buffer
lock.
Another process might be in ffs_copyonwrite(), trying to allocate a new block
for a copy. It would hold snaplk while trying to get the cg map buffer lock.
Release the cg map buffer early and use the copy for most of the cgaccount
processing to avoid this deadlock.
skipping the call from ffs_snapremove() if the block number is zero.
Simplify snapshot locking in ffs_copyonwrite() and ffs_snapblkfree() by using
the same locking protocol for low block numbers as for larger block numbers.
This removes a lock leak that could happen if vn_lock() succeeded after
lockmgr() failed in ffs_snapblkfree().
Check if snapshot is gone before retrying a lock in ffs_copyonwrite().
reclamation. If the vnode previously was a fifo then v_op would point to
ffs_fifoops[12] instead of the expected ffs_vnodeops[12], causing a panic at
the end of ffsext_strategy.
the UDF specification specifies a logical sectorsize of 2048.
Instead, get it from GEOM.
- When reading the UDF Anchor Volume Descriptor, use the logical
sectorsize of 2048 when calculating the offset to read from, but
use the actual sectorsize to determine how much to read.
- works with reading a DVD disk and a DVD disk image file via mdconfig
- correctly returns EINVAL if we try to mount_udf an audio CD, instead
of panicking inside GEOM when INVARIANTS is set
a functional change. I know this is a big no no, so this is a forced commit
to note the functional changes from my previous revision:
@@ -196,7 +176,7 @@ wait_for_lock(const char *name, int flag
int fd;
if ((fd = open(name, O_CREAT|O_RDONLY|O_EXLOCK|flags, 0666)) == -1) {
- if (errno == ENOENT || errno == EINTR || errno == EAGAIN)
+ if (errno == EINTR || errno == EAGAIN)
return (-1);
err(EX_CANTCREAT, "cannot open %s", name);
}
17+17+24 bit pi/2 must only be used when subtraction of the first 2
terms in it from the arg is exact. This happens iff the the arg in
bits is one of the 2**17[-1] values on each side of (float)(pi/2).
Revert to the algorithm in rev.1.7 and only fix its threshold for using
the 3-term pi/2. Use the threshold that maximizes the number of values
for which the 3-term pi/2 is used, subject to not changing the algorithm
for comparing with the threshold. The 3-term pi/2 ends up being used
for about half of its usable range (about 64K values on each side).
a maximum error of 2.905 ulps for cosf(), but the algorithm for cosf()
is good for < 1 ulps and happens to give perfect rounding (< 0.5 ulps)
near +-pi/2 except for the bug. The extra relative errors for tanf()
were similar (slightly larger). The bug didn't affect sinf() since
sinf'(+-pi/2) is 0.
For range reduction in ~[-3pi/4, -pi/4] and ~[pi/4, 3pi/4] we must
subtract +-pi/2 and the only complication is that this must be done
in extra precision. We have handy 17+24-bit and 17+17+24-bit
approximations to pi/2. If we always used the former then we would
lose up to 24 bits of accuracy due to cancelation of leading bits, but
we need to keep at least 24 bits plus a guard digit or 2, and should
keep as many guard bits as efficiency permits. So we used the
less-precise pi/2 not very near +-pi/2 and switched to using the
more-precise pi/2 very near +-pi/2. However, we got the threshold for
the switch wrong by allowing 19 bits to cancel, so we ended up with
only 21 or 22 bits of accuracy in some cases, which is even worse than
naively subtracting pi/2 would have done.
Exhaustive checking shows that allowing only 17 bits to cancel (min.
accuracy ~24 bits) is sufficient to reduce the maximum error for cosf()
near +-pi/2 to 0.726 ulps, but allowing only 6 bits to cancel (min.
accuracy ~35-bits) happens to give perfect rounding for cosf() at
little extra cost so we prefer that.
We actually (in effect) allow 0 bits to cancel and always use the
17+17+24-bit pi/2 (min. accuracy ~41 bits). This is simpler and
probably always more efficient too. Classifying args to avoid using
this pi/2 when it is not needed takes several extra integer operations
and a branch, but just using it takes only 1 FP operation.
The patch also fixes misspelling of 17 as 24 in many comments.
For the double-precision version, the magic numbers include 33+53 bits
for the less-precise pi/2 and (53-32-1 = 20) bits being allowed to
cancel, so there are ~33-20 = 13 guard bits. This is sufficient except
probably for perfect rounding. The more-precise pi/2 has 33+33+53
bits and we still waste time classifying args to avoid using it.
The bug is apparently from mistranslation of the magic 32 in 53-32-1.
The number of bits allowed to cancel is not critical and we use 32 for
double precision because it allows efficient classification using a
32-bit comparison. For float precision, we must use an explicit mask,
and there are fewer bits so there is less margin for error in their
allocation. The 32 got reduced to 4 but should have been reduced
almost in proportion to the reduction of mantissa bits.
the modified interface that they use. Changes include:
- Register a different interrupt handler for the new interface. This one is
INTR_MPSAFE, not INTR_FAST, and directly processes completions and AIFs.
- Add an event registration and callback mechanism for the ioctl and CAM
modules can know when a resource shortage clears. This condition was
previously fatal in CAM due to programming oversights.
- Fix locking to play better with newbus.
- Provide access methods for talking to cards with the NEWCOMM interface.
- Fix up the CAM module to be better suited for dealing with newer firmware
on the PERC Si/Di series that requires talking to plain SCSI via aac.
- Add a whole slew of new PCI Id's.
Thanks to Adaptec for providing an initial version of this work and for
answering countless questions about it. There are still some rough edges in
this, but it works well enough to commit and test for now.
Obtained from: Adaptec, Inc.
o Rather than just try to turn off EXCA_INTR_RESET, set the entire register
to 0. This is slightly faster, and a better hammer.
o Move attempted clearing of the output enable (EXCA_PWRCTL_OE) back to
after we turn off the power. Modify it to write 0 so that we don't get
Bad Vcc messages on TI bridges (untested, but ru@ sent me a similar patch)
while at the same time avoiding interrupt storms on Ricoh bridges (tested
by me on my Sony).
# Many of my observations of 'breakage' for this patch are due to some bug
# in the load/unload of cbb.ko unlreated to this change. I'll be investigating
# and fixing that bug in the fullness of time.
an allocation. This fixes the malloc 'use after free' panic on boot that
many were seeing. It doesn't solve the problem of the allocations being
cached and then written past their bounds later. That will take more work.
Submitted by: kan
in 1993 in rev.1.5 of the i386 a.out version (csu/i386/crt0.c).
Profiling uses a magic label "eprol" to delimit the start of the part
of the text section covered by profiling. This label must be placed
before the call to main() to get main() properly profiled. It was
placed there in rev.1.1 of crt0.c. Rev.1.5 imported the initial
implementation of shared libraries in FreeBSD and misplaced the label.
Fortunately, the misplaced label was misspelled and the old label
wasn't removed, so the new label had no effect. Unfortunately, when
profiling was implemented for the ELF in 1998 in rev.1.2 of
csu/i386-elf/crt1.c, only the incorrectly placed label was copied
(after fixing its name). The bug was then copied to all other arches.
The label seems to be still misplaced in NetBSD for most arches. It
is in common.c for most arches so it is even further from being inside
the function that calls main().
I think "eprol" is short for "end of prologue", but it must be placed
before the end of the prologue so that it covers main(). crt0.c has
it before the calls atexit(_mcleanup) and monstartup(...), but it
cannot affect these calls so I moved it after the call to monstartup().
It now also covers the call to _init() but not the newer call to
_init_tls(). Profiling of _init() seems to be harmless, and the call
to _init_tls() seems to be misplaced.
Reviewed by: jdp (long ago, for a slightly different i386 version)
