- Change the transmit logic to send from a pre-allocated buffer instead
of doing DMA out of the mbufs directly. Timing shows that it takes
4 times longer to map in the mbufs via the sgmap routines instead of
just copying the data! This must be fixed sometine in the future.
- A bunch of other minor fixes.
FIOASYNC ioctrl command had been forwarded from wsmux to each mouse devices
and the devices had set their own event queue async mode. But no one had took
care of the event queue's mode of wsmux itself. Wsmux should make it's own
event queue async mode when it receive FIOASYNC command.
doing a cpu_set_kpc(), just pass the entry point and argument all
the way down the fork path starting with fork1(). In order to
avoid special-casing the normal fork in every cpu_fork(), MI code
passes down child_return() and the child process pointer explicitly.
This fixes a race condition on multiprocessor systems; a CPU could
grab the newly created processes (which has been placed on a run queue)
before cpu_set_kpc() would be performed.
and QUEUE FULL cases; this is already done for all BAD_STATUS cases.
Make sure to requeue the SCB in the above cases internally in the driver,
the SCSI layer doesn't know how to deal with it properly.
some devices may not be found if the BIOS (which would normally
do the reset at startup) is disabled. Should really be done from the SCSI
layer.
Implement the SCBUSIORESET ioctl.
sys_semconfig into a placebo system call, to avoid giving folks an
easy way to wedge processes which use semaphores.
NOTE: unlike 386bsd and freebsd, processes which did not have
semaphore undo records would not be affected by this problem (reducing
it from a serious local denial-of-service problem to a largely
cosmetic problem, since virtually nobody uses semaphores). But the
code is just Wrong so we're ripping it out anyway.
renegades, and must be handled correctly. In particular, they should
be added to their old auto-config set, but then immediately released.
Failing to do otherwise means that they potentialy end up in a
different (and competing!) RAID set which may auto-configure in place
of the correct one, and cause all sorts of chaos at auto-configure
time.
by Anders Magnusson.
Honor elem_count in the KERN_PROC2 case, as well as overall buffer
space. The only user-land code to use this set the elem_count to
"buffer_space / elem_size", so we've had no incorrect behaviour to date.
- Change ktrace interface to pass in the current process, rather than
p->p_tracep, since the various ktr* function need curproc anyway.
- Add curproc as a parameter to mi_switch() since all callers had it
handy anyway.
- Add a second proc argument for inferior() since callers all had
curproc handy.
Also, miscellaneous cleanups in ktrace:
- ktrace now always uses file-based, rather than vnode-based I/O
(simplifies, increases type safety); eliminate KTRFLAG_FD & KTRFAC_FD.
Do non-blocking I/O, and yield a finite number of times when receiving
EWOULDBLOCK before giving up.
- move code duplicated between sys_fktrace and sys_ktrace into ktrace_common.
- simplify interface to ktrwrite()
All the dirop vnops now mark the inodes with a new flag, IN_ADIROP, which
is removed as soon as the dirop is done (as opposed to VDIROP which stays
until the file is written). To address one issue raised in PR#9357.
state into global and per-CPU scheduler state:
- Global state: sched_qs (run queues), sched_whichqs (bitmap
of non-empty run queues), sched_slpque (sleep queues).
NOTE: These may collectively move into a struct schedstate
at some point in the future.
- Per-CPU state, struct schedstate_percpu: spc_runtime
(time process on this CPU started running), spc_flags
(replaces struct proc's p_schedflags), and
spc_curpriority (usrpri of processes on this CPU).
- Every platform must now supply a struct cpu_info and
a curcpu() macro. Simplify existing cpu_info declarations
where appropriate.
- All references to per-CPU scheduler state now made through
curcpu(). NOTE: this will likely be adjusted in the future
after further changes to struct proc are made.
Tested on i386 and Alpha. Changes are mostly mechanical, but apologies
in advance if it doesn't compile on a particular platform.
This bug is only visible on some disks, and I have verified that it works
correctly on VAX. However, du to the author, it is a problem on other
machines/disks as well. Here's the author's comments:
> The MI code for the NCR5380 has a small bug in it The RZ56 wants to
> negotiate for Synchronous Data Transfers and when it does, the current code
> looks like it will send a REJECT message as the spec calls for when we cant
> handle the request. However, a couple of lines of code in the
> ncr5380_msg_in routine are missing wrt the PARITY message and anything that
> wants to send a REJECT message. The ATN setup that is done in
> ncr_sched_msgout is immediately negated by the final byte ACK sequence in
> this routine. This causes bad things to happen to an RZ56.
kernel should only be used for experimentation
comment out some stuff not present or used on my development machine
g/c ef
add (commented out) aha attachment
db_stack_trace_print(__builtin_frame_address(0),...), to printf() the
stack trace to the message bufffer and console. Idea from SunOS/Solaris.
Useful when dumping fails.
re-enqueueing an already free ccb.
Prepare the background to have a unique initialization function for
all of the supported host adapters.
Rename a lot of #define in a more sane way.
Move Carriers initialization and defines in adwmcode.{c,h}
Don't lose Carrier nodes in case of a BUS/Chip reset explicitly invoked
after a DMA failure.
XXX - DMA failure still arise when AdvanSys U[2]W host adapters are used
in conjunction with Intel 82443BX Host Bridge/Controller (rev. 0x03).
!?!Have to understand why!?!
case that write verf is changed. Suggested by mycroft@netbsd.org.
- Reset wcred to NULL (i.e., write credential isn't decieded) everytime
before gathering buffer for new commit, so that there is a chance to
the commit request is merged.
from db_stack_trace_cmd() to db_stack_trace_print(),
and add an additional argument, a function pointer for an
output routine (i.e. printf() or db_printf()).
Add db_stack_trace_cmd() in db_command.[ch], calling
db_stack_trace_print() with db_printf() as the printer.
Move count==-1 special handling from db_stack_trace_print() [nee
db_stack_trace_cmd()] to db_stack_trace_cmd() [nascent here].
Again, I'm unable to test compilation on all affected platforms,
so advance apologies for potential brokenness.
errors from ps(1) and some other kernel grovellers, and return some
data that has previously only been accessable with /dev/kmem read
access. The sysctls are:
+ KERN_PROC2 - return an array of fixed sized "struct kinfo_proc2"
structures that contain most of the useful user-level data in
"struct proc" and "struct user". The sysctl also takes the size of
each element, so that if "struct kinfo_proc2" grows over time old
binaries will still be able to request a fixed size amount of data.
+ KERN_PROC_ARGS - return the argv or envv for a particular process id.
envv will only be returned if the process has the same user id as the
requestor or if the requestor is root.
+ KERN_FSCALE - return the current kernel fixpt scale factor.
+ KERN_CCPU - return the scheduler exponential decay value.
+ KERN_CP_TIME - return cpu time state counters.
With input and suggestions from many people on tech-kern.
which indicates that the process is actually running on a
processor. Test against SONPROC as appropriate rather than
combinations of SRUN and curproc. Update all context switch code
to properly set SONPROC when the process becomes the current
process on the CPU.
Lots of the work was done by Adam Ciarcinsky.
Currently, this only supports CyberPPC boards by Phase 5. Blizzard PPC
expected later.
The kernel is useless but for demonstrating that it starts... especially
interupts, and most of MMU support, is not in yet. Builtin console works,
however, and you can look at the kernel startup messages.
pointer indicating how to print the symbol. This allows db_printsym()
to called in places where db_printf() is not an appropriate output
function.
While straightforward, apologies in advance if I've introduced any minor
syntax errors; I was unable to test compilation this on all the affected
platforms.
scsibus0: waiting 2 seconds for devices to settle...
message during boot crashes the system.
This is because ite_filter() is called before init(8) opens the console.
- Add 16 bytes to the stack on entry to _mcount so we don't
overflow it.
- Use inline interrupt {dis,en}abling instead of calling
profiled function in locore.
