even if we've detected a 'root on raid' autoconfigure. This change is really
only cosmetic, since setroot() will still do the right thing and honor
the 'root on foo' setting.
with spl used to protect other allocations and frees, or datastructure
element insertion and removal, in adjacent code.
It is almost unquestionably the case that some of the spl()/splx() calls
added here are superfluous, but it really seems wrong to see:
s=splfoo();
/* frob data structure */
splx(s);
pool_put(x);
and if we think we need to protect the first operation, then it is hard
to see why we should not think we need to protect the next. "Better
safe than sorry".
It is also almost unquestionably the case that I missed some pool
gets/puts from interrupt context with my strategy for finding these
calls; use of PR_NOWAIT is a strong hint that a pool may be used from
interrupt context but many callers in the kernel pass a "can wait/can't
wait" flag down such that my searches might not have found them. One
notable area that needs to be looked at is pf.
See also:
http://mail-index.netbsd.org/tech-kern/2006/07/19/0003.htmlhttp://mail-index.netbsd.org/tech-kern/2006/07/19/0009.html
unsigned for now. This prevents rf_reasonable_label() from rejecting
a valid label when these fields have an integer overflow. The reality
is that these need to be 64-bit quantities, but that will come later.
- struct timeval time is gone
time.tv_sec -> time_second
- struct timeval mono_time is gone
mono_time.tv_sec -> time_uptime
- access to time via
{get,}{micro,nano,bin}time()
get* versions are fast but less precise
- support NTP nanokernel implementation (NTP API 4)
- further reading:
Timecounter Paper: http://phk.freebsd.dk/pubs/timecounter.pdf
NTP Nanokernel: http://www.eecis.udel.edu/~mills/ntp/html/kern.html
original RAIDframe code had the same bug with dag_h being used when
possibly NULL. Use dagList as the starting point for any potential
dag_h's. Move the initialization of dag_h in this part to a
little later. Loop now runs through in equivalent lock-step with the
construction of the dagList earlier in the function.
Addresses Coverity CID 1129 (id=6841 Run 5).
pool each time a new array was configured. This causes grief
with things like 'vmstat -m' by causing it to loop. Make RAIDframe
only initialize PSS bits once.
Pointed out by simonb@. Fix tested by simonb@. Thanks!
"private" structure in struct buf that can be used to keep track of
the request associated with this buffer (the buffer used here is one
allocated from rf_CreateDiskQueueData(), so it's ours to do with what
we please). Shrinks code a little, reduces the run-time memory
footprint a bit, and simplifies both rf_DispatchKernelIO() and
KernelWakeupFunc().
Thanks to yamt for his "why is rf_DispatchKernelIO using another buf"
question which prompted me to revisit this code.
Commit fixes problem reported (privately) by Jeff Rizzo.
XXX: Note that we are still are unable to allow a reconstruct of
a failed spare -- it's easy enough to attempt to do it, but the
code in rf_ContinueReconstructFailedDisk() isn't anywhere close to
ready to allow it. Yet another reason why the whole disk/spare thing
needs to be re-worked. *sigh*
- rather than embedding bufq_state in driver softc,
have a pointer to the former.
- move bufq related functions from kern/subr_disk.c to kern/subr_bufq.c.
- rename method to strategy for consistency.
- move some definitions which don't need to be exposed to the rest of kernel
from sys/bufq.h to sys/bufq_impl.h.
(is it better to move it to kern/ or somewhere?)
- fix some obvious breakage in dev/qbus/ts.c. (not tested)
- introduce rf_buf_queue_check() which checks to see if there
is work to do in the incoming buffer queue
- rf_RaidIOThread() is now responsible for calling raidstart(), and is
also now the only place that calls raidstart()
- raidstrategy() now just queues requests in buf_queue
and signals rf_RaidIOThread() that work has arrived
Hopefully addresses PR#30233
