Move the interupt handler to a driver_intr_t type function as it was trying
to do way to much for a lightweight filter interrupt function.
Introduce much more locking around fc->mtx. Tested this for lock reversals
and other such lockups. Locking seems to be working better, but there
is much more to do with regard to locking. The most significant lock is
in the BUS RESET handler. It was possible, before this checkin, to set
a bus reset via "fwcontrol -r" and have the BUS RESET handler fire before
the code responsible for asserting BUS RESET was complete. This locking
fixes that issue.
Move some of the memory allocations in the fc struct to the attach function
in firewire.c
Rework the businfo.generation indicator to be merely a on/off bit now.
It's purpose according to spec is to notify the bus that the config ROM
has changed. That's it.
Catch and squash a possible panic in SBP where in the SBP_LOCK was held
during a possible error case. The error handling code would definitely
panic as it would try to acquire the SBP_LOCK on entrance.
Catch and squash a camcontrol/device lockup when firewire drives go away.
When a firewire device was powered off or disconnected from the firewire
bus, a "camcontrol rescan all" would hang trying to poll removed devices
as they were not properly detached. Don't do that.
Approved by: scottl
MFC after: 2 weeks
channel, only accept a real 11g channel; this fixes a problem where we were
wrongly promoting 11b to a Dynamic Turbo G channel which broke scanning on
channel 6
from the inet6 stack along with statistics and make sure we
properly free the rt in all cases.
While the current situation is not better performance wise it
prevents panics seen more often these days.
After more inet6 and ipsec cleanup we should be able to improve
the situation again passing the rt to ip6_forward directly.
Leave the ip6_forward_rt entry in struct vinet6 but mark it
for removal.
PR: kern/128247, kern/131038
MFC after: 25 days
Committed from: Bugathon #6
Tested by: Denis Ahrens <denis@h3q.com> (different initial version)
symlinks after setting the owner. As a result, mode
and timestamp were not restored. This patch corrects the
problem by simply removing the short-circuit for symlinks
and using lchown()/lchmod()/lutimes() always for restoring
metadata.
PR: bin/91316
Submitted by: Jaakko Heinonen
Reviewed by: Joerg Sonnenberger
MFC after: 14 days
that the annotated function returns a pointer that doesn't alias any
extant pointer. This results in a 50%+ speedup in microbenchmarks such
as the following:
char *cp = malloc(1), *buf = malloc(BUF);
for (i = 0; i < BUF; i++) buf[i] = *cp;
In real programs, your mileage will vary. Note that gcc already
performs this optimization automatically for any function called
`malloc', `calloc', `strdup', or `strndup' unless -fno-builtins is
used.
So we are no longer interested in the error returned from
the *fs_doio() functions. With that we can remove the
error variable as its value is unused now.
Submitted by: Christoph Mallon christoph.mallon@gmx.de
a locked route. Thus we have to use RTFREE_LOCKED(9) to get it unlocked
and rtfree(9)d rather than just rtfree(9)d.
Since the PR was filed, new places with the same problem were added
with new code. Also check that the rt is valid before freeing it
either way there.
PR: kern/129793
Submitted by: Dheeraj Reddy <dheeraj@ece.gatech.edu>
MFC after: 2 weeks
Committed from: Bugathon #6
This ensures that the value written is both compatible with
older mtree versions (which expect the value after the period
to be an integer count of nanoseconds after the whole second)
and is a correct floating-point value.
Leave the parsing code unchanged so it will continue to read
older files.
Leave then in struct vinet6 to not break the ABI with kernel modules
but mark them for removal so we can do it in one batch when the time
is right.
MFC after: 1 month
Bluetooth Network Access Point (NAP), Group Ad-hoc Network (GN) and
Personal Area Network User (PANU) profiles.
Obtained from: NetBSD
MFC after: 1 month
the example script of the manpage feeds awk(1) with values larger
than UINT32_MAX. Then awk prints a negative value, and this
messes up $BPFPROG. Trying to load the resulting bpf byte codes
with ngctl then fails.
For example, the output for PATTERN="udp and dst net 255.255.0.0/16"
should be (all in one line):
bpf_prog_len=10
bpf_prog=[
{ code=40 jt=0 jf=0 k=12 }
{ code=21 jt=7 jf=0 k=34525 }
{ code=21 jt=0 jf=6 k=2048 }
{ code=48 jt=0 jf=0 k=23 }
{ code=21 jt=0 jf=4 k=17 }
{ code=32 jt=0 jf=0 k=30 }
{ code=84 jt=0 jf=0 k=4294901760 }
{ code=21 jt=0 jf=1 k=4294901760 }
{ code=6 jt=0 jf=0 k=8192 }
{ code=6 jt=0 jf=0 k=0 }
]
The two k=4294901760 values are displayed as k=-2147483648 by awk.
Replace the awk script of the manpage example with a slower but
safer version, that doesn't really attempt to convert the byte
code printed by tcpdump from string to number and back.
PR: docs/123255
Submitted by: Eugenio Maffione, eugenio.maffione at telecomitalia.it
MFC after: 3 days
but the man page describes conceptual information about the process of
adding a user, thus it should belong to section 7.
- Remove HISTORY and BUGS sections because of the aforementioned reason.
PR: docs/130151
Submitted by: Marian Cerny <jojo@matfyz.cz>
MFC after: 3 days
