| Age | Commit message (Collapse) | Author |
|---|
|
move some of the work from the system task queue to the driver-specific
task queue.
ok jsg@
|
|
address widths in 1st-gen amd64 cpus.
ok kettenis, deraadt, guenther
|
|
sending window for L2TP control is full when the control is terminating
(by a L2TP keepalive failure or other reasons). In that case, if the
L2TP peer didn't respond at all, npppd had kept some PPP sessions
forever.
|
|
|
|
event_loop() returns abnormally which is unexpected.
diff from Yuuichi Someya
|
|
|
|
|
|
the problems it tickled by working outside the biglock on archs
with mutex and clock interaction have been fixed, as evidenced by
the softnet taskq.
ok deraadt@
|
|
Problem noted by patrick keshishian (sidster (at) boxsoft.com)
ok deraadt@
|
|
there are zero groups to remove any that we might otherwise inherit (as
pointed out by grawity at gmail.com) and since the 2nd argument to
setgroups is always a static global it's always valid to dereference in
this case. ok deraadt@ djm@
|
|
|
|
function should be treated as a single negation. From FreeBSD
via Liviu Daia.
|
|
MULTIPROCESSOR, and quote the alpha ARM to explain why; while there, make the
failure messages a bit more detailed.
|
|
call, instead of &curproc. The copy routine may sleep and we may resume on
a different processor. This has been plaguing the alpha MULTIPROCESSOR kernels
since the very beginning; it's amazing this did not cause more havoc.
Joint debugging and hair pulling with dlg@ and deraadt@; ok dlg@
|
|
|
|
|
|
- this testsuite covers patterns
|
|
|
|
ok reyk@
|
|
it was deprecated in lua code, but here the code is new. The documentation
don't mention it either.
ok reyk@
|
|
ok jung@ jmc@
|
|
With important help on the pattern matcher from semarie@
OK semarie@
|
|
elf_symloadx() are reinitialised.
Else show_file() in nm.c will used these variables, even if they has
just been freed. (nm.c +689).
Problem found by afl.
ok miod@
|
|
big enough to fill at least one Elf_Shdr.
While here, inverts calloc() arguments to be calloc(nmemb, size),
according to fread() call after.
This problem was found with afl, with e_shentsize=1.
ok miod@
|
|
after the added bound check).
Found with afl.
ok miod@
|
|
after kernel bss but before end of the image, the page tables used
the read-only mapping of the hole. When booting a small non-generic
kernel, this resulted in a crash, while writing to the page tables
later.
Make sure that the page tables are created after esym and after
end.
OK mlarkin@ deraadt@
|
|
- if section header table is be present
- consistency of section header table size
ok miod@
|
|
Most of the ARP layer already take an ifp pointer and this makes clear
wich chunks of code are messing with ac_enaddr.
Note that our Ethernet code assume that these pointer are interchangeable
since the first element of the "struct arpcom" is a "struct ifnet".
|
|
|
|
|
|
Move bridge_input() outside of ether_input() in order to duplicate packets
flowing through a bridge port before applying any transformation on mbufs.
This saves a various m_adj(9)/M_PREPEND(9) dances and remove the bridge(4)
hack from vlan(4).
Tested by mxb <mxb AT alumni DOT chalmers DOT se> and kettenis@
ok bluhm@
|
|
|
|
it receives such packets when the errno is not EAGAIN or EINTR, it had
closed all L2TP sessions and stoppped the L2TP server. Also fix the
receiving GRE packet since it potentially has the same problem.
|
|
diff from Fabian Raetz.
|
|
Previously, CBS_dup() had its own offset. However, it is more consistent
to copy everything.
ok miod@ jsing@
|
|
The statements were chained together with OR which makes it more annoying
to debug. Also, it was short circuiting all tests as soon as one function
failed. Since the functions are independent, they should each run until
error.
Discussed with miod@ and jsing@
|
|
|
|
this may have caused the monitor to refuse multiple authentications
that would otherwise have successfully completed; ok markus@
|
|
with acpi.
OK mlarkin@
|
|
done by introducining a magic offset. Pages below this offset are mapped
with default memory attributes. Above this offset pages are mapped write
combining.
ok mlarkin@
|
|
from tilo stritzky
thanks miod for help with the diff, and who also noted that
leading whitespace gets stripped too;
|
|
Do rule counter increments after state has been successfully
installed. This has an additional benefit of making error
handling a bit simpler.
OK mpi, bluhm
|
|
|
|
be reinstated when toggled off then on again; this seems to make sense,
but posix is not explicit.
we know that gnu ed and freebsd do reinstate the user defined prompt, but
we do not. no one has yet stepped up to change this, so i'm adding a note
to this effect until such a time as we do/
i have not tweaked the text for -p or "P" to allow for an easy back out
if behaviour does change;
|
|
when an isochronous transfer is done.
Frame lengths are just input values and no driver mess with them.
ok ratchov@, jmatthew@
|
|
won't try to deref the pointer. Based on a patch from mail at quitesimple.org,
ok djm deraadt
|
|
$REMOTE_USER before using them in the Location.
From Sebastien Marie (semarie)
|
|
There's not bit to indicate the speed of a USB3.0 device attached to a hub
port so do not abuse the PORT_TEST bit. Instead make the xhci(4) root hub
report the PORT_POWER_SS bit when appropriate and use it to determin the
speed of a new device.
While here make the root hub report the link state and config error, from
FreeBSD.
|
|
Pass the length to free(9), do not violate the radix/route layer and
set the gateway of a route a bit later to simplify error code path.
ok claudio@
|
|
back to EHCI at suspend. So route the ports back to xHCI at resume.
Problem reported by Adam Wolk, thanks!
|
