Any call of tty_close() with a tty refcount of <= 1 is wrong and we will
free the tty in this case. This is a workaround until the underlaying
devfs/tty problems are fixed.
Nuke OSSV4_EXPERIMENT ifdefs from kludgy MIXER_SIZE definition
since the exclusion of sizeof(oss_mixer_enuminfo) is the real
cause of overflow (worth of +3000 bytes).
The addition of few more 32bit integer into struct snd_mixer causing
an accidental MIXER_SIZE overflow. Use 8bit integer instead of 32bit
to reduce the kludge size.
andre [Fri, 29 Sep 2006 21:38:12 +0000 (21:38 +0000)]
Back out rev. 1.152 as it was breaking vlan tag insertion when vlan tag
stripping was disabled due to being in promisc mode. This is a hardware
bug. Update comment to explicitly state the reason the manual vlan tag
insertion in this case. See rev. 1.53 for further information as well.
andre [Fri, 29 Sep 2006 13:47:38 +0000 (13:47 +0000)]
Remove manual vlan header insertion in em_encap(). It is unnecessary as the
generic vlan_start() takes care of it when vlan hardware insertion is disabled.
In em_set_promisc() add a note that BPF may also be enabled without going into
promisc mode.
andre [Fri, 29 Sep 2006 13:37:26 +0000 (13:37 +0000)]
Change em_transmit_checksum_setup() to deal with already inserted vlan headers,
IP options and add skeleton IPv6 support. The new code structure can also be
easily enhanced to support new/more protocols (SCTP) in the future.
andre [Fri, 29 Sep 2006 13:33:30 +0000 (13:33 +0000)]
Change em_tso_setup() to deal with already inserted vlan headers, IP options
and add skeleton IPv6 support. The new code structure can also be easily
enhanced to support new/more protocols (SCTP) and IP fragmentation in the
future.
In em_encap() only try to do TSO if 'dotso' is true.
andre [Fri, 29 Sep 2006 13:12:38 +0000 (13:12 +0000)]
Handle all error cases from bus_dmamap_load_mbuf_sg(). Those are:
- EFBIG means the mbuf chain was too long and bus_dma ran out of segments.
Defragment the mbuf chain and try again. (Already existed, not changed.)
- ENOMEM means bus_dma could not obtain enough bounce buffers at this point
in time. Defer sending and try again later.
- All other errors, in particular EINVAL, are fatal and prevent the mbuf
chain from ever going through. Drop it and report error.
- Checking (nsegs == 0) is unnecessary as bus_dmamap_load_mbuf_sg() always
reports an error if it is < 1.
This prevents broken packets from clogging the interface queue indefinately.
Free tty struct after last close. This should fix the pty-leak by numbers.
Remove workarounds for tty_refcount beeing 0, this will be fixed differently
later.
Free tty struct after last close. This should fix the pty-leak by numbers.
Remove workarounds for tty_refcount beeing 0, this will be fixed differently
later.
Back out rev 1.145 since we initialize the tty struct from scratch and bad
things can't happen anymore.
Rather than autoscaling the number of TIME_WAIT sockets to maxsockets / 5,
scale it to min(ephemeral port range / 2, maxsockets / 5) so that people
with large gobs of memory and/or large maxsockets settings will not
exhaust their entire ephemeral port range with sockets in the TIME_WAIT
state during periods of heavy load.
Those who wish to tweak the size of the TIME_WAIT zone can still do so with
net.inet.tcp.maxtcptw.
Do not chdir(2) until after setuid(2), otherwise "su - username" would fail
when root doesn't have the permission to enter target user's home directory.
If set, PAM environment variable HOME will be used in chdir(2) instead of
pwd->pw_dir, this allows pam_chroot module to continue to function.
Tweak the code to handle intercepting BIOS calls to int 0x15 to shave
another 16 bytes off of BTX (and thus boot2):
- Compare against the value of %eax that is saved on the stack instead of
loading it into %eax (which requires saving the current %eax on the
stack).
- Use %ch to examine the keyboard flag state in the BIOS to see if
Ctrl-Alt-Del is pressed instead of %al so we don't have to save %eax on
the stack anymore.
Optimize the int 15/87 handler for space to shave another 16 bytes off of
BTX (and thus boot2):
- Don't bother saving %eax, %ebx, or %ecx as it is not necessary.
- Use a more compact sequence to load the base value out of a GDT entry
by loading the contiguous low 24 bits into the upper 24 bits of %eax,
loading the high 8 bits into %al, and using a ror to rotate the bits
(2 mov's and a ror) rather than loading the pieces in smaller chunks
(3 mov's and a shl).
- Use movzwl + leal instead of movl + movw + shll + addl.
- Use 'xchgl %eax,%foo' rather than 'movl %eax,%foo' for cases where
it's ok to trash %eax. xchgl %eax, foo is a 1-byte opcode whereas the
mov is a 2-byte opcode.
- Use movzwl rather than xorl + movw.
Ignore a sub-topic match if it is inside the command description.
Otherwise, merge-help can get confused by a command description that
includes a word that starts with a capital S.
andre [Thu, 28 Sep 2006 18:02:46 +0000 (18:02 +0000)]
When tcp_output() receives an error upon sending a packet it reverts parts
of its internal state to ignore the failed send and try again a bit later.
If the error is EPERM the packet got blocked by the local firewall and the
revert may cause the session to get stuck and retry indefinitely. This way
we treat it like a packet loss and let the retransmit timer and timeouts
do their work over time.
The correct behavior is to drop a connection that gets an EPERM error.
However this _may_ introduce some POLA problems and a two commit approach
was chosen.
Various fixups, especially for the upcomming High Definition Audio
commit.
1) sys/dev/sound/pcm/sound.h
sys/dev/sound/pcm/channel.c
* Be more specific: SD_F_SOFTVOL -> SD_F_SOFTPCMVOL
2) sys/dev/sound/pcm/mixer.[ch]
* Implement
mix_setparentchild()
mix_setrealdev()
mix_getparent()
mix_getchild()
The purpose of these functions is implement relative volume
adjustment, such as to tie two or more mixer device into a
single logical device. Usefull for the upcoming HDA driver
and few AC97 codec (such as AD1981B) where the master volume
"vol" need to be implemented using this logical manner.
3) sys/dev/sound/pcm/ac97_patch.[ch]
* Patch for AD1981B codec to enable (automuting) headphone jack sense.
4) sys/dev/sound/pcm/ac97.c
* Implement proper logical master volume for AD9181B codec
through various mix_set{parentchild,realdev}(). Tie both
"ogain" (headphone volume) and "phone" (speaker/lineout) to
a logical "vol".
5) sys/dev/sound/pcm/usb/uaudio_pcm.c
* ditto, for "vol" -> { "pcm" }.
Fix compilation with gcc 4.1. This is imported on the vendor branch as it
was applied in the mainstream source and a later complete import of
4.2.2p3 will complete the fix.
Fix the IPv4 multicast routing detach path. On interface detach whilst
the MROUTER is running, the system would panic as described in the PR.
The fix in the PR is a good start, however, the other state associated
with the multicast forwarding cache has to be freed in order to avoid
leaking memory and other possible panics.
The IPv4 code should clean up multicast group state when an interface
goes away. Without this change, it leaks in_multi (and often ether_multi
state) if many clonable interfaces are created and destroyed in quick
succession.
The concept of this fix is borrowed from KAME. Detailed information about
this behaviour, as well as test cases, are available in the PR.
Fix our ioctl(2) implementation when the argument is "int". New
ioctls passing integer arguments should use the _IOWINT() macro.
This fixes a lot of ioctl's not working on sparc64, most notable
being keyboard/syscons ioctls.
Full ABI compatibility is provided, with the bonus of fixing the
handling of old ioctls on sparc64.
Emulate moving cr0, cr2, cr3, or cr4 into any i386 general register
rather than just emulating mov cr0, eax. This fixes some Compaq/HP BIOS
with DMA (as the BIOS tried to read cr3 so it could translate addresses
if paging was enabled).
Since ULE doesn't honor hlt_cpus_mask don't compile code that prevents
timer interrupt servicing for disabled HTT cores in ULE case. Should be
probably fixed in ULE code instead, but we have no real maintainer for
ULE to do it.
Bump MAX_INITIATORS to 1024- the LSI-Logic can go even higher than
this for 'initiator id'- this is a stopgap until a sparse map is
added.
Make compat defines for offset format (FreeBSD 5 or less).
Add no-asyncio flag. There's some breakage with ASYNC I/O that every
now and then drops us into an infinite loop. This also then does
a fallback to no-asyncio if the AIO option isn't loaded/compiled into
the kernel.
A number of other chanes to try and track some breakage.