Split the 'video' ACPI lock up into two locks to resolve a LOR with the
sysctl lock. The 'video' lock now protects the 'bus' of video output
devices attached to a graphics adapter. It is used when iterating over
the list of outputs, etc. The 'video_output' lock is used to lock the
output-specific data similar to a driver lock for the individual video
outputs.
Don't do an IPv6 operation when the kernel doesn't have
an IPv6 support.
Reported by: Alexander Best <alexbestms__at__math.uni-muenster.de>
Confirmed by: Paul B. Mahol <onemda__at__gmail.com>,
Alexander Best <alexbestms__at__math.uni-muenster.de>
1) Restore the PCI Express control register after a watchdog
reset. This is required because the device will come out
of watchdog reset with the pectl reg at its default state,
and important BIOS configuration (like max payload size)
could be lost.
2) Call mxge_start_locked() for every tx queue before dropping
the lock in the watchdog handler. This is required, as
the queue's buf ring may have filled during the reset.
Correct a bug that could lead to a kernel panic if a user attempted to
perform 802.11 operations directly on the ndis0 interface before the
first VAP (wlan0) had been created. This would lead to a NULL-pointer
dereference in the kernel.
Submitted by: Paul B. Mahol <onemda@gmail.com>
MFC after: 3 days
When releasing a read/shared lock we need to use a write memory barrier
in order to avoid, on architectures which doesn't have strong ordered
writes, CPU instructions reordering.
Diagnosed by: fabio
Reviewed by: jhb
Tested by: Giovanni Trematerra
<giovanni dot trematerra at gmail dot com>
Reserve system call numbers for Capsicum security framework capabilities,
capability mode, and process descriptors: cap_new, cap_getrights, cap_enter,
cap_getmode, pdfork, pdkill, pdgetpid, and pdwait.
Add some bits of HDMI/DisplayPort support from later specification updates.
It may be not enough to make them work, but at least should give some
information about these beasts.
The first 96 bytes may not be zeroes. It can contain trivial boot
code that merely emits an error and waits for a key press before
rebooting. The error being that extended partitions are not
bootable. The origin is presumed to be Windows 2000; Windows XP
does not do this...
For now, ignore the first 96 bytes when checking that the EBR is
(for the most part) all zeroes.
Tested by: Mario Lobo <mlobo@digiart.art.br>
MFC after: 1 week
For AR8132 fast ethernet controller, do not report 1000baseT
capability to mii(4). Even though AR8132 uses the same model/
revision number of F1 gigabit PHY, the PHY has no ability to
establish 1000baseT link. I have no idea why Atheros use the same
device/model id for this PHY.
With this change atphy(4) does not report 1000baseT media
capability and manual 1000baseT configuration is also disabled
which is more desirable behavior for 10/100Mbps PHY.
Add DGE-560SX(Yukon XL) to the supported device list. Many thanks
to "Eugene Perevyazko <john <> dnepro dot net>" who kindly gave
remote access to system with DGE-560SX.
Add hack to pass controller specific information to phy driver.
Unlike most other PHYs there is no easy way to know which media
type the PHY supports on Marvell PHYs. MIIF_HAVEFIBER flags is now
passed via bus-specific instance variable of a device. While I'm
here add 88E1112 specific work around to set SIGDET polarity low.
Many thanks "Eugene Perevyazko <john <> dnepro dot net>" who kindly
gave remote access to system with DGE-560SX.
It seems some 82559ER controllers do not support Rx checksum
offloading. Datasheet said nothing about the limitation of 82559ER
except WOL. Explicitly disable Rx checksum offloading for
controllers that is known to lack the capability.
PR: kern/138135
Tested by: Gooderum, Mark < mgooderum <> websense dot com >
Temporarily disable the use of 1GB page mappings by the direct map. There
are currently two problems with the use of 1GB page mappings by the direct
map. First, at least one device driver uses pmap_extract() rather than
DMAP_TO_PHYS() to translate a direct map address to a physical address.
Unfortunately, neither pmap_extract() nor pmap_kextract() yet support 1GB
page mappings. Second, pmap_bootstrap() needs to interrogate the MTRRs to
ensure that a 1GB page mapping doesn't span two MTRRs of different types.
Reported and tested by: Daniel O'Connor
MFC after: 3 days
- F_READAHEAD: specify the amount for sequential access. The amount is
specified in bytes and is rounded up to nearest block size.
- F_RDAHEAD: Darwin compatible version that use 128KB as the sequential
access size.
A third argument of zero disables the read-ahead behavior.
Please note that the read-ahead amount is also constrainted by sysctl
variable, vfs.read_max, which may need to be raised in order to better
utilize this feature.
Thanks Igor Sysoev for proposing the feature and submitting the original
version, and kib@ for his valuable comments.
Use correct sizeof() object for klist 'list'. Currently, struct klist
contained only SLIST_HEAD as its member, thus sizeof(struct klist) would
equal to sizeof(struct klist *), so this change makes the code more
correct in terms of semantics, but should be a no-op to compiler at this
time.
Add extra safety locking when clobbering xfer->flags_int.started in start and
stop functions, because xfer->flags_int is also updated by the USB controller,
under the controller lock.
- clean up USB detach logic. There seems to be some problems detaching multiple
USB HUBs connected in series from the root.
- after this patch the rule is:
1) Always use device_detach() on the USB HUB first.
2) Never just device_delete_child() on the USB HUB, because that function
will traverse to all the device leaves and free them first, and then the USB
stack will free the devices twice which doesn't work very well.
- make sure the did DMA delay gets set after the timeout has elapsed to make
logic more clear. There is no functional difference.
Silence warning printed by getfsspec(3) when /etc/fstab does not exist
fstab: /etc/fstab:0: No such file or directory
and from dump(8) when setfsent(3) fails due to /etc/fstab not existing:
DUMP: Can't open /etc/fstab for dump table information: No such...
This makes daily and security periodic runs somewhat cleaner in jails
which lack /etc/fstab files.
Add a simple C program to check mmap calls to various different addresses.
The most important test is the mapping fixed at address 0 depending on the
new sysctl.
Things will be updated and possibly converted to m4/.t style once the
details about the kernel patch will be shaken out.
simon [Sun, 27 Sep 2009 21:01:07 +0000 (21:01 +0000)]
- When we run our trap cleanup handler, echo that we are running this
handler to make it more clear why we are 'suddenly' running df,
umount, and mdconfig.
- Remove trap handler again after we have unconfigured the memory
device etc. Before we could end up running the trap handler if a
later stage failed, which was a bit confusing and not really useful.
ed [Sun, 27 Sep 2009 18:19:41 +0000 (18:19 +0000)]
Add support for VT200-style mouse input.
Right now if applications want to use the mouse on the command line,
they use sysmouse(4) and install a signal handler in the kernel to
deliver signals when mouse events arrive. This conflicts with my plan to
change to TERM=xterm, so implement proper VT200-style mouse input.
Because mouse input is now streamed through the TTY, it means you can
now SSH to another system on the console and use the mouse there as
well. The disadvantage of the VT200 mouse protocol, is that it doesn't
seem to generate events when moving the cursor. Only when pressing and
releasing mouse buttons.
There are different protocols as well, but this one seems to be most
commonly supported.
Reported by: Paul B. Mahol <onemda gmail com>
Tested with: vim(1)
simon [Sun, 27 Sep 2009 14:49:51 +0000 (14:49 +0000)]
Do not allow mmap with the MAP_FIXED argument to map at address zero.
This is done to make it harder to exploit kernel NULL pointer security
vulnerabilities. While this of course does not fix vulnerabilities,
it does mitigate their impact.
Note that this may break some applications, most likely emulators or
similar, which for one reason or another require mapping memory at
zero.
This restriction can be disabled with the security.bsd.mmap_zero
sysctl variable.
des [Sat, 26 Sep 2009 23:05:01 +0000 (23:05 +0000)]
printerr_reply() has never been used for as long as we've had this code in
our tree (13+ years). This is an excellent argument for aggressive use
of "static".
Fix several logic bugs in the previous IPv6 variable change and
re-add $ipv6_enable support for backward compatibility. From
UPDATING:
1. To use IPv6, simply define $ifconfig_IF_ipv6 like $ifconfig_IF
for IPv4. For aliases, $ifconfig_IF_aliasN should be used.
Note that both variables need the "inet6" keyword at the head.
Do not set $ipv6_network_interfaces manually if you do not
understand what you are doing. It is not needed in most cases.
$ipv6_ifconfig_IF and $ipv6_ifconfig_IF_aliasN still work, but
they are obsolete.
2. $ipv6_enable is obsolete. Use $ipv6_prefer and/or
"inet6 accept_rtadv" keyword in ifconfig(8) instead.
If you define $ipv6_enable=YES, it means $ipv6_prefer=YES and
all configured interfaces have "inet6 accept_rtadv" in the
$ifconfig_IF_ipv6. These are for backward compatibility.
3. A new variable $ipv6_prefer has been added. If NO, IPv6
functionality of interfaces with no corresponding
$ifconfig_IF_ipv6 is disabled by using "inet6 ifdisabled" flag,
and the default address selection policy of ip6addrctl(8)
is the IPv4-preferred one (see rc.d/ip6addrctl for more details).
Note that if you want to configure IPv6 functionality on the
disabled interfaces after boot, first you need to clear the flag by
using ifconfig(8) like:
ifconfig em0 inet6 -ifdisabled
If YES, the default address selection policy is set as
IPv6-preferred.
The default value of $ipv6_prefer is NO.
4. If your system need to receive Router Advertisement messages,
define "inet6 accept_rtadv" in $ifconfig_IF_ipv6. The rc(8)
scripts automatically invoke rtsol(8) when the interface becomes
UP. The Router Advertisement messages are used for SLAAC
(State-Less Address AutoConfiguration).
Make malloc(3) superpage aware. Specifically, if getpagesizes(3) returns
a large page size that is greater than malloc(3)'s default chunk size but
less than or equal to 4 MB, then increase the chunk size to match the large
page size.
Most often, using a chunk size that is less than the large page size is not
a problem. However, consider a long-running application that allocates and
frees significant amounts of memory. In particular, it frees enough memory
at times that some of that memory is munmap()ed. Up until the first
munmap(), a 1MB chunk size is just fine; it's not a problem for the virtual
memory system. Two adjacent 1MB chunks that are aligned on a 2MB boundary
will be promoted automatically to a superpage even though they were
allocated at different times. The trouble begins with the munmap(),
releasing a 1MB chunk will trigger the demotion of the containing superpage,
leaving behind a half-used 2MB reservation. Now comes the real problem.
Unfortunately, when the application needs to allocate more memory, and it
recycles the previously munmap()ed address range, the implementation of
mmap() won't be able to reuse the reservation. Basically, the coalescing
rules in the virtual memory system don't allow this new range to combine
with its neighbor. The effect being that superpage promotion will not
reoccur for this range of addresses until both 1MB chunks are freed at some
point in the future.
ed [Sat, 26 Sep 2009 15:26:32 +0000 (15:26 +0000)]
Add 256 color support.
It is quite inconvenient that if an application for xterm uses 256 color
mode, text suddenly starts to blink (because of ;5; in the middle).
We'd better just implement 256 color mode and add a conversion routine
from 256 to 8 color mode, which doesn't seem to be too bad in practice.
Remapping colors is done quite simple. If one of the channels is most
actively represented, primary colors are used. If two channels are most
actively represented, secondary colors are used. If all three channels
are equal (gray), it picks between black and white.
ed [Sat, 26 Sep 2009 15:07:11 +0000 (15:07 +0000)]
Properly get out of origin mode if the cursor has to move outside of it.
In some cases events may occur that move the cursor outside the
scrolling region while in origin mode, which is normally not possible.
Events like these include:
lindev(4) [1] is supposed to be a collection of linux-specific pseudo
devices that we also support, just not by default (thus only LINT or
module builds by default).
While currently there is only "/dev/full" [2], we are planning to see more
in the future. We may decide to change the module/dependency logic in the
future should the list grow too long.
This is not part of linux.ko as also non-linux binaries like kFreeBSD
userland or ports can make use of this as well.
Handle cases where virtual (GFS) vnodes are referenced when doing forced
unmount. In that case we cannot depend on the proper order of invalidating
vnodes, so we have to free resources when we have a chance.
PR: kern/139062
Reported by: trasz
MFC after: 3 days
- Don't depend on value returned by gfs_*_inactive(), it doesn't work
well with forced unmounts when GFS vnodes are referenced.
- Make other preparations to GFS for forced unmounts.
PR: kern/139062
Reported by: trasz
MFC after: 3 days
- Use x86bios_offset() instead of BIOS_PADDRTOVADDR() macro.[1]
- Clear all registers before calling real mode interrupt handlers as we did
for dpms and vesa and re-enable the function as it should be fixed by this.
- Tidy up register access. For example, when we call INT 0x15, AH=0xc0,
we used to initialize AX=0xc000 to clear AL at the same time but it is
very confusing. We don't have to do this any more because we are explicitly
clearing all registers now.
- Check size of system configuration table although it is almost always 8.
This is to make sure we are not reading some random low physical memory.
Hopefully it is just zero in that case. :-)
- Fix some style nits and add more comments.