hyperv/hn: Suspend and resume the backend properly upon MTU change.
Suspend:
- Prevent the backend from being touched on TX path.
- Clear the RNDIS RX filter, and wait for RX to drain.
- Make sure that NVS see the chimney sending buffer and RXBUF
disconnection, before unlink these buffers from the channel.
Resume:
- Reconfigure the RNDIS filter.
- Allow TX path to work on the backend.
- Kick start the TX eof task, in case the OACTIVE is set.
This fixes various panics, when the interface has traffic and MTU
is being changed.
MFC after: 1 week
Sponsored by: Microsoft
Differential Revision: https://reviews.freebsd.org/D8046
loader command interpreter should reset command_errmsg
The command interpreter does leave command_errmsg as is after printing its
content, assuming the next command will reset it in bf_command(). However,
in case the forth native word is defined as builtin, the bf_command is not
used and forth words will also end up the command_errmsg content printed.
Since command_errmsg is pointer to actual error message, which can be static
read only string, we can not just set *command_errmsg = '\0', instead we need
to reset the pointer itself.
Summary:
This enables some features of the DIU, using a static configuration,
specified either via a 'edid' property on the 'display' FDT node, or a
'video-mode' environment variable (bootarg). 'video-mode' was chosen because it
matches u-boot's naming, so it can be set with:
setenv bootargs video-mode=${video-mode}
at the u-boot CLI.
Mouse cursor is not supported currently, as a hardware cursor is not supported
by framebuffer VT yet. Currently it only supports a 32bpp ARGB (actually BGRA)
format, and only a single composite plane, at up to 1280x1024.
Elantech trackpads are found in some laptops like the Asus UX31E. They
are "synaptics compatible" but use a slightly different protocol.
Elantech hardware support is not enabled by default and just like
Synaptic or TrackPoint devices it should be enabled by setting
tunable, in this case hw.psm.elantech_support, to non-zero value
For machines which support PCID but not have INVPCID instruction,
i.e. SandyBridge and IvyBridge, correct a race between pmap_activate()
and invltlb_pcid_handler().
Reported by and tested by: Slawa Olhovchenkov <slw@zxy.spb.ru>
MFC after: 1 week
hiren [Mon, 26 Sep 2016 10:13:58 +0000 (10:13 +0000)]
In sendit(), if mp->msg_control is present, then in sockargs() we are allocating
mbuf to store mp->msg_control. Later in kern_sendit(), call to getsock_cap(),
will check validity of file pointer passed, if this fails EBADF is returned but
mbuf allocated in sockargs() is not freed. Fix this possible leak.
Portability changes:
1. macOS nor Linux have MAP_NOCORE nor MAP_NOSYNC. Define as 0.
2. macOS doesn't have SEEK_DATA nor SEEK_HOLE. Define as -1
so that lseek will return -1 (with errno set to EINVAL).
3. gcc correctly warns that error is assigned but not used in
image_copyout_region(). Fix by returning on the first error.
Avoid depending on the <sys/endian.h> header for le*enc and be*enc.
Not only is the header unportable, the encoding/decoding functions
are as well. Instead, duplicate the handful of small inlines we
need into a private header called endian.h.
Aside: an alternative approach is to move the encoding/decoding
functions to a separate system header. While the header is still
nonportable, such an approach would make it possible to re-use the
definitions by playing games with include paths. This may be the
preferred approach if more (build) utilities need this. This
change does not preclude that. In fact, it makes it easier.
Eliminate the use of EDOOFUS. The error code was used to signal
programming errors, but is really a poor substitute for assert.
And less portable as well.
Replace the use of linker sets with constructors for both the
formats and schemes. Formats and schemes are registered at
runtime now, rather than collected at link time.
(1) Print the default segment %ss before adresses relative to %bp.
This is too cluttered for me, but so is printing some other default
prefixes, and this is a reasonable reminder that %ss is quite
likely to be different from %ds in 16-bit mode.
db_disasm still handles prefixes poorly, by trying to discard
redundant ones. This loses information, and sometimes the result
is wrong or misleading.
Clean up nearby initializations and dead code.
(2) Fix decoding of operand and address size prefixes in 16-bit mode.
They reverse the default in all modes.
Obtained from: (1) is partly from r1.4 (2003/11/08) in DFlyBSD (?)
Determine the operand/address size of %cs in a new function
db_segsize().
Use db_segsize() to set the default operand/address size for
disassembling. Allow overriding this with the "alternate" display
format /I. The API of db_disasm() should be debooleanized to pass a
more general request (amd64 needs overrides to sizes of 16, 32, and
64, but this commit doesn't implement anything for amd64 since much
larger changes are needed to restore the amd64 disassmbler's support
for non-default sizes).
Fix db_print_loc_and_inst() to ask for the normal format and not the
alternate in normal operation.
This is most useful for vm86 mode, but also works for 16-bit protected
mode.
Use db_segsize() to avoid trying to print a garbage stack trace if %cs
is 16 bits. Print something like the stack trace termination message
for a trap boundary instead.
Document that the alternate format is now useful on i386.
Fix vm86 initialization, part 3 of 2 and a half. (Actually, just fix
early printfs and debugging of vm86 initialization and some other early
initialization in some cases.) Add an option debug.late_console (with
default 1=off) to move console and kdb initialization back where it was.
Do the same for amd64 although there is no vm86 there.
On my test system, debug.late_console=0 works for the syscons, sio and
uart console drivers on amd64 and i386, and for vt on i386 but not on
amd64.
The early printfs fixed by debug.late_console=0 are:
- on i386, the message about lost memory above 4G
- with -v in otherwise normal use, about 20 printfs for SMAP
- other debugging messages for memory sizing. Mostly under -v and
not printed in normal use.
Document in a comment how much earlier the initialization and early
printf()s can be. That is very early for the console. Not much more
than curthread is needed. kdb use obviously needs to be not so early,
since it needs IDT initialization and that is done relatively late
for convenience and historical reasons.
andrew [Sun, 25 Sep 2016 07:48:08 +0000 (07:48 +0000)]
Add the start of a GENERIC armv6 kernel config. This supports the Allwinner
SMP SoCs and qemu virt. Further SoCs can be supported if they support the
PLATFORM, PLATFORM_SMP, and MULTIDELAY options.
cxgbe(4): Use the port's top speed to figure out whether it is "high
speed" or not (for the purpose of calculating the number of queues etc.)
This does the right thing for 25Gbps and 100Gbps ports.
When MAKEOBJDIRPREFIX points to a case-insensitive file system, the
build can break when different source files create the same target
files (case-insensitivity speaking). This is the case for object
files compiled with -fpic and shared libraries. The former uses
an extension of ".So", and the latter an extension ".so". Rename
shared object files from *.So to *.pico to match what NetBSD does.
Update local variable 'block' after calling capacity_resize(),
otherwise format_resize(), which is called right after, isn't
getting the current/actual image size. Rather than rounding up,
format_resize() could end up truncating the size and we don't
allow that by design.
When MAKEOBJDIRPREFIX points to a case-insensitive file system, the
build can break when different source files create the same target
files (case-insensitivity speaking). This is the case for object
files compiled with -fpic and shared libraries. The former uses
an extension of ".So", and the latter an extension ".so". Rename
shared object files from *.So to *.pico to match what NetBSD does.
cxgbe(4): Support SIOGIFXMEDIA so that ifconfig displays correct media
for 25Gbps and 100Gbps ports. This should have been part of r305713,
which is when the driver first started reporting extended media types.
fix vnode lock assertion for extended attributes directory
Background. In ZFS a file with extended attributes has a special
directory associated with it where each extended attribute is a file.
The attribute's name is a file name and its value is a file content.
When the ownership of a file with extended attributes is changed, ZFS
also changes ownership of the special directory. This is where the bug
was hit.
The bug was introduced in r209158.
Nota bene. ZFS vnode locks are typically acquired before
z_teardown_lock (i.e., before ZFS_ENTER). But this is not the case for
the vnodes that represent the extended attribute directory and files.
Those are always locked after ZFS_ENTER. This is confusing and fragile.
PR: 212702
Reported by: Christian Fuss to FreeNAS
Tested by: mav
MFC after: 1 week
Fragmented UDP and ICMP packets were corrupted if a firewall with reassembling
feature (like pf'scrub) is enabled on the bridge. This patch fixes corrupted
packet problem and the panic (triggered easly with low RAM) as explain in PR
185633.
bridge_pfil and bridge_fragment relationship:
bridge_pfil() receive (IN direction) packets and sent it to the firewall The
firewall can be configured for reassembling fragmented packet (like pf'scrubing)
in one mbuf chain when bridge_pfil() need to send this reassembled packet to the
outgoing interface, it needs to re-fragment it by using bridge_fragment()
bridge_fragment() had to split this mbuf (using ip_fragment) first then
had to M_PREPEND each packet in the mbuf chain for adding Ethernet
header.
But M_PREPEND can sometime create a new mbuf on the begining of the mbuf chain,
then the "main" pointer of this mbuf chain should be updated and this case is
tottaly forgotten. The original bridge_fragment code (Revision 158140,
2006 April 29) came from OpenBSD, and the call to bridge_enqueue was
embedded. But on FreeBSD, bridge_enqueue() is done after bridge_fragment(),
then the original OpenBSD code can't work as-it of FreeBSD.
'compat' can never be NULL, because the compatible check loop ends when
compat->ocd_str is NULL. This causes ds1307 to attach to any unclaimed i2c
device.
bhnd(4): Implement common API for IOST/IOCTL register access and core reset
- Added bhnd(4) bus APIs for per-core ioctl/iost register access.
- Updated reset/suspend bhnd(4) APIs for compatibility with ioctl/iost
changes.
- Implemented core reset/suspend support for both bcma(4) and siba(4).
- Implemented explicit release of all outstanding PMU requests at the bus
level when putting a core into reset.
Approved by: adrian (mentor, implicit)
Differential Revision: https://reviews.freebsd.org/D8009
adrian [Sat, 24 Sep 2016 01:21:42 +0000 (01:21 +0000)]
[iwm] fix default antenna selection a bit; fix 5GHz rate control handling.
* Don't do the antenna switching when setting up the rate table - we don't
take into account whether it's an active antenna or not (eg shared with BT.)
I'll look into this a bit more later.
* The default antenna is still 1, I'll look into that a bit more later.
(So no, this doesn't fix it for Larry who needs ANT_B to be active, not
ANT_A.)
* My changes to the rate control setup used the wrong method for finding
a suitable rate, which led to 1M CCK frames being queued for 11a operation.
This is .. sub-optimal. Change the rate control lookup path to use
the global table instead of the per-node table, which won't be setup
until the node rate table is setup (which happens way too late in this
game.)
cxgbe(4): Fix netmap with T6, which doesn't encapsulate SGE_EGR_UPDATE
message inside a FW_MSG. The base NIC already deals with updates in
either form.
If the kernel is not compiled with the CAPABILITIES kernel options
fget_unlocked doesn't return the sequence number so fd_modify will
always report modification, in that case we got infinity loop.
fget_cap_locked returns a referenced file, but the fgetvp_rights does
not need it. Instead, due to the filedesc lock being held, it can
ref the vnode after the file was looked up.
Fix up fget_cap_locked to be consistent with other _locked helpers and not
ref the file.
Add a table of vnode locks and use them along with bucketlocks to provide
concurrent modification support. The approach taken is to preserve the
current behaviour of the namecache and just lock all relevant parts before
any changes are made.
Lookups still require the relevant bucket to be locked.
Re-check the systrace probe ID before calling dtrace_probe().
Otherwise there exists a narrow window during which a syscall probe can be
disabled and cause a concurrently-running thread to call dtrace_probe()
with an invalid probe ID.
amdsbwd, intpm: unify bits specific to AMD chipsets (FCHs, southbridges)
AMD chipsets have proprietary mechanisms for dicovering resources.
Those resources are not discoverable via plug-and-play mechanisms
like PCI configuration registers or ACPI.
For this reason a chipset-specific knowledge of proprietary registers
is required.
At present there are two FreeBSD drivers that require the proprietary
resource discovery. One is amdsbwd which is a driver for the watchdog
timer in the AMD chipsets. The other is intpm SMBus driver when it
attaches to the newer AMD chipsets where the resources of the SMBus HBA
are not described in the regular PCI way.
In both cases the resources are discovered by accessing AMD PMIO space.
Thus, many definitions are shared between the two drivers.
This change puts those defintions into a common header file.
As an added benefit, intpm driver now supports newest FCHs built into
AMD processors of Family 15h, models 70h-7Fh and Family 16h, models
30h-3Fh.
Fix regression from r297400, which truncates headers in case of low socket
buffer and put a small optimization for low socket buffer case:
- Do not hack uio_resid, and let m_uiotombuf() properly take care of it. This
fixes truncation of headers at low buffer.
- If headers ate all the space, jump right to the end of the cycle, to
avoid doing single page I/O and allocating zero length mbuf.
- Clear hdr_uio only if space is positive, which indicates that all uio
was copied in.