When running as a guest operating system, the FreeBSD kernel must assume
that the virtual machine monitor has enabled machine check exceptions.
Unfortunately, on AMD Family 10h processors the machine check hardware
has a bug (Erratum 383) that can result in a false machine check exception
when a superpage promotion occurs. Thus, I am disabling superpage
promotion when the FreeBSD kernel is running as a guest operating system
on an AMD Family 10h processor.
framebuffer driver, etc. work on FreeBSD/amd64.
A significant amount of improvements were done by jkim@ during the recent
months to make vesa(4) work better, over the initial code import. This
work is based on OpenBSD's x86emu implementation and contributed by
paradox <ddkprog yahoo com> and swell.k at gmail com.
Hopefully I have stolen all their work to 8-STABLE :)
All bugs in this commit are mine, as usual.
Allow user programs to execute mfpvr instructions. Linux allows this, and
some math-related software like GMP expects to be able to use it to pick
a target appropriately.
Reported by: Jakob van Santen <vansanten at wisc dot edu>
Enable NETIF_OPEN_CLOSE_ONCE on PowerPC OFW. This fixes netbooting on
PowerPC Book-S hardware, which had been broken for a very long time.
Submitted by: Andreas Tobler
GPIB overhaul, part #2: make the tnt4882 driver work with the newer
TNT5004 IC. This involved a major rewrite of a number of things, as
this chip no longer supports the NAT7210 legacy mode but requires the
host to use the (more modern) FIFO mode.
In theory, this also ought to work on the older TNT4882C chip. I'll
probably add this as optional support (perhaps by a device.hints flag)
later on. By now, FIFO mode is *only* activates iff a TNT5004 chip
has been detected (where the old code didn't work at all), while
everything else is supposed to use the old code.
Make sure that FTS_COMFOLLOW is not set when the -P option is in effect.
Otherwise the -i option will show the inode number of the referenced file
for symbolic links given on the command line. Similarly, the file color
was printed according to the link target in colorized output.
PR: bin/102394
Document the interaction between /etc/devfs.conf and
/etc/defaults/devfs.conf
PR: docs/117308
Submitted by: Mel <mel.xyzzy rachie.is-a-geek.net> (partially)
Since the existing implementation searches ':' backward, a path which
includes ':' could not be mounted. You can now mount such path by
enclosing an IP address by '[]'.
Though we should change to search ':' forward, it will break
'ipv6addr:path' which is currently working. So, it still searches ':'
backward, at least for now.
Add rc.d script for the rtsold(8) daemon.
The rtsol(8) handles just one RA then exit. So, the OtherConfig flag
may not be handled well by rtsol(8) in the environment where there are
multiple RA servers on the segment. In such case, rtsold(8) will be
your friend.
- Add idempotency guards so the structures can be used in other utilities.
- Update bpb structs with reserved fields.
- In direntry struct join deName with deExtension. Although a
fix was attempted in the past, these fields were being overflowed,
Now this is consistent with the spec, and we can now share the
WinChksum code with NetBSD.
These fixes correct a problem in the file system that treats large
inode numbers as negative rather than unsigned. For a default
(16K block) file system, this bug began to show up at a file system
size above about 16Tb.
These fixes also update newfs to ensure that it will never create a
filesystem with more than 2^32 inodes.
They also update libufs, tunefs, and growfs so that they properly
handle inode numbers as unsigned.
Reported by: Scott Burns, John Kilburg, and Bruce Evans
Followup by: Jeff Roberson
PR: 133980
Add interface description capability as inspired by OpenBSD. Thanks for
rwatson@, jhb@, brooks@ and others for feedback to the old implementation!
Sponsored by: iXsystems, Inc.
On FreeBSD, time_t is 64-bit for all platforms except i386 and powerpc,
where the type is 32-bit. ZFS can handle 64-bit timestamp internally
but zfs_setattr() would check if the time value can fit, we change the
checking macros to match 64-bit timestamp if the platform supports it.
This change has some downsides like, while you can import zfs on 32-bit
platforms, the timestamp would overflow if they are out of the range.
This fixes the Y2.038K issue on platforms using 64-bit timestamps.
Reviewed by: pjd
Replace the static NGROUPS=NGROUPS_MAX+1=1024 with a dynamic
kern.ngroups+1. kern.ngroups can range from NGROUPS_MAX=1023 to
somewhere in the neighborhood of INT_MAX/4 one a system with sufficent
RAM and memory bandwidth. Given that the Windows group limit is
1024, this range should be sufficient for most applications
r202342:
Only allocate the space we need before calling kern_getgroups instead
of allocating what ever the user asks for up to "ngroups_max + 1". On
systems with large values of kern.ngroups this will be more efficient.
The now redundant check that the array is large enough in
kern_getgroups() is deliberate to allow this change to be merged to
stable/8 without breaking potential third party consumers of the API.
Fix mptutil's method for locating disk devices attached to a specific
mpt(4) controller. Previously, the code assumed that multiple match
patterns provided to an XPT_DEV_MATCH request were ANDed together.
Instead, they are ORed. Instead, to match peripherals for a specific bus,
one query needs to be performed to lookup the path ID of the bus. A second
query can then be performed matching peripherals attached to that path.
This approach also makes the code a bit cleaner as the returned match
results do not mix buses and perphierals.
- Don't emit a warning in 'show adapter' if the IOC2 or IOC6 pages are not
present. mpt(4) controllers that do not support RAID do not have an IOC6
page, for example.
- Correct a check for a missing page error in a debug function.
Fixes for ls(1) long format (-l) output:
- Allow -h option to work if the listing contains at least one device
file.
- Align major and minor device numbers correctly to the size field.
PR: bin/125678
Teach the (gpt)zfsboot and zfsloader raidz code to use its buffers
more efficiently.
Before this patch, in the worst case memory use would increase
exponentially on the number of drives in the raidz vdev.
Submitted by: Matt Reimer <mattjreimer@gmail.com>
Sponsored by: VPOP Technologies, Inc.
Silence from: dfr
- Return EAFNOSUPPORT instead of EINVAL for unsupported address family,
this matches the Linux behavior.
- Check if we have sufficient space allocated for socket structure, which
fixes a buffer overflow when wrong length is being passed into the
emulation layer. [1]
PR: kern/138860
Submitted by: Mateusz Guzik <mjguzik gmail com>
Reported by: Alexander Best [1]
module so it's not present twice.
- Move nfs_realign() from the NFS client to the shared NFS code and
remove the NFS server version in order to reduce code duplication.
The shared version now uses a second parameter how, which is passed
on to m_get(9) and m_getcl(9) as the server used M_WAIT while the
client requires M_DONTWAIT, and replaces the the previously unused
parameter hsiz.
- Change nfs_realign() to use nfsm_aligned() so as with other NFS code
the alignment check isn't actually performed on platforms without
strict alignment requirements for performance reasons because as the
comment suggests unaligned data only occasionally occurs with TCP.
- Change fha_extract_info() to use nfs_realign() with M_DONTWAIT rather
than M_WAIT because it's called with the RPC sp_lock held.
