Andrew Turner [Thu, 24 Dec 2015 09:40:29 +0000 (09:40 +0000)]
Ads support to the xhci pci attachment to use MSI-X interrupts when
available. As with MSI interrupts these can be disabled by setting
hw.usb.xhci.msix to 0 in the loader.
MSI-X interrupts are needed on some hardware, for example the Cavium
ThunderX only supports them, and with this we don't fall back to polling.
John Baldwin [Thu, 24 Dec 2015 06:22:41 +0000 (06:22 +0000)]
Fix remaining direct tests of the carry flag in the v86 %eflags via a
magic number to use V86_CY() instead. These should have been fixed as
part of the cleanup in r226746 but were missed.
The md5 sums of the object files were unchanged, so there should be no
functional change.
Justin Hibbits [Thu, 24 Dec 2015 04:30:15 +0000 (04:30 +0000)]
Extend Book-E to support >4GB RAM
Summary:
With some additional changes for AIM, that could also support much
larger physmem sizes. Given that 32-bit AIM is more or less obsolete, though,
it's not worth it at this time.
Navdeep Parhar [Thu, 24 Dec 2015 02:35:59 +0000 (02:35 +0000)]
MFP r291227, r291228, and r292618.
r291227:
s/is->is_conn/ic to shorten things a bit.
r291228:
Do not generate PDUs with payload greater than max_data_segment_length.
It is perhaps preferable to have a separate limit for send instead of
reusing the receive limit. I'll discuss with trasz@ and mav@ before
pulling this into head.
John Baldwin [Wed, 23 Dec 2015 21:41:42 +0000 (21:41 +0000)]
Move shared variables from {amd64,i386}/initcpu.c to x86/identcpu.c.
While here, move the common bits of <machine/cputypes.h> to
<x86/cputypes.h> as well.
Enji Cooper [Wed, 23 Dec 2015 18:48:32 +0000 (18:48 +0000)]
Increase the timeout for resolv_test from the default (300 seconds) to
450 seconds
This is required on slower network connections, and on older releases
(stable/10 seems to be slower as far as name resolution goes.. not sure
why yet).
Remove an outdated comment in the Makefile from when I was working on
this code over a year ago on github
Andrew Turner [Wed, 23 Dec 2015 17:54:19 +0000 (17:54 +0000)]
Be stricter on which functions we can probe with FBT. We now only check the
first instruction to see if it's either a pushm with lr, or a sub with sp.
The former is the common case, with the latter used with va_args.
This removes 12 probes. These are all hand-written assembly, with a few C
functions with no stack usage.
Submitted by: Howard Su <howard0su@gmail.com>
Differential Revision: https://reviews.freebsd.org/D4419
Andrew Turner [Wed, 23 Dec 2015 15:22:44 +0000 (15:22 +0000)]
Increase the size and alignment of the setjmp buffer. This will allow for
possible future CPU extentions with larger registers.
jmp_buf's size and alignment are baked into the ABI of third party libraries
and thus are very hard to change later so it is best to waste a small amount
of space now.
Bjoern A. Zeeb [Wed, 23 Dec 2015 10:23:47 +0000 (10:23 +0000)]
Remove superfluous return (1) missed in r292601.
Reported by: Matthew D. Fuller (fullermd over-yonder.net),
Kevin Bowling (kevin.bowling kev009.com)
MFC after: 13 days
X-MFC with: r292601
Sponsored by: The FreeBSD Foundation
Enji Cooper [Wed, 23 Dec 2015 09:59:31 +0000 (09:59 +0000)]
Move mac_bsdextended check up before running the test_libugidfw_strings testcases
I realize that these tests could be run before mac_bsdextended is loaded, but
it would overcomplicate things to special case handle the testcases before doing
the mac_bsdextended(4) feature check
The testcases will be split up so they can be run separately in the near future
Warner Losh [Wed, 23 Dec 2015 06:49:18 +0000 (06:49 +0000)]
Copy all the dtb files that we build as part of the kernel build from
boot/dtb to the fat partition. They seem to be needed.
Create an empty uEnv.txt file
Check for packet_length is greater than 60 bytes as well as packet_length is
greater than len_on_bd, before invoking the routine to handle jumbo over SGL
(bxe_service_rxsgl()).
Add counters for number of jumbo_over_SGL packets (rx_bxe_service_rxsgl) and
erroneous jumbo_over_SGL packets (rx_erroneous_jumbo_sge_pkts)
Make it possible for the cdevsw d_close() driver method to detect last
close and close due to revoke(2)-like operation.
A new FLASTCLOSE flag indicates that this is last close. FREVOKE is
set for revokes, and FNONBLOCK is also set, same as is already done
for VOP_CLOSE() call from vgonel().
The flags reuse user open(2) flags which are never stored in f_flag,
to not consume bit space in the ABI visible way. Assert this with the
static check.
Requested and reviewed by: bde
Sponsored by: The FreeBSD Foundation
MFC after: 2 weeks
John Baldwin [Tue, 22 Dec 2015 20:33:49 +0000 (20:33 +0000)]
Move the mkioctls script to libsysdecode and use it to generate a
sysdecode_ioctlname() function. This function matches the behavior
of the truss variant in that it returns a pointer to a string description
for known ioctls. The caller is responsible for displaying unknown
ioctl requests. For kdump this meant moving the logic to handle unknown
ioctl requests out of the generated function and into an ioctlname()
function in kdump.c instead.
Ed Schouten [Tue, 22 Dec 2015 18:12:11 +0000 (18:12 +0000)]
Let tsearch()/tdelete() use an AVL tree.
The existing implementations of POSIX tsearch() and tdelete() don't
attempt to perform any balancing at all. Testing reveals that inserting
100k nodes into a tree sequentially takes approximately one minute on my
system.
Though most other BSDs also don't use any balanced tree internally, C
libraries like glibc and musl do provide better implementations. glibc
uses a red-black tree and musl uses an AVL tree.
Red-black trees have the advantage over AVL trees that they only require
O(1) rotations after insertion and deletion, but have the disadvantage
that the tree has a maximum depth of 2*log2(n) instead of 1.44*log2(n).
My take is that it's better to focus on having a lower maximum depth,
for the reason that in the case of tsearch() the invocation of the
comparator likely dominates the running time.
This change replaces the tsearch() and tdelete() functions by versions
that create an AVL tree. Compared to musl's implementation, this version
is different in two different ways:
- We don't keep track of heights; just balances. This is sufficient.
This has the advantage that it reduces the number of nodes that are
being accessed. Storing heights requires us to also access all of the
siblings along the path.
- Don't use any recursion at all. We know that the tree cannot 2^64
elements in size, so the height of the tree can never be larger than
96. Use a 128-bit bitmask to keep track of the path that is computed.
This allows us to iterate over the same path twice, meaning we can
apply rotations from top to bottom.
Inserting 100k nodes into a tree now only takes 0.015 seconds. Insertion
seems to be twice as fast as glibc, whereas deletion has about the same
performance. Unlike glibc, it uses a fixed amount of memory.
I also experimented with both recursive and iterative bottom-up
implementations of the same algorithm. This iterative top-down version
performs similar to the recursive bottom-up version in terms of speed
and code size.
For some reason, the iterative bottom-up algorithm was actually 30%
faster for deletion, but has a quadratic memory complexity to keep track
of all the parent pointers.
Andrew Turner [Tue, 22 Dec 2015 17:18:40 +0000 (17:18 +0000)]
Don't adjust the program counter to an invalid address after reaching a
breakpoint. The value doesn't need to be adjusted as it is already
correctly returned from the kernel.
This allows lldb to set breakpoints, and stop on them, however more work
is needed, for example single stepping fails to stop.
Bjoern A. Zeeb [Tue, 22 Dec 2015 15:03:45 +0000 (15:03 +0000)]
If vnets are torn down while ifconfig runs an ioctl to say, destroy an
epair(4), we may hit if_detach_internal() without holding a lock and by
the time we aquire it the interface might be gone.
We should not panic() in this case as it is our fault for not holding
the lock all the way. It is not ideal to return silently without error
to user space, but other callers will all ignore the return values so
do not change the entire KPI for little benefit for now.
The ifp will be dealt with one way or another still.
Sponsored by: The FreeBSD Foundation
MFC after: 2 weeks
Reviewed by: gnn
Differential Revision: https://reviews.freebsd.org/D4529
Bjoern A. Zeeb [Tue, 22 Dec 2015 15:00:04 +0000 (15:00 +0000)]
If bootverbose is enabled every vnet startup and virtual interface
creation will print extra lines on the console. We are generally not
interested in this (repeated) information for each VNET. Thus only
print it for the default VNET. Virtual interfaces on the base system
will remain printing information, but e.g. each loopback in each vnet
will no longer cause a "bpf attached" line.
Sponsored by: The FreeBSD Foundation
MFC after: 2 weeks
Reviewed by: gnn
Differential Revision: https://reviews.freebsd.org/D4531
Bjoern A. Zeeb [Tue, 22 Dec 2015 14:53:19 +0000 (14:53 +0000)]
Since r256624 we've been leaking routing table allocations
on vnet enabled jail shutdown. Call the provided cleanup
routines for IP versions 4 and 6 to plug these leaks.
Sponsored by: The FreeBSD Foundation
MFC atfer: 2 weeks
Reviewed by: gnn
Differential Revision: https://reviews.freebsd.org/D4530
Guard against the same process being both CUSE server and client at
the same time. This can easily lead to a deadlock when destroying the
character devices nodes.
Ian Lepore [Tue, 22 Dec 2015 03:07:38 +0000 (03:07 +0000)]
Set env vars from values on the efi loader command line.
Examine each cmdline arg and if it contains an '=' convert it to ascii and
pass it to putenv(). This allows var=value settings to come in on the
command line.
This will allow overriding dhcp server-provided data in loader(8), as
discussed in PR 202098
Ian Lepore [Tue, 22 Dec 2015 03:02:52 +0000 (03:02 +0000)]
Allow dhcp/bootp server-provided values to be overriden from environment
variables in loader(8) and other libstand applications.
Sometimes a dhcp server provides incorrect information along with the IP
address. It would be useful to have a way to override this with
locally-supplied information, such as command line parameters passed from a
prior-stage bootloader. This change allows pre-existing env vars to take
precedence over values delivered by the dhcp or bootp server.
The bootp/dhcp code in libstand automatically creates environment variables
from the data provided by the server (dhcp.root-path, dhcp.domain-name,
etc). It also transcribes the values to some global variables such as
'rootpath' and 'hostname'.
This change does two things:
When adding dhcp.* vars to the environment, don't replace existing
vars/values.
When setting the global vars rootpath and hostname, use the
dhcp.root-path and dhcp.host-name env var values if they exist.
This allows the platform-specific part of loader(8) to obtain override
values in some platform-specific way and store them in the environment
before opening the network device. The set of values that can be overriden
is currently limited to just string options. The values that are delivered
as binary data are things that probably shouldn't be overridden (IP,
netmask, gateway, etc).
The original patch this evolved from was submitted by martymac@
Ian Lepore [Tue, 22 Dec 2015 00:53:19 +0000 (00:53 +0000)]
Fix the detection of IO/memory space changing across busses when the bus
is not pci (and thus where, ironically, the whole situation is meaningless).
This was not an error in the original code, it was introduced during my
refactoring to commonize the routine. A small change a few lines above
drove the need to make this change, and the error didn't show up on the
platforms I initially tested with.
Enji Cooper [Tue, 22 Dec 2015 00:43:22 +0000 (00:43 +0000)]
Don't dump core files with lib/libc/ssp/ssp_test and lib/libc/gen/assert_test
The default `sysctl kern.corefile` value is compatible with `kyua test` (FreeBSD
will dump to the current directory). If it's set to an absolute path however,
`kyua test` will not be able to clean up the corefiles after the fact
The corefiles have little value when testing the behavior of feature behavior,
so just disable corefile generation
Ian Lepore [Mon, 21 Dec 2015 23:47:49 +0000 (23:47 +0000)]
Include machine/_bus.h so that bus_space_[tag|handle]_t will be available.
It appears that all platforms except aarch64 are getting the file via
various header pollution, and ensuring _bus.h is included before any
openfirmware headers in every consumer of ofw/fdt stuff seems like more of
a career path than a task, so I'm taking this easy way out.
Ed Maste [Mon, 21 Dec 2015 22:42:03 +0000 (22:42 +0000)]
boot1.efi: show EFI error number, not full status value
EFI return values set the high bit to indicate an error. The log
messages changed here are printed only in the case of an error,
so including the error bit is redundant. Also switch to decimal to
match the error definitions (in sys/boot/efi/include/efierr.h).
MFC after: 1 week
Sponsored by: The FreeBSD Foundation
Add i.MX 6 IPU driver and enable it in IMX6 config
Current functionality is somewhat limited: driver assumes that there
is only one active IPU unit (IPU1) and that video output is DI0 and
video mode is 1024x768. For more advanced functionality driver requires
proper clock management which is work in progress. At the moment driver
assumes that pixel clock is configured by u-boot for 1026x768 mode.
Alan Somers [Mon, 21 Dec 2015 22:19:22 +0000 (22:19 +0000)]
Fix "mount -a" for NFS and ZFS filesystems with shared mountpoints
sbin/mount.c
Check whether an fstab entry has the same fstype as a mounted
filesystem before declaring it to be mounted. This will allow NFS
filesystems that share a mountpoint with a local filesystem to be
automatically mounted at boot.
This is not such an unusual situation. For example, if somebody uses
the standard installer with a ZFS root, he'll get a /usr/home
filesystem, even though he may choose to mount /usr/home over NFS.
Enji Cooper [Mon, 21 Dec 2015 21:24:03 +0000 (21:24 +0000)]
Integrate tools/regression/mac/mac_bsdextended and
tools/regression/mac/mac_portacl into the FreeBSD test suite as
tests/sys/mac/bsdextended and tests/sys/mac/portacl, respectively
Enji Cooper [Mon, 21 Dec 2015 21:15:23 +0000 (21:15 +0000)]
Make the mac_portacl testcases work / more robust
- A trap(1) call has been added to the test scripts to better
ensure that the tests do a better job at trying to restore the
test host state at the end of the tests (if the test was
interrupted before it would leave the system in an odd state,
potentially making the test results for subsequent runs
non-deterministic).
- Add root user checks
- Fix nc(1) usage:
-- -o is deprecated
-- Using `-w 10` will make the call timeout after 10 seconds so it
doesn't block indefinitely
- Use local variables
- Be more terse in the error messages
- Parameterize out "127.0.0.1"