dim [Wed, 12 Nov 2014 20:01:10 +0000 (20:01 +0000)]
Pull in r221709 from upstream llvm trunk (by Frédéric Riss):
Totally forget deallocated SDNodes in SDDbgInfo.
What would happen before that commit is that the SDDbgValues associated with
a deallocated SDNode would be marked Invalidated, but SDDbgInfo would keep
a map entry keyed by the SDNode pointer pointing to this list of invalidated
SDDbgNodes. As the memory gets reused, the list might get wrongly associated
with another new SDNode. As the SDDbgValues are cloned when they are transfered,
this can lead to an exponential number of SDDbgValues being produced during
DAGCombine like in http://llvm.org/bugs/show_bug.cgi?id=20893
Note that the previous behavior wasn't really buggy as the invalidation made
sure that the SDDbgValues won't be used. This commit can be considered a
memory optimization and as such is really hard to validate in a unit-test.
This should fix abnormally large memory usage and resulting OOM crashes
when compiling certain ports with debug information.
pfg [Wed, 12 Nov 2014 16:23:56 +0000 (16:23 +0000)]
ifdef ext2_print_inode which is not really used.
ext2_print_inode is not really used but it was nice to
have for initial development work. #ifdef it under a
new EXT2FS_DEBUG knob so that we don't spend time
compiling it.
ian [Wed, 12 Nov 2014 02:37:27 +0000 (02:37 +0000)]
Bugfixes for the imx5/imx6 iomux fdt_pinctrl driver.
I originally overlooked a couple flag bits defined in the fdt binding docs.
One flag suppresses the pad configuration (pullup/pulldown/etc). The other
one requires that the SION (set input on) flag be set in the mux register.
Also, it appears from the data involved that if the input register
address in the config tuple is zero, there is no input configuration. The
old code was writing to register zero, which contains a collection of misc
control bits (having nothing to do with input configuration) that probably
shouldn't get overwritten arbitrarily. The bindings doc doesn't explictly
mention this.
marcel [Wed, 12 Nov 2014 00:10:27 +0000 (00:10 +0000)]
SEEK_DATA has interesting behaviour for sparse files on ZFS. A sparse file
with 128K of random data and truncated to 800K can have SEEK_DATA return -1
when given an offset of 128K. On UFS, the SEEK_DATA returns 800K (the size
of the file). SEEK_HOLE on ZFS seems to behave the same as UFS.
To handle this, map -1 to the size of the file (`end') when lseek returns
this for either SEEK_HOLE or SEEK_DATA. When sparse files are not supported
by the file system both `hole' and `data' will now be equal to `end' and we
will treat the entire file as data. This way, the -1 return for SEEK_DATA
on ZFS will end up doing the right thing.
grehan [Tue, 11 Nov 2014 22:03:11 +0000 (22:03 +0000)]
Fix incorrect reading of 32-bit modinfo by 64-bit loaders.
The various structures in the mod_metadata set of a FreeBSD kernel and
modules contain pointers. The FreeBSD loader correctly deals with a
mismatch in loader and kernel pointer size (e.g. 32-bit i386/ppc
loader, loading 64-bit amd64/ppc64 kernels), but wasn't dealing with
the inverse case where a 64-bit loader was loading a 32-bit kernel.
Reported by: ktcallbox@gmail.com with a bhyve/i386 and ZFS root install
Differential Revision: https://reviews.freebsd.org/D1129
Reviewed by: neel, jhb
MFC after: 1 week
glebius [Tue, 11 Nov 2014 20:32:46 +0000 (20:32 +0000)]
Remove SF_KQUEUE code. This code was developed at Netflix, but was not
ever used. It didn't go into stable/10, neither was documented.
It might be useful, but we collectively decided to remove it, rather
leave it abandoned and unmaintained. It is removed in one single
commit, so restoring it should be easy, if anyone wants to reopen
this idea.
jhb [Tue, 11 Nov 2014 19:44:59 +0000 (19:44 +0000)]
Move NFS and TFTP filesystems before the synthetic filesystems (bzip,
gzip, and split). "Real" filesystems should always be listed first so
that the "bare" filename is tried before alternate filenames. For PXE
booting in particular this can remove a lot of spurious pathname lookups.
While here, move splitfs to the bottom after the bzip and gzip filesystems
as it is the least often used.
dim [Tue, 11 Nov 2014 18:54:57 +0000 (18:54 +0000)]
Change kbdb's kthr::cpu field into an int, to avoid gcc warnings about
comparing it with NOCPU, which became -1 recently. While here, avoid
using it for address calculations if it is negative.
jhb [Tue, 11 Nov 2014 18:15:05 +0000 (18:15 +0000)]
Use the callout(9) API instead of timeout(9). To do this more cleanly,
convert a global timer to a per-controller timer. This works much better
with locking and removes the need for several global lookup tables.
adrian [Tue, 11 Nov 2014 17:14:11 +0000 (17:14 +0000)]
Use the correct device (child) when asking the bus layer about which power
state said device should go into.
This was a snafu introduced in the ACPI/PCI awareness separation.
When putting a device into a power state, the bus (and thus firmware,
eg ACPI) should be asked before hand to check whether the device
can indeed go into that power state.
There's a set of nodes in ACPI under each device - the _SxD nodes - which
state which ACPI power state to put the device into when the system is
going into power save state 'x'. So when going into S3, the existence
of an _S3D node would override whatever the system was trying to do.
By default the PCI code wants to put devices into D3 before suspending.
I have a laptop here (Asus Zenbook - check the PR) whose EHCI controller
really wants to be in D2 during suspend, not D3. So if we put it into
D3 and then try to enter S3, everything hangs. The device itself
can go into D3 - it just can't be there when the call to ACPI to enter
S3 occurs. The PCI patch fixes this.
jkim@ noticed that the same is needed for the ACPI child device
enumeration.
Thankyou to Matt Dillon (the programmer, not the actor) for buying me
this particular laptop so I could debug the issues with the Atheros
AR9485 that is in it. It's his fault that I ended up with this
laptop and was sufficiently annoyed by the lack of USB suspend
to go down this rabbit hole.
Tested:
* Thinkpad T400
* Thinkpad X230
* Thinkpad T42
* Thinkpad T60
* Asus Zenbook (see PR)
* Asus EEEPC 701
* Asus EEEPC 1001PX
TODO:
* Figure out what we should do about devices we unload drivers for
that want to be in a specific state when entering S3 / S4 -
the "put devices into D3 if they're not bound to a driver" option
may also mess with things.
emaste [Tue, 11 Nov 2014 14:59:46 +0000 (14:59 +0000)]
Add workaround for vt efifb's early use of PHYS_TO_DMAP
In vt_efifb_init the framebuffer's physaddr is passed to PHYS_TO_DMAP
before the DMAP is setup. The result is not actually accessed until
after the mapping is setup, though. Loosen the assertion in PHYS_TO_DMAP
for now, to allow use when dmaplimit == 0.
Reviewed by: kib
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D1142
kib [Tue, 11 Nov 2014 12:52:45 +0000 (12:52 +0000)]
Based on some BIOS configuration (GGC register in host bridge, bit 1),
IGP may declare subclass as either VGA-compatible, or non-VGA. The
difference is that in the later case, IGP does not claim VGA cycles.
Other than that, the device functions normally, and agp_i810 should
attach to it.
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
pjd [Tue, 11 Nov 2014 04:48:09 +0000 (04:48 +0000)]
Add missing privilege check when setting the dump device. Before that change it
was possible for a regular user to setup the dump device if he had write access
to the given device. In theory it is a security issue as user might get access
to kernel's memory after provoking kernel crash, but in practise it is not
recommended to give regular users direct access to storage devices.
Rework the code so that we do privileges check within the set_dumper() function
to avoid similar problems in the future.
melifaro [Tue, 11 Nov 2014 02:52:40 +0000 (02:52 +0000)]
Kill custom in_matroute() radix mathing function removing one rte mutex lock.
Initially in_matrote() in_clsroute() in their current state was introduced by
r4105 20 years ago. Instead of deleting inactive routes immediately, we kept them
in route table, setting RTPRF_OURS flag and some expire time. After that, either
GC came or RTPRF_OURS got removed on first-packet. It was a good solution
in that days (and probably another decade after that) to keep TCP metrics.
However, after moving metrics to TCP hostcache in r122922, most of in_rmx
functionality became unused. It might had been used for flushing icmp-originated
routes before rte mutexes/refcounting, but I'm not sure about that.
So it looks like this is nearly impossible to make GC do its work nowadays:
in_rtkill() ignores non-RTPRF_OURS routes.
route can only become RTPRF_OURS after dropping last reference via rtfree()
which calls in_clsroute(), which, it turn, ignores UP and non-RTF_DYNAMIC routes.
Dynamic routes can still be installed via received redirect, but they
have default lifetime (no specific rt_expire) and no one has another trie walker
to call RTFREE() on them.
So, the changelist:
* remove custom rnh_match / rnh_close matching function.
* remove all GC functions
* partially revert r256695 (proto3 is no more used inside kernel,
it is not possible to use rt_expire from user point of view, proto3 support
is not complete)
* Finish r241884 (similar to this commit) and remove remaining IPv6 parts
luigi [Tue, 11 Nov 2014 00:13:28 +0000 (00:13 +0000)]
in the Linux section, properly define the NMG_LOCK type.
Also import WITH_GENERIC in preparation to adding fine-grained
options to disable specific netmap components.
emaste [Mon, 10 Nov 2014 18:20:46 +0000 (18:20 +0000)]
Add /usr/lib/debug directory to hier(7)
The canonical standalone debug directory established by the GNU
toolchain is /usr/lib/debug, and we use it when WITH_DEBUG_FILES is set.
Mention it in the file system hierarchy page.
Reviewed by: bcr
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D1134
ae [Mon, 10 Nov 2014 16:12:51 +0000 (16:12 +0000)]
Add sa6_checkzone_ifp() function. It checks correctness of struct
sockaddr_in6, usually obtained from the user level through ioctl.
It initializes sin6_scope_id using given interface.
des [Mon, 10 Nov 2014 09:55:35 +0000 (09:55 +0000)]
I just realized that the previous commit message makes no sense: the
first sentence should have read "Constify the AES and SHA-256 code and
wrappers". This allows us to feed zero_region (which is const) to the
hash function during reseeding and thereby implement the FS&K version of
SHAd-256 instead of the older F&S version.
delphij [Mon, 10 Nov 2014 08:20:21 +0000 (08:20 +0000)]
MFV r274273:
ZFS large block support.
Please note that booting from datasets that have recordsize greater
than 128KB is not supported (but it's Okay to enable the feature on
the pool). This *may* remain unchanged because of memory constraint.
Limited safety belt is provided for mounted root filesystem but use
caution is advised.
mav [Sun, 9 Nov 2014 22:43:29 +0000 (22:43 +0000)]
Handle PREEMPT AND ABORT service action equal to PREEMPT.
With command serialization used in CTL, there are no other commands to abort
when PREEMPT AND ABORT gets to run, so it is practically equal to PREEMPT.
melifaro [Sun, 9 Nov 2014 21:33:01 +0000 (21:33 +0000)]
Renove faith(4) and faithd(8) from base. It looks like industry
have chosen different (and more traditional) stateless/statuful
NAT64 as translation mechanism. Last non-trivial commits to both
faith(4) and faithd(8) happened more than 12 years ago, so I assume
it is time to drop RFC3142 in FreeBSD.
bryanv [Sun, 9 Nov 2014 20:04:12 +0000 (20:04 +0000)]
Enable LRO by default when available on vtnet interfaces
The prior change to not enable LRO by default has confused several
people. The configurations where LRO is problematic is not the
typical use case for VirtIO, and due to other issues, this often
requires checksum offloading to be disabled anyways.
Initialize tqent_flags in the userland taskq implementation. Without
this the assertion of tq->tq_freelist != NULL may fail in taskq_destroy.
The problem is that tqent_flags is never initialized in the userland
implementation while the kernel one does initialize it. Without proper
initialization, the flag may have its lowest bit set, making it treated
as TQENT_FLAG_PREALLOC and never removing taskq_ent_t from tq_freelist.
Remove ip6_getdstifaddr() and all functions to work with auxiliary data.
It isn't safe to keep unreferenced ifaddrs. Use in6ifa_ifwithaddr() to
determine ifaddr corresponding to destination address. Since currently
we keep addresses with embedded scope zone, in6ifa_ifwithaddr is called
with zero zoneid and marked with XXX.
Also remove route and lle lookups from ip6_input. Use in6ifa_ifwithaddr()
instead.
kib [Sat, 8 Nov 2014 11:56:26 +0000 (11:56 +0000)]
MFi386 r253328:
Create a proper stack frame for amd64 version of bcopy(). Note that
this also makes the stack properly aligned in the function, despite it
is not strictly needed.
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
delphij [Sat, 8 Nov 2014 07:30:40 +0000 (07:30 +0000)]
MFV r274271:
Improve zdb -b performance:
- Reduce gethrtime() call to 1/100th of blkptr's;
- Skip manipulating the size-ordered tree;
- Issue more (10, previously 3) async reads;
- Use lighter weight testing in traverse_visitbp();
ngie [Sat, 8 Nov 2014 03:20:56 +0000 (03:20 +0000)]
Use PROGS instead of PROG and remove unnecessary SRCS?= assignment
Using PROG instead of PROGS will in cases of high -j with -DNO_ROOT cause
the PROG to show up more than once as it's handling the SCRIPTS install case
in a recursive manner, separate from the non-recursive case
After the recent batch of commits to bsd.progs.mk to fix behavior with how
variables are defaulted to, explicitly setting SRCS for a PROG is no longer
required
jilles [Fri, 7 Nov 2014 21:30:16 +0000 (21:30 +0000)]
sh(1): Mention portability issue with shifting zero positional parameters.
Per Austin Group issue #459, shifting zero positional parameters may or may
not be considered an operand error (which causes the shell to exit in most
cases).