Pull in r353378 from upstream lld trunk (by George Rimar):
[LLD][ELF] - Set DF_STATIC_TLS flag for X64 target
This is the same as D57749, but for x64 target.
"ELF Handling For Thread-Local Storage" p41 says
(https://www.akkadia.org/drepper/tls.pdf):
R_X86_64_GOTTPOFF relocation is used for IE TLS models.
Hence if linker sees this relocation we should add DF_STATIC_TLS flag.
This adds support to lld for the DF_STATIC_TLS flag in shared objects,
which signals to the dynamic linker that the shared object requires
static thread local storage.
See also: https://reviews.freebsd.org/D19072
MFC after: 1 week
bde [Thu, 21 Feb 2019 17:31:33 +0000 (17:31 +0000)]
Move scterm_teken.c from 6 MD files lists to the MI files list so that it
is easier to configure. It is MI, unlike some of the other syscons files
already in the MI list.
Move scvtb.c similarly. It is needed whenever sc is configured, and is
more MI than most of the files already in the MI list.
This only changes the combined list for arm64 and mips. These arches
already cannot build sc or even NOTES.
mav [Thu, 21 Feb 2019 16:47:36 +0000 (16:47 +0000)]
Fix few issues in ioat(4) driver.
- Do not explicitly count active descriptors. It allows hardware reset
to happen while device is still referenced, plus simplifies locking.
- Do not stop/start callout each time the queue becomes empty. Let it
run to completion and rearm if needed, that is much cheaper then to touch
it every time, plus also simplifies locking.
- Decouple submit and cleanup locks, making driver reentrant.
- Avoid memory mapped status register read on every interrupt.
- Improve locking during device attach/detach.
- Remove some no longer used variables.
Reviewed by: cem
MFC after: 1 week
Sponsored by: iXsystems, Inc.
Differential Revision: https://reviews.freebsd.org/D19231
emaste [Thu, 21 Feb 2019 15:08:25 +0000 (15:08 +0000)]
svn: support building with WITH_PIE
Subversion builds and links against its own .a archives using local
rules, so did not benefit from with the WITH_PIE library support added
in r344179. Apply the same _pie suffix locally.
Reviewed by: cem
Differential Revision: https://reviews.freebsd.org/D19246
tuexen [Thu, 21 Feb 2019 10:35:32 +0000 (10:35 +0000)]
The receive buffer autoscaling for TCP is based on a linear growth, which
is acceptable in the congestion avoidance phase, but not during slow start.
The MTU is is also not taken into account.
Use a method instead, which is based on exponential growth working also in
slow start and being independent from the MTU.
This is joint work with rrs@.
Reviewed by: rrs@, Richard Scheffenegger
Sponsored by: Netflix, Inc.
Differential Revision: https://reviews.freebsd.org/D18375
tuexen [Thu, 21 Feb 2019 09:34:47 +0000 (09:34 +0000)]
This patch addresses an issue brought up by bz@ in D18968:
When TCP_REASS_LOGGING is defined, a NULL pointer dereference would happen,
if user data was received during the TCP handshake and BB logging is used.
A KASSERT is also added to detect tcp_reass() calls with illegal parameter
combinations.
Reported by: bz@
Reviewed by: rrs@
MFC after: 3 days
Sponsored by: Netflix, Inc.
Differential Revision: https://reviews.freebsd.org/D19254
cem [Thu, 21 Feb 2019 02:41:57 +0000 (02:41 +0000)]
fuse: Fix a regression introduced in r337165
On systems with non-default DFLTPHYS and/or MAXBSIZE, FUSE would attempt to
use a buf cache block size in excess of permitted size. This did not affect
most configurations, since DFLTPHYS and MAXBSIZE both default to 64kB.
The issue was discovered and reported using a custom kernel with a DFLTPHYS
of 512kB.
sef [Thu, 21 Feb 2019 01:30:37 +0000 (01:30 +0000)]
* Handle SIGPIPE in gssd
We've got some cases where the other end of gssd's AF_LOCAL socket gets
closed, resulting in an error (and SIGPIPE) when it tries to do I/O to it.
Closing without cleaning up means the next time nfsd starts up, it hangs,
unkillably; this allows gssd to handle that particular error.
* Limit the retry cound in gssd_syscall to 5.
The default is INT_MAX, which effectively means forever. And it's an
uninterruptable RPC call, so it will never stop.
dab [Wed, 20 Feb 2019 22:05:44 +0000 (22:05 +0000)]
Complete fix for CID 1007454, CID 1007453: Resource leak in newsyslog
The result of a strdup() was stored in a global variable and not freed
before program exit. This is a follow-up to r343906. That change
attempted to plug these resource leaks but managed to miss a code path
on which the leak still occurs. Plug the leak on that path, too.
tsoome [Wed, 20 Feb 2019 21:07:09 +0000 (21:07 +0000)]
loader: really fix cd9660 dirmatch
The cd9660_open() does pass whole path to dirmatch() and we need to
compare only the current path component, not full path.
Additinally, skip over duplicate / (if any) and check if the last component
in the path was meant to be directory (having trailing /). If it is in fact
a file, error out.
dim [Wed, 20 Feb 2019 21:06:11 +0000 (21:06 +0000)]
Pull in r354515 from upstream libc++ trunk:
Fix the build with gcc when `-Wredundant-decls` is passed
Summary:
gcc warns that `__throw_runtime_error` is declared both in
`<__locale>` and `<stdexcept>`, if `-Wredundant-decls` is passed on
the command line; this is the case with FreeBSD when ${WARNS} == 6.
Since `<__locale>` gets its first declaration via a transitive
include of `<stdexcept>`, and the second declaration is after the
first invocation of `__throw_runtime_error`, delete that second
declaration.
Signed-off-by: Enji Cooper <yaneurabeya@gmail.com>
Reviewers: kristina, MaskRay, EricWF, ldionne, ngie
dim [Wed, 20 Feb 2019 20:17:54 +0000 (20:17 +0000)]
Fix more AddressSanitizer violations in usr.bin/top
In line_update(), set lastcol correctly after moving to any non-zero
column, so the "overwrite old stuff" part does not attempt to address
negative offsets in the current line.
Rewrite setup_buffer() to always allocate at least 80 characters,
otherwise various calls to summary_format() will overwrite the end of
the buffers, if the screen width gets small enough.
avg [Wed, 20 Feb 2019 13:19:08 +0000 (13:19 +0000)]
fix userland illumos taskq code to pass relative timeout to cv_timedwait
Unlike illumos, FreeBSD cv_timedwait requires a relative timeout. That
applies both to the kernel illumos compatibility code and to the
userland "fake kernel" code.
kib [Wed, 20 Feb 2019 09:38:19 +0000 (09:38 +0000)]
Implement rangesets.
The data structure implements non-intersecting intervals over the [0,
UINT64_MAX] range, and supports fast insert, predicated clearing of
subrange, and lookup of an interval containing the specified address.
Internally it is a pctrie over the interval start addresses.
Implementation provides additional guarantees over the structure state
in case of memory allocation failures.
Reviewed by: markj
Tested by: pho
Sponsored by: The FreeBSD Foundation
MFC after: 2 weeks
Differential revision: https://reviews.freebsd.org/D18893
bapt [Wed, 20 Feb 2019 06:40:52 +0000 (06:40 +0000)]
calendar: use iconv to respect the output encoding
calendar(1) can have input in various encoding, specifying
LANG=<locale_name> to enable calendar(1) to determine which one to use.
The problem is the content of the calendar itself is exposed as is making it
unreadable in many cases. For example french calendar which is encoded
ISO8859-1 is rendered badly in a fr_FR.UTF-8 environment.
Using iconv allows to solve this issue.
This will also allow to keep only 1 encoding in base for those files without
breaking user existing setup
Reported by: many
MFC after: 1 month
Differential Revision: https://reviews.freebsd.org/D19221
ian [Wed, 20 Feb 2019 03:00:55 +0000 (03:00 +0000)]
Fix the handling of legacy-format devices in the u-boot loaderdev variable.
When I added support for the standard loader(8) disk0s2a: type formats,
the parsing of legacy format was broken because it also contains a colon,
but it comes before the slice and partition. That would cause disk_parsedev()
to return success with the slice and partition set to wildcard values.
This change examines the string first, and if it contains spaces, dots, or
a colon at any position other than the end, it must be a legacy-format
string and we don't even try to use disk_parsedev() on it.
cem [Wed, 20 Feb 2019 02:49:26 +0000 (02:49 +0000)]
Fuse: whitespace and style(9) cleanup
Take a pass through fixing some of the most egregious whitespace issues in
fs/fuse. Also fix some style(9) warts while here. Not 100% cleaned up, but
somewhat less painful to look at and edit.
bde [Wed, 20 Feb 2019 02:40:38 +0000 (02:40 +0000)]
Fix hangs in r341810 waiting for AP startup.
idle_td is dereferenced without thread-locking it to make its contents is
invariant, and was accessed without telling the compiler that its contents
is invariant. Some compilers optimized accesses to the supposedly invariant
contents by moving the critical checks for changes outside of the loop that
waits for changes. Fix this using atomic ops.
This bug only showed up for the following configuration: a Turion2
system, amd64 kernels, compiled by gcc, and SCHED_4BSD. clang fails
to do the optimization with all CFLAGS that I tried, because it doesn't
fully optimize the '__asm __volatile' for cpu_spinwait() although this
asm has no memory clobber. gcc only does the optimization with most
CFLAGS. I mostly used -Os with all compilers. i386 works because gcc
-m32 -Os only moves 1 or the 2 accesses outside of the loop.
Non-Turion2 systems and SCHED_ULE worked due to different timing (when
all APs start before the BP checks them outside of the loop).
bde [Wed, 20 Feb 2019 02:14:41 +0000 (02:14 +0000)]
Attempt to complete fixing programmable function keys for syscons.
The flag for the driver capability of supporting the fix is independent
of the flag for cons25 mode so that it can be managed independently, but
I forget to preserve it when resetting the terminal.
pjd [Tue, 19 Feb 2019 23:35:55 +0000 (23:35 +0000)]
The way ZFS searches for its vdevs is the following: first it looks for
a vdev that has the same name as the one stored in metadata and that has
all VDEV labels in place. If it cannot find a GEOM provider with the given
name and all VDEV labels it will scan all GEOM providers for the best match
(the most VDEV labels available), but here the name is ignored.
In case the ZFS pool is created, eg. using GPT partition label:
# zpool create tank /dev/gpt/tank
everything works, and on every import ZFS will pick /dev/gpt/tank and
not /dev/da0p4.
The problem occurs when da0p4 is extended and ZFS is unable to find all
VDEV labels in /dev/gpt/tank anymore (the VDEV labels stored at the end
of the partition are now somewhere else). In this case it will scan all
GEOM providers and will pick the first one with the best match, ie. da0p4.
Fix this problem by checking the VDEV/provider name even if we get the same
match. If the name is the same as the one we have in pool's metadata, prefer
this GEOM provider.
Reported by: oshogbo, Michal Mroz <m.mroz@fudosecurity.com>
Tested by: Michal Mroz <m.mroz@fudosecurity.com>
Obtained from: Fudo Security
ngie [Tue, 19 Feb 2019 22:19:31 +0000 (22:19 +0000)]
Make `server_cat(..)` handle short receives
In short, the prior code was far too simplistic when it came to calling recv(2)
and failed intermittently (or in the case of Jenkins, deterministically).
Handle short recv(2)s by checking the return code and incrementing the window
into the buffer by the number of received bytes. If the number of received
bytes <= 0, then bail out of the loop, and test the total number of received
bytes vs the expected number of bytes sent for equality, and base whether or
not the test passes/fails on that fact.
Remove the expected failure, now that the hdtr testcases deterministically pass
on my host after this change [1].
ed [Tue, 19 Feb 2019 21:58:23 +0000 (21:58 +0000)]
Place an upper bound on the number of iterations for REP.
Right now it's possible to invoke the REP escape sequence with a maximum
of tens of millions of iterations. In practice, there is never any need
to do this. Calling it more frequently than the number of cells in the
terminal hardly makes any sense. By placing a limit on it, we can
prevent users from exhausting resources in inside the terminal emulator.
As support for this escape sequence is not present in any of the stable
branches, there is no need to MFC.
markj [Tue, 19 Feb 2019 21:33:02 +0000 (21:33 +0000)]
Limit the number of entries allocated for a REPORT_ZONES command.
The DIOCGETZONE ioctl can be used to fetch the zone list of an SMR
drive, and the caller specifies the number of entries it wants to fetch.
Clamp the caller's request to a sane limit so that a user cannot attempt
large allocations. Callers already need to invoke the ioctl multiple
times to fetch the full list in general, so there's no harm in limiting
the number of entries returned.
Fix style while here.
admbug: 807
Reported by: Ilja Van Sprundel <ivansprundel@ioactive.com>
Reviewed by: asomers, ken
Tested by: ken
MFC after: 1 week
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D19249
markj [Tue, 19 Feb 2019 21:22:22 +0000 (21:22 +0000)]
Impose a limit on the number of GEOM_CTL arguments.
Otherwise a privileged user can trigger a memory allocation of
unbounded size, or an integer overflow in the subsequent
geom_alloc_copyin() call, leading to out-of-bounds accesses.
Hard-code a large limit to circumvent this problem.
admbug: 854
Reported by: Anonymous of the Shellphish Grill Team
Reviewed by: ae
MFC after: 1 week
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D19251
thj [Tue, 19 Feb 2019 19:57:55 +0000 (19:57 +0000)]
When dropping a fragment queue count the number of fragments in the queue
When dropping a fragment queue, account for the number of fragments in the
queue. This improves accounting between the number of fragments received and
the number of fragments dropped.
imp [Tue, 19 Feb 2019 19:37:02 +0000 (19:37 +0000)]
Remove the i915 and radeon drivers.
Per discussions on arch@ and elsewhere, the maintenance of this code
has moved to the drm-kmod and drm-legacy-kmod ports. Remove the i915
and radeon drivers from the tree.
imp [Tue, 19 Feb 2019 19:36:43 +0000 (19:36 +0000)]
Per discussions on arch@ and elsewhere, retire drm module / drives.
Retire the drm modules / drivers. These are now handled by the
drm-legacy-kmod port and/or the drm-kmod port. All future
development and maintanace will be handled there.
markj [Tue, 19 Feb 2019 15:46:43 +0000 (15:46 +0000)]
Move a racy assertion in filt_pipewrite().
EVFILT_WRITE knotes for pipes live on the knlist for the other end of the
pipe. Since they do not hold a reference on the corresponding file
structure, they may be removed from the knlist by pipeclose() while still
remaining active. In this case, there is no knlist lock acquired before
filt_pipewrite() is called, so the assertion fails.
Fix the problem by first checking whether that end of the pipe has been
closed. These checks are memory safe since the knote holds a reference
on one end of the pipe, and the pipe structure is not freed until both
ends are closed. The checks are not racy since PIPE_EOF is never cleared
after being set, and pipe_present is never set back to PIPE_ACTIVE after
pipeclose() has been called.
PR: 235640
Reported and tested by: pho
Reviewed by: kib
MFC after: 2 weeks
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D19224
trasz [Tue, 19 Feb 2019 12:45:37 +0000 (12:45 +0000)]
Work around the "nfscl: bad open cnt on server" assertion
that can happen when rerooting into NFSv4 rootfs with kernel
built with INVARIANTS.
I've talked to rmacklem@ (back in 2017), and while the root cause
is still unknown, the case guarded by assertion (nfscl_doclose()
being called from VOP_INACTIVE) is believed to be safe, and the
whole thing seems to run just fine.
trasz [Tue, 19 Feb 2019 11:07:02 +0000 (11:07 +0000)]
Bump the default kern.rpc.gss.client_max from 128 to 1024.
The old value resulted in bad performance, with high kernel
and gssd(8) load, with more than ~64 clients; it also triggered
crashes, which are to be fixed by a different patch.
glebius [Mon, 18 Feb 2019 21:27:13 +0000 (21:27 +0000)]
With r343051 UMA switched from atomic counts to counter(9) and now kernel
reports snap counts of how much a zone alloced and how much it freed. It
may happen that snap values doesn't match, e.g alloced - freed < 0.
Workaround that in memstat library.
dim [Mon, 18 Feb 2019 19:07:15 +0000 (19:07 +0000)]
Add one additional file to libllvmminimal, since in some cases (e.g.
upgrading from stable/10 to stable/11) symbols from it are needed to
link llvm-tblgen and clang-tblgen.
ian [Mon, 18 Feb 2019 17:12:30 +0000 (17:12 +0000)]
Allow the u-boot loaderdev env var to be formatted in the "usual" loader(8)
way: device<unit>[s|p]<slice><partition>. E.g., disk0s2a or disk3p12.
The code first tries to parse the variable in this format using the
standard disk_parsedev(). If that fails, it falls back to parsing the
legacy format that has been supported by ubldr for years.
In addition to 'disk', all the valid uboot device names can also be used:
mmc, sata, usb, ide, scsi. The 'disk' device serves as an alias for all
those types and will match the Nth storage-type device found (where N is
the unit number).
vmaffione [Mon, 18 Feb 2019 14:21:41 +0000 (14:21 +0000)]
netmap: don't schedule kqueue notify task when kqueue is not used
This change adds a counter (kqueue_users) to keep track of how many
kqueue users are referencing a given struct nm_selinfo.
In this way, nm_os_selwakeup() can schedule the kevent notification
task only when kqueue is actually being used.
This is important to avoid wasting CPU in the common case where
kqueue is not used.
ian [Mon, 18 Feb 2019 04:44:52 +0000 (04:44 +0000)]
Make uboot_devdesc properly alias disk_devdesc, so that parsing the u-boot
loaderdev variable works correctly.
The uboot_devdesc struct is variously cast back and forth between
uboot_devdesc and disk_devdesc as pointers are handed off through various
opaque interfaces. uboot_devdesc attempted to mimic the layout of
disk_devdesc by having a devdesc struct, followed by a union of some
device-specific stuff that included a struct that contains the same fields
as a disk_devdesc. However, one of those fields inside the struct is 64-bit
which causes the entire union to be 64-bit aligned -- 32 bits of padding
is added between the struct devdesc and the union, so the whole mess ends
up NOT properly mimicking a disk_devdesc after all. (In disk_devdesc there
is also 32 bits of padding, but it shows up immediately before the d_offset
field, rather than before the whole collection of d_* fields.)
This fixes the problem by using an anonymous union to overlay the devdesc
field uboot network devices need with the disk_devdesc that uboot storage
devices need. This is a different solution than the one contributed with
the PR (so if anything goes wrong, the blame goes to me), but 95% of the
credit for this fix goes to Pawel Worach and Manuel Stuhn who analyzed the
problem and proposed a fix.
emaste [Mon, 18 Feb 2019 03:49:16 +0000 (03:49 +0000)]
kdump: expand comment on reasons for CAPFAIL_LOOKUP
Comment for CAPFAIL_LOOKUP refered only to paths containing ".." but
it is returned for other restricted VFS lookup cases, such as absolute
paths or openat(AT_FDCWD, ...).
kevans [Mon, 18 Feb 2019 02:59:47 +0000 (02:59 +0000)]
lualoader: only clear the screen before first password prompt
This was previously an unconditional screen clear, regardless of whether or
not we would be prompting for any passwords. This is pointless, given that
the screen clear is only there to put our screen into a consistent state
before we draw the prompts and do cursor manipulation.
This is also the only screen clear besides that to draw the menu. One can
now see early pre-loader and loader output with the menu disabled, which may
be useful for diagnostics.
ian [Sun, 17 Feb 2019 23:32:09 +0000 (23:32 +0000)]
Restore loader(8)'s ability for lsdev to show partitions within a bsd slice.
I'm pretty sure this used to work at one time, perhaps long ago. It has
been failing recently because if you call disk_open() with dev->d_partition
set to -1 when d_slice refers to a bsd slice, it assumes you want it to
open the first partition within that slice. When you then pass that open
dev instance to ptable_open(), it tries to read the start of the 'a'
partition and decides there is no recognizable partition type there.
This restores the old functionality by resetting d_offset to the start
of the raw slice after disk_open() returns. For good measure, d_partition
is also set back to -1, although that doesn't currently affect anything.
I would have preferred to make disk_open() avoid such rude assumptions and
if you ask for partition -1 you get the raw slice. But the commit history
shows that someone already did that once (r239058), and had to revert it
(r239232), so I didn't even try to go down that road.
oshogbo [Sun, 17 Feb 2019 18:26:27 +0000 (18:26 +0000)]
libnv: fix double free
In r343986 we introduced a double free. The structure was already
freed fixed in the r302966. This problem was introduced
because the GitHub version was out of sync with the FreeBSD one.
Submitted by: Mindaugas Rasiukevicius <rmind@netbsd.org>
MFC with: r343986
projects / FreeBSD / FreeBSD.git / log