target. Due to a way issignal() selects the next signal to deliver
and report, if the simultaneous or already pending another signal
exists, that signal might be reported by the next waitpid(2) call.
This causes minor annoyance for debuggers, which must be prepared to
take any signal as the first event, then filter SIGSTOP later.
More importantly, for tools like gcore(1), which attach and then
detach without processing events, SIGSTOP might leak to be delivered
after PT_DETACH. This results in the process being unintentionally
stopped after detach, which is fatal for automatic tools.
The solution is to force SIGSTOP to be the first signal reported after
the attach. Attach code is modified to set P2_PTRACE_FSTP to indicate
that the attaching ritual was not yet finished, and issignal() prefers
SIGSTOP in that condition. Also, the thread which handles
P2_PTRACE_FSTP is made to guarantee to own p_xthread during the first
waitpid(2). All that ensures that SIGSTOP is consumed first.
Additionally, if P2_PTRACE_FSTP is still set on detach, which means
that waitpid(2) was not called at all, SIGSTOP is removed from the
queue, ensuring that the process is resumed on detach.
In issignal(), when acting on STOPing signals, remove the signal from
queue before suspending. Otherwise parallel attach could result in
ptracestop() acting on that STOP as if it was the STOP signal from the
attach. Then SIGSTOP from attach leaks again.
As a minor refactoring, some bits of the common attach code is moved
to new helper proc_set_traced().
Reported by: markj
Reviewed by: jhb, markj
Tested by: pho
Sponsored by: The FreeBSD Foundation
MFC after: 2 weeks
Differential revision: https://reviews.freebsd.org/D7256
Instead of changing the whole course to another POSIX-permitted way
for consistency and uniformity I decide to completely ignore missing
regex fucntionality and focus on fixing bugs in what we have now,
too many small obstacles we have choicing other way, counting ports.
Corresponding libc changes are backed out in r302824.
On machines where SIZE_T_MAX exceeds OFF_MAX (signed 64-bit), permit seeking
character devices to negative off_t values. This enables dd(1) to interact
with kernel KVA in /dev/kmem on amd64, for example.
Sponsored by: EMC / Isilon Storage Division
This is the same issue as r297997, but was missed in it.
The WARNS value changes between 'build-tools' (MK_WARNS=no) and
'everything' resulting in a rebuild of this file.
Approved by: re (implicit)
Sponsored by: EMC / Isilon Storage Division
The fix in r300649 was not sufficient to convince Coverity that the
buffer was NUL terminated, even with the buffer pre-zeroed. Swap
the size and nmemb arguments to fread() so that a valid lenght is
returned, which we can use to terminate the string in the buffer
at the correct location. This should also quiet the complaint about
the return value of fread() not being checked.
Reported by: Coverity
CID: 1019054, 1009614
MFC after: 1 week
The first file in these lists will generate everything else so only
it should be getting a .meta file. With bmake's missing=yes meta
feature these would otherwise cause a rebuild without the
.NOMETA hint.
Sponsored by: EMC / Isilon Storage Division
ki_flag and ki_tdflag have been 'long', not 'int', since 2000 and 2005,
respectively.
Submitted by: Shawn Wills <swills at isilon dot com>
Sponsored by: EMC / Isilon Storage Division
case of \c in the prompt format string is a no-op. We already passed
this test at the top of the loop, and i has not yet been incremented in
this path. Change this test to (i < PROMPTLEN - 2).
Reported by: Coverity
CID: 1008328
Reviewed by: cem
MFC after: 1 week
This will make things more flexible if the program path changes in the future,
and the test in and of itself doesn't call /usr/bin/bc -- it just calls bc
MFC after: 1 week
Sponsored by: EMC / Isilon Storage Division
There is no need to to call strdup() on the value returned by fmt().
The latter calls fmt_argv() which always returns a dynamically
allocated string, and calling strdup() on that leaks the memory
allocated by fmt_argv(). Wave some const magic on ki_args and
ki_env to make the direct assignment happy. This requires a tweak
to the asprintf() case to avoid a const vs. non-const mismatch.
Reported by: Coverity
CID: 1011370
MFC after: 1 week
Replace the magic constant 127 in the loop interation count with
"PROMPTLEN - 1".
gethostname() is not guaranteed to NUL terminate the destination
string if it is too short. Decrease the length passed to gethostname()
by one, and add a NUL at the end of the buffer to make sure the
following loop to find the end of the name properly terminates.
The default: case is the likely cause of Coverity CID 1008328. If
i is 126 at the top of the loop interation where the default case
is triggered, i will be incremented to 127 by the default case,
then incremented to 128 at the top of the loop before being compared
to 127 (PROMPTLENT - 1) and terminating the loop. Then the NUL
termination code after the loop will write to ps[128]. Fix by
checking for overflow before incrementing the index and storing the
second character in the buffer.
These fixes are not guaranteed to satisfy Coverity. The code that
increments i in the 'h'/'H' and 'w'/'W' cases may be beyond its
capability to analyze, but the code appears to be safe.
Reported by: Coverity
CID: 1008328
Reviewed by: jilles, cem
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D6482
- Use arc4random_uniform(3).
- Prevent a segmentation fault when ed receives a signal
while being in getpass(). [1]
Obtained from: OpenBSD [1] (CVS Rev. 1.15)
MFC after: 2 weeks
When getline(3) in 2009 was added a _WITH_GETLINE guard has also been added.
This rename is made in preparation for the removal of this guard
Obtained from: NetBSD
after r298107
Summary of changes:
- Replace all instances of FILES/TESTS with ${PACKAGE}FILES. This ensures that
namespacing is kept with FILES appropriately, and that this shouldn't need
to be repeated if the namespace changes -- only the definition of PACKAGE
needs to be changed
- Allow PACKAGE to be overridden by callers instead of forcing it to always be
`tests`. In the event we get to the point where things can be split up
enough in the base system, it would make more sense to group the tests
with the blocks they're a part of, e.g. byacc with byacc-tests, etc
- Remove PACKAGE definitions where possible, i.e. where FILES wasn't used
previously.
- Remove unnecessary TESTSPACKAGE definitions; this has been elided into
bsd.tests.mk
- Remove unnecessary BINDIRs used previously with ${PACKAGE}FILES;
${PACKAGE}FILESDIR is now automatically defined in bsd.test.mk.
- Fix installation of files under data/ subdirectories in lib/libc/tests/hash
and lib/libc/tests/net/getaddrinfo
- Remove unnecessary .include <bsd.own.mk>s (some opportunistic cleanup)
Document the proposed changes in share/examples/tests/tests/... via examples
so it's clear that ${PACKAGES}FILES is the suggested way forward in terms of
replacing FILES. share/mk/bsd.README didn't seem like the appropriate method
of communicating that info.
MFC after: never probably
X-MFC with: r298107
PR: 209114
Relnotes: yes
Tested with: buildworld, installworld, checkworld; buildworld, packageworld
Sponsored by: EMC / Isilon Storage Division
If the hostname is empty and \h is used in $PS1,
the remainder of the prompt following \h will be empty.
Likewise for $PWD and \w. Fix it.
Reviewed by: jilles
MFC after: 1 week
Sponsored by: Dell Inc.
Differential Revision: https://reviews.freebsd.org/D6188
Fix a related typo while here.
Note, this change results in the Kyuafile inclusion in the runtime
package, which needs to be fixed, however addresses the PR as far
as I can tell in my tests.
PR: 209114
Submitted by: ngie
Sponsored by: The FreeBSD Foundation
is invalid. It is used to find the size of allocated memory. As malloc may
allocate memory with mmap it will fail to take this memory into account.
Obtained from: brooks
This avoids 'build command changed' due to CFLAGS/CC changes during the
normal build. Without this the build-tools targets end up rebuilding
for the *target* rather than keeping the native versions built in
build-tools.
Sponsored by: EMC / Isilon Storage Division
At least for ncurses this fixes a build error due to it trying to run
'ccache --version' to work around a gcc 5 bug using the fix in r287205.
Sponsored by: EMC / Isilon Storage Division
libraries (such as libthr) which maintain their own signal state. This
change adds the tcsh SAVESIGVEC option to save and restore the sigvecs for
the signals the child modifies before it execs.
Reviewed by: kib, rwatson
Reported by: kib
breaking the ABI. Special value is stored in the lock pointer to
indicate shared lock, and offline page in the shared memory is
allocated to store the actual lock.
Reviewed by: vangyzen (previous version)
Discussed with: deischen, emaste, jhb, rwatson,
Martin Simmons <martin@lispworks.com>
Tested by: pho
Sponsored by: The FreeBSD Foundation
(when they actually get committed, that is), and might also come in handy
in other situations.
Reviewed by: wblock@ (man page)
MFC after: 1 month
Sponsored by: The FreeBSD Foundation
These are no longer needed after the recent 'beforebuild: depend' changes
and hooking DIRDEPS_BUILD into a subset of FAST_DEPEND which supports
skipping 'make depend'.
Sponsored by: EMC / Isilon Storage Division
The intent of the previous code in that case was to force
an explicit write, but the implementation was incorrect, and
as a result the write was never performed. This new implementation
instead uses ftruncate(2) to extend the file with a trailing hole.
Also introduce regression tests for these cases.
PR: 189284
(original PR whose fix introduced this bug)
PR: 207092
Differential Revision: D5248
Reviewed by: sobomax,kib
MFC after: 2 weeks
Note: tcsh(1) has a MK_TCSH=no test, so this should be a separate
package, which requires pre-install/post-install scripts, to be
added later.
Sponsored by: The FreeBSD Foundation
The symbols of libiconv from ports were changed to
have prefixed.
Since we have iconv in our libc these days, we don't
need it on 10.X and later. However, 9.X still need
this.
Spotted by: Yoshihiko Sarumaru
MFC after: 1 days
Builtins (including variable assignments without command word), function
calls and redirected compound commands need to restore file descriptors
to their original state after execution. This is handled by allocating a
redirtab structure. These mallocs and frees show up heavily in pmcstat.
Only allocate a redirtab if there are actually redirections and maintain a
count of how many levels of REDIR_PUSH there are without redirtabs.
A simple loop without external programs like
sh -c 'i=0; w=$(printf %0100d 7); while [ "$i" -lt 1000000 ]; do
i=$((i+1)); done'
is over 25% faster on an amd64 bhyve VM.
The linear search using strcmp() shows up in pmcstat for several percent.
Split the operators into lengths and whether they start with '-' and compare
bytes using == instead of strcmp().
A simple test
sh -c 'i=0; w=$(printf %0100d 7); while [ "$i" -lt 1000000 ]; do
v=$(printf %sx%s "$w" "$w"); i=$((i+1)); done'
is over 4% faster on an amd64 bhyve VM.
