The GLIBC extension AT_EMPTY_PATH allows the functions fchownat
and fstatat to operate on dirfd alone, if the given pathname is an
empty string. This also allows to operate on any file type, not
only directories.
Commit fa84aa4dd2 broke this. It only allows dirfd to be a
directory in calls to these two functions.
Fix that by handling AT_EMPTY_PATH right in gen_full_path_at.
A valid dirfd and an empty pathname is now a valid combination
and, noticably, this returns a valid path in path_ret. That
in turn allows to remove the additional path generation code
from the callers.
Fixes: fa84aa4dd2 ("Cygwin: fix errno values set by readlinkat")
Reported-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Tested-by: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
Convert gen_full_path_at to take flag values from the caller, rather
than just a bool indicating that empty paths are allowed. This is in
preparation of a better AT_EMPTY_PATH handling in a followup patch.
Reviewed-by: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
Cygwin needs an internal flag to allow specifying an empty pathname
in utimesat (GLIBC extension). We define it in _default_fcntl.h to
make sure we never introduce a value collision accidentally.
While at it, define the values as 16 bit hex values.
Reviewed-by: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
The check if the local variable p is NULL is useless. The preceeding
code always sets p to a valid pointer, or it crashes if path_ret is
invalid (which would be a bug in Cygwin).
Fixes: c57b57e5c4 ("* cygwin.din: Sort.")
Reviewed-by: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
Also: workaround a problem with actions/checkout's post-run step using
Cygwin git but being incompatible with it. (This would be better solved
by cygwin-install-action having a post-run step to unwind it's PATH
changes, but that's not currently possible)
Signed-off-by: Jon Turney <jon.turney@dronecode.org.uk>
A more sophisticated (and modern) test harness would probably be useful,
but switching to Automake's built-in test harness gets us parallel test
execution, colourization of failures, simplifies matters, seems adequate
for the current testuite, and means we don't need to write any icky Tcl.
Signed-off-by: Jon Turney <jon.turney@dronecode.org.uk>
When creating the POSIX ACL rewrite, the code merging permissions from
everyone/group to group/user ACEs was accidentally called for newly
generated files as well.
This could result in broken permissions, if umask used unusual values
like "0100", granted permissions to everyone/group not granted to
group/user.
Make sure to skip permission merging if the file got just created and
we only want to set correct permissions for the first time.
Fixes: bc444e5aa4 ("Reapply POSIX ACL changes.")
Reported-by: Jon Turney <jon.turney@dronecode.org.uk>
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
Move strace output to fix uninitalized use of fh introduced in previous commit.
../../../../src/winsup/cygwin/syscalls.cc: In function ‘int stat_worker(path_conv&, stat*)’:
../../../../src/winsup/cygwin/syscalls.cc:1971:69: error: ‘fh’ may be used uninitialized [-Werror=maybe-uninitialized]
Fixes: 42b44044b3 ("Cygwin: Fix Windows file handle leak in stat("file", -1)")
Signed-off-by: Jon Turney <jon.turney@dronecode.org.uk>
Don't leak a Windows file handle if stat() is called with a valid
filename, but invalid stat buffer pointer.
We do not destroy fh (which closes a Windows handle it has opened) if an
exception happens in the __try block.
Avoid this by re-ordering things so that we don't construct the fhandler
object until after we've attempted to use the struct stat buffer.
Fixes: 73151c54d5 ("syscalls.cc (stat_worker): Don't call build_fh_pc with invalid pc.")
Signed-off-by: Jon Turney <jon.turney@dronecode.org.uk>
Drop setting TDIRECTORY, just use /tmp in the 'test installation' now
that we have it.
This effectively reverts f3ed5f2fe0
Signed-off-by: Jon Turney <jon.turney@dronecode.org.uk>
Starting with commit 42faed4128 ("* thread.h (class pthread): Add bool
member canceled."), pthread::testcancel waits infinitely on cancel_event
after it checked if the canceled variable is set. However, this might
introduce a deadlock, if the thread calling pthread_cancel is terminated
after setting canceled to true, but before calling SetEvent on cancel_event.
In fact, it's not at all necessary to wait infinitely. By definition,
the thread is only canceled if cancel_event is set. The canceled
variable is just a helper to speed up code. We can safely assume that
the thread hasn't been canceled yet, if canceled is set, but cancel_event
isn't.
Fixes: 42faed4128 ("* thread.h (class pthread): Add bool member canceled.")
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
Despite our efforts, sometimes the async cancellation gets deferred.
Notice this by calling pthread_testcancel(), and then try to work out if
async cancellation was ever successful by checking if all threads ran
for the full expected time, or if some were stopped early.
Also, increase the time we allow for the async cancellation to get
delivered to 30 seconds.
Signed-off-by: Jon Turney <jon.turney@dronecode.org.uk>
Add back the restoration of signal handlers modified during system() on
thread cancellation.
Removed in 3cb9da14 which describes it as 'ill-conceived' (additional
context doesn't appear to be available).
We use the internal implementation helpers for the pthread cleanup
chain, so we can neatly tuck it inside the object, and keep the point
when we restore the signal handlers the same. (The
pthread_cleanup_push/pop() functions are implemented as macros which
must appear in the same lexical scope.)
Fixes: 3cb9da1461 ("Put signals on hold and use system_call_cleanup
class to set and restore signals rather than doing it prior to to
running the program. Remove the ill-conceived pthread_cleanup stuff.")
Signed-off-by: Jon Turney <jon.turney@dronecode.org.uk>
Take note of schedparam in any pthread_attr_t passed to pthread_create.
postcreate() (racily, after the thread is actually created), sets the
scheduling priority if it's inherited, but precreate() doesn't store any
scheduling priority explicitly set via a non-default attr to
pthread_create, so schedparam.sched_priority has the default value of 0.
(I think this is another long-standing bug exposed by 4b51e4c1. Now we
don't lie about the actual thread priority, it's apparent it's not
really being set in this case.)
Fixes testcase priority2.
Signed-off-by: Jon Turney <jon.turney@dronecode.org.uk>
Test access05 and symlink03 expect operations to fail which succeed when
we have Adminstrator privileges.
There's perhaps a bit of incoherency here: some XFAILed tests expect to
run as root (so maybe we need the ability to selectively cygdrop?).
Signed-off-by: Jon Turney <jon.turney@dronecode.org.uk>
Change TCIDs to they match the filename
Fix use of "%0" rather than "%o"
Record failure on mismatched permissions, rather than immediately breaking
See ltp commits fa31d55d, 923b23ff and b846e7bb
fa31d55d34923b23ff1fb846e7bb9c
Signed-off-by: Jon Turney <jon.turney@dronecode.org.uk>
full_path needs to hold a overlong pathname of length PATH_MAX+1, plus a
terminating null.
See ltp commit 44d51c3f
44d51c3f06
Signed-off-by: Jon Turney <jon.turney@dronecode.org.uk>
These tests async thread cancellation of a thread that doesn't have any
cancellation points.
Unfortunately, since 2b165a45 the async cancellation silently fails when
the thread is inside the kernel function Sleep(), so it just exits
normally after 10 seconds. (See the commentary in pthread::cancel() in
thread.cc, where it checks if the target thread is inside the kernel,
and silently converts the cancellation into a deferred one)
Work around this by busy-waiting rather than Sleep()ing for 10 seconds.
This is still somewhat fragile: the async cancel could still fail, if it
happens to occur while we're inside the kernel function that time()
calls.
v2:
Do nothing more efficiently
Signed-off-by: Jon Turney <jon.turney@dronecode.org.uk>
Since commit 4b51e4c1, we return the actual thread priority, not what we
originally stored in the thread attributes.
Windows only supports 7 thread priority levels, which we map onto the 32
required by POSIX. So, only a subset of values will be returned exactly
by by pthread_getschedparam() after pthread_setschedparam().
Adjust tests priority1, priority2 and inherit1 so they only check for
round-tripping priority values which can be exactly represented.
For CI, this needs to handle process priority class "below normal
priority" as well.
Also check that the range of priority values is at least 32, as required
by POSIX.
Signed-off-by: Jon Turney <jon.turney@dronecode.org.uk>
Check direct call to system(), as well as one in a subprocess.
(This is a lot easier to debug when it's completely broken by the
environment the test is running in)
Signed-off-by: Jon Turney <jon.turney@dronecode.org.uk>
Astonishingly, we don't have this already, so tests which hang just stop
the testsuite dead in it's tracks...
Signed-off-by: Jon Turney <jon.turney@dronecode.org.uk>
Do some setup in the Cygwin 'installation' at testsuite/testinst/:
* Ensure /tmp exists
* Use BusyBox to provide executables needed by tests which use system()
(sh, sleep, ls)
This enables tests which use system(), or require /tmp to exist to pass.
Signed-off-by: Jon Turney <jon.turney@dronecode.org.uk>
The current version of <sys/cpuset.h> cannot be compiled by Clang due to
the use of builtin versions of malloc, free, and memset. Their presence
here was a dubious optimization anyway, so their usage has been
converted to standard library functions.
The use of __builtin_popcountl remains because Clang implements it just
like gcc does. If/when some other compiler (Rust? Go?) runs into this
issue we can deal with specialized handling then.
The "#include <sys/cdefs>" here to define __inline can be removed since
both of the new includes sub-include it.
Addresses: https://cygwin.com/pipermail/cygwin/2023-July/253927.html
Fixes: 9cc910dd33a5 (Cygwin: Make <sys/cpuset.h> safe for c89 compilations)
Signed-off-by: Mark Geisert <mark@maxrnd.com>
While st_rdev returned by fstat() for /dev/tty should be FH_TTY,
the current cygwin1.dll returns FH_PTYS+minor or FH_CONS+minor.
Similarly, fstat() does not return correct value for /dev/console,
/dev/conout, /dev/conin or /dev/ptmx.
This patch fixes the issue by:
1) Introduce dev_referred_via in fhandler_termios.
2) Add new argument, which has dev_t value referred by open(),
for constructors of fhandler_pty_slave and fhandler_pty_master to
set the value of dev_referred_via.
3) Set st_rdev using dev_referred_via in fhandler_termios::fstat()
if it is available.
Reviewed-by: Corinna Vinschen <corinna@vinschen.de>
Signed-off-by: Takashi Yano <takashi.yano@nifty.ne.jp>
As reported in
https://cygwin.com/pipermail/cygwin/2023-June/253888.html,
"Bad address" error occurs when stat() is called after the commit
3721a756b0 ("Cygwin: console: Make the console accessible from
other terminals.").
There are two problems in the current code. One is fhandler_console::
fstat() calls get_ttyp()->getsid(). However, fh_alloc() in dtable.cc
omits to initialize the fhandler_console instance when stat() is
called. Due to this, get_ttyp() returns NULL and access violation
occurs. The other problem is fh_alloc() assigns fhandler_console
even if the CTTY is not a console. So the first problem above occurs
even if the CTTY is a pty.
This patch fixes the issue by:
1) Call set_unit() to initialize _tc if the get_ttyp() returns NULL.
2) Assign fhandler_pty_slave for /dev/tty if CTTY is a pty in fh_alloc().
Fixes: 3721a756b0 ("Cygwin: console: Make the console accessible
from other terminals.").
Fixes: 23771fa1f7 ("dtable.cc (fh_alloc): Make different decisions
when generating fhandler for not-opened devices. Add kludge to deal
with opening /dev/tty.")
Reported-by: Bruce Jerrick <bmj001@gmail.com>
Reviewed-by: Corinna Vinschen <corinna@vinschen.de>
Signed-off-by: Takashi Yano <takashi.yano@nifty.ne.jp>
Four modifications to include/sys/cpuset.h:
* Change C++-style comments to C-style also supported by C++
* Change "inline" to "__inline" on code lines
* Add "#include <sys/cdefs.h>" to make sure __inline is defined
* Don't declare loop variables on for-loop init clauses
Tested by first reproducing the reported issue with home-grown test
programs by compiling with gcc option "-std=c89", then compiling again
using the modified <sys/cpuset.h>. Other "-std=" options tested too.
Addresses: https://cygwin.com/pipermail/cygwin-patches/2023q3/012308.html
Fixes: 315e5fbd99ec ("Cygwin: Fix type mismatch on sys/cpuset.h")
Signed-off-by: Mark Geisert <mark@maxrnd.com>
Currently, _my_tls.tid is set to pthread_null if pthread::self()
is called before pthread::init_mainthread(). As a result, pthread::
init_mainthread() does not set _my_tls.tid appropriately. Due to
this, pthread_join() fails in LDAP environment if the program is
the first program which loads cygwin1.dll.
https://cygwin.com/pipermail/cygwin/2023-June/253792.html
With this patch, _my_tls.tid is re-initialized in pthread::
init_mainthread() if it is pthread_null.
Reported-by: Mümin A. <muminaydin06@gmail.com>
Reviewed-by: Corinna Vinschen <corinna@vinschen.de>
Signed-off-by: Takashi Yano <takashi.yano@nifty.ne.jp>
For example, for GCC/GCN target (AMD GPUs), target libraries are built
individually per supported hardware ISA ('-march=[...]'). Testing such a
toolchain via, for example:
$ make RUNTESTFLAGS='--target_board=[...]/-march=gfx90a' check[...]
... does work fine for all 'check-gcc-[...]' as well as GCC-provided target
libraries, 'check-target-[...]'. Just for 'check-target-newlib', for the
example above, not the '-march=gfx90a' newlib libraries are linked in, but
instead always the default ones, which results in link FAILure. This is cured
simply by skipping use of 'newlib/testsuite/lib/flags.exp', so that the
standard search paths as determined by GCC, DejaGnu are used for newlib, too.
In test for AMD/Intel Control flow Enforcement Technology user mode
shadow stack support replace Windows version tests with test of wincap
member addition has_user_shstk with Windows version dependent value
Fixes: 41fdb869f9 ("fhandler/proc.cc(format_proc_cpuinfo): Add Linux 6.3 cpuinfo")
Signed-off-by: Brian Inglis <Brian.Inglis@Shaw.ca>
In the very early code path where `dll_crt0_1 ()` calls
`user_shared->initialize ()`, the Cygwin runtime calls `internal_pwsid ()`
to initialize the user name in preparation for reading the `fstab` file.
In case `db_home: env` is defined in `/etc/nsswitch.conf`, we need to
look at the environment variable `HOME` and use it, if set.
When all of this happens, though, the `pinfo_init ()` function has had no
chance to run yet (and therefore, `environ_init ()`). At this stage,
therefore, `getenv ()`'s `findenv_func ()` call still finds `getearly ()`
and we get the _verbatim_ value of `HOME`. That is, the Windows form.
But we need the "POSIX" form.
To add insult to injury, later calls to `getpwuid (getuid ())` will
receive a cached version of the home directory via
`cygheap->pg.pwd_cache.win.find_user ()` thanks to the first
`internal_pwsid ()` call caching the result via
`add_user_from_cygserver ()`, read: we will never receive the converted
`HOME` but always the Windows variant.
So, contrary to the assumptions made in 27376c60a9 (Allow deriving the
current user's home directory via the HOME variable, 2023-03-28), we
cannot assume that `getenv ("HOME")` returned a "POSIX" path.
This is a real problem. Even setting aside that common callers of
`getpwuid ()` (such as OpenSSH) are unable to handle Windows paths in the
`pw_dir` attribute, the Windows path never makes it back to the caller
unscathed. The value returned from `fetch_home_env ()` is not actually
used as-is. Instead, the `fetch_account_from_windows ()` method uses it
to write a pseudo `/etc/passwd`-formatted line that is _then_ parsed via
the `pwdgrp::parse_passwd ()` method which sees no problem with
misinterpreting the colon after the drive letter as a field separator of
that `/etc/passwd`-formatted line, and instead of a Windows path, we now
have a mere drive letter.
Let's detect when the `HOME` value is still in Windows format in
`fetch_home_env ()`, and convert it in that case.
For good measure, interpret this "Windows format" not only to include
absolute paths with drive prefixes, but also UNC paths.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
The account under which Azure Web Apps run is an IIS APPOOL account that
is generated on the fly.
These are special because the virtual machines on which thes Apps run
are not domain-joined, yet the accounts are domain accounts.
To support the use case where such a Web App needs to call `ssh` (e.g.
to deploy from a Git repository that is accessible only via SSH), we do
need OpenSSH's `getpwuid (getuid ())` invocation to work.
But currently it does not. Concretely, `getuid ()` returns -1 for these
accounts, and OpenSSH fails to find the correct home directory
(_especially_ when that home directory was overridden via a `db_home:
env` line in `/etc/nsswitch.conf`).
This can be verified e.g. in a Kudu console (for details about Kudu
consoles, see https://github.com/projectkudu/kudu/wiki/Kudu-console):
the domain is `IIS APPPOOL`, the account name is the name of the Azure
Web App, the SID starts with 'S-1-5-82-`, and
`pwdgrp::fetch_account_from_windows()` runs into the code path where
"[...] the domain returned by LookupAccountSid is not our machine name,
and if our machine is no domain member, we lose. We have nobody to ask
for the POSIX offset."
Since these IIS APPPOOL accounts are relatively similar to AzureAD
accounts in this scenario, let's imitate the latter to support also the
former.
Reported-by: David Ebbo <david.ebbo@gmail.com>
Helped-by: Corinna Vinschen <corinna@vinschen.de>
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
The commit 9fc746d17d does not fix transferring input at exit
appropriately. If the more than one non-cygwin apps are executed
simultaneously and one of them is terminated, the pty master failed
to send input to the other non-cygwin apps. This patch fixes that.
Fixes: 9fc746d17d ("Cygwin: pty: Fix transferring type-ahead input between input pipes.")
Signed-off-by: Takashi Yano <takashi.yano@nifty.ne.jp>
After the commit e5fcc5837c, transferring type-ahead input between
the pipe for cygwin app and the pipe for non-cygwin app will not be
done appropriately when the stdin of the non-cygwin app is not pty.
Due to this issue, sometimes the keyboard input might be lost which
should be sent to cygwin app. This patch fixes the issue.
Fixes: e5fcc5837c ("Cygwin: pty: Fix reading CONIN$ when stdin is not a pty.")
Signed-off-by: Takashi Yano <takashi.yano@nifty.ne.jp>
Disable at least m68010 and m68020. These processors certainly do not
like unaligned accesses.
Signed-off-by: Remy Bohmer <linux@bohmer.net>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>
Signed-off-by: Austin Morton <austinpmorton@gmail.com>
Signed-off-by: Chris Packham <judge.packham@gmail.com>
Corinna Vinschen
Jon Turney
Jeff Johnston
Jordi Sanfeliu
Alexey Lapshin
Mark Geisert
Takashi Yano
Thomas Schwinge
Mingye Wang
Brian Inglis
Philippe Cerfon
Johannes Schindelin
Chris Packham
Remy Bohmer