Up to Vista CreateProcessAsUser only worked with primary tokens,
so convert S4U impersonation token to primary token. MSDN still
documents it that way, but actually an impersonation token is
sufficient since Windows 7.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
Under WOW64 on 64 bit Windows 7, MsV1_0S4ULogon appears to be
unimplemented, probably under Vista as well. Re-enable
create_token method, to allow basic seteuid on W7 WOW64 and
Vista as well.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
for domain accounts we try KerbS4ULogon first, MsV1_0S4ULogon
second. But we only fetch the package id for the supporting
authentication package (Kerberos/MsV1_0) once at the start.
Duplicate LsaLookupAuthenticationPackage call and move into the
Kerb/MsV1_0 branches so that it fetches the correct package id
for the method we call next.
Curious enough this worked before. Apparently both methods
work with the MICROSOFT_KERBEROS_NAME_A package id. However,
requesting and using the right authentication package id is
the prudent thing to do.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
This reverts commit 7c34811440.
This potentially allows to circumvent OpenSSHs user/group name matching,
unless the Admin knows to add every local user twice or to use patterns,
e.g.:
Match user MACHINE+user,.+user
Match user *+user
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
Commit c1023ee353 changed the way
path_conv::binmode() works. Rather than returning three states,
O_BINARY, O_TEXT, 0, it only returned 2 states, O_BINARY, O_TEXT. Since
mounts are only binary if they are explicitely mounted binary by setting
the MOUNT_BINARY flag, textmode is default.
This introduced a new bug. When inheriting stdio HANDLEs from native
Windows processes, the fhandler and its path_conv are created from a
device struct only. None of the path or mount flags get set this way.
So the mount flags are 0 and path_conv::binmode() returned 0.
After the path_conv::binmode() change it returned O_TEXT since, as
explained above, the default mount mode is textmode.
Rather than just enforcing binary mode for path_conv's created from
device structs, this patch changes the default mount mode to binary:
Replace MOUNT_BINARY flag with MOUNT_TEXT flag with opposite meaning.
Drop all explicit setting of MOUNT_BINARY. Drop local set_flags
function, it doesn't add any value.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
When recognizing a negative pid, optind is off by one. The
code correcting this has been erroneously removed by commit
8de660271f. Revert that.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
The solution from commit 9a3cc77b2a
didn't work for foreign domain accounts. Rather than calling
LookupAccountSid we now use the info when we fetch it anyway
via LDAP or Net*GetInfo. Only in case of domain groups we have
to add an LDAP call explicitly.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
Make sure a domain+username fits into the local name buffer.
The former buffer size didn't take adding a domain name to
a really_really_long_user_name into account.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
When looking up valid accounts by name, LookupAccountName returns
a SID and a case-correct domain name. However, the name was input
and LookupAccountName is case-insensitive, so the name is not
necessarily written the same way as in SAM or AD.
Fix that by doing a reverse lookup on the just fetched SID. This
fetches the account name in the correct case. Override the
incoming name with the case correct name from LookupAccountSid.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
This is a really old and crappy API, as the previous commit shows.
Use NtQueryInformationFile(FilePositionInformation) here instead.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
dtable::set_file_pointers_for_exec is called from
child_info_spawn::worker to move the file position of O_APPEND
files to EOF if the child is a native child.
However, this only works correctly for the first O_APPEND
file descriptor:
- set_file_pointers_for_exec calls SetFilePointer. The higher
4 bytes of the desired file offset are given to SetFilePointer
as pointer to a DWORD value. On return, SetFilePointer returns
the higher 4 bytes of the new file position in this DWORD.
- So for the second and subsequent descriptors the higher 4 byte
of the file position depend on what the actual file position
of the previous file has been set to:
- If the file is > 2 Gigs, the high offset will not be 0 anymore.
- If the desciptor points to a non-seekable file (i.e., a pipe
or socket), SetFilePosition returns an error and sets the high
position to -1.
Fix this by calling SetFilePointerEx instead, which does not
modify the incoming position value.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
When calling spawnve, in contrast to execve, the parent has
to create the pid for the child. With the old technique
this was simply the Windows pid, but now we have to inform
the child about its new pid.
Add a cygpid member to class child_info_spawn. Set it in
child_info_spawn::worker, just prior to calling CreateProcess
rather than afterwards. Overwrite cygheap->pid in
child_info_spawn::handle_spawn before calling pinfo::thisproc.
Make sure pinfo::thisproc knows the pid is already set by
setting the handle argument to INVALID_HANDLE_VALUE.
Also set procinfo->dwProcessId to myself_initial.dwProcessId
instead of to myself_initial.pid for clarity.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
When the current process has renamed (to bin) a readonly dll, we get
STATUS_TRANSACTION_NOT_ACTIVE for unknown reason when subsequently
creating the forkable hardlink. A workaround is to open the original
file with FILE_WRITE_ATTRIBUTES access, but that fails with permission
denied for users not owning the original file.
* forkable.cc (dll::create_forkable): Retry hardlink creation using the
original file's handle opened with FILE_WRITE_ATTRIBUTES access when the
first attempt fails with STATUS_TRANSACTION_NOT_ACTIVE.
* Rename cygwin_shared->prefer_forkable_hardlinks to
forkable_hardlink_support, with values
0 for Unknown, 1 for Supported, -1 for Unsupported.
Upon first dll loaded ever, dll_list::forkable_ntnamesize checks the
/var/run/cygfork directory to both exist and reside on NTFS, setting
cygwin_shared->forkable_hardlink_support accordingly.
* Replace enum forkables_needs by bool forkables_created: Set
to True by request_forkables after creating forkable hardlinks.
To avoid the need for each process to check the filesystem to detect
that hardlink creation is impossible or disabled, cache this fact in
shared memory. Removing cygfork directory while in use does disable
hardlinks creation. To (re-)enable hardlinks creation, the cygfork
directory has to exist before the first cygwin process does fork.
* forkable.cc (dll_list::forkable_ntnamesize): Short cut
forkables needs to impossible when disabled via shared memory.
(dll_list::update_forkables_needs): When detecting hardlink
creation as impossible (not on NTFS) while still (we are the
first one checking) enabled via shared memory, disable the
shared memory value.
(dll_list::request_forkables): Disable the shared memory value
when hardlinks creation became disabled, that is when the
cygfork directory was removed.
* include/cygwin/version.h: Bump CYGWIN_VERSION_SHARED_DATA 6.
* shared_info.h (struct shared_info): Add member
prefer_forkable_hardlinks. Update CURR_SHARED_MAGIC.
* shared.cc (shared_info::initialize): Initialize
prefer_forkable_hardlinks to 1 (Yes).
To support in-cygwin package managers, the fork() implementation must
not rely on .exe and .dll files to stay in their original location, as
the package manager's job is to replace these files. Instead, when the
first fork try fails, and we have NTFS, we use hardlinks to the original
binaries in /var/run/cygfork/ to create the child process during the
second fork try, along the main.exe.local file to enable the "DotLocal
Dll Redirection" feature for the dlls.
The (probably few) users that need an update-safe fork manually have to
create the /var/run/cygfork/ directory for now, using:
mkdir --mode=a=rwxt /var/run/cygfork
* child_info.h: Bump CURR_CHILD_INFO_MAGIC.
(enum child_status): Add _CI_SILENTFAIL flag.
(struct child_info): Add silentfail setter and getter.
* winsup.h (child_copy): Add bool silentfail parameter.
* cygheap.cc: Pass silentfail parameter to child_copy.
* dcrt0.cc: Ditto.
* dll_init.h (struct dll): Define public inline method forkedntname.
(struct dll_list): Declare private method find_by_forkedntname.
* dll_init.cc (struct dll_list): Implement find_by_forkedntname.
(dll_list::alloc): Use find_by_forkedntname when in load after fork.
(dll_list::load_after_fork_impl): Load dlls using dll::forkedntname.
* fork.cc (frok::parent): Set silentfail child info flag. Pass
silentfail parameter to child_copy. Use forkedntname of
dlls.main_executable.
(fork): When first dofork run failed and did not use forkables,
run dofork again with_forkables set to true.
(child_copy): Use debug_printf if silentfail is true,
system_printf otherwise.
In preparation to protect fork() against dll- and exe-updates, create
hardlinks to the main executable and each loaded dll in subdirectories
of /var/run/cygfork/, if that one exists on the NTFS file system.
The directory names consist of the user sid, the main executable's NTFS
IndexNumber, and the most recent LastWriteTime of all involved binaries
(dlls and main executable). Next to the main.exe hardlink we create the
empty file main.exe.local to enable dll redirection.
The name of the mutex to synchronize hardlink creation/cleanup also is
assembled from these directory names, to allow for synchronized cleanup
of even orphaned hardlink directories.
The hardlink to each dynamically loaded dll goes into another directory,
named using the NTFS IndexNumber of the dll's original directory.
* Makefile.in (DLL_OFILES): Add forkable.o.
* dll_init.h (struct dll): Declare member variables fbi, fii,
forkable_ntname. Declare methods nominate_forkable,
create_forkable.
(struct dll_list): Declare enum forkables_needs. Declare member
variables forkables_dirx_size, forkables_dirx_ntname,
forkables_mutex_name, forkables_mutex. Declare private methods
forkable_ntnamesize, prepare_forkables_nomination,
update_forkables_needs, update_forkables, create_forkables,
denominate_forkables, close_mutex, try_remove_forkables,
set_forkables_inheritance, request_forkables. Declare public
static methods ntopenfile, read_fii, read_fbi. Declare public
methods release_forkables, cleanup_forkables. Define public
inline method setup_forkables.
* dll_init.cc (dll_list::alloc): Allocate memory to hold the
name of the hardlink in struct dll member forkable_ntname.
Initialize struct dll members fbi, fii.
(dll_list::load_after_fork): Call release_forkables method.
* fork.cc: Rename public fork function to static dofork, add
with_forkables as bool pointer parameter. Add new fork function
calling dofork. (struct frok): Add bool pointer member
with_forkables, add as constructor parameter.
(frok::parent): Call dlls.setup_forkables before CreateProcessW,
dlls.release_forkables afterwards.
* pinfo.cc (pinfo::exit): Call dlls.cleanup_forkables.
* syscalls.cc (_unlink_nt): Rename public unlink_nt function to
static _unlink_nt, with 'shareable' as additional argument.
(unlink_nt): New, wrap _unlink_nt for original behaviour.
(unlink_nt_shareable): New, wrap _unlink_nt to keep a binary
file still loadable while removing one of its hardlinks.
* forkable.cc: New file.
Implement static functions mkdirs, rmdirs, rmdirs_synchronized,
stat_real_file_once, format_IndexNumber, rootname, sidname,
exename, lwtimename. Define static array forkable_nameparts.
(struct dll): Implement nominate_forkable, create_forkable.
(struct dll_list): Implement static methods ntopenfile,
read_fii, read_fbi. Implement forkable_ntnamesize,
Even for the main executable and cygwin1.dll store the file name as full
NT path. Create the child process using the main executable's file name
converted from the full NT path stored before.
* dll_init.cc (dll_list::alloc): Search for DLL_SELF type entry
with module name like for DLL_LINK, use full NT path to search
for DLL_LOAD type only. For DLL_SELF type do not indicate
having a destructor to be called.
(dll_list::find): Ignore DLL_SELF type entries.
(dll_list::init): Ditto. Call track_self method.
(dll_list::track_self): New.
(dll_list::load_after_fork): Call track_self method.
* dll_init.h (enum dll_type): Add DLL_SELF, for the main
executable and cygwin1.dll.
(struct dll_list): Declare private method track_self. Declare
member variable main_executable.
* fork.cc (frok::parent): Use ntname from dlls.main_executable
to create child process, converted to short path using
dll_list::buffered_shortname.
Store loaded dll's file name as full NT path.
* dll_init.h (struct dll): Rename member variable name to ntname.
(struct dll_list): Declare private static member variable
nt_max_path_buffer. Declare public static methods form_ntname,
form_shortname. Define public static methods nt_max_path_buf,
buffered_shortname.
(dll_list::operator []): Use PCWCHAR rather than const PWCHAR.
(dll_list::find_by_modname): Ditto.
* dll_init.cc (in_load_after_fork): Define earlier in file.
(struct dll_list): Rename member variable name to ntname.
Define nt_max_path_buffer variable.
Implement static methods form_ntname, form_shortname.
(dll_list::operator []): Use PCWCHAR rather than const PWCHAR.
(dll_list::find_by_modname): Ditto.
(reserve_at): Ditto.
(release_at): Ditto.
(dll_list::alloc): Use nt_max_path_buf method instead of local
buffer. Store module file name as full NT path, convert using
the form_ntname static method.
(dll_list::load_after_fork): Call load_after_fork_impl only when
reload_on_fork is set.
* fork.cc (frok::child): Call dlls.load_after_fork even without
need to dynamically load dlls.
(frok::parent): Move syscall_printf into the retry loop.
Trying to delete in-use executables and DLLs using
FILE_DISPOSITION_POSIX_SEMANTICS returns STATUS_CANNOT_DELETE.
Fall back to the former method if that error occurs to allow
unlinking these files.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
The symlink target of /proc/PID/fd files pointing to pipes and
sockets are just artificial filenames referencing the object using
some internal number. The pipe open code expects a path specifying
process pid and the internal number so it access the right process
and pipe.
- Set the posix path of the pipe to the simple pipe name only,
as it shows up in /proc/PID/fd. A /proc/self prefix is just
as wrong as a /dev/fd prefix.
- Revert thinko in fhandler_pipe::open expecting the name as
/proc/self/fd/... In fact this should never happen.
- Fix up the path before re-opening the pipe instead.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
The reopen code neglected to pass along the requested open
mode correctly. This may end up reopening the file with
incorrect access mask, or duplicating the wrong pipe handle.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
Move TerminateProcess call into cleanup code to make sure child
doesn't linger in some border cases.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
When fork finally fails although both CreateProcess and creating the
"cygpid.N" shared memory section succeeded, we have to release that
shared memory section as well - before releasing the process handle.
Otherways we leave an orphan "cygpid.N" shared memory section, and any
subsequent cygwin process receiving the same PID fails to initialize.
* fork.cc (frok::parent): Call child.allow_remove in cleanup code.
- If the execve'ed process is a non-Cygwin process, we have to
create the matching winpid symlink and remove the old one
ourselves.
- If we spawn a child, the winpid symlink has to be maintained
by the child process, otherwise it disappears if the parent
process exits.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
The arguments are not used anyway, so drop them. When called,
procinfo->dwProcessId is already set right, so we don't have
to access myself_initial.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
This may end up killing the wrong process. Only allow Cygwin PID.
Slightly clean up code: Remove outdated W95 considerations. Fix
a bug in commandline argument processing.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
Since commit b5e1003722, native
Windows processes not started by Cygwin processes don't have a
Cygwin PID anymore. This breaks ps -W and kill -f <WINPID>.
Introduce MAX_PID (65536 for now).
Cygwin processes as well as native Windows processes started
from a Cygwin process get a PID < MAX_PID. All other native
Windows processes get a faked Cygwin PID >= MAX_PID.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
Using the Windows PID as Cygwin PID has a few drawbacks:
- the PIDs on Windows get reused quickly. Some POSIX applications choke
on that, so we need extra code to avoid too quick PID reuse.
- The code to avoid PID reuse keeps parent process handles and
(depending on a build option) child processes open unnecessarily.
- After an execve, the process has a split personality: Its Windows PID
is a new PID, while its Cygwin PID is the PID of the execve caller
process. This requires to keep two procinfo shared sections open, the
second just to redirect process info requests to the first, correct
one.
This patch changes the way Cygwin PIDs are generated:
- Cygwin PIDs are generated independently of the Windows PID, in a way
expected by POSIX processes. The PIDs are created incrementally in
the range between 2 and 65535, round-robin.
- On startup of the first Cygwin process, choose a semi-random start PID
for the first process in the lower PID range to make the PIDs slightly
unpredictable. This may not be necessary but it seems kind of inviting
to know that the first Cygwin process always starts with PID 2.
- Every process not only creates the shared procinfo section, but also a
symlink in the NT namespace, symlinking the Windows PID to the Cygwin
PID. This drops the need for the extra procinfo section after execve.
- Don't keep other process handles around unnecessarily.
- Simplify the code creating/opening the shared procinfo section and
make a clear distinction between interfaces getting a Cygwin PID and
interfaces getting a Windows PID.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
If a process is just exiting, requesting memory info may fail
with STATUS_PROCESS_IS_TERMINATING. Right now the code just bails
out if fetching mem info fails. However, the rest of the info is
still valuable for procps, so just carry on.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
After creating a pthread, the stack gets moved to the desired memory
location. While the 32 bit thread wrapper copies the exception handler
information to the new stack (so we have at least *some* exception
handler present), the x86_64 code didn't perform any exception handler
magic. Fix that.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
VirtualQueryEx, called by fixup_mmaps_after_fork, requires
PROCESS_QUERY_INFORMATION permissions per MSDN. However, testing
shows that PROCESS_QUERY_LIMITED_INFORMATION is sufficient when
running the same code on Windows 8.1 or Windows 10. Fix the code
to give the forked child always PROCESS_QUERY_INFORMATION perms
on Windows Vista/7 and respective server releases.
Revert now unneeded patch to check_token_membership as well.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
POSIX requires that raise(3) is equivalent to
pthread_kill(pthread_self(), sig);
in multi-threaded applications. Our raise just called kill(sig).
Fix that.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
Keeping an inheritable handle open results in that handle being
spilled over into grandchild processes, which is not desired.
Duplicate original parent handle into a non-inheritable one with
minimal SYNCHRONIZE permissions and close the original handle.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
- Exec'ed/spawned processes don't need PROCESS_DUP_HANDLE. Remove that
permission from the parent handle.
- PROCESS_QUERY_LIMITED_INFORMATION doesn't work for Windows 7 if the
process is started as a service. Add PROCESS_QUERY_INFORMATION for
pre-Windows 8 in that case.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
The version info only depends on the object files. This results
in the version info not being rebuild immediately if a source
file is changed. Rather, the version info is only rebuilt on the
next make run.
Fix that by making the version info build rule dependent on the
source files.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
- This simple and official method replaces cyglsa and "create token"
methods. No network share access, same as before.
- lsaauth and create_token are disabled now. If problems crop up,
they can be easily reactivated. If no problems crop up, they
can be removed in a while, together with the lsaauth subdir.
- Bump Cygwin version to 3.0.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
The previous patch failed with password-less auth because in
that case the return code from get_server_groups wasn't tested.
Fix that. Also make sure that get_server_groups does not
check if the account is disabled or locked out when just fetching
the group list for initgroups or getgrouplist.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
So far seteuid could change uid to any existing account, given
sufficient permissions of the caller. This is kind of bad since
it disallows admins to refuse login to disabled or locked out
accounts.
Add check for the account's UF_ACCOUNTDISABLE or UF_LOCKOUT flags
and don't let the user in, if one of the flags is set.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
Use info from same source (GetNetworkParams).
Also move getdomainname near gethostname in source.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
If gethostname() fails we call GetComputerNameEx with
ComputerNameDnsFullyQualified. This is wrong, gethostname should return
the hostname only, not the FQDN. Fix this by calling GetComputerNameEx
with ComputerNameDnsHostname.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
When reusing a cygthread, the stub method fails to set the thread name
to the new name. The name is only set when actually creating the
thread. Fix that by moving the SetThreadName call right in front of the
thread function call.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
Combine with a bit of cleanup:
- Drop overrun_event_running in favor of overrun_count being -1.
- Fix include guard in posix_timer.h.
- Drop ununsed function timespec_to_us.
- Don't use Interlocked functions without need.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
Allocating on the cygheap would copy information of the tracker into
the child process. A forked child knows the timer id and could simply
still access the (free'd but still valid) timer_tracker on the heap,
which is dangerous and very certainly doesn't reflect POSIX semantics.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
- Rename files timer.* to posix_timer.*.
- Reimplement using an OS timer rather than a handcrafted wait loop.
- Use a Slim R/W Lock for synchronization.
- Drop timer chaining. It doesn't server a purpose since all timers
are local only.
- Rename ttstart to itimer_tracker to better reflect its purpose.
It's not the anchor for a timer chain anymore anyway.
- Drop fixup_timers_after_fork. Everything is process-local, nothing
gets inherited.
- Rename timer_tracker::disarm_event to disarm_overrun_event for
better readability.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
- When correcting the next expiration timestamp, the number of
expirations gets computed correctly, just the expiration timestamp
itself is then only incremented by a single interval, rather than
the just computed expired intervals. Fix that.
- drop the local clock variable in timerfd_tracker::create. It doesn't
serve any purpose.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
- split into to __try/__except blocks to make sure
leave_critical_section is always called when required.
- Actually fill time_spec in settime so it_interval is returned
correctly.
- Return all 0 if timer is disarmed.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
The value returned by reading from a timerfd is not an overrun
count in the same sense as for posix timers, it's an expiry counter.
Reflect that in the name.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
- Drop erroneous initial computation of overrun count in settime
for absolute non-realtime clocks. It's repeated in thread_func
and thus counted twice.
- Fix overrun computation for timestamp offsets being a multiple of
the timer interval. The timestamp has to be corrected after the
first offset, otherwise the correction loop counts the intervals
again.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
Non-CLOCK_REALTIME counters always use a relative DueTime in NtSetTimer.
However, relative DueTime has to be negative, but the code
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
- On systems with inexact realtime clock, the current timestamp may
be fractionally smaller than the desired timestamp. This breaks the
computation for incrementing overrun_count so overrun_count may end
up as 0. Expiring the timer with an overrun_count of 0 is a no-go.
Make sure we always increment overrun_count by at least one after
timer expiry.
- Do not expire the timer when another process deletes its timer_tracker.
This, too, may result in a 0 overrun_count.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
Using posix timers "timer_tracker" as base class for timerfd was flawed.
Posix timers are not inherited by child processes and don't survive
execve. The method used by posix timers didn't allow to share timers
between processes. The timers were still per-process timers and worked
entirely separate from each other. Reading from these timers via
different descriptors was only synchronized within the same process.
This does not reflect the timerfd semantics in Linux: The per-file
timers can be dup'ed and survive fork and execve. They are still just
descriptors pointing to the same timer object originally created by
timerfd_create. Synchronization is performed between all descriptor
instances of the same timer, system-wide.
Thus, reimplement timerfd using a timer instance in shared memory,
a kernel timer, and a handful of sync objects.
Every process maintains a per-process timerfd struct on the cygheap
maintaining a per-process thread. Every process sharing the same
timerfd will run this thread checking the state of the timer, similar
to the posix timer thread, just working on the shared objects and
synchronizing its job with each other thread.
Drop the timerfd implementation in the posix timer code and move the
public API to fhandler_timerfd.c. Keep the ttstart timer_tracker
anchor out of "NO_COPY" since the fixup_after_fork code should run to
avoid memory leakage.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
Linux returns EINVAL, "fd is attached to an object which is unsuitable
for writing". If we don't handle write locally, write returns EBADF.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
- Puzzeling: Commit ec98d19a08
changed ttstart to NO_COPY but kept all the code to handle
fixup after fork. Revert to not-NO_COPY and make timerfd
fork work.
- On fixup_after_fork, keep timerfd timers and restart thread
if they were armed in the parent.
- Move timerfd timer_trackers to cygheap. Overload timer_tracker
new and delete methods to handle timers accordingly. This is not
exactly required for fork, but exec will be grateful.
- Give up on TFD_TIMER_CANCEL_ON_SET for now. There's no easy way
to recognize a discontinuous change in a clock.
- Be paranoid when cleaning out ttstart.
- Fix some minor issues.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
- Drop initial overrun computation from timer_tracker::settimer.
It's performed in timer_tracker::thread_func anyway.
- Fix regression in returning correct overrun count narrowed down to
int from timer_getoverrun. This has been introduced by changing
overrun_count_curr to LONG64.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
First cut of a timerfd implementation.
Still TODO:
- fork/exec semantics
- timerfd_settime TFD_TIMER_CANCEL_ON_SET flag
- ioctl(TFD_IOC_SET_TICKS)
- bug fixes
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
In case sigwait_common returns EINTR, read wrongly ignores it,
so read can't be interrupt by a signal. Fix that.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
While FileRenameInformationEx is defined starting with Windows
10 1709 per MSDN, it only starts working in W10 1809, apparently.
Users of 1803 report "Function not implemented".
Introduce wincap_10_1809 and change the version check in
wincapc::init accordingly. Split has_posix_file_info into
has_posix_unlink_semantics and has_posix_rename_semantics.
Enable the latter only starting with W10 1809.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
Allow the signal thread to recognize we're called in consequence of
select on a signalfd. If the signal is part of the wait mask, don't
call any signal handler and don't remove the signal from the queue,
so a subsequent read (or sigwaitinfo/sigtimedwait) still gets the
signal. Instead, just signal the event object at
_cygtls::signalfd_select_wait for the thread running select.
The addition of signalfd_select_wait to _cygtls unearthed the alignment
problem of the context member again. To make sure this doesn't get lost,
improve the related comment in the header file so that this (hopefully)
doesn't get lost (again).
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
First cut of a signalfd implementation.
Still TODO: Non-polling select.
This should mostly work as on Linux except for missing support
for some members of struct signalfd_siginfo, namely ssi_fd,
ssi_band (both SIGIO/SIGPOLL, not fully implemented) and ssi_trapno
(HW exception, required HW support).
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
There's a long-standing bug in select. If we have poll-only
descriptors in the fd set, select overwrites the incoming
fd sets with the polling result. If none of the fds is ready,
select has to loop again. But now the fd sets are set to all
zero and select hangs.
Fix this by utilizing the local fd sets r, w, e as storage for
the incoming fd sets and use them to initialize select_stuff.
If we have to loop, overwritung the incoming fd sets doesn't matter.
While at it, rename r, w, e to readfds_in, writefds_in, exceptfds_in.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
- use int64_t instead of long long
- make is_timer_tracker const
- improve copyright header comment
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
So far we check the recycler name all the time, and the last interation
also only managed to handle two ways to write the recycler. However,
an adventurous user might change the case of the recycler arbitrarily.
Fix this problem by keeping track of the name in a somewhat relaxed
fashion. Use camel back on drive C by default, all upper case elsewhere.
Only if the rename op fails do we fix the recycler name on the fly
when trying to create it, and it turns out it already existed.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
If the first rename fails, we reopen the rootdir for creating a subdir.
The rootdir handle can change its value at this point, but the code
doesn't take this into account. The subsequent rename then fails with
STATUS_INVALID_HANDLE. Fix this by copying the new rootdir value to
pfri->RootDirectory.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
So far we reopened the file if it was opened case sensitive to
workaround the problem that the recycler could be named in
camel back or all upper case, depending on who created it.
That's a problem for O_TMPFILE on pre-W10. As soon as the
original HANDLE gets closed, delete-on-close is converted to full
delete disposition and all useful operations on the file cease to
work (STATUS_ACCESS_DENIED or STATUS_FILE_DELETED).
To avoid that problem drop the reopen code and check for the exact
recycler filename, either $Recycle.Bin or $RECYCLE.BIN, if the file
has been opened case sensitive.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
On pre-W10 systems there's no way to reopen a file by handle if
the delete disposition is set. We try to get around with
duplicating the handle.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
Corinna Vinschen
Yaakov Selkowitz
Michael Haubenwallner
Jozef Lawrynowicz
Ken Brown